© Springer-Verlag Berlin Heidelberg 2003

An Assessment of the Potential Environmental RisksPosed by Phthalates in Soil and Sediment

Thomas F. Parkerton 1 · Charles A. Staples 2

1 ExxonMobil Biomedical Sciences Inc., Hermeslaan 2, 1831 Machelen, Belgium.E-mail: thomas.f.parkerton@exxonmobil.com

2 Assessment Technologies, Inc., 10201 Lee Highway, Suite 580, Fairfax, VA 22030, USA.E-mail: castaple@ix.netcom.com

To assess the potential environmental concerns associated with phthalate esters (PEs) in sed-iments and native- as well as sludge-amended soils a screening risk assessment was performedusing the risk quotient paradigm. Five single isomers, dimethyl, diethyl, di-n-butyl, butylben-zyl and di-2-ethylhexyl, and two commercial mixed isomers, di-isononyl and di-isodecyl, werespecifically investigated. Application of statistical extrapolation techniques to aquatic effectsdata coupled with Equilibrium Partitioning (EqP) theory were used to derive Predicted No Ef-fect Concentrations (PNECs) intended to protect terrestrial and benthic organisms from directtoxicity posed by PEs in soil or sediment. The resultant PNECs were found to be protectivewhen compared to the wealth of available soil and sediment toxicity data for these compounds.PNECs intended to protect wildlife consumers from indirect effects associated with exposurevia the terrestrial/benthic food chain were also calculated for each PE. Comparison of risk-based criteria revealed that direct toxicity to soil or sediment-dwelling organisms dictatesPNEC derivation for low molecular weight PEs while potential indirect effects on wildlife con-sumers via food chain exposure determine PNECs for higher molecular weight PEs. A com-prehensive literature review indicated extensive field monitoring data are available character-izing PE concentrations in sediments from Europe, North America and Japan. While lessexposure data were available for characterizing the soil compartment, predicted and observedconcentrations were lower than in sediments. Results of the screening risk assessment foundthat for all PEs investigated, none of the observed soil concentrations exceeded risk-based lim-its even in the case of soils that were heavily amended with sewage sludge. Similarly, no studyreported concentrations in field sediments that exceeded the PNEC for either BBP or DINP. Forthe remaining PEs, at least one study indicated a maximum sediment concentration above thePNEC. However, the number of sediment samples exceeding the PNEC was typically less than1% of the available monitoring database. It is concluded that the environmental concerns posedby soil and sediment-associated PEs are at worst, restricted to infrequent, localized hot spotsof contaminated sediment. The conservative assumptions invoked in this screening risk analy-sis and implications of this work in future regulatory decision-making are also discussed.

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318

2 Hazard Characterization . . . . . . . . . . . . . . . . . . . . . . . 319

2.1 Direct Effects on Soil and Sediment-Dwelling Organisms . . . . . 3192.1.1 Soil and Sediment Toxicity Tests . . . . . . . . . . . . . . . . . . . 3192.1.2 Extrapolation from Aquatic Toxicity Data Using EqP Theory . . . 3272.1.3 Association-Based Methods Based on Field Data . . . . . . . . . . 328

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 317–349DOI 10.1007/b11471

2.1.4 PNEC Selection for Direct Effects . . . . . . . . . . . . . . . . . . 3292.2 Indirect Effects on Wildlife via the Food Chain . . . . . . . . . . . 329

3 Exposure Characterization . . . . . . . . . . . . . . . . . . . . . . 334

3.1 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3343.1.1 Native Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3343.1.2 Sludge-Amended Soil . . . . . . . . . . . . . . . . . . . . . . . . . 3343.1.3 Field Monitoring Data for Soil . . . . . . . . . . . . . . . . . . . . 3363.2 Sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3363.2.1 Field Monitoring Data for Sediments . . . . . . . . . . . . . . . . 337

4 Risk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . 339

5 Summary and Discussion . . . . . . . . . . . . . . . . . . . . . . 342

6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345

1Introduction

During the 1950s the potential commercial benefit of phthalate esters (PEs) became increasingly recognized. Due to excellent performance as cost-effectiveplasticizers in a broad range of applications, demand for PEs burgeoned. As PEuse continued increasing concern was raised regarding the possible risks thatmade-made chemicals could pose to the environment. As a result, this class ofimportant industrial chemicals has repeatedly been the focus of environmentalresearch for several decades.

The aquatic toxicity database for phthalate esters is extensive [1]. These datahave recently been used to develop a number of species and endpoint-specificquantitative structure activity relationships (QSARs) that describe PE aquatictoxicity. Application of statistical extrapolation procedures to these data has en-abled risk-based surface water concentrations (i.e. Predicted No Effect Concen-trations or PNECs) to be developed for four commercially important PEs: di-methyl (DMP), diethyl (DEP), di-n-butyl (DBP) and butylbenzyl (BBP) phthalate[2]. To assess the potential risks that these substances pose to the aquatic en-vironment, Staples et al. [3] prepared a comprehensive compilation of historicalexposure monitoring data. Comparison of observed or predicted surface waterconcentrations to PNECs indicated environmental concentrations that were typically several orders of magnitude below risk-based environmental quality objectives.

For higher molecular weight PEs such as di-2-ethylhexyl phthalate (DEHP),no acute or chronic toxicity is evident at the water solubility limit. This lack ofaquatic toxicity hazard may be explained by the combined role of low water sol-ubility and limited bioconcentration potential due to biotransformation. Thesetwo factors prevent the accumulation of tissue residues above a critical thresh-old. Thus, aqueous exposure is not expected to result in an internal critical bodyresidue that elicits adverse effects. Consequently, surface water concentrations ofthese substances are not expected to pose a direct concern to aquatic life [2].

318 T.F. Parkerton and C.A. Staples

The above studies indicate PEs are unlikely to pose harm to aquatic biota.However, due to the range of physico-chemical properties exhibited by PEs, soiland sediments may also serve as a significant, if not predominant compartmentdictating environmental fate behavior [4]. Moreover, ingestion of contaminatedsoil or sediment by terrestrial or benthic organisms, respectively, may serve as anadditional route of exposure relative to that provided by pore water, especially forpoorly water soluble substances such as high molecular weight phthalates. Sincedegradation rates in soils and sediments typically are slower than in surface wa-ter [5], these compartments also have the potential to serve as long-term sourcesof indirect exposure via trophic transfer (e.g. via the food chain). Therefore, thepotential environmental risks posed by phthalates in soil and sediment logicallywarrant further investigation.

The objective of the present study is to provide an assessment of the directrisks posed by soil and sediment-associated phthalates on benthic and terrestrialorganisms as well as the indirect risks (via the food chain) to wildlife. In additionto the single isomer PEs reported in the surface water risk assessment by Stapleset al. [3], two additional mixed isomers, diisononyl (DINP) and diisodecyl (DIDP)phthalate are considered in the present study. These substances are included dueto commercial significance and the expected importance that soil and sedimentcompartments play in the environmental fate of these poorly water soluble PEs.The remainder of this paper is organized into the following sections. First, the di-rect and indirect hazard of soil and sediment-associated phthalates is criticallyreviewed. Based on this analysis, risk-based soil and sediment quality objectivesare derived. Environmental monitoring data obtained from field studies are com-piled to characterize soil and sediment concentrations of the selected PEs in dif-ferent regions of the world. This information is then used as the basis for riskcharacterization.A discussion of the assumptions and uncertainties in this analy-sis are also presented in the concluding section.

2Hazard Characterization

2.1Direct Effects on Soil and Sediment-Dwelling Organisms

2.1.1Soil and Sediment Toxicity Tests

A compilation of available soil and sediment toxicity test data is provided inTable 1. Results are tabulated for broad taxonomic groups representing microbe,plant, invertebrate, vertebrate and multi-species (mesocosm) tests by endpointtype (i.e. L/EC50, LOEC, NOEC). Test duration, and when available, soil organiccarbon content (or soil type) is also provided.

Available acute or short-term chronic data for DMP and DEP, while limited, arein the range of 100 to >1000 mg/kg dry. In the case of DBP, considerable toxicitydata are available across trophic levels. Acute toxicity is observed at concentra-tions similiar to DMP and DEP. Several chronic NOECs for DBP based on growth

An Assessment of the Potential Environmental Risks Posed by Phthalates 319

320 T.F. Parkerton and C.A. Staples

Tabl

e1.

Sum

mar

y of

soil/

sedi

men

t tox

icit

y te

st d

ata

for

phth

alat

e es

ters

Test

spe

cies

(com

mon

nam

e)Te

st d

urat

ion

Test

end

poin

tSo

il/se

dim

ent

EC o

r LC

50LO

EC/N

OEC

R

ef.

in d

ays

exce

pt

orga

nic

carb

on

(mg/

kg d

ry)

(mg/

kg d

ry)

whe

re s

tate

d(%

dry

)

Dim

ethy

l pht

hala

te (D

MP)

Mic

roor

gani

sms

Soil

mic

robe

s1

bact

eria

num

ber

3.8

1000

*N

R[6

]Pl

ants

Spin

acea

ole

race

a (s

pina

ch)

16se

edlin

g he

ight

NR

<10

00*

NR

[7]

Pisi

um s

ativ

um (p

eas)

14se

edlin

g he

ight

NR

ca.1

000*

NR

[7]

Soil

inve

rteb

rate

sA

llolo

boph

ora

tube

rcul

ata

14su

rviv

alO

ECD

soi

l?33

35N

R[8

]Ei

seni

a fo

etid

a14

surv

ival

OEC

D s

oil?

3160

NR

[8]

Eudr

ilus

euge

niae

14su

rviv

alO

ECD

soi

l?20

00N

R[8

]Pe

rion

yx e

xcav

atus

14su

rviv

alO

ECD

soi

l?10

64N

R[8

]

Die

thyl

pht

hala

te (D

EP)

Mic

roor

gani

sms

Soil

mic

robe

s1

bact

eria

num

ber

3.8

1000

*10

00*/

100*

[6]

Plan

tsLa

ctuc

a sa

tiva

(let

tuce

)7

shoo

t wei

ght

1.4

106

NR

[9]

Lact

uca

sati

va (l

ettu

ce)

14sh

oot w

eigh

t1.

413

4N

R[9

]Sp

inac

ea o

lera

cea

(spi

nach

)16

seed

ling

heig

htN

R>

1000

NR

[7]

Pisi

um s

ativ

um (p

eas)

14se

edlin

g he

ight

NR

>10

00N

R[7

]Se

dim

ent i

nver

tebr

ates

Chi

rono

mus

tent

ans

(mid

ge)

10su

rviv

al,g

row

th2.

45>

3100

3100

/843

[10]

Dib

utyl

pht

hala

te (D

BP)

Plan

tsZ

ea m

ays

(cor

n)21

seed

ger

min

atio

nSa

nd

>20

,000

*[1

1]Z

ea m

ays

(cor

n)21

heig

ht,s

hoot

leng

thSa

nd

2000

*/20

0*[1

1]La

ctuc

a sa

tiva

(let

tuce

)7

shoo

t wei

ght

1.4

387

NR

[9]

Lact

uca

sati

va (l

ettu

ce)

14sh

oot w

eigh

t1.

4>

1000

NR

[9]

An Assessment of the Potential Environmental Risks Posed by Phthalates 321

Tabl

e1

(con

tinu

ed)

Test

spe

cies

(com

mon

nam

e)Te

st d

urat

ion

Test

end

poin

tSo

il/se

dim

ent

EC o

r LC

50LO

EC/N

OEC

R

ef.

in d

ays

exce

pt

orga

nic

carb

on

(mg/

kg d

ry)

(mg/

kg d

ry)

whe

re s

tate

d(%

dry

)

Dib

utyl

pht

hala

te (D

BP)

Spin

acea

ole

race

a (s

pina

ch)

16se

edlin

g he

ight

NR

>10

00[7

]Pi

sium

sat

ivum

(pea

s)14

seed

ling

heig

htN

R>

1000

[7]

Soil

inve

rteb

rate

sFo

lsom

ia fi

met

aria

(spr

ingt

ails

)21

adul

t sur

viva

l<

1.5

305

33**

[12]

Fols

omia

fim

etar

ia (s

prin

gtai

ls)

21ad

ult s

urvi

val

<1.

527

734

**[1

2]Fo

lsom

ia fi

met

aria

(spr

ingt

ails

)21

adul

t rep

rodu

ctio

n<

1.5

6814

**[1

2]Fo

lsom

ia fi

met

aria

(spr

ingt

ails

)21

adul

t rep

rodu

ctio

n<

1.5

8450

**[1

2]Fo

lsom

ia fi

met

aria

(spr

ingt

ails

)42

juve

nile

sur

viva

l<

1.5

19,4

11.3

**[1

2]Fo

lsom

ia fi

met

aria

(sp

ring

tails

)42

juve

nile

gro

wth

<1.

5>

10[1

2]Fo

lsom

ia fi

met

aria

(sp

ring

tails

)42

juve

nile

dev

elop

men

t<

1.5

>10

1.0/

0.5*

*[1

2]Se

dim

ent i

nver

tebr

ates

Chi

rono

mus

tent

ans

(mid

ge)

10su

rviv

al,g

row

th2.

4582

631

5/50

[10]

Chi

rono

mus

tent

ans

(mid

ge)

10su

rviv

al,g

row

th4.

816

6430

90/4

23[1

0]C

hiro

nom

us te

ntan

s (m

idge

)10

surv

ival

,gro

wth

14.1

4730

3550

/508

[10]

Hya

lella

azt

eca

(am

phip

od)

10su

rviv

al,g

row

th2.

45>

17,4

00[1

0]H

yale

lla a

ztec

a (a

mph

ipod

)10

surv

ival

,gro

wth

4.8

>29

,500

[10]

Hya

lella

azt

eca

(am

phip

od)

10su

rviv

al,g

row

th14

.1>

71,9

00[1

0]M

ulti

-spe

cies

Sedi

men

t Mes

ocos

m8

wks

com

mun

ity

stru

ctur

eN

R10

00/1

00?

[13]

Ben

zyl b

utyl

pht

hala

te (B

BP)

Soil

inve

rteb

rate

sEi

seni

a fo

etid

a14

surv

ival

and

gro

wth

Art

ifici

al?

>10

00*

[14]

Di-

2et

hylh

exyl

pht

hala

te (D

EHP)

Mic

roor

gani

sms

Soil

mic

robe

s8

hre

spir

atio

n in

hibi

tion

NR

49,0

00/N

R[1

5]

322 T.F. Parkerton and C.A. Staples

Tabl

e1

(con

tinu

ed)

Test

spe

cies

(com

mon

nam

e)Te

st d

urat

ion

Test

end

poin

tSo

il/se

dim

ent

EC o

r LC

50LO

EC/N

OEC

R

ef.

in d

ays

exce

pt

orga

nic

carb

on

(mg/

kg d

ry)

(mg/

kg d

ry)

whe

re s

tate

d(%

dry

)

Di-

2et

hylh

exyl

pht

hala

te (D

EHP)

Soil

mic

robe

s1–

16st

ruct

ural

and

func

tion

al d

iver

sity

3.8

>10

0,00

0*[6

]So

il m

icro

bes

94re

spir

atio

n in

hibi

tion

1.8

>25

0[1

6]So

il m

icro

bes

28re

spir

atio

n in

hibi

tion

2.3

>57

3[1

7]So

il m

icro

bes

28re

spir

atio

n in

hibi

tion

5.9

>82

9[1

7]So

il m

icro

bes

60ni

trog

en m

iner

aliz

atio

n in

hibi

tion

1.8

>25

0[1

6]So

il m

icro

bes

14,2

8ni

trog

en m

iner

aliz

atio

n in

hibi

tion

1.8

>73

1[1

8]So

il m

icro

bes

14,2

8ni

trog

en m

iner

aliz

atio

n in

hibi

tion

5.9

>68

6[1

8]So

il m

icro

bes

7,28

dehy

drog

enas

e in

hibi

tion

1.8

>57

3[1

9]So

il m

icro

bes

7,28

dehy

drog

enas

e in

hibi

tion

5.9

>82

9[1

9]Se

dim

ent m

icro

bes

NR

resp

irat

ion

inhi

biti

on?

NR

>10

0[2

0]Se

dim

ent m

icro

bes

2.5

resp

irat

ion

inhi

biti

on9.

2#84

+[2

1]

Plan

tsFe

stuc

a ar

undi

nace

a (t

all f

escu

e)lif

e cy

cle

grow

th1

>14

[22]

Lact

uca

sati

va (l

ettu

ce)

life

cycl

egr

owth

1>

14[2

2]La

ctuc

a sa

tiva

(let

tuce

)7

shoo

t wei

ght

1.4

>10

00N

R[9

]La

ctuc

a sa

tiva

(let

tuce

)14

shoo

t wei

ght

1.4

>10

00N

R[9

]D

anuc

us c

arot

a L.

(car

rot)

life

cycl

egr

owth

1>

14[2

2]C

apsi

cum

ann

um L

.(ch

ili p

eppe

r)lif

e cy

cle

grow

th1

>14

[22]

Trit

icum

aes

tivu

m (w

heat

)14

germ

inat

ion,

shoo

t wei

ght

OEC

D>

100*

[23,

24]

Lepi

dium

sat

ivum

(cre

ss)

14ge

rmin

atio

n,sh

oot w

eigh

tO

ECD

>10

0*[2

3,24

]Br

assi

ca n

apas

(mus

tard

)14

germ

inat

ion,

shoo

t wei

ght

OEC

D>

100*

[23,

24]

Bras

sica

rap

a (t

urni

p)14

shoo

t wei

ght

NR

>10

00*

>10

00*

[25]

Aven

a sa

tiva

(oat

s)14

shoo

t wei

ght

NR

>10

00*

10/1

00–

1000

*x[2

5]Sp

inac

ea o

lera

cea

(spi

nach

)16

seed

ling

heig

htN

R>

1000

[7]

Pisi

um s

ativ

um (

peas

)14

seed

ling

heig

htN

R>

1000

[7]

Soil

inve

rteb

rate

s

An Assessment of the Potential Environmental Risks Posed by Phthalates 323

Tabl

e1

(con

tinu

ed)

Test

spe

cies

(com

mon

nam

e)Te

st d

urat

ion

Test

end

poin

tSo

il/se

dim

ent

EC o

r LC

50LO

EC/N

OEC

R

ef.

in d

ays

exce

pt

orga

nic

carb

on

(mg/

kg d

ry)

(mg/

kg d

ry)

whe

re s

tate

d(%

dry

)

Di-

2et

hylh

exyl

pht

hala

te (D

EHP)

Eise

nia

foet

ida

14su

rviv

alO

ECD

>10

00*

[23]

Fols

omia

fim

etar

ia (s

prin

gtai

ls)

21ad

ult s

urvi

val a

nd r

epro

duct

ion

<1.

5>

5000

[12]

Fols

omia

fim

etar

ia (s

prin

gtai

ls)

42ju

veni

le s

urvi

val,

grow

th a

nd

<1.

5>

1000

[12]

deve

lopm

ent

Sedi

men

t inv

erte

brat

esA

eshn

a (d

rago

nfly

larv

ae)

40pr

edat

ion

effic

ieny

16N

R14

68[2

6]C

hiro

nom

us te

ntan

s (m

idge

) 28

emer

genc

e,se

x ra

tio

3.6

>10

,000

[27]

Chi

rono

mus

tent

ans

(mid

ge)

10su

rviv

al,g

row

th4.

8>

3070

[10]

Hya

lella

azt

eca

(am

phip

od)

10su

rviv

al,g

row

th4.

8>

3170

[10]

Sedi

men

t ver

tebr

ates

Ran

a ar

valis

(moo

r fr

og)

30eg

g ha

tchi

ng (5

C)

8.4–

13.2

#ca

.450

+N

R[2

1]R

ana

arva

lis (m

oor

frog

)60

tado

ple

surv

ival

(5C

)8.

4–13

.2#

>26

00[2

1]R

ana

arva

lis (m

oor

frog

)14

egg

hatc

hing

(10

C)

1.2#

>20

5***

[28]

Ran

a ar

valis

(moo

r fr

og)

14eg

g ha

tchi

ng (1

0C

)9.

0#>

433*

**[2

8]R

ana

arva

lis (m

oor

frog

)14

egg

hatc

hing

(10

C)

16.8

#>

699*

**[2

8]R

ana

arva

lis (m

oor

frog

)14

egg

hatc

hing

(10

C)

30.6

#>

255*

**[2

8]R

ana

arva

lis (m

oor

frog

)29

tado

ple

surv

ival

and

gro

wth

(10

C)

1.2#

>20

5***

[28]

Ran

a ar

valis

(moo

r fr

og)

29ta

dopl

e su

rviv

al a

nd g

row

th (1

0C

)9.

0#>

433*

**[2

8]R

ana

arva

lis (m

oor

frog

)29

tado

ple

surv

ival

and

gro

wth

(10

C)

16.8

#>

699*

**[2

8]R

ana

arva

lis (m

oor

frog

)29

tado

ple

surv

ival

and

gro

wth

(10

C)

30.6

#>

255*

**[2

8]R

ana

arva

lis (m

oor

frog

)22

–25

egg

hatc

hing

,hat

chin

g ti

me

(5C

)16

>99

9[2

9]R

ana

arva

lis (m

oor

frog

)35

tadp

ole

surv

ival

,gro

wth

and

16

>99

9[2

9]de

velo

pmen

t (5

C)

Ran

a ar

valis

(moo

r fr

og)

22–

25eg

g ha

tchi

ng,h

atch

ing

tim

e (5

C)

17.3

>10

28[2

9]R

ana

arva

lis (m

oor

frog

)35

tadp

ole

surv

ival

,gro

wth

and

17

.3>

1028

[29]

deve

lopm

ent (

5C

)

324 T.F. Parkerton and C.A. Staples

Tabl

e1

(con

tinu

ed)

Test

spe

cies

(com

mon

nam

e)Te

st d

urat

ion

Test

end

poin

tSo

il/se

dim

ent

EC o

r LC

50LO

EC/N

OEC

R

ef.

in d

ays

exce

pt

orga

nic

carb

on

(mg/

kg d

ry)

(mg/

kg d

ry)

whe

re s

tate

d(%

dry

)

Di-

2et

hylh

exyl

pht

hala

te (D

EHP)

Ran

a ar

valis

(moo

r fr

og)

9–21

egg

hatc

hing

,hat

chin

g ti

me

(10

C)

16>

844

[29]

Ran

a ar

valis

(moo

r fr

og)

26ta

dpol

e su

rviv

al,g

row

th a

nd

16>

844

[29]

deve

lopm

ent (

10C

)R

ana

arva

lis (m

oor

frog

)9–

21eg

g ha

tchi

ng,h

atch

ing

tim

e (1

0C

)17

.3>

1164

[29]

Ran

a ar

valis

(moo

r fr

og)

26ta

dpol

e su

rviv

al,g

row

th a

nd

17.3

>11

64[2

9]de

velo

pmen

t (10

C)

Mul

ti-s

peci

esSe

dim

ent m

esoc

osm

30co

mm

unit

y st

ruct

ure

NR

>6.

2[3

0]

Di-

ison

onyl

pht

hala

te (D

INP)

Mic

roor

gani

sms

Soil

mic

robe

s33

inhi

biti

on o

fglu

cose

uti

lizat

ion

1.7

>96

16[3

1]Pl

ants

Lact

uca

sati

va (l

ettu

ce)

5se

ed g

erm

inat

ion

4.0#

<10

,000

[32]

Lact

uca

sati

va (l

ettu

ce)

5se

ed g

erm

inat

ion

1.7

<10

,000

[32]

Lact

uca

sati

va (l

ettu

ce)

5se

ed g

erm

inat

ion

4.0#

3000

/100

0[3

3]La

ctuc

a sa

tiva

(let

tuce

)5

seed

ger

min

atio

n1.

730

00/1

000

[33]

Lact

uca

sati

va (l

ettu

ce)

28se

ed g

erm

inat

ion,

grow

th1.

7>

1387

[34]

Loliu

m sp

.(ry

e gr

ass)

5se

ed g

erm

inat

ion

4.0#

>10

,000

[32]

Loliu

m sp

.(ry

e gr

ass)

5se

ed g

erm

inat

ion

1.7

>10

,000

[32]

Soil

inve

rteb

rate

sEi

seni

a fo

etid

a14

surv

ival

4.0#

>90

00[3

5]Ei

seni

a fo

etid

a14

surv

ival

1.7

>79

00[3

5]Se

dim

ent i

nver

tebr

ates

Chi

rono

mus

tent

ans

(mid

ge)

10su

rviv

al,g

row

th4.

8>

2680

[10]

Hya

lella

azt

eca

(am

phip

od)

10su

rviv

al,g

row

th4.

8>

2900

[10]

Sedi

men

t ver

tebr

ates

An Assessment of the Potential Environmental Risks Posed by Phthalates 325Ta

ble

1(c

onti

nued

)

Test

spe

cies

(com

mon

nam

e)Te

st d

urat

ion

Test

end

poin

tSo

il/se

dim

ent

EC o

r LC

50LO

EC/N

OEC

R

ef.

in d

ays

exce

pt

orga

nic

carb

on

(mg/

kg d

ry)

(mg/

kg d

ry)

whe

re s

tate

d(%

dry

)

Di-

ison

onyl

pht

hala

te (D

INP)

Ran

a ar

valis

(moo

r fr

og)

9–21

egg

hatc

hing

,hat

chin

g ti

me

(10

C)

16>

707

[29]

Ran

a ar

valis

(moo

r fr

og)

26ta

dpol

e su

rviv

al,g

row

th a

nd

16>

707

[ 29]

deve

lopm

ent (

10C

)R

ana

arva

lis (m

oor

frog

)9–

21eg

g ha

tchi

ng,h

atch

ing

tim

e (1

0C

)17

.3>

1009

[29]

Ran

a ar

valis

(moo

r fr

og)

26ta

dpol

e su

rviv

al,g

row

th a

nd

17.3

>10

09[2

9]de

velo

pmen

t (10

C)

Di-

isod

ecyl

pht

hala

te (D

IDP)

Plan

tsLa

ctuc

a sa

tiva

(let

tuce

)5

seed

ger

min

atio

n4.

0#>

10,0

00[3

2]La

ctuc

a sa

tiva

(let

tuce

)5

seed

ger

min

atio

n1.

7>

10,0

00[3

2]Lo

lium

sp.(

rye

gras

s)5

seed

ger

min

atio

n4.

0#>

10,0

00[3

2]Lo

lium

sp.(

rye

gras

s)5

seed

ger

min

atio

n1.

7>

10,0

00[3

2]So

il in

vert

ebra

tes

Eise

nia

foet

ida

14su

rviv

al4.

0#>

9000

[35]

Eise

nia

foet

ida

14su

rviv

al1.

7>

7900

[35]

Sedi

men

t inv

erte

brat

esC

hiro

nom

us r

ipar

ius

(mid

ge)

28em

erge

nce,

sex

rati

o3.

6>

10,0

00[2

7]C

hiro

nom

us te

ntan

s (m

idge

) 10

surv

ival

,gro

wth

4.8

>26

30[1

0]H

yale

lla a

ztec

a (a

mph

ipod

)10

surv

ival

,gro

wth

4.8

>20

90[1

0]Se

dim

ent v

erte

brat

esR

ana

arva

lis (m

oor

frog

)14

egg

hatc

hing

(10

C)

9.0#

>65

7***

[28]

Ran

a ar

valis

(moo

r fr

og)

29ta

dopl

e su

rviv

al a

nd g

row

th (1

0C

)9.

0#>

657*

**[2

8]

Not

e:Ex

posu

re c

once

ntra

tion

s ar

e re

port

ed a

s m

ean

valu

es b

ased

on

anal

ytic

al m

easu

rem

ents

repo

rted

ove

r th

e ex

posu

re p

erio

d un

less

oth

erw

ise

in-

dica

ted.

NR

=N

ot r

epor

ted.

*Ba

sed

on n

omin

al e

xpos

ure

conc

entr

atio

ns r

epor

ted.

** N

OEC

indi

cate

d co

rres

pond

s to

rep

orte

d EC

10.

***

Mea

sure

d ex

posu

re c

once

ntra

tion

inse

dim

ent a

t end

oft

est.

#O

rgan

ic c

arbo

n co

nten

t est

imat

ed b

y m

ulti

plyi

ng %

loss

on

igni

tion

by

0.4.

+ R

epor

ted

fres

h w

eigh

t con

cent

rati

ons

corr

ecte

d to

dry

wei

ght b

y as

-su

min

g a

0.4

dry

to w

et w

eigh

t rat

io.

x –

A n

on-d

ose

depe

nden

t red

ucti

on in

gro

wth

was

obs

erve

d in

bot

h 10

0an

d 10

00m

g/kg

trea

tmen

ts.

or reproduction endpoints are in the range of 10–100 mg/kg dry. A statisticallysignificant NOEC value below 1 mg/kg dry is reported for development of juve-nile springtails (i.e. number of cuticles). However, the authors question the eco-logical significance of this reported effect. Moreover, interpretation of this end-point was further complicated by the high variation in molting frequencyobserved in control animals. For BBP, although limited toxicity data are available,no acute toxicity was reported in earthworms exposed to a soil concentration of1000 mg/kg dry.

Considerable soil and sediment toxicity data are available for high molecularweight PEs.With a few exceptions discussed below, no acute or chronic effects arereported at the highest concentrations investigated, typically >100 mg/kg dry.

Early studies by Swedish investigators reported that DEHP caused adverse ef-fects on microbial respiration and hatching of moor frog eggs at sediment con-centrations below 100 mg/kg dry [21, 36]. However, a critical review of these stud-ies reveals a number of technical problems. In these tests DEHP was spiked to wetsediment by first dissolving the test substance into ethanol. The introduction ofethanol to wet sediment is known to significantly alter the nature of sediment or-ganic carbon as evidenced by a marked increase in the concentration of dissolvedorganic carbon in the pore water (David Mount, USEPA, personal communica-tion). Thus, ethanol functions as a solvent to extract particulate organic carbonfrom sediment particles. This perturbation of the test sediment can significantlyalter normal partitioning behavior and confound toxicity test interpretation. Fur-thermore, since no analytical measurements were provided at the start of toxic-ity tests it is possible that considerable heterogeneity in sediment concentrationsresulted as a result of this spiking procedure, again complicating test interpreta-tion. Given the non-standardized nature of these tests with this species and lackof experience with normal control variation in the toxicity test endpoints exam-ined the reliability of these tests is uncertain. To address these concerns, subse-quent toxicity studies with microbes [6, 17] and moor frogs [28, 29] have been re-ported. In these follow-up studies in which the use of ethanol as a carrier solventwas typically avoided, no effects were observed. Moreover, further experimentsusing ethanol as a carrier solvent did not replicate any of the findings reportedin the original studies [28]. Consequently, the early studies by Thuren and co-workers cannot be regarded as reliable for risk assessment purposes nor serve asan appropriate technical basis for derivation of environmental risk limits as re-cently proposed [37].

An analogous situation is represented by the soil toxicity study conducted by Stanley and Tapp [25] since anomalous test results were reported relative tonumerous other test data available (Table 1). These authors spiked 1, 10, 100 and1000 mg/kg of DEHP to quartz sand and then examined shoot growth of pre-germinated seeds of turnips (Brassica rapa) and oats (Avena sativa) after 14 daysrelative to an untreated control group. No test substance related effects were reported for turnips up to 1000 mg/kg dry but statistical analysis of the raw shootweight data indicated that the growth of oats was significantly reduced at boththe 100 and 1000 mg/kg dry DEHP treatments. However, no concentration-dependent response was evident since both concentrations elicited the same degree of growth reduction (ca. 30%) questioning the interpretation and relia-

326 T.F. Parkerton and C.A. Staples

bility of these findings. If one excludes as unreliable the studies mentioned above,none of the numerous soil and sediment toxicity tests available for DEHP demon-strated an adverse effect at the highest concentration tested (Table 1).

As in the case of DEHP, numerous soil and sediment toxicity studies show noadverse effects for DINP and DIDP at the highest concentrations tested. However,one exception has been reported for DINP in studies with lettuce. Lettuce seedgermination after 5 days was significantly reduced in a concentration-dependentmanner in two soils resulting in a NOEC and LOEC of 1000 and 3000 mg/kg,respectively. A follow-up 28-day chronic toxicity study with lettuce seeds failedto reveal any growth effects at the highest DINP concentration tested(i.e. 1387 mg/kg dry).

The above review of the available ecotoxicological data suggests that high mol-ecular weight phthalates may cause adverse effects on plants at extreme exposureconcentrations (e.g. >1000 mg/kg dry). Curiously, such effects are however notreported for DIDP (Table 1). If the effects observed are genuinely test substancerelated, it is hypothesized that such effects are likely due to a physical explana-tion (e.g. hydrophobic effect on soil influencing water uptake by seeds) ratherthan a systemic toxicity mechanism. Such physical effects have been reportedpreviously for soils contaminated with petroleum hydrocarbons [38].

2.1.2Extrapolation from Aquatic Toxicity Data Using EqP Theory

The extensive aquatic toxicity database that is available for PEs can be extrapo-lated to predict the hazard to soil and sediment-dwelling organisms using theEquilibrium Partitioning (EqP) model:

PNEC(direct) =Koc PNECaquatic (1)

Where:

PNEC(direct) predicted no effect concentration in soil/sediment (mg/kg oc)Koc organic carbon-normalized partition coefficient (l/kg oc)PNECaquatic predicted no effect concentration in surface water (mg/l)

The PNECsoil/sediment can be expressed on a dry weight basis by simply multiplyingby the organic carbon fraction of the soil or sediment. The technical basis sup-porting this approach for deriving sediment or soil quality criteria has previouslybeen described [39–42]. This approach is currently used in a variety of regula-tory programs in both North America and Europe [43, 44].

The Koc in Eq. (1) can be estimated from the octanol-water partition coeffi-cient (Kow) using the correlation reported by Seth et al. [45]:

Koc =0.35 Kow (2)

For the low molecular weight phthalates DMP, DEP, DBP and BBP, a statistical ex-trapolation procedure has been recently applied to available aquatic toxicity datato derive PNECaquatic [2]. However, in the case of higher molecular weight phtha-lates, i.e. alkyl chain length of six or more carbons, no aquatic toxicity is observedat aqueous solubility. As noted earlier, the lack of hazard is attributed to the

An Assessment of the Potential Environmental Risks Posed by Phthalates 327

combined role of low aqueous solubility and limited bioconcentration potentialthat prevent achieving tissue concentrations in biota needed to elicit adverse effects [2]. Consequently, for these substances the water solubility limit can besubstituted into Eq. (1) to estimate a lower-bound concentration below which anecotoxicity effect in soil or sediment is precluded. Water solubility and Log Kowvalues required in these calculations were taken from Cousins and Mackay [46].

In the case of a low organic carbon fraction (0.01) soil or sediment in whichhigh bioavailability is expected, the PNEC(direct) is calculated to range from0.44 mg/kg dry for DMP to 6.67 mg/kg dry for BBP (Table 2). In contrast forhigher molecular weight phthalates, EqP predictions indicate that chronic effectsare not expected at concentrations in the hundred parts per million range evenin soils or sediments with low organic carbon content.

2.1.3Association-Based Methods Based on Field Data

Association-based methods have also been used to derive sediment quality cri-teria for chemicals including selected phthalates (Table 3) as summarized in theU.S. EPA’s national sediment quality survey [44]. These methods are based on theempirical association between a specific biological endpoint (sediment toxicity,benthic diversity) and the concentration of the sediment contaminant deter-mined in concurrent field samples. Barrick et al. [47] developed apparent effectthresholds (AETs) for several phthalates using concurrent chemical and biolog-ical effect data from the Puget Sound Estuary. AETs were defined for each bio-logical indicator as the highest detected concentration among sediment samplesthat did not exhibit statistically significant effects. In other words, AETs charac-terize the highest observed sediment concentration for a given chemical that istolerated without empirical evidence of adverse effect. A somewhat differentmethod was used by the Florida Department of Environmental Protection [48]to calculate a probable effect level (PEL) for DEHP. The PEL was defined as thegeometric mean of the 50th percentile concentration of the effects data (sediment

328 T.F. Parkerton and C.A. Staples

Table 2. Derivation of PNECsoil/sediment (direct) based on equilibrium partitioning theory

PE Aquatic PNEC Log Kow Soil/sediment(mg/L)

PNEC PNEC a

(mg/kg OC) (mg/kg dry)

DMP 3.109 1.61 44 0.44DEP 0.865 2.54 105 1.05DBP 0.043 4.27 280 2.80BBP 0.038 4.70 667 6.67DEHP 2.49 E-03 b 7.73 46802 >468DINP 3.08 E-04 b 8.60 42916 >429DIDP 3.80 E-05 b 9.46 38358 >384

a Assuming an organic carbon content of 1%.b Water solubility limit.

samples exhibiting biological effects) and the 85th percentile concentration of thenon-effects data (sediment samples not exhibiting a statistically significant bio-logical response).

2.1.4PNEC Selection for Direct Effects

Comparison of association-based PNECs (Table 3) with causality-based PNECsderived using EqP indicate that the former values are 2 to 360 times lower (i.e.more conservative) with DEHP showing the greatest discrepancy. Association-based PNECs for DEHP are clearly inconsistent with the results of laboratory tox-icity tests summarized in Table 1 and thus do not provide a sound basis for riskassessment. The principle limitation of association-based PNECs is that causal relationships between concentration and biological responses cannot be estab-lished due to the confounding influence of other contaminant and non-contam-inant factors that influence biological endpoints in field samples.Additional flawsin this methodology based on statistical considerations have recently been de-scribed by von Stackelberg and Menzie [49].

In contrast, lower-bound PNECs derived using EqP for DEHP, DINP and DIDPare fully consistent with the lack of toxicity observed for high molecular weightPEs. Moreover, soil and sediment toxicity test results, summarized in Table 1 pro-vide empirical evidence that the PNECs presented in Table 2 for lower molecu-lar weight PEs are protective for terrestrial and benthic species. For example, thePNEC for DBP is estimated to be 280 mg/kg oc whereas the chronic NOAEL for themost sensitive test species (Springtails) is >750 mg/kg oc. Further support for theuse of EqP in the derivation of sediment PNECs is provided by Call et al. [10].Basedon the above discussion, PNECs obtained by extrapolation of aquatic toxicity datausing EqP theory were used to quantify risks posed by direct effects.

2.2Indirect Effects on Wildlife via the Food Chain

To assess the hazard posed to wildlife that consume terrestrial or benthic or-ganisms that have been exposed to PEs in soil or sediment a no observed adverse

An Assessment of the Potential Environmental Risks Posed by Phthalates 329

Table 3. Comparison of predicted no effect concentrations for sediment

PE Causality-based Association-based ReferencePNEC (mg/kg dry) a PNEC (mg/kg dry) b

DMP 0.44 0.16 [47]DEP 1.05 0.2 [47][47] 2.80 1.4 [47]BBP 6.67 0.9 [47]DEHP >468 1.3–1.9 [47]

2.65 [48]

a Derived using Eq. 1 and data provided in Table 2.b Derived using field data.

effect level (NOAELwildlife) must be defined. This value should be derived fromlong-term dietary toxicity studies with mammals or birds and be based on effectendpoints relevant to wildlife populations, i.e. survival, growth and reproduction.David et al. [50], has recently provided a detailed review of the available labora-tory toxicity studies for PEs with mammals. Long-term dietary toxicity studieswith rats are available for all the PEs considered in this analysis. Based on theavailable toxicological database, a NOAELrat was selected from the most relevantstudy that demonstrated population-based effects (Table 4). In contrast to the ex-tensive toxicological database available for PEs in mammals, limited toxicologi-cal data are available in avian species. As in the case of mammals, PEs are re-ported to exhibit low acute toxicity to birds [58]. Chronic data in avian species are available for DEHP. O’Shea and Stafford [59] reported no adverse effects onsurvival or growth of European starlings fed DEHP at a dietary concentration of 250 mg/kg for 30 days corresponding to a NOAEL >30 mg/kg body wt/day.In a 4-week feeding study with chickens, egg production and growth were de-creased at ca. 300 mg/kg body wt/day [60]. In another 230 day feeding study withchickens, cessation of egg production was reported at ca. 600 mg/kg body wt/day[61]. These studies suggest long-term effects for DEHP in avian species occur inthe same range as reported for rats (i.e. NOAELrat =113 mg/kg body wt/day,Table 4).

The derivation of NOAELwildlife from NOAELrat requires extrapolation factorsfor allometric scaling of dose as well as uncertainty regarding species sensitiv-ity. These considerations can be expressed in equation form [62] as:

NOAELrat Wrat0.33

NOAELwildlife =394 �94� (3)UF Wwildlife

Where:

Wrat body weight of rat (kg)Wwildlife body weight of wildlife (kg)UF uncertainty factor for interspecies sensitivity

A recent review of ecological risk assessments conducted in the U.S. revealed an UF of 10 is typically assumed for interspecies extrapolation [62]. This study

330 T.F. Parkerton and C.A. Staples

Table 4. Long-term laboratory toxicity studies with rats

PE Study type Effect endpoint a NOAELrat Ref.(mg/kg/day)

DMP Cancer Growth 1440 [51]DEP Teratology Material survival 1800 [52]DBP 2-Gen. Repro.b Litter size 60 [53]BBP 2-Gen. Repro.b No effects >100 [54]DEHP 2-Gen. Repro.b Pup survival during lactation 113 [55]DINP 2-Gen. Repro.b No effects >600 [56]DIDP 2-Gen. Repro.b Pup survival during post-partum 108 [57]

a Most sensitive population-based endpoint reported to show a significant effect.b Two-generation reproductive toxicity test.

also reported that the weight of mammalian wildlife receptors ranged from 0.025 (deer mouse) to 100 kg (harbor seal).

For derivation of a risk-based PNECsoil/sediment (indirect) intended to protectwildlife from indirect exposure via the food chain, the NOAELwildlife is equated tothe dietary dose derived via this pathway:

NOAELwildlife =Iprey Rwd Flip BSAF PNEC(indirect) (4)Where:

Iprey wildlife ingestion rate to prey (kg prey dry/kg wildlife/day)Rwd wet to dry weight ratio of prey (kg wet/kg dry)Flip lipid fraction of prey (kg lipid/kg wet)BSAF Biota to soil/sediment accumulation factor normalized to lipid and or-

ganic carbon (kg oc/kg lipid)

The ingestion rate of prey can be estimated based on the allometric equation pro-vided by Nagy [63]:

Iprey =0.07 Wwildlife–0.18 (5)

Where:

Wwildlife is the body weight of the wildlife receptor in kg wet

Substituting Eqs. (3) and (5) into (4) and solving for the PNEC yields:

14 NOAELrat Wrat0.33

PNEC(indirect) =389993 (6)W 0.15

wildlife UF Rwd Flip BSAF

Due to the susceptibility of PEs to biotransformation, these substances are not ex-pected to undergo biomagnification [5]. In fact in a recent field study decreasingconcentrations of phthalates in biota (i.e. biodilution) have been demonstratedwith increasing trophic position for high molecular weight PEs [64]. Conse-quently, organisms at the base of the food web possessing limited metabolic capability (e.g. mollusks) are expected to exhibit the highest concentration ofPEs. For these organisms, the Equilibrium Partitioning model provides a con-servative characterization of the BSAF in Eq. (6).

In order to apply Eq. (6), typical values are assumed for all input parametersexcept Wwildlife in which the maximum value reported by Duke and Taggart [62]is selected. An extreme value for this input was chosen to ensure calculatedPNECs are conservative.

Based on the following assumptions:

Wrat 0.48 kgWwildlife 100 kgRwd 5 kg wet/kg dryFlip 0.01 kg lipid/kg wetUF 10BSAF 1

Substitution into Eq. (6) results in the following approximation:

PNEC(indirect) = 10 NOAELrat (7)

An Assessment of the Potential Environmental Risks Posed by Phthalates 331

An alternative approach to wildlife effect assessment is provided by the EuropeanTechnical Guidance Document of new and existing substances [65]. The approach used to assess the potential for “secondary poisoning” via the foodchain first involves calculation of a predicted no effect concentration in the dietof a wildlife consumer:

NOAELrat CFPNECoral =399 (8)

AFWhere:

PNECoral Predicted no effect concentration in the diet (mg/kg diet)CF Conversion factor (kg body wt – day/kg diet)NOAELrat No adverse effect level from chronic rat study (mg/kg body wt/day)AF Application factor to account for interspecies variation and lab to field

extrapolations

The default value for the conversion factor varies between 10–20 for rats de-pending on test animal size while the default application factor of 30 is appliedto a rat chronic study for extrapolation purposes. If the Equilibrium Partitioningmodel is applied in conjunction with Eq. (8) the following equation is obtainedfor soil/sediment:

NOAELrat CFPNEC(indirect) =399 (9)

AF Flip BSAF

As in Eq. (7), the PNEC is expressed on an organic carbon basis and other vari-ables are as previously defined.

Given the following inputs:

CONV 10Flip 0.01 kg lipid/kg wetAF 30BSAF 1

Substitution into Eq. (9) yields the following result:

PNEC(indirect) =33 NOAELrat (10)

Hence the EU TGD approach for wildlife effect assessment gives a similar, albeitslightly less conservative, PNEC than obtained using the methodology outlinedfor deriving Eq. (7).

The above analysis has focused on potential adverse effects to wildlife preda-tors that ingest soil or sediment-dwelling biota. However herbivores should alsobe considered. Past research suggests that phthalates are very inefficiently trans-ferred from soil to plants hence this is not expected to be a significant wildlife ex-posure pathway of concern [1]. However, herbivorous wildlife or domestic live-stock may ingest significant amounts of soil. In a recent study by Rhind et al. [66],the amount of DEHP ingested by sheep via soil from pastures amended withsewage sludge was investigated. This study found that sheep weighting 25–80 kgingested 28 to 135 g dry soil per day depending on season.A maximum daily soilingestion rate of 314 g dry soil was also reported.Applying Eq. (3) for derivation

332 T.F. Parkerton and C.A. Staples

of a NOAEL for sheep and equating this to the maximum dose that a sheep wouldreceive via soil ingestion enables a soil PNEC to be derived:

NOAELrat Wrat0.33

PNEC(indirect) =394 �91 � (11)UF Isoil foc Wsheep

Where:

Isoil Soil ingestion rate for sheep (kg soil/kg body wt/day)ƒoc Organic carbon fraction of soil (kg oc/kg dry)

Applying the same defaults for Wrat and UF as previously described and assum-ing a 0.01 organic carbon fraction and a maximum soil ingestion rate of0.314 kg/day for a 25 kg sheep yields:

PNEC(indirect) =216 NOAELrat (12)

Comparison of Eqs. (12) with (7) suggests that the risks posed to wildlife by in-gestion of contaminated prey will be greater than that posed via direct soil in-gestion thus dictating PNEC derivation.

PNECs intended to protect wildlife derived using Eq. (7) are compared toPNECs intended to protect soil and sediment-dwelling organisms in Table 5.Results indicate that direct effects drive environmental concerns for lower molecular weight PEs while indirect effects dictate the environmental hazard forthe higher molecular weight PEs.

An Assessment of the Potential Environmental Risks Posed by Phthalates 333

Table 5. Comparison of PNECsoil/sediment for direct and indirect effects a

PE PNECdirectb (mg/kg dry) PNECindirect

c (mg/kg dry)

DMP 0.44 144DEP 1.05 180DBP 2.80 6.0BBP 6.67 >10.0DEHP >468 11.3DINP >429 >60DIDP >383 10.8

a PNEC values are normalized to a 1% organic carbon content.b Derived using Eq. 1, Table 1.c Derived using Eq. 7.

3Exposure Characterization

3.1Soil

3.1.1Native Soil

The principle source of PEs to soils that are not amended with sewage sludge isatmospheric deposition. Several studies have reported atmospheric depositionrates of PEs in different locations (Table 6). Based on these data an atmosphericdeposition of 1 µgdry m–2 d–1 is typical for DMP, DEP, DBP and DEHP. Given thisestimate and assuming an average mixing depth for non-agricultural soil of0.05 m and a soil density of 1700 kgdry m–3 [65] and ignoring the mitigating roleof biodegradation the resulting annual background soil concentration is esti-mated to be 0.004 mg kgdry

–1 . Since atmospheric deposition of BBP appears to beabout an order of magnitude lower, even lower background soil concentrationsare expected.

3.1.2Sludge-Amended Soil

Sludge from municipal wastewater treatment plants is typically disposed of viaincineration, placement in landfills, or via land application to agricultural fields,forested land or other sites e.g. parks, golf courses, and reclamation projects.This latter disposal method is often viewed as the most cost-effective and envi-ronmentally beneficial option [72]. The enhanced use of sludge for agriculturalpurposes is also a policy endorsed by the EU [73].

Certain PEs are commonly detected in sewage sludge from municipal waste-water treatment plants (Table 6). Thus, an examination of potential exposure andrisks to soil-dwelling organisms and terrestrial wildlife that results from sludgeapplication is warranted.

Sludge application rates differ regionally and by type of application. For ex-ample in the US, typical sludge application rates to agricultural soils are1 kgdry m–2 yr–1 (=10 t ha–1 yr–1) while a higher rate of 1.8 kgdry m–2 yr–1 is used for

334 T.F. Parkerton and C.A. Staples

Table 6. Atmospheric deposition rates reported for phthalate esters

Location Deposition Deposition flux (mg/m2/d)

Type DMP DEP DBP BBP DEHP Ref.

US Great Lakes Wet+Dry NR NR 0.53 NR 0.53 [67]Sweden Wet+Dry NR NR 0.56 NR 0.79 [68]Denmark Wet+Dry NR NR 0.31 0.05 0.56 [69]Germany Wet NR NR 0.66 NR 1.56 [70]Germany Wet 1.15 1.07 1.57 0.10 2.88 [71]

NR = Not reported.

forested or public lands [72]. In Canada, sludge application of 0.8 kgdry m–2 yr–1 isallowed over a 5-year period. In Europe, typical sludge application rates to agri-cultural soil and grassland are 0.5 and 0.1 kgdry m–2 yr–1, respectively [65] with application rates of as high as 1.7 kgdry m–2 yr–1 also reported [74]. Neglecting therole of degradation processes and background concentration in native soil, a conservative estimate of the soil concentration resulting from sludge amendmentis given by:

Csludge XappCsoil =99 (13)Zsoil rsoil

An Assessment of the Potential Environmental Risks Posed by Phthalates 335

Table 7. Summary of phthalate concentrations in sludges and estimated upper-bound soil con-centrations in sludge amended soils

PE Location Year Average sludge No. of Ref. Estimated soil concentration Samples concentration (mg/kgdry) (mg/kgdry) a

DMP Canada 93/94 0.030 72 [75] <0.001Denmark 95/96 0.034 11 [76] <0.001Europe b 99 <0.040 b 35 [77] <0.001

DEP Canada 93/94 0.228 72 [75] <0.001Denmark 95/96 0.238 11 [76] <0.001Europe b 99 <0.160 c 35 [77] <0.001

DBP Canada 93/94 6.84 72 [75] 0.034Germany 93/94 22.5 50 [78] 0.112Germany 97 0.5 15 [79] 0.002Germany 97 1.83 7 [80] 0.009Denmark 95/96 3.88 20 [76] 0.019Europe b 99 6.50 c 35 [77] 0.032

BBP Canada 93/94 2.97 72 [75] 0.015Germany 93/94 11.7 50 [78] 0.058Germany 97 <0.02 15 [79] <0.001Germany 97 0.82 28 [80] 0.004Denmark 95/96 0.18 20 [76] <0.001Europe b 99 3.6 c 35 [77] 0.018

DEHP Canada 93/94 150 72 [75] 0.750Germany 93/94 23.4 50 [78] 0.117Germany 97 67.3 15 [79] 0.336Germany 97 19.2 7 [80] 0.096Denmark 95/96 37.9 20 [76] 0.189Norway NR 58 36 [81] 0.290Europe b 99 85.4 c 35 [77] 0.427

DINP Germany 97 9.1 5 [80] 0.045Europe b 99 <0.5 c 35 [77] 0.003

DIDP Europe b 99 <1.3 c 35 [77] 0.007

a Soil concentration estimated based on application of Eq. (13) using a upper bound sludge application rate of 1.8 kgdry m–2 yr–1.

b Survey included sludge samples collected from UK, France, Germany, Sweden and TheNetherlands.

c Median value.NR=Not reported.

Where:

Csoil Concentration in sludge amended soil (mg kg–1dry)

Csludge Concentration in sewage sludge (mg kg–1dry)

Xapp Sludge application rate (kgdry m–2 yr–1)Zsoil Soil mixing depth (m)rsoil Soil bulk density (kgdry m–3)

Assuming a typical agricultural soil mixing depth of 0.2 m and a bulk density of1700 kgdry m–3 [65], sludge concentrations will be diluted by two hundred-fold atan upper-bound sludge application rate of 1.7 kgdry m–2 yr–1. Repeated applicationin subsequent years is not expected to result in a progressive accumulation overtime due to the biodegradable nature of PEs. Given average PE concentrations insludge and the assumptions outlined above, upper-bound concentrations insludge amended soil are calculated in Table 7. Based on a comparison to pre-dicted exposure concentrations discussed in Sect. 3.1.1, these estimates suggestsludge amendments do not serve as a major contributor to DMP or DEP in soil.In contrast, sludge addition appears to be the dominant source of the more hy-drophobic, higher molecular weight PEs.

3.1.3Field Monitoring Data for Soil

Several recent field studies have reported PE concentrations in European soils.Monitoring data for individual PEs are summarized graphically in Figs. 1–6. Ina Dutch regional monitoring survey [83], median soil concentrations of all PEswere about an order of magnitude lower than observed in an industrial area [82].A similar range of DEHP soil concentrations was reported in sludge-amendedsoil in the United Kingdom [66]. The large variation in DBP, BBP, DEHP andDINP soil concentrations reported in Denmark [97] reflects the dramatic differ-ences in sludge application rates to the soils investigated.

For the higher molecular weight PEs (DBP and higher), the observed range in soil concentrations across these studies appears to correspond with the range in concentrations expected from inputs derived from background at-mospheric deposition to sludge application. Moreover, DMP and DEP con-centrations reported in the Dutch monitoring survey are consistently low and at a level reflective of atmospheric sources as anticipated. However, ob-served DMP and DEP soil concentrations reported in Germany are much higher than expected based on exposure calculations presented in the previoussections. This discrepancy suggests that the overall magnitude of emissions tosoil in this local industrialized area is significantly higher than presumed forthese PEs.

3.2Sediment

Sources of phthalates to the aquatic environment include industrial and domes-tic wastewater effluents as well as non-point source inputs such as urban runoff

336 T.F. Parkerton and C.A. Staples

and atmospheric deposition [69]. While phthalates can undergo anaerobicbiodegradation, as is the case for many organic chemicals, degradation generallyoccurs more slowly than in aerobic soils [5]. Differences in expected half-lives be-tween soil and sediment largely explain the higher concentrations predicted insediment by multimedia fugacity models [4].

3.2.1Field Monitoring Data for Sediments

Field data published from the 1990’s are shown in Figs. 1–6. Data compilationwas restricted to recent studies to reduce the confounding problem of laboratorycontamination that has plagued interpretation of early PE measurements whilealso facilitating direct comparison of contemporary sediment measurementswith recent soil surveys discussed in Sect. 3.1.3.

An Assessment of the Potential Environmental Risks Posed by Phthalates 337

Fig. 1. Median and range of dimethyl phthalate (DMP) soil and sediment concentrations re-ported in field surveys. The total number of samples analysed in each study is provided on theright hand axis. The predicted no effect concentration (PNEC) is indicated by a solid verticalline. Values in parenthesis indicate the number of samples for a given study that exceed thePNEC. BDL indicates all samples examined were below the analytical detection limit(s). Letterscorrespond to the following references: A=UMEG 1999 [82]; B=Alcontrol 1999 [83]; C=Furt-mann 1993 [71]; D=Alcontrol 1999 [83]; E=Vethaak et al. 2002 [84]; F=Niva 1996 [85];G=Parkmann and Remberger 1995, 1996 [86, 92]; H=Tan 1995 [87]; I=Lopes and Furlong 2001[88]; J=USEPA 1997 [44]; K=Garrett 2002 [89]; L=Mackintosh et al. 2002 [90]

Concentration (mg/kg dry wt)

Several generalizations are apparent from this analysis. First, an extensivemonitoring database for single isomer PE concentrations in sediments fromNorth America, Europe and Asia is available. Limited data are also available forthe mixed isomers, DINP and DIDP. Median sediment concentrations of the in-dividual PEs follow the general trend: DEHP >DBP, DINP, DIDP >BBP, DMP, DEP.

Second, PE sediment concentrations often exhibit a several order of magnituderange both within and between monitoring studies. This variance reflects site-specific factors associated with sampling locations (e.g. vicinity of point sourceinputs, degree of industrial activity, receiving water dilution), the non-persistentnature of PEs (e.g. degradation as one proceeds from a source) as well as poten-tial differences in analytical methodology (e.g. method detection limits, extrac-tion efficiencies). Third, there are no obvious geographical differences in sedi-ment concentrations which is likely due to the large scatter in results reported

338 T.F. Parkerton and C.A. Staples

Fig. 2. Median and range of diethyl phthalate (DEP) soil and sediment concentrations reportedin field surveys. The total number of samples analysed in each study is provided on the righthand axis. The predicted no effect concentration (PNEC) is indicated by a solid vertical line.Values in parenthesis indicate the number of samples for a given study that exceed the PNEC.BDL indicates all samples examined were below the analytical detection limit(s). Letters cor-respond to the following references: A=UMEG 1999 [82]; B=Alcontrol 1999 [83]; C=Furtmann1993 [71]; D=Vitali et al. 1997 [71]; E=Alcontrol 1999 [83]; F=Vethaak et al. 2002 [84]; G=Niva1996 [85]; H=Parkmann and Remberger 1995, 1996 [92] [86]; I=Tan 1995 [87]; J=JAE 1993[93]; K=MOC 1999a, b, c [94–96]; L=Lopes and Furlong 2001 [88]; M=USEPA 1997 [44];N=Garrett 2002 [89]; O=Mackintosh et al. 2002 [90]

Concentration (mg/kg dry wt)

across different world regions. Lastly, median sediment concentrations typicallybracket median soil concentrations. However, comparison of the regional surveyof sediment and soil PE concentrations in the Netherlands clearly shows that me-dian concentrations in sediment are higher than soil [83]. Moreover, maximumconcentrations exceed those reported in soils often by orders of magnitude(Figs. 1–6).

4Risk Assessment

To assess the potential risk that soil and sediment associated phthalates pose toterrestrial and benthic organisms by direct exposure or to wildlife by indirect ex-

An Assessment of the Potential Environmental Risks Posed by Phthalates 339

Fig. 3. Median and range of dibutyl phthalate (DBP) soil and sediment concentrations reportedin field surveys. The total number of samples analysed in each study is provided on the righthand axis. The predicted no effect concentration (PNEC) is indicated by a solid vertical line.Values in parenthesis indicate the number of samples for a given study that exceed the PNEC.Letters correspond to the following references: A=UMEG 1999 [82]; B=Alcontrol 1999 [83];C=Vikelsoe et al. 1999 [97]; D=Vitali et al. 1997 [71]; E=Furtmann 1993 [71]; F=Fromme etal. 2002 [79]; G=Vondracek et al. 2001 [98]; H=Alcontrol 1999 [83]; I=Vethaak et all 2002 [84];J=Vikelsoe et al. 2001 [99]; K=Parkmann and Remberger 1995, 1996 [86, 92]; L=NIVA 1996[85]; M=Tan 1995 [87]; N=JAE 1993 [93]; O=MOC 1999a, b, c [94–96]; P=Lopes and Furlong2001 [88]; Q=USEPA 1997 [44]; R=Garrett 2002 [89]; S=Mackintosh et al. 2002 [90]

Concentration (mg/kg dry wt)

posure via the food chain, observed concentrations were compared to the low-est PNEC given in Table 5. Results of this comparison are illustrated in Fig. 1-6 foreach of the PEs investigated. In cases where the maximum concentration for agiven study exceeds the PNEC (denoted by a solid vertical line), the number ofexceedances (indicated in parenthesis) is specified next to the total number ofsamples analyzed (shown on the right hand axis of the plot).

For all PEs investigated, none of the observed soil concentrations exceededrisk-based limits even in the case of soils that were heavily amended with sewagesludge. Similarly, for BBP and DINP, none of the maximum sediment concentra-tions reported in field surveys exceeded the PNEC. For the remaining phthalates,at least one monitoring study indicated maximum reported sediment concen-trations above the PNEC. In the case of DMP, studies from the US and the Nether-

340 T.F. Parkerton and C.A. Staples

Fig. 4. Median and range of butylbenzyl phthalate (BBP) soil and sediment concentrations re-ported in field surveys. The total number of samples analysed in each study is provided on theright hand axis. The predicted no effect concentration (PNEC) is indicated by a solid verticalline. Values in parenthesis indicate the number of samples for a given study that exceed thePNEC. Letters correspond to the following references: A = UMEG 1999 [82]; B = Alcontrol 1999[83]; C = Vikelsoe et al. 1999 [97]; D = Vitali et al. 1997 [71]; E = Furtmann 1993 [71]; F = From-me et al. 2002 [79]; G = Alcontrol 1999 [83]; H = Vethaak et all 2002 [84]; I = Vikelsoe et al. 2001[99]; J = NIVA 1996 [85]; K = Parkmann and Remberger 1995, 1996 [86, 92]; L = JAE 1993 [93];M = MOC 1999a, b, c [94–96]; N = Lopes and Furlong 2001 [88]; O = USEPA 1997 [44];P = Garrett 2002 [89]; Q = Mackintosh et al. 2002 [90]

Concentration (mg/kg dry wt)

lands indicated 34 sediment samples were above the PNEC thus representing lessthan 2% of the reported sediment measurements included in Fig. 1. For DEP, twostudies from North America yielded 5 samples exceeding the PNEC thus repre-senting less than 0.3% of the available sediment monitoring database. Four mon-itoring studies from both North America and Europe indicated 14 sediment sam-ples with DBP concentrations above the PNEC which translates to 0.6% of thereported sediment measurements. In the case of DEHP, six studies representingfield data sets from North America, Europe and Asia were shown to include a lim-ited number of measurements that exceeded the PNEC. The elevated concentra-tions reported in Sweden corresponded to samples taken near a production site.Collectively across all monitoring studies, 28 sediment samples were found to beabove the PNEC representing 1.4% of the reported DEHP measurements. Three

An Assessment of the Potential Environmental Risks Posed by Phthalates 341

Fig. 5. Median and range of di-2-ethylhexy phthalate (DEHP) soil and sediment concentrationsreported in field surveys. The total number of samples analysed in each study is provided onthe right hand axis. The predicted no effect concentration (PNEC) is indicated by a solid ver-tical line. Values in parenthesis indicate the number of samples for a given study that exceedthe PNEC. Letters correspond to the following references: A = UMEG 1999 [82]; B = Alcontrol1999 [83]; C = Rhind et al. 2002 [66]; D = Vikelsoe et al. 1999 [97]; E = Vitali et al. 1997 [71];F = Furtmann 1993 [71]; G = Fromme et al. 2002 [79]; H = Vondracek et al. 2001 [98]; I = Al-control 1999 [83]; J = Vethaak et all 2002 [84]; K = Long et al. 1998 [100]; L = Boutrup et al. 1998[101]; M = Vikelsoe et al. 2001 [99]; N = NIVA 1996 [85]; O = Parkmann and Remberger 1995,1996 [86, 92]; P = Tan 1995 [87]; Q = JAE 1993 [93]; R = MOC 1999a, b, c [94–96]; S = USEPA1997 [44]; T = Lopes and Furlong 2001 [88]; U = Garrett 2002 [89]; V = Mackintosh et al. 2002 [90]

Concentration (mg/kg dry wt)

sediment samples in the immediate vicinity of a European DIDP productionplant with limited wastewater treatment facilities were found to exceed the PNECfor this substance. However median concentrations reflecting the regional ex-posure situation for DIDP are 2–3 orders of magnitude lower.

5Summary and Discussion

Risk assessment should form the logical basis for rationale management of con-taminated soils and sediments. Derivation of transparent, scientifically defensi-ble, causal, risk-based soil and sediment quality criteria is a critical aspect of therisk assessment process. In this study, the technical basis used in developingPNECs for each PE is explained. These values are then used for screening avail-able monitoring data using the simple hazard quotient (i.e. PEC/PNEC) para-digm. Our analysis suggests that in the case of the low molecular weight PEs,direct toxicity to soil or sediment-dwelling organisms dictates PNEC derivation.

342 T.F. Parkerton and C.A. Staples

Fig. 6. Median and range of di-isononyl and di-isodecyl phthalate (DINP, DIDP) soil and sed-iment concentrations reported in field surveys. The total number of samples analysed in eachstudy is provided on the right hand axis. The predicted no effect concentration (PNEC) is in-dicated by a solid vertical line.Values in parenthesis indicate the number of samples for a givenstudy that exceed the PNEC. BDL indicates all samples examined were below the analytical detection limit(s). Letters correspond to the following references: A, E = Mackintosh et al. 2002[90]; B, D, F, G = Alcontrol 1999 [83]; C = Vikelsoe et al. 1999 [97]

Concentration (mg/kg dry wt)

In contrast, indirect effects on wildlife via food chain exposure determine PNECsfor higher molecular weight PEs. However, despite the hydrophobicity of highermolecular weight PEs, trophic transfer is limited by metabolism at higher trophiclevels [5, 64, 102, 103]. As a result, the PNECs presented are intended to protectwildlife that exclusively consume benthic and terrestrial prey which are at thebase of the food chain and lack metabolic capabilities (e. g. mollusks,oligochaetes). Screening calculations suggest dietary exposure via ingestion ofcontaminated prey is a more important pathway then direct ingestion of conta-minated soil or sediment. These calculations were based on the extreme soil in-gestion rate reported for sheep. However, weight normalized soil ingestion ratesfor other domestic animals appear similar to sheep further supporting this gen-eral conclusion [102].

It is important to highlight some key assumptions invoked in the PNEC de-rivation since these are likely to contribute conservatism in the numerical valuesobtained. First, PNECs were calculated using an organic carbon content of only1%. Often, contaminated soil or sediments have a much higher organic carboncontent. Therefore, in accordance with the EqP theory, this assumption is likelyto overstate the bioavailability of highly contaminated field samples. This con-servatism can be illustrated by reexamining the number of sediment measure-ments reported in the USEPA national sediment quality inventory that exceed theorganic carbon normalized PNEC. In the case of DMP, only 8 sediment samplesout of 580 where organic carbon measurements were also reported, exceeded the PNEC. This represents a decline in the exceedance frequency from 3%(30/1001 on a dry weight basis) to 1.4% (8/580 on an organic carbon basis). In thecase of DBP, only 2 out of 520 organic carbon normalized sediment concentra-tions were above the PNEC reflecting a reduction in the exceedance frequencyfrom 0.9% (dry weight basis) to 0.2% (organic carbon basis). None of the organiccarbon normalized concentrations of either DEP or DEHP exceeded the corre-sponding PNEC in contrast to dry weight based concentrations reported in thisdata set (Figs. 2, 5).

A second assumption likely to introduce conservatism is that factors affectingthe bioavailability of PEs in field samples have been ignored. While the EqP theory may be applicable to freshly spiked laboratory tests [104], such assump-tions may introduce significant conservatism when applied to field samples thathave been subjected to sequestration processes and/or contain PEs in an inertform. The importance of these considerations in risk assessment of soils and sed-iments is well recognized [105–107]. Recent laboratory and field studies clearlydemonstrate the reduction in DEHP bioavailability to soil microbes over time[74, 108]. The relationship between PE bioavailability and environmental per-sistence is further reviewed by Peterson and Staples [109]. Recent work also sug-gests that certain PEs may be present in sewage sludge as abraded PVC particlesthus occurring in an occluded state within the polymer matrix [110]. In this form,PEs may be less bioavailable than EqP predictions for either toxicity or bioaccu-mulation indicate.

A third aspect of the PNEC derivation affording conservatism relates specifi-cally to the protection of wildlife consumers. In the screening risk assessmentpresented, the spatial scale of PE contamination in soil/sediment is not consid-

An Assessment of the Potential Environmental Risks Posed by Phthalates 343

ered. However, PNECs are based on the assumption that the wildlife consumersreceive their entire diet from an area corresponding to the PNEC concentration.Since individual grab samples may represent only very localized “hot spots”, theactual exposure that resident wildlife consumers receive can in fact be muchlower since prey from less contaminated areas also comprise a portion of the diet.Thus, neglecting spatial scale may overstate potential risks to wildlife consumers.

With respect to exposure assessment, considerable field data are availablecharacterizing PEs concentrations in sediments in different regions. Less infor-mation is generally available on PE concentrations in soil although availablemonitoring data suggest maximum concentrations found in sediments exceedconcentrations observed in the soil compartment. For a number of the moni-toring studies available, sampling has specifically targeted or at least includedsites that are expected to exhibit high PE contamination (e.g. production sites).Thus, observed exposure measurements reflect both regional and local exposurescenarios which in part explains the wide range of concentrations reported asnoted previously. Differences in exposure concentrations may also reflect differ-ences in analytical protocol for PE determination that were taken to preventingor correcting for background laboratory contamination. In a number of the stud-ies little information is provided on these critical analytical details. For example,in the study by Long [100], the issue of blank determination is not even men-tioned. It is interesting to note that median sediment concentrations reported inthis study are higher than in any of the other studies. The uncertainty in the an-alytical methods employed in this study at least questions the reliability of thesedata for use in the context of risk assessment.

Despite the conservatism in PNEC derivation and the potential bias in ex-posure characterization, the key conclusion from this screening risk assessmentis that environmental concerns posed by soil and sediment-associated PEs are indeed low.

Given the available hazard and exposure data, environmental concerns are,at worst, restricted to infrequent, localized hot spots of contaminated sediments.At these sites, a more detailed risk assessment would be needed to refine the con-servative assumptions used in the present analysis. Additional exposure data onPE concentration in soils and mixed isomer concentrations in sediments (par-ticularly from the US and Japan) are suggested as future needs for compliment-ing the existing monitoring database and further confirming the conclusionreached in the present study. The general conclusion of this screening risk analy-sis for the soil/sediment compartment is consistent with the findings of a simi-lar PE risk assessment focused on the surface water compartment [3]. This studyconcluded that PEs do not pose a threat to aquatic organisms in North Americanand Western European surface waters.

The results of this study indicate that current pollution controls are generallyquite effective in limiting PEs in soil and sediment to concentrations that are wellbelow risk-based thresholds. Consequently, future regulatory initiatives that areintended to improve environmental quality by imposition of additional controlsor restrictions, specifically on phthalates, are expected to provide little benefit.For example, based on the analysis presented in this study the proposed EU limitfor DEHP of 100 mg/kg in sewage sludge [73] not only lacks a sound risk-based

344 T.F. Parkerton and C.A. Staples

justification but also restricts the practical societal use of a cost-effective and en-vironmentally beneficial waste disposal option. Similarly, the risk analysis pre-sented in this study has shown that application of association-based effect thresh-olds for phthalates can lead to misdirected policy priorities for the managementof contaminated sediments. These examples highlight the role that a environ-mental risk assessment can serve in guiding rational decision-making.

6References

1. Staples CA, Parkerton TF, Adams WJ, Gorsuch JW, Biddinger GR, Reinhert K (1997a) Environ Toxicol Chem 16:875

2. Parkerton TF, Konkel WJ (2000) Ecotox Environ Saf 45:613. Staples CA, Parkerton TF, Peterson TR (2000) Chemosphere 40:8854. Cousins I, Mackay D, Parkerton TF (2002) Physical Chemical Properties and Evaluative

Fate Modeling of Phthalate Esters, Chapter submitted to Springer-Verlag EnvironmentalChemistry Handbook Volume for Phthalate Esters

5. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997b) Chemosphere 35:6676. Cartwright CD, Thomas IP, Burns RC (2000) Environ Sci Technol 19:12537. Herring R, Bering L (1988) Contam Toxicol 40:6268. Neuhauser EF, Durkin PR, Malecki MR, Anatra M (1986) Comparative Biochemistry and

Physiology. 83C.1.197–2009. Hulzebos EM, Adema DMM, Dirven-Van Breemen EM, Henzen L, van Dis WW, Her-

bold HA, Hoekstra JA, Baerselman R, van Gestel CAM (1993) Environ Toxicol Chem12:1079

10. Call DJ, Cox CA, Geiger DL, Genisot KI, Markee TP, Brooke LT, Polkinghorne CN, VandeVenter FA, Robillard KA, Gorsuch JW, Parkerton TF, Reiley MC, Ankley GT, Mount DR(2001) Environ Toxicol Chem 20:1805

11. Shea PJ, Weber JB, Overcash MR (1982) Bull Environ Contam Toxicol 29:15312. Jensen J, van Langevelde J, Putze G, Henning Krough P (2001) Environ Toxicol Chem

20:108513. Tagatz ME, Plaia GR, Deans CH (1986) Bull Environ Contam Toxicol 37:141.14. Huntingdon Life Science (SLU 001/983882) Butylbenzylphthalate, acute toxicity (LC50) to

the earthworm (Eisenia foetida) (1998)15. Mathur SP (1974) J Environ Quality 3 :20716. Kirchmann H, Aström H, Jönsäll G (1991) Swedish J Agric Res 21:10717. Kaiser T, Felgentreu D (2000) Determination of the influence of Di-2 ethylhexyl phthalate

(DEHP) on soil quality – Part I: Determination of soil microflora activity. Berlin, FederalBiological Research Centre for Agriculture and Forestry; Institute of Ecological Chemistry

18. Kaiser T, Felgentreu D (2000a) Determination of the influence of Di-2 ethylhexyl phtha-late (DEHP) on soil quality – Part II: Determination of dehydrogenase activity. Berlin,Federal Biological Research Centre for Agriculture and Forestry; Institute of EcologicalChemistry

19. Kaiser T, Felgentreu D (2000b) Determination of the influence of Di-2 ethylhexyl phthalate (DEHP) on soil quality – Part III: Determination of nitrogen mineralization in soils and the influence of DEHP on these processes. Berlin, Federal Biological ResearchCentre for Agriculture and Forestry; Institute of Ecological Chemistry

20. Mutz RC, Jones JR (1977) The effects of phthalate esters on geochemical cycles in fresh-water hydrosoil. Transactions of the Missouri Academy of Science. Vol 10 and 11, p 296

21. Larsson P, Thurén A, Gahnström G (1986) Environ Poll (Series A) 42:22322. Aranda JM, O’Connor GA, Eiceman GA (1989) J Environ Qual 18:4523. Difenbach R (1998a) Bestimmung der Auswirkungen von Vestinol AH auf Regenwürmer

(Eisenia foetida), RW-71, Huls Infracor GmbH, Marl, Germany, 15 pp

An Assessment of the Potential Environmental Risks Posed by Phthalates 345

24. Difenbach R (1998b) Bestimmung der Auswirkungen von Vestinol AH auf das WachstumTerrestrischer Planzen, Huls Infracor GmbH, PF-56, Marl, Germany, 20 pp

25. Stanley RD, Tapp JF, Williams BRH (1982) An assessment of ecotoxicological test meth-ods: part VIII. The effect of nine chemicals on the growth of Avena sativa and Brassicarapa, Brixham Laboratory, Imperial Chemical Industries, Report BL/A/2164, 26 pp

26. Woin P, Larsson P (1987) Bull Environ Contam Toxicol 38:22027. Brown D,Thompson RS,Stewart KM,Croudace CP,Gillings E (1996) Chemosphere 32:217728. Wennberg L, Parkman H, Remberger M,Viktor T,Williams C (1997) The influence of sed-

iment-associated phthalate esters (DEHP and DIDP) on hatching and survival of themoorfrog, Rana arvalis. IVL, Box 21060, S-100 31 Stockholm, Sweden. IVL report B1260,26 pp

29. Solyom P, Remberger M, Viktor T (2001) Further Investigations on the Influence ofSediment-Associated Phthalate Esters (DEHP and DINP) on Hatching and Survival of theMoorfrog Rana arvalis, IVL Swedish Environmental Research Institute Report No.A20173, 42 pp

30. Perez KT, Davey EW, Lackie NF, Morrison GE, Murphy PG, Soper AE,Windslow DL (1983)Environmental assessment of a phthalate ester, Di(2-ethylhexyl) phthalate (DEHP),derived from a marine microcosm. Aquatic Toxicology and Hazard Assessment: SixthSymposium,ASTM STP 802,WE Bishop, RD Cardwell, and BB Heidolph (Eds),AmericanSociety for Testing and Materials, Philadelphia, pp 180–191. 1983 (ASTM Special techni-cal publ. 0066–0558:802)

31. Exxon Biomedical Sciences, Inc. (2002) Microbial Respiration Inhibition in Soil. Unpub-lished draft report 199694B. May 7, 1999

32. Exxon Biomedical Sciences, Inc. (1996a) “Seed germination limit test with rye grass andlettuce, East Millstone, NJ. Unpublished report No 199674

33. Exxon Biomedical Sciences, Inc. (1996b) “Seed germination test with lettuce, East Mill-stone, NJ. Unpublished report No 199674A

34. Exxon Biomedical Sciences, Inc. (2000) Lettuce, Seed Germination and Growth Test,Unpublished draft report 199674C, June 23, 1999

35. Exxon Biomedical Sciences, Inc. (1996) “Earthworm limit test,” East Millstone, NJ. Un-published report No 1996 92

36. Larsson P, Thurén T (1987) Environ Toxicol Chem 6:41737. Van Wezel AP, van Vaardingen P, Posthumus R, Crommentuijn GH, Sijm DT (2000) Eco-

tox Environ Saf 46:30538. Yeung P (1990) A method for measuring water repellent soil, Proceedings of the 27th

Annual Alberta Soil Science Workshop, Edmonton, Alberta, Canada, pp 59–6439. Van Gestel CAM, Ma M (1990) Chemosphere 21:102340. Di Toro DM, Zarba CS, Hansen DJ, Berry WJ, Swartz RC, Cowan CE, Pavlou SP, Allen HE,

Thomas NA, Paquin PR (1991) Environ Toxicol Chem 10:154141. van Leeuwen CJ, Van der Zandt PTJ, Aldenberg T, Verhaar HJM, Hermens JLM (1992)

Environ Toxicol Chem 11:26742. Belfroid AC, Seinen W, Gestel CAM, Hermens JLH, van Leeuwen CJ (1995) Environ Sci

Pollut Res 2 :543. EC. 1996. Commission of the European Communities. Technical Guidance Documents in

Support of the Commission Directive 93/67/EEC on risk assessment for new substancesand the Commission Regulation (EC) No 1488/94 on risk assessment for existing sub-stances. Office of Publications of the European Communities, Luxembourg

44. USEPA (1997) The incidence and seventy of sediment contamination in surface waters ofthe United States, volume 1: National Sediment Quality Survey, EPA 823/R-97/006,Wash-ington DC, USA

45. Seth R, Mackay D, Muncke J (1999) Environ Sci Technol 33:239046. Cousins IT, Mackay D (2000) Chemosphere 41:138947. Barrick R, Becker S, Brown L, Beller H, Pastouk R (1988) Sediment quality values re-

finement: 1988 Update and evaluation of Puget Sound AET, volume 1, Puget Sound estu-ary program, office of Puget Sound, Seattle, WA, USA

346 T.F. Parkerton and C.A. Staples

48. Florida Department of Environmental Protection (1994) Approach to the assessment ofsediment quality in Florida coastal waters, volume 1, Development and evaluation ofsediment quality assessment guidelines, office of water policy, Talahasse, FL, USA

49. Von Stackelberg, Menzie CA (2002) Environ Toxicol Chem 21:46650. David RM, McKee RH, Butala JH, Barter RA, Kayser M (2001) In: Bingham E, Cohrosen

B, Powell CH (eds) Esters of aromatic mono-, di-, and tricarboxylic acids, aromatic diacidsand di-, tri-, or polyalcohols, Pattys Toxicology, 5th edn, vol 6, pp 635–932

51. Lehman, AJ (1955) Insect repellents. Food and Drug Officials of the United States, Quar-terly Bulletin 19, pp 87–99

52. Field AE et al. (1993) Teratology 48:3353. Patel R, Wolfe G, Rouselle S, Cherian G, Ko K, Moore R, Bishop J, Chapin R (2001) The

Toxicologist 60:38554. Nagao T, Ohta R, Marumo H, Shindo T, Yoshimura S, Ono H (2000) Reproductive Toxi-

cology 14:51355. Schilling K, Gembart C, Hellwig J (2001) Di-2-ethylhexyl phthalate – Two-generation re-

production toxicity study in Wistar rats. Continuous Dietary Administration. Unpublishedlaboratory report

56. Waterman S, Keller L, Trimmer G, Freeman J, Nikiforov A, Harris S, Nicolich M, McKee R(2000) Reproductive Toxicology 14:21

57. Hushka L, Waterman S, Keller L, Trimmer G, Freeman J, Ambroso J, Nicolich M, McKee R(2001) Reproductive Toxicology 15:153

58. Hill EF, Heath RG, Spann JW,Williams JD (1975) Lethal dietary toxicities of environmentalpollutants to birds, Washington DC, US Fish and Wildlife Service, Special Scientific Report Wildlife No. 191, pp 61

59. O’Shea TJ, Stafford CJ (1980) Bull Environ Contam Toxicol 25:34560. Wood DL, Bitman J (1984) Poultry Sci 36:46961. Ishida M, Suyama K, Adachi S (1981) J Agriculture Food Chem 29:7262. Duke LD, Taggart M (2000) Environ Toxicol Chem 19:166863. Nagy KA (1987) Ecol Monogn 57:11164. Gobas FAPC, Mackintosh CE, Webster G, Ikonomou M, Parkerton TF, Robillard K (2002)

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs, Chapter submitted toSpringer-Verlag Environmental Chemistry Handbook Volume for Phthalate Esters

65. ECB (2002) Draft Revision of the Technical Guidance 7 Document on Risk Assessment. En-vironmental Risk Assessment, Chapter 3, European Chemicals Bureau, Ispra, Italy, 211 pp

66. Rhind SM, Smith A, Kyle CE, Telfer G, Martin G, Duff E, Mayes RW (2002) J Environ Monit4 :142

67. Eisenreich SJ, Looney BB, Thornton JD (1981) Environ Sci Technol 15:3068. Thuren A, Larsson P (1990) Environ Sci Technol 24:55469. Vikelsoe J, Thomsen M, Johansen E (1998) Sources of phthalates and nonylphenol in

municipal waste water: a study in a local environment, NERU Technical Report No 225,Ministry of Environment and Energy, National Environmental Research Institute, Ros-kilde, Denmark, 52 pp

70. Schleyer R, Renner I, Mühlhausen D (1991) Immisionsbelastung (1991) – Konsequenzenfür die Grundwasserqualität BGA

71. Furtmann K (1993) Phthalates in the Aquatic Environment, PhD dissertation, Regional Water and Waste Water Authority, Nordrhein-Westfalen

72. USEPA (1995) Process Design Manual: Land Application of Sewage Sludge and Domes-tic Seepage, Office of Research and Development, US Environmental Protection Agency,Washington, DC, EPA 625/K-95/001, 301 pp

73. EC (2001) Organic Contaminants in Sewage Sludge for Agricultural Use, European Com-mission Joint Research Center, Institute for Environment and Sustainability and WasteUnit, 73 pp

74. Vikelsoe J, Thomsen M, Carlsen L (2002) Sci Tot Environ (in press)75. Weber MD, Nichols JA (1995) “Organic and Metal Contaminants in Canadian Municipal

Sludges and a Sludge Compost.”Wastewater Technology Centre, Burlington, ON, Canada

An Assessment of the Potential Environmental Risks Posed by Phthalates 347

76. Kristensen P, Torslov J, Samsoe-Peterson L, Rasmussen JO (1996) “Sludge and Waste Prod-ucts to be used on Arable Land.” MilJoProject No. 328. MilJo-09, Energiministeriet,Miljostyvelsen. Copenhagen, Denmark,165 pp

77. Research Institute for Chromatography (RIC) (2000) Kortrijk, Belgium. RIC reportECPI\2000–05–0007

78. Schnaak W, John T (1994) “Investigations of organic pollutants in sewage sludge and theecotoxicological evaluation of sewage sludge amended soils.” Prepared by Fraunhofer – Institut for Umweltchemie und Okotoxikologie for Landesumweitamt Brandenburg,Germany

79. Fromme H, Kuchler T, Otto T, Pilz K, Muller J, Wenzel A (2002) Wat Res 36:142980. Kolb M, Welte K, Mettenleiter S, Trinkmann A, Bestimmung von Phthalsäureestern in

Klärschlämmen mittels GC/MS, Wasser and Boden 49:5/199781. Paulsrud BA, Wien and Nedland KT (2000) A Survey of Toxic Organics in Norwegian

Sewage Sludge, Compost and Manure. – Aquateam, Norwegian Water Technology CentreASOSLO

82. UMEG (1999) Bodenzustandsbericht Grossraum Stuttgart. – Herausgeber: Ministeriumfür Umwelt und Verkehr, Baden-Württemberg, 107S

83. AL Control Biochem Laboratoria (1999) The Analysis Of Phthalates In Soil And Sediment,Hoogvliet, The Netherlands, 11 pp + appendices

84. Vethaak AD, Rijs GBJ, Schrap SM, Ruiter H, Gerristen A, Lahr J (2002) Estrogens and Xeno-estrogens in the Aquatic Environment of the Netherlands, Occurrence, Potency and Biological Effects, Directorate-General for Public Works and Waste Management,RIZA/RIKZ Report no 2002.001, 293 pp

85. Norwegian Institute for Water Research (NIVA), “Occurrence of Phthalates and Organ-otins in Sediments and Water in Norway,” Report SNO 3552–96

86. Parkman H, Remberger M (1996) “Phthalates in Water and Sediments in Major Cities andRemote Lakes in Sweden,” IVL, Swedish Environmental Research Institute, Stockholm,Sweden

87. Tan GH (1995) Bull Environ Contam Toxicol 54:17188. Lopes TJ, Furlong ET (2001) Environ Toxicol Chem 20:72789. Garrett CL (2002) Phthalate Esters in Harbour Areas of South Coastal British Columbia.

Environment Canada, Environmental Protection Branch Pacific and Yukon Region, Re-gional Program Report, 100 pp

90. Mackintosh CE, Gobas FAPC, Ikonomou M (2002) Distribution of Phthalate Esters in a Marine Food Web, Report Prepared for the American Chemistry Council, Washington,DC, 209 pp

91. Vitali M, Guidotti M, Macilenti G, Cremisini C (1997) Env Int 23:33792. Parkman H, Remberger M (1995) “Phthalates in Swedish Sediments,” IVL, Swedish Envi-

ronmental Research Institute, Stockholm, Sweden93. Japan Environment Agency (JAE) (1999) “Results of measurements of endocrine dis-

rupting chemicals in the environment”, Tokyo, Japan94. Ministry of Construction (MOC) (1999a) “Fiscal 1999 Fact-Finding Study of Endocrine

Disrupting Chemical Substances in the Water Environment (Spring and Summer Stud-ies)” River Bureau, Sewerage and Sewage Purification Department, City Bureau, Ministryof Construction, Japan. November

95. Ministry of Construction (MOC) (1999b) “Fiscal 1999 Fact-Finding Studies Regarding Endocrine Disrupting Chemical Substances in the Water Environment (Spring Study)”River Bureau, Sewerage and Sewage Purification Department, City Bureau, Ministry ofConstruction, Japan

96. Ministry of Construction (MOC) (1999c) “Fiscal 1999 Results of the Fact-Finding Stud-ies Regarding Endocrine Disrupting Chemical Substances in the Water Environment(Summer Studies)”River Bureau, Sewerage and Sewage Purification Department, City Bu-reau, Ministry of Construction, Japan

97. Vikelsoe J, Thomsen M, Johansen E, Carlsen L (1999) “Phthalates and Nonylphenols inSoil. A Field Study of Different Soil Profiles.” National Environmental Research Institute,Denmark. 128 pp. NERI Technical Report No. 268

348 T.F. Parkerton and C.A. Staples

98. Vondracek J, Machala M, Minksova K, Blaha L, Murk AJ, Kozubik A, Hofmanova J,Hilscherova K, Ulrich R, Ciganek M, Neca J, Svrckova D, Holoubek I (2001) Environ Tox-icol Chem 20:1499

99. Vikelsoe J, Fauser P, Sorensen PB, Carlsen L (2001) “Phthalates and Nonylphenols inRoskilde Fjord. A Field Study and Mathematical Modelling of Transport and Fate in Water and Sediment. The Aquatic Environment.” National Environmental Research In-stitute, Roskilde. 106 pp. NERI Technical Report No. 339

100. Long J, House W, Parker A, Rae J (1998) Sci Tot Environ 210/211:229101. Boutrup S, Erichsen PC, Wiggers L, Jensen A (1998) Miljöfremmende stoffer in Århus

Amt – fase 2 og 3, 1997–1998. Århus Amt Natur og Miljökontoret, Lyseng Allé,DK-8270 Höjbjerg, Denmark

102. Fries G (1996) Sci Tot Environ 185:93103. Cousins I, MacKay D (2003) Multimedia Mass Balance Modeling of two Phthalate Esters

Using the Regional Population-Based Model (RPm), Chapter submitted to Springer-Ver-lag Environmental Chemistry Handbook Volume for Phthalate Esters

104. Sverdrup L, Nielsen ET, Krogh PH (2002) Environ Sci Technol 36:2429105. Alexander M (1995) Environ Sci Technol 29:2713106. Beck AJ, Johnson DL, Jones KC (1996) Sci Tot Environ 185:125107. Kraaij RH, Tolls J, Sijm D, Cornelissen G, Heikens A, Belfroid A (2002) Environ Toxicol

Chem 21:752108. Madsen PL, Thyme JB, Moldrup P, Roslev P (1999) Environ Sci Technol 33:2601109. Peterson DR, Staples CA (2003) Degradation of Phthalate Esters in the Environment,

Chapter submitted to Springer-Verlag Environmental Chemistry Handbook Volume forPhthalate Esters

110. Tienpont B, David F, Vanwalleghem F, Sandra P (2001) J Chromatogr A 911:235

An Assessment of the Potential Environmental Risks Posed by Phthalates 349

© Springer-Verlag Berlin Heidelberg 2003

Analytical Methods Review

Frank David 1 · Pat Sandra 2 · Bart Tienpont 2 · Freddy Vanwalleghem 3

Michael Ikonomou 4

1 Research Institute for Chromatography, Pres. Kennedypark 20, 8500 Kortrijk, BelgiumE-mail: frank.david@richrom.com

2 Laboratory of Organic Chemistry, University of Gent, Krijgslaan 281 S4, 9000 Gent, Belgium3 Proviron, NV, Stationstraat 123, 8400 Oostende, Belgium4 Contaminants Science Section, Institute for Ocean Sciences, BC, Canada

In this chapter, an overview is given of the analytical methods developed for the determina-tion of phthalates in different environmental samples, including water, soil, sediments, sludges,air and biota. The analytical methods based on gas chromatography coupled to mass spec-troscopy (GC-MS) and liquid chromatography coupled to mass spectroscopy (LC-MS) are described and typical applications are presented. Dedicated sample preparation methods aredescribed and compared. Special attention is paid to contamination problems.

Keywords. Phthalates, GC-MS, LC-MS, Contamination, Sample preparation

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2 Analysis of Phthalates . . . . . . . . . . . . . . . . . . . . . . . . 12

2.1 Phthalate Analysis by Capillary Gas Chromatography . . . . . . . 122.1.1 Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.1.2 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.1.3 Mixed Isomer Phthalates . . . . . . . . . . . . . . . . . . . . . . . 172.1.4 Chemical Ionisation Mass Spectrometry . . . . . . . . . . . . . . 172.1.5 Quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.2 Phthalate Analysis by HPLC-MS . . . . . . . . . . . . . . . . . . . 192.3 Analysis of Phthalic Acid Mono-Esters . . . . . . . . . . . . . . . 222.3.1 Derivatisation and GC-MS . . . . . . . . . . . . . . . . . . . . . . 222.3.2 Analysis of Phthalic Acid Mono-Esters by HPLC-MS . . . . . . . . 23

3 The Blank Problem . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.1 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263.2 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . 273.3 Chromatographic Analysis . . . . . . . . . . . . . . . . . . . . . . 27

4 Analysis of Phthalates in Water Samples . . . . . . . . . . . . . . 28

4.1 Liquid-Liquid Extraction (LLE) . . . . . . . . . . . . . . . . . . . 284.2 Solid-Phase Extraction (SPE) . . . . . . . . . . . . . . . . . . . . 294.3 Solventless Extraction Methods (SPME and SBSE) . . . . . . . . . 31

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 9–56DOI 10.1007/b11461

5 Analysis of Phthalates in Sediments, Soils and Sewage Sludges . . 33

5.1 Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335.2 Clean-Up Procedures . . . . . . . . . . . . . . . . . . . . . . . . . 355.3 Determination of Phthalates in Sewage Sludge . . . . . . . . . . . 35

6 Analysis of Phthalates in Air . . . . . . . . . . . . . . . . . . . . . 35

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356.2 Analysis of Total Phthalate Concentrations in Air . . . . . . . . . . 376.3 Passive Sorptive Sampling of Phthalates in Air . . . . . . . . . . . 416.4 Dust Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

7 Analysis of Phthalates in Biota (Vegetation, Milk, Fish) . . . . . . 46

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467.2 Analytical Procedure for the Determination of Phthalates

in Vegetation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477.3 Analytical Procedure for the Determination of Phthalates in Milk

or Edible Oils and Fat . . . . . . . . . . . . . . . . . . . . . . . . . 487.4 Analytical Procedure for the Determination of Phthalates in Fish . 49

8 Sample Preparation Methods for Phthalic Acid Mono-Esters . . . 50

9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

1Introduction

Phthalic acid diesters (PAEs) are widely spread in the environment. During thepast years, interest in monitoring this class of industrial chemicals has increased.Several methods have been developed for their determination in a very broadrange of matrices, including water (drinking water, surface water, waste water),soil, sediment, sludge, dust, indoor and outdoor air, and biota (vegetation, milk,fish, etc.).

In this chapter an overview is given of different techniques used for the analy-sis of phthalates in these matrices. Particular attention is paid to the determina-tion of the most relevant phthalates in terms of their environmental distribution.These phthalates include diisobutyl phthalate (DiBP), dibutyl phthalate (DBP),butylbenzyl phthalate (BBzP), bis(2-ethylhexyl) phthalate (DEHP), diisononylphthalate (DiNP) and diisodecyl phthalate (DiDP). Although the last two phthalates have industrial importance, they are often not included in genericmethods for environmental analysis, such as EPA or ISO methods [1, 2]. The mainreason for this is that these phthalates are not pure chemicals, but mixtures ofseveral isomers. Consequently, the analysis is more difficult. Special attention willbe paid to their analysis and alternative methods will be discussed.An overviewof the most important phthalates, including molecular formula, abbreviation,molecular weight and CAS number, is given in Table 1.

10 F. David et al.

Table 1. List of phthalates

Phthalates Formula Abbreviation Molar mass CAS No

Single isomer phthalatesDimethyl phthalate b C10H10O4 DM b 194.2 131-11-3Diethyl phthalate b C12H14O4 DE b 222.4 84-66-2Diallyl phthalate C14H14O4 DalP 246.3 131-17-9Di(1-methylethyl) phthalate C14H18O4 DiPP 250.3 605-45-8

(diisopropyl phthalate)Dipropyl phthalate C14H18O4 DPP 250.3 131-16-8Butyl-2-methylpropyl phthalate C16H22O4 BMPP 278.4 17851-53-5Dibutyl phthalate b C16H22O4 DBP b 278.4 84-74-2Di(2-methylpropyl) phthalate C16H22O4 DiBP b 278.4 84-69-5

(diisobutyl phthalate) b

Di(2-ethoxyethyl) phthalate C16H22O6 DeoEP 310.4 605-54-9Dicyclopentyl phthalate C18H22O4 DCPeP 302.4 18699-38-2Butylcyclohexyl phthalate C18H24O4 BCHP 304.4 84-64-0Di(3-methylbutyl) phthalate C18H26O4 DMBP 306.4 605-50-5

(diisopentyl phthalate)Dipentyl phthalate C18H26O4 DPeP 306.4 131-18-0Butylbenzyl phthalate b C19H20O4 BBzP b 312.4 85-68-7Diphenyl phthalate C20H14O4 DPhP 318.3 84-62-8Dicyclohexyl phthalate C20H26O4 DCHP 330.4 84-61-7Butyl 2-ethylhexyl phthalate C20H30O4 BEHP 334.5 85-69-8Butyloctyl phthalate C20H30O4 BOP 334.5 84-78-6Di(2-ethylbutyl) phthalate C20H30O4 DEBP 334.5 7299-89-0Dihexyl phtalate C20H30O4 DHP 334.5 84-75-3Dibenzyl phthalate C22H18O4 DBzP 346.3 523-31-9Dimethyl cyclohexyl phthalate C22H30O4 DMCHP 358.5 27987-25-3Butyldecyl phthalate C22H34O4 BDcP 362.6 89-19-0Diheptyl phthalate C22H34O4 DHpP 362.5 3648-21-3Di(5-methylhexyl) phthalate a C22H34O4 DMHP 362.5 41451-28-9Benzyl 2-ethylhexyl phthalate C23H28O4 BzEHP 368.6 18750-05-5Di(2-ethylhexyl) phthalate b C24H38O4 DEHP b 390.6 117-81-7Dioctyl phthalate C24H38O4 DOP 390.6 117-84-0Hexyldecyl phthalate C24H38O4 HDcP 390.6 25724-58-7Octyldecyl phthalate C26H26O4 ODcP 418.6 119-07-3Di(7-methyloctyl) phthalate a C26H42O4 DMOP 418.6 28553-12-0Dinonyl phthalate C26H42O4 DNP 418.6 84-76-4Di(3,3,5-trimethylhexyl) phthalate C26H42O4 DTMHP 418.6 4628-60-8Didecyl phthalate C28H46O4 DDcP 446.7 84-77-5Di(8-methylnonyl) phthalate a C28H46O4 DMNP 446.7 89-16-17Diundecyl phthalate C30H50O4 DUP 474.7 3648-20-2Didodecyl phthalate C32H54O4 DDDP 502.8 2438-90-8

Isomeric mixture phthalatesDiisoheptyl phthalate C22H34O4 DiHpP 362.5Diisononyl phthalate b C26H42O4 DiNP b 418.6Diisodecyl phthalate b C28H46O4 DiDP b 446.7

Phthalic acid monoestersMonomethyl phthalate C9H8O4 MMP 180.2 4376-18-5Monoethyl phthalate C10H10O4 MEP 194.2 2306-33-4Monobutyl phthalate C12H14O4 MBP 222.2 131-70-4Mono(2-ethylhexyl) phthalate C16H22O4 MEHP 278.4 4376-20-9

a These compounds are present in the isomeric mixtures of DiHpP, DiNP and DiDP respec-tively.

b These are the most important phthalates.

The analysis of phthalic acid esters is mostly performed by gas chromatogra-phy (GC). Phthalates ranging from the most volatile dimethyl phthalate (DMP)to didodecyl phthalate can be analysed by capillary gas chromatography (CGC)as they are sufficiently volatile and thermostable. High-pressure liquid chro-matography (HPLC) can be used as an alternative technique and is especially use-ful for the analysis of isomeric mixtures. The possibilities of CGC and HPLC forthe analysis of phthalates are discussed below.

Recent environmental and toxicological concerns have also focused on the de-termination of phthalate metabolites. The primary metabolites of the diesters arethe monoesters, formed by the hydrolysis of one ester function to give phthalicacid monoesters. These monoesters contain a free acid function and require de-rivatisation before GC analysis. These methods are also presented.

The major problem in phthalate analysis is, however, not the analysis itself, butthe risk of contamination. Contamination can occur in every stage of the wholeanalytical procedure including sampling, sample preparation (extraction, clean-up, concentration) and analysis. An overview is given below of possible sourcesof contamination (instrument, solvents, air etc.) and recommendations are givenon how to avoid or control the contamination problem.

Finally, an overview is given of different sample preparation techniques that are used to extract and isolate phthalates from different environmental matrices.

2Analysis of Phthalates

2.1Phthalate Analysis by Capillary Gas Chromatography

Capillary gas chromatography is the most widely used analytical technique forthe determination of phthalates. Different capillary columns, inlet systems anddetectors are used. For the gas chromatographic separation, capillary columnscoated with apolar stationary phases (polydimethylsiloxanes or polymethyl-phenylsiloxanes) are preferred, since they provide sufficient resolution, a highermaximum operating temperature and a lower bleeding than columns coated with polar stationary phases such as poly(ethylene glycol)s (Wax- columns) orcyanopropyl stationary phases. An example of the separation of the most im-portant phthalates is given in Fig. 1. Peak identification is given in Table 2. Thisseparation is performed on a 30 m ¥ 0.25 mm i.d. capillary column coated witha 0.25 µm film of 5% phenyl-95% methyl polysiloxane (HP-5MS). The analyticalconditions are summarised in Table 3. These conditions are typical for the analy-sis of phthalates by CGC-MS and result in sufficient resolution of the most im-portant phthalates. These conditions are also a good compromise between reso-lution and speed of analysis. Longer columns or slower temperature programscan give higher resolution, but for most environmental applications this is notneeded. As demonstrated in Fig. 1, a boiling point separation is obtained on theapolar column.Around 22–24 min and at the end of the chromatogram, two un-resolved groups of peaks are observed. These groups of peaks result from the

12 F. David et al.

mixed isomer phthalates diisoheptyl phthalate, diisononyl phthalate and di-isodecyl phthalate. Long capillary columns and slow temperature programs cangive better resolution, but no column is able to fully separate all isomers. There-fore, a rather fast oven program is preferred, since this results in compression ofthe groups of C7-, C9- and C10-isomers and a better signal-to-noise ratio. Thesensitivity is therefore higher than with slow temperature programs, in whichbroader isomer distribution patterns are obtained.

Analytical Methods Review 13

Fig.

1.G

C-M

S ch

rom

atog

ram

ofa

pht

hala

te s

tand

ard

mix

ture

(ide

ntifi

cati

on:s

ee T

able

2,co

ndit

ions

:see

Tab

le3)

Tim

e

2.1.1Injection

For sample introduction, cool on-column injection, whereby the liquid sample isdirectly introduced into the capillary column, is the best choice in terms of solutedegradation and discrimination. Moreover, in the case of phthalate analysis, coolon-column injection also results in the lowest possible instrumental phthalatecontamination. The major limitation of cool on-column injection is, however, thecontamination of the capillary column with non-volatile material. This non-volatile material can be present in extracts from complex samples, such as waste-water, sludges, food or biota. In practice, cool on-column injection is therefore re-stricted to the analysis of clean samples, such as extracts from drinking water (orsurface water). For routine analyses of other sample extracts, splitless injection

14 F. David et al.

Table 2. Peak identification for Fig. 1 and ions monitored in GC-MS analysis a

Peak Nr (Fig. 1) Phthalate Abbreviation Target ion Qualifier ion

1 Dimethyl phthalate DMP 163 1942 Diethyl phthalate DEP 149 1773 Diisobutyl phthalate DiBP 149 2234 Dibutyl phthalate DBP 149 2235a, 5b, 5c Diisopentyl phthalate (isomers) DPeP 149 2376 Butylbenzyl phthalate BBzP 149 91, 2067 Diisoheptyl phthalate DiHpP 149 2658 Dicyclohexyl phthalate DCHP 149 167, 2499 Di(2-ethylhexyl) phthalate DEHP 149 279

10 Diisononyl phthalate DiNP 149 29311 Diisodecyl phthalate DiDP 149 307

Internal standardsNot included D4-ring-dibutyl phthalate D4-DBP 153 –Not included D4-ring-dioctyl phthalate D4-DOP 153 –Not included Diallyl phthalate DalP 149 189

a For high-resolution mass spectroscopy, the monitored masses are 149.0239 (phthalates) and153.0490 for the d4-labelled phthalates [3, 4].

Table 3. Analytical conditions for GC-EIMS

Instrument Agilent 6890 GC – Agilent 5973 MSDColumn 30 m ¥ 0.25 mm i.d.¥ 0.25 µm HP-5MS

(5% phenyl–95% methyl silicone)Carrier gas 1 mL min–1 helium, constant flow

(50 kPa at 50 °C)Injection 1 µL splitless, 280 °C, 0.75 min purge delayOven temperature 50 °C–1 min–10 °C min–1 –320 °C–2 minDetection MS in SIM mode, Monitored ions: see Table 2

is the method of choice. Modern split/splitless inlets in combination with new au-tosamplers minimise sample degradation and discrimination. Splitless injectioncan be used for the injection of contaminated sample extracts, since most of thenon-volatile contaminants remains in the liner. Recently, also programmed tem-perature vaporising (PTV) inlets have been used. PTV inlets can be compared tosplit/splitless inlets, but injection is performed at a relatively low temperature(around the boiling point of solvent), followed by a rapid heating (typically10 °C s–1) to 250–300 °C for the vaporisation of the sample. The main advantageof PTV injection is the possibility of performing large volume injection. Injec-tions up to 1000 µL can be performed by using the solvent vent mode. Duringsample introduction and an initial hold time at low temperature, the solvent isevaporated and vented via the split vent (at a high split ratio).After solvent evac-uation, the split vent is closed (splitless mode) and the inlet temperature is in-creased to vaporise the solutes.

2.1.2Detection

Detection of phthalates can be done by flame ionisation detection (FID), electroncapture detection (ECD) or mass spectrometry (MS). GC-FID is not frequentlyused, since the detector is not specific for phthalates. Some official methods (USEPA methods 606 and 8060) describe the use of ECD for phthalate analysis.Although ECD detectors are relative sensitive for phthalates, the specificity is restricted, since ECD responds much more sensitively towards halogenated compounds. The most important detector for phthalate analysis is mass spec-trometric detection. All types of MS analysers, including quadrupole analysers,triple quadrupole analysers, ion traps and magnetic sector instruments havebeen used. Benchtop quadrupole systems are generally preferred due to their ro-bustness, stability, linear dynamic range and low cost. Ion trap detectors havesimilar sensitivity, but the dynamic range is often lower and at higher solute con-centration, as is often encountered in phthalate analysis, spectral deviations havebeen observed. High-resolution magnetic sector instruments are also used [3, 4],but the high resolution is not required for most applications.

For most applications, a benchtop single quadrupole system provides suf-ficient sensitivity in selected ion monitoring (SIM) mode in electron impact (EI) ionisation. Typical sensitivities obtained on state-of-the-art instruments are in the low picogram (1–10 pg) range for the single isomer phthalates.Due to the presence of several isomers, the sensitivity for isomeric mixture phthalates (DiHpP, DiNP, DiDP) is typically one order of magnitude lower(10–50 pg).

The mass spectra for all phthalates (except dimethyl phthalate) are very sim-ilar. The mass spectrum of DEHP is shown in Fig. 2. The main ion is at m/z 149,resulting from fragmentation with loss of the alkyl ester groups and a furan ringformation (Fig. 3). Besides the most abundant ion at m/z 149, the spectrum israther poor. The molecular ion (m/z 390) is not detected. The second most im-portant ion is at m/z 279. This ion results from fragmentation by loss of one alkylgroup. This ion is therefore different for each phthalate and can be used to dif-

Analytical Methods Review 15

16 F. David et al.

Fig.

2.El

ectr

on im

pact

mas

s sp

ectr

um o

fDEH

P

ferentiate phthalates. Further fragmentation of the ion at m/z 279 results in ionat m/z 167. A similar fragmentation pattern is found for the other phthalates,except for dimethyl phthalate. In the mass spectrum of dimethyl phthalate, themolecular ion is detected at m/z 194. The most abundant ion is at m/z 163, cor-responding to the loss of a methoxy group (M–31). A list of the most importantions that can be used for mass spectroscopic detection of phthalates is includedin Table 2.

2.1.3Mixed Isomer Phthalates

A specific problem is the quantitative determination of the mixed isomer phthalates. No capillary column is able to separate all isomers, or to separate allC9-isomers from the C10-isomers. Therefore, extracted ion chromatograms onthe specific ions m/z 293 (DiNP) and m/z 307 (DiDP) can be used. This is illus-trated in Fig. 4. Although the groups of peaks overlap, the extracted ion chro-matograms can differentiate between both compounds, and the sum of the peakareas measured on each extracted ion trace can be used for quantification.

2.1.4Chemical Ionisation Mass Spectrometry

As an alternative to electron impact ionisation, phthalates can also be analysedby CGC-MS by using chemical ionisation [5]. Positive ion chemical ionisation results in a softer ionisation and therefore the fragmentation is reduced and therelative abundance of high-mass ions is increased. This is illustrated for dibutylphthalate in Fig. 5. With methane as the reagent gas, the fragmentation is stillstrong and the most abundant ion is still at m/z 149. However, now a pseudo-mol-ecular ion at M+1 (m/z 279) is clearly detected.With ammonia as the reagent gas,the possibilities of positive ion chemical ionisation are clearly demonstrated. Now

Analytical Methods Review 17

Fig. 3. Mass fragmentation of DEHP

Fig. 4. Extracted ion chromatograms for DEHP (m/z 279), DiNP (m/z 293) and DiDP (m/z 307)

Fig. 5 A, B. PCI mass spectra of dibutyl phthalate (A methane reagent gas, B ammonia reagentgas). Published with permission from Ref. [5]

Abundance

Time Æ

18 F. David et al.

the most abundant ion is the pseudo-molecular ion at m/z 279 (M+1). Positivechemical ionisation is therefore an interesting alternative, especially for the de-termination of the mixed isomer phthalates.

2.1.5Quantification

The quantification of phthalates can be done by external or internal standard calibration. Internal standard calibration is preferred, since the internal standardcan be used to correct for variations in injection, analysis (column performance)and condition of the mass spectrometer. The best internal standards for phthalate analysis by mass spectrometric detection are isotopically labelled phthalates. Both C13-labelled or ring D4-labelled (deuterated phthalates) arecommercially available. If added before sample preparation, these internal stan-dards can also be used to correct for extraction efficiency. The most importantinternal standards used in the analysis of phthalates are included in Table 2.

2.2Phthalate Analysis by HPLC-MS

Although CGC-MS is widely used, LC-MS can provide interesting alternatives,especially in the analysis of the isomeric mixture phthalates and monoesters.An analytical method based on LC-electrospray ionisation mass spectrometry forthe determination of phthalates in marine water, sediment and biota samples waspresented by Lin et al. [6]. The LC-MS instrument is operated in the positive ionmonitoring mode. The separation was achieved on a 25 cm ¥ 2 mm i.d. C8 col-umn with a mobile phase of 90% methanol and 10% 0.5 mM sodium acetate. Themobile phase flow rate was 0.22 mL min–1 and a 1/10 flow split ratio (0.022 mLmin–1 to MS) was used before the mass spectrometer (VG Quattro triple quadru-pole). The C8 (octyl-silicagel) column is preferred because the isomeric mixtureselute as single, well-defined and narrower peaks than on a classical C18 (oc-tadecyl silicagel, ODS) column. Phthalate esters are monitored as sodiated mol-ecular adducts (M+Na, at m/z MW+23) under SIM conditions. The stable iso-tope dilution method was used for quantification. Compared to GC-MS analysis,a lower sensitivity is obtained, but the LC-MS approach was found to give the fol-lowing advantages: superior selectivity with molecular weight information forthe isomeric mixtures, more reliable quantification of the phthalate ester iso-meric mixtures (such as DiHpP, DiNP and DiDP), simpler clean-up proceduresand shorter analysis times.

An example of a separation of mixed isomer phthalates by LC-MS is given inFig. 6. This separation was obtained under slightly different conditions [7]. Theanalytical conditions are summarised in Table 4. The separation of the isomericmixtures obtained on the cyanopropyl silicagel (CN) column is very good. On thiscolumn, the isomeric mixtures are also separated as well-defined peaks, allowinggood quantification. In comparison to GC-MS, less resolution is obtained, but forquantification this is an advantage. The mass spectra obtained by electrosprayionisation MS for DEHP and DiDP are given in Fig. 7. The most abundant ions

Analytical Methods Review 19

20 F. David et al.

Fig.

6.LC

-MS

anal

ysis

ofi

som

eric

mix

ture

pht

hala

tes b

y el

ectr

ospr

ay io

nisa

tion

.(Io

ns m

onito

red:

m/z

363

for D

iHpP

,m/z

391

for D

EHP,

m/z

419

for D

iNP

and

m/z

447

for

DiD

P) (c

ondi

tion

s:se

e Ta

ble

4)

Table 4. Analytical conditions for LC-ESMS analysis of isomeric mixture phthalates

Instrument Agilent 1100 LC–Agilent 1100 SL MSDColumn 25 cm L¥4.6 mm i.d.¥5 µm Alltima CN (cyanopropyl silica)Mobile Phase A. 0.5% ammonium acetate in water

B. methanolGradient 60% B for 5 min, to 100% B at 35 minFlow 1 mL min–1

Injection 10 µL Detection Ionisation mode: API-ES (electrospray)

Polarity: positiveMonitored Ions: M+1 (363 (DiHP), 391 (DEHP), 394 (IS: d4-DEHP),419 (DiNP), 447 (DiDP))

Fig. 7. Positive ion electrospray spectra obtained for DEHP and DiDP (conditions Table 4)

DEHP

DiDP

Analytical Methods Review 21

are respectively ions at m/z 391 and m/z 447, corresponding to the M+H adducts(at m/z M+1). In addition, the acetate-adducts are detected (as M+59, respec-tively at m/z 449 and m/z 505). In this method, ammonium acetate is used asbuffer in the mobile phase. This analysis was performed on an Agilent 1100 LC-MSD using an orthogonal electrospray interface. No effluent splitting was usedand the entire column flow (1 mL min–1) is introduced into the mass spectrom-eter. Since contamination of the interface and ion source is reduced by using avolatile buffer, an ammonium acetate buffer is preferred in this method over asodium salt buffer.

In addition to LC-MS, LC-MS/MS can also be used. By monitoring daughterions from a secondary fragmentation of an isolated parent ion, an additional levelof specificity is added. LC-MS/MS is especially suitable in the determination ofphthalates in complex matrices such as biota extracts [6].

2.3Analysis of Phthalic Acid Mono-Esters

Phthalic acid mono-esters are the primary metabolites of phthalates. In additionto the analysis of the phthalic ester diesters, recent analytical work also focuseson the analysis of the monoesters. The determination of monoethylhexyl phthalate (MEHP) was first described in biological samples (mainly plasma), asit was observed that during blood transfusion DEHP was leached out of PVCblood storage bags and tubing for haemodialysis and extracorporeal blood cir-culation [8–10]. Only later have methods for the determination of monoesters inother biomatrices been described. More recently, methods have been developedfor the analysis of the monoesters in environmental samples such as water, soil,and biota.

2.3.1Derivatisation and GC-MS

In comparison to the analysis of the phthalic acid diesters, the analysis of the mo-noesters is more difficult. The free fatty acid group makes the compound morepolar and this leads to adsorption during gas chromatographic analysis. For theanalysis of trace levels of monoesters, it is therefore not possible to obtain quan-titative data without blocking the acidic function prior to gas chromatographicanalysis. Derivatisation of the acid (–COOH) function can be done by esterifica-tion (–COOR) or by silylation (–COOSiR3). In selecting an appropriate derivati-sation technique, it is important that possible hydrolysis and/or transesterifica-tion are avoided. Monoethylhexyl phthalate can for instance be derivatised by classical esterification methods used for fatty acids (BF3/methanol,KOH/methanol, H2SO4/methanol, etc.) but these methods will lead to further hydrolysis, with partial or complete formation of phthalic acid dimethyl ester(DMP). It should also be kept in mind that in real samples the monoesters arenormally present in combination with the diesters. Some derivatisation reagentsalso lead to (partial) hydrolysis of the diesters and therefore lead to false positiveresults.

22 F. David et al.

Sjöberg et al. [8–10] described the use of a mixture of pentafluoropropanoland pentafluoropropionic anhydride as derivatisation reagent. The analysis wasperformed by GC-MS under conditions similar to those described for phthalateanalysis. Marschall and Egestad [11, 12] used silylation for the derivatisation ofMEHP and conjugates prior to GC-MS analysis. Esterification by triethyloxoniumtetrafluoroborate was used by Haam et al. [13] for the derivatisation of fourmetabolites, including MEHP, 5-carboxy-2-ethylpentyl phthalate, 2-ethyl-5-oxo-hexyl phthalate and 2-ethyl-5-hydroxyhexyl phthalate in urine extracts. Suzukiet al. [14] used diazomethane to prepare the methyl esters of the monoestersprior to GC-MS analysis. All these derivatisation techniques work well, but theymust be performed in anhydrous conditions. First the monoesters have to be ex-tracted from the matrix (blood, urine, water, suspended solids) and subsequentlythe extracts are dried under nitrogen. Then, the reagent is added and the de-rivatisation is performed under heating (pentafluoropropanol/pentafluoropro-pionic anhydride or silylation) or at room temperature (diazomethane). Thesederivatisation reactions cannot be performed in aqueous matrices. Recently,ethylchloroformate was used for the derivatisation (ethylation) of monoesters inaqueous media [15]. This derivatisation works in situ and allows subsequent extraction of the derivatised phthalates as described below.

2.3.2Analysis of Phthalic Acid Mono-Esters by HPLC-MS

Alternatively, phthalic acid monoesters can be analysed without derivatisation byHPLC [12, 16–19]. HPLC in combination with mass spectroscopic detection ispreferred especially for the analysis of isomeric mixtures of phthalic acid mo-noesters. For spectrometric detection, atmospheric pressure chemical ionisation(APCI) gives the highest sensitivity [18, 19].An example of a separation obtainedby LC-APCI-MS for a standard mixture is given in Fig. 8. The analytical condi-tions are summarized in Table 5. The mass spectra obtained by APCI-MS forMEHP and monoisodecyl phthalate (MiDP) are given in Fig. 9. The most abun-

Analytical Methods Review 23

Table 5. Analytical conditions for LC-APCIMS analysis of phthalic acid monoesters

Instrument Agilent 1100 LC–Agilent 1100 SL MSDColumn 25 cm L ¥ 4.6 mm i.d.¥ 5 µm Alltima CN (cyanopropyl silica)Mobile Phase A. 0.5% ammonium acetate in water

B. methanolGradient Isocratic, 50% A–50% BFlow 0.5 mL min-1

Injection 50 µL Detection Ionisation mode: APCI (atmospheric pressure chemical ionisation)

Polarity: negativeFragmentor: 70 VVaporizer: 325 °C; Drying gas: 5 L min–1 nitrogenMonitored Ions: M–1 (277 (DEHP), 281 (IS: d4-MEHP), 291 (MiNP),305 (MiDP))

24 F. David et al.

Fig.

8.LC

-MS

anal

ysis

ofp

htha

lic a

cid

mon

oest

ers b

y at

mos

pher

ic p

ress

ure

chem

ical

ioni

sati

on.(

Ions

mon

itore

d:m

/z27

7fo

r MEH

P,m

/z29

1fo

r MiN

Pan

d m

/z30

5fo

r M

iDP)

(ana

lyti

cal c

ondi

tion

s:se

e Ta

ble

5)

Analytical Methods Review 25

Fig. 9. Mass spectra obtained for MEHP and MiDP by APCI ionisation

26 F. David et al.

dant ions are detected at M–1 (m/z 277 for MEHP and m/z 305 for MiDP). In the spectrum of MiDP, a trace of monoisoundecyl phthalate is also detected(m/z 319). Blount et al. [18] also used LC-APCI-MS for the analysis of phthalatemetabolites in urine extracts. The analyses were performed on a ThermoQuestTSQ 7000 triple quadrupole analyzer. The mass spectrometer was operated inMS/MS mode. The parent ions at M–1 were isolated and specific daughter ionsobtained after collision-induced dissociation with argon were monitored. Thismethod allows very selective detection of phthalic acid monoesters in complexmatrices.

Prior to the analysis, the monoesters need to be extracted from the sample ma-trix. Examples of sample preparation methods for monoesters are given below.

3The Blank Problem

The major problem in phthalate analysis is the contamination problem, resultingin false positive results or over-estimated concentrations. The risk of con-tamination is present in the whole analytical scheme, including sampling, sam-ple preparation and chromatographic analysis. Due to the fact that phthalates are widely used, they are present in air, water, organic solvents, plastics and adsorbed on glass or other materials. Some typical sources of possible contam-ination are listed below and recommendations are given for minimising conta-mination.

3.1Sampling

Sampling is the critical first step. The sampling of liquids and solids is preferablydone in glass containers.All plastic materials should be avoided.Although somematerials do not contain phthalates as additives, phthalates might be adsorbed onthe surface and cleaning can be difficult. Specially cleaned glass containers canbe purchased (e.g. I-Chem sample bottles). Otherwise glass containers should berinsed with solvents and dried at 400 °C. Containers should not be left open, sincethey can adsorb phthalates from laboratory air onto the wall surface. Also, stop-pers for bottles or container lids can contain phthalates. The stoppers or lidsshould also be cleaned or blank checked.

During sampling, contact between the sample and hands or plastic glovesshould be avoided. Metal spatulas are preferred over plastic materials.After sam-pling, the containers should be closed. Storage of liquid samples can be done at4 °C. Biota samples or soils and sediments are stored at –20 °C. Since phthalatesundergo biodegradation, the storage of aqueous samples at 4 °C should be notlonger than 4 days. Chemical preservation may be performed by addition of500 mg sodium azide per litre sample [20]. A detailed description of samplingmethods for water and sediments is given by Parkman and Remberger [21] andby Braaten et al. [22].

3.2Sample Preparation

The most important rule is that the risk of contamination is reduced if the sam-ple preparation is kept to a minimum, with minimal extraction steps, minimal ex-tract concentration and minimal glassware use. For the different sample prepa-ration techniques, the risks of contamination associated with each method arediscussed. In general, glassware and solvents are the most important sources ofcontamination. Glassware can be cleaned by solvent rinsing and thermal treat-ment at 400 °C for 1–2 h [20]. After cooling, the glassware should be stored in aclosed container or wrapped in aluminium foil to avoid adsorption of phthalatesfrom the air. Prior to use, the glassware should be rinsed with a small portion ofblank-tested organic solvent (cyclohexane or isooctane) to deactivate the surface.Due to the thermal treatment at 400 °C, the glass surface is more active and phthalates can be (irreversibly) adsorbed.

Organic solvents and laboratory grade water also contain traces of phthalates.Even commercially available solvents for trace analysis (e.g. pesticide analysisgrade) can contain ppb (µg L–1) levels of phthalates. Some laboratories use in-house-distilled solvents, but this is not always possible in routine analysis andcontamination of the solvents during and after distillation is still possible. Dueto the ppb level of phthalates in solvents, concentration of the extracts should beminimised, since the concentration of phthalates will increase proportionally. Iffor instance liquid-liquid extraction is used and a 1 L sample is extracted threetimes with 40 mL solvent, and the extract is concentrated to 1 mL, a 1 µg L–1 con-tamination of phthalates in the extraction solvent will result in a 0.12 ppb back-ground value. The limit of quantification can therefore not be lower than0.12 ppb.Also, reagents need to be checked.Anhydrous sodium sulfate can, for in-stance, contain up to 0.5 mg kg–1 of DEHP.

Contamination of glassware and solvents is also likely to occur due to laboratory air. Phthalate concentrations in laboratory air are often in the100– 500 ng m–3 range and the phthalates can be concentrated in open solventbottles or adsorbed on glassware. Glassware can be protected by aluminium foil,while solvent bottles should remain closed until use. As far as possible, plasticmaterials should be removed from the laboratory. Cleaning agents used in thelaboratory (floor or furniture cleaning, glassware cleaning, etc.) may con-tain volatile phthalates (especially DEP, DiBP and DBP) and may severely con-taminate the laboratory air. Finally, it should be noted that personal hygiene materials (soaps, handcreams, cosmetics) often contain phthalates (mainly DMPand DEP).

3.3Chromatographic Analysis

Phthalates can be present in the chromatographic system. The most importantcontamination is located in the inlet and gas supply system. Split/splitless inletsmay contain septa, liners and O-rings that are contaminated with phthalates. Dur-ing splitless injection, a large solvent vapour cloud is created and these solvent

Analytical Methods Review 27

vapours can escape from the liner and perform phthalate extraction from septaand O-rings. This problem is less important with cool on-column or PTV inlets,since the temperature at which injection is made is lower and no large vapourvolumes are created.When using splitless injection, care should be taken that thevolume of the sample vapour does not exceed the liner volume. If splitless injec-tion is used, the quality of the septum should also be evaluated. Some septa con-tain phthalates and these should be avoided. Septumless inlets can be an alter-native (e.g. Merlin Microseal). Another critical factor is the quality of caps forautosampler vials. These caps can also contain phthalates. The best solution isnever to inject twice from the same vial. Once a vial cap has been punctured,phthalates leach out into the sample (in the organic solvent). It can be observedthat the phthalate concentration in an extract will increase with the number ofinjections made from the same vial.

In case special sample introduction systems such as thermal desorption areused, these systems also need to be checked for blanks. Cryogenic traps will con-centrate phthalates and even very low concentrations in blank desorption tubes,in gas lines, etc. will lead to false positive results.

4Analysis of Phthalates in Water Samples

For the determination of phthalates in water samples, published methods can bedivided in two classes: methods based on liquid-liquid extraction (LLE) andmethods based on solid-phase extraction (SPE). Both methods can be used suc-cessfully, but depending on the expected phthalate concentration and on thecomplexity of the matrix (drinking water versus wastewater) one technique ispreferred over the other. Recent developments in analytical methods for aqueousmatrices include micro liquid-liquid extraction (in combination with large vol-ume injection), solid-phase micro-extraction (SPME) and stir bar sorptive ex-traction (SBSE). These techniques are briefly discussed below.

4.1Liquid-Liquid Extraction (LLE)

The extraction of relatively large volumes of water samples (1–2 L) with an apolar non-miscible solvent is the most straightforward method for the extrac-tion of phthalates from aqueous samples. Good extraction recoveries are ob-tained by using dichloromethane, cyclohexane or hexane [22]. The ratio of thevolume of the water sample to the volume of the organic solvent should besmaller than 20. Higher ratios will still result in high recovery for the high-mol-ecular weight phthalates (log Kow = 7.73 for DEHP [23]), while low-molecularweight phthalates are more water-soluble and consequently the recovery drops(log Kow = 1.61 for DMP). After extraction, the organic phase is dried (over an-hydrous sodium sulfate) and concentrated. If complex-contaminated samples areextracted, several compounds such as hydrocarbons, detergents and plant mate-rial (sterols) are co-extracted. The extract can then be purified on an activatedaluminium oxide column. A glass column or cartridge is packed with 1 g alu-

28 F. David et al.

minium oxide (50–200 µm, neutral, activated for 4 h at 400 °C) and rinsed with2 mL extraction solvent. The column is dried under nitrogen and the extract iseluted through the cartridge and collected in a tube for further concentration oranalysis.

In general, liquid-liquid extraction is limited due to the presence of trace lev-els of phthalates in commercially available solvents, even in solvents for traceanalysis (pesticide residue analysis). Therefore the concentration factor is limited.A contamination of 1 ng L–1 in the solvent and a concentration factor of 50 (ex-traction with 50 mL solvent and concentration to 1 mL) leads to a 50 ng L–1 back-ground value. Accurate determinations below 100 ng L–1 (0.1 ppb) are thereforequestionable with this method. Vikelsoe et al. [3] reported a 90 ng L–1 detectionlimit and 770 ng L–1 background values for DEHP by a liquid-liquid extractionmethod applied to wastewater samples.

Recently, micro liquid-liquid extraction techniques have been presented [24].These techniques have the advantage that the solvent consumption is reducedand the whole extraction can be performed in a small (10–50 mL vial). Micro liq-uid-liquid extraction can be combined with large volume injection, resulting inmaintained sensitivity. Extraction of 25 mL with 5 mL solvent and injection of50 µL gives the same sensitivity as extraction of 250 mL with 50 mL solvent, con-centration to 1 mL and a 1-µL injection. However, with micro-LLE and large vol-ume injection, the purity of solvents is again a limiting factor.

Liquid-liquid extraction offers the advantage that water samples with particlesor heavily contaminated water samples (waste water) can be extracted. For theanalysis of relatively pure water samples (drinking water, river water, surface wa-ter, seawater) solid-phase extraction is preferred.

4.2Solid-Phase Extraction (SPE)

In comparison with liquid-liquid extraction, the amount of solvent can be re-duced drastically by using solid-phase extraction. The water sample (100 mL to2 L) is passed through an apolar sorbent and the phthalates are concentrated onthe sorbent. The most widely used sorbent for enrichment of phthalates is octadecyl silicagel (ODS, C18). For elution, different solvents can be used(dichloromethane, ethyl acetate, etc.). Solid-phase extraction can be performedwith classical cartridges containing 100–500 mg ODS material with an averageparticle size of 40 µm. These cartridges are commercially available from differ-ent suppliers. Most commercial cartridges are made from polyethylene orpolypropylene barrels and this can cause relatively high background levels. Fortrace analysis, glass barrels are available that can be packed with ODS material.As an alternative to classical cartridges, solid-phase extraction disks can be used.These disks are composed from small particle (7 µm) ODS material that is in-corporated in a Teflon fibril. The advantage of the disks is that the sampling flowis higher and that (very) large samples can be processed faster.

A typical SPE sample preparation method is proposed as a draft ISO method[2]. The method is based on the work of Furtmann et al. [20]. Glass SPE cartridgesare packed with 250 mg of RP C18 material. The cartridges are installed on a SPE

Analytical Methods Review 29

vacuum manifold and conditioned with ethyl acetate (one bed volume) and thendried under nitrogen (for 10 s). Nitrogen is preferred over air to reduce contam-ination. Then, the cartridges are conditioned with methanol (two bed volumes).After this conditioning, the ODS bed is not allowed to run dry.A (300 mL) reser-voir is connected on top of the cartridge and the sample (250 mL) is loaded in thereservoir. Another pre-treated ODS cartridge is installed above the reservoir toreduce contamination by air. With the aid of vacuum, the sample is elutedthrough the cartridge at a flow rate of about 2–10 mL min–1. After elution, thecartridge is dried with nitrogen for about 5 min. Finally, the phthalate fraction iseluted from the cartridge into a collection tube with 2 mL of internal standard so-lution (in ethyl acetate) by using vacuum. The collected fraction can be analyseddirectly or an additional clean-up can be used (see above).

The advantage of solid-phase extraction is that a large concentration factorcan be obtained without (or with limited) solvent concentration, and con-sequently without the concentration of contaminants in the organic solvent.If a 250 mL water sample is extracted and the final volume is 2 mL, a con-centration factor of 125 is obtained without solvent concentration. Back-ground levels below 10 ng L–1 and detection limits below 50 ng L–1 can be obtained.

The major limitation of the SPE method is the extraction of water samplescontaining solids or heavily contaminated samples. If the amount of particulatesis low, the SPE method can still be used if the particles do not block the car-tridges. If the eluate is allowed to soak the particulates for a few minutes (afterthe eluting solvent is added), quantitative extraction is obtained. If a high amountof particles is present, it is advised that the sample is filtered first and the phthalates measured separately on the aqueous and the solid phases. If the sam-ple is highly contaminated with other organic material (solvents, oil) or containshigh amounts of detergents, recovery may be low do to incomplete enrichmenton the C18 material. For these samples liquid-liquid extraction is advised.

Several groups have used similar SPE techniques successfully. Van der Veldeet al. [24] used 500 mg C18 Polar Plus cartridges for the extraction of 250 mLsamples. The cartridges were eluted with 5 mL of a 70% pentane-30% methyltert-butyl ether mixture. The reported recoveries were between 63% (DEHP) and99% (BBzP). The detection limit for the single isomer phthalates was around0.1 µg L–1. Letinski [25] used C18 extraction disks for the analysis of 3 L watersamples. Elution was done by 10 mL of an ethyl acetate/dichloromethane mix-ture. The extract was concentrated to 0.5 mL. The detection limit for DiNP andDiDP was 0.1 µg L–1 and the recovery for DiNP and DiDP was 94% and 99%, re-spectively. Lin et al. [6] also used C18 extraction disks.Very high sample volumes(30 L) were pumped through the cartridge (47 mm C18 3 M Empore disk) at 30 mL min–1. After extraction, the disks were extracted by sonication, as described for sediment samples. In this way, very low concentrations (down to5 ng L–1) could be measured in seawater samples.

Normally extracts from water samples (drinking water, river water, seawater,rainwater) obtained by solid-phase extraction do not require additional clean-up.Only wastewater samples can require additional clean-up (on aluminium oxide,see above).

30 F. David et al.

4.3Solventless Extraction Methods (SPME and SBSE)

Recently solventless extraction methods have been described for the ex-traction of semi-volatile compounds, such as pesticides, PAHs and phthalates,from aqueous matrices. Pawliszyn [26] developed solid-phase micro-extrac-tion (SPME). In this extraction technique, a fused silica fibre coated with a thin(7–100 µm) layer of polymer phase (usually polydimehylsiloxane) is immersedin the aqueous sample, while the sample is stirred. After a certain extraction time (varying from minutes to one hour), the coated fibre is retracted in a holder, transferred to the GC and desorbed in a hot inlet (for instance a split-less inlet at 250 °C). The sorbed compounds are desorbed again and injected intothe column for separation and analysis. This simple technique has been appliedto different environmental applications. However, the major limitation is thesmall amount of coating on the fibre. The very small volume of the coating(around 0.5 µL) results in a very high ratio between sample volume (typically10–20 mL) and the extraction phase. Consequently, only solutes with very high (Kow > 4) octanol-water partition coefficients can be extracted with high re-covery.

Baltussen et al. [27] introduced magnetic stir bars coated with silicones as al-ternative solventless extraction method (stir bar sorptive extraction, SBSE). Thestir bars are placed in 10–50 mL samples and are stirred on a magnetic stirrer for30–120 min.After extraction, the stir bar is removed from the sample and placedin a thermal desorption unit for thermal desorption and on-line GC-MS analy-sis. The same methodology is used for air monitoring (see below). For semi-volatile compounds, such as phthalates, a 10 mm-long stir bar coated with a 0.5 mm film of dimethylpolysiloxane (PDMS volume = 24 µL) is used. Theamount of coating is 50 times higher and consequently higher recoveries are ob-tained for compounds with lower Kow values. Stir bar sorptive extraction has beenused for various environmental samples and was also used for the extraction ofsemi-volatile compounds, including phthalates, in beverages. The possibilities ofstir bar sorptive extraction, followed by thermal desorption-GC-MS are demon-strated in Fig. 10.A 10 mL rainwater sample was collected directly in a 20 mL vial.The PDMS-coated stir bar was introduced and extraction was performed over2 h. The resulting chromatogram (extracted ion chromatogram at m/z 149) showsthe presence of DiBP (16.01 min), DBP (18.39 min) and DEHP (29.59 min). Theconcentrations were 4.60 ppb (DiBP), 0.53 ppb (DBP) and 0.25 ppb (DEHP).These data clearly illustrate that transport through air is an important pathwayof distribution of phthalates in the environment. This analysis also shows thatvery high sensitivities can be obtained by SBSE. In this application, the mass spec-trometer was operated in scan mode and the signal-to-noise ratio for DEHP ishigher than 40. The detection limit is thus lower than 25 ppt (ng L–1) in scanmode and lower than 5 ppt in SIM mode.

Analytical Methods Review 31

32 F. David et al.

Fig.

10.

Extr

acte

d io

n ch

rom

atog

ram

ofs

tir

bar

sorp

tive

ext

ract

ion

ofph

thal

ates

from

rai

nwat

er (1

6.01

min

=D

iBP,

18.3

9m

in=

DBP

,29

.59

min

=D

EHP)

Analytical Methods Review 33

5Analysis of Phthalates in Sediments, Soils and Sewage Sludges

In comparison with water samples, the determination of phthalates in solid (orsemi-solid) samples usually requires two steps: an extraction step and a clean-upstep. Solid samples include soils, sediments, sludges and solid waste. Soil samplesare normally collected as grab samples by using a stainless steel drill, which allows sampling at different depths. Typical concentration levels for phthalates in soil are in the order of 10–1000 µg kg–1 dry mass.

Sampling of sediments is used for the study of historical phthalate contami-nation and for the determination of local contamination and biodegradation.Sampling techniques that can be used for sediment sampling are described byParkman and Remberger [21] and by Braaten et al. [22]. Concentration levels ofphthalates largely depend on the sampling site. The average concentrations areof the same order of magnitude as the soil samples.

Sludge samples from wastewater treatment are important samples to monitorinput sources and to study biodegradation. Moreover, determination ofphthalates in treated (digested and dried) sludge may be important. In at leastone European country (Denmark) there is a defined maximum allowable level forDEHP in sludge to be used as an agricultural fertiliser. Phthalates entering thewastewater treatment plant via the wastewater influent are in general well re-moved from the aqueous stream in part by biodegradation and in part by removal in the sludge. Typical phthalate concentrations in treated sewage sludgeare between 10–100 mg kg-1 dry mass. Dried sludges usually have a 20–30% drymass content and can be analysed in a similar way to soils and sediments. Wetsludges (dry mass < 5%) can be analysed as such or after concentration of thesolids and removal of the aqueous phase (centrifugation, filtering). Solid samplescan be stored at –20 °C during several weeks.

5.1Extraction

For the extraction of phthalates from solid (or semi-solid) matrices, various tech-niques have been used. Soxhlet extraction, whereby an organic solvent is heatedand the condensed vapours are percolated through the solid samples held in a fil-ter cartridge (thimble), is still considered as the reference method for the ex-traction of semi-volatile pollutants from solid environmental samples. Soxhletextraction is a slow extraction procedure (4–24 h) and uses relatively largeamounts of solvent (100–250 mL per sample). Due to the large solvent con-sumption, contamination is therefore a major issue with Soxhlet extraction.

Soxhlet extraction has been used by Steffen and Lach [28]. Sediment sampleswere first freeze-dried. Then 30–50 g samples were extracted with 200 mLtoluene over 8 h. For some samples, sulfur compounds were removed prior to GC-MS analysis by sonication of the extract in combination with copper powder.

During the past years, automated and miniaturised extraction techniques areslowly replacing classical Soxhlet extraction. These techniques include automatedSoxhlet extraction (Soxtec), shaking, ultrasonic extraction, microwave assisted

extraction (MASE), accelerated solvent extraction (ASE) and supercritical fluidextraction (SFE).Automated miniaturized Soxhlet extraction (Soxtec, Soxtherm)is faster than Soxhlet extraction and solvent consumption is reduced. Further re-duction of solvent consumption is possible by using supercritical fluid extraction.In SFE, pressurized carbon dioxide is brought to a pressure and temperatureabove its critical pressure (75 bar) and critical temperature (35 °C), resulting ina supercritical fluid. A supercritical fluid can be considered as a dense gas, withgas-like viscosity and flow characteristics and liquid-like solvating characteris-tics. Supercritical fluid extraction has been used successfully for a broad rangeof applications. SFE was used by Kolb et al. [29] for the extraction of phthalatesfrom sewage sludge. The extraction was performed at 60 °C with carbon dioxidemodified with 5% hexane as the extraction solvent and a three-step pressure pro-gram (5 min at 200 bar, 5 min at 300 bar and 20 min at 400 bar). The extractionefficiencies were higher than 90% for DBP, BBzP and DEHP, and were 84% forDiNP and 79% for DiDP. The standard deviation was 2–7%.

Currently, supercritical fluid extraction is replaced by accelerated solvent ex-traction (ASE). In ASE, similar equipment is used, but the supercritical fluid is re-placed by a classical organic solvent (dichloromethane, hexane, etc.). The solventis pumped into an extraction vessel containing the sample. The solvent in thethimble is pressurized (up to 5000 psi) and heated (typically at 100°C). After ex-traction, the solvent is eluted in a vial. The extraction is completely automatedand in comparison to Soxhlet extraction, solvent consumption is drastically re-duced. Accelerated solvent extraction was used by Lettinski et al. [25]. Typically15 g soil sample is mixed with anhydrous sodium sulfate (or Hydromatrix). Thesample is placed in an extraction thimble and extracted with dichloromethaneat 2000 psi and 100 °C for15 min. Recoveries of phthalates were 80–100%. Themain problem was the relatively high background values that were typicallyaround 36 µg kg–1 for DiBP, 24 µg kg–1 for DBP and 35 µg kg–1 for DEHP.

Probably one of the most simple and cheap extraction methods is shaking orsonication. Vikelsoe et al. [4] used a simple shaking method for the determina-tion of phthalates in soil samples. The samples were not dried before extraction.After addition of internal standard, a 50 g sample was extracted with 100 mLdichloromethane by shaking for 4 h. The blank values were around 6 µg kg–1 forDEHP.

Low solvent consumption in combination with relatively short extractiontimes are also obtained by ultrasonic extraction. Braaten et al. [22] used sonica-tion for the determination of phthalates in sediments by using 5 g wet sample.Firstly, 2 mL acetonitrile was added and a 10 min ultrasonic extraction was per-formed. Then, 2 mL acetonitrile and 3 mL hexane were added and the ultrasonicextraction was repeated for 30 min. The organic phase was isolated, washed withwater and the hexane phase was cleaned on aluminium oxide. A similar methodwas applied by Parkman and Remberger [21] to the determination of phthalatesin Swedish and Dutch sediments. Liu [6] used a triple sonication extraction of 2 g sample (mixed with 20 g sodium sulfate) with each time 20 mL1 : 1 dichloromethane : hexane for the determination of phthalates in marine sed-iments. Paxeus [30] also used ultrasonic extraction of dried sludge with MTBEor a 1 : 1 mixture of hexane and acetone. Recoveries were higher than 90% by

34 F. David et al.

using three 10 min extractions. The organic phase (upper layer) is removed after each extraction and replaced by fresh solvent. Zurmuhl [31] demonstratedthat similar extraction recoveries are obtained by ultrasonic extraction in com-parison to Soxhlet extraction.

5.2Clean-Up Procedures

Since extracts of soils, sediments and sludges often contain other contaminantsor co-extracted compounds such as plant sterols, clean-up is often needed. A simple clean-up method based on solid-phase extraction is described by Braatenet al. [22]. The extract from sediments are applied on a 500 mg neutral aluminacolumn (activated at 400 °C, deactivated with 9% water). The column is rinsed with 3 mL hexane. The phthalate fraction is eluted with 3 mL of a 75%hexane/25% MTBE mixture. Liu [6] used 15 g alumina column (300 ¥10 mm i.d.,alumina deactivated with 15% water) with 1–2 cm anhydrous sodium sulfate ontop. After applying the sample, the column is eluted with 30 mL hexane (waste),then with 30 mL 10% dichloromethane in hexane (PCB fraction) and finally with30 mL 50% dichloromethane in hexane. This last fraction contains the phthalates.In some cases, an additional clean-up on a 7.5 g Florisil column was used. Thecolumn was rinsed with 30 mL dichloromethane (waste) and with 30 mL 5% ace-tone in dichloromethane. The phthalates are present in this last fraction. Themethod blanks for a 2 g sediment sample were: 1.3 µg kg–1 DiBP, 28.4 µg kg–1 DBPand 24.9 µg kg–1 DEHP. The detection limits were thus determined by the methodblanks for these compounds. The reported recoveries were higher than 70% forDMP, higher than 80% for DEP, DiNP and DiDP, and higher than 90% for DiBP,DBP, BBzP and DEHP.

5.3Determination of Phthalates in Sewage Sludge

The determination of phthalates in a sludge sample is demonstrated in Fig. 11.The sludge was obtained from a wastewater treatment plant. The extraction wasperformed by ultrasonic extraction with acetone : hexane according to the pro-cedure described by Paxeus [30]. In this sample, DiBP, DBP, BBzP and DEHP areclearly detected. At the end of the chromatogram, the “hump” corresponding toDiNP and DiDP is also detected. In this sample, the concentration of DEHP, themost abundant phthalate, was around 30 mg g–1 dry mass.

6Analysis of Phthalates in Air

6.1Introduction

Transport in the atmosphere is an important pathway for the distribution ofphthalates in the environment [32, 33]. Most of the phthalates are adsorbed onto

Analytical Methods Review 35

36 F. David et al.

Fig.

11.

GC

-MS

chro

mat

ogra

m o

btai

ned

for

a w

aste

wat

er tr

eatm

ent p

lant

slu

dge

extr

act

Analytical Methods Review 37

particulate matter [34–36] and earlier methods used for monitoring phthalatesin air often focused on the collection and analysis of dust samples [37].Althoughthe vapour pressures of phthalates are relatively low [23], phthalates can still beregarded as semi-volatiles and are also present in the aerosol or vapour phase.

Since phthalates are present at trace levels (ng–µg m–3 level), direct analysisof air samples is not possible and enrichment techniques are needed. Enrichmentof compounds in air can be done by active or passive concentration of the soluteson adsorbents or sorbents. For the determination of phthalates adsorbed on par-ticulates, active or passive sampling on filters is used.

6.2Analysis of Total Phthalate Concentrations in Air

For the determination of total phthalate concentrations (gaseous + aerosols +particulates), sampling on a sorbent tube is used. Most widely used adsorbentsare activated carbon, porous polymers like Tenax (2,6-diphenylphenylene oxidepolymer) and resins (XAD-2) [38–47]. Recommended methods for the samplingand analysis of phthalates (OSHA CIM, OSHA 104, NIOSH 5020) also includetrapping of the phthalates present as aerosols or in the vapour phase on a com-bination of a polyurethane foam (PUF) plug and a resin cartridge. Phthalates adsorbed to particulate material can also be collected first on a glass fibre filter.In these generic methods, typically 60–240 L of air are sampled at a rate of1000 mL min–1. It should be noted that with these methods contamination prob-lems can occur due to PVC filter holders or plasticized rings that are used to holdthe glass fibre filters. These materials can contain phthalates and consequentlylead to high blank values.

After sampling, the phthalates are desorbed from the adsorbents by liquid ex-traction, or by thermal desorption. In the case of particulates concentrated on fil-ters, liquid extraction is used. The extraction procedure for particulates on filtermedia is very similar to the extraction of phthalates from solid samples (soil, sed-iment). Ultrasonic extraction, Soxhlet extraction, accelerated solvent extractionand others can be used [41, 47]. The advantage of liquid desorption is that quan-titative extraction of the solutes can easily be obtained, the extract can be furtherfractionated or purified and the final extract can be analysed several times.Liquid desorption, however, lacks sensitivity. Assuming a limit of detection of10 pg per compound with mass spectroscopic detection, a concentration of min-imum 10 ppb is required in the final extract. By using liquid extraction of the sor-bent or filter, concentration to 1 mL and a 1 µL injection, at least 1000 L of airshould be sampled to quantify 10 ng m–3. Otake et al. [46] used for instance100 mg charcoal tubes and sampled for 72 h at 1 L min-1 (total sample 4.3 m3).After sampling, desorption was done with 1 mL toluene and the extract wasanalysed by GC-MS. A detection limit of 4 ng m–3 could be obtained. Thermaldesorption, on the other hand, has the advantage of higher sensitivity because allsorbed compounds can be quantitatively transferred to the GC and the detector.The same sensitivity (10 pg in detector) can thus be obtained by sampling only1 L of air. The limiting factor in thermal desorption is the desorption efficiency.The adsorption of high-molecular weight compounds is strong and a high energy

(high temperature) is needed for quantitative desorption especially when classical adsorbents are used (e.g. carbon-based materials, Tenax and Porapak).Tenax adsorption traps can be used successfully for the low-molecular weight phthalates (typically DEP, DiBP, DBP), but for the high-molecular weight isomericmixtures (DiNP, DiDP) poor recoveries have been observed.

Pre-concentration by sorptive enrichment with silicones offers a useful alter-native and was introduced by Baltussen et al. [48, 49]. Air was sampled througha tube packed with a bed of 100% polydimethylsiloxane (PDMS). The polymericmaterial is above its glass transition temperature (Tg) at sampling temperatures(0–30 °C) and the solutes are, in contrast to classical sampling systems, sorbedinto (dissolved in) the liquid stationary PDMS phase rather than adsorbed ontoan active surface. The sampling traps allow sufficient enrichment at a relativehigh sampling speed and quantitative thermal desorption can be performed atmoderate temperatures. Moreover, the material is highly inert and has a highthermal stability. However, for the enrichment of phthalates, it was observed that100% PDMS traps resulted in too many bleeding compounds (low-molecularweight silicones) that interfered in the extracted ion chromatograms at m/z 149[50]. This is the most abundant and the quantification ion for phthalate deter-mination. The dimethylsiloxane oligomers give m/z 73 and m/z 147 as mostabundant ions but because of the silicium isotopes the ion m/z 149 is also ob-served in the spectra.An alternative method was found by coating a silicone layeron an inert support in a 5% (w/w) concentration of 5%. Thermal desorptiontubes (4 mm i.d. ¥ 180 mm) were filled with 100 mg of the material and were con-ditioned at 300 °C for 2 h. The ‘blank’ profile showed only some traces of volatilesilicon fragments, but these did not interfere with the phthalate peaks. By usingthese traps, ng m–3 concentrations of phthalates can be measured. Since pre-con-centration on PDMS tubes is based on sorptive enrichment, the breakthroughvolumes (Vb) of phthalates can be calculated directly from the theory developedby Lövkist and Jönsson [51]. The values for some important phthalates are sum-marised in Table 6 and vary between 17 L for dimethyl phthalate up to 75 ¥ 106 Lfor diisodecyl phthalate when sampling is performed at a rate of 500 mL min–1.If the sample volume is limited to 15 L, all phthalates are quantitatively trapped.Assuming a mass spectrometric detection limit of 10 pg, the sensitivity of thethermal desorption-GC-MS method is 10 pg/15 L or 0.7 ng m–3. In practice, a

38 F. David et al.

Table 6. Breakthrough volumes (L) of several phthalates whensampling is performed on PDMS tubes (100 mg 5% PDMS) at aflow rate of 500 mL min–1 at room temperature (20 °C)

Compound Breakthrough volume (L)

DMP 17DEP 23DIBP 13 ¥ 102

DBP 19 ¥ 102

DEHP 17 ¥ 106

DIDP 75 ¥ 106

limit of quantitation of 3 ng m–3 was obtained due to typical background levelsof DiBP, DBP and DEHP around 20–30 pg.

A typical sampling and analysis procedure based on sorptive enrichment andthermal desorption-GC-MS is summarized as follows [50]. Samples are aspiratedthrough a thermal desorption tube containing 100 mg 5% PDMS sorbent by us-ing a universal air sampling pump (SKC, Dorset, UK) at a nominal flow of 500 mLmin–1 for 30 min (sample volume = 15 L).After sampling, the sampling tubes areclosed and stored in an airtight container or wrapped in aluminium foil. Aftersampling, the tubes are placed in a thermal desorption unit (TDS-A, GerstelGmbH, Muelheim, Germany), mounted on an Agilent 6890 GC coupled to an Ag-ilent 5973 mass selective detector. Desorption is started by programming the tubeto 300 °C. The released solutes are transferred from the sampling tube through aheated fused silica capillary into a cryotrap (PTV inlet). During thermal des-orption, the PTV is cooled to –100 °C and the phthalates are trapped in an emptyliner. After 10 min, the thermal desorption is completed and the solutes are in-jected into the capillary column by rapidly heating the PTV to 300 °C at 10 °C s–1.The chromatographic and mass spectroscopic conditions are the same as de-scribed above (Table 3).

As for other sample matrices, a potential problem in the trace analysis ofphthalates in air is contamination. It is very important that the contact betweenstored sample tubes and laboratory air is minimised, since DiBP and DBP werefound to be two important contaminants originating from the surrounding labatmosphere. A typical chromatogram of the analysis of laboratory indoor air is presented in Fig. 12. The measured concentrations of DiBP and DBP are200–700 ng m–3. If tubes are left unprotected in the laboratory, they will adsorbphthalates and this leads to overestimated data.

The limit of detection (LOD) of the method with a sampling speed of 500 mLmin–1 and a sampling time of 30 min (15 L sample) is of the order of 1 ng m–3

for the single isomer phthalates and around 10 ng m–3 for the mixed isomer phthalates. For DiBP, DBP and DEHP, background values under optimised (clean)conditions can range between 2–3 ng m–3 and values below these limits are ques-tionable for these compounds. The sorptive enrichment-thermal desorption-GC-MS method was validated in the concentration range 3 to 3000 ng m–3. The cor-relation coefficients of the linearity curves were higher than 0.9970 for the singleisomer phthalates DMP, DEP, DiBP, DBP, BBzP and DEHP, and 0.9786 for DiNPand 0.9782 for DiDP. The relative standard deviation of five replicate analyses ata level of 100 ng m–3 was between 2.3% (for DEP) and 9.2% (for DEHP).

As a typical example of phthalate monitoring in air, the concentration ofphthalates was measured in a greenhouse and in the surrounding atmosphere.Outside samples were taken at 1, 10 and 100 m distances. The resulting data aresummarized in Fig. 13. Inside the greenhouse, the concentrations of the detectedphthalates were respectively: 226 ng m–3 DiBP, 156 ng m–3 DBP, 48 ng m–3 BBzPand 309 ng m–3 DEHP. At 100 m distance, only background levels of DiBP, DBPand DEHP were measured and the concentrations decrease with increasing dis-tance from the greenhouse.

It should also be noted that dynamic (active) sampling by using either adsor-bents or sorbents measures total phthalate concentrations, since particulates will

Analytical Methods Review 39

Fig.

12.

Phth

alat

e of

labo

rato

ry a

ir a

t m/z

149

on a

5%

PD

MS

tube

.Pea

ks 1

DEP

,2D

iBP,

3D

BP,4

DEH

P

40 F. David et al.

also be trapped on the material that acts as a filter. If a glass filter is placed be-fore the (ad)sorbent tube, the particulates can be monitored separately. However,the differentiation between particle-bound (and adsorbed) phthalates and thephthalates in gaseous and aerosol phase by such a sampling train (filter+sorbent)is questionable. It can be expected that phthalates initially present in the gaseousphase can also be concentrated on a filter or on dust collected on the filter dur-ing sampling due to adsorption (so the concentration in the sorption tube is toolow). In contrast, phthalates initially adsorbed on particles and trapped on a glassfibre filter can be purged out towards the sorption tube (giving too high valuesfor the “gaseous+aerosol fraction”). This effect has already been clearly demon-strated by Schulz and Püttmann [37]. With a 1 h sampling time, measured phthalate concentrations were higher than the values obtained by using a 24 hsampling. This effect was called the “blowing off” problem. Therefore, we recommend the use of adsorbents or sorbents for the determination of total phthalate concentrations.

6.3Passive Sorptive Sampling of Phthalates in Air

Passive sampling is a very suitable technique for the measurement of air qualityin indoor and working environment [52]. For passive sampling, a sorbent is putin a holder. This may be placed in a specific place in the room or used as a per-sonal sampler. Sampling is performed over several hours or days and concen-trations are expressed as time weight averages (TWA). The (ad)sorbed solutes aredesorbed with a suitable solvent and the extract is analysed. Recently, solid-phasemicro-extraction (SPME) with a PDMS coating immobilized on a fused silica

Analytical Methods Review 41

Fig. 13. Phthalate concentrations (ng m–3) in air samples in and outside a greenhouse

fibre, was introduced as an alternative sampling device for passive diffusion sampling [53–55]. After sampling, the fibre is introduced in a hot GC inlet andthe concentrated compounds are thermally desorbed in the GC. The major con-straint of the SPME method is the relatively small amount of sorbent that is pre-sent on the fibre (0.5 µL). For this reason, the recently developed PDMS-coatedstir bars can be an interesting alternative. It has been demonstrated that these stirbars can be used to sample volatiles in the headspace of liquid or solid samples[56]. In passive sampling with sorptive extraction, the PDMS polymer is exposedto an air sample during a relative long period (hours, days) and the pollutantspresent at an initial concentration CS

0 are enriched until equilibrium is reachedbetween the gas and the PDMS polymer phase. This distribution can be definedfor a given component at a given temperature as the distribution coefficientKPDMS/air:

C •PDMSKPDMS/S = 9 (1)C •

S

where C •PDMS and C •

S are the solute concentrations in the PDMS phase and airsample at equilibrium state, respectively. Since passive sampling is usually per-formed in open areas, where the sample volume is much larger than the volumeof the PDMS phase, the concentration C•

S at equilibrium does not significantly dif-fer from the initial concentration CS

0. This last value can thus easily be calculated:n

CS0 = 968 (2) or n = VPDMS · CS

0 · KPDMS/S (3)VPDMS · KPDMS/S

where n the absolute amount of solute that is enriched in the PDMS polymer andVPDMS is the volume of the PDMS phase. Equation (3) shows that n is directly pro-portional to the amount of sorbent VPDMS . This implies that if a coated bar con-taining 50 µL PDMS is used instead of an SPME fibre (0.5 µL PDMS) the sensi-tivity is increased by a factor 100.

Passive sampling with a PDMS-coated bar was applied to the determination ofphthalates in a room with PVC flooring. A clean PDMS-coated bar containing50 µL PDMS was exposed to the indoor environment for 24 h.After sampling, thephthalates were analysed by thermal desorption-GC-MS. The resulting extractedion chromatogram at m/z 149 is shown in Fig. 14. The high abundance of DiBP,DBP and DEHP illustrates the extremely high enrichment of the phthalates in thePDMS polymer phase. Since phthalates adsorbed on particulates are not enrichedby this sampling method, passive sampling with sorptive extraction allows dif-ferentiation between phthalates that are present in the vapour or aerosol phaseand phthalates adsorbed onto particulates. Only the phthalates in vapour oraerosol phase are measured. Therefore, passive sorptive sampling is comple-mentary to dynamic sorptive enrichment.

6.4Dust Analysis

An important part of the human phthalate intake from the atmosphere may beattributed to the inhalation of dust particles on which plasticizers are adsorbed.

42 F. David et al.

Analytical Methods Review 43

Fig.

14.

Pass

ive

diff

usio

n ex

trac

tion

(PSS

E,50

µL P

DM

S) o

fpht

hala

tes

in a

car

pete

d ho

use

44 F. David et al.

Hence, attention has been paid to the determination of phthalates in dust. Thehazardous impact of phthalates adsorbed on dust particles depends on the par-ticle size of the dust.According to the American Conference of Governmental In-dustrial Hygienists (ACGIH) [57] and European regulation [58], the inhalabledust fraction is defined as the fraction of particulates with a median particle sizesmaller than 100 µm. The thoraric fraction is defined as the fraction of particu-lates with a median particle size smaller than 10 µm. The respirable fraction is de-fined as the fraction of particulates with a median particle size smaller than4 µm. Inhalable dust is deposited in the respiratory trajectory (mouth and gul-let), while respirable particles are deposited anywhere in the gas-exchange re-gion, including the lung bladders. Thoraric dust is retained before the lung area.

Dust sampling can be done by using vacuum cleaners, glass fibre filters incombination with high-volume air sampling pumps or with special dust sam-plers. A detailed description of dust sampling is given in VDI method 4300 [59].Special dust samplers allow differentiation between the total inhalable and res-pirable dust filtration [60]. By using these samplers, the total inhalable dust frac-tion is aspirated at a flow of about 2 L min–1 onto a glass fibre filter (pore size1.0 µm) through a broad cassette inlet. This sampler discriminates larger partic-ulate matter. Smaller respirable particles can be collected with a cyclone-typesampler. The recent models are constructed of plasticizer-free plastics that elim-inate electrostatic problems. This avoids repellence of the particles and contributes to high sampling efficiencies. In both cases, sampling is typically performed over several hours. Total dust concentrations are measured gravi-metrically. Phthalates adsorbed on the dust fraction can be determined after liq-uid extraction of the filter. However, in our experience, it is extremely difficult toobtain reliable data on phthalates on the dust fraction collected in these specialsamplers.

More reliable data are obtained by sampling larger amounts of dust by high-volume sampling, followed by dust fractionation by sieving and phthalate deter-mination by liquid extraction and GC-MS. Fractionation of dust samples can bedone by using stainless steel sieves or by centrifugation. Extraction of phthalatesfrom dust can be performed by the methods described for solid samples [47, 61].Alternatively, thermal extraction can be used, especially if limited amounts ofsample are available. A few mg of sample can be placed in a thermal desorptiontube and the tube is analysed in the same way as for air samples. An example is given in Fig. 15, showing the thermal extraction-GC-MS analysis of a housedust sample. The measured phthalate concentrations were 90 mg kg–1 DBP,200 mg kg–1 DiHP, 700 mg kg–1 DEHP and 300 mg kg–1 DiNP. This sample clearlyshows the presence of the C7-isomeric mixture of diisoheptyl phthalate elutingbefore DEHP.

Analytical Methods Review 45

Fig.

15.

The

rmal

ext

ract

ion-

GC

-MS

anal

ysis

ofh

ouse

dus

t

7Analysis of Phthalates in Biota (Vegetation, Milk, Fish)

7.1Introduction

The determination of phthalates in biological matrices such as vegetation, fish ormilk is more difficult due to the complexity of the matrix. In general, the method-ology for the determination of phthalates in biota can be divided into two classes,depending on the fat content of the matrix.

For samples with a relatively low fat content (< 1%), phthalates can be ex-tracted by using the same methods as described for soils, sediment and othersolid samples. During the extraction, other constituents such as sterols, pigments,flavanoids, waxes, fatty acids, etc will be co-extracted. After extraction, a clean-up method is needed. As clean-up methods, column chromatography or solid-phase extraction can be used. In these clean-up methods, the separation ofphthalates and co-extracted compounds is based on differences in polarity.

For fatty matrices, the main problem is the co-extraction of fats. Since lipidshave a polarity similar to the polarity of phthalates, it is difficult to remove themwith methods based on column chromatography or solid-phase extraction. Thefat matrix can be removed by a clean-up method based on size exclusion chro-matography (gel permeation chromatography, GPC). In size exclusion chro-matography, compounds are separated according to the molecular volume(ª molecular mass). Larger molecules cannot enter the pores of the GPC pack-ing material and elute before smaller molecules that can enter the small pores.GPC was first described for the fractionation of pesticides from lipid matrices[62–63]. However, this liquid chromatography method is more and more con-sidered as a very valuable sample preparation technique for all determinationsof organic contaminants in lipid matrices. On GPC columns, phthalates with mol-ecular weights in the range 200–400 Da are separated from lipids (MW around800 Da). Classical GPC separations are performed on large (40 cm ¥ 25 mm i.d.)low-pressure columns, operated at 5 mL min–1 [62–63]. The method can beminiaturized and solvent consumption can be largely reduced by using HPLCequipment and high-pressure GPC columns. Automated GPC clean-up can beperformed with a system consisting of an isocratic HPLC pump, an autosamplerallowing the injection of 500 µL fat extract, a temperature-controlled columnoven, a variable UV detector (optional) and a fraction collector. The separationis performed on small-bore columns (e.g. 300 mm ¥ 7.5 mm i.d. PL-Gel with5 µm particles and 5 nm pore size), operated at 1 mL min–1 dichloromethane. The5 nm pore size is important, since it allows the separation of organic compoundsin the 100–1000 mass range. Best resolution is obtained on two columns in series (total length = 60 cm). Solvent consumption is approximately 5–10 timeslower in comparison to the classical GPC method. The phthalate fraction (3–5 mL) is collected automatically.After collection, the fraction is concentratedand this clean-up step can eventually be followed by an additional clean-upmethod according to polarity by using column chromatography or SPE.

46 F. David et al.

7.2Analytical Procedure for the Determination of Phthalates in Vegetation

Vegetation samples normally do not contain high levels of lipids. For the deter-mination of phthalates in plant material, first a liquid-solid extraction is used.Alltechniques described for solid samples (Soxhlet, shaking, ultrasonic extraction,ASE, SFE) can also be used for the extraction of phthalates from plant material.High recoveries are obtained by using a simple shaking or ultrasonic extractionmethod. These methods are also fast, cheap and relatively small amounts of sol-vents are used. For the extraction, two approaches can be used.

In the first approach [6], vegetation samples are first homogenised with ablender. Approximately 5 g (wet) sample is spiked with internal standard(s) andmixed with 30 g anhydrous sodium sulfate (pre-baked at 450 °C) in a mortar un-til a dry powder is obtained. The dried sample is extracted with 30 mL of a1 : 1 mixture of hexane and dichloromethane by sonication for 10 min. After thesuspended particles are settled, the supernatant is removed. The extraction is re-peated twice with fresh solvent and finally the three organic fractions are com-bined and concentrated to 1 mL under nitrogen.

In the second approach [7], the sample is not dried with sodium sulfate. Ap-proximately 10 g homogenised (wet) sample is weighed in a 40 mL I-Chem vial.After addition of the internal standard(s) and 10 mL acetone, the sample is ex-tracted for 15 min by sonication. Then 10 mL cyclohexane is added and the son-ication extraction is repeated for another 15 min. After this first extraction, thevials are placed on a shaking machine for 30 min. Finally, the vials are againplaced in the ultrasonic bath for 15 min. After completion, the extraction vial iscentrifuged. An aliquot (5 mL) of the (upper) cyclohexane phase is transferredto another tube and concentrated to 1 mL

Clean-up of the extracts can be done by column chromatography or by solid-phase extraction. The procedures are identical to the procedures described forsolid samples [6,22]. By using column chromatography, the concentrated extractis transferred onto an alumina column packed with 15 g deactivated alumina(15% water w/w) and with a 1–2 cm bed of anhydrous sodium sulfate on top. Thecolumn is eluted with 30 mL hexane, followed by 30 mL 10% dichloromethane inhexane and finally 30 mL 50% dichloromethane in hexane. This last fraction con-tains the phthalates. This fraction is concentrated to 1 mL and analysed. Theclean-up can be miniaturized by using a 500 mg alumina cartridge. After apply-ing the extract, the cartridge is washed with 3 mL hexane and the phthalates areeluted with 3 mL 75% hexane-25% MTBE.

The limit of detection of these methods is in the order of 2–10 µg kg–1 wetsample for the single isomer phthalates (and 20–100 µg kg–1 wet sample for themixed isomer phthalates). However, method blanks are typically in the order of10–20 ppb, especially for DiBP, DBP and DEHP. For these compounds, the limitof quantitation is around 20–40 µg kg–1 wet sample (set to two times the back-ground level).

Analytical Methods Review 47

7.3Analytical Procedure for the Determination of Phthalates in Milk or Edible Oils and Fat

For the determination of phthalate in milk samples the following method was de-veloped [7]. Milk samples are first homogenised by shaking. Milk powders aredissolved in water (1 g/10 mL). Approximately 5 g milk sample is then extractedwith 20 mL of a 1 : 1 cyclohexane/acetone mixture in a 40 mL I-Chem vial. Thevials are shaken on a shaking machine for 30 min.After completion, the vials arecentrifuged and the supernatant is transferred to a pre-weighed I-Chem vial. Thesolvent is evaporated under nitrogen and the fat content is measured. The residueis dissolved in dichloromethane and internal standard (d4-DEHP) is added. Theamounts of solvent and internal standard are adjusted to give approximately50 mg fat and 100 ng internal standard per mL dichloromethane. The solutionwas then homogenised in a vortex agitator. The separation of the fat from the phthalate containing fractions is then performed by gel permeation chromatog-raphy (GPC).

For the determination of phthalates in edible oils and fat, the oil or fat samplecan by diluted directly in dichloromethane to a concentration of 50 mg fat permL. If the sample solution is not clear, some water might be present and this canbe removed by adding some anhydrous sodium sulfate to the sample. The clearsolution is also fractionated by GPC.

For the GPC separation of the dichloromethane solution, 500 µL is injectedonto a two column combination. This combination consisted of a 5 cm ¥ 7.5 mmi.d. PL-Gel pre-column and two 30 cm ¥ 7.5 mm i.d.¥ 5 µm PL-Gel 5 nmcolumns. The mobile phase is dichloromethane at 1 mL min–1 flow rate. UV de-tection at 220 nm is used to monitor the effluent. Phthalates typically elute in awindow between 20 and 23 min (3 mL fraction). This fraction is automaticallycollected. The total run time is 30 min. To the collected fraction, 100 µL cyclo-hexane is added and the extract is concentrated to 100 µL. This extract is analysedby GC-MS.

The limit of detection of this method is around 20 ng g–1 fat. DiBP, DBP andDEHP are detected in the method blanks, but the levels are constant around20–50 ng g–1 fat. Hence, the limit of quantification for these phthalates is40–100 ng g–1 fat (two times background level). For milk samples with 3% fatcontent, this corresponds to a quantification limit of 1–3 ng g–1 milk.

The recovery of the GPC clean-up method was tested by spiking an olive oilsample at a 500 ppb level with DMP, DEP, DiBP, DBP, BBzP and DEHP and at a 5 ppm level with DiNP and DiDP. The olive oil was tested before and no phthalates were detected at concentrations above 50 ppb (500 ppb for DiNP andDiDP). The results from the non-spiked oil were therefore considered as the “pro-cedure blanks”. D4-DEHP (internal standard) was added to the dichloromethanesolution. The sample was analysed in triplicate. The linearity was tested by spiking an olive oil sample at seven levels (100, 250, 500, 1000, 2500, 5000 and7500 ng g–1 fat) (¥ 10 for mixed isomers). The mean recovery, standard deviation(RSD %) and linearity (correlation coefficient r2) are listed in Table 7. In general,good recoveries (>80%) are obtained, except for butylbenzyl phthalate for whichthe recovery is lower. This is probably due to a slightly different behaviour of this

48 F. David et al.

compound in sample clean-up (GPC). The correlation coefficients are better than0.99 for all phthalates, except for DiNP (0.98). This correlation shows that themethod can be used in the concentration range from 100–7500 ppb for the sin-gle isomer phthalates and from 1–75 ppm for the mixed isomer phthalates.

An alternative and automated method for the determination of phthalates inoil and fat matrices was described by Pacciarelli et al. [64]. In this method, an on-line HPLC-GC method was used. The sample fractionation was performed byHPLC in straight (normal) phase mode. Fractionation of the triglycerides wasdone on a silica column (100 ¥ 4.6 mm i.d. Lichrosorb SI-100, 5 µm) using1 : 1 dichloromethane/cyclohexane (with 0.5% acetonitrile) as mobile phase at1 mL min–1. The fraction containing DEHP was automatically transferred in theGC and the triglyceride matrix is eluted afterwards.

7.4Analytical Procedure for the Determination of Phthalates in Fish

Depending on the type of fish and available sample size, the fish samples can beanalysed as whole fish, as fillet (muscle with skin removed) or the analysis canbe performed on specific organ samples (liver, stomach). For diet studies, the fil-let sample represents the edible part of the fish. For contamination studies, wholefish samples or organ analyses give relevant information. Due to their lipophiliccharacter, phthalates accumulate in the fat. The analytical methods for the de-termination of phthalates in fish consist in general of a fat (+phthalate) extrac-tion step, followed by a clean-up step. Different extraction techniques can beused. These include ultrasonic extraction, Soxhlet extraction,ASE, etc. For clean-up, gel permeation chromatography and column chromatography or solid phaseextraction can be used. A fish matrix is quite complex and it is advisable to useboth GPC plus an additional alumina column clean-up based on polarity differ-ences to remove fats and other interfering compounds that are co-extracted.

Liu et al. [6] used ultrasonic extraction. The samples are homogenised un-frozen. Analytical subsamples, 5 g wet weight, are spiked with d4-labelled stan-dards and mixed with 30 g anhydrous sodium sulfate in a mortar until a drypowder is obtained. This homogenised and dried sample is extracted with 30 mLdichloromethane by sonication for 10 min and shaking for another 10 min.After the suspended particles are settled, the supernatant is removed and the extraction is repeated twice with fresh solvent. Finally, the three dichloromethanefractions are combined and concentrated to 1 mL. The extracts are first cleanedby GPC. The phthalate fraction is concentrated and then transferred onto an alu-

Analytical Methods Review 49

Table 7. Validation data for the determination of phthalates in oil

DMP DEP DiBP DBP BBzP DEHP DiNP DiDP

Recovery (%) 103 134 104 138 69 127 85 81RSD (%) 15.7 15.7 9.5 11.8 16.0 12.1 11.6 10.6Linearity (r2) 0.9956 0.9946 0.9992 0.9962 0.9947 0.9978 0.9841 0.9903

mina column packed with 15 g deactivated alumina (15% water w/w) and with1–2 cm bed of anhydrous sodium sulfate on top. The column is eluted with 30 mLhexane, 30 mL 10% dichloromethane in hexane and finally 30 mL 50% di-chloromethane in hexane. This last fraction contains the phthalates. This fractionis concentrated to 1 mL and analysed. By using this method, Liu et al. [6] obtained detection limits around 2 ng g–1. Due to method blanks, the limit ofquantification for DBP was around 20 ng g–1 and around 40 ng g–1 for DEHP. Therecovery was 70–95% for the single isomer phthalates, measured by GC-MS and75–110% for the mixed isomer phthalates measured by LC-MS.

The same preparation procedure was used by David et al. [7]. In this proce-dure, 5 g homogenised fish sample was mixed with internal standard (d4-DEHP)and with 15 g pre-cleaned sodium sulfate. Extraction was performed by sonica-tion with 20 mL cyclohexane for 15 min. The extract was centrifuged and the su-pernatant was removed. The extraction was performed another time with freshsolvent and the combined extracts were concentrated under nitrogen and frac-tionation by GPC, using the method described for milk samples. An example ofthe GPC separation for a fish extract is given in Fig. 16. The chromatogram ob-tained for a DEHP reference standard is overlayed. The DEHP peak elutes on thetail of the lipid peak. The fraction between 15 and 17 min is collected, concen-trated and analysed by GC-MS. In this fraction, some sterols elute that cannot be separated from DEHP. These compounds do not interfere with the furtheranalysis. The GC-MS chromatograms are shown in Fig. 17. The extracted ionchromatogram at m/z 149 shows the presence of DEHP. The extracted ion chro-matogram at m/z 153 shows the presence of the internal standard (d4-DEHP). Inthis case, no additional clean-up by column chromatography or by SPE was used.The recovery for DEHP measured at three spike levels between 2 ng g–1 (wet sam-ple) and 40 ng g–1 (wet sample) were 89–93%. The relative standard deviation at40 ng g-1 (wet sample) was 9.7%.

Letinski [25] used accelerated solvent extraction for the extraction of phtha-lates from fish tissue.Approximately 1 g fish sample was mixed with sodium sul-fate and extracted with a 90% hexane/10% ethyl acetate mixture at 120 °C and1600 psi. This extract was purified on alumina. The limit of quantification was150 ng g–1 fish sample and recovery was 98% for the recovery matrix spike.

8Sample Preparation Methods for Phthalic Acid Mono-Esters

Phthalic acid monoesters are more polar than phthalates and therefore the sam-ple preparation methods must be adapted. For the determination of monoestersin blood, plasma and urine samples, several methods are described in the liter-ature [8–10, 12, 13, 17, 18]. For the determination of monoesters in environ-mental samples, only limited data are available. Suzuki et al. [14] used solid-phaseextraction on polymeric phase disks (styrene divinylbenzene) for the determi-nation of monoesters in riverwater samples. The waters were acidified to pH 2before extraction to ensure the monoesters are in the protonated form (notionised). Extraction efficiency was higher than 72% (except for monomethyl phthalate).

50 F. David et al.

After extraction, the monoesters must be derivatised prior to GC analysis.Alternatively, HPLC of non-derivatised monoesters can be performed (seeabove).

Recently, a method has been developed for the analysis of phthalic acid mo-noesters in water samples by using an in situ derivatisation, followed by stir barsorptive extraction and thermal desorption GC-MS [19]. One mL of water sam-ple is placed in a 20 mL glass vial and spiked with internal standard solu-tion. 500 µL of a 2 : 1 (v/v) mixture of ethanol and pyridine is added and the

Analytical Methods Review 51

Fig.

16.

GPC

frac

tion

atio

n of

phth

alat

es in

fish

ext

ract

52 F. David et al.

Fig. 17. GC-MS chromatograms of DEHP and internal standard (d4-DEHP) extracted from fish

mixture was homogenised by vortex agitation. After addition of 100 µL of ethylchloroformate, the mixture was again vortex homogenised and placed in an ul-trasonic bath for 15 min. During vortex and ultrasonic agitation gas evolved from the vial. The mixture was finally diluted with 10 mL water. The derivatisedmonoesters are extracted by stir bar sorptive extraction (SBSE), followed bythermal desorption-GC-MS. The limit of detection (LOD) is in the order of0.05 µg L–1 for MBP and MEHP. The analysis of a sample spiked at 1 µg L–1 levelsis shown in Fig. 18. The monoesters MBP, MEHP, MiNP and MiDP are easily detected.

Analytical Methods Review 53

Fig.

18.

Det

erm

inat

ion

ofph

thal

ic a

cid

mon

oest

ers

in w

ater

by

in s

itu

deri

vati

sati

on a

nd S

BSE-

ther

mal

des

orpt

ion-

GC

-MS

9Conclusions

During the past years, various sample preparation and analytical methods havebeen developed for the determination of phthalic acid diesters in different envi-ronmental samples. The major problem in phthalate analysis is the risk of con-tamination. Precautions have to be taken into account to minimize this risk andto control the background values.

Based on the authors’ personal experience, the following methods are recom-mended for the different matrices. Clean water samples can be extracted bymethods based on solid-phase extraction. For contaminated water samples,miniaturized liquid-liquid extraction can be used. Sediments, sludges and soilsamples can be extracted by using ultrasonic extraction. Biota samples can alsobe extracted by sonication. The extracts are purified by gel permeation chro-matography, followed by column chromatography or solid-phase extraction. Theanalysis of the single isomer phthalates can be done by GC-MS. HPLC-MS offersan interesting alternative for the analysis of mixed isomer phthalates.

For the analysis of monoesters in environmental samples, new methods arecurrently under development. The extraction and clean-up method used for phthalates need to be adapted. Specifically, the monoesters must be derivatisedprior to GC-MS analysis. Alternatively, they can be analysed without derivatisa-tion by using LC-MS.

10References

1. US EPA methods 525 (drinking water), 606 and 625 (waste water) and 8060 and 8270 (solidwaste). US Environmental Protection Agency, Cincinnati, Ohio 45268, USA

2. ISO draft international standard 18856, ISO. Geneva, Switzerland3. Vikelsoe J, Thomsen M, Johansen E (1998) Sources of phthalates and nonylphenols in

municipal waste water, technical report no 225 and 268. National Environmental ResearchInstitute (NERI), Roskilde, Denmark

4. Vikelsoe J, Thomsen M, Johansen E, Carlsen L (1999) Phthalates and nonylphenols in soil,technical report no 268. National Environmental Research Institute (NERI), Roskilde,Denmark

5. George C, Prest H (2001) Agilent technologies application note Nr 5988–2244EN6. Lin, Zhong-Ping, Ikonomou MG, Hongwu J, Macintosh C, Gobas FAPC (2002) Environ Sci

Technol submitted for publication7. David F, Tienpont B, Sandra T, Sandra P (2002) submitted for publication8. Sjöberg P, Bondesson U (1985) J Chromatogr 344:1679. Sjöberg P, Bondesson U, Sedin E, Gustafsson J (1985) Eur J Clin Invest (1985) 15:430

10. Sjöberg P, Bondesson U, Sedin E, Gustafsson J (1985) Transfusion 25:42411. Marschall H-M, Green G, Egestad B, Sjövall J (1988) J Chromatogr 452:45912. Egestad B, Green G, Sjöberg P, Klasson-Wehler E, Gustafsson J (1996) J Chromatogr B 677:9913. Haam D, Vandenbroek P, Jongeneelen F (1993) Int Arch Occupat Environ Health 64:55514. Suzuki T, Yaguchi K, Suzuki S, Suga T (2001) Environ Sci Technol 35:375715. Tienpont B, David F, Sandra P (2002) submitted for publication16. Barry Y, Labow R, Keon W, Tocchi M, Rock G (1989) J Thorac Cardiovasc Surg 97, 90017. Shintani H (2000) Chromatographia 52:72118. Blount B, Milgram K, Silva M, Malek N, Reidy J, Needham L, Brock J (2000) Anal Chem

72:4127

54 F. David et al.

19. Tienpont B, David F, Sandra P (2002) submitted for publication20. Furtmann K (1994) Fresenius J Anal Chem 348:29121. Parkman H, Remberger M (1995) Phthalates in Swedish sediments, IVL Report Publ no

1167, p 21+appendix22. Braaten B, Berge JA, Berglind L, Baekken T (1996) Occurence of Phthalates and Organotins

in sediments and water in Norway. Norwegian Institute for Water Research (NIVA) reportSNO 3552–96, p 45

23. Cousins I, Mackay D (2000) Chemosphere 41:138924. Van der Velde E, Korte de G, Versteegh A (1998) Determination of phthalate esters in

water by SPE or in-vial extraction with GC-MS analysis: how to avoid the contaminationproblem. Paper presented at 20th international symposium on capillary chromatography,Riva del Garda, Italy, 26–29 May 1998

25. Letinski DJ (1999) Lecture presented at Workshop Nov 4–526. Pawliszyn J (1999) Applications of solid phase micro-extraction, RSC chromatography

monographs. Royal Society of Chemistry, Letchworth, UK (ISBN 0–85404–525–2)27. Baltussen E, Sandra P, David F, Cramers C (1999) J Microcolumn Sep 11:73728. Steffen D, Lach G (2000) Phthalate und Trichlosan in Sedimenten und Schwebstof-

fen Niedersächsischen Gewässer, Niedersächsisches Landesamt für Ökologie, report10/2000)

29. Kolb M, Welte K, Mettenleiter S, Trinkmann A (1997) Wasser Boden 49:5730. Paxéus N (1999) submitted for publication31. Zurmühl T (1990) Analyst 115:117132. Thomas GH (1973) Environ Health Perspect 3 :2333. Giam CS, Chanm HS, Neff GS, Atlas EL (1978) Science 199:41934. Cautreels W, van Cauwenberghe K (1978) Atm Environ 12:113335. Cautreels W, van Cauwenberghe K, Guzman LA (1993) Sci Total Environ 8 :4736. Thúren A, Larsson P (1990) Environ Sci Technol 24:55437. Schulz H-M, Püttmann W (1993) Analysis of saturated hydrocarbons, fatty acids and

phthalic acid esters in air particulate matter of a city area (Aachen). Wissenschaft undUmwelt 2/1993, p 131

38. Chang LW, Atlas E, Giam CS (1985) Int J Environ Anal Chem 19:14539. Figge K, Rabel W, Wieck A (1987) Fresenius J Anal Chem 327:26140. Vainiotalo S, Pfaffli P (1990) Ann Occup Hyg 34:58541. California Environmental Protection Agency (1992) Monitoring of phthalates and PAHs in

indoor and outdoor air samples in riverside, California. Contract no A933-144, Dec 199242. Fisher J, Ventura K, Prokes B, Jandera P (1993) Chromatographia 37:4743. Fujimoto T, Takeda N, Taira T, Iikawa R (1995) Kurin Tekunoroji 5 :4544. Bartulewicz J, Bartulewicz E, Gawlowski J, Niedzielski J (1996) J Chem Anal (Warsaw)

41:75345. Weiling G, Xiku W (1997) Huanjing Huaxue 16:38246. Otake T, Yoshinaga J, Yanagisawa Y (2001) Environ Sci Technol 35:309947. Rudel RA, Brody JG, Sprengler JD, Vallarino J, Geno PW, Sun G, Yau A (2001) J Air Waste

Manage Assoc 51:49948. Baltussen E, Janssen HG, Sandra P, Cramers CA (1997) J High Resol Chromatogr 20:38549. Baltussen E, David F, Sandra P, Janssen HG, Cramers CA (1998) J High Resol Chromatogr

21:33250. Tienpont B, David F, Sandra P, Vanwalleghem F (2000) J Microcolumn Sep 12:19451. Lövkist P, Jönsson JA (1987) Anal Chem 59:81852. Namiesnik J, Gorecki T (2000) LC-GC Europe, September 2000, pp 67853. Martos PA, Pawliszyn J (1997) Anal Chem 69:20654. Martos PA, Pawliszyn J (1999) Anal Chem 71:151355. Khaled A, Pawliszyn J (2000) J Chromatogr 892:45556. David F, Sandra P (2001) Passive sorptive sampling for indoor air monitoring using poly-

dimethylsiloxane coated stir bas, paper 740, The Pittsburgh conference on analytical chem-istry and applied spectroscopy, 4–9 March 2001

Analytical Methods Review 55

57. ACGIH technical committee on air sampling procedures (1984) Particle size-selection sam-pling in the workplace. ACGIH, Cincinnati, Ohio

58. EN 481 (1993) European committee for standardisation CEN 1993-07–2759. VDI 4300 (1999) Measurement of indoor air pollution, sampling of house dust. Verein

Deutscher Ingenieure, Düsseldorf60. Mark D, Vincent JH (1986) Ann Occup Hyg 30:8961. Bruns-Weller E, Pfordt J (2000) Z Umweltchem Okotox 12:12562. Specht W, Tilkes M (1980) Frezenius Z Anal Chem 301:30063. Thier H-P, Zeumer H (1987) Manual of pesticide residue analysis, vol 1. VCH, Weinheim,

p 7564. Pacciarelli B, Müller E, Schneider R, Grob K, Steiner W, Fröhlich D (1988) J High Resol

Chromatogr 11:135

56 F. David et al.

© Springer-Verlag Berlin Heidelberg 2003

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters

Ian T. Cousins 1 · Donald Mackay 1 · Thomas F. Parkerton 2

1 Canadian Environmental Modelling Centre, Environmental and Resource Studies,Trent University, Peterborough, Ontario, K9J 7B8, Canada. E-mail: dmackay@trentu.ca

2 Exxon Mobil Biomedical Sci. Inc., Hermeslaan 2, 1831, Machelen, Belgium

A review is presented of the physical-chemical properties and reactivity of the phthalate estersincluding a discussion of how these properties control their partitioning and fate in the envi-ronment. The air and water solubilities decrease by orders of magnitude from the short alkylchain phthalates such as dimethyl phthalate (DMP) to the long alkyl chain phthalates such asdi-2-ethylhexyl phthalate (DEHP). The octanol-water partition coefficient, which is a measureof hydrophobicity, increases by orders of magnitude with increasing alkyl chain length and thisincrease is mainly controlled by the reduction in water solubility rather than an increase in oc-tanol solubility. This increase in hydrophobicity results in strong sorption of the higher mol-ecular weight phthalates to organic matter.Air-water partition coefficients (or Henry’s law con-stants) also increase with increasing alkyl chain length. However, the greater evaporativepotential of higher molecular weight phthalate esters from water is offset by sorption to sus-pended matter in surface waters. Phthalates have high values of KOA suggesting that they willbe appreciably sorbed to aerosol particles, soil and vegetation. From available data obtained un-der environmental conditions, half-lives of phthalates in environmental media are proposed.Systematic differences in reactivity or half-life are apparent, with the primary biodegradationhalf-life tending to increase with increasing alkyl chain length. In contrast, the opposite pat-tern is observed for the air oxidation half-life. A series of evaluative modelling calculations isdescribed to illustrate how the physical-chemical properties result in differences in environ-mental partitioning behaviour, persistence and transport potential. In comparison to other or-ganic chemical classes, model results indicate that phthalates are not environmentally persis-tent or subjected to significant long-range transport. Although the overall environmentalpersistence of the higher molecular weight phthalates tends to increase, KOA and thus thepropensity to partition to aerosols, vegetation and soils also increases, thereby reducing the po-tential for long-range transport. Recommendations for future research on physical-chemicalproperties of phthalate esters for environmental fate assessment are discussed.

Keywords. Phthalate ester, Structure, Physical-chemical property, Model, Fate

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

2 Structure-Property Analysis of Physical-Chemical Properties . . 58

3 Observations on Physical-Chemical Properties . . . . . . . . . . 64

3.1 Physical State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643.2 Solubility in Water . . . . . . . . . . . . . . . . . . . . . . . . . . 643.3 Vapour Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 57–84DOI 10.1007/b11463

3.4 Air-Water Partition Coefficient . . . . . . . . . . . . . . . . . . . 673.5 Octanol-Water Partition Coefficient . . . . . . . . . . . . . . . . . 683.6 Octanol-Air Partition Coefficient . . . . . . . . . . . . . . . . . . 713.7 Data Gaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

4 Degrading Reactions . . . . . . . . . . . . . . . . . . . . . . . . . 72

5 Evaluative Fate Modelling with the EQC Model . . . . . . . . . . . 74

5.1 EQC Level I Modelling . . . . . . . . . . . . . . . . . . . . . . . . 755.2 EQC Level II Modelling . . . . . . . . . . . . . . . . . . . . . . . . 765.3 EQC Level III Modelling . . . . . . . . . . . . . . . . . . . . . . . 765.4 Estimating Persistence and Long-Range Transport Potential

with the TaPL3 Model . . . . . . . . . . . . . . . . . . . . . . . . 80

6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

1Introduction

Phthalate esters are widely used as plasticizers, serving as important additivesthat impart flexibility to polymers including poly(vinyl chloride) (PVC), poly-vinylacetates, cellulosics and polyurethanes [1]. The stability, fluidity and lowvolatility of high-molecular mass phthalate esters make them ideal for use asplasticizers.

The variety of possible chemical structures of phthalate esters results in a widerange of physical-chemical properties and hence environmental partitioning be-haviour for this class of compounds. This wide range of properties is principallya result of the variation in the length of the alkyl chains substituted on the diestergroups. The names, molecular formulae, molar masses and melting points of22 phthalate esters are listed in Table 1.

The objectives of this chapter are to review the published physical-chemicaland reactivity data for the phthalate esters, seek relationships between chemicalstructure and properties and determine how these properties will influence par-titioning between abiotic media in the environment with the use of evaluative en-vironmental fate models. The accumulation of phthalate esters in biotic media(i.e. food webs) is the focus of a separate chapter in this volume.

2Structure-Property Analysis of Physical-Chemical Properties

Physical-chemical properties which can be measured readily in the laboratorywith a view to determining environmental partitioning include: solubility in water, vapour pressure, the Henry’s law constant (H), the octanol-water partitioncoefficient (KOW) and the octanol-air partition coefficient (KOA). There are few direct measurements of Henry’s law constants for the phthalate esters and no

58 I.T. Cousins, D. Mackay and T.F. Parkerton

measurements of KOA, although, as discussed later, both these properties are eas-ily estimated as ratios of solubilities in air, water and octanol. Solubilities in water, vapour pressures and KOW for 22 phthalate esters, which were taken fromthe review by Cousins and Mackay [2], are summarized in Table 2.

For a specific physical-chemical property the published values in Table 2sometimes vary by several orders of magnitude. For example, the reported aqueous solubilities of di-n-octyl phthalate at 25 °C vary between 0.4 and 3000 µgL–1, that is by a factor of 7500. This variation in reported physical-chemical prop-erties is initially surprising because the values are physical constants that shouldbe known precisely. However, such variation in reported values is not unusual[36] and occurs because there are a wide range of different measurement tech-niques used for a given property, a wide range of laboratories undertaking thetests, a number of measurement difficulties that need to be overcome (e.g. analy-sis of concentrations close to detection limits, emulsion formation, sorption toglassware etc.) and a number of errors or oversights made by researchers.

Selecting a “best value” for a given physical-chemical property is problematic;approaches include: (1) expert judgment, (2) using a list or checklist of criteriato assess the quality of published value (e.g. see ref. [37]) or (3) using structure-

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 59

Table 1. List of phthalate esters studied and their associated molar masses, molar volumes andmelting points

Phthalate ester Abbre- Molar Le Bas molar Meltingviation mass volume point

(g mol–1) (cm3 mol–1) (°C)

Dimethyl phthalate DMP 194.2 206.4 5.5Diethyl phthalate DEP 222.2 254.0 –40Diallyl phthalate DAP 246.2 283.6 –Dipropyl phthalate DPP 250.3 298.4 –Di-n-butyl phthalate DnBP 278.4 342.8 –35Disiobutyl phthalate DIBP 278.4 342.8 –58Di-n-propyl phthalate DnPP 250.3 387.2 –Butylbenzyl phthalate BBP 312.4 364.8 –35Diisohexyl phthalate DHP 334.4 431.6 –27.5Di-n-heptyl phthalate DIHpP 362.5 476.0 –Di-n-octyl phthalate DnOP 390.6 520.4 –Butyl 2-ethylhexyl phthalate BOP 334.4 416.6 –37Di(n-hexyl, n-octyl, n-decyl) phthalate a 610P 404.6 542.6 –4Di(2-ethylhexyl) phthalate DEHP 390.6 520.4 –46Diisooctyl phthalate DIOP 390.6 520.4 –46Di-n-nonyl phthalate DnNP 418.6 564.8 –Diisononyl phthalate DINP 418.6 564.8 –48Di-n-decyl phthalate DnDP 446.7 609.2 –Diisodecyl phthalate DIDP 446.7 609.2 –46Di(heptyl, nonyl, undecyl) phthalate a D711P 418.7 564.8 <–50Diundecyl phthalate DUP 447.7 653.6 –9Ditridecyl phthalate DTDP 530.8 742.4 –37

a These are mixtures of three phthalate esters. Le Bas molar volumes were calculated using themethod in ref. [3]. Melting points were taken from ref. [4].

60 I.T. Cousins, D. Mackay and T.F. Parkerton

Table 2. Physical-chemical properties of phthalate esters at 25 °C that were used in the “three-solubility analysis” [2]

Phthalate ester CSWL

a (mg L–1) PLb (Pa) Log KOW

c

DMP 2810 [5] 0.22 [6] 1.46 [5]4000 [6] 1.47 [4]4248 [7] 1.53 [6]4290 [8] 1.56 [5]4320 [9] 1.61 [5]

1.61 [10]1.62 [11]1.66 [10]1.74 [10]1.90 [10]1.60 [12]

DEP 680 [13] 0.052 [14] 2.21 [11]896 [9] 0.0813 [14] 2.21 [11]928 [8] 0.22 [6] 2.24 [6]1080 [6] 2.29 [6]

2.35 [8]2.67 [7]3.00 [15]2.42 [12]

DAP 182 [8] 3.23 [8]DPP 108 [8] 3.27 [8]DnBP 9.15 [16] 2.27 ¥10–3 [17] 3.74 [8]

10.1 [8] 2.53¥10–3 [18] 4.08 [19]11.2 [6] 2.77¥10–3 [20] 4.11 [21]13 [9] 4.67¥10–3 [22] 4.13 [12]

4.80¥10–3 [23] 4.30 [24]5.47¥10–3 [19] 4.39 [18]9.73¥10–3 [6] 4.56 [19]

4.57 [8]4.72 [25]4.72 [16]4.79 [6]5.15 [7]4.50 [11]

DIBP 6.2 [26] 4.11 [8]20 [27]20.3 [8]

DnPP 5.62 [12]BBP 0.7 [26] 1.15 ¥10–3 [27] 3.57 [6]

2.69 [6] 1.20 ¥10–3 [6] 3.97 [11]2.82 [8] 4.05 [7]2.9 [9] 4.11 [21]40.2 [7] 4.75 [7]

4.77 [21]4.77 [27]4.91 [8]4.73 [12]

DHP 0.24 [6] 2.40 ¥10–4 [22] 5.65 [6]7.00 ¥10–2 [28] 1.87 ¥10–3 [6] 5.93 [6]4.60 ¥10–2 [29] 6.82 [12]

property relationships (SPRs) to identify potential errors in reported data. Thestructure-property method is particularly attractive because it is less subjectivethan the other two approaches and is well suited to analysing structurally simi-lar compounds or a homologous series of compounds such as the phthalate es-ters. Recently, Thomsen et al. [38] used SPR concepts for correlating water solu-bilities and partition coefficients of phthalate esters against a variety of moleculardescriptors and Cousins and Mackay [2] used the “three-solubility” analysis,which correlates solubilities of phthalates in air, water and octanol against mo-lar volume.

In this chapter, we include the solubilities and partition coefficients that wereestimated by Cousins and Mackay [2] by using the “three-solubility” approach

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 61

Table 2 (continued)

Phthalate ester CSWL

a (mg L–1) PLb (Pa) Log KOW

c

DIHpP 1.70 ¥10–2 [28]DnOP 0.02 [16] 2.53 ¥10–2 [18] 5.22 [8]

0.02 [26] 7.06 [30]3 [9] 8.06 [15]4.00 ¥10–4 [28] 8.10 [12]5.10 ¥10–4 [29] 8.18 [29]

610P 0.9 [6] 6.53 ¥10–4 [6] 7.25 [6]DEHP 0.041 [8] 5.47 ¥10–4 [31] 5.22 [32]

0.315 [16] 1.31 ¥10–5 [21] 5.11 [33]0.285 [26] 3.73 ¥10–5 [18] 7.06 [30]0.34 [24] 4.40 ¥10–5 [21] 7.14 [30]0.4 [13] 9.47 ¥10–5 [17] 7.45 [30]1.9 ¥10–3 [9] 8.53 ¥10–4 [6] 7.45 [32]

8.35 [34]8.05 [32]8.05 [32]7.94 [6]8.06 [15]7.27 [12]

DIOP 0.09 [6] 1.87 ¥10–4 [18]7.47 ¥10–4 [6]

DnNP 1.3 ¥10–4 [9]DINP 6.00 ¥10–4 [35] 7.20 ¥10–5 [6]

0.2 [6]1.1 ¥10–4 [28]

DIDP 0.28 [27] 6.80 ¥10–6 [23]1.7 ¥10–4 [28]1.19 [6]

DnDP 5.00 ¥10–5 [28] 8.91 [29]2.20 ¥10–4 [29]

D711P 6.00 [6]DUP 1.1 [6]DTDP 0.34 [26]

a CSWL is the solubility of liquid phthalate in water.

b PL is the liquid vapour pressure.c KOW is the octanol-water partition coefficient.

[39] (Table 3). Briefly, in the “three-solubility” approach correlations are soughtbetween the solubilities or “apparent-solubilities” of liquid-state compounds inair, water and octanol and molecular descriptors. Cousins and Mackay [2] usedLe Bas molar volume [3] as the molecular descriptor, which is readily calculatedby summing atomic volumes with adjustment for the volume decrease arisingfrom ring formation. Figures 1–3 show the relationships between molar volumesand the three solubilities with linear regression trend-lines plotted through thepoints. Data points that were identified as unreliable by Cousins and Mackay [2]have been removed from these plots. The air solubility regression line has beenextrapolated beyond the data range to show the expected solubility in air of heav-ier phthalates. It is currently not possible to reliably measure such low vapourpressures; thus, estimates are used based on extrapolation of the linear regressionequation. From the correlations between solubilities and Le Bas molar volume,the partition coefficients KAW (air-water), KOW (octanol-water) and KOA (octanol-air) were deduced as the ratios of the “three-solubility” correlations. Further de-tails of the analysis of the physical-chemical properties of phthalate esters by the“three-solubility” approach are described in ref. [2].

The “three-solubility” analysis was particularly useful in highlighting the mea-surement errors in vapour pressures and solubilities in water for the longer chainphthalate esters. Furthermore, it validated some of the recent meticulous mea-surements of solubility in water and log KOW carried out by Letinksi et al. [28] and

62 I.T. Cousins, D. Mackay and T.F. Parkerton

Table 3. Calculated physical-chemical properties of phthalate esters at 25 °C [2]

Phthalate CSWL PL Log KOW Log KOA Log KAW H

ester (mg L–1) (Pa) (Pa m3 mol–1)

DMP 5220 0.263 1.61 7.01 –5.40 9.78 ¥10–3

DEP 591 6.48 ¥10–2 2.54 7.55 –5.01 2.44 ¥10–2

DAP 156 2.71 ¥10–2 3.11 7.87 –4.76 4.28 ¥10–2

DPP 77 1.75 ¥10–2 3.40 8.04 –4.64 5.69 ¥10–2

DnBP 9.9 4.73 ¥10–3 4.27 8.54 –4.27 0.133DIBP 9.9 4.73 ¥10–3 4.27 8.54 –4.27 0.133DnPP 1.3 1.28 ¥10–3 5.12 9.03 –3.91 0.302BBP 3.8 2.49 ¥10–3 4.70 8.78 –4.08 0.205DHP 0.159 3.45 ¥10–4 6.00 9.53 –3.53 0.726DnHP 0.159 3.45 ¥10–4 6.00 9.53 –3.53 0.726DIHpP 2.00 ¥10–2 9.33 ¥10–5 6.87 10.04 –3.17 1.69DnOP 2.49 ¥10–3 2.52 ¥10–5 7.73 10.53 –2.80 3.95BOP 0.385 5.37 ¥10–4 5.64 9.37 –3.73 0.466610P 8.76 ¥10–4 1.31 ¥10–5 8.17 10.78 –2.61 6.05DEHP 2.49 ¥10–3 2.52 ¥10–5 7.73 10.53 –2.80 3.95DIOP 2.49 ¥10–3 2.52 ¥10–5 7.73 10.53 –2.80 3.95DnNP 3.08 ¥10–4 6.81 ¥10–6 8.60 11.03 –2.43 9.26DINP 3.08 ¥10–4 6.81 ¥10–6 8.60 11.03 –2.43 9.26DnDP 3.81 ¥10–5 1.84 ¥10–6 9.46 11.52 –2.06 21.6DIDP 3.81 ¥10–5 1.84 ¥10–6 9.46 11.52 –2.06 21.6D711P 3.08 ¥10–4 6.81 ¥10–6 8.60 11.03 –2.43 9.26DUP 4.41 ¥10–6 4.97 ¥10–7 10.33 12.02 –1.69 50.5DTDP 7.00 ¥10–8 3.63 ¥10–8 12.06 13.01 –0.95 275

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 63

Fig. 1. Relationship between solubility of phthalate esters in water and molar volume

Fig. 2. Relationship between solubility of phthalate esters in air and molar volume

Ellington [29], which were found to be consistent with the values estimated by the“three-solubility” analysis (compare measured data listed in Table 2 with esti-mated values in Table 3).

3Observations on Physical-Chemical Properties

3.1Physical State

The phthalate esters contained in Table 1 are liquids at typical environmentaltemperatures. Melting points lie between 5.5 °C and –58 °C, and boiling points are between 230 and 486 °C [4]. Thus at low environmental temperatures some phthalates have the potential to be present in the solid state.

3.2Solubility in Water

A declining trend in solubility in water is observed with increasing alkyl chainlength or molar volume (Table 3 and Fig. 1). The phthalates have a remarkablylarge variation in their solubility in water with DMP being moderately soluble atª 5 g L–1 and DTDP having an estimated solubility some 11 orders of magnitude

64 I.T. Cousins, D. Mackay and T.F. Parkerton

Fig. 3. Relationship between solubility of phthalate esters in octanol and molar volume

lower at 7 ¥10–11 g L–1. The variability in reported solubility data (Table 2) tends to increase as the measured solubilities decrease, mainly as a result of increased difficulty of measurement. Measuring solubilities in water can be undertaken fairly accurately by standard methods for compounds with solubilities greater than 1.0 mg L–1, but below this limit problems occur;measurement below 1.0 µg L–1 is often unreliable. As discussed by Staples et al. [4], measurement problems are caused by emulsion formation during flask shaking, contamination from apparatus and solvents, and insufficiently low analytical detection limits. Each of these problems leads to measurementsthat overestimate the true water solubility. Furthermore, Thomsen et al. [40] have argued that phthalate esters with long alkyl chains exhibit weak surface activity, which may cause them to have measured solubilities that are higher than the actual unimeric saturation due to formation of micelles or emulsions.While the water solubility estimate of 15 µg L–1 reported for DEHP by Thomsenet al. based on surface tension measurements is much lower than found in early studies, this technique provides only an indirect measure of solubility. Thisis due to the fact that surface tension measurements can be affected by im-purities present in the test sample (i.e. traces of unreacted monoester). Solubil-ities in water were directly measured by a “slow-stir” technique to prevent the formation of emulsions and “state-of-the-art” analytical methodology to preventcontamination and to obtain low detection limits [28–30]. “Slow-stir” studies are believed to provide the most accurate measurements of unimeric satura-tion for phthalates with long alkyl chains, particularly because experimental data were validated by the “three-solubility” analysis of Cousins and Mackay [2]. Turner and Rawling [41] have recently provided additional experimental confirmation concluding that the true water solubility of DEHP is in the few mg L–1 range.

Ellington [29] has suggested that phthalate esters with long alkyl side chains(e.g. DnDP) may rotate and fold in aqueous solution to assume conformations oflower energy that more closely resemble a branched-alkyl chain. Furthermore,it is argued that the “effective” molar volume of the “folded” configuration ofthese phthalates would be less than the molar volume of the unfolded compoundsand thus their solubility in water would be higher. It is not possible at this point,with the available data, to properly test this hypothesis, but Cousins and Mackay[2] report a strong linear correlation with the Le Bas Molar volume, which doesnot take folding into account, for both short- and long-chain phthalates,suggesting that the “folding effect” does not significantly reduce solubilities.Branched-chain isomers of chemicals in general have greater water solubilitiesthan the straight-chain isomers [42], but it was also not possible to observe systematic differences (Table 2) between the straight- and branched-chain phthalates, possibly because measurement error between the various studies is greater than the actual difference in solubility. However, in the study by Letinksiet al. [28], branched isomers were found to systematically exhibit higher so-lubilities. For example, di-2-ethylhexyl phthalate was found to have a four-fold greater measured solubility than di-n-octyl phthalate. Similar trends were observed for linear and branched analogues with alkyl chains of nine or ten carbons.

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 65

The influence of salinity on the aqueous solubility of non-ionic organic chem-icals is given by the equation [43]:

log [Csat, w/Csat, salt] = Ks Ms (a)

where Csat,w denotes the solubility in deionised water, Csat, salt denotes the solubil-ity in the presence of dissolved salts, KS is the Setschenow constant and MS is thetotal molar concentration of dissolved salts. Turner and Rawling [41] have re-cently conducted careful solubility experiments with DEHP at different salinitiesby using radiotracer techniques. Experimental data were fitted to the above equa-tion yielding an estimate for KS of 1.25 L mol–1.Assuming that seawater has a dis-solved salt concentration of 0.5 M application of the above equation indicates theDEHP is approximately four times less soluble in salt versus fresh water. This pre-diction is in reasonable agreement with the ratio of measured values reported forDEHP in lab water (1.9 µg L–1) and seawater (0.6 µg L–1) [20, 21].

3.3Vapour Pressure

Vapour pressures similarly show a declining trend with increasing alkyl chainlength or molar volume, although the decreasing trend is not as pronounced as for aqueous solubilities. DMP has a vapour pressure of ª 0.3 Pa, that is some seven orders of magnitude higher than the estimated vapour pressure of4 ¥ 10–8 Pa for DTDP. Measured vapour pressures of a given compound are alsohighly variable, with variability increasing with increasing alkyl chain length.Measurement problems are encountered below 10–4 Pa and measurement below10–6 Pa is often unreliable. As discussed by Cousins and Mackay [2], it is proba-bly more accurate to estimate vapour pressures below 10–4 by extrapolation of therelationship between vapour pressure and molar volume than to rely on mea-sured values. It is also possible to extrapolate data from higher temperatures.Unpublished vapour pressure measurements at various temperatures above 60 °Cwere obtained for DEHP, DINP and DIDP from industry. These data were ob-tained by using a dynamic vapour pressure balance that is suitable for vapourpressure determinations for these substances at elevated temperatures. Experi-mental data were fitted to the Clausius-Clapeyron equation:

ln PL = A + B/T (b)

where PL is the vapour pressure in Pa, T is the absolute temperature (K) and A andB are empirical constants derived by linear regression. Results of regressionanalyses are summarized in Table 4 and indicated a good fit to the above equa-tion as shown in Fig. 4. Extrapolation of the resulting equations to 25 °C yieldedvapour pressure estimates which can be compared with recommended estimatesderived from the “three-solubility”approach as shown in Table 4.While estimatesobtained from high-temperature extrapolation differed by only a factor of threefor DEHP, extrapolated values were more than an order of magnitude higher thanvalues recommended by Cousins and Mackay [2].

Given the discrepancy in the vapour pressure estimates obtained from theseindependent approaches, it is recommended that vapour pressures of high-

66 I.T. Cousins, D. Mackay and T.F. Parkerton

molecular weight phthalates be investigated by using “state-of-the-art” methodsat environmentally relevant temperatures. For example, a generator columnmethod could be used [44] in which large volumes of air are passed through atemperature-controlled glass column containing glass wool coated with liquidphthalate. The flow rate through the column is such that the air becomes satu-rated with phthalate and a sorbent trap collects the phthalate in the air on exit-ing the column.

3.4Air-Water Partition Coefficient

The equilibrium distribution of a substance between air and water is known asthe air-water partition coefficient (KAW) and is used to indicate the tendency ofa substance to escape from water to air. It is noteworthy that volatility of the puresubstance is characterized by vapour pressure.Volatility from water is expressedas KAW and for volatilisation from organic media KOA is preferred. KAW is moreusually measured in the form of a Henry’s law constant (H, Pa mol m–3), which

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 67

Fig. 4. Vapour pressure data for selected phthalates at elevated temperature

Table 4. Results of regression analysis for vapour pressure measurements obtained at elevatedtemperatures and comparison of estimates extrapolated to 25 °C to reported literature

Phthalate A B r2 n PL @25 °C (Pa) PL @25 °C (Pa)Extrapolated Recommendeda

DEHP 31.08 12,090 0.998 10 7.6 ¥10–5 2.5¥10–5

DINP 28.06 11,281 0.999 87 5.6¥10–5 6.8¥10–6

DIDP 27.17 11,072 0.997 34 4.6¥10–5 1.8¥10–6

a Based on “three-solubility” correlation approach [2].

can also be calculated by dividing the vapour pressure (Pa) by the solubility inwater (mol m–3). Cousins and Mackay [2] used the “three-solubility” analysis (seeTable 3) to calculate values for H and derive values of KAW (from the relationshipKAW = H/RT, where R is the gas constant J mol–1 K–1 and T is the absolute tem-perature). The lower molecular weight phthalate esters have fairly high vapourpressures, but because of their high solubility in water they have very low H orKAW values. Thus, they volatilise fairly rapidly from the pure state but only veryslowly from aqueous solution. Because solubilities in water decrease more thanvapour pressures (see Figs. 1 and 2) with increasing alkyl chain length (or mo-lar volume), the air-water partition coefficients apparently increase with in-creasing molecular weight. Thus, the higher molecular weight phthalate esterswill potentially evaporate more rapidly from water, but this behaviour is miti-gated by sorption to suspended matter in the water column. To illustrate the com-bined effects of KAW and sorption on air-water distribution it is useful to calcu-late the concentration in air CA in equilibrium with unit total concentration inwater CTW, that is the total of both dissolved and sorbed forms. Now, CA is KAW CW,where CW is the dissolved concentration. Furthermore, CW/CTW or Fdis is the frac-tion dissolved and can be evaluated by

Fdis = 1/(1 + Msusp foc Koc) (c)

where Msusp is the suspended solids concentration in kg L–1, foc is the organic car-bon (OC) fraction of the suspended solids and Koc is the organic carbon nor-malized solids-water partition coefficient in L kg–1 OC. It follows that

CA = KAW Fdis CTW (d)

Using typical Msusp and foc values of 1.5 ¥ 10–5 kg L–1 and 0.1 and assuming as afirst approximation KOC = 0.35 KOW [45], log CA is plotted in Fig. 5 as a function ofmolar volume for CTW = 1.0. The concentration in air therefore rises at lower molar volumes (because of increasing KAW), but above 400 cm3 mol–1 it drops because Fdis becomes smaller. The net result is that higher molecular weight phthalate esters volatilise only slowly from water.

3.5Octanol-Water Partition Coefficient

The equilibrium distribution of a substance between water and octanol is knownas the octanol-water partition coefficient and is often used to predict the ex-pected partitioning in the environment between water and animal/plant lipidsand water and sediment/soil organic matter. There is an abundance of correla-tions between KOW and measured environmental partition coefficients (e.g.bioconcentration factors, see review by Gobas et al. [46]; soil/sediment organiccarbon-water partition coefficients, see reviews by Seth et al. [46] and Doucette[47]; and plant lipid-water partition coefficients, see review by McLachlan [48]).

The most common method of measuring KOW is to shake the test substance in a two-phase mixture of octanol and water and to measure the resulting equi-librium concentration in both phases [21]. One of the problems with this technique is that for low solubility, hydrophobic compounds, shaking promotes

68 I.T. Cousins, D. Mackay and T.F. Parkerton

Fig. 5. Preliminary analysis illustrating influence of sorption processes on volatilisation po-tential of phthalates from surface waters. CA is the concentration in air in equilibrium with unittotal (dissolved and sorbed) concentration in water

the formation of emulsions, which will cause concentrations in the water phaseto be higher than the solubility in water. To minimize the formation of emul-sions the “slow-stir” method has been developed and is believed to give more reliable results for the phthalate esters [29]. Several alternative methods for mea-suring KOW are available including the use of high-performance liquid chro-matography [49].

Log KOW of phthalate esters increases with increasing alkyl chain length or mo-lar volume indicating greater hydrophobicity. For example, the KOW of DMP ofª 66 is about ten orders of magnitude lower than the estimated KOW of DTDP of1012. Cousins and Mackay [2] have shown that solubilities in octanol are muchless sensitive to changes in molar volume than solubilities in water and air. Sol-ubilities in octanol decrease by only about one order of magnitude from DMP toDTDP, whereas solubilities in water decrease by 11 orders of magnitude. There-fore, the reason for the increase in KOW with increasing alkyl chain length is thatsolubilities in water decrease more per unit volume in molar volume than the sol-ubilities in octanol. It is noteworthy that lipophilicity (which may be thought ofas lipid or octanol solubility) does not increase with increasing alkyl chain lengthas is often wrongly asserted, but actually decreases slightly.

The high KOW values of the phthalates indicate these substances are very hydrophobic and will sorb strongly to organic matter and surfaces. However, thepotential for bioaccumulation is mitigated by biotransformation as discussed inthe chapter focussing on bioaccumulation of phthalates.

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 69

One of the major uses of KOW is to estimate the organic-carbon water partitioncoefficient (KOC) for soils and sediments and thus the soil/sediment-water dis-tribution coefficient (KD). Staples et al. [2] reviewed KOC measurements for phthalates. When KOC and KOW are correlated there is a linear relationship be-tween KOW and KOC for the lower molecular weight phthalates which follows therelationship proposed by Seth et al. [46]: KOC = 0.35 KOW. However, the linear relationship fails for phthalates with a KOW greater than 106 with the result thatKOC is poorly predicted by KOW (Fig. 6).

70 I.T. Cousins, D. Mackay and T.F. Parkerton

Fig.

6.R

elat

ions

hip

betw

een

KO

Wan

d ex

peri

men

tal K

OC

data

.The

stra

ight

line

show

s the

rela

tion

ship

KO

C=

0.35

KO

W[4

6].

Expe

rim

enta

l KO

C d

ata

wer

e ta

ken

from

ref

.[4]

This discrepancy may be the result of problems in measuring the truly dis-solved concentration in water. The phthalates may also be in the form of emul-sions or attached to colloidal material, thus the dissolved concentration is over-estimated and KOC is underestimated. Recent support for the important role thatcolloids may serve in reducing apparent KOC values obtained in laboratory testsis provided by Zhou and Liu [50].An alternative explanation is given by Williamset al. [51] who have hypothesized that KOC is inversely dependent on solids con-centrations due to particle-particle collisions that induce desorption. These authors calculated particle-corrected KOC values, which were one to three ordersof magnitude higher than measured values.An inverse relationship between theDEHP solids-water partition coefficient and solids concentration has been re-ported in recent lab studies in both surface and marine water [41, 50]. This trendis also apparent from analysis of field monitoring data for DEHP presented byLong et al. [52]. However, it is difficult to differentiate the relative role of colloidbinding versus particle interactions, since colloid and solids concentrations typ-ically co-vary. A third explanation is that the time necessary to achieve equilib-rium sorption in the organic carbon is long (i.e. months or years) because of thelarge molar volume of the phthalate and its correspondingly slow diffusion [53].However, since both lab and field studies produce KOC values below theoreticalestimates derived from KOW correlations, the previous two explanations appearmore plausible.

Turner and Rawling [41] also found that the higher KOC values obtained in sea-water versus those in freshwater could not solely be explained by a “salting-out”effect. These authors have hypothesized that additional interactions between saltions and organic matter facilitate partitioning of DEHP to solids.

3.6Octanol-Air Partition Coefficient

Harner and Mackay [54] have suggested using the octanol-air partition coeffi-cient (KOA) to describe the partitioning of organic chemicals between air and or-ganic phases in soils, plants and atmospheric aerosols. Recently, highly statisti-cally significant correlations have been found between KOA and measuredgas-particle [55, 56], soil-air [57] and plant-air partition coefficients [58, 59] fora range of persistent organic pollutants.Although it is preferable to measure KOAdirectly, it can be calculated as the ratio KAW/KOW. Measured values are preferredbecause as a result of the partial miscibility of the octanol-water system KOW isnot a true ratio of the solubilities in pure octanol and pure water, but rather the ratio of solubilities in octanol-saturated water and water-saturated octanol[60]. This complication can cause measured KOA values to be up to an order ofmagnitude higher than values estimated from KAW/KOW. However, since KOAmeasurements of phthalate esters have not yet been undertaken, the cal-culation method must be used here to give an indication of their magnitude.Table 3 shows that calculated values of KOA vary between 107 and 1013 and increase with increasing alkyl chain length or molar volume. These high KOAvalues result in a strong tendency of phthalate esters to partition to aerosols,plants and soils.

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 71

3.7Data Gaps

Although there are an abundance of physical-chemical property data for the phthalates, the above analysis has revealed that there are many poor data in theliterature and still some data gaps that it would be advisable to fill. For example,for phthalates with alkyl chains of more than six carbon atoms long, limited pub-lished vapour pressure data are available and new reliable measurements areneeded using state-of-the-art methods. Furthermore, it would be advisable tomeasure KOA values in the laboratory, since calculated estimates may be in errorby an order of magnitude. Finally, we recommend that such properties are mea-sured at a range of temperatures representative of environmental conditions (i.e.–40 to +40 °C) because the environment is rarely at 25 °C. From this informationit will be possible to calculate enthalpies of phase change, which allow calculationof a property at any temperature. Further research is also needed to better quan-tify the influence of particulate and colloidal organic carbon in freshwater andmarine waters on KOC values selected for evaluative environmental fate modelling.

4Degrading Reactions

Biodegradation is the dominant loss process for phthalate esters in all media ex-cept the atmosphere where they are likely to be susceptible to rapid photo-oxi-dation by hydroxyl radicals [4]. Both aerobic and anaerobic microbial breakdownof phthalate esters is initiated by ester hydrolysis to form the monoester and thecorresponding alcohol. The monoester will be further enzymatically degraded tophthalic acid, which will be mineralised following a series of steps. A full reviewof the reaction pathways has been reported by Ejlertsson and Svensson [61].

It is now common practice to simplify environmental reaction kinetics and as-sume that a first-order rate constant (or half-life) can be applied to estimate theloss from each medium. This is necessarily an approximation of the truth and in-volves a judgment that, in a particular type of soil or water, the compound is sub-ject to biodegradation with a half-life of x hours [62]. Furthermore, only thetransformation of the parent compound to the primary metabolite is considered;thus, primary rather than ultimate degradation half-lives are estimated. It is noteworthy that degradation half-lives of organic compounds cannot be viewedin the same way as radionuclide half-lives, which are a fundamental, reproducibleproperty of the radionuclide. Degradation half-lives are functions of both thechemical and the environment. In a series of handbooks by Mackay and co-workers [62], half-lives are assigned on a semi-decade logarithmic scale into one of nine classes (Class 1: 5 h (3–10 h), 2: 17 h (10–30 h), 3: 55 h (30–100 h),4: 170 h (100 – 300 h), 5: 550 h (300 – 1000 h), 6: 1700 h (1000 – 3000 h),7: 5500 h (3000–10,000 h) 8: 17,000 h (10,000–30,000 h) and 9: 55,000 h(30,000–100,000 h)). A chemical can be assigned for example to a half-life class3 with a geometric mean of 55 h and a range of 30–100 h, even though at differ-ent times and places the half-life may fall into class 2 or 4, with perhaps a 5%probability of reaching class 1 or 5.

72 I.T. Cousins, D. Mackay and T.F. Parkerton

Photo-oxidation half-lives, based on hydroxyl radical attack, for phthalate esters were predicted by Staples et al. [4] by using the Atmospheric OxidationProgram (AOP) [63] and these varied between 0.2-2 days for DEHP and9.3–93 days for DMP. There is one experimental study to validate these predic-tions, which reported an atmospheric half-life for DEHP of about 1 day [64]. Forthe Level II and III simulations presented in this paper, half-lives in air were assigned based on the AOP predicted data.

Primary aerobic biodegradation half-lives of phthalate esters in natural watersand soils have been estimated by Staples et al. [4] by an analysis of available mea-sured data. Aerobic biodegradation half-lives in natural waters and soils tend toincrease with increasing alkyl chain length. DMP, DEP, DnBP and DEHP wereshown to have aerobic biodegradation half-lives in natural waters of 0.2–10,0.3–12, 1.7–2.5 and 2–22 days, respectively, and half-lives in soils of 1–40, 1–75,0.4–80 and 25–250 days, respectively. Rates of degradation in aerobic soils wereshown to be 2-5 times slower than in aerobic aquatic environments. This isthought to be due to reduced bioavailability of phthalates resulting from sorptionto soil organic matter. Staples et al. [4] reported that there are only limited dataavailable on biodegradation of phthalate esters in sediments, but the data suggestthat primary biodegradation of phthalate esters in sediments is slower than insoils and of the order of several months (>100 days). BBP is a special case in thatit does not contain two straight alkyl chains in its structure and thus the mech-anism for primary degradation is likely to be different. Measured data for BBPfrom Staples et al. [4] suggest that its aerobic biodegradation half-lives in naturalwaters, soils and sediments are 0.4 day to 8 ¥ 104 days, 9.6 days and 1.6–2.2 days,respectively. The value of 8 ¥ 104 days seems to be erroneous because there arefive other data points in the range 0.4–1.4 days. There is only one data point foraerobic biodegradation in soil for BBP, which is particularly disappointing be-cause soil is the primary medium of accumulation for BBP. Degradation rates ofphthalate esters in anaerobic media are slower, but the models used in this chap-ter only treat aerobic environmental media. Only surface soils (top 5 cm) and sed-iments (top 3 cm) are treated.

The above analysis of measured biodegradation half-life data from Stapleset al. [4] has been used to allocate approximate half-lives by using a semi-decadelogarithmic scale for water, soil and sediment compartments in the EQC Level IIand III simulations (Table 5). This approximate allocation takes account of thelarge uncertainty in measured biodegradation half-lives. We have taken a con-servative approach in our allocation of half-lives and assigned half-lives that arenear to the top of the range reported by Staples et al. [4]. This conservative approach results in estimated half-lives that are higher than those suggested inthe chapter of this handbook focussing on environmental degradation rates ofphthalates. However, it is believed that a conservative approach is appropriate for allocation of half-lives because degradation studies are often conducted at aconstant 25 °C, whereas the environment is often at a lower temperature, somestudies use inoculums and some allow the microbial population to become acclimated. Furthermore, some microcosm studies may not separate losses from degradation from losses due to partitioning to sediments and volatili-sation.

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 73

5Evaluative Fate Modelling with the EQC Model

Conducting evaluative assessments can provide invaluable insights into the char-acteristics of chemical behaviour in the environment. Because the environmentconsidered is purely evaluative or hypothetical, there is no possibility of valida-tion, but the equations used to describe partitioning, transport and transfor-mation are identical to those used successfully in validated models of chemicalfate in more defined environments. The aim is to establish the general featuresof chemical behaviour, namely, into which media the chemical will tend to par-tition, the primary loss mechanisms, the tendency for intermedia transport,the tendency to bioaccumulate, the tendency to undergo long-range transportand environmental persistence. Multimedia models of this type are widely usedby the scientific community as useful tools for providing information on chem-ical fate and have also found acceptance in regulatory practice in a number ofcountries.

The Equilibrium Criterion or EQC model, the model of choice here, has beendescribed fully elsewhere [65]. Briefly, this model in the form of a computer pro-gram, deduces the fate of a chemical in Level I, II and III evaluative environmentsby using principles described by Mackay [36]. The EQC evaluative environmentis an area of 100,000 km2 that is regarded as being representative of a jurisdic-tional region such as the US state of Ohio, or the country of Greece. EQC can sim-ulate the chemical fate of a variety of different chemical class types, classified according to the data requirements to run a model simulation. Phthalate esterspartition to all environmental media and are thus classified as type 1 chemicals

74 I.T. Cousins, D. Mackay and T.F. Parkerton

Table 5. Allocation of half-lives for use in Level II and III EQC simulations

Phthalate ester Assumed reaction half-lives (h)

Air Water Soil Sediment

DMP Class 5 4 5 6Mean 550 170 550 1700Range 300–1000 100–300 300–1000 1000–3000

DEP Class 4 4 5 6Mean 170 170 550 1700Range 100–300 100–300 300–1000 1000–3000

DnBP Class 3 4 6 7Mean 55 170 1700 5500Range 30–100 100–300 1000–3000 3000–10,000

BBP Class 2 3 6 6Mean 17 55 1700 1700Range 10–30 30–100 1000–3000 1000–3000

DEHP Class 2 5 7 7Mean 17 550 5500 5500Range 10–30 300–1000 3000–10,000 3000–10,000

for which all partition coefficients and Z values (fugacity capacities) must be de-fined [36].

5.1EQC Level I Modelling

EQC Level I modelling has been performed for the 22 phthalate esters listed inTable 1 for which physical-chemical properties have previously been estimated(Table 3). Level I EQC model results indicate that under equilibrium, steady stateconditions, with no reaction, the vast majority of phthalates released will residein soil, sediment or water with over 99% being distributed to these three media(Table 6). The low vapour pressures ensure that only small percentages partitionto air. Phthalate esters with alkyl chains containing greater than five carbons par-tition almost exclusively to the organic carbon component of soil and sediment,whereas those with short alkyl chains (<4 carbons), and hence lower hydropho-bicities, partition readily to water.

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 75

Table 6. Summary of results of Level I and II simulations using the EQC model

Phthalate Level I distribution Level II distributionester

Air Water Soil Sediment Persistence Loss by reaction(%) (%) (%) (%) (days) (%)

DMP 0.2 96.3 3.5 0.1 8.4 80DEP 0.4 75.8 23.3 0.5 9.9 81DAP 0.4 46.0 52.4 1.2 – –DPP 0.4 30.4 67.7 1.5 – –DnBP 0.2 5.6 92.2 2.1 28 95DIBP 0.2 5.6 92.2 2.1 – –DnPP 0.1 0.8 96.9 2.2 – –BBP 0.1 2.2 95.6 2.1 27 98DHP <0.1 0.1 97.6 2.2 – –DnHP <0.1 0.1 97.6 2.2 – –DIHpP <0.1 <0.1 97.7 2.2 – –DnOP <0.1 <0.1 97.8 2.2 – –BOP <0.1 0.3 97.5 2.2 – –610P <0.1 <0.1 97.8 2.2 – –DEHP <0.1 <0.1 97.8 2.2 34 100DIOP <0.1 <0.1 97.8 2.2 – –DnNP <0.1 <0.1 97.8 2.2 – –DINP <0.1 <0.1 97.8 2.2 – –DnDP <0.1 <0.1 97.8 2.2 – –DIDP <0.1 <0.1 97.8 2.2 – –D711P <0.1 <0.1 97.8 2.2 – –DUP <0.1 <0.1 97.8 2.2 – –DTDP <0.1 <0.1 97.8 2.2 – –

5.2EQC Level II Modelling

“Persistence in the environment” can be operationally defined as the overall res-idence time at steady state in a multimedia environment [66]. Level II and LevelIII calculations can both give estimates of persistence. Level II calculations givea preliminary estimate of persistence based on equilibrium partitioning,whereas Level III provides a more sophisticated albeit data-intensive estimate ofpersistence that accounts for medium of discharge and intermedia transportrates. A Level II approach is becoming an increasingly popular method becauseof its simplicity and low input data requirements [67]. Level II and III EQC sim-ulations have been performed on five phthalate esters: DMP, DEP, DBP, DBP and DEHP. These representatives possess the best data sets for characteriz-ing compartmental half-lives that are required for these calculations while span-ning a wide range in physical chemical properties characteristic of this chemi-cal class.

Level II EQC modelling predicts an increasing environmental persistence withincreasing alkyl chain length (Table 6). Increasing half-lives and tendency to par-tition to soils and sediments reduce advection and combine to increase the over-all residence time as the alkyl chain length increases.

5.3EQC Level III Modelling

Level III modelling is useful for determining how the medium of release affectsenvironmental fate. Level III fugacity calculations allow non-equilibrium condi-tions to exist between connected media as steady state, and illustrate importanttransport and transformation processes. The tendency of chemicals to migratebetween media can be assessed by modelling emissions to each individualmedium and calculating the amount present at steady state. Table 7 shows theamount of chemical present in each medium of the EQC model environment forindividual emissions of 1000 kg h–1 to the air, water and soil compartments, aswell as a “total” for simultaneous emissions of 1000 kg h–1 to each compartment.Figures 7 and 8 show two example diagrams of the model output for two phtha-lates (DMP and DEHP) with contrasting alkyl chain lengths.

The Level III analysis shows that DEHP tends to partition out of air and wa-ter and into soil and sediments respectively and this is likely to be the pattern forother high-molecular weight phthalates with long alkyl chains. The lower mole-cular weight phthalates tend to remain in the medium of release if emitted to airand water, but if emitted to air, they are deposited and accumulate in soil. Im-portant transport processes for the phthalates that should be the focus of re-search are atmospheric deposition (both wet and dry) and sediment deposition.Important associated processes that should also be studied are gas-particle andsediment-water partitioning.

The Level III model can also be used to derive residence times in the modelworld, which can be thought as “characteristic times” describing the dynamics ofphthalates in the system (Table 7). For example, the model predicts that when to-

76 I.T. Cousins, D. Mackay and T.F. Parkerton

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 77

Tabl

e7.

EQC

Lev

el II

I res

ults

:che

mic

al a

mou

nts

in e

ach

med

ium

bas

ed o

n si

ngle

and

mul

tipl

e em

issi

ons

Phth

alat

e Em

issi

on

Am

ount

at s

tead

y st

ate

(kg)

(per

cent

in b

rack

ets)

Res

iden

ce

Tim

e to

97%

ste

ady

este

rm

ediu

mti

me

(d)

stat

e/cl

eara

nce

(d)a

Air

Wat

erSo

ilSe

dim

ent

DM

PA

ir26

,400

(5.8

)42

,800

(9.4

)38

5,00

0 (8

4.8)

99 (<

0.1)

1957

Wat

er5

(<0.

1)19

7,00

0 (9

9.7)

75 (<

0.1)

455

(0.2

)8.

225

Soil

755

(0.1

)43

,400

(6.6

)61

2,00

0 (9

3.3)

100

(<0.

1)27

81“T

otal

”27

,100

(2.1

)28

3,00

0 (2

1.6)

997,

000

(76.

2)65

4 (0

.1)

1854

DEP

Air

39,3

00 (1

1.3)

15,6

00 (4

.5)

292,

000

(84.

2)93

.8 (<

0.1)

1545

Wat

er19

(<0.

1)19

7,00

0 (9

9.3)

141

(0.1

)11

80 (0

.6)

8.3

25So

il33

4 (<

0.1)

10,0

00 (1

.3)

749,

000

(98.

6)60

(<0.

1)32

96“T

otal

”39

,600

(3.0

)22

2,00

0 (1

7.0)

1,04

0,00

0 (7

9.8)

1,34

0 (0

.1)

1854

DnB

PA

ir38

,700

(13.

8)56

50 (2

.0)

234,

000

(83.

5)19

40 (0

.7)

1236

Wat

er99

(<0.

1)19

5,00

0 (7

4.3)

601

(0.2

)66

,700

(25.

5)11

33So

il27

.8 (<

0.1)

699

(<0.

1)2,

440,

000

(>99

.9)

239

(<0.

1)10

030

0“T

otal

”38

,900

(1.3

)20

1,00

0 (6

.7)

2,68

0,00

0 (8

9.7)

68,9

00 (2

.3)

4213

0BB

PA

ir18

,600

(14.

4)98

4 (0

.7)

109,

000

(84.

6)33

5 (0

.3)

5.4

16W

ater

26.7

(<0.

1)72

,600

(74.

6)51

(<0.

1)24

,700

(25.

4)4.

112

Soil

5 (<

0.1)

108

(<0.

1)2,

450,

000

(>99

.9)

36.8

(<0.

1)10

030

0“T

otal

”18

,600

(0.7

)73

,700

(2.8

)2,

560,

000

(95.

6)25

,100

(0.9

)37

110

DEH

PA

ir13

,200

(0.5

)5,

460

(0.2

)2,

340,

000

(93.

4)14

8,00

0 (5

.9)

100

300

Wat

er20

(<0.

1)16

1,00

0 (3

.9)

3,52

0 (<

0.1)

4,35

0,00

0 (9

6.4)

190

570

Soil

<1

(<0.

1)12

8 (<

0.1)

7,93

0,00

0 (>

99.9

)3,

470

(<0.

1)33

010

00“T

otal

”13

,200

(0.1

)16

6,00

0 (1

.1)

10,3

00,0

00 (6

8.7)

4,50

0,00

0 (3

0.1)

210

630

aT

he r

esid

ence

tim

e m

ulti

plie

d by

3.

tal emissions of DEHP to the environment are 1000 kg h–1 to air (or 8,760,000 kgyear–1), the total inventory in the environment is 2,507,000 kg.As a result the char-acteristic time is 2,507,000/8,760,000 or 0.29 years or about 104 days. A dynamicmodel of DEHP under initial conditions of zero concentrations followed by sus-tained constant emissions would display an approach to a steady state that wouldbe essentially 97% complete after three characteristic times or 0.85 years or312 days. Similarly, if DEHP emissions were stopped entirely the system would be

78 I.T. Cousins, D. Mackay and T.F. Parkerton

Fig.

7.D

iagr

am s

how

ing

EQC

Lev

el II

I out

put f

or D

MP26

349

kg (5

.472

%)

Fug.

= 3

.362

mPa

263

mg/m

3

3.66

e+5

kg (7

6.0%

)Fu

g. =

1.4

63mP

a13

.6ng

/g88

977

kg (1

8.5%

)Fu

g. =

0.0

22mP

a44

5ng

/L

31.6

kg (6

.57

e-3%

)Fu

g. =

2.8

5e·

3mP

a0.

049

ng/g

Tota

l Mas

s =

4.82

e+5

kg

Pers

iste

nce

=16

1h

= 6.

688

days

EQ

C M

od

el v

. 1.0

Leve

l III

Che

mic

al: D

MP

263

kg/h

33.2

kg/

h

0.08

8 kg

/h

79.9

kg/

h

641

kg/h

17.4

kg/

h

1000

kg/

h

1000

kg/

h89

.0 k

g/h

1121

kg/

h

6.33

e · 4

kg/

h

0.39

9 kg

/h

0.05

7 kg

/h0.

456

kg/h

263

kg/h

1000

kg/

h14

93 k

g/h

Air

Soil

Wat

er

Sedi

men

t

Lege

nd:

EMIS

SION

REAC

TION

ADVE

CTIO

N

INTE

RMED

IAEX

CHAN

GE

clear of DEHP almost entirely in 0.85 years. The implication is that, since DEHPhas been in use for many years, it must have reached a steady-state condition inthe environment, thus the use of a steady-state model is justified for regionalmass balance calculations. This argument applies equally for all the phthalates(Table 7). The medium of release greatly affects the estimated environmental res-idence time with emissions to the relatively slowly reacting soil compartment re-sulting in the longest residence times.

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 79

Fig.

8.D

iagr

am s

how

ing

EQC

Lev

el II

I out

put f

or D

EHP13

238

kg (0

.341

%)

Fug.

= 0

.146

mPa

132

mg/m

3

3.18

e+6

kg (8

1.9%

)Fu

g. =

3.3

8mP

a11

8ng

/g1.

54+

5kg

(3.9

80%

)Fu

g. =

0.2

13mP

a77

2ng

/L

5.34

e+5

kg (1

3.8%

)Fu

g. =

0.02

6mP

a83

5ng

/g

Tota

l Mas

s =

3.88

e+6

kg

Pers

iste

nce

=12

93h

= 53

.9da

ys

EQ

C M

od

el v

. 1.0

Leve

l III

Che

mic

al: D

EH

P

132

kg/h

540

kg/h

1.44

6 kg

/h

33.8

kg/

h

296

kg/h

1.25

e · 3

kg/

h

1000

kg/

h

1000

kg/

h15

4 kg

/h

195

kg/h

10.7

kg/

h

673

kg/h

10.7

kg/

h69

4 kg

/h

0.31

9 kg

/h

1000

kg/

h12

95 k

g/h

Air

Soil

Wat

er

Sedi

men

t

Lege

nd:

EMIS

SION

REAC

TION

ADVE

CTIO

N

INTE

RMED

IAEX

CHAN

GE

5.4Estimating Persistence and Long-Range Transport Potential with the TaPL3 Model

The TaPL3 model, which was developed specifically for estimating environmen-tal persistence and long-range transport (LRT) potential [66, 68], is identical inproperties to EQC, but it is set to mimic a closed system, from which no advec-tive losses occur (via air or water flows). By considering only irreversible losses,a conservative estimate of persistence is obtained that may better reflect the re-moval of chemicals from the environment as a whole. Environmental persistenceis evaluated by running various hypothetical releases to the TaPL3 model, usually1000 kg h–1 to air, water and soil, one at a time, and then 1000 kg h–1 to all threemedia simultaneously. These hypothetical scenarios are used because they allowcomparison with other chemicals on an equal basis and because release in-formation is often not available. For phthalates, Parkerton and Konkel [69] haveestimated the amounts of phthalates released to various environmental mediaand have concluded that the vast majority of phthalates are released to the envi-ronment from product end use to the atmosphere. Thus, the emission to air sce-nario may give the most realistic estimate of phthalate persistence in the envi-ronment. The calculated overall environmental residence times, which we defineas “environmental persistence”, for emission to air and water are presented inTable 8 and compared to estimates for well-known persistent organics pollutant(POPs). It is concluded that phthalates are not persistent to the same extent as,for example, PCBs or dioxins.

McLachlan and Horstmann [70] have suggested that uptake of organic com-pounds by vegetation from the atmosphere is important for compounds with octanol-air partition coefficients (KOA) greater than 108. Phthalate esters have KOA values that vary between 107 and 1013; thus, uptake by vegetation is likely tobe important. The environmental persistence may be reduced due to the high re-activity of some organic compounds on vegetation surfaces. Unfortunately, thereis currently a lack of data on metabolism rates of phthalate esters in vegetationmaking it difficult to assess the effect of vegetation uptake on persistence.

80 I.T. Cousins, D. Mackay and T.F. Parkerton

Table 8. Overall environmental persistence and characteristic travel distances for total inputsinto either air or water calculated using TaPL3 [66, 68]

Substance Overall persistence Characteristic Overall persistence Characteristic for emission to air travel distance for emission travel distance (d) in air to water in river water

(km) (d) (km)

DMP 26 520 10 880DEP 24 1000 10 880DBP 19 910 14 870BBP 6.6 330 4.4 280DEHP 120 220 260 700HCB 1,300 110,000 1,600 2,600 a-HCH 150 12,000 230 14,000tetra-PCB 2,500 8,900 2,900 2,900 2,3,7,8-TCDD 860 810 1,800 1,300

There are two main factors which control a chemical’s LRT potential: (1) per-sistence in the atmosphere or water column which can be characterized by a half-life and (2) “stickiness”, the propensity for a chemical to partition to the terres-trial surface if transported in air, or its propensity to partition to sediment iftransported in water. TaPL3 [68] is used here to calculate the characteristic traveldistance of the phthalate esters in air and water (Table 8). The comparison be-tween phthalates and other compounds in Table 8 indicates that phthalates as aclass do not exhibit unfavourable long-range transport characteristics. The lowermolecular weight phthalates are reactive and the heavier phthalates are “sticky”ensuring that they do not travel far in air or water. The fact that phthalates are frequently detected at low concentrations in environmental media is likely to bea function of their high production volume and widespread usage in polymericmaterials rather than a high potential for transport or persistence in the envi-ronment.

6Conclusions

The phthalate esters display a remarkable range of properties extending over6–11 orders of magnitude. Especially important are the marked decreases involatility and solubility in water with increased molar volume or alkyl chainlength. Measuring these properties for the higher molecular weight phthalates isextremely difficult and new methodologies are required for reliable measure-ment. The very low solubilities in water of the octyl and high-molecular weightphthalates will cause bioconcentration and toxicity experiments to be extremelydemanding, especially because the high KOW values will result in strong sorptionto dissolved organic carbon and surfaces. The wide variations in physical-chem-ical properties are reflected in significant differences in environmental parti-tioning, which have been illustrated using an evaluative fate model. The lowermolecular weight phthalate esters are quite volatile, but owing to their very lowKAW values they will volatilise fairly rapidly from the pure state but only veryslowly from aqueous solution. The log KOW values vary from 1.61 to 12.06; thus,the high-molecular weight esters are very hydrophobic and will sorb strongly toorganic matter and surfaces. The high values of KOA suggest that any higher mol-ecular weight esters present in the atmosphere will be appreciably sorbed toaerosol particles and to soil and vegetation. Air-water partition coefficients in-crease with increasing molecular weight; thus, the higher molecular weight phthalate esters will potentially evaporate more rapidly from water, but this will be mitigated by sorption to suspended matter in the water column. The phthalate esters also show significant and systematic differences in reactivity orhalf-life, with the primary biodegradation half-life increasing with increasingalkyl chain length and the photo-oxidation half-life showing the opposite trend.These changes in environmental persistence of phthalates have been illustratedby using the TaPL3 model. Based on these calculations it is concluded that phthalates as a class are not as persistent as other well-known chlorinated organicpollutants. Although the environmental persistence of higher molecular weightphthalates tends to be greater, KOA and thus propensity to partition to aerosols,

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 81

vegetation and soils also increases. This increased “stickiness” tends to limit therelative ability of the heavier phthalates to be transported in air or water as hasbeen demonstrated with the TaPL3 model. The evaluative modelling exercise undertaken here is valuable for assisting the evaluation of experimental resultson the fate of phthalates. Ultimately, these models require verification from well-designed monitoring programs, which seek to determine the actual fate ofphthalates in our complex, variable and real environment. Evaluation of the per-formance of mass balance models in predicting environmental partitioning andfate is the focus of a later chapter.

Acknowledgement. We are grateful to the Phthalates Esters Panel of the American ChemistryCouncil (ACC) for funding this research, and to NSERC and the consortium of chemical com-panies that provide finances to support the Canadian Environmental Modelling Centre.

7References

1. ECETOC (1988) Technical Report No 292. Cousins IT, Mackay D (2000) Chemosphere 41:13893. Reid RC, Prausnitz JM, Polling BE (1987) The properties of gases and liquids, 4th edn.

McGraw-Hill, New York4. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:6675. Nielsen NM, Bundgaard H (1989) J Med Chem 32:7276. Howard PH, Banerjee S, Robillard KH (1985) Environ Toxicol Chem 4:6537. Veith GD, Macek KJ, Petrocelli SR, Carroll J (1980) An evaluation of using partition co-

efficients and water solubility to estimate bioconcentration factors for organic chemicals.In: Eaton JG, Parrish PR, Hendricks AC (eds) Fish, aquatic toxicology ASTM STP 707.Amer-ican Society for testing and Materials, pp 116–129

8. Leyder F, Boulanger P (1983) Bull Environ Contam Toxicol 30:1529. Wolfe NL, Steen WC, Burns LA (1980) Chemosphere 9 :403

10. Renberg LO, Sundström SG, Rosen-Olofsson A-C (1985) Toxicol Environ Chem 10:33311. McDuffie B (1981) Chemosphere 10:7312. Ellington JJ, Floyd TL (1996) Octanol/water partition coefficients for eight phthalate esters.

United States Environmental Protection Agency: Environmental Research Brief. Report noEPA/600/S-96/006

13. Russell DJ, McDuffie B (1986) Chemosphere 15:100314. Grayson BT, Forsbraey LA (1982) Pest Sci 13:26915. De Kock AC, Lord DA (1987) Chemosphere 16:13316. DeFoe DL, Holcombe GW, Hammermeister DE (1990) Environ Toxicol Chem 9:62317. Dobbs AJ (1984) Chemosphere 13:68718. Werner AC (1952) Ind Eng Chem 44:273619. Harnisch M, Möckel HJ, Schulze G (1983) J Chromat 282:31520. Gückel W, Kästel R, Lewerenz J, Synnatschke G (1982) Pest Sci 13:16121. OECD guidelines for testing of chemicals (1981) “Partition coefficient (n-octanol-water)

flask shaking method”, Section 107. Organisation for Economic Cooperation and Devel-opment, Paris, France

22. Frissel WJ (1955) Ind Eng Chem 18:109623. Quackenbos HM Jr (1954) Ind Eng Chem 46:133524. Haky JE, Leja B (1986) Anal Lett 19:12325. Hansch C, Leo A (1987) LogP database. Pomona College, Claremont, CA26. Hollifield HC (1979) Bull Environ Contam Toxicol 23:57927. Gledhill WE, Kaley RG, Adams WJ, Hicks O, Michael PR, Saeger VW (1980) Environ Sci

Technol 14:301

82 I.T. Cousins, D. Mackay and T.F. Parkerton

28. Letinksi DJ, Connelly MJ, Parkerton TF (1999) Slow-stir water solubility measurements forphthalate ester plasticisers. Paper presented at SETAC Europe, 9th Annual Meeting, Leipzig,Germany, May 25–29, p 159

29. Ellington JJ (2000) J Chem Eng Data 44:1414–141830. De Bruijn J, Busser F, Seinen W, Hermens J (1989) EnvironToxicol Chem 8:49931. Dobbs AJ, Cull MR (1982) Environ Poll (Series B) 28932. Brooke D, Nielsen I, De Bruijn J, Hermens H (1990) Chemosphere 21:11933. OECD guidelines for testing of chemicals (1981) “Vapour pressure curve (dynamic method

– static method – isoteniscope – vapour pressure balance – gas saturation method)”,Section 104. Organisation for Economic Cooperation and Development, Paris, France

34. Klein W, Kördel W, Weiss M, Poremski HJ (1988) Chemosphere 17:36135. Exxon Biomedical Sciences Inc. (1996) Water solubility study no 199638. East Millstone, NJ,

USA36. Mackay D (2001) Multimedia environmental models: the fugacity approach, 2nd edn.

Lewis, Boca Raton, London37. Kollig HP (1988) Toxicol Environ Chem 17:28738. Thomsen M, Rasmussen AJ, Carlsen L (1999) Chemosphere 38:261339. Cole JG, Mackay D (2000) Environ Toxicol Chem 19:26540. Thomsen M, Carlsen L, Hvidt S (2001) Environ Toxicol Chem 20:12741. Turner A, Rawling MC (2000) Marine Chem 68:20342. Howard PH (2000) Biodegradation. In: Boethling S, Mackay D (eds) Handbook of pro-

perty estimation methods for chemicals: environmental and health sciences. Lewis,Boca Raton

43. Schwarzenbach RP, Gschwend PM, Imboden DM (1993) Environmental organic chemistry.Wiley and Sons, New York

44. Wania F, Shiu WY, Mackay D (1994) J Chem Eng Data 39:57245. Gobas FAPC, Morrison HA (2000) Bioconcentration and biomagnification in the environ-

ment. In: Boethling S, Mackay D (eds) Handbook of property estimation methods forchemicals: environmental and health sciences. Lewis, Boca Raton

46. Seth R, Mackay D, Muncke J (1999) Environ Sci Technol 33:239047. Doucette WJ (2000) Soil and sediment sorption coefficients. In: Boethling S, Mackay D (eds)

Handbook of property estimation methods for chemicals: environmental and health sciences. Lewis, Boca Raton

48. McLachlan MS (2000) Vegetation-air partition coefficients. In: Boethling S, Mackay D (eds)Handbook of property estimation methods for chemicals: environmental and health sciences. Lewis, Boca Raton

49. Leo A (2000) Octanol-water partition coefficients. In: Boethling S, Mackay D (eds) Hand-book of property estimation methods for chemicals: environmental and health sciences.Lewis, Boca Raton

50. Zhou JL, Liu YP (2000) Marine Chem 71:16551. Williams MD, Adams WJ, Parkerton TF, Biddinger GR, Robillard KA (1995) Environ Toxi-

col Chem 14:147752. Long JLA, House WA, Parker A, Rae JE (1998) Sci Tot Environ 210/211:22953. Cornelissen G, van Noor PCM, Govers HAJ (1997) Environ Toxicol Chem 16:135154. Harner T Mackay D (1995) Environ Sci Technol 29:159955. Finizio A, Mackay D, Bidleman T, Harner T (1997) Atmos Environ 31:228956. Harner T, Bidleman T (1998) Environ Sci Technol 32:149457. Hippelein M, McLachlan MS (1998) Environ Sci Technol 32:31058. Kömp P, McLachlan, MS (1997) Environ Sci Technol 31:294459. Hiatt MH (1999) Environ Sci Technol 33:412660. Chiou CT, Schmedding DW, Manes M (1982) Environ Sci Technol 16:461. Ejlertsson J, Svensson B (1995) A review of the possible degradation of polyvinyl chloride

(PVC) plastics and its component phthalic acids and vinyl chloride under anaerobic conditions prevailing in landfills. Dept of Water and Environmental Studies, Linkoping University, Sweden

Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters 83

62. Mackay D, Shiu W-Y, Ma K-C (1999) Physical-chemical properties and environmental fatehandbook, CRC netBASE CD-ROM. Chapman and Hall/CRC Press, Boca Raton, FL, USA

63. Meylan WM, Howard PH (1993) Chemosphere 26:229364. Behnke W, Nolting F, Zetzsch C (1987) An aerosol smog chamber for testing abiotic degra-

dation. In: Greenhalgh R, Roberts TR (eds) Pesticide science biotechnology proceedings ofsixth international congress on pesticide chemistry. Blackwell, Oxford, UK, pp 401–404

65. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Environ Toxicol Chem 15:162766. Webster E, Mackay D, Wania F (1998) Environ Toxicol Chem 17:214867. Gouin T, Mackay D, Webster E, Wania F (2000) Environ Sci Technol 34:88168. Beyer A, Mackay D, Matthies M, Wania F, Webster E (2000) Environ Sci Technol 34:69969. Parkerton TF, Konkel WJ (2000) Evaluation of the production, consumption, end use and

potential emissions of phthalate esters. Prepared for the American Chemistry Council(Draft of November 2000) by Exxon Mobil Biomedical Sciences Inc (EMBSI), Annandale,NJ, USA

70. McLachlan MS Horstmann M (1998) Environ Sci Technol 32:413

84 I.T. Cousins, D. Mackay and T.F. Parkerton

© Springer-Verlag Berlin Heidelberg 2003

Degradation of Phthalate Esters in the Environment

Dennis R. Peterson 1 · Charles A. Staples 2

1 ExxonMobil Biomedical Sciences, Inc., 1545 Route 22 East, P.O. Box 971, Annandale,New Jersey 08801–0971, USA. E-mail: dennis.r.peterson@exxonmobil.com

2 Assessment Technologies, Inc., 10201 Lee Highway, Suite 580, Fairfax, VA 22030, USA

This chapter reviews the degradation data that are appropriate for modeling the environmen-tal concentration of phthalate diesters. The necessary data are the first-order rates of primarydegradation in air, water, and soil. The predominant fate of PDEs is biodegradation in waste-water, aerobic aquatic (water/sediment) environments, and soil. Anaerobic degradation mayalso play an important role for some phthalates under certain conditions. Of the abioticprocesses, the rates of hydrolysis and direct photolysis are too low to have a significant influ-ence on overall phthalate fate. On the other hand, indirect photolysis in air due to hydroxyl rad-ical attack in both vapor phase and particle-sorbed phthalates may have a significant role inthe overall environmental degradation of some phthalates. The literature data for these envi-ronmental degradation rates are evaluated with regard to their relevance to actual environ-mental conditions and recommendations are presented for the most likely degradation ratesfor specific phthalate diesters. The rates of further degradation of phthalate diester degrada-tion products are also briefly discussed.

Keywords. Phthalate ester, Hydrolysis, Photolysis, Biodegradation, First-order rate

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

2 Selection of Data for the Evaluation of Environmental Persistence 87

2.1 Realism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 872.2 Primary Versus Ultimate Degradation . . . . . . . . . . . . . . . . 87

3 Abiotic Degradation . . . . . . . . . . . . . . . . . . . . . . . . . 88

3.1 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883.2 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

4 Biodegradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

4.1 Surface Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 904.1.1 Reported Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 904.1.2 Bioavailability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 954.1.3 Biodegradation of Very Low Concentrations . . . . . . . . . . . . 974.2 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984.2.1 Reported Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 85–124DOI 10.1007/b11464

4.2.2 Conditions Affecting Soil Biodegradation . . . . . . . . . . . . . . 1004.3 Wastewater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1054.4 Anaerobic Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1104.5 Solid-Phase Biodegradation . . . . . . . . . . . . . . . . . . . . . 113

5 Degradation of Phthalate Metabolites . . . . . . . . . . . . . . . . 114

6 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . 118

7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

Abbreviations

BBP Butylbenzyl phthalateDBP Dibutyl phthalateDAP Diamyl phthalateDDP Didecyl phthalateDEHP Bis(2-ethylhexyl) phthalateDEP Diethyl phthalateDHP Dihexyl phthalateDIBP Diisobutyl phthalateDIDP Diisodecyl phthalateDIHP Diisoheptyl phthalateDINP Diisonyl phthalateDIUP Diisoundecyl phthalateDMP Dimethyl phthalateDnOP Di-n-octyl phthalateDNP Dinonyl phthalateDOP Dioctyl phthalateDPP Di-n-pentyl phthalateDPrP Dipropyl phthalate

1Introduction

Phthalate diesters (PDEs) are commonly released into the environment and arerapidly degraded. The main degradation route of PDEs is thought to be biodegra-dation. The biodegradability of PDEs has been reviewed extensively in the past.The main subject of this review will be data on the rate of degradation of PDEsin the environment, with particular emphasis on biodegradation. The rate ofdegradation of chemicals in the environment is important for modeling theiroverall environmental fate and, in particular, their overall degree of persistenceand transport. The modeling of the environmental fate and concentration ofPDEs is covered elsewhere in this volume.

Recent reports on the results of standard biodegradation tests will only bementioned briefly and more emphasis will be placed upon evaluation of the rateof primary biodegradation measured under environmentally realistic situations.Furthermore, the physical properties of PDEs generally result in their association

86 D.R. Peterson and C.A. Staples

with particulate matter in air and water and also to their strong partitioning tosoil and sediment. Thus, particular attention will be paid to the studies of degra-dation rates in soil and sediment.

2Selection of Data for the Evaluation of Environmental Persistence

2.1Realism

The rate data need to be as relevant as possible to the natural environment, sincethese data are used to evaluate environmental persistence and provide data forenvironmental fate models. Data from simulation studies, microcosms, and fieldstudies are the most applicable. Because the rate of biodegradation of a chemi-cal is dependent both upon its concentration and the concentration of degrad-ing microorganisms, the most useful results are those that are done at environ-mentally realistic concentrations of the chemical and with microorganismdensities that are typical of a particular environmental compartment.

Environmental conditions such as temperature, nutrient availability, oxygenavailability, and substrate bioavailability all affect the rate of biodegradation. Theinfluence of experimental conditions will be evaluated when reviewing rate data.

2.2Primary Versus Ultimate Degradation

The degradation data that are of use for modeling and for evaluation of envi-ronmental persistence are the pseudo-first-order disappearance rates of thePDEs. Environmental fate models utilize physical-chemical properties such aswater solubility and vapor pressure to calculate the partitioning behavior of achemical in the environment. These physical chemical properties differ greatlybetween the “parent” PDE and its degradation products. The overall decrease inconcentration of the parent chemical in the environment is evaluated from itsdegradation rates in the various environmental compartments. In the terminol-ogy of biodegradation, it is the rate of primary degradation that is needed. Forabiotic degradation, such as hydrolysis and photolysis, the standard tests alsomeasure the rate of decrease in parent chemical concentration, that is, primarydegradation.

In addition to the use of primary degradation rate data in environmental fatemodeling, evaluation of the property of persistence within the context of “per-sistent organic pollutants” also implies primary degradation, since a chemicalstructure that is degraded cannot also persist. This does not mean that the degra-dation products are of no consequence. In assessing the overall risk of a chemi-cal to the environment, the hazard and fate of the degradation products are important considerations. But these “daughter” chemicals will have differentphysical-chemical properties than the parent substance. In the case of PDEs,the biodegradation products are known and the further degradation of these willbe discussed.

Degradation of Phthalate Esters in the Environment 87

3Abiotic Degradation

3.1Air

Direct photolysis probably does not play a major role in the atmospheric de-gradation of PDEs [1]. However, indirect photolysis by hydroxyl radical oxida-tion may contribute appreciably to overall fate. Staples et al. [1] calculated a rangeof air oxidation half-lives for PDEs based on the air oxidation program (AOP) of Atkinson [2] and a range of hydroxyl radical concentrations from 3 ¥ 105 to3 ¥ 106 molecule cm–3. The atmospheric half-lives of a number of PDEs have been recalculated with an updated version of AOP (AOPWIN 1.89) as imple-mented in EPIWIN [3]. In this half-life calculation we also used a hydroxyl radical concentration of about 1 ¥ 106 molecule cm–3, a global average tropo-spheric concentration adjusted for diurnal and seasonal variation [4]. Althoughhydroxyl radical concentrations vary considerably depending upon the degree of atmospheric pollution and intensity of sunlight, Atkinson has recommendedthe use of this average value for global photooxidation estimates [4]. EPIWINuses a 12-hour daylight period and a hydroxyl radical concentration of 1.5 ¥ 106.The rate constants (kOH) for hydroxyl radical attack and the calculated half-livesare shown in Table 1. Only DMP and DEP have appreciable (> 2 d) atmospherichalf-lives. The half-lives shown in Table 1 for indirect photolysis in air are at thelow end of those given by Staples et al. [1]. These values are still conservative,since emissions of PDEs are expected to be mainly in urbanized areas where hy-droxyl radical concentrations are 3–10 times higher than the value used in thecalculation.

The less volatile PDEs, such as BBP and DOP, are reportedly about equally distributed in air between gas and particulate phases [5]. So the question arises

88 D.R. Peterson and C.A. Staples

Table 1. Atmospheric photodegradation rates and half-lives of selected PDEs

Phthalate CAS No. k T1/2 a T1/2 (¥10–12 cm3 molecule–1 s–1) (d) (h)

DMP 131-11-3 0.574 14.41 346DEP 84-66-2 3.466 2.39 57.3DBP 84-74-2 9.277 0.89 21.4DIBP 84-69-5 9.260 0.89 21.4BBP 85-68-7 11.049 0.75 18.0DHP 84-75-3 14.929 0.55 13.3DIHP 18.719 0.44 10.6DEHP 117-81-7 21.955 0.38 9.0DnOP 117-84-0 20.581 0.40 9.6DINP 68515-48-0 23.408 0.35 8.5DIDP 68515-49-1 26.217 0.32 7.6DIUP 3648-20-2 31.847 0.26 6.2

a Based on a global, seasonal, and diurnal average hydroxyl radical concentration of 1 ¥106 molecule cm–3 [4].

as to the photolysis half-lives of particulate-sorbed PDEs. Behnke et al. [6] have investigated the photolysis rate for DEHP adsorbed to various particulateaerosols. They report a first-order rate constant of 1.4 ¥ 10–11 cm3 molecule–1 s–1

for the reaction of DEHP with hydroxyl radicals when adsorbed as a mono-layer on Fe2O3 or SiO2 aerosols. Photolysis on TiO2 was at a much higher rate due to catalytic activity of the carrier. This rate for inert particle sorbed photo-lysis corresponds to a half-life of 0.6 d, using the global average hydroxyl radicalconcentration of 9.7 ¥ 105 molecule cm–3. This half-life is not much longer thanwith the calculated gas-phase value of 0.38 d. Thus, sorption onto aerosol particles seems not to have a large effect on the overall rate of indirect photo-lysis of PDEs.

3.2Water

Jin et al. [7] reported first-order aqueous photolysis rates for DBP and DEHP of0.23 and 0.9 h–1, corresponding to half-lives of 3 and 0.75 h, respectively. ThePDEs were present in a surface micro-layer in mg L–1 quantities. The reaction wasdependent upon light intensity and oxygen. The rate was stimulated by the pres-ence of TiO2 and H2O2 and the optimum pH was 6.0. The results under artificiallight (72,000 lux) were essentially the same as with natural sunlight (83,000 lux).The photodegradation rates were higher in natural water than in simulated systems.

The importance of these findings is difficult to assess but they certainly war-rant further investigation. Due to their low water solubilities, surface films ofPDEs are potentially quite important, as observed by Södergren [8] in laboratoryecosystems.Various estimates of the aquatic photolysis half-lives of PDEs are inthe range of months to a year or more [1, 9]. Gledhill [10] reported a measureddegradation of BBP exposed to sunlight of only 5% in 28 days.A possible reasonfor the apparent discrepancy between these values and the measurement by Jinet al. [7] is that the latter value, measured in natural water, is the consequence ofindirect photolysis. Indirect photolysis in natural water may proceed through either hydroxyl radical or the photoactivation of organic matter or nitrate.Moreover, indirect aqueous photolysis half-lives of most classes of chemicals areexpected to be of the order of days to weeks [11].

The hydrolysis rate of a number of PDEs has been reported by Wolfe et al. [12]and, with the exception of BBP, half-lives of many years are expected [1]. Thesefinding are in general agreement with what is known regarding ester hydrolysis[13]. Thus, abiotic hydrolysis is not expected to play a significant role in deter-mining the environmental fate of PDEs.

Degradation of Phthalate Esters in the Environment 89

4Biodegradation

4.1Surface Water

4.1.1Reported Rates

There are a large number of reports that PDEs undergo rapid ultimate biodegra-dation (mineralization) in laboratory biodegradation tests and these have beenreviewed by Staples et al. [1]. Since that review, Scholz et al. [14] published a report showing that DBP, DEHP, and DINP met all of the criteria for readybiodegradability, including the 10-day window, in tests strictly adhering to reg-ulatory guidelines and criteria for ready biodegradability. Meeting the criteria forready biodegradability implies that the chemical will rapidly degrade in the en-vironment, but such tests are essentially pass/fail tests and do not provide quan-titative data on the rate of environmental biodegradation [15, 16].

There are a number of reasons why these laboratory test systems are not pre-dictive of environmental degradation rates:

– The tests are designed to measure the extent of biodegradation and not therate.

– The tests measure the ultimate degradation (mineralization) of the test chem-ical and not its primary degradation rate.

– The tests measure the population dynamics of batch culture of the degradingorganisms, covering the adaptation, growth, and senescence phases, and notthe steady-state kinetics attained in the environment.

– The chemical being tested is the only organic chemical added and forms thesole carbon source for growth and metabolism of the microorganisms, whilein the environment many carbon sources are available for growth.

– The microorganisms in the test system are standardized as to type (aerobicsewage derived) and number and not intended to simulate typical surface wa-ter populations in type or number.

– Adapted microbial cultures are not allowed in the ready test, but in the envi-ronment continuous emissions of a chemical would lead to adaptation.

– The concentration of test chemical employed is generally much higher thanwould occur in the environment.

The rate of biodegradation of a chemical is, in general, not a first-order reaction.However, under certain conditions it may approach first order (“pseudo-first-or-der”) and these conditions are often approximated in the environment. Mathe-matically, the degradation rate of substrate in a batch system may be describedby the integrated Monod equation [15]:

– ds/dt = mmax S(S0 + X0 – S)/(Ks + S) (1)

Where mmax is the maximum specific growth rate of the degrading organisms, Sis the substrate concentration at time t, S0 is the initial substrate concentration,

90 D.R. Peterson and C.A. Staples

X0 is the initial bacterial density divided by the growth yield, and Ks is the satu-ration constant, substrate concentration giving growth rate of mmax/2.

Biodegradation rates depend upon both the concentration of the chemical andthe concentration of the degrading organism and would be expected to be sec-ond order. However, when the number of organisms is rate limiting, their con-centration is rapidly changing when they grow and multiply at the expense ofsubstrate (test chemical). At low or intermediate chemical concentrations, whenthe test chemical concentration is limiting, as its concentration declines, uptakeand enzyme binding of the chemical result in complex kinetics (Monod kinetics).At high test chemical concentrations, when its concentration is not rate limitingbecause the uptake and degradation mechanisms of the microorganisms are sat-urated, degradation rates will be zero order or logarithmic (depending only onmicroorganism number).

Thus, if the number of degrading microorganisms is high in relation to testchemical concentration and the enzyme saturation constant is high in relation tothe substrate concentration, then pseudo-first-order rates may be observed. Thatis, the degradation rate (k1) will be observed to be solely dependent on substrateconcentration and a plot of the logarithm of substrate concentration versus timeyields a straight line with a slope equal to k1 (and a half-life, t1/2, equal to ln 2/k1).Under these conditions, the rate may be expressed as:

– ds/dt = k1 S (2)

Where k1 is equal to µmax X0/Ks.These conditions may be achieved in systems in which substrate concentra-

tions are of the order of tens to hundreds of micrograms per liter and the popu-lation of degrading bacteria is about 108 cells L–1 or higher [15].

A secondary problem with regard to measured biodegradation rates of PDEsrelates their hydrophobic nature. First-order reactions relate to the concentrationof the chemical in solution, but PDEs may exist predominantly bound to partic-ulates or dissolved organic carbon (DOC) and not truly in solution. Or they maybe simply tested at concentrations above their water solubility. All of these situ-ations are treated below under the topic of “bioavailability”.

Considering the foregoing, it would seem that the best way to establish the rateof biodegradation in the environment is to rely upon measured rates of degra-dation of the chemical within the environment or within simulations of the en-vironment. Aronson and Howard [17] have also made this recommendation.However, for chemicals in general there are few rate data measured under suchcircumstances and few standardized methods for collecting such data. There aremuch more available data from ready biodegradation tests or other screeningtests. As a consequence of the lack of measured rate data, some regulatory authorities have established criteria for extrapolation from screening test resultsto environmental degradation rates. Within the EU risk assessment context, achemical that passes the ready biodegradability test is assigned an aerobicaquatic biodegradation rate of 0.047 d–1 and one passing an inherent test is as-signed a rate of 0.0047 d–1, corresponding to half-lives of 15 and 150 days. In thecontext of persistence for PBT chemicals, Boethling [18] has suggested rates of0.14 d–1 and 0.0069 d–1 for chemicals that are readily and inherently biodegraded,

Degradation of Phthalate Esters in the Environment 91

respectively, corresponding to half-lives of 5 and 100 days. He also suggests in-termediate rates of 0.069 d–1 and 0.023 d–1 for chemicals that give intermediate re-sults in the ready test.As mentioned earlier, PDEs are readily biodegradable andthus, by EU or Boethling criteria would have assumed half-lives in the aquatic en-vironment of 15 or 5 days.

There are a number of published data for PDE pseudo-first-order biodegra-dation rates in aquatic microcosms. Most of these systems contain an added sed-iment layer or, as a result of containing field-collected water, contain suspendedsediment and microorganisms. Since all of the PDEs have high adsorption coef-ficients for sediments and suspended solids [1], it is expected that a major frac-tion of PDEs in aquatic systems will be sediment associated and therefore,aquatic degradation studies for water and aerobic sediment are considered to-gether.

Saeger and Tucker [19] reported on the primary degradation of BBP, DEHP,DUP, and an isomeric mixture of heptyl, nonyl, and undecyl alcohol PDEs (“San-itizer 711” or di-711) in Mississippi River water. The river water was settled andthe incubations conducted unstirred in the dark at PDE concentrations of 1 mgL–1. The half-lives may be estimated from their graphic presentation of the chem-ical recovery data as about 1.5 d for BBP, 3.3 weeks for DUP, 4.3 weeks for DEHP,and 5.1 weeks for di-711. There was no apparent lag phase. These half lives cor-respond to rates of 0.46, 0.030, 0.023, and 0.019 d–1, respectively (see Table 2). Asimilar study was reported by Carson et al. [20] for BBP and di-711. They compared the results of a Mississippi River water biodegradation study design(river die-away) with two types of microcosms: (1) A lake/sediment system using a slow-stirred sediment-water system from a spring-fed freshwater lake(Lake 34, Bush wildlife area, St. Charles Co. MO), and (2) an unstirred river water/sediment (“ecocore” from Illinois River) design. PDE concentrations were at 10 and 100 µg L–1. They reported half-lives of 5 d and <3 d for BBP in the two systems compared with 2 d in the river die-away system. The di-711 hadhalf-lives of 6–8 d in the river die-away design and 1–5 d in the lake water/sediment microcosm. It is likely that the differences are due to differences in thenumber of degrading organisms in water/sediment from the two sample locations. Adams and Saeger [21] also reported a half-life of 1.4 d for BBP in alake microcosm.

Regional differences in the ability of water and sediment microorganisms todegrade DBP were investigated by Walker et al. [22]. They collected water andsediment samples from a number of rivers and estuaries in Florida, Mississippi,and Louisiana. Incubations were carried out in shake flasks (140 rpm, 25 °C) withDBP concentration of 0.5 mg L–1. Of the eight sites tested, seven of the water onlysystems showed overall (including lag phase) first-order rate constants of0.07–0.98 d–1. The eighth system degraded too quickly to determine the rate. Themean of these seven rate constants was 0.472 d–1 and the median 0.512 d–1. For theeight sediment samples, rates ranged from 0.092 to 1.09 d–1 with a mean of0.38 d–1 and a median of 0.29 d–1. There was no apparent relationship betweenrates in water and sediment from the same site or with the salinity at the sites. Itis also evident that there is little difference in degradation rates between systemswith and without suspended sediments.

92 D.R. Peterson and C.A. Staples

Furtmann [23, 24] measured the rate of primary biodegradation of a series of15 different PDEs in Rhine River water. Concentrations were in the low mg L–1

range for the lower PDEs and less than 1 µg L–1 for the less soluble PDEs. Incu-bations were performed in shake flasks at 25 °C. The first-order rate constants determined rates that ranged from a high of 2.2 d–1 for BBP to a low of 0.2 d–1

for DDP. There were no long lag phases as all of the PDEs tested degraded by 50% in <3 days. Ye and Tian [25] have reported the rates of primary biodegra-dation of DMP, DEP, DBP, and DAP. However, they used an inoculum derived fromsoil and sewage supernatant similar to that used in the MITI test. Half-lives of0.23, 0.16, 0.22, and 0.22 d were reported for these four PDEs, corresponding torates of 3.0, 4.3, 3.2, and 3.2 d–1.

Degradation of Phthalate Esters in the Environment 93

Table 2. Summary of first-order PDE biodegradation rates in aerobic aquatic environments

Phthalate First-order rate (d–1) Half-life (d) Test conditions Ref.

DMP 0.5 1.39 RDAa, shaken, 25 °C ª 1 µg L–1 [23] 0.23 3.0 MITI inoculum [25]

DEP 1.8 0.39 RDA, shaken, 25 °C ª 1 µg L–1 [23] 0.16 4.33 MITI inoculum [24]0.98 b 0.71 Microcosm periphiton [27]

DPrP 1.3 0.53 RDA, shaken, 25 °C ª 1 µg L–1 [23]DBP 0.8 0.87 RDA, shaken, 25 °C ª 1 µg L–1 [23]

0.51 c 1.36 RDE, estuarine & river [22]0.29 c 2.40 Sediment microcosm [22]0.22 3.15 MITI inoculum [25]0.14 d 4.95 RDA, low sediment [29]0.12 d 5.78 RDA, water only [29]

DIBP 0.8 0.87 RDA, shaken, 25 °C ª 1 µg L–1 [23]BBP 0.46 b 1.5 RDA, unstirred, 1 mg L–1 [19]

0.50 b 1.4 Microcosm, lake [21]0.35 b 2 RDA [20]0.14 b 5 Microcosm, un-impacted [20]

>0.23 b <3 Microcosm, Illinois river [20]2.2 0.32 RDA, shaken, 25 °C ª 1 µg L–1 [23]

DPP 1.3 0.53 RDA, shaken, 25 °C ª 1 µg L–1 [23]DOP 0.7 1.0 RDA, shaken, 25 °C ª 1 µg L–1 [23]DEHP 0.023 b 30 RDA, unstirred, 1 mg L–1 [19]

0.2 3.5 RDA, shaken, 25 °C ª 1 µg L–1 [23]1.73 0.4 Field data, estuarine sediment [26]

Di-711 0.019 36 RDA, unstirred, 1 mg L–1 [19]0.13–0.09 b 6–8 RDA [20]0.69–0.14 b 1–5 Microcosm, lake [20]

DDP 0.5 1.4 RDA, shaken, 25 °C ª 1 µg L–1 [23]DIUP 0.030 b 23 R DA, unstirred, 1 mg L–1 [19]

a RDA “River Die-Away” study, flask containing un-inoculated river water.b Value calculated from data presented in the referenced paper.c Median result from 7 to 8 samples from different locations.d Calculated from second-order rates.

A different approach to the estimation of degradation rates has been made byFauser et al. [26]. By using the concentration versus depth profile of DEHP in sed-iments in Roskilde Fjord, Denmark, they derived a model which incorporatedrates of sedimentation and of aerobic and anaerobic biodegradation. Through fit-ting the model to observed concentrations, they arrived at a first-order aerobicdegradation rate for DEHP in sediment of 2 ¥ 10–6 s–1 or 1.73 d–1, equivalent to ahalf-life of 0.4 d.

The rates of chemical degradation in periphyton (aufwuchs) from microcosmswere reported by Lewis et al. [27]. They correlated the rate of degradation withthe presence of total bacteria and degrading bacteria. In the flow-through mi-crocosms, DEP showed second-order rate constants of 0.03–2.2 nL cell–1 h–1. Themean of all of the experiments was 0.41 nL cell–1 h–1 (= 0.41 ¥ 10–9 L cell–1 h–1).These experiments contained 107–109 total bacteria based on total plate count;thus, by using an average of 108 bacterial cells L–1, the average pseudo-first-orderrate is 0.041 h–1 or 0.98 d–1 (t1/2 = 0.7 d). The typical number of bacteria in surfacewater is 108–1010 cells L–1 [28], so that a microorganism number of 108 is also con-veniently close to the average for surface water.

Second-order rate constants for DBP primary biodegradation in MississippiRiver water, with and without added sediments, were determined by Steen et al. [29]. Initial DBP concentrations were 1–2 mg L–1. Plate counts of total organisms were used to convert the observed pseudo-first-order rate constantsto second-order rate constants. These ranged from 3.1 ¥ 10–11 L cell–1 h–1 in water alone to 6.1 ¥ 10–13 L cell–1 h–1 in the high-sediment content flasks. Unfor-tunately, the cell numbers are not given so the corresponding pseudo-first-orderrates cannot be directly calculated. However, a logarithmic graph of the DBP con-centrations versus time is presented and the slopes may be estimated to deter-mine the pseudo-first-order rates. These first-order rates are calculated to beabout 0.12, 0.14, and 0.07 d–1 for water only, low-sediment, and high-sediment systems, respectively. These rates correspond to half-lives of 5.8, 4.9, and 10.1 d.The larger differences between the reported second-order rates for water and forthe high-sediment system than between these first-order rates we calculate forthese same systems is evidently due to higher cell numbers in the sedimentamended systems.

A summary of these reported PDE biodegradation rates in surface water ispresented in Table 2. The pseudo-first-order rates of primary biodegradation un-der environmentally realistic conditions, such as model ecosystems or incubationin river water, are in the range of 0.2–2.0 d–1 for most PDEs studied. The studyof Saeger and Tucker [19] gave somewhat lower rates for the higher phthalates(DEHP, D711, and DIUP) but these studies were conducted at concentrations wellabove the solubilities of these phthalates and the systems were settled and not agitated at all. Studies conducted at more realistic environmental concentrationsof PDEs, such as those of Furtman [23], show rates in the range of 0.2–2.0 d–1 forthe higher molecular weight PDEs as well.

94 D.R. Peterson and C.A. Staples

4.1.2Bioavailability

Biodegradation cannot occur if for some reason the chemical is unavailable to themicroorganisms or if the concentration truly in solution is less than the total con-centration, the biodegradation rate may be reduced. Factors that reduce a chem-ical’s effective concentration with regard to biodegradation and/or toxicity aresaid to reduce the “bioavailability” of the chemical. Factors affecting bioavail-ability are solubility, dissolution rate, and sorption to dissolved or particulate or-ganic matter. Due to the relative hydrophobicity of the higher PDEs (DBP andlarger), their biodegradation rates may be reduced by low bioavailability.

One might expect that the phthalate esters would all undergo primarybiodegradation at about the same rate. Kurane et al. [30] investigated the degra-dation rate of a series of PDEs by one (Pseudomonas acidovorans) of a numberof isolated PDE degrading organisms. They found that for a series of nine n-alkylPDEs, ranging in length from methyl to tridecyl, the rates were all between about0.1 and 0.3 d–1. Most of the PDEs including DMP and DOP through DTDP hadrates near 0.1 d–1, but higher rates were observed for DBP and DHpP. In a laterstudy [31], the PDE hydrolyzing enzyme from another organism (Nocardia ery-thropolis) was isolated and its substrate specificity and other properties were de-termined. The enzyme had many of the properties of a typical lipase, includinggood degradation of triglycerides (olive oil and tributryn). This enzyme also de-grades all of the phthalates at reasonably good rates, with the highest rates forDBP and DEHP. Based on enzyme specificity alone, one would expect to see onlysmall differences in biodegradation rates among PDE, certainly not differencesof orders of magnitude. The fact that in some test systems the higher molecularweight PDEs, such as DEHP, show much lower biodegradation rates than the low-molecular weight PDEs, such as DMP, may be ascribed to a reduced bioavail-ability of the higher PDEs. This lower bioavailability may be due to lack of solu-bility or to sorption or sequestration by organic matter.

Regarding solubility, the ready biodegradability screening tests, with the ex-ception of the closed-bottle test, use chemical concentrations in the 10–100 mg L–1 range. Within this concentration range, for the higher PDEs, a sig-nificant fraction is insoluble. Nyholm [32] reported on the results of bio-degradation studies with a number of techniques to increase solubilization ofsome poorly water-soluble substances, including DEHP, in a manometric screen-ing test. He found that by using either emulsifiers or increasing the surface areathrough the use of silica gel of glass fiber filters as carriers generally increasedthe biodegradability of these substances. For DEHP, it is evident from the degra-dation curves that the use of the solid carriers may have increased the early rateof degradation. However, the 28-day extent of degradation seems to be decreasedby these materials, perhaps through adsorption. Scholz et al. [14] also noted thattesting biodegradability at concentrations well above water solubility might havelead to some of the variable results reported for PDEs. On the other hand, Gib-bons and Alexander [33] reported that some bacteria (Mycobacterium sp. and Nocardia sp.) excrete products that increase the solubility of DHP, DEHP, DIOP,and DIDP.

Degradation of Phthalate Esters in the Environment 95

Achinger et al. [34] reported on the use of respirometry to evaluate the effectsof solubility on bacterial growth kinetics. They used enrichment cultures in abatch reactor, relating oxygen uptake to maximum growth rate and growth yield(Monod kinetics). They reported maximal growth rates of about 0.1 h–1 andgrowth yields of about 0.5 mg mg–1 for DMP, DEP, DBP, BBP, DnOP, and DEHP.The kinetic model fit the experimental data rather well for DMP, DEP, DBP, andBBP, even when BBP was tested above and below the its water solubility. On theother hand, DnOP and DEHP data fit the model more poorly as degradation pro-ceeded. The authors concluded that this was most likely due to the rate of uti-lization exceeding the rate of solubilization.A modified model to account for sol-ubility effects gave a better fit to the data. Rate constants for substrate utilizationcalculated from the data presented by Achinger et al. [34] in different experimentsranged from about 14 to 3.7 d–1 (t1/2 = 0.05–0.2 d) with little difference among thePDEs tested. Wang et al. [35, 36] reported on the maximal growth rate of DBP ina continuous culture system. They reported a value of 0.38 h–1 with substrate con-centrations of 0.5–1 g L–1, well above the solubility of DBP of about 10–2 g L–1,which may be the reason the rate is so much lower than those of Achinger et al. [34].

Another factor that may reduce bioavailability, and hence apparent biodegra-dation rates, is adsorption to organic matter. Wang and Grady [37] extended thestudy of Achinger et al. [34] to include both dissolution kinetics and sorption tobiomass in the kinetic analysis of DBP utilization by microorganisms. A bac-terium (Pseudomonas fluorescens) which did not degrade DBP was added tobatch cultures at a concentration of 2 g L–1 and degrading organisms from an en-richment culture were added at about 10–2 g L–1. Degradation parameters wereevaluated from 14C mass balance in evolved CO2, liquid phase, and biomass, byusing various concentrations of carbonyl-labeled 14C-DBP. They found that whenDBP was added at concentrations greater than its solubility, the presence of theliving carrier organisms increased the rate of biodegradation compared to sys-tems without carrier. They concluded that desorption from carrier was fasterthan the rate of dissolution of insoluble DBP. It is also apparent that sorption-desorption is reversible and relatively fast compared to either the rate of disso-lution or the rate of biodegradation.

In the study by Steen at al. [29] mentioned earlier, the addition of sediment totest systems in amounts that reduced the dissolved concentrations of DBP by 15%and 90% had only a small effect on the pseudo-first-order biodegradation rates,increasing the rate slightly at 15% and reducing it only by half at 90%. However,this minimal effect of added sediment may possibly be due to a concomitant in-crease in the number of microorganisms present, since there was a much greaterreduction in the second-order rate constant as a result of sediment addition. Theyalso reported that the second-order rates were all approximately the same if theywere corrected for the fraction of BBP actually in solution. In this study the con-centration of non-degraders was much higher than the concentration of de-graders. Of course, when the organism to which the PDE is sorbed is capable ofPDE degradation as well, such sorption would be expected to increase the rate ofdegradation. Yan et al. [38] reported a study on the bioconcentration of DMP,DEP, and DBP by the algae Chlorella pyrenoidosa. They found that the algae not

96 D.R. Peterson and C.A. Staples

only sorbed but also degraded the PDEs at an appreciable rate and the increasein algal biomass resulted in logistic rates.

It seems likely that biodegradation does occur to some extent in the sorbedstate. The higher phthalates in aqueous systems are predominantly sorbed to suspended solids [1, 23]. If they were only biodegraded in the non-sorbed state,the effective rate of biodegradation would be expected to be reduced pro-portionately to the extent of adsorption as pointed out by Steen et al. [29] (as-suming rapid, reversible adsorption). Since DEHP and DnOP have adsorption coefficients to organic carbon (Koc) about three orders of magnitude higher thanthose of DMP and DEP [1], one would expect considerably longer measured half-lives for the higher PDEs, which does not seem to be the case. A possible explanation is that PDEs sorbed to biomass still undergo biodegradation. Thistopic will be addressed again when discussing biodegradation rates in soil andsewage.

4.1.3Biodegradation of Very Low Concentrations

Furtmann [23] has drawn attention to the fact that many PDEs are bio-degraded in laboratory river die-away studies down to a level of several ng L–1

and then degradation stops (in particular DEP, DHxP, and DEHP). The levelsreached often appear to be similar to background levels originally found in the water sample. The very low rate of biodegradation at very low concentrationshas been observed by many for a range of chemicals including glucose. This phenomenon has been discussed by Rubin et al. [39] who also observed it forDEHP. As an explanation of this, they suggested the existence of two types oforganisms involved in chemical degradation: eutrophs and oligotrophs. Eutrophsare capable of growing on relatively high concentrations of the chemical but have a low affinity for it and cannot sustain growth and metabolism at very lowconcentrations. Oligotrophs can degrade very low concentrations of a variety ofchemicals but are less specific and may degrade other chemicals present in thewater rather than the chemical of interest. Another explanation suggested byPagga [40] is that bacteria are unable to produce the necessary degradative enzymes either because a minimum substrate concentration is necessary to induce them or because the substrate is at too low a level to be transported intothe cell. Other possibilities for PDEs are that the chemically detected concen-trations are not bioavailable to the bacteria due to binding to suspended or dis-solved organic carbon or that measured background levels in the water are ac-tually PDE backgrounds introduced during sample analysis do to the ubiquity ofPDEs in the laboratory. The existence of such trace levels is often used as evidencethat PDEs are persistent. However, Furtmann [24] concludes that the trace back-ground levels of PDEs in water, which do not appear to be further biodegraded,are normally below 1 µg L–1 and that these levels are not relevant for toxicologi-cal concerns.

Degradation of Phthalate Esters in the Environment 97

4.2Soil

4.2.1Reported Rates

In soil mobility studies on DMP, DEP, and DBP, Russell et al. [41] noted that thesechemicals biodegraded quickly in soil. They conducted biodegradability studieson these PDEs in soil from two locations in Broome County, New York that hadbeen treated with landfill leachate. Soil suspensions of about 1–2 g soil in50–60 mL of water were shaken at 25 °C with 1-3 mg L–1 PDE.At various time in-tervals, aliquots were extracted and analyzed by gas chromatography (GC). Sim-ilar experiments were done with uncontaminated soil from a nature preserve.The landfill leachate-treated soil gave biodegradation rates of about 0.06 h–1 forDMP, DEP, and 0.04 h–1 for DBP (half-lives of 0.5 and 0.7 d, respectively). The soilfrom the nature preserve gave biodegradation rates of 0.015, 0.016, and 0.067 h–1

for DMP, DEP, and DBP, respectively (t1/2 =1.9, 1.8, and 0.4 d). In a similar ex-periment, Kirchmann et al. [42] found DEHP to follow zero-order kinetics in soil with rates of 0.20 and 1.76 mg kg–1 d–1 at initial concentrations of 5 and250 mg kg–1.

The soil suspension method used by Russell [41] could be criticized for not being an exact simulation, since it is essentially an aqueous system with soil bacteria. Rüdel et al. [43] investigated the soil biodegradation rate of DEHP insimulation systems. They compared the first-order rates using 1 mg kg–1 DEHPin two different soils: a silty sand and a silty loam. They also compared rates be-tween laboratory systems under standardized conditions of soil moisture (40%of holding capacity) and temperature (20 °C) with outdoor lysimeter experimentsand laboratory flasks with varied moisture and temperature regimes to mimicoutdoor conditions. The reported DEHP half-lives under laboratory conditionswere 20 d for the loam and 68 d for the sand. The higher rate in loam (0.035 d–1)than sand (0.010 d–1) was thought to be due to the approximately threefold higherbiomass in the loam. Bioavailability differences must not have played as great arole, since the loam had twice the organic carbon as the sand (2% vs. 1%), andwould have been expected to reduce bioavailable concentrations proportionallyand give a lower rate in the loam. Moisture and temperature had a significant effect on the rates, since the half-lives in the simulated outdoor systems were 31 d(first-order rate = 0.022) for the loam and 170 d (rate = 0.004) for the sand and thehalf-lives in the lysimeters were 21 d (rate = 0.033) and 54 d (rate = 0.013) in thetwo soils. The presence of plant growth (barley) had an uneven effect on DEHPdegradation. These workers also reported on some noteworthy aspects of DEHAmetabolism in soil. By using 14C-labeled DEHP, they found no metabolites besidesCO2, that is, the undegraded residue was probably parent DEHA. However, theyalso found that after long incubation times, a significant fraction of the radioac-tivity was non-extractable. An average extractability in the laboratory systemswas 34 and 39% in 64 and 100 d, respectively, for the silty loam and an average of11% and 13% in 64 and 100 d for the silty sand. Other chemicals (biocides) theytested showed similar behavior. The implications of this finding will be addressedlater.

98 D.R. Peterson and C.A. Staples

Shanker et al. [44] published on the biodegradation of DMP, DBP, and DEHPin soil. They incubated samples of garden soil containing a reasonably high con-centration of PDEs (500 mg kg–1) in Erlenmeyer flasks at 30 °C and 60% moistureholding capacity for varying time periods and then analyzed by HPLC for par-ent PDEs and metabolites. Their quantitative data may be analyzed to evaluatepseudo-first-order rates. The sample intervals were at time zero and then at 5 dayintervals for up to 30 days. The rates we calculate from their data for DMP, DBP,and DEHP are 0.40, 0.39, and 0.12 d–1, respectively (t1/2 =1.7, 1.8, and 5.6 d). Therewas a lag phase of about 5 days for DEHP but not for the other two PDEs. It thelag phase is omitted, the rate for DEHP is 4.3 d–1. They found no appreciableamounts of metabolites, except for PA (phthalic acid), which was found in lowconcentrations during intermediate stages of degradation and never ap-proached the concentrations of the parent PDE in the soil.

The biodegradation rates of DBP and DEHP in three different soils were re-ported by Chen et al. [45]. They also used 500 mg kg–1 of PDE, incubating the soilsat 30% soil moisture and 28 °C. They extracted and analyzed the PDEs by GC-FIDat 5-day intervals up to 30 days. They found that the data fit first-order kineticswith no appreciable lag phase. The biodegradation rates were 0.103, 0.062, and0.044 d–1 for DBP and 0.040, 0.019, and 0.015 d–1 for DEHP. The highest rates forboth PDEs were in the soil with the intermediate organic carbon (2%), while thehighest organic carbon (3.3%) soil gave the intermediate rates. There were noclear relationships between biodegradation rates and other soil properties. AsRüdel et al. [43] had earlier reported that they found biodegradation rates in-creased with increased soil temperature. They also reported that in sterile soil, ex-tractable PDEs decreased with time, particularly for DEHP, which decreased bynearly 20% in 30 days in the low-organic carbon soil. Wang et al. [46] reportedon the biodegradation of DBP in soil microcosms at 60% moisture holding ca-pacity and 25 °C. DBP was initially present at 100 mg kg–1 soil. They report con-centration versus time data that can be used to evaluate the first-order rate. Therate determined from their data is 0.036 d–1. Inoculation of the soil with a DBP-degrading bacterium greatly enhanced this rate. The microbial population of theun-inoculated soil was about 4 ¥105 colony forming units g–1 and only decreasedslightly by the end of the 30-day incubation period.

Cartwright et al. [47] recently reported on the effects of DEP and DEHP on soilmicroorganisms.As part of the study, they investigated the degradation of thesetwo PDEs by indigenous soil organisms. The soil was sandy clay loam soil con-taining 3.78% organic carbon and the PDEs were added at concentrations of 0.1,1, and 10 g kg–1. Sub-samples of the amended soils were incubated in 28 mLsealed bottles at 20 °C and 50% soil moisture. The bottles containing 1 and 10 gkg–1 PDE were opened for an hour every 5 days to allow for aeration. Separatecontainers were extracted and analyzed by HPLC for remaining PDE at varioustime intervals. They report that DEP degraded rapidly with “half-times” of 0.75,5, and 17 d at 0.1, 1, and 10 g kg–1. DEHP only degraded by about 10% in 70 d.They proposed reduced bioavailability of DEHP, due to greater soil adsorption,as the reason for the difference in biodegradation rate between the two PDEs.Toxicity to soil organisms at the higher levels was not a reason for the differencebetween the biodegradability of the two PDEs, since the DEHP-containing soil

Degradation of Phthalate Esters in the Environment 99

showed no significant concentration related difference in total bacteria (colonyforming units, about 3 ¥ 107 g–1 soil) at any of the sample intervals. The half-timesfor degradation are evidently not first-order half-lives but the time to reach 50%of nominal additions. The authors give a graph of the PDE versus time values forthe 0.1 g kg–1 experiment and re-plotting values from this graph for DEP on a nat-ural log scale gives a reasonably straight line (r2 = 0.98) with a rate constant of0.28 d–1 (t1/2 = 2.5 d). The DEHP curve is quite interesting; after a decline to about80% of its initial concentration in the first two weeks, it showed no furtherchanges. This is further discussed in the next section.

4.2.2Conditions Affecting Soil Biodegradation

The extent of biodegradation of DEHP over time observed by Cartwright et al.[47], as discussed above, shows an unexpected pattern. If adsorption to soil wereresponsible for decreased biodegradability, one would expect to see a continuingbiodegradation at a rate consistent with the concentration of the equilibriumdesorbed concentration over the entire 70 days, unless the adsorption was irre-versible. In another recent paper (Jensen et al. [48]) on the toxicity of DBP andDEHP to a soil invertebrate (Folsomia fimetaria), the authors also measured DBP concentrations over time for soil with initial concentrations of 50 and500 mg kg–1. The data over 25 days very closely fit a model the authors proposedwherein adsorption and degradation occur simultaneously but wherein desorp-tion is negligible. The model predicts that after 30 days about 50% of the DBP will be sequestered and unavailable for biodegradation.

The lack of linearity of logarithmic plots of PDE biodegradation in soil hasbeen examined in a number of studies on PDE mineralization [49–52]. Thesestudies are not strictly applicable to determining rates of primary biodegrada-tion. When primary degradation and mineralization are measured in the samestudy, mineralization, as measured by carbon dioxide evolution, generally pro-ceeds at a slower rate and to a lesser extent than primary degradation. This dif-ference may be ascribed to the formation of metabolites that are further de-graded at a slower rate than the parent substance. Another reason is theincorporation of carbon into growing biomass. For instance, Roslev et al. [51]found that a significant portion of 14C-DEHP was incorporated into phospholipidfatty acids of soil organisms. They also observed that a significant fraction of CO2in long-term degradation assays originated from turnover of biomass. Never-theless, such studies of mineralization may provide information on the condi-tions affecting the overall biodegradation process. They also constitute a lowerlimit for biodegradation rate, since the first biodegradation step (primary) willalways be as fast or faster than complete conversion to CO2.

Fairbanks et al. [49] observed that mineralization of DEHP in three New Mex-ico soils gave an initial rapid rate, with little or no lag, followed by a slowing ofthe rate with time. They stated that they did not include first-order rate constantsin their report due to the complexity of the curves, but only presented log DEHPconcentration (calculated based on the percent of 14CO2 produced) versus timecurves. The initial points of these curves are relatively linear and correspond to

100 D.R. Peterson and C.A. Staples

rates of about 0.035, 0.069, 0.058 d–1 at an initial soil concentration of 2 mg kg–1.At 20 mg kg–1, initial rates are 2- to 4-fold slower. These workers also reporteddata for rates in sewage sludge amended soil, which gave a similar pattern.

Dörfler et al. [50] did analyze the kinetics and derived a complicated kineticexpression describing DEHP mineralization. Roslev et al. [51] studied DEHPmineralization in sludge-amended soils, including an analysis of the complex kinetics. They report that the mineralization of DEHP could be divided into two distinct kinetic phases: an initial phase over 28 days that followed first-orderkinetics and a late phase that was slower and a function of an exponent of time(t–0.387). Half-time for the initial phase was 58 d (first-order rate = 0.012 d–1) andfor the late phase was 147 d. This same group later published a more detailed re-port on the rates of the two phases [52]. The rates were similar for untreated andsludge-amended soil.At 20 °C the initial first-order rate of DEHP biodegradationin soil was 0.0134 d–1 (t1/2 = 52 d) and in sludge-amended soil it was 0.0127 d–1

(t1/2 = 55 d).The second phase of mineralization observed by these workers is difficult to

interpret, since they are not measuring primary degradation of DEHP but pro-duction of CO2. A number of other slow processes may limit ultimate degrada-tion to CO2 of a fraction of the 14C originally in the DEHP, particularly since, asnoted in the Roslev et al. [51] paper, a significant portion of the 14C was incor-porated into lipids and eventually was released through turnover of the biomass.Another explanation of the second slow phase of release is the possible “seques-tration” of the parent DEHP within soil solids. This phenomenon has been re-ported on and investigated by numerous researchers [53–56]. A wide range ofhydrophobic chemicals have been shown to undergo such sequestration,wherein after aging of the chemicals in soil (or sediment), they penetrate intosolid particles and are no longer surface adsorbed and available to some extrac-tion solvents. The chemicals are also not available for biodegradation or toxicity.Unlike sorption-desorption equilibrium, which is relatively fast in relation tobiodegradation rate, diffusion of the entrapped chemical out of the solid is a slowprocess with a half-life of the order of ten days or more [55].

The rates of first-order reactions are concentration independent. The abovestudies on mineralization of DEHP showed a considerable dependence of rateson initial test material concentration. Fairbanks et al. [49] used initial concen-trations of 2 and 20 mg kg–1 and half-lives were 2- to 4-fold longer at the higherconcentration in the three soils. Dörfler et al. [50] studied 0.5 and 10 mg kg–1 con-centrations and report that higher initial concentration resulted in lower percentdegradation in all three soils tested. Madsen et al. [52] reported very similar first-order rate constants for DEHP at initial concentrations of 1.6, 3.2, and 9.9 mg kg–1

in sludge-amended soil. They reported rates of 0.0087, 0.0081, and 0.0078 d–1,respectively. However, an initial concentration of 35.1 mg kg–1 gave an order ofmagnitude higher rate of 0.090 d–1.

The rate of biodegradation of a chemical in soil will differ considerably de-pending on how the concentrations are measured and how rapid desorption rel-ative to the rate of biodegradation. Assuming for the moment that a chemicaldoes not biodegrade when tightly adsorbed to soil solids but only when it is inpore water (bioavailable), then one would expect that biodegradation would be

Degradation of Phthalate Esters in the Environment 101

much slower for chemicals with a high Koc. For instance, comparing DEHP withDMP, the Koc value of DEHP is more than three orders of magnitude higher, soits bioavailable concentration in soil is expected to be three orders of magnitudelower. As a consequence, the apparent biodegradation rate, based on total soilconcentration change would be three orders of magnitude lower if the two PDEshave the same aqueous biodegradation rate. If one measures only the concen-tration change in the aqueous phase, and if the desorption of DEHP from the soilis rapid relative to the biodegradation rate, then the apparent degradation rate isstill slow because as DEHP is degraded from the water, it is quickly replenishedfrom the sorbed phase. However, if the desorption is irreversible, then the ap-parent biodegradation rate is much faster if one is measuring concentrationchange in the water phase, as shown in Fig. 1a. Moreover, if one measures the to-tal decrease in soil concentration when desorption is irreversible, a logarithmicplot of the data versus time gives a non-linear curve with a decreasing rate,Fig. 1b. If one assumes a slow rate of desorption relative to biodegradation, thenthat rate becomes limiting. In fact this probably explains the kinetics observedby Dörfler et al [50], Roslev et al. [51], and Madsen et al. [52].

However, it is apparent that there must also be some biodegradation of thePDEs in the sorbed phase. In soil, with water making up about 12% of its weightand organic matter 2%, one would expect that for the higher PDEs with Koc val-ues greater than 10,000, the fraction dissolved in soil pore water would be lessthan 0.01%. One would expect that biodegradation rates of these PDEs would bemany orders of magnitude lower in soil than in water if only this dissolved frac-tion were biodegradable, since the numbers of microorganisms in soil is not1000-fold higher than in water. But, in fact the biodegradation rates of the PDEsin soil are very similar to those in water. There are also a number of reports ofcertain organisms degrading PDE in the solid phase, as will be described later.

Temperature plays an important role in biodegradation rate. Rüdel et al. [43]ascribed the major reason for the differences between laboratory and simulatedoutdoor soil degradation systems to temperature. When they corrected for tem-perature differences by using a factor of 50% decrease in rate with a 6.5 °C de-crease in temperature, the two systems agreed well for DEHP and the pesticidesand biocides tested. Chen et al. [45] studied the biodegradation of DEHP at threedifferent temperatures: 10, 28, and 35 °C. Unfortunately they did not report therates at these temperatures; however, the extent of degradation at the three tem-peratures was 21.5, 28.5, and 33.2%, respectively. Madsen et al [52] also investi-gated the effect of temperature on the rate of mineralization of DEHP in sludge-amended and un-amended soil at 5, 10, and 20 °C. Each doubling of thetemperature resulted in a doubling of the rate.

Soil moisture has been reported in many of the studies as being an importantcondition for soil biodegradation, with higher soil moisture resulting in a greaterrate of biodegradation or mineralization. This is expected since bacteria live inthe water phase and require water for diffusion of nutrients into and out of thecell. It is also likely that the phthalates are degraded most rapidly in the aqueousphase and that increased volumes of interstitial water in the soil will result in agreater dissolved amount PDE. However, if the soil is flooded and its maximummoisture holding capacity exceeded, there is a possibility that anaerobic condi-

102 D.R. Peterson and C.A. Staples

Degradation of Phthalate Esters in the Environment 103

Fig. 1. a Modeled slurry biodegradation, concentration change with time in the water phase,comparing reversible (dashed line) and irreversible (solid line) adsorption. Conditions:Koc = 104, solids 600 g L–1, 5% organic carbon, k1 = 1 d–1. b Modeled slurry biodegradation, con-centration change with time in entire system, comparing reversible (dashed line) and irre-versible (solid line) adsorption. Conditions the same as Fig. 1a

a

b

104 D.R. Peterson and C.A. Staples

tions will occur.A lack of oxygen will be expected to result in a slower biodegra-dation rate in soil; for instance, Madsen et al. [52] report a rate of anaerobicdegradation of DEHP in sludge-amended soil of 0.0023 d–1 as compared with0.0127 d–1 in the same soil under aerobic conditions. Anaerobic biodegradationrate in sediment and soil is treated below as a separate topic.

The literature values for the pseudo-first-order half-lives for PDEs in soil dis-cussed in this section are summarized in Table 3.Although the data are sparse forDMP and DEP, it is expected that these two PDEs should not differ greatly in pri-mary degradation rate due to their similar chemical properties. Their biodegra-dation rates range from 0.36 to 0.40 d–1. There are more data for DBP, rangingfrom 0.036 to 1.6 d–1. The value of 1.6 was determined on an aqueous suspensionof soil rather than soil alone. A value of about 0.1 d–1 would seem to be repre-sentative for the remaining DBP rate data. Clearly, the higher extent of soil sorp-tion of DEHP results in somewhat lower bioavailability and lower biodegradationrate. However, the primary rate data would indicate that degradation might oc-cur to some extent in the sorbed phase. There are considerable data for a primarybiodegradation rate of DEHP in the range of 0.01–0.1 d–1.A biodegradation rateof 0.03 d–1 would seem to be representative of these data. It must be rememberedthat the use of first-order rates is a simplification of the process of degradation.As we have seen, DEHP can become sequestered in soil and unavailable forbiodegradation. In which case, biodegradation of this sequestered fraction wouldseem to be limited by the rate of reversal of this sequestration process.

Table 3. Summary of pseudo-first-order PDE biodegradation rates in aerobic soil

Phthalate First-order Half-life Test conditions Ref.rate (d–1) (d)

DMP 0.36 a 1.93 Aqueous suspension, agitated [41]0.40 1.7 Flask, 30 °C, 60% WHC b, garden soil [44]

DEP 0.38 a 1.83 Aqueous suspension, agitated [41]DBP 1.61 a 0.43 Aqueous suspension, agitated [41]

0.39 1.8 Flask, 30 °C, 60% WHC, garden soil [44]0.103 6.7 Flask, 28 °C, 30% WHC, 2% OC c [45]0.044 11.2 Flask, 28 °C, 30% WHC, 3.3% OC [45]0.062 15.8 Flask, 28 °C, 30% WHC, 1.6% OC [45]0.036 19.3 Flask, 25 °C, 60% WHC [46]

DEHP 0.035 2.0 Flask, 20 °C 49% WHC, loam [43]0.010 69.3 Flask, 20 °C 49% WHC, sand [43]0.033 21 Outdoor lysimeter, loam [43]0.013 53.3 Outdoor lysimeter, sand [43]0.12 5.6 Flask, 30 °C, 60% WHC, garden soil [44]0.040 17.3 Flask, 28 °C, 30% WHC, 2% OC [45]0.019 36.5 Flask, 28 °C, 30% WHC, 3.3% OC [45]0.015 46.2 Flask, 28 °C, 30% WHC, 1.6% OC [45]0.012 58 Sludge amended loam, 75% WHC [51]

a Value calculated from data presented in the referenced paper.b WHC water holding capacity.c OC organic carbon content of the soil.

Degradation of Phthalate Esters in the Environment 105

4.3Wastewater

A number of studies have shown that PDEs degrade rapidly in wastewater (for in-stance Furtmann [23]). The standard screening tests for biodegradability alsogenerally use aerobic sludge from a wastewater treatment plant (WWTP) as aninoculum source and the ready biodegradability of PDEs would indicate thatcompetent organisms are present in this sludge. The issue this section seeks to ad-dress is the rate and/or extent of biodegradation in a wastewater treatment sys-tem, which of course is dependent on the specific system employed. Wastewatertreatment systems have a much higher concentration of microorganisms thanstandard screening test systems. One method that has been used to evaluate po-tential removal of chemicals from wastewater in a WWTP is a semi-continuousactivated sludge (SCAS) test. Saeger and Tucker [19] evaluated removal of a num-ber of PDEs (MBP, BBP, DEHP, di-711, and DUP) in a SCAS system. The systemis a cylinder containing aerated activated sludge from a domestic treatment plant.Once daily, aeration is stopped and the sludge is allowed to settle. Supernatant liq-uid is drawn off and replaced with water containing nutrients (synthetic sewage)and test chemical.After an acclimation period, samples are withdrawn at varioustimes subsequent to chemical addition, and the extent of degradation (or re-moval) evaluated.At PDE addition rates up to 200 mg d–1, Saeger and Tucker [19]found >99% primary biodegradation for MBP and BBP, as determined by sludgeremoval and analysis. The degradation of DEHP, di-711, and DUP was 78, 52, and45%, respectively, at 5 mg d–1 additions. They also reported the time course ofbiodegradation of BBP, which declined to about 50% in slightly over an hour andby about 98% in 6 h. Soluble by-products reached a maximum by 5 h and weredegraded by 14 h. Results of SCAS testing on DBP were also reported by Wang et al. [57]. They found essentially the same degradation rate at concentrations of50–200 mg L–1 and concluded they were at zero-order kinetics and that the rateconstant could be equated with the saturation biodegradation rate. The rate con-stant they determined was 0.015 h–1 (or 0.36 d–1).

SCAS systems are not strictly simulations of WWTP because they are fed in-termittently rather than continuously. Petrasek et al. [58] studied the removal ef-ficiencies of a number of PDEs (DMP, DEP, DBP, BBP, DnOP, and DEHP) in a lab-oratory sewage simulation system at nominal concentrations of 50 µg L–1. Thesystem was fed with raw sewage from a nearby domestic WWTP. Sludge retentiontime was seven days. The average concentrations of the PDEs in the activatedsludge effluent was around 1 µg L–1 for all but DEHP, with a number of non-de-tects averaged in as the detection limit. DEHP was detected in all samples of ac-tivated sludge effluent at an average concentration of 11.3 µg L–1. For the morevolatile DMP and DEP, a significant portion of the removal may be attributableto air stripping. Tokuz [59] has also reported on the removal of DMP and DEP ina laboratory-scale sewage treatment system. The system was fed with syntheticsewage. The mixed liquid suspended solids (MLSS) were at about 1100 mg L–1.Addition of DEP and DMP was slowly increased to 410 mg L–1 and 540 mg L–1,respectively, by day 19. The total influent COD (chemical oxygen demand) rose from about 600 to 2000 mg L–1, while the effluent COD remained low

< 100 mg L–1. MLSS increased to 1600 mg L–1, indicating that a considerable por-tion of the PDEs was converted to biomass. Although indicating that DMP andDEP were appreciably removed by treatment, quantitative removals or rates arenot possible for these data, since specific analyses were not conducted.

Another approach to evaluating the removal rate of chemicals in a WWTP isto evaluate influent and effluent concentrations in an actual WWTP. Paxéus et al.[60] conducted such a study over the course of three years (1989–1991) at theGöteborg (Sweden) Regional Sewage Works. They found average influent waste-water concentrations in the low µg L–1 for DMP, DEP, BBP, and DnOP. Annual average influent concentrations were 36–86 µg L–1 for DBP and 30–40 µg L–1 forDEHP. Effluent concentrations were below detection levels for all PDEs exceptthese last two. Effluent concentrations of these were 0.1–2.0 µg L–1 for DBP and0.3–2.0 µg L–1 for DEHP. Removals from wastewater appear to be in the range of90% or greater but the contribution of biodegradation to the total removal can-not be evaluated from these data. Furtmann [23] also analyzed influent and efflu-ent concentrations of a number of PDEs from two WWTPs, one treating domes-tic sewage and the other industrial sewage. DEP, DIBP, DBP, BBP, DEHP, and DOPwere all in the 1–10 µg L–1 concentration range in the influent to both plants.DEHP was at 25 µg L–1 in the domestic sewage and 71 µg L–1 in the industrialsewage. DOP concentration was low (1.3 µg L–1) in the domestic sewage and high(36 µg L–1; average of two sampling dates) in the industrial sewage. All of thesePDEs were removed by 98% in the effluents of both plants, with the exception ofDEP and DBP. These were removed by 88% (DEP) and 83% (DBP) in the do-mestic plant and 93% and 95% in the industrial WWTP. Of course, it is likely thatthe higher PDEs partition primarily with the waste sludge [23]. There are a num-ber of papers on the occurrence of PDEs in sludges from WWTPs [24, 61, 62] butthese data shed little light on the issue of biodegradation rate without having con-current influent and effluent concentrations. At best, it can be concluded that aportion of the higher phthalates remains sorbed to the waste sludge. For instance,Zurmühl [61] reported sludge concentrations (in units of mg kg–1 dry weight(d.w.)) ranging from 2.6 to 263 for DBP, non-detectable to 0.7 for BBP, and65.8–481 for DEHP, while no DMP or DEP was detected. Calculated mean valueswere 35, 0.3, and 179 for DBP, BBP, and DEHP. The DBP mean is considerably in-fluenced by a single high value, as the median is 4.1 mg kg–1.

Mickelson et al. [63] used a WWTP computer model (SimpleTreat) and com-pared the results with removals in three Danish WWTPs to evaluate the biodegra-dation rate constant. They assumed the rate constant was the major source of er-ror in the model. Observed removals for DEHP in three WWTPs averaged 85%and the modeled results for the same three plants averaged 93%.A calculation ofthe biodegradation rate necessary to obtain this result, gave 0.55 L gss

–1 d–1 at 20 °C.Assuming a standard sludge retention time of 8.5 d, the extent of biodegradationfor a system with no primary settler was 11% and for a system with a primary set-tler was 6%. They calculated that the major portion of the influent DEHP, 78%and 56% for these two treatment systems, respectively, goes to waste sludge. Clarket al. [64] published results on a model that evaluates fate of PDEs in a WWTP(STP). They estimated half-lives of DBP and DEHP of 100 h in a 2 g L–1 MLSS sys-tem. This is equivalent to a rate of 0.083 L gss

–1 d–1, somewhat lower than that

106 D.R. Peterson and C.A. Staples

Degradation of Phthalate Esters in the Environment 107

determined by Mickelson et al. [63]. The resulting modeled removals were 81%for DBP and 91% for DEHP, with biodegradation accounting for 27% of the re-moval of each.

We have also used an updated spreadsheet version (3.0) of SimpleTreat [65] tomodel the data given by Furtmann [23] for WWTPs, treating industrial wastewaterand domestic wastewater. Furtmann analyzed the influent and effluent wastewaterconcentrations in addition to activated sludge concentrations for DEP, DIBP, DBP,BBP, DEHP, and DOP. A mass balance is not strictly possible, since the analyseswere performed on grab samples that Furtman characterizes as “random,” al-though they were all taken on the same day. The problem is that the differencesin sludge and water retention times may result in a lengthy delay before a changein influent concentration manifests itself in sludge concentration, as may be seenin the DEHP concentrations in the domestic WWTP in Table 4. The influent con-centration of DEHP is quite low, while the sludge concentration is quite high.For the industrial WWTP, analyses from two sampling dates were averaged.Table 4 shows the concentrations in these two plants in influents, effluents, andsludges. Standardized European conditions [66], including the default WWTPbiodegradation rate for ready biodegradable chemicals of 1 h–1, together withphysical properties of the PDEs from Cousins and Mackay [67], were used in themodel. The calculated concentrations shown in Table 4 were obtained for aWWTP without primary sedimentation. The default biodegradation rate of1.0 h–1 greatly underestimates the total effluent concentrations, except for DEP.The amount calculated as un-biodegraded is greater for the higher molecularweight PDEs, although the rates are all the same. This is a result of the assump-tion that no degradation occurs in the adsorbed phase.

The SimpleTreat model (version 3) also has the option of assuming the samefirst-order kinetic rate in both the aqueous and solid phase. By using the modelin this mode, one must assume a lower rate constant to fit the measured data because 1.0 h–1 results in an overestimation of total removal. Adjustment of therate constant to 0.03 h–1 gives approximately correct concentrations in sludge(77 mg kg–1) and effluent (2.3 µg L–1, mostly associated with suspended solids).However, the assumptions of degradation in the solid phase and a rate of 0.03 h–1

gives too low a total removal for DEP with 77% of the influent concentration stillin the aqueous effluent. The rate has to be raised to nearly 1 h–1 to get agreementwith the measured removals of DEP. Thus, the truth probably lies somewhere inthe middle, with biodegradation occurring in solid and liquid phase at differentrates. The main point is that degradation is probably occurring in the sorbedphase. This may be partly a result of sorption to organisms capable of degradingPDEs. It also may be a result of the fact that interfacial surfaces may serve to con-centrate both the PDE and degrading organisms [8]. Wang et al. [68] have re-ported that microbial cells immobilized within gel beads biodegrade DBP fasterthan freely suspended cells.

Another phase that may contain PDEs in sewage is small particles of abradedPVC. There is evidence that plastic particles, probably abraded from surfaces bymechanical action, are found in wastewater and sludge. Teinpoint et al. [69] usedpyrolysis GC-MS to quantify PVC in sludges from nine different WWTPs. Theconcentration of PVC ranged from 18 to 508 mg kg–1 (d.w.). In these sludge sam-

108 D.R. Peterson and C.A. Staples

Tabl

e4.

PDE

rem

oval

by

WW

TPs

,mea

sure

d an

d m

odel

ed

Was

tew

ater

sou

rce:

Influ

ent

Dom

esti

cIn

dust

rial

a

Con

cent

rati

ons:

(µL–1

)Ef

fluen

tSl

udge

Influ

ent

Efflu

ent

Slud

ge(µ

L–1)

(mg

kg–1

)(µ

L–1)

(µL–1

)(m

g kg

–1)

PDE

Valu

eD

EPM

easu

red

0.52

0.06

–1.

00.

06<

0.01

Mod

eled

b ,kb1

=1.

0h–1

0.52

0.07

0.01

1.0

0.01

0.02

Mod

eled

,kb2

=0.

03h–1

0.52

0.40

0.05

1.0

0.77

0.10

DIB

PM

easu

red

2.3

0.07

0.3

5.4

0.08

0.02

Mod

eled

,kb1

=1.

0h–1

2.3

0.33

2.0

5.4

0.78

4.6

Mod

eled

,kb2

=0.

03h–1

2.3

0.36

2.1

5.4

1.3

7.4

DBP

Mea

sure

d1.

30.

220.

88.

20.

301.

4M

odel

ed,k

b1=

1.0

h–11.

30.

191.

18.

21.

26.

9M

odel

ed,k

b2=

0.03

h–11.

30.

211.

28.

21.

37.

4BB

PM

easu

red

0.8

n.d.

c1.

02.

4n.

d.<

0.01

Mod

eled

,kb1

=1.

0h–1

0.8

0.12

1.4

2.4

0.35

4.3

Mod

eled

,kb2

=0.

03h–1

0.8

0.07

0.8

2.4

0.20

2.4

DO

PM

easu

red

25.0

n.d.

1436

.0n.

d.2.

3M

odel

ed,k

b1=

1.0

h–125

.03.

8012

836

.05.

518

4M

odel

ed,k

b2=

0.03

h–125

.00.

8027

36.0

1.2

38D

EHP

Mea

sure

d1.

30.

0315

371

.00.

0892

Mod

eled

,kb1

=1

h–11.

30.

206.

671

.010

.936

0M

odel

ed,k

b2=

0.03

h–11.

30.

041.

471

.02.

376

aM

ean

ofm

easu

red

valu

es fr

om tw

o in

dust

rial

WW

TPs

.b

k b1

first

-ord

er b

iode

grad

atio

n ra

te a

pplie

d on

ly to

aqu

eous

pha

se,k

b2ap

plie

d to

bot

h liq

uid

and

solid

pha

ses.

cn.

d.no

t det

ecte

d.

Degradation of Phthalate Esters in the Environment 109

ples, the sum of DEHP, DINP, and DIDP (phthalates commonly used as plasti-cizers in PVC) concentrations ranged from 14 to 297 mg kg–1. Furthermore, thePDE concentration (sum of these three) correlated very closely with the con-centration of PVC (r2 = 0.89). If PDEs are contained within a plastic matrix, onewould expect that they are not very bioavailable for biodegradation. The degra-dation of PDEs in the solid phase will be discussed further as a separate topic.

For purposes of modeling the rate of biodegradation of PDEs in wastewater, theassumption of a biodegradation rate of 1 h–1 seems to provide a conservative es-timate for the more water-soluble PDEs (e.g., DMP, DEP, DBP, and BBP). However,if the model algorithm does not provided for degradation in the sorbed phase, thisrate provides too low an estimate of biodegradation for the higher PDEs. A moreappropriate model is one that allows degradation in the sorbed phase, in whichcase, a rate of 0.03 h–1 for overall degradation seems to fit the higher PDE.

Recently, Fauser et al. [70] published a report on a detailed investigation of thebehavior a number of phthalates (and two surfactants) in a specific, municipalwastewater treatment plant (Bjermarken WWTP in Roskilde, Denmark.) Theyanalyzed influent and effluent, together with sludge, for concentrations of DEHP,DPP, DBP, BBP, DOP, and DNP. Influent concentrations were daily composites, in-fluent were daily grab samples, and sludge was a single day sample. Total hy-draulic retention time of the plant is 46 h (19 h for biological treatment) and thesludge age (residence time) is 20 days. The analytical results are shown in Table 5.DEHP was present in influent wastewater at 35 µg L–1 and the other PDEs wereall at or below about 1 µg L–1. Removals were >95% for all but BBP and DBP. Apurpose of the Fauser et al. [70] study was to evaluate the applicability of a con-tinuous flow model like SimpleTreat to a site specific model of a WWTP with al-ternating aerobic and anoxic cycles (4 h cycle) from separate treatment units.They found that the variation from the constant steady-state removal ratebrought about by the cyclic nature of the process was far less for substances thatare hydrophobic and adsorb to the sludge than for water-soluble substances.Their modeling also showed that for substances with half-lives longer than about

Table 5. Measured concentrations and modeled biodegradation in Roskilde [70]

Phthalate Measured Concentrations k1a Disposition (%)

(h–1)Inlet Outlet II° Sludge Biode- Aqueous Sludgeb

(µg L–1) (µg L–1) (mg kg–1) graded effluent effluent

DBP 1.03 0.91 0.16 – c – – –BBP 0.39 0.13 0.01 0.009 48 15 37DPP 0.07 0.008 n.d.d 0.053 71 8 21DEHP 35.4 0.96 3.51 0.032 70 2 28DOP 0.57 0.013 0.05 0.024 63 3 34DNP 0.44 0.013 0.05 0.021 61 2 37

a Pseudo-first order biodegradation rate for aerobic degradation obtained by modeling mea-sured data in aggregate flow model.

b Disposition values for primary and secondary sludge were combined.c Due to insufficient data, DBP was not modeled.d n.d. not detected.

2 h, a simpler model of the aggregate flows could be used. They assumed thatbiodegradation occurred only in the soluble state and represented the aerobicpseudo-first-order rate constant as k1/R where R is the sorbed fraction, equal to1+ kd · [particulate mater]. Fitting the model to the measured data, they deter-mined half-lives equivalent to the aggregate biodegradation rates shown inTable 5. These rates range from 0.009 h–1 for BBP to 0.053 h–1 for DPP. The corre-sponding fraction of the total influent biodegraded is also shown in Table 5.These ranged from 48% for BBP to 71% for DPP.

4.4Anaerobic Rates

The fate of PDEs in anaerobic sewage sludge is important, since many WWTPsuse an anaerobic digestion process prior to sludge disposal. Shelton et al. [71] reported on the biodegradation of a number PDEs in laboratory digesters containing undiluted sludge and 10% sludge. The systems were purged with a gasmix of 90% N2 and 10% CO2. Initial PDE concentrations were 20 mg L–1.Aliquotswere removed at various intervals, dried, and extracted with solvent, and thePDEs were determined by GC. Although they did not report rates of biodegra-dation, rates may be evaluated from their graphs of the data. For the undilutedsludge, DMP, DEP, and DBP all were essentially completely degraded by the firstsampling interval (7 d), so the rates for these were >0.3 d–1. BBP had a degrada-tion rate of 0.056 d–1. DEHP and DOP degraded considerably slower, with ratesof 0.001 d–1 and 0.006 d–1, respectively. In the 10% sludge experiment the systemswere sampled more frequently and the rates were slower; thus, rates of 0.42, 0.069,0.073, and 0.096 d–1were calculated for DMP, DEP, DBP, and BBP, respectively.DEHP and DOP had quite inconsistent results, with all concentrations, includingtime zero, varying between 40 and 60% of nominal. There must have been somedegradation in the 10% sludge, since both gave about 10% of theoretical methaneproduction, while the other PDEs gave 80–100%. Subsequently, Ziogou et al. [72]determined the rates of biodegradation of six PDEs in anaerobic sewage sludge,amending the sludge at PDE concentrations in the range of 30–600 mg kg–1. Theyreported pseudo-first-order rates of 0.21, 0.14, 0.25, and 0.16 d–1 for DMP, DEP,DBP, and BBP, respectively. Except for DMP, these rates are about twice as fast asthose derived from the data of Shelton et al. [71]. DOP and DEHP did not degrademeasurably during the 32 day test.

More recently, Painter and Jones [73] studied the anaerobic biodegradation ofDBP, BBP, and DEHP in an anaerobic biodegradation test system by using 10%inocula from various sources. The PDEs were incubated at 0.2 mM and analyzedby hexane extraction-GC. The anaerobic sludge inoculated system gave degra-dation of DBP and BBP but not DEHP. The rates calculated from their data are0.025 d–1 for DBP and 0.013 d–1 for BBP. Wang et al. [74] also studied the rate ofPDE biodegradation in undiluted anaerobic digester sludge from a domesticWWTP. They report first-order rate constants of 0.696 d–1 for DMP, 0.518 d–1 forDBP, and 0.0336 d–1 for DOP.

Johnson and Lulves [75] studied the biodegradation of DBP in anaerobic pond sediment. They used 14C-DBP and separated parent form metabolites

110 D.R. Peterson and C.A. Staples

Degradation of Phthalate Esters in the Environment 111

by thin-layer chromatography. They present data on the percent recovery ofDBP versus control after 1, 5, 7, and 14 days of incubation that may be used to calculate a first-order rate of degradation of 0.27 d–1. The anaerobic degrada-tion of DMP, DBP, and DEHP in garden soil was studied by Shanker et al.[44]. They amended the soil with 500 mg kg–1 of PDE. The soil was then flooded with water and the flasks stoppered to achieve anoxic conditions.The PDEs and PA were extracted and analyzed at various intervals. Although they did not calculate rates, these may be calculated from their tabulated data. The calculated first-order rates are 0.033 d–1, 0.036 d–1, and 0.013 d–1 forDMP, DBP, and DEHP, respectively. Painter and Jones [73] reported on the use of freshwater and salt marsh inocula (10% v/v) in the anaerobic test system that they used for sludge. The rates of anaerobic degradation calculated from their data are 0.13 d–1 and 0.076 d–1 for DBP and BBP, respectively, with the freshwater sediment. The saltwater sediment showed similar activity,higher for DBP (0.31 d–1) and lower for BBP (0.051 d–1). Madsen et al [76] alsostudied the anaerobic biodegradation in sediment of a number of chemicals,including DMP and DIBP, but they only measured ultimate degradation by usingmethane production. This end point is unlikely to be equivalent to parent chemical degradation, since sulfate- and nitrate-reducing organisms may be present in sediment.

Chauret et al. [77] reported on the biodegradation of DBP in subsurface soilmicrocosms under aerobic and nitrate-, iron-, and sulfate-reducing redox con-ditions. They report a rate of 2.05 nmol g-sediment–1 d–1 in aerobic systems andapparently zero-order rates of 0.86, 0.50, and 0.18 nmol g-sediment–1 d–1 for ni-trate-, iron-, and sulfate-reducing conditions, respectively. It is apparent by com-parison with the aerobic rate, that anaerobic biodegradation via nitrate- and sul-fate-reducing pathways may be appreciable.Wang et al. [78] have reported on theisolation of a nitrate-reducing organism growing on and completely degradingDBP. For reasons of lack of environmental relevance, we omit these rates in lab-oratory systems using isolated pure strains of bacteria.

Ejlertsson et al. [79–81] studied anaerobic degradation in laboratory systemsinoculated with diluted (1%) samples from a digester treating municipal solidwaste. For DEP, after an initial adaptation phase of up to 20 d, degradation pro-ceeded rapidly at a rate of 0.7 mg g–1 d–1 [79]. In experiments terminated at278 days they found complete removal of DBP, 77% removal of BBP, and 19% re-moval of DEHP. Ejlertsson et al. [82] studied the ultimate anaerobic degradation(methane production) of a range of PDEs and of their constituent alcohols. Theyfound that PDE with water solubilities above about 50 µg L–1, such as DBP, DBP,and DHxP, were degraded within the 35–100 d incubation period, while DEHP,DOP, and DDP, with solubilities below about 3 µg L–1 showed no conversion tomethane. The higher alcohols such as 2-ethylhexanol, octanol, and decanol werewell degraded. The authors ascribed the difference in PDE biodegradability tolack of solubility of the higher PDEs, although it may as well be due to a lack ofbioavailability due to their hydrophobicity and strong adsorption to the highsolids content in anaerobic test systems.

As discussed earlier, Fauser et al. [26] developed a model to reflect the con-centration of DEHP in the depth profile of sediments in Roskilde Fjord, Den-

mark. The model takes into account the rates of sedimentation, aerobicbiodegradation, and anaerobic biodegradation. Fitting of the model to experi-mental sediment concentrations gave an aerobic rate constant for degradation of2 ¥ 10–5 s–1 and an anaerobic rate constant of 8 ¥ 10–6 s–1 (below 5 cm.). This lat-ter rate constant converts to a DEHP anaerobic biodegradation rate of 0.69 d–1

(t1/2 = 1.0 d).The reported rates of anaerobic biodegradation of PDEs are summarized in

Table 6. For the less hydrophobic PDEs, the pseudo-first-order rates ranged from

112 D.R. Peterson and C.A. Staples

Table 6. Summary of biodegradation rates of PDEs in anaerobic environments

Phthalate First-order Half-life Test conditions Ref.rate (d–1) (d)

Anaerobic WWTP digester sludge, batch incubationDMP >0.3 <2.3 Undiluted sludge, 20 mg L–1 DMP [71]

0.25 a 2.8 Undiluted sludge, 0.5–10 mg L–1 DMP [72]0.696 1.0 Undiluted sludge, 10 mg L–1 DMP [74]0.42 3.3 10% diluted sludge, 20 mg L–1 DMP [71]

DEP >0.3 <2.3 Undiluted sludge, 20 mg L–1 DEP [71]0.14 a 5.0 Undiluted sludge, 0.5–10 mg L–1 DEP [72]0.069 10.0 10% diluted sludge, 20 mg L–1 DEP [71]

DBP >0.3 <2.3 Undiluted sludge, 20 mg L–1 DBP [71]0.26 a 2.7 Undiluted sludge, 0.5–10 mg L–1 [72]0.581 1.19 Undiluted sludge, 10 mg L–1 DBP [74]0.073 9.5 10% diluted sludge, 20 µg mL–1 DBP [71]0.025 27.7 10% diluted sludge, 0.2 mM DBP [73]

BBP 0.056 12.4 Undiluted sludge, 20 mg L–1 BBP [71]0.19 a 3.7 Undiluted sludge, 0.5–10 mg L–1 BBP [72]0.096 7.2 10% diluted sludge, 20 mg L–1 BBP [71]

DEHP 0.001 693 Undiluted sludge, 20 mg L–1 DEHP [71]Nil – Undiluted sludge, 0.5–10 mg L–1 DEHP [72]Nil – 10% sludge, 20 mg L–1 DEHP [71]Nil – 10% sludge, 0.2 mM DEHP [73]

DOP 0.006 115 Undiluted sludge, 20 mg L–1 DOP [71]Nil – Undiluted sludge, 0.5–10 mg L–1 DOP [72]0.0336 20.6 Undiluted sludge, 10 mg L–1 DO [74]

Anaerobic soil and sedimentDMP 0.033 21.0 Flooded soil, 500 mg kg–1 [44]DBP 0.036 19.3 Flooded soil, 500 mg kg–1 [44]

0.27 2.6 Pond sed.: water (1 :2), 1 mg L–1 DBP [75]0.13 5.3 10% freshwater sed., 0.2 mM DBP [73]0.31 2.2 10% salt marsh sed., 0.2 mM DBP [73]

BBP 0.076 9.1 10% freshwater sed., 0.2 mM DBP [73]0.051 13.6 10% salt marsh sed., 0.2 mM DBP [73]

DEHP 0.013 53.3 Flooded soil, 500 mg kg–1 [44]Nil – Pond sed.: water (1:2), 1 mg L–1 DEHP [75]0.69 1.0 Field data, sed. DEHP concentrations [26]

a Mean of determinations at three PDE concentrations, 0.5, 1, and 10 mg L–1.

Degradation of Phthalate Esters in the Environment 113

about 0.2 to 0.6 d–1 for undiluted anaerobic sludge, corresponding to half-lives ofabout 1-5 days.A 10% sludge inoculum in an anaerobic, aqueous system resultedin somewhat lower rates. Anaerobic sludge biodegradation of BBP was a bitslower, with rates of about 0.06–0.2 d–1, while the more hydrophobic DEHP andDOP were degraded very slowly in anaerobic sewage sludge with reported half-lives of nearly two years (693 d) for DEHP and 21–115 d for DOP.Anaerobic sed-iment studies of DBP and BBP gave biodegradation rates within the same rangesas seen for anaerobic sludge biodegradation even though most of these were car-ried out at 10% inoculum. The flooded soil studies of Shanker et al. [44] gavesomewhat lower rates for DMP (0.033 d–1) and DBP (0.036 d–1) than was seenwith anaerobic sludge, while DEHP showed a higher rate (0.013 d–1) than re-ported for sludge.

4.5Solid-Phase Biodegradation

There have been numerous reports of the growth of microbes on the surface offabricated PVC plastic materials. This susceptibility of PVC to biodeteriorationis principally due to the presence of biodegradable plasticizers in the polymer[83]. Webb et al. [83] isolated and identified a large number of fungi that colo-nized the surface of dioctyl adipate (DOA)-plasticized PVC. The most activestrains were able to utilize the plasticized PVC as a sole source of carbon, pro-duced an extra-cellular esterase, and caused weight loss of the PVC. Other in-vestigators have also isolated large numbers of fungi and yeasts capable of uti-lizing PDEs as a source of carbon [84, 85].

El-Sharouny [86] isolated plasticizer-degrading thermophilic fungi from NileRiver mud.All but four of the 19 fungi isolated could degrade DBP and DOP, thetwo phthalate plasticizers used. He then measured weight loss in plastic stripscontaining 33% plasticizer, in liquid cultures inoculated with each of the six mostactive degraders. The percent weight loss was corrected for simple diffusionalloss with a non-inoculated poisoned control. Losses in 14 days ranged from 2.3 to11.9% for DBP and 2.8–12.5% for DOP.Yabannavar and Bartha [87] studied thebiodegradability of a number of plasticized plastic films in soil microcosms. Oneof these films contained DOP (24.8%) as the plasticizer. The soil was a sandyloam containing 5% organic matter and maintained at 50% moisture capacityand 27 °C. The film was amended to the soil as 2¥3 mm pieces at 1% w/w and thesoil incubated aerobically for three months.Weight determinations were made bySoxhlet extraction of the PVC with methyl ethyl ketone (MEK), and the plasticwas precipitated and weighed. The PVC film containing DOP lost 20.8% of itsweight. This is likely to all be DOP loss because the authors report that no plas-ticizer was detectable by GC in the MEK extracts. On a weight basis, this is 84%biodegradation of the DOP.Although the time course of biodegradation was notdetermined on the DOP-plasticized film, it was determined on dioctyl adipate-plasticized films, which showed similar weight loss in three months. On average,these lost 64% in the first month and then very little in the two subsequentmonths. The authors speculate that this result may be due to a residual fractiontrapped within the film.

Gumargalieva et al. [88] evaluated the kinetics of PDE loss from PVC by usinga culture of Aspergillus niger isolated from PVC insulation of wires. The overallrate of PDE loss depends on both diffusion through the PVC to the surface anddesorption from the surface. They reported that under their conditions, desorp-tion from the surface of the plastic was slower than diffusion. They measuredoverall loss rate of PDE from PVC over 12 months with and without surfacegrowth of the fungus. They found rates of 8.6 ¥ 10–4 d–1 (t1/2 = 800 d) in the pres-ence of the fungus and 1.38 ¥ 10–4 d–1 (t1/2 = 5000 d or 13.8 y) in its absence. Bycomparing these rates with calculated rates for diffusion and desorption in theabsence of the fungus, it was shown that the volatilization of the PDE is rate lim-iting without the fungus. With the fungus, the rate of loss was limited by diffu-sion out of the PVC. The loss mechanism in the presence of the fungus is thusbiodegradation rather than volatilization.

Obviously, such solid-phase biodegradation of PDEs is very slow in compar-ison to their biodegradation rates in soil or water. The point is though, that PDE-containing plastic articles in soil, landfills, and even in the open air are likely tobe surface colonized by phthalate-degrading microorganisms and the PDE de-graded at the surface as rapidly as it diffuses out. This diffusion rate is slow anddependent on the size and shape of the plastic items (Osmon et al. [89]). Thereis also the possibility that in some polymers a fraction of PDE that is trappedwithin the polymer matrix and cannot diffuse out [87]. The implication, however,is that plastics in soil and landfills are not likely to be a source of PDE emissionto water or air because their surfaces may be colonized by slow growing fungi,yeast, and bacteria that can biodegrade the PDE as it diffuses out of the item.

5Degradation of Phthalate Metabolites

The metabolic pathways for PDE biodegradation have been reviewed [1]. Thereis abundant evidence that the pathway of PDE metabolism, at least by microor-ganisms, is via a stepwise hydrolysis of the two ester bonds, first giving the monoester plus the free alcohol, followed by hydrolysis of the second ester bondto give PA and alcohol. The PA is degraded aerobically via hydroxylation and decarboxylation to give protochatechuic acid, which is further degraded to car-bon dioxide through either ortho or meta cleavage of the aromatic ring. PA is degraded anaerobically via decarboxylation to benzoic acid, which is further degraded through ring saturation.

Most of the studies delineating these pathways were conducted using isolatedPDE degrading bacteria [90–93]. In these studies, there were differences betweenthe microorganisms in their further metabolism of the immediate degradationproducts, phthalate monoester (PME), and PA. For instance, Englehardt et al. [92]reported that four isolates from enrichment cultures (Penicillium lilacinum andthree bacteria, two gram positive and one gram negative) together with three or-ganisms from stock cultures (Corynebacterium petrophilum, Arthrobacter hy-drocarboglutamicus, and Mycobacterium phlei) all formed MBP from DBP almostquantitatively. MBP was isolated by thin-layer chromatography (TLC) and noother metabolites were detected. On the other hand, they found that three other

114 D.R. Peterson and C.A. Staples

Degradation of Phthalate Esters in the Environment 115

strains of coryneform bacteria degraded DBP quickly with only a transient ap-pearance of MBP in the medium.All three strains could be grown on MBP or PA.Other workers have reported on individual microorganisms growing on PDEswith only a transient appearance of PME or PA in the medium [92, 93]. Kuraneet al. [30] followed the time course of degradation of DEHP by Pseudomonas aci-dovarans and saw a transient appearance of soluble metabolites. They analyzedthese by TLC and identified the soluble metabolites as PA and protocatechuicacid. Kurane et al. [31] also analyzed the intermediates produced from DEHP hy-drolysis by a purified phthalate ester hydrolase from Nocardia erythropolis. Theyfound only PA and no MEHP and concluded that “the purified enzyme rapidlyconverts phthalate diesters into phthalic acid without phthalate monoesters ac-cumulating.” This implies that the hydrolysis is accomplished by the same en-zyme and that hydrolysis of the monoester is faster than of the diester. However,studies on single strains of microorganisms in growth medium are not entirelyrelevant to the rate of biodegradation of PMEs under environmental conditions.

There are some studies available on mixed inocula or microcosms, wherein thedegradation of PME is reported, including its decrease in concentration overtime. The mineralization of 14C-labeled DBP, MBP, and PA in soil was studied byInman et al. [94]. Their graphs indicate that both MBP and DBP produced 14CO2at about the same rate, while PA produced it much more quickly. These data seemto indicate that metabolism of the phthalic acid through decarboxylation is fasterthan hydrolysis of the ester linkages and that hydrolysis of DBP is as fast or fasterthan hydrolysis of MBP. On the other hand, Shanker et al. [44] only detected par-ent PDE and phthalic acid in garden soil degradation studies of DMP, BBP, andDEHP.At each time interval, for all three PDEs, the PA was always much lower inconcentration than the PDE. In the aerobic studies, the maximum concentrationof PA never exceeded 3% by weight of the original amount of PDE added. In thestudies of the fate of DEHP in soil by Schmitzer et al. [95], radiolabeled DEHP wasemployed and soil residues were extracted and analyzed by TLC. In laboratorystudies, after seven days, 89.8% of the radioactivity was unchanged DEHP andthe other residues were too low in amount to quantify. In outdoor studies, after111 days, 3% of the residual radioactivity was DEHP, 0.14% was MEHP, and0.35% was PA. In laboratory soil microcosms and outdoor lysimeters, Rüdel et al.[43] were able to detect only DEHP and no MEHP or other metabolites by usingradiolabeled DEHP and HPLC separation.All of these studies, with the exceptionthat of Inman et al. [94], indicate that in aerobic soil, degradation of the PME isfaster than degradation of the PDE.

A simple first-order model may be constructed for the hydrolysis of PDE toPME and of PME to PA. If the rates of the two reactions are equal, and assumingirreversibility, the concentration of PME will reach a maximum of about 37% ofthe starting concentration of PDE. This is also the crossover point, after which theconcentration of PME will exceed the concentration of PDE. To achieve a situa-tion in which the concentration of PME is always less than the concentration ofPDE, the rate of PME hydrolysis must be at least twice the rate of PDE hydroly-sis. These situations are illustrated in Fig. 2a and 2b. Of course, if the enzyme sys-tems for intermediate breakdown need to be induced, a lag in their further ca-tabolism may occur. For the Micrococcus strain studied by Eaton and Ribbons

116 D.R. Peterson and C.A. Staples

Fig. 2. a Modeled time course of degradation for a PDE (solid line) and intermediates. Rates:diester – 1.0 d–1, monoester – 1.0 d–1, phthalic acid – 1.0 d–1. b Modeled time course of degra-dation for a PDE (solid line) and intermediates. Rates: diester – 1.0 d–1, monoester – 2.0 d–1,phthalic acid – 1.0 d–1

a

b

[91], the esterase enzyme activities were constitutive but the enzyme catalyzingthe hydroxylation of PA was inducible. In the pure species (Nocardia sp.) work ofEngelhardt and Wallnöfer [93], curves for DBP disappearance and transient con-centrations of MBP appeared as in Fig. 2b, in which the rate of monoester hy-drolysis is twice that of the diester. On the other hand, a similar graph presentedby Karegoudar and Pujar [96] for DEP degradation by a Micrococcus sp. lookedmore like Fig. 2a, in which the two hydrolysis rates are equal.

The initial, first-order rate of PA degradation in sludge-amended soil was de-termined by Roslev et al. [51] as 0.32 d–1, under aerobic conditions. Under anaer-obic conditions, it appears that such a lag in PA degradation does occur. Sheltonet al. [71] presented a graph of micromolar concentrations of BBP, MBuP, and PAversus time during anaerobic incubation in 10% sludge. Only a small amount ofMBuP was found, reaching a maximum on day 3 of about 5% of the initial con-centration of BBP (concentrations are not presented for MBnP). PA did not reacha maximum until day 10, when it was about 80% of initial BBP. Analysis of therate of BBP disappearance (linear regression of the natural logarithm of con-centration versus time) gives a rate of BBP degradation of 0.21 d–1. Curve “strip-ping” and regression analysis of the terminal rate of PA degradation gives a rateof 0.31 d–1. By using these rates in the kinetic model described in the previousparagraph, the maximum reached by the PA concentration would be 32% and oc-curs on day 4. However, if one assumes a 9-day lag before PA degradation begins,the maximum occurs on that day at about 80% of the initial BBP concentration.The data of Shanker et al. [44], on the other hand, show that for DMP, DBP, andDEHP in anaerobic soil systems, the transient appearance of PA never reachedmore than 10% of the amount of PDE present or more than 6% of the initialamount. Similarly, Johnson and Lulves [75] reported only low concentrations ofPA during DBP biodegradation in anaerobic freshwater sediment.

Ejlertsson et al. [79, 80] have observed very similar results to those of Sheltonet al. [71] in diluted (1%) anaerobic municipal solid waste digester (biogas reac-tor). By using DEP in concentration ranges of 50–250 mg L–1, they found pro-portional amounts (70–75%) of PA produced which took considerable time todegrade beyond complete DEP degradation [79]. In addition, they observed a lagof about 20 days in the degradation of DEP. Degradation of PA commenced onabout day 30 and was not complete until day 60–80. Only small amounts of MEP(<15%) were produced and quickly disappeared. However, this same group [80]also reported that in an incubation of DEP with anaerobic solid waste from amodel landfill, only 43% was degraded in 110 days and more MEP was produced(40%) than PA (5.7%). By using this same inoculum, anaerobic degradation ofBBP resulted in 78% degradation to MBuP (29%), MBnP (43%), and PA (14%).By contrast, PA itself was completely degraded in 110 days. Thus, the source of in-oculum appears to influence the relative rates of biodegradation of parent PDE,PME, and PA.

Ejlertsson and Svendsson [81] also studied the anaerobic degradation ofmetabolites of DEHP, namely MEHP, 2-ethylhexanol (2-EH), and 2-ethylhexanoicacid (2-EHA). Again, by using inocula from digestion of solid waste (biogas sys-tem), they found that MEHP degraded quickly with no lag phase and a half-lifeof about 7 days. The PA produced reached a stoichiometric amount and did not

Degradation of Phthalate Esters in the Environment 117

begin to degrade until about day 40. It then degraded rapidly in about 15 days.Rates are roughly 0.1 d–1 for each. 2-EH and 2-EHA both showed lags of 6 and 8 d,respectively, before rapid degradation began. After this lag phase, 2-EHA de-graded in 4 days and in 5 days once degradation began. 2-EH also showed the appearance of a stoichiometric amount of 2-EHA as a transient intermediate,reaching a maximum at day 10 and completely degrading at day 15. The degra-dation of the intermediates, both the monoester and the alcohols are more rapidthan the degradation of the parent DEHA. This may be partially due to thegreater bioavailability of these more polar and more water-soluble metabolites.

Given the difference in hydrophobicity between PMEs and PDEs, especially forthe higher phthalates, models predict that the PMEs will partition much morestrongly to water than to solids in WWTP effluents. Thus, one would expect muchhigher concentrations of PMEs than PDEs in aqueous effluent if PMEs were pre-sent at steady state. In fact, Paxéus [97] has reported on the concentrations ofPMEs and PDEs from three large WWTPs in Sweden. Two of the plants had de-tectable amounts of DEP, DBP, BBP, and DEHP. All concentrations were in therange 3–22 µg L–1. Only two of the corresponding PMEs were detected, MEP andMBuP. At one plant these were present at lower concentrations than the corre-sponding PDEs (0.5 and 4 versus 3 and 11) and at the other at about similar con-centrations (11 and 20 versus 8 and 22). At the third WWTP, the only PDE de-tected in aqueous effluent was DBP (6 µg L–1), while MBuP and MEP weredetected at low levels (0.5 µg L–1, each).Again, these data suggest that much lowerconcentrations of PMEs, as compared to PDEs, appear to be present in activatedsludge during wastewater treatment and that this must be a consequence of amore rapid degradation rate of PME than PDE.

In conclusion, in almost every circumstance in which the kinetics of metabo-lite formation and disappearance has been quantified, the rate of PME degrada-tion is twice that of the corresponding PDE or more.

Routes of degradation of the PMEs in the environment, other than biodegra-dation, are not expected to be significant. As discussed previously, abiotic hy-drolysis of PDEs is quite slow and not expected to be significant. Thus, the for-mation of PMEs will be mainly due to metabolism and will occur within biotaand in aqueous systems. Since the PDEs are carboxylic acids, they will be mainlydissociated at neutral pH and hence are not expected to be volatile. Abiotic hydrolysis rates of the PMEs are at least an order of magnitude slower than hydrolysis of the parent PDE [12].

6Summary and Conclusions

Owing to their major use as components in fabricated plastic articles, it is ex-pected that PDEs will be released to the environment through slow volatilizationor leaching. It is apparent that the diffusion of PDEs out of such articles is slowand also that organisms may colonize surfaces of plastics and metabolize PDEsas fast as they diffuse to the surface. Such articles when exposed to the atmos-phere may also emit PDEs through volatilization or leaching by rainwater. Indi-rect photooxidation in surface films may also play a role in PDE degradation.

118 D.R. Peterson and C.A. Staples

Phthalates emitted in household wastewater may still be contained within smallparticles of abraded plastic. PDEs contained within plastics in landfills are un-likely to migrate due to their hydrophobicity.

In surface water, hydrolysis is not expected to contribute very greatly to theoverall degradation of PDEs (or PMEs), since their hydrolysis half-lives are of theorder of a year or longer. Photolysis on the other hand may be a significant path-way for abiotic degradation. In aqueous systems, PDEs in surface layers may bedegraded at a significant rate (1–3 h–1) due to indirect photolysis. However, thebulk of PDEs in the aquatic environment are expected to be particulate bound.No data are available on the rate of photodegradation in this state but this is alsounlikely to be a major fate of PDEs. For the more volatile, low-molecular weightPEs (DMP, DEP, and DBP) and for all of the PEs when released to air, indirect pho-todegradation via hydroxyl radicals may also be a significant route of abioticdegradation. Photodegradation half-lives, with the exception of DMP, are in therange of a day or less in typical urban air where they are likely to be emitted except for DMP. Although a sizeable percentage of the PEs in air are likely to beassociated with particulates, data for DEHP show that indirect photolysis of par-ticulate sorbed PDE is about the same as in the vapor phase [7]. Thus, it is rea-sonable to use the modeled values shown in Table 1 as the likely half-lives for PDEdegradation in air.

The relatively low vapor pressures of PDEs (and PMEs) are expected to resultin a major portion of releases going to wastewater, surface water, and soil. ThePDEs in surface water are also likely to be mainly associated with suspended par-ticulates and sediments. Thus, the rates of biodegradation of PDEs in WWTPs,surface water, sediment, and soil are expected to be major determinants of theiroverall environmental fate. The rate of aqueous biodegradation of all of the PDEsis in the approximate range of 0.2–2.0 d–1 in test systems that are intended to sim-ulate environmental conditions, including the presence of suspended sediments.There are data on the biodegradation rates in soil for fewer of the PDEs than inwater or sediment. However, soil biodegradation rates appear to be lower thanaquatic ecosystem rates and to be lower for the higher PDEs. The biodegradationrates in soil of DMP and DEP are about 0.4 d–1, while that of DBP is in the rangeof 0.04–0.4 d–1 and DEHP rates are lower, at about 0.01–0.1 d–1. This lower ratefor DEHP biodegradation in soil is likely to be due to its reduced bioavailabilityas a result of adsorption or sequestration. The rates of anaerobic biodegradationdo not seem to vary greatly between sludge, sediment, and soil for the same PDEsand were about 0.6–0.06 d–1 in these systems for DMP, DBP, and BBP. On theother hand, DEHP and DOP show a relatively low rate of anaerobic biodegrad-ability.Again, lower bioavailability is likely to be responsible for these lower rates.A range of 0.01–0.001 d–1 was taken as representative for these higher PDEs.

Regarding the rate of biodegradation in wastewater treatment plants, the de-fault rate given in the EU Technical Guidance Documents [66] of 1 h–1 for readybiodegradable chemicals, seems approximately correct when applied to DEP. Thatis, the use of this rate in the SimpleTreat model results in calculated effluent con-centrations in water and sludge that are in approximate agreement with mea-sured values. However, for the other PDEs, the model overestimates the concen-trations in aqueous effluent and sludge. This is the result of an underestimation

Degradation of Phthalate Esters in the Environment 119

of total biodegradation, since biodegradation of adsorbed chemical is not in-cluded. Running the model in a mode that utilizes the same rate irrespective ofthe degree of sludge adsorption, results in too high an extent of removal at a rateof 1 d–1. A rate of about 0.03 d–1 in this mode gives a better result for DOP andDEHP but overestimates effluent concentrations for DEP.

The recommended ranges for half-lives of phthalates derived in this review are presented in Table 7. The data are grouped by “low”- and “high”-molecularweight PDEs, since the values are quite similar within these two groups and sincethe amount of data does not support separate values for each individual PDE.

Boethling et al. [98] analyzed biodegradability data for a number of chemicalsand concluded that degradation rates are generally about as fast in surface soilas in freshwater. They also observed that half-lives of chemicals were 2–4 timeslonger under anaerobic conditions than aerobic. The values in Table 7 are in ap-

120 D.R. Peterson and C.A. Staples

Table 7. Recommended environmental degradation rates of PDEs

Environmental Route Phthalate Rate (d–1) Half-life (d) Half-life (h)compartment

Surface water biodegradation low MW a 0.2–2.0 3.5–0.35 84–8.4& sediment (aerobic)

biodegradation high MW b 0.2–2.0 3.5–0.35 84–8.4(aerobic)biodegradation PMEs 0.4–4.0 1.7–0.17 42–4.2(aerobic)hydrolysis all <7 ¥10–4 >103 >2 ¥104

photolysis all ? ? ?(indirect)

Soil biodegradation low MW 0.1–0.4 6.9–1.7 166–41(aerobic)biodegradation high MW 0.01–0.1 69–7 1663–166(aerobic)

Soil & sediment biodegradation low MW 0.06–0.6 11.6–1.2 278–28(anaerobic)biodegradation high MW 0.006-.01 116–69 2784–1656(anaerobic)biodegradation PMEs 0.1–1.0 6.9–0.7 166–17(anaerobic)

Wastewater water phase low MW 24 0.029 0.7slurry phase high MW 0.75 0.92 22

Air photolysis DMP 0.048 14.4 346(indirect)photolysis DBP 0.29 2.4 58(indirect)photolysis low MW 0.8–1 0.9–0.7 22–17(indirect)photolysis high MW 3.5–1.2 0.6–0.3 14.4–7.2(indirect)

a The low-MW PDEs are considered those with C1 to C4 alcohol side-chains (DMP, DEP, DBP)plus BBP.

b The high-MW PDEs are those with C6 alcohol side-chains and above.

Degradation of Phthalate Esters in the Environment 121

proximate agreement with these conclusions; however, soil biodegradation half-lives are about 2–4 times longer than aquatic for the lower PDEs and 20-foldlonger for the higher molecular weight PDEs. This lower soil biodegradability forthe higher PDEs is most likely due to their sorption to soil organic matter, ren-dering them unavailable for biodegradation. Soil anaerobic biodegradation isabout the same as in aerobic soil for the lower PDEs but has 2–5 times longerhalf-lives for the higher PDEs.

The differences in measured soil and sediment half-lives between the lowerand higher PDEs do not agree with the EU Technical Guidance Documents(TGD) values, which are based, in part, on soil sorption coefficient (Kp) [66].For “readily” biodegradable chemicals with a Kp of <100 L kg–1, the TGD spec-ifies a half-life in soil of 30 days. The TGDs also specify an increase of one order of magnitude in half-life for each order of magnitude increase in Kp.Thus, for a chemical with a Kp of between 100,000–1,000,000 L kg–1, the spec-ified half-life is 300,000 d or 822 y. DMP has a Kp of <100 L kg–1 while that ofDEHP is between about 100,000 and 1,000,000. This assumed low rate is clearlynot correct, based on the measured rates presented in this chapter. The TGD specified increase in half-life with increase in Kp is apparently based on the assumption that adsorbed chemicals are not degraded. The data and discussionson bioavailability and biodegradation of sorbed PDEs clearly contradict thispremise.

The ranges for rates and half-lives given in Table 7 account approximately forthe variation in results from laboratory tests. In the environment the actual ex-pected range of variation in these rates is likely to be much larger due to the largerange of types and numbers of bacteria in different locations and variation inconditions such as temperature and pH. However, to use such fate data in mod-eling one must assume a standard condition of the model that is typical or rep-resentative. The studies that were considered in this review were environmentallyrelevant, generally using natural water or inocula from areas likely to be impactedby PDE emissions.

The degradation half-lives used in the Modeling Chapter are longer than thoserecommended here. These half-lives were based on the upper end of the range ofdegradation rates reviewed by Staples et al. [1]. The use of these half-lives in mod-eling is expected to result in conservative estimates of the persistence and con-centrations of PDEs in the environment.

7References

1. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:6672. Atkinson R (1988) Environ Toxicol Chem 7:4353. Meylan WM, Howard PH (1993) Chemosphere 26:2293; AOPWIN version 1.89 is part of

EPIWIN, a QSPR computer program available from Syracuse Research Corporation, NorthSyracuse, NY

4. Atkinson R (2000) Atmospheric oxidation. In: Boethling RS, Mackay D (eds) Handbook ofproperty estimation methods for chemicals, environmental and health sciences. Lewis,Boca Raton, FL, p 335, chap 14

5. Ligocki MP, Pankow JF (1988) Environ Sci Technol 23:75

6. Behnke W, Nolting F, Zetzsch C (1987) An aerosol smog chamber for testing abiotic degra-dation of compounds with low volatility. In: Greenlegh R, Roberts TR (eds) Pestic SciBiotechnol, Proc Int Congr Pestic Chem Blackwell, Oxford, UK, p 401

7. Jin Z, Huang G, Chai Y, Zhong Y, Wang D, Li H (1999) Huanjing Huaxue 18:1098. Södergren A (1982) Environ Poll A27:2639. Howard P (1991) Handbook of environmental degradation rates. Lewis, Chelsea, MI

10. Gledhill WE, Kaley RG, Adams WJ, Hicks O, Michael PR, Saeger VW, LeBlanc GA (1980) Environ Sci Technol 14:301

11. Mill T (2000) Photoreactions in surface waters. In: Boethling RS, Mackay D (eds) Handbookof property estimation methods for chemicals, environmental and health sciences. Lewis,Boca Raton, FL, p 355, chap 15

12. Wolfe NL, Steen WC, Burns LA (1980) Chemosphere 9 :40313. Wolfe NL, Jeffers PM (2000) Hydrolysis. In: Boethling RS, Mackay D (eds) Handbook of

property estimation methods for chemicals, environmental and health sciences. Lewis,Boca Raton, FL, p 311, chap 13

14. Scholz N, Diefenbach R, Rademacher I, Linnemann D (1997) Bull Environ Contam Toxicol58:527

15. Painter HA (1995) Detailed review paper on biodegradability testing. OECD guidelines forthe testing of chemicals. Environment monograph No 98. Organisation for Economic Co-operation and Development, Paris

16. Blok J, Balk F (1994) Guidance document for the interpretation of biodegradability testdata. Report to the European Communities, Contract No B-3040/93/001114. BKH Con-sulting Engineers, Delft, The Netherlands

17. Aronson D, Howard PH (1999) Evaluating potential POB/PBT compounds for en-vironmental persistence. Report No SRC-TR-99–020. Syracuse Research Corp. North Syracuse NY.

18. Boethling B (2000) HPVC-screening tool: using ready and inherent biodegradability datato derive input data for the EQC model. In: environment Canada, environmental catego-rization for persistence bioaccumulation and inherent toxicity of substances on the Do-mestic Substance List using QSARs, results of an international workshop hosted by Chem-icals Evaluation Division of Environment Canada in Philadelphia, PA 11–12 Nov 1999.Final Report, Appendix 10. Environment Canada

19. Saeger VW, Tucker ES (1976) Appl Environ Microbiol 31:2920. Carson DB, Saeger VW, Gledhill WE (1990) Use of microcosms versus conventional

biodegradation testing for estimating chemical persistence. In: Landis WG, van der SchalieWH (eds) Aquatic toxicology and risk assessment: thirteenth volume. ASTM STP 1096.American Society for testing and Materials, Philadelphia PA, p 48

21. Adams WJ, Saeger VW (1993) Utility of microcosms for predicting the environmental fateof chemical. In: Gorsuch JW, Dwyer FJ, Ingersol DG, La Pointe TW (eds) Environmental tox-icology and risk assessment. American Society for testing and Materials, Philadelphia PA,p 48

22. Walker WW, Cripe CR, Pritchard PH, Bourquin AW (1984) Chemosphere 13:128323. Furtmann K (1993) Phthalate in der aquatischen Umwelt. PhD Thesis, Universität

Gesamthochschule Duisenberg. English Translation prepared for European Council forPlasticizers and Intermediates, Brussels, 1996.

24. Furtmann K (1995) Anal Methods Instrum 2:25425. Ye C, Tian K (1990) Water Treat 5 :47426. Fauser P, Sørensen PB, Vikelsøe J, Carlsen L (2000) Fate of di-2-ethylhexyl) phthalate

(DEHP) in Roskilde Fjord. Poster presented at the 20th international symposium on halo-genated environmental pollutants & POPs, Dioxin 2000, Monterey CA, August 2000

27. Lewis DL, Kellogg RB, Holm HW (1985) Comparison of microbial transformation rate coefficients of xenobiotic chemicals between field-collected and laboratory microbiota. In:Boyle TP (ed) Validation and predictability of laboratory methods for assessing the fate andeffects of contaminants in aquatic ecosystems. American Society for Testing and Materi-als, Philadelphia PA, p 3

122 D.R. Peterson and C.A. Staples

Degradation of Phthalate Esters in the Environment 123

28. Skow KM (1982) Rate of biodegradation. In: Lyman WJ, Reehl, WF, Rosenblatt DH (eds)Handbook of chemical property estimation methods. McGraw-Hill, New York, chap 9

29. Steen WC, Paris DF, Baughman GL (1980) Effects of sediment sorption on microbial degra-dation of toxic substances. In: Baker RA (ed) Contaminants in sediments, vol 1 : Fate andtransport, case studies, modeling toxicology. Ann Arbor Science, Ann Arbor MI, p 477

30. Kurane R, Suzuki T, Takahara Y (1977) Agric Biol Chem 41:211931. Kurane R, Suzuki T, Fukuolka S (1984) Appl Microbiol Biotechnol 20:37832. Nyholm N (1990) Chemosphere 21:147733. Gibbons JA, Alexander M (1989) Environ Toxicol Chem 8:28334. Aichinger G, Grady CPL Jr, Tabak HH (1992) Water Environ Res 64:89035. Wang J, Liu P, Quan Y (1995) Water Treat 10:33136. Wang J, Liu P, Shi H, Qian Y (1998) Chemosphere 37:25737. Wang X, Grady CPL Jr (1995) Water Environ Res 67:86338. Yan H, Ye C, Yin C (1995) Environ Toxicol Chem 14:93139. Rubin HE, Subba-rao RV, Alexander M (1982) Appl Environ Microbiol 43:113340. Pagga U (1987) Wasser Abwasser Forsch 20:10141. Russell DJ, McDuffie B, Fineberg S (1985) J Environ Health A20:92742. Kirchmann H, Aastroem H, Joensaell G (1991) Swed J Agric Res 21:10743. Rüdel H, Schmidt S, Kördel W, Klein W (1993) Sci Total Environ 132:18144. Shanker R, Ramakrishna C, Seth PK (1985) Environ Poll A39:145. Chen Y, Shen D, Hu Z, Liu X,Wu D, Zhao D, Zhang J (1997) Huanjing Kexue Xuebao 17:34046. Wang J, Liu P, Shi H, Qian Y (1997) Chemosphere 35:174747. Cartwright CD, Thompson IP, Burns RG (2000) Environ Toxicol Chem 19:125348. Jensen J, van Langevelde J, Pritzl G, Krogh PH (2001) Environ Toxicol Chem 20:108549. Fairbanks BC, O’Connor GA, Smith SE (1985) J Environ Qual 14:47950. Dörfler U, Haala R, Matthies M, Scheunert I (1996) Ecotoxicol Environ Saf 34:21651. Roslev P, Madsen PL, Thyme JB, Henriksen K (1998) Appl Environ Microbiol 64:471152. Madsen PL, Thyme JB, Moldrup P, Roslev P (1999) Environ Sci Technol 33:260153. Hatzinger PB, Alexander M (1997) Environ Toxicol Chem 16:221554. Cornelissen G, van Noort PCM, Govers HAJ (1997) Environ Toxicol Chem 16:135155. Pignatello JJ, Xing B (1996) Environ Sci Technol 30:156. Northcott GL, Jones KC (2001) Environ Sci Technol 35:110357. Wang J, Liu P, Shi H, Qian Y (1997) Process Biochem (Oxford) 32:56758. Petrasek AC, Kugelman IJ,Austern BM, Pressley TA,Winslow LA,Wise RH (1983) J Wat Poll

Cont Fed 55:128659. Tokuz RY (1991) Hazard Waste Hazard Matter 8 :24560. Paxéus N, Robinson P, Balmer P (1992) Water Sci Technol 25:24961. Zurmühl T (1990) Analyst 115:117162. Merkel D, Appuhn H (1996) Korrespondenz Abwasser 43:57863. Mikkelsen J, Nyholm N, Jacobsen BN, Fredenslund FC (1996) Water Sci Technol 33:

27964. Clark B, Henry GLH, Mackay D (1995) Environ Sci Technol 29:148865. Struijs J, Stoltenkamp J, van de Meent D (1991) Water Res 25:891 (SimpleTreat Ver 3.0 from

Struijs J, 1997)66. European Commission (1996) Technical guidance documents in support of the Commis-

sion Directive 93/67/EEC on the risk assessment of new notified substances and the Com-mission Regulation (EC) 1488/94 on risk assessment of existing substances. EuropeanChemicals Bureau, Ispra, Italy

67. Cousins I, Mackay D (2000) Chemosphere 41:138968. Wang J, Liu P, Quan Y (1997) Environ Int 23:77569. Tienpont B, David F, Vanwalleghem F, Sandra P (1995) J Chromatogr A911:23570. Fauser P, Sørensen PB, Carlsen L, Vikelsøe J (2001) Phthalates, nonophenol and LAS in

Roskilde wastewater treatment plant. NERI technical report no 354. National Environ-mental Research Institute, Denmark

71. Shelton DR, Boyd SA, Tiedje JM (1984) Environ Sci Technol 18:93

72. Zigou K, Kirk PWW, Lester JN (1989) Water Res 23:74373. Painter SE, Jones WJ (1990) Environ Technol 11:101574. Wang J, Chen L, Shi H, Qian Y (2000) Chemosphere 41:124575. Johnson BT, Lulves W (1975) J Fish Res Board Canada 32:33376. Madsen T, Rasmussen HB, Nilsson L (1995) Chemosphere 31:424377. Chauret C, Inniss WE, Mayfield CI (1996) Ground Water 35:79178. Wang J, Liu P, Qian Y (1999) Process Biochem (Oxford) 34:74579. Ejlertsson J, Houwen FP, Svensson BH (1996) Swed J Agric Res 26:5380. Ejlertsson J, Meyerson U, Svensson BH (1996) Biodegradation 7 :34581. Ejlertsson J, Svensson BH (1996) Biodegradation 7 :50182. Ejlertsson J, Alnervik M, Jonsson S, Svensson BH (1997) Environ Sci Technol 31:276183. Webb JS, Nixon M, Eastwood IM, Greenhalgh M, Robson GD, Handley PS (2000) Appl

Environ Microbiol 66:319484. Jackson MA, Labeda DP, Becker LA (1996) J Ind Microbiol 16:30185. Peciulyte D (1997) Biologija 2 :2986. El-Sharouny HMM (1993) Sohag Pure Appl Sci Bull 9 :11787. Yabannavar A, Bartha R (1993) Soil Biol Biochem 25:146988. Gumargalieva KZ, Zaikov GG, Semenov SA, Zhdanova OA (1999) Polym Degrad Stab

63:11189. Osmon JL, Klausmeier RE, Jamison EI (1972) Rate-limiting factors in biodeterioration of

plastics. In: Biodegradation of materials, vol 2. Applied Science, London, p 6690. Keser P, Pujar BG, Eaton RW, Ribbons DW (1976) Environ Health Persp 18:15691. Eaton RW, Ribbons DW (1982) J Bacteriol 151:4892. Engelhardt G, Wallnofer PR, Hutzinger O (1975) Bull Environ Contam Toxicol 13:34293. Engelhardt G, Wallnofer PR (1978) Appl Environ Microbiol 35:24394. Inman JC, Strachan SD, Sommers LE, Nelson DW (1984) J Environ Sci Health B19:24595. Schmitzer JL, Scheunert I, Korte F (1988) J Agric Food Chem 36:21096. Karegoudar TB, Pujar BG (1984) Current Microbiol 11:32197. Paxéus N (1996) Water Res 30:111598. Boethling RS, Howard PH, Beauman JA, Larosche ME (1995) Chemosphere 30:741

124 D.R. Peterson and C.A. Staples

© Springer-Verlag Berlin Heidelberg 2003

Observed Concentrations in the Environment

Kathryn Clark 1 · Ian T. Cousins 2 · Donald Mackay 2 · Kentaro Yamada 3

1 BEC Technologies Inc., 61 Catherine Avenue, Aurora, Ontario, L4G 1K6, CanadaE-mail: kclark@pathcom.com

2 Canadian Environmental Modelling Centre, Environmental and Resource Studies,Trent University, Peterborough, Ontario, K9J 7B8, Canada

3 CG Ester Corporation, Landic Bldg. 8F, 2-16-13, Nihonbashi, Chuo-ku,Tokyo 103-0027, Japan

Measured concentrations of six phthalate esters in seven environmental media are compiledand analyzed. The data are predominantly from Europe, the United States, and Japan. The sixphthalate esters are dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate(DBP), butylbenzyl phthalate (BBP), bis(2-ethylhexyl) phthalate (DEHP), and di-n-octyl phthalate (DnOP). The media addressed are water, sediment, soil, air, dust, food, wastewater,sewage sludges, and rainwater.

The reported concentrations vary widely as a result of several factors including analyticalerror, sample contamination, and proximity to a variety of past and present sources. To gainan impression of the absolute levels and distributions, histograms are prepared with binningon a semi-decade logarithmic scale. Cumulative histograms are also prepared to convey an im-pression of cumulative distribution. To gain an appreciation of the relative concentrations invarious media, fugacities are estimated and plotted, thus revealing the relative equilibrium sta-tus between media and any biomagnification. These plots suggest that phthalate esters are notpersistent in the environment and do not biomagnify, as they are rapidly metabolized in or-ganisms.

Keywords. Phthalate ester, Concentrations, Fugacity, Persistence

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

2.1 Database Generation . . . . . . . . . . . . . . . . . . . . . . . . . 1282.2 Histograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1292.3 Transforming Concentrations into Fugacities . . . . . . . . . . . . 130

3 Dimethyl Phthalate (DMP) . . . . . . . . . . . . . . . . . . . . . . 132

3.1 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1323.1.1 Surface Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1323.1.2 Groundwater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1323.1.3 Drinking Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363.2 Sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363.3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363.4 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363.5 Dust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 125–177DOI 10.1007/b11465

3.6 Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363.7 Other Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1373.7.1 Wastewater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1373.7.2 Sludge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

4 Diethyl Phthalate (DEP) . . . . . . . . . . . . . . . . . . . . . . . 137

4.1 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1374.1.1 Surface Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1374.1.2 Groundwater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1374.1.3 Drinking Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.2 Sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.4 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.5 Dust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.6 Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1424.7 Other Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1424.7.1 Wastewater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1424.7.2 Sludge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

5 Dibutyl Phthalate (DBP) . . . . . . . . . . . . . . . . . . . . . . . 142

5.1 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1425.1.1 Surface Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1425.1.2 Groundwater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1475.1.3 Drinking Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1475.2 Sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1475.3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1475.4 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1485.5 Dust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1485.6 Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1485.7 Other Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1485.7.1 Wastewater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1485.7.2 Sludge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

6 Butylbenzyl Phthalate (BBP) . . . . . . . . . . . . . . . . . . . . . 149

6.1 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1496.1.1 Surface Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1496.1.2 Groundwater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1496.1.3 Drinking Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1496.2 Sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536.3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536.4 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536.5 Dust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536.6 Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536.7 Other Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1546.7.1 Wastewater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1546.7.2 Sludge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

126 K. Clark et al.

7 Bis(2-Ethylhexyl) Phthalate (DEHP) . . . . . . . . . . . . . . . . . 1547.1 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1547.1.1 Surface Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1547.1.2 Groundwater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1597.1.3 Drinking Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1597.2 Sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1597.3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1597.4 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1607.5 Dust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1607.6 Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1607.7 Other Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1607.7.1 Wastewater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1607.7.2 Sludge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1617.7.3 Rainwater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

8 Di-n-Octyl Phthalate (DnOP) . . . . . . . . . . . . . . . . . . . . 1618.1 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1618.1.1 Surface Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1618.1.2 Drinking Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1618.2 Sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1648.3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1648.4 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1648.5 Dust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1648.6 Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1648.7 Other Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1648.7.1 Sludge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1648.7.2 Wastewater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

9 Discussion and Recommendations . . . . . . . . . . . . . . . . . 1659.1 Histograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1659.2 Fugacities in Various Media . . . . . . . . . . . . . . . . . . . . . 1729.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

Abbreviations

BBP Butylbenzyl phthalateC Concentration (mol m–3)DBP Dibutyl phthalateDEHP Bis(2-ethylhexyl) phthalateDEP Diethyl phthalateDMP Dimethyl phthalateDnOP Di-n-octyl phthalatef Fugacity (Pa)fOM Organic matter content of the aerosolfD Fraction of chemical in the dissolved phase of water

Observed Concentrations in the Environment 127

fOC Fraction of organic carbon in the suspended sedimentH Henry’s law constant (Pa m3 mol–1)K KelvinKOW Octanol-water partition coefficientKOA Octanol-air partition coefficientKOC Organic carbon-water partition coefficientKP Gas-particle partition coefficient (m3 µg–1)M Molar mass (g mol–1)R Gas constant (Pa m3 mol–1 K–1)T Absolute temperature (K)TSP Total suspended particle concentration (µg m–3)vSP Volume fraction of suspended particles in waterZ Fugacity capacity of the phase (mol m–3 Pa–1)r Density of the phase (kg m–3)j Fraction of chemical associated with particles in air

1Introduction

Measured concentrations of six phthalate esters in seven environmental media are compiled and analyzed. The six phthalate esters are dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP), butylbenzylphthalate (BBP), bis(2-ethylhexyl) phthalate (DEHP), and di-n-octyl phthalate(DnOP). The media addressed are water, sediment, soil, air, dust, food, and “othermedia” including wastewater, sewage sludges, and rainwater. The data are pre-dominantly from Europe, the United States, and Japan. The monitoring data havebeen collected for a variety of reasons and by different groups (e.g., by regula-tors to support development of regulations, by industry for compliance purposes,by researchers to support modeling efforts, etc.). Due to the varying interests ofthe organizations that have collected the data, there is variation in the proximityof the measurements to sources of the phthalate esters. Where possible, data lo-cations are classified as rural or urban to assist in evaluation of the data.

2Methodology

2.1Database Generation

Numerous data sources have been reviewed to determine the ranges and distri-butions of phthalate esters in the environment. A compilation of data was pro-duced by Exxon Mobil Biomedical Sciences Inc. (EMBSI) [1] for five phthalate esters: DEP, DMP, DBP, BBP, and DEHP. The data were categorized by EMBSI intoregions including Canada, United States, Europe, and Japan/Asia. The data werealso categorized by EMBSI in terms of data quality. The following ranking schemewas employed:

128 K. Clark et al.

1. Reliable without restrictions2. Reliable with restrictions3. Not reliable4. Unassignable

The EMBSI database is used as the starting point for the database presentedherein. Additional data, available as of August 31, 2000, have been added to thisdatabase. These data include recent measurements of phthalates in air, dust, fish,milk, and vegetation in the Netherlands [2, 3], phthalates in surface water andwastewater in Germany [4], DEHP and DBP in milk, breast milk, baby food, anddust [5], DBP in surface water and wastewater in Europe [6], DEHP in surface wa-ter, sediments, and sludge in Europe [7], phthalates in Canadian drinking waterand surface water [8, 9] and sludge [10, 11], and data from Environment Canadaand Health Canada (EC&HC) [12–15]. In addition to the five phthalate esters inthe original EMBSI database, some information on DnOP concentrations in theenvironment is available in the Canadian Environmental Protection Act (CEPA)Priority Substances List assessment report [16, 17] as well as from Alberta Envi-ronment [8, 9].

Measured concentrations, particularly for phthalates other than DEHP, are of-ten reported as “not detected”. As a result, treatment of the detection limit sig-nificantly influences the characteristics (i.e., mean and standard deviation) of thedata. For the results presented herein, the non-detectable data have been set toone-half the detection limit. Tables and histograms that summarize the data arepresented herein. The complete database, with references, is presented in a reportto the American Chemistry Council [18].

2.2Histograms

It is difficult to visualize the distribution of the monitoring data by examining theraw data contained in the monitoring database or by examining the summary tables and statistics. Therefore, it was decided to display the monitoring datagraphically using both frequency histograms and cumulative distribution plotsderived from these histograms. These are termed “cumulative histograms”.

Environmental concentrations in each medium are divided logarithmicallyinto classes or bins, with a factor of approximately three between adjacent bins.For example, for surface water concentrations, the following bins are allocated:

Bin Range (µg L–1) Midpoint of range (µg L–1)

1 0.01–0.03 0.0172 0.03–0.1 0.0553 0.1–0.3 0.174 0.3–1.0 0.555 1.0–3.0 1.76 3.0–10 5.57 10–30 178 30–100 55

Observed Concentrations in the Environment 129

The bins allocated can be adjusted to fit the range of environmental concentra-tions in the database.

The frequency histograms are devised by recording the number of study av-erages in the monitoring database that fall within each bin.A weighted frequencyhistogram is devised by adding up the number of samples that contribute to eachstudy average within a particular bin. For example, if there are three study aver-ages, within a bin ranging from 0.1 to 0.3 µg L–1, of 0.12 µg L–1 (n=10), 0.20 µg L–1

(n=1), and 0.25 µg L–1 (n=5), then this bin will have a frequency of 3 in the unweighted histogram and a frequency of 16 in the weighted histogram.

The cumulative histograms are devised by calculating the cumulative per-centage contribution of each successive bin to the total number of study averagesin the case of the unweighted cumulative histogram, or to the total number ofsamples in the case of the weighted cumulative histogram.

Weighted and unweighted frequency and cumulative histograms are plottedfor DEHP, surface water, sediment, and air concentrations in Europe and are dis-cussed in Section 9.0. Insufficient data are available to produce a histogram forconcentrations in soil.

2.3Transforming Concentrations into Fugacities

The study of both aquatic and terrestrial ecosystems has shown that one usefulapproach for studying food chain bioaccumulation is through transforming en-vironmental concentrations into fugacities [19]. Fugacities offer the advantage ofbeing able to use a single currency to compare levels of contamination in differ-ent environmental media and organisms.

The fugacity f (Pa) of a compound in a particular phase can be calculated fromthe concentration C (mol m–3) by using the following equation:

f = C/Z (1)

where Z is the fugacity capacity of the phase for the compound (mol m–3 Pa–1).Hence, if the fugacity capacities are known, then the fugacities can be calculatedfrom measured concentrations on a volume basis. It is possible to derive con-centrations in units of mol m–3 by using the molecular mass M (g mol–1) and den-sity of the phase r (kg m–3). The fugacity capacities or Z values for the phases canbe estimated by using methods outlined by Mackay [20]. In this report, concen-trations of phthalates in Europe are transformed into their corresponding fu-gacities for the media of air, surface water, sewage sludge, vegetation, soil, sedi-ments, fish, cow’s milk, and human milk.We use only concentration data from therecent RIVM/ECPI monitoring campaign [2, 3] for calculating fugacities, exceptfor surface water concentrations [4] and human milk concentrations [5, 21]. Thishigh quality subset of multimedia concentrations is selected to minimize datavariability. Unfortunately, it is only possible to calculate the fugacities of DEHPand DBP with this limited data set. The equations used to estimate the Z valuesfor these phases are briefly described below. Physical-chemical data used in thecalculations such as KOW , H, and KOA are taken from Cousins and Mackay [22].Environmental parameters used in the calculations are listed in Table 1.

130 K. Clark et al.

The fugacity capacity of pure air ZA is given by

ZA = 1/(RT) (2)

where R is the gas constant (Pa m3 mol–1 K–1) and T is the absolute temperature(K).Air concentrations given in the database are total air concentrations (the sumof the amount in gaseous and particle phases). To calculate the fraction on theparticles (j) and from this the gaseous phase concentration ((1–j) multiplied bythe total air concentration), the following equation is used:

j = KP (TSP)/[1+KP (TSP)] (3)

where KP is the gas-particle partition coefficient (m3 µg–1) and TSP is the totalsuspended particle concentration (µg m–3). KP is estimated from

KP = 1.23 ¥ 10–12 fOM KOA (4)

where fOM is the organic matter content of the aerosol and KOA is the octanol-airpartition coefficient.

The fugacity capacity of pure water ZW is given by

ZW = 1/H (5)

where H is the Henry’s law constant (Pa m3 mol–1). Surface water concentrationsgiven in the database are total water concentrations (the sum of the amount indissolved and suspended particle phases). To calculate the dissolved water con-centration the following equation is used:

fD = 1/(1 + fOC nSP KOC) (6)

where fD is the fraction in the dissolved phase, fOC is the fraction of organic car-bon in the suspended sediment, vSP is the volume fraction of suspended particles,and KOC is the organic carbon/water partition coefficient; it is assumed that KOCis equal to KOW.

Observed Concentrations in the Environment 131

Table 1. Environmental parameters assumed for calculating media-specific fugacities

Parameter Value

Total suspended particle concentration in air (µg m–3) 80Fraction of organic matter in air particles 0.20Volume fraction of suspended particles in water 0.000015Fraction of soil organic carbon 0.02Fraction of sediment organic carbon 0.05Gas constant (Pa m3 mol–1 K–1) 8.314Environmental temperature (K) 273Temperature of cow’s milk (K) 310Temperature of human milk (K) 310Fraction of plant lipid 0.01Fraction of lipid in fish 0.05Fraction of lipid in cow’s milk 0.033Fraction of lipid in human milk 0.04

Concentrations in soil, sediments, cow’s milk, human milk, fish, and sewagesludge are all reported on a mass per lipid or mass per organic carbon basis.These concentrations are converted to units of mol m–3 (assuming a density of1000 kg m–3 for all phases). Concentrations in plants are not given on a lipid ba-sis, thus they are first converted to units on a mass per lipid basis by assumingthe volume fraction of lipid in the plant is 0.01. The fugacity capacity of thelipid/organic carbon phase (ZO) in each medium is calculated as

ZO = KOW ZW (7)

This assumes that KOW is equivalent to the lipid-water and organic carbon-wa-ter partition coefficients. The fugacities of the phthalates in each medium arethen calculated by using Eq. (1).

3Dimethyl Phthalate (DMP)

A summary of the reported concentrations of DMP is presented in Table 2.

3.1Water

3.1.1Surface Water

The overall mean concentration of DMP in surface water in Canada (1.40 µg L–1)is significantly higher than that calculated for the United States (0.0017 µg L–1)and Europe (0.034 µg L–1). As indicated in Table 2, the concentrations measuredin rural and urban regions in Canada do not differ significantly. The maximummeasured concentration in the United States surface water (0.003 µg L–1) is muchless than the maximum measured concentration in Canada (33 µg L–1). The maximum concentration for water with substantial industrial sources in Canada(9 µg L–1) is slightly lower than the overall maximum concentration; however, themean concentration for water with substantial industrial sources is 2.76 µg L–1,which significantly affects the overall mean concentration. No data are reportedfor Japan/Asia.

3.1.2Groundwater

A maximum concentration of 355 µg L–1 is reported for the United States as a mean value; however, the median (27 µg L–1) is much lower than the mean.Only one reference is available for Europe, which indicates concentrations lessthan 0.1 µg L–1. Canadian data are available in a drinking water database re-ported below. Although some of these data represent Canadian groundwater,the data does not differentiate between surface water supplies and groundwatersupplies.

132 K. Clark et al.

Observed Concentrations in the Environment 133Ta

ble

2.Su

mm

ary

ofco

ncen

trat

ions

for

dim

ethy

l pht

hala

te

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Surf

ace

wat

er (µ

g L–1

)C

anad

a1.

40.

0533

0.15

0.35

1.2

1962

data

pri

mar

ily fr

om A

lber

ta E

nvir

onm

ent [

9]ru

ral

0.8

0.05

33N

Ac

0.5

NA

1026

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itur

ban

0.86

0.05

13N

A0.

5N

A18

4m

edia

n=

0.5;

repr

esen

ts 1

/2th

e de

tect

ion

limit

indu

stri

al s

ourc

es2.

760.

19

NA

2.6

NA

625

USA

0.00

170.

001

0.00

30.

0013

0.00

170.

002

509

Euro

pe0.

034

<0.

001

20.

009

0.00

90.

055

138

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Gro

undw

ater

(µg

L–1)

Can

ada

NA

NA

NA

NA

NA

NA

Can

adia

n da

ta r

epre

sent

ed in

dri

nkin

g w

ater

sum

mar

yU

SA27

NA

NA

NA

NA

NA

2Eu

rope

NA

<0.

1<

0.1

NA

NA

NA

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Dri

nkin

g w

ater

(µg

L–1)

Can

ada

0.5

0.5

1N

A0.

5N

A12

73da

ta p

rim

arily

from

Alb

erta

Env

iron

men

t [8]

rura

l0.

50.

51

NA

0.5

NA

637

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itur

ban

0.5

0.5

1N

A0.

5N

A63

6M

edia

n=

0.5;

repr

esen

ts 1

/2th

e de

tect

ion

limit

USA

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esEu

rope

NA

<0.

1<

0.1

NA

NA

NA

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Sedi

men

ts (m

g kg

–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esU

SA3.

60.

004

590

0.17

1.6

11.8

1651

Euro

pe0.

011

0.00

010.

20.

0035

0.00

350.

045

51Ja

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

134 K. Clark et al.

Tabl

e2

(con

tinu

ed)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Soil

(µg

kg–1

)C

anad

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

USA

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esEu

rope

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esJa

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Air

(ng

m–3

)C

anad

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

USA

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esEu

rope

outd

oor

2.55

<1

230.

51

530

indo

or20

.2<

112

92.

110

3926

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Dus

t (in

door

s) (µ

g kg

–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esU

SAN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Euro

pe17

30<

200

5000

340

500

5000

7de

tect

ed in

2of

7sa

mpl

esJa

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Oth

er m

edia

Was

tew

ater

(µg

L–1)

incl

udes

sew

age,

slur

ry,e

fflu

ent,

and

influ

ent

Can

ada

2.76

0.1

9N

A2.

6N

A62

5U

SA13

.713

.720

7N

AN

AN

A1

Euro

pe1.

5<

0.02

130.

030.

142.

176

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Observed Concentrations in the Environment 135Ta

ble

2(c

onti

nued

)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Slud

ge (µ

g kg

–1)

prim

ary

sew

age

slud

geC

anad

a30

<10

130

NA

NA

NA

72U

SA39

,800

NA

NA

NA

NA

NA

2Eu

rope

863

2019

,500

NA

NA

NA

38Ja

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Food

(µg

g–1)

Bev

erag

esN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Cer

eal

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esD

airy

N

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

(exc

l.m

ilk)

Eggs

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esFa

t & o

ilsN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Fish

0.00

5N

AN

AN

A0.

005

NA

25m

ean

repr

esen

ts 1

/2th

e de

tect

ion

limit

Frui

tsN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Gra

ins

0.00

30.

002

0.00

30.

002

0.00

30.

003

3da

ta c

ateg

oriz

ed a

s no

t rel

iabl

eM

eat

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esM

ilk0.

0012

5N

AN

AN

A0.

0012

5N

A29

mea

n re

pres

ents

one

hal

fdet

ecti

on

limit

Nut

s &

N

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

bean

sO

ther

food

sN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Poul

try

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esPr

oces

sed

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esm

eat

Vege

tabl

es0.

005

NA

NA

NA

0.00

5N

A48

mea

n re

pres

ents

1/2

the

dete

ctio

n lim

itIn

fant

N

AN

AN

AN

AN

Ano

dat

a re

fere

nces

form

ula-

pow

der

Brea

st m

ilkN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

aT

he 1

0th

perc

enti

le,m

edia

n,an

d 90

th p

erce

ntile

val

ues

wer

e ca

lcul

ated

by

usin

g th

e m

eans

oft

he s

tudi

es,n

ot th

e in

divi

dual

dat

apoi

nts.

bT

he n

umbe

r of

data

poi

nts

repr

esen

ts th

e nu

mbe

r of

data

poi

nts

avai

labl

e w

ith

suff

icie

nt in

form

atio

n to

cal

cula

te a

n ar

ithm

etic

mea

n.c

NA

not

ava

ilabl

e.

3.1.3Drinking Water

The calculated mean concentration of DMP in Canadian drinking water is equalto one-half the detection limit (0.5 µg L–1). One reference for Europe reports con-centrations less than the detection limit (0.1 µg L–1).

3.2Sediment

The mean concentration of DMP in sediment is highest in the United States(3.6 mg kg–1) and is two orders of magnitude lower for Europe (0.011 mg kg–1).The maximum concentration reported in the United States is 590 mg kg–1. Nodata references are reported for Canada or Japan/Asia.

3.3Soil

No data are available for DMP measured in soil.

3.4Air

Concentrations measured in air are only identified for Europe. The calcu-lated mean concentration in indoor air is 20.2 ng m–3, with the data ranging from <1 ng m–3 to a maximum of 129 ng m–3. The calculated mean concen-tration in outdoor air is 2.55 ng m–3, with the data ranging from <1 ng m–3 to23 ng m–3.

3.5Dust

Measurements of DMP in dust are only available for Europe, where concentra-tions range between <200 µg kg–1 and 5000 µg kg–1. The calculated mean con-centration is 1730 µg kg–1.

3.6Food

Measurements of DMP concentrations in food are very limited. There are notenough data to prepare a summary by region, nor to complete each of the foodcategories reported for the other phthalates. DMP has not been detected in fish,milk, or vegetables.

136 K. Clark et al.

3.7Other Media

3.7.1Wastewater

The calculated mean concentrations for Canada and Europe are comparable(2.76 µg L–1 and 1.5 µg L–1, respectively).A maximum concentration of 207 µg L–1

was reported for the United States, which is an order of magnitude higher thanthat of Canada or Europe. The calculated mean concentration for the UnitedStates is 13.7 µg L–1.

3.7.2Sludge

Only one reference is available for DMP measured in sludge in the United States.This reference presents a mean concentration of 39,800 µg kg–1 with no reportedrange. In a study of 72 samples of sewage sludge from 12 locations across Canada,seven of the locations report non-detectable results [11]. The measured concen-trations of DMP in sludge range between the detection limit (assumed equal to10 µg kg–1) and 130 µg kg–1 with an overall calculated mean of 30 µg kg–1. Sludgedata for Europe range between 20 µg kg–1 and 19 500 µg kg–1 with a calculatedmean concentration of 863 µg kg–1.

4Diethyl Phthalate (DEP)

A summary of the reported concentrations of DEP is presented in Table 3.

4.1Water

4.1.1Surface Water

The overall mean concentration of DEP in surface water in Canada (1.42 µg L–1)is comparable with that calculated for the United States (1.33 µg L–1).As indicatedin Table 3, the calculated mean concentration for Canada is influenced by the datafor surface water with substantial industrial sources. The mean concentrationsfor rural and urban Canada are less than one-half of the calculated mean for theUnited States. The European and Japan/Asian mean concentrations (0.087 µg L–1

and 0.1 µg L–1, respectively) are lower than the North American data.

4.1.2Groundwater

Only two references are available for measurements of DEP in groundwater in theUnited States. A maximum concentration of 147 µg L–1 was reported for the

Observed Concentrations in the Environment 137

138 K. Clark et al.Ta

ble

3.Su

mm

ary

ofco

ncen

trat

ions

for

diet

hyl p

htha

late

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Surf

ace

wat

er (µ

g L–1

)C

anad

a1.

420.

0555

NA

cN

AN

A17

42da

ta p

rim

arily

from

Alb

erta

Env

iron

men

t [9]

rura

l0.

560.

055

NA

0.5

NA

963

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itur

ban

0.5

0.05

2.5

NA

0.5

NA

171

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itin

dust

rial

sou

rces

3.04

0.05

55N

A3.

5N

A60

8m

edia

n=

3.5

USA

1.33

0.01

550.

013

0.01

95.

765

7Eu

rope

0.08

7<

0.00

84

0.00

90.

030.

1432

1Ja

pan/

Asi

a0.

1N

AN

AN

AN

AN

A5

aver

age

base

d on

1/2

the

dete

ctio

n lim

it

Gro

undw

ater

(µg

L–1)

Can

ada

NA

NA

NA

NA

NA

NA

Can

adia

n da

ta r

epre

sent

ed b

y dr

inki

ng

wat

er s

umm

ary

USA

143

0.87

147

NA

NA

NA

351

refe

renc

e w

ith

34da

ta p

oint

s (m

edia

n=

11;m

ean

=14

7)Eu

rope

NA

0.1

0.2

NA

NA

NA

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Dri

nkin

g w

ater

(µg

L–1)

Can

ada

0.52

0.08

43

NA

NA

NA

1150

data

pri

mar

ily fr

om A

lber

ta E

nvir

onm

ent [

8]ru

ral

0.53

0.08

43

NA

0.5

NA

587

med

ian

=0.

5;re

pres

ents

1/2

dete

ctio

n lim

itur

ban

0.52

0.42

3N

A0.

5N

A56

3m

edia

n=

0.5;

repr

esen

ts 1

/2th

e de

tect

ion

limit

USA

6.57

0.01

100.

028

0.1

5.4

67Eu

rope

0.05

0.05

0.1

NA

NA

NA

1av

erag

e ba

sed

on 1

/2th

e de

tect

ion

limit

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Sedi

men

ts (m

g kg

–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esU

SA3.

35<

0.00

259

00.

030.

078.

316

74Eu

rope

0.00

75<

0.00

72

0.00

350.

004

0.01

46Ja

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Observed Concentrations in the Environment 139Ta

ble

3(c

onti

nued

)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Soil

(µg

kg–1

)C

anad

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

USA

7.50

¥10

4N

AN

AN

A62

0N

A37

Euro

peN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Air

(ng

m–3

)C

anad

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

USA ou

tdoo

rN

A0.

255

NA

NA

NA

indo

orN

AN

AN

AN

A34

0N

AEu

rope

outd

oor

38.6

<1

242

0.55

11.5

7632

indo

or62

125

3234

119

171

333

26Ja

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Dus

t (in

door

s) (µ

g kg

–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esU

SAN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Euro

pe27

0090

070

0050

030

0052

007

dete

cted

in 5

of7

sam

ples

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Oth

er m

edia

W

aste

wat

er (µ

g L–1

)in

clud

es s

ewag

e,sl

urry

,eff

luen

t,an

d in

fluen

tC

anad

a3.

040.

0555

NA

NA

NA

608

USA

6.24

0.01

1220

NA

NA

NA

14Eu

rope

4.73

<0.

0276

.90.

030.

219.

151

Japa

n/A

sia

NA

<0.

23.

7N

AN

AN

A

140 K. Clark et al.Ta

ble

3(c

onti

nued

)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Slud

ge (µ

g kg

–1)

prim

ary

sew

age

slud

geC

anad

a33

7<

2050

00N

AN

AN

A77

USA

290

NA

NA

NA

NA

NA

1Eu

rope

507

<0.

0229

000.

0111

920

5018

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Food

(µg

g–1)

not s

epar

ated

into

geo

grap

hica

l loc

atio

nsB

ever

ages

0.02

7<

0.05

0.09

0.02

50.

025

0.03

840

dete

cted

in 1

of40

sam

ples

Cer

eal

0.11

0.04

0.19

0.04

30.

110.

184

dete

cted

in 3

of4

sam

ples

Dai

ry

0.05

NA

NA

NA

NA

NA

6m

ean

repr

esen

ts 1

/2th

e de

tect

ion

limit

(exc

l.m

ilk)

Eggs

0.05

NA

NA

NA

NA

NA

1m

ean

repr

esen

ts o

ne h

alfd

etec

tion

lim

itFa

t & o

ils0.

25N

AN

AN

AN

AN

A3

mea

n re

pres

ents

1/2

the

dete

ctio

n lim

itFi

sh0.

059

<0.

001

0.5

0.00

70.

050.

563

Frui

ts0.

076

<0.

040.

73N

AN

AN

A13

dete

cted

in 2

of13

sam

ples

Gra

ins

0.05

NA

NA

NA

NA

NA

9de

tect

ed in

1of

9sa

mpl

esM

eat

0.05

NA

NA

NA

NA

NA

9m

ean

repr

esen

ts 1

/2th

e de

tect

ion

limit

Milk

0.00

2<

0.00

25<

0.01

0.00

275

0.00

50.

005

33m

ean

repr

esen

ts 1

/2th

e de

tect

ion

limit

Nut

s &

0.

045

NA

NA

NA

NA

NA

3m

ean

repr

esen

ts 1

/2th

e de

tect

ion

limi

bean

stO

ther

Foo

ds0.

58<

0.01

5.3

0.00

50.

051.

519

dete

cted

in 5

of19

sam

ples

Poul

try

0.05

NA

NA

NA

NA

NA

2m

ean

repr

esen

ts 1

/2th

e de

tect

ion

limit

Proc

esse

d 0.

05N

AN

AN

AN

AN

A15

mea

n re

pres

ents

one

hal

fdet

ecti

on li

mit

mea

tVe

geta

bles

0.00

5<

0.01

0.02

6N

AN

AN

A48

Infa

nt

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esfo

rmul

a-po

wde

rBr

east

milk

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

aT

he 1

0th

perc

enti

le,m

edia

n,an

d 90

th p

erce

ntile

val

ues

wer

e ca

lcul

ated

by

usin

g th

e m

eans

oft

he s

tudi

es,n

ot th

e in

divi

dual

dat

apoi

nts.

bT

he n

umbe

r of

data

poi

nts

repr

esen

ts th

e nu

mbe

r of

data

poi

nts

avai

labl

e w

ith

suff

icie

nt in

form

atio

n to

cal

cula

te a

n ar

ithm

etic

mea

n.c

NA

not

ava

ilabl

e.

United States. Only one reference is available for Europe and the reported con-centrations range between 0.1 µg L–1 and 0.2 µg L–1. Canadian data are availablein a drinking water database reported below. Although some of these data rep-resent Canadian groundwater, the data does not differentiate between surface wa-ter supplies and groundwater supplies.

4.1.3Drinking Water

The highest calculated mean concentration is for the United States (6.57 µg L–1).The calculated mean concentrations did not vary significantly between urbanand rural areas in Canada (0.52 µg L–1 and 0.53 µg L–1, respectively). These meanvalues are comparable to the European maximum of 0.1 µg L–1. No data are avail-able for Japan/Asia.

4.2Sediment

References for DEP reported in sediments are only available for the United Statesand Europe. The mean concentration for DEP in sediment is highest in the UnitedStates (3.35 mg kg–1) and is over two orders of magnitude lower in Europe(0.0075 mg kg–1). The maximum reported concentration was in the United States(590 mg kg–1).

4.3Soil

Only one reference is available reporting measurements of DEP in soil. This ref-erence reports 37 samples with a mean concentration of 75,000 µg kg–1 and a me-dian of 620 µg kg–1. The range of the data is not reported.

4.4Air

Measured concentrations indoors are higher than concentrations outdoors.The maximum mean indoor air concentration is for Europe (621 ng m–3),compared with a median concentration of 340 ng m–3 in the United States. TheEuropean data indicated outdoor concentrations ranging between <1 ng m–3 and242 ng m–3 with a calculated mean of 38.6 ng m–3.

4.5Dust

Measurements of DEP in dust are only available for Europe. Reported concen-trations range between 900 µg kg–1 and 7000 µg kg–1. The calculated mean con-centration is 2700 µg kg–1.

Observed Concentrations in the Environment 141

4.6Food

Several references present DEP concentrations measured in food; however, thereare insufficient food data to prepare a summary by region. Concentrations mea-sured in the available food groups reveal the highest maximum concentrationmeasured in other foods (5.3 µg g–1 in miscellaneous packaged food). DEP wasnot detected in dairy products, eggs, fats and oils, meats, milk, nuts and beans,poultry, or processed meat. No measurements of DEP in infant formula or breastmilk are available.

4.7Other Media

4.7.1Wastewater

Canada, Europe, and the United States have comparable calculated mean con-centrations (3.04 µg L–1, 4.73 µg L–1 and 6.24 µg L–1, respectively). A maximumconcentration of 1220 µg L–1 was reported for the United States, which is morethan an order of magnitude higher than that of Canada and Europe. The lowestmaximum concentration reported for Japan/Asia was 3.7 µg L–1.

4.7.2Sludge

The measured concentrations of DEP in sludge range between a trace amountand 5000 µg kg–1, with an overall calculated mean of 337 µg kg–1. One referencefor the United States reports a mean concentration of 290 µg kg–1. The Europeandata result in a calculated mean concentration of 507 µg kg–1.

5Dibutyl Phthalate (DBP)

A summary of the reported concentrations of DBP is presented in Table 4.

5.1Water

5.1.1Surface Water

The calculated mean concentration of DBP in surface water for Canada(1.42 µg L–1) is slightly higher than that calculated for the other regions. The max-imum measured concentration was 350 µg L–1 (for Japan/Asia). The maximumfor Canadian waters with substantial industrial sources (64 µg L–1) is compara-ble with the United States maximum (63 µg L–1).

142 K. Clark et al.

Observed Concentrations in the Environment 143Ta

ble

4.Su

mm

ary

ofco

ncen

trat

ions

ofd

ibut

yl p

htha

late

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Surf

ace

wat

er (µ

g L–1

)C

anad

a1.

420.

0003

100

0.01

0.09

1.4

1672

data

pri

mar

ily fr

om A

lber

ta E

nvir

onm

ent [

9],

EC&

HC

[14]

rura

l0.

680.

015.

2N

Ac

0.5

NA

874

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itur

ban

0.45

0.00

036.

98N

A0.

5N

A21

8m

edia

n=

0.5

for A

lber

ta d

atas

et (1

65va

lues

into

tal)

;rep

rese

nts

1/2

the

dete

ctio

n lim

itin

dust

rial

sou

rces

2.9

0.01

64N

A1.

0N

A58

0m

edia

n=

1.0

USA

0.41

0.00

0763

0.00

190.

014

1.0

1180

Euro

pe0.

240.

002

170

0.00

90.

110.

723

4Ja

pan/

Asi

a0.

790.

0235

00.

50.

50.

511

9

Gro

undw

ater

(µg

L–1)

Can

ada

NA

NA

NA

NA

NA

NA

Can

adia

n da

ta r

epre

sent

ed b

y dr

inki

ng

wat

er s

umm

ary

USA

4.5

0.5

50N

A4.

74N

A10

44Eu

rope

0.25

0.12

0.46

NA

NA

NA

1Ja

pan/

Asi

a2.

140.

512

0.5

0.5

0.5

21m

edia

n=

0.5;

repr

esen

ts 1

/2th

e de

tect

ion

limit

Dri

nkin

g w

ater

(µg

L–1)

Can

ada

0.8

0.14

8N

AN

AN

A70

5da

ta p

rim

arily

from

Alb

erta

Env

iron

men

t [8]

mea

sure

men

ts in

dri

nkin

g w

ater

rura

l0.

730.

148

NA

0.5

NA

354

med

ian

=0.

5;re

pres

ents

1/2

dete

ctio

n lim

itur

ban

0.85

0.18

7N

A0.

5N

A35

1m

edia

n=

0.5;

repr

esen

ts 1

/2th

e de

tect

ion

limit

USA

4.6

0.01

500.

030.

15.

811

11Eu

rope

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esJa

pan/

Asi

a0.

9<

1.0

3.1

0.5

0.5

0.5

36

144 K. Clark et al.

Tabl

e4

(con

tinu

ed)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Sedi

men

ts (m

g kg

–1)

Can

ada

NA

0.01

10.4

NA

NA

NA

2re

fere

nces

;una

ble

to c

alcu

late

ave

rage

USA

14.6

0.00

007

3333

0.00

760.

1218

3073

Euro

pe0.

220.

0001

28.3

0.00

350.

016

0.25

94Ja

pan/

Asi

a0.

12<

0.00

52.

30.

030.

080.

287

Soil

(µg

kg–1

)C

anad

aN

A0.

027

1.52

NA

NA

NA

3re

fere

nces

wit

h no

dat

a po

ints

;una

ble

to

calc

ulat

e av

erag

eU

SA24

1.5

280

NA

NA

NA

1nu

mbe

r of

data

poi

nts

unkn

own;

aver

age

repo

rted

from

ref

eren

ce

Euro

pe18

40.

0156

0N

AN

AN

A13

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Air

(ng

m–3

)C

anad

aou

tdoo

r4.

90.

170

02.

54.

35.

744

indo

or2.

9N

AN

AN

AN

AN

A9

USA ou

tdoo

r2.

50.

0892

0.33

2.6

1738

indo

or0.

20.

242

0N

AN

AN

A6

Euro

peou

tdoo

r15

.80.

238

03.

28

8683

indo

or10

32<

394

4515

655

159

0026

Japa

n/A

sia

outd

oor

960.

0021

370

NA

NA

NA

2in

door

0.00

09N

AN

AN

AN

AN

A1

Observed Concentrations in the Environment 145Ta

ble

4(c

onti

nued

)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Dus

t (in

door

s) (µ

g kg

–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esU

SAN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Euro

pe5.

67¥

104

56.

78¥

105

2500

5.6¥

104

1.6¥

105

26Ja

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Oth

er m

edia

Was

tew

ater

(µg

L–1)

incl

udes

sew

age,

slur

ry,e

fflu

ent,

and

influ

ent

Can

ada

2.8

0.01

100

NA

NA

NA

599

USA

4.7

0.00

422

650.

030.

087.

139

Euro

pe22

<0.

0210

000.

070.

2736

70Ja

pan/

Asi

a0.

10.

125

0N

AN

AN

A10

Slud

ge (µ

g kg

–1)

prim

ary

sew

age

slud

ge

Can

ada

1.29

¥10

420

01.

61¥

105

1.68

¥10

45.

67¥

104

5.69

¥10

482

USA

5.60

¥10

420

3.20

¥10

649

324

001.

74¥

105

7Eu

rope

2.10

¥10

40.

144.

30¥

105

0.36

2480

2.82

¥10

470

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esR

ainw

ater

(µg

L–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esU

SAN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Euro

pe0.

160.

034.

50.

200.

291.

076

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Food

(µg

g–1)

not s

epar

ated

into

geo

grap

hica

l loc

atio

nsB

ever

ages

0.1

0.02

50.

560.

025

0.02

50.

325

12de

tect

ed in

4of

12sa

mpl

esC

erea

l0.

30.

10.

50.

10.

30.

55

Dai

ry

0.04

<0.

01<

0.1

0.04

0.05

0.05

12(e

xcl.

milk

)Eg

gs0.

090.

050.

10.

059

0.08

90.

099

11Fa

t & o

ils2.

5<

0.5

110.

481.

86.

528

146 K. Clark et al.

Tabl

e4

(con

tinu

ed)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Food

(µg

g–1)

not s

epar

ated

into

geo

grap

hica

l loc

atio

nsFi

sh0.

23<

0.00

135

0.00

050.

0045

0.34

43Fr

uits

0.03

30.

010.

16N

AN

AN

A14

dete

cted

in 3

of14

sam

ples

Gra

ins

0.26

0.03

1.9

0.04

60.

050.

469

dete

cted

in 4

of9

sam

ples

Mea

t0.

920.

037.

30.

050.

051.

39

Milk

0.01

20.

003

0.2

0.00

360.

015

0.04

550

Nut

s &

0.

18<

0.09

0.57

0.04

50.

045

0.23

6de

tect

ed in

3of

6sa

mpl

esbe

ans

Oth

er fo

ods

0.16

<0.

0162

0.00

850.

090.

6138

Poul

try

0.13

<0.

10.

20.

050.

050.

174

Proc

esse

d 0.

54<

0.1

6.2

NA

NA

NA

13m

eat

Vege

tabl

es0.

17<

0.01

100.

020.

088.

067

Infa

nt

0.07

<0.

050.

40.

025

0.06

50.

2353

form

ula-

pow

der

Infa

nt

0.00

29<

0.00

30.

007

NA

NA

NA

4de

tect

ed in

1of

4sa

mpl

esfo

rmul

a-liq

uid

Brea

st M

ilk0.

030.

010.

08N

AN

AN

A10

Baby

Foo

d0.

028

<0.

010.

03N

AN

AN

A12

aT

he 1

0th

perc

enti

le,m

edia

n,an

d 90

th p

erce

ntile

val

ues

wer

e ca

lcul

ated

by

usin

g th

e m

eans

oft

he s

tudi

es,n

ot th

e in

divi

dual

dat

a po

ints

.b

The

num

ber

ofda

ta p

oint

s re

pres

ents

the

num

ber

ofda

ta p

oint

s av

aila

ble

wit

h su

ffic

ient

info

rmat

ion

to c

alcu

late

an

arit

hmet

ic m

ean.

cN

A n

ot a

vaila

ble.

5.1.2Groundwater

Groundwater data are primarily available for the United States and Japan/Asia.A maximum concentration of 50 µg L–1 is reported for the United States. Thehighest mean concentration is also calculated for the United States (4.5 µg L–1),which is approximately twice the mean for Japan/Asia (2.14 µg L–1). Only one ref-erence for Europe reported a mean concentration (0.25 µg L–1). In addition, themaximum concentration for Europe (0.46 µg L–1) is considerably lower than themaximum concentrations for the United States and Japan/Asia. Canadian dataare available in a drinking water database reported below.Although some of thesedata represent Canadian groundwater, the data do not differentiate between sur-face water supplies and groundwater supplies.

5.1.3Drinking Water

The drinking water data reported for the United States results in a calcu-lated mean concentration (4.6 µg L–1) higher than the mean concentrations cal-culated for Canada (0.8 µg L–1) and Japan/Asia (0.9 µg L–1). The maximum concentration was reported for the United States (50 µg L–1). No data are avail-able for Europe.

5.2Sediment

The mean concentration of DBP in sediment is highest in the United States(14.6 mg kg–1). The United States data represents the largest data set (3073 re-ported samples). The mean concentration is two orders of magnitude lower inEurope (0.22 mg kg–1) and Japan/Asia (0.12 mg kg–1), although there are signifi-cantly less data references. Two data references are reported for Canada [15],however, a mean concentration is not reported. The Canadian data range between0.01 mg kg–1 and 10.4 mg kg–1, which are comparable with the European andJapan/Asian data. The maximum concentration was reported for the UnitedStates (3333 mg kg–1).

5.3Soil

Few references report measured concentrations of DBP in soil. An average con-centration in soil of 24 µg kg–1 was reported for the United States, which is lowerthan the European mean of 184 µg kg–1. The maximum measured concentrationwas recorded in Europe (560 µg kg–1). Maximum concentrations are comparablefor the United States and Europe, while the Canadian maximum (1.52 µg kg–1) isconsiderably lower.

Observed Concentrations in the Environment 147

5.4Air

Concentrations measured in air are reported to be the highest in Europe. The cal-culated mean indoor air concentration in Europe is 1032 ng m–3, while the max-imum is 9445 ng m–3. The calculated means in the other regions are lower,although they are based upon less data.

The maximum mean concentration of DBP in outdoor air is 96 ng m–3 (inJapan/Asia). The calculated mean concentration in Europe is 15.8 ng m–3, whilethe mean concentrations in the United States and Canada are 2.5 ng m–3 and4.9 ng m–3, respectively.

5.5Dust

References reporting measured concentrations in dust are only available for Europe. The calculated mean concentration is 56,700 µg kg–1, with data rangingbetween 5 µg kg–1 and 6.78 ¥105 µg kg–1.

5.6Food

Several references present DBP concentrations measured in food; however, thereare not sufficient food data to prepare a summary by region. Concentrationsmeasured in the available food groups reveal the highest maximum concentra-tion measured in “other foods” (62 µg g–1, measured in gravy and Parmesancheese) and fish (maximum of 35 µg g–1). The highest mean concentration is forfats and oils (2.5 µg g–1).

5.7Other Media

5.7.1Wastewater

The maximum mean concentration is calculated for Europe (22 µg L–1), althoughthis mean is comparable with the mean concentrations calculated for the otherregions. The lowest calculated mean is for Japan/Asia (0.1 µg L–1). The Canadianmaximum concentration of 100 µg L–1 is the same order of magnitude as theJapan/Asian maximum of 250 µg L–1. The highest reported maximum concen-tration was measured in the United States (2265 µg L–1).

5.7.2Sludge

The calculated mean concentrations of DBP in sludge are in the same order ofmagnitude (104 µg kg–1) for the three regions referencing data (Canada, United

148 K. Clark et al.

States, and Europe). The highest concentration is reported for the United States(3.2 ¥106 µg kg–1). No data are referenced for Japan/Asia.

6Butylbenzyl Phthalate (BBP)

A summary of the reported concentrations of BBP is presented in Table 5.

6.1Water

6.1.1Surface Water

The overall calculated mean concentration of BBP in surface water in Canada(1.51 µg L–1) is higher than that calculated for the United States (0.38 µg L–1). Themajority of the data points for the United States are non-detectable. Similarly, themedian concentrations for Canadian rural and urban waters are both 0.5 µg L–1,which represent one half of the detection limit.

The maximum concentration (66 µg L–1) for the United States is comparablewith the Canadian maximum for rural and surface waters with substantial in-dustrial input (50 µg L–1 and 84 µg L–1, respectively). The maximum measuredconcentration in urban Canadian surface water (6 µg L–1) is comparable with themaximum measured concentration in Europe (13.9 µg L–1). Only one data refer-ence is available for Japan/Asia.

6.1.2Groundwater

Very little data have been identified for BBP in groundwater. A maximum con-centration of 38 µg L–1 is reported in one of two references for the United States.These references did not report an average concentration. Canadian data areavailable in a drinking water database reported below. Although some of thesedata represent Canadian groundwater, the data do not differentiate between sur-face water supplies and groundwater supplies.

6.1.3Drinking Water

The maximum measured concentration in the United States is 38 µg L–1. Analy-sis of the Canadian data is primarily dependent upon treatment of the detectionlimit, as most measurements were less than the detection limit. The maximummeasured value is 2.8 µg L–1. One reference for the Netherlands reports non-de-tectable concentrations (<0.1 µg L–1).

Observed Concentrations in the Environment 149

150 K. Clark et al.Ta

ble

5.Su

mm

ary

ofco

ncen

trat

ions

ofb

utyl

benz

yl p

htha

late

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Surf

ace

wat

er (µ

g L–1

)C

anad

a1.

510.

0284

NA

c2.

47N

A18

84da

ta p

rim

arily

from

Alb

erta

Env

iron

men

t [9]

rura

l0.

610.

0550

NA

0.5

NA

1010

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itur

ban

0.55

0.05

6N

A0.

5N

A18

3m

edia

n=

0.5;

repr

esen

ts 1

/2de

tect

ion

limit

indu

stri

al s

ourc

es3.

050.

0284

NA

1.0

NA

629

med

ian

=1.

0U

SA0.

380.

001

660.

002

0.00

30.

5927

45Eu

rope

0.06

<0.

0113

.90.

009

0.00

90.

2185

Japa

n/A

sia

NA

<0.

21

NA

NA

NA

Gro

undw

ater

(µg

L–1)

Can

ada

NA

NA

NA

NA

NA

NA

Can

adia

n da

ta r

epre

sent

ed b

y dr

inki

ng w

ater

sum

mar

y be

low

USA

NA

ND

d38

NA

NA

NA

dete

ctio

n lim

it n

ot p

rovi

ded

Euro

peN

A0.

040.

24N

AN

AN

AJa

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Dri

nkin

g w

ater

(µg

L–1)

Can

ada

0.5

<0.

0000

22.

8N

AN

AN

A12

39da

ta p

rim

arily

from

Alb

erta

Env

iron

men

t [8]

rura

l0.

50.

51

NA

0.5

NA

619

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itur

ban

0.5

0.5

1N

A0.

5N

A62

0m

edia

n=

0.5;

repr

esen

ts 1

/2th

e de

tect

ion

limit

USA

NA

NA

38N

AN

AN

AEu

rope

<0.

1<

0.1

<0.

1N

AN

AN

AJa

pan/

Asi

aN

AN

AN

AN

AN

AN

A

Sedi

men

ts (m

g kg

–1)

Can

ada

NA

<10

037

0N

AN

AN

AU

SA0.

190.

005

5.5

0.01

20.

099

0.27

1485

Euro

pe0.

059

<0.

007

18.2

0.00

250.

007

0.08

86Ja

pan/

Asi

aN

AN

D0.

02N

AN

AN

Ade

tect

ion

limit

not

pro

vide

d

Observed Concentrations in the Environment 151Ta

ble

5(c

onti

nued

)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Soil

(µg

kg–1

)C

anad

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

USA

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esEu

rope

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esJa

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Air

(ng

m–3

)C

anad

aou

tdoo

rN

A0.

381.

78N

AN

AN

Ain

door

ND

NA

NA

NA

NA

NA

dete

ctio

n lim

it n

ot p

rovi

ded

USA ou

tdoo

r9.

51

201.

94.

5515

.589

min

.and

max

.are

ref

eren

ced

aver

ages

indo

orN

AN

A14

0N

A35

120

250

indo

or a

ir s

ampl

esEu

rope

outd

oor

1.7

<1

<10

0.5

1.0

4.7

32in

door

35<

346

51.

513

163

26Ja

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Dus

t (in

door

s) (µ

g kg

–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esU

SAN

AN

AN

AN

AN

AN

Ano

dat

a re

fere

nces

Euro

pe3.

33¥

105

<10

001.

7¥10

622

501.

9¥10

41.

1¥10

68

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Oth

er m

edia

Was

tew

ater

(µg

L–1)

incl

udes

sew

age,

slur

ry,e

fflu

ent,

and

influ

ent

Can

ada

3.05

0.02

84N

A1.

0N

A62

9U

SA29

90.

144

9N

AN

AN

A3

Euro

pe0.

76<

0.06

300.

071

0.61

1.72

47Ja

pan/

Asi

a0.

1<

0.2

1.5

NA

NA

NA

10

152 K. Clark et al.Ta

ble

5(c

onti

nued

)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Slud

ge (µ

g kg

–1)

prim

ary

sew

age

slud

ge

Can

ada

3200

501.

40¥

104

NA

NA

NA

79U

SA1.

12¥

105

NA

NA

NA

NA

NA

3Eu

rope

1800

0.14

2.10

¥10

51.

367

076

0062

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Food

(µg

g–1)

not s

epar

ated

into

geo

grap

hica

l loc

atio

nsB

ever

ages

0.03

9<

0.05

0.11

0.02

50.

025

0.06

86

Cer

eal

0.05

NA

NA

NA

0.05

NA

4m

ean

repr

esen

ts 1

/2th

e de

tect

ion

limit

Dai

ry

0.4

<0.

11.

60.

050.

051.

16

(exc

l.m

ilk)

Eggs

0.08

<0.

10.

09N

AN

AN

A3

Fat &

oils

7.4

<0.

547

.80.

250.

648.

523

Fish

0.01

<0.

001

0.01

80.

020.

050.

0531

Frui

ts0.

02N

AN

AN

AN

AN

A13

mea

n re

pres

ents

1/2

the

dete

ctio

n lim

itG

rain

s0.

05N

AN

AN

A0.

05N

A9

mea

n re

pres

ents

1/2

the

dete

ctio

n lim

itM

eat

0.13

<0.

10.

80.

050.

050.

149

Milk

0.00

12<

0.00

250.

008

0.00

160.

005

0.00

535

Nut

s &

0.

045

NA

NA

NA

0.04

5N

A3

mea

n re

pres

ents

1/2

the

dete

ctio

n lim

itbe

ans

Oth

er fo

ods

0.09

<0.

010.

480.

005

0.05

0.13

19de

tect

ed in

1of

19sa

mpl

esPo

ultr

y0.

040.

03<

0.1

0.03

40.

050.

054

Proc

esse

d 0.

05N

AN

AN

AN

AN

A3

mea

n re

pres

ents

1/2

the

dete

ctio

n lim

itm

eat

Vege

tabl

es0.

005

<0.

010.

018

NA

NA

NA

48In

fant

0.

044

<0.

001

0.25

0.00

150.

0065

0.12

57fo

rmul

a-po

wde

rBr

east

milk

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

aT

he 1

0th

perc

enti

le,m

edia

n,an

d 90

th p

erce

ntile

val

ues

wer

e ca

lcul

ated

by

usin

g th

e m

eans

oft

he s

tudi

es,n

ot th

e in

divi

dual

dat

a po

ints

.b

The

num

ber

ofda

ta p

oint

s re

pres

ents

the

num

ber

ofda

ta p

oint

s av

aila

ble

wit

h su

ffic

ient

info

rmat

ion

to c

alcu

late

an

arit

hmet

ic m

ean.

cN

A n

ot a

vaila

ble.

dN

D n

ot d

etec

ted.

6.2Sediment

The mean concentration for sediment is higher in the United States(0.19 mg kg–1) as compared with that of Europe (0.059 mg kg–1), although thereare less data references for Europe. Three references for Canada are available thatcite a maximum concentration of 370 mg kg–1. This is higher than the maximumconcentrations reported for the United States (5.5 mg kg–1) and Europe(18.2 mg kg–1). Only two references are reported for Japan/Asia, with significantlylower values (maximum concentration of 0.02 mg kg–1).

6.3Soil

No data are available for BBP measured in soil.

6.4Air

Concentrations in indoor air are higher than concentrations in outdoor air. Thecalculated mean indoor air concentration is 35 ng m–3 in Europe, which corre-sponds with the median of a United States study. The maximum indoor air con-centration was in Europe and is 465 ng m–3. A concentration of 120 ng m–3 is re-ported as the 90th percentile of a study of indoor air in California.

The mean outdoor concentration in Europe is 1.7 ng m–3, while the calculatedmean concentration in the United States is 9.5 ng m–3. Canadian data range be-tween 0.38 ng m–3 and 1.78 ng m–3.

6.5Dust

Measurements of BBP in dust are only available for Europe. The concentrationsrange between <1000 µg kg–1 and 1.75 ¥106 µg kg–1, with a calculated mean con-centration of 3.33 ¥105 µg kg–1.

6.6Food

Several references report BBP concentrations measured in food; however,there are insufficient data to prepare a summary by region. Concentrations mea-sured in the available food groups reveal the highest mean concentration for fats and oils (7.4 µg g–1). For the food categories of cereal, fruits, grains, nuts and beans, and processed meats, BBP was not detected and the mean value inTable 5 represents one half the detection limit. BBP was detected in only 1 of19 samples of “other foods”. No measurements of BBP in breast milk are avail-able.

Observed Concentrations in the Environment 153

6.7Other Media

6.7.1Wastewater

The comparative results of BBP measured in various wastewaters, including in-dustrial effluents and influents, leachates, and storm water, etc., are summarizedin Table 5. The maximum mean concentration is calculated for the United States(299 µg L–1). This mean value is two orders of magnitude higher than the meanconcentrations calculated for the other regions. The lowest calculated mean is forJapan/Asia (0.1 µg L–1). The Canadian maximum concentration of 84 µg L–1 is thesame order of magnitude as the European maximum of 30 µg L–1. The highest re-ported maximum was measured in the United States (449 µg L–1).

6.7.2Sludge

The majority of the reported data for sludges and composts are available forCanada and Europe. The maximum mean concentration is calculated for theUnited States (112,000 µg kg–1). The calculated mean for Canada is less(3200 µg kg–1), as is the calculated mean concentration for Europe (1800 µg kg–1).The maximum measured concentration is reported for Europe (210,000 µg kg–1);however, a maximum concentration is not reported in the study referenced forthe United States.

7Bis(2-Ethylhexyl) Phthalate (DEHP)

A summary of the reported concentrations of DEHP is presented in Table 6.

7.1Water

7.1.1Surface Water

The lowest mean concentration (0.21 µg L–1) is for the United States. The calcu-lated mean concentrations for Europe and Japan/Asia (0.93 µg L–1 and 0.9 µg L–1,respectively) are within the range of the Canadian urban and rural mean con-centrations of 1.15 µg L–1 and 0.78 µg L–1, respectively. The overall Canadian meanis 3.78 µg L–1. Note that the Canadian means are strongly influenced by the as-sumption that non-detect values are equal to one half the detection limit. The me-dian concentration for both urban and rural waters is 0.5 µg L–1, which representsone half the detection limit. The detection limits for the United States and Euro-pean data are more than an order of magnitude lower than the lowest Canadiandetection limit.

154 K. Clark et al.

Observed Concentrations in the Environment 155Ta

ble

6.Su

mm

ary

ofco

ncen

trat

ions

ofb

is(2

-eth

ylhe

xyl)

pht

hala

te

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Surf

ace

wat

er (µ

g L–1

)C

anad

a3.

780.

0533

60.

292.

876.

610

81da

ta p

rim

arily

from

Alb

erta

Env

iron

men

t [9]

;hi

gh/lo

w d

ata

valu

es r

emov

edru

ral

0.78

0.05

14N

Ac

0.5

NA

367

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itur

ban

1.15

0.05

11N

A0.

5N

A78

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itin

dust

rial

sou

rces

5.69

0.05

336

NA

3.0

NA

607

med

ian

=3.

0U

SA0.

21<

0.00

213

70.

012

0.05

1.6

1309

Euro

pe0.

93<

0.00

850

0.01

80.

211.

939

7Ja

pan/

Asi

a0.

9<

0.02

150.

50.

53.

219

8

Gro

undw

ater

(µg

L–1)

Can

ada

NA

NA

NA

NA

NA

NA

Can

adia

n da

ta r

epre

sent

ed in

dri

nkin

g w

ater

sum

mar

y be

low

USA

15.7

ND

d47

04.

315

.727

2de

tect

ion

limit

not

spe

cifie

dEu

rope

0.26

<0.

071.

40.

190.

671.

19

Japa

n/A

sia

0.79

<1.

018

.40.

50.

511

61m

axim

um is

pre

sent

ed a

s an

ave

rage

Dri

nkin

g w

ater

(µg

L–1)

Can

ada

1.72

1.0

540.

003

0.25

1.4

959

data

pri

mar

ily fr

om A

lber

ta E

nvir

onm

ent [

8]

rura

l1.

560.

454

NA

0.5

NA

422

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itur

ban

1.97

0.5

42N

A1.

0N

A49

9m

edia

n=

1.0

USA

0.55

0.16

170

0.22

0.55

0.88

4Eu

rope

NA

0.18

3.5

NA

NA

NA

NA

rang

e on

ly a

vaila

ble

Japa

n/A

sia

0.91

1.0

4.7

0.5

0.5

3.0

61

Sedi

men

ts (m

g kg

–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

avai

labl

eU

SA1.

40.

0002

721

80.

041

0.16

3.4

83Eu

rope

4.9

0.00

0148

70.

019

0.29

8.2

405

Japa

n/A

sia

0.48

0.00

917

NA

NA

NA

1

156 K. Clark et al.Ta

ble

6(c

onti

nued

)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Soil

(µg

kg–1

)C

anad

aN

A<

0.1

11N

AN

AN

AN

A30

data

poi

nts;

only

ran

ge a

vaila

ble

USA

0.03

0.03

1280

NA

NA

NA

1Eu

rope

484

5100

2950

663

Japa

n/A

sia

NA

1050

0N

AN

AN

AN

A

Air

(ng

m–3

)C

anad

aou

tdoo

r2.

00.

55.

0N

AN

AN

A3

indo

orN

A<

500

3100

NA

NA

NA

NA

USA ou

tdoo

r5.

0<

0.4

650.

762.

316

34in

door

109

2024

027

5599

107

Euro

peou

tdoo

r21

.90.

2810

901.

317

.512

685

indo

or24

518

1046

2011

139

826

Japa

n/A

sia

outd

oor

1450

<17

2800

NA

1450

NA

2in

door

1000

NA

NA

NA

NA

NA

1

Dus

t (in

door

s) (µ

g kg

–1)

Can

ada

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

esU

SA3.

24¥

106

2.38

¥10

64.

10¥

106

NA

NA

NA

2Eu

rope

6.62

¥10

520

004.

58¥

106

3.67

¥10

44.

71¥

105

2.09

¥10

655

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

refe

renc

es

Vege

tati

on (µ

g kg

–1dr

y w

t.)Eu

rope

NA

1200

11,3

00N

AN

AN

A

Observed Concentrations in the Environment 157Ta

ble

6(c

onti

nued

)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Oth

er M

edia

Was

tew

ater

(µg

L–1)

incl

udes

sew

age,

slur

ry,e

fflu

ent,

and

influ

ent

Can

ada

5.72

<0.

133

62.

65.

28.

866

2U

SA27

0.01

4400

3.0

8.3

618

Euro

pe34

.40.

068

1800

0.23

1.74

5376

Japa

n/A

sia

5.6

<0.

240

NA

NA

NA

2Sl

udge

(µg

kg–1

)pr

imar

y se

wag

e sl

udge

C

anad

a1.

39¥

105

3000

4.40

¥10

51.

90¥

104

6.80

¥10

41.

60¥

105

84U

SA3.

01¥

105

420

5.83

¥10

7N

AN

AN

A16

Euro

pe5.

31¥

104

900

2.60

¥10

61.

40¥

104

4.80

¥10

41.

60¥

105

62Ja

pan/

Asi

a48

817

0N

AN

AN

A1

1se

t ofm

easu

rem

ents

in 1

974

Rai

nwat

er (µ

g L–1

)C

anad

a0.

006

0.00

40.

01N

AN

AN

A1

USA

0.17

0.00

40.

680.

039

0.17

0.26

22Eu

rope

140.

0083

540.

430.

8538

4Ja

pan/

Asi

a6.

10.

0053

16.5

0.4

1.8

143

Food

(µg

g–1)

not s

epar

ated

into

geo

grap

hica

l loc

atio

nsB

ever

ages

0.07

70.

006

1.7

0.01

50.

043

0.15

72C

erea

l0.

530.

021.

70.

032

0.05

1.3

5D

airy

1.

50.

059

16.8

0.07

60.

962.

510

7(e

xcl.

milk

)Eg

gs0.

21<

0.01

0.6

0.01

90.

120.

474

Fat &

oils

4.1

0.7

11.9

1.0

2.4

4.6

36Fi

sh0.

465.

00¥

10–5

320.

001

0.02

1.3

64Fr

uits

0.03

<0.

020.

110.

020.

020.

062

15G

rain

s0.

5<

0.1

1.5

0.05

0.14

1.1

11M

eat

0.35

<0.

010.

80.

050.

050.

7527

Milk

0.08

<0.

005

1.4

0.01

0.03

50.

1410

8N

uts

&

bean

s0.

21<

0.08

0.8

0.04

50.

045

0.27

6

158 K. Clark et al.

Tabl

e6

(con

tinu

ed)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Food

(µg

g–1)

not s

epar

ated

into

geo

grap

hica

l loc

atio

nsO

ther

food

s0.

28<

0.01

250.

005

0.05

1.3

63Po

ultr

y1.

10.

052.

60.

250.

92.

24

Proc

esse

d 0.

94<

0.1

4.32

0.11

0.45

1.9

25m

eat

Vege

tabl

es0.

170.

0098

2.2

0.04

50.

048

1.1

84m

inim

um r

epre

sent

s m

edia

n fr

om o

ne s

tudy

Infa

nt

0.2

<0.

012

0.98

0.00

60.

120.

6466

form

ula-

pow

der

Infa

nt

0.00

7<

0.00

50.

150.

006

0.00

60.

007

9fo

rmul

a-liq

uid

Brea

st m

ilk0.

062

0.01

0.6

NA

0.06

2N

A10

Baby

food

0.12

0.01

0.6

0.03

90.

120.

2316

aT

he 1

0th

perc

enti

le,m

edia

n,an

d 90

th p

erce

ntile

val

ues

wer

e ca

lcul

ated

by

usin

g th

e m

eans

oft

he s

tudi

es,n

ot th

e in

divi

dual

dat

apoi

nts.

bT

he n

umbe

r of

data

poi

nts

repr

esen

ts th

e nu

mbe

r of

data

poi

nts

avai

labl

e w

ith

suff

icie

nt in

form

atio

n to

cal

cula

te a

n ar

ithm

etic

mea

n.c

NA

not

ava

ilabl

e.d

ND

not

det

ecte

d.

The maximum concentration measured in Canada is 336 µg L–1 (for surfacewater with substantial industrial sources). The maximum concentrations, mea-sured in the United States, Europe, and Japan/Asia, are 137 µg L–1, 50 µg L–1, and15 µg L–1, respectively.

7.1.2Groundwater

The database for groundwater is considerably more limited compared to that forsurface water. The highest maximum and mean concentrations for DEHP ingroundwater have been measured in the United States as 470 µg L–1 and15.7 µg L–1, respectively. The mean concentration for Europe is 0.26 µg L–1, whichis comparable to the Japan/Asian calculated mean concentration of 0.79 µg L–1.Canadian data are available in a drinking water database reported below; al-though some of these data represent Canadian groundwater, the data do not dif-ferentiate between surface water supplies and groundwater supplies.

7.1.3Drinking Water

The calculated mean concentration of DEHP in drinking water is highest inCanada (1.72 µg L–1).Although, considerably less data are available for the otherregions, the calculated mean concentrations in the United States and Japan/Asiaare slightly less (0.55 µg L–1 and 0.91 µg L–1, respectively). However, the maximumconcentration reported for the United States (170 µg L–1) is three times higherthan the Canadian maximum (54 µg L–1). The European and Japan/Asian maxi-mums are 3.5 µg L–1 and 4.7 µg L–1, respectively.

7.2Sediment

The mean concentration in sediment is highest in Europe (4.9 mg kg–1). Themean concentrations are somewhat lower in Japan/Asia (0.48 mg kg–1) and theUnited States (1.4 mg kg–1) although there are less data references. Maximum ref-erenced concentrations range between 17 mg kg–1 (Japan/Asia) and 487 mg kg–1

(Europe).

7.3Soil

Very little data are available for DEHP measured in soil. Of the data referenced,DEHP concentrations in soil are higher in Europe with a mean concentration of48 µg kg–1. The maximum concentrations range between 11.0 µg kg–1, reportedfor Canada, and 5100 µg kg–1, reported for Europe. Data for the United States arereported to range between 0.03 µg kg–1 and 1280 µg kg–1. Only one reference is re-ported for Japan/Asia, with concentrations in soil ranging between 10 µg kg–1 and500 µg kg–1.

Observed Concentrations in the Environment 159

7.4Air

In general, measured indoor concentrations are higher than concentrations out-doors. The Canadian and Japan/Asian datasets are very limited compared to theUnited States and European datasets. The mean outdoor concentration rangesfrom 2.0 ng m–3 in Canada to 1450 ng m–3 in Japan/Asia. The mean indoor concentration ranges from 109 ng m–3 in the United States to 1000 ng m–3 inJapan/Asia.

7.5Dust

DEHP concentrations measured in dust are only available for Europe and theUnited States. Europe reports concentrations ranging between 2000 µg kg–1 and4.58 ¥106 µg kg–1, with a mean of 6.62 ¥105 µg kg–1. The United States reports con-centrations ranging between 2.38 ¥106 µg kg–1 and 4.10 ¥106 µg kg–1, with a meanof 3.24 ¥106 µg kg–1.

7.6Food

DEHP concentrations measured in food are the most extensive of all of the phthalates reported; however, there are insufficient food data to prepare a sum-mary by region. DEHP concentrations measured in the various food groups inTable 6 reveal the highest mean concentration for fats and oils (4.1 µg g–1). Thehighest maximum concentrations are in fish (32 µg g–1) and “other foods”(25 µg g–1). Note that the data have not been separated by year of sampling,although it should be noted that food processing practices may have changedsince the time of sampling.

7.7Other Media

7.7.1Wastewater

The comparative results of DEHP measured in various wastewaters, including industrial effluents and influents, leachates, and storm water, etc., are summa-rized in Table 6. The maximum mean concentration is calculated for Europe(34.4 µg L–1). This mean is only slightly higher than the mean concentrations cal-culated for the other regions. Japan/Asia and Canada have the lowest calculatedmeans (5.6 µg L–1 and 5.72 µg L–1, respectively). The highest reported maximumwas measured in the United States (4400 µg L–1).

160 K. Clark et al.

7.7.2Sludge

The majority of the reported data for sludges and composts are available forCanada and Europe. The maximum mean concentration is calculated for theUnited States (301,000 µg kg–1). The calculated mean for Canada is in the sameorder of magnitude (139,000 µg kg–1), while the calculated mean concentrationfor Europe is slightly lower (53,000 µg kg–1). A single reference is available forJapan/Asia, which reports a much lower mean concentration (48 µg kg–1).

7.7.3Rainwater

The highest calculated mean concentration is for Europe (14 µg L–1), while thelowest mean concentration is for Canada (0.006 µg L–1). The highest maximumconcentration was also measured in Europe (54 µg L–1).

8Di-n-Octyl Phthalate (DnOP)

A summary of the reported concentrations of DnOP is presented in Table 7.

8.1Water

8.1.1Surface Water

The overall mean concentration for Canada is calculated to be 1.35 µg L–1. Thedata include rural and urban locations, as well as surface waters with substantialindustrial sources. The maximum concentration (45 µg L–1) was recorded in anurban area; however, the highest mean concentration (5.5 µg L–1) is calculated forsurface water with substantial industrial input. The calculated mean concentra-tion for urban areas (1.10 µg L–1) is comparable with that of the rural areas(0.91 µg L–1).

8.1.2Drinking Water

Measured concentrations in Canada range between the detection limit (<1.0 µgL–1) and 11 µg L–1. The mean concentrations are 0.54 µg L–1 and 0.53 µg L–1, for urban and rural areas, respectively.Analysis of 25 drinking water supplies in theUnited States is reported by Health Canada [17]. The results did not exceed thedetection limit. DnOP was detected, but not quantified, in three aquifer-deriveddrinking water supplies in New York State [17]. DnOP was detected in drinkingwater in Europe; however, levels were not quantified [17].

Observed Concentrations in the Environment 161

162 K. Clark et al.

Tabl

e7.

Sum

mar

y of

conc

entr

atio

ns o

fdi-

n-oc

tyl p

htha

late

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Surf

ace

wat

er (µ

g L–1

)C

anad

a1.

350.

0545

0.00

360.

0071

1.0

1901

data

pri

mar

ily fr

om A

lber

ta E

nvir

onm

ent [

9],

EC&

HC

[16]

and

HC

[17]

rura

l0.

910.

0518

NA

c0.

5N

A11

15m

edia

n=

0.5;

repr

esen

ts 1

/2th

e de

tect

ion

limit

urba

n1.

100.

0545

NA

0.5

NA

165

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itin

dust

rial

sou

rces

5.50

0.1

10N

A5.

0N

A62

1m

edia

n=

5.0

USA

NA

NA

NA

NA

NA

NA

no d

ata

repo

rted

Euro

peN

AN

AN

AN

AN

AN

Ano

dat

a re

port

edJa

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

port

ed

Dri

nkin

g w

ater

(µg

L–1)

Can

ada

0.53

0.5

11N

AN

AN

A12

76da

ta p

rim

arily

from

Alb

erta

Env

iron

men

t [8]

,EC

&H

C [1

6] a

nd H

C [1

7]ru

ral

0.53

0.5

9N

A0.

5N

A62

5m

edia

n=

0.5;

repr

esen

ts 1

/2th

e de

tect

ion

limit

urba

n0.

540.

511

NA

0.5

NA

657

med

ian

=0.

5;re

pres

ents

1/2

the

dete

ctio

n lim

itU

SAN

AN

AN

AN

AN

AN

Ano

dat

a re

port

edEu

rope

NA

NA

NA

NA

NA

NA

no d

ata

repo

rted

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

repo

rted

Sedi

men

t (m

g kg

–1)

Can

ada

15<

0.01

515

NA

NA

NA

1U

SAN

AN

AN

AN

AN

AN

Ano

dat

a re

port

edEu

rope

NA

NA

NA

NA

NA

NA

no d

ata

repo

rted

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

repo

rted

Observed Concentrations in the Environment 163

Tabl

e7

(con

tinu

ed)

Reg

ion

Med

ium

Con

cent

rati

onN

o.D

ata

Com

men

tsPo

ints

b

Mea

nM

ini-

Max

i-10

thM

edia

na

90th

mum

mum

Per-

Per-

cent

ilea

cent

ilea

Air

(ng

m–3

)C

anad

aN

AN

AN

AN

AN

AN

Ano

dat

a re

port

edU

SAN

AN

AN

AN

AN

AN

A1

refe

renc

e fo

r fly

ash

only

Euro

peN

AN

AN

AN

AN

AN

Ano

dat

a re

port

edJa

pan/

Asi

aN

AN

AN

AN

AN

AN

Ano

dat

a re

port

ed

Oth

er m

edia

Slud

ge (µ

g kg

–1)

prim

ary

sew

age

slud

geC

anad

a54

20<

1063

,000

NA

NA

NA

79U

SAN

AN

AN

AN

AN

AN

Ano

dat

a re

port

edEu

rope

NA

NA

NA

NA

NA

NA

no d

ata

repo

rted

Japa

n/A

sia

NA

NA

NA

NA

NA

NA

no d

ata

repo

rted

aT

he 1

0th

perc

enti

le,m

edia

n,an

d 90

th p

erce

ntile

val

ues

wer

e ca

lcul

ated

by

usin

g th

e m

eans

oft

he s

tudi

es,n

ot th

e in

divi

dual

dat

a po

ints

.b

The

num

ber

ofda

ta p

oint

s re

pres

ents

the

num

ber

ofda

ta p

oint

s av

aila

ble

wit

h su

ffic

ient

info

rmat

ion

to c

alcu

late

an

arit

hmet

ic m

ean.

cN

A n

ot a

vaila

ble.

8.2Sediment

Environment Canada and Health Canada [16] includes three references to theanalyses of DnOP in sediments. One of the references reports data collected0.5 km and 1 km downstream of a sewage outfall. The data range between<0.015 mg kg–1 and 15 mg kg–1.

8.3Soil

No data have been identified.

8.4Air

The unquantified presence of DnOP in air in Europe is reported [17]. In addition,a study of air quality from a coal-fired power station in the United States reportslevels of DnOP of 517,000 ng m–3 and 751,000 ng m–3 [17].

8.5Dust

No data have been identified.

8.6Food

Concentrations of 0.081 µg g–1 and 0.11 µg g–1 are reported in two out of ten fishsamples; however, the results are considered suspect due to possible sources ofcontamination in sampling and analysis [16].

8.7Other Media

8.7.1Sludge

Webber and Nichols [11] evaluated 72 samples of sewage sludge analyzed from12 locations across Canada.Webber and Lesage [10] reported seven sludge sam-ples analyzed for DnOP. The measured concentrations of DnOP in sludge rangebetween the detection limit (assumed to equal 10 µg kg–1) and 63,000 µg kg–1 withan overall calculated mean of 5420 µg kg–1.

164 K. Clark et al.

8.7.2Wastewater

As noted for surface water, the industrial surface water data reported by AlbertaEnvironment [9] range between 0.1 µg L–1 and 10 µg L–1 with a calculated meanconcentration of 5.5 µg L–1.

9Discussion and Recommendations

An in-depth analysis and discussion of the DEHP concentration ranges in Europeis presented here. DEHP is selected for this analysis as there are more monitor-ing data available for DEHP than for any other phthalate ester. The analysis canbe applied to other phthalate esters and different geographical regions, where suf-ficient monitoring data are available.

9.1Histograms

Four types of histograms are plotted to show the distributions of concentrationsof DEHP in air, surface water, and sediment in Europe, that is, unweighted fre-quency, weighted frequency, unweighted cumulative, and weighted cumulative.Thus, in total, 12 histograms are plotted for DEHP (see Figs. 1–6). The locationof the calculated weighted average is shown on each histogram. Where possible,data corresponding to a laboratory detection limit are identified.

The weighting of study averages to the number of samples in the study has alarge effect on the distributions in the histograms and can dictate the magnitudeof the calculated weighted average. Furthermore, as the weighted average is anarithmetic average, a few high concentrations can bias the concentrations up-wards. This is evidenced by the cumulative histograms, which show that the 50thpercentile is usually at a lower concentration than the weighted average concen-tration. If all of the raw data from every study were available (and not just thestudy arithmetic mean and number of samples), the preferred approach wouldbe to calculate a geometric mean (i.e., a mean of the logarithms of each reportedconcentration). This approach would help to reduce the effect of outliers on thecalculated average and would make weighting unnecessary.

The histograms show that DEHP concentrations are characterized as extend-ing over a very wide range, which makes interpretation and averaging difficult.For example, water concentrations vary from 0.003 to 10 µg L–1, that is, by a fac-tor of 3000. We believe that this wide distribution is due to variable proximity tosources, that is, the concentrations on the right-hand side of the histograms areprobably representative of impacted sites and those on the left-hand side arelikely to be representative of regional background environmental concentrations.Separation of the monitoring data into groups based on proximity to sourceswould require going back to the source references and allocating concentrationsto different classifications such as “urban/industrial”or “rural/background”. Evenafter careful scrutiny of source references, separation of the data may not be pos-

Observed Concentrations in the Environment 165

166 K. Clark et al.

Fig.

1.Fr

eque

ncy

hist

ogra

m o

fthe

dis

trib

utio

n of

DEH

P w

ater

con

cent

rati

ons

in E

urop

e fr

om p

ublis

hed

stud

ies

Observed Concentrations in the Environment 167

Fig.

2.C

umul

ativ

e hi

stog

ram

oft

he d

istr

ibut

ion

ofD

EHP

wat

er c

once

ntra

tion

s in

Eur

ope

from

pub

lishe

d st

udie

s

168 K. Clark et al.

Fig.

3.Fr

eque

ncy

hist

ogra

m o

fthe

dis

trib

utio

n of

DEH

P se

dim

ent c

once

ntra

tion

s in

Eur

ope

from

pub

lishe

d st

udie

s

Observed Concentrations in the Environment 169

Fig.

4.C

umul

ativ

e hi

stog

ram

oft

he d

istr

ibut

ion

ofD

EHP

sedi

men

t con

cent

rati

ons

in E

urop

e fr

om p

ublis

hed

stud

ies

170 K. Clark et al.

Fig.

5.Fr

eque

ncy

hist

ogra

m o

fthe

dis

trib

utio

n of

DEH

P ai

r co

ncen

trat

ions

in E

urop

e fr

om p

ublis

hed

stud

ies

Observed Concentrations in the Environment 171

Fig.

6.C

umul

ativ

e hi

stog

ram

oft

he d

istr

ibut

ion

ofD

EHP

air

conc

entr

atio

ns in

Eur

ope

from

pub

lishe

d st

udie

s

sible. Therefore, we have sought to interpret the monitoring data based on the cumulative histograms as follows:

– One third represent “remote from source” samples.– One third represent “close to source” samples.– One third represent intermediate samples.

Based on the cumulative histograms for DEHP (Figs. 2, 4, and 6), concentrationranges and average values are assigned to the first two categories for surface waters, sediments, and air in Europe. The corresponding fugacities are also calculated. It is not possible to undertake a similar analysis for soil because of thelack of monitoring data available.

For water, 90% of the reported data lie between 0.01 and 3.00 µg L–1 with aweighted average of 0.93 µg L–1. We interpret the distribution to indicate that, inwater which is not subject to direct discharge of DEHP, the concentration prob-ably lies in the range 0.01–0.10 µg L–1, that is, typical values of 0.03 µg L–1 plus orminus a factor of three. In waters that are in the vicinity of discharges, concen-trations probably lie in the range 0.3–3.0 µg L–1 with a typical value of 1.0 µg L–1,plus or minus a factor of three. The corresponding fugacities of these averages are 12 and 390 nPa for the “remote from source” and “close to source” samples,respectively.

For sediments, 90% of the reported data lie between 0.001 and 10 mg kg–1 witha weighted average of 4.9 mg kg–1. We interpret the distribution to indicate that,in water that is not subject to direct discharge of DEHP, the concentration prob-ably lies in the range 0.001–0.1 mg kg–1, that is, typical values of 0.01 mg kg–1 plusor minus a factor of ten. In sediments that are in the vicinity of discharges,concentrations probably lie in the range 1.0–10 mg kg–1 with a typical value of3.0 mg kg–1, plus or minus a factor of three. The corresponding fugacities of theseaverages are 0.12 and 37 nPa for the “remote from source” and “close to source”samples, respectively.

For air, 90% of the reported data lie between 0.3 and 100 ng m–3 with aweighted average of 21.9 ng m–3. We interpret the distribution to indicate that,in air which is not subject to direct discharge of DEHP, the concentration prob-ably lies in the range 0.3–3.0 ng m–3, that is, typical values of 1.0 ng m–3 plus orminus a factor of three. In air that is in the vicinity of discharges, concentrationsprobably lie in the range 10–100 ng m–3 with a typical value of 30 ng m–3, plus orminus a factor of three. The corresponding fugacities of these averages are3.7 and 110 nPa for the “remote from source” and “close to source” samples,respectively.

9.2Fugacities in Various Media

A plot of DEHP and DBP fugacities for a range of environmental media is shownin Fig. 7. The ranges of the reported environmental concentrations in the Euro-pean dataset are represented by the bars on Fig. 7.

Fugacities of DEHP in different media decrease by a factor of about1000 from left to right on the plot, from air to fish. The fugacities of DEHP in

172 K. Clark et al.

Observed Concentrations in the Environment 173

Fig. 7. Estimated fugacities of DEHP and DBP in environmental media

Fig. 8. Comparison of food chain biomagnification for two contrasting organic compounds

fish, cow’s milk, and human breast milk are considerably lower than in other media. We hypothesize that this decrease in fugacity represents a biodilution effect, as DEHP is transferred through aquatic and terrestrial food chains. As discussed by McLachlan [19], three factors combine to cause these biodilution effects:

1. Chemical degradation.2. Reduced absorption of highly hydrophobic compounds in the digestive tract.3. Kinetically limited uptake of involatile, hydrophobic compounds by plants.

It is believed that for DEHP, all three factors are likely to have some contributionto the observed biodilution effects.

McLachlan [19] focused on the air/soil-plant-cow-milk-human food chain,which is important for human exposure to polychlorinated biphenyls (PCBs) andpolychlorinated dibenzo-p-dioxins and furans (PCDD/Fs), as well as the phtha-late esters. In Fig. 8, the fugacities of DEHP in European air, soil, plants, cow’smilk, and human milk are compared to fugacities reported by McLachlan [19] forhexachlorobenzene (HCB), in samples collected in Germany. This provides an in-teresting comparison of different chemicals in the same food chain. The fugaci-ties of HCB clearly increase by a factor of 300 (or 2.5 log units) between air andhuman breast milk indicating strong biomagnification. In contrast, DEHP biodi-lutes by a factor of 1000 (or 3.0 log units). There is strong biodilution from air toplant (a factor of 13) and from plant to cow (a factor of 32). The drop in fugacityfrom air to plant is probably the result of the third factor listed above: kineticallylimited uptake by plants. The drop in fugacity from plant to cow may be a com-bination of the first and second factors: degradation and reduced absorption inthe digestive tract. McLachlan [19] found that the 2,3,7,8-substituted polychlo-rinated dibenzo-p-dioxins and furans, which have similar hydrophobicities (KOWof 106 –108) to DEHP (KOW of 107.65), but are highly persistent, may biodilute bya factor of 10–100 between air and human breast milk. This is lower than thebiodilution effect observed for DEHP of about 1000 and suggests that for DEHPdegradation (i.e., biotransformation) contributes to the biodilution effect(Fig. 9).

In Fig. 7, the fugacities of DEHP and DBP in the same media are compared.This is the only meaningful comparison that we can make with another phthalate ester because of a lack of comparable quality data for other phthalates.DBP fugacities only drop by a factor of 10 from air to fish and are thus close toequifugacity in all media. Because DBP has a lower hydrophobicity than DEHP,its moderate biodilution is likely to be mediated by metabolism only.

The observed biodilution of DEHP must be regarded as fortuitous because itreduces human exposure, which would be much greater if the food supply werein equilibrium with the air. Furthermore, the lack of DEHP biomagnification willlimit the exposure of upper trophic level mammals to the monoethylhexyl ester(MEHP), a metabolite of DEHP, and the putative toxic species in mammals.MEHP is formed by enzymatic hydrolysis of DEHP in the intestine and liver oforganisms. MEHP is water-soluble and will not bioaccumulate or biomagnify inthe food chain, but will instead be excreted (or conjugated and excreted) in theurine of the organism in which it is formed.

174 K. Clark et al.

This approach of plotting fugacity as a function of medium can, we believe,provide interesting insights into chemical fate. As discussed in the chapter byCousins and Mackay entitled “multimedia mass balance modeling of two phthalate esters by the regional population-based model (RPM)”, for DEHP andDBP, emissions to air represent more than 90% of the total emissions to the en-vironment. A hydrophobic substance that is less persistent (i.e., more reactive)will display a decreasing fugacity from source to sink because the rate of loss in“sink” media is sufficient to reduce fugacities, since there is insufficient time forthe solute to diffuse in adequate quantities to equalize the fugacity. A fugacitygradient approach of this type thus indicates two features of the chemical be-havior:

1. It is likely that the highest fugacity medium is the primary source and the low-est fugacity media tend to be sinks subject to transport from the source.

2. The downward source to sink slope is an indication of reactivity, that is, a highnegative slope corresponds to high reactivity and short half-lives or persis-tence.

9.3Conclusions

For all phthalates, there is a considerable amount of data reporting measuredconcentrations in surface waters and sediments in Europe and the United States.However, there is a lack of outdoor air data available and a severe paucity of datareporting measured concentrations in soil. Given the estimated importance of

Observed Concentrations in the Environment 175

Fig. 9. Comparison of the biodilution of two compounds with similar hydrophobicities but dif-ferent metabolism rates

soil in terms of environmental fate (i.e., it is the primary medium of accumula-tion for phthalate esters), additional measurements in soil would be helpful formodel verification.

Concentrations of phthalate esters are characterized by a very wide range,which makes interpretation and averaging difficult.We believe that this wide dis-tribution is due to variable proximity to sources. Therefore, we have sought to in-terpret the monitoring data as follows: one third represent “remote from source”samples, one third represent “close to source”samples, and one third represent in-termediate samples.

The EUSES (European Union System for the Evaluation of Substance) soft-ware, which is the preferred tool in Europe for the risk assessment of chemicals,is designed to undertake a three-tier risk assessment, that is, local-scale, regional-scale, and continental-scale [23, 24]. This three-tier approach helps to determinerisks to humans and organisms in differently impacted regions. The EUSES mod-eling approach would fit quite well with the proposed separation of monitoringdata into different classifications; however, the approach is most successful whenlarge amounts of high quality data are available.

By transforming environmental concentrations into fugacities, it becomesclear that DEHP appreciably biodilutes as it is transferred through aquatic andterrestrial food chains. Three factors combine to cause these biodilution effects:chemical degradation, reduced absorption of highly hydrophobic compounds inthe digestive tract, and kinetically limited uptake of involatile, hydrophobic com-pounds by plants. However, degradation is thought to have the largest influenceon the observed biodilution of phthalate esters. DBP does not biodilute as ap-preciably.

Acknowledgement. We are grateful to the Environmental Research Task Group (ERTG) of thePhthalate Ester Panel of the American Chemistry Council (ACC) for funding this research.

10References

1. Exxon Mobil Biomedical Sciences Inc (1999) Compilation of data for five phthalate esters.East Millstone, NJ, Prepared for American Chemistry Council, Arlington, VA

2. RIC (1999) DBP risk assessment – Environmental sampling3. RIC (2000) RIVM monitoring program for milk, vegetation, cattle feed, fish and air. Sub-

project E2. ECPI\2000–07 and 2000–144. Alberti J, Brull U, Furtmann K, Braun G (2000) Occurrence of Phthalates in German sur-

face and wastewater. Presented at SETAC World Congress, Brighton5. Bruns-Weller E, Pfordt J (2000) Z Umweltchem Okotox 12:1256. VROM (1998) Risk assessment – Dibutylphthalate. Prepared for the Ministry of Housing,

Spatial Planning and the Environment (VROM). Prepared by the Netherlands Organiza-tion for Applied Scientific Research (TNO) and the National Institute of Public Health andthe Environment (RIVM). Draft, 17 November

7. National Chemicals Inspectorate (NCI) (2000) Risk assessment – Bis(2-ethylhexyl) phthalate. Sweden, Draft, May

8. Alberta Environment (1999) Data analysed for phthalates in drinking water in Alberta.Enforcement and Monitoring Division. Edmonton, AB, Canada

9. Alberta Environment (1999) Data analysed for phthalates in surface water in Alberta.Water Sciences Branch, Water Data Management Section. Edmonton, AB, Canada

176 K. Clark et al.

10. Webber MD, Lesage S (1989) Waste Manage Res 7 :6311. Webber MD, Nichols JA (1995) Organic and Metal contaminants in Canadian municipal

sludges and a sludge compost. Wastewater Technology Centre, Burlington, ON, February12. Environment Canada and Health Canada (2000) Canadian Environmental Protection Act,

Priority Substances List assessment report – Butylbenzyl phthalate. Ottawa, ON13. Environment Canada and Health Canada (1994) Canadian Environmental Protection Act,

Priority Substances List assessment report – Bis(2-ethylhexyl) phthalate. Ottawa, ON14. Environment Canada and Health Canada (1994) Canadian Environmental Protection Act,

Priority Substances List assessment report – Dibutyl phthalate. Ottawa, ON15. Health Canada (1994) Canadian Environmental Protection Act, Priority Substances List –

Supporting documentation health-related sections – Di-n-Butyl phthalate. Ottawa, ON16. Environment Canada and Health Canada (1993) Canadian Environmental Protection Act,

Priority Substances List assessment report – Di-n-octyl phthalate. Ottawa, ON17. Health Canada (1994) Canadian Environmental Protection Act, Priority Substances List –

Supporting documentation health-related sections – Di-n-octyl Phthalate. Ottawa, ON18. Clark K, Cousins I, Mackay D (2001) Multimedia modelling and exposure assessment for

phthalate esters – Observed concentrations in the environment. Prepared for AmericanChemistry Council

19. McLachlan MS (1996) Environ Sci Technol 30:25220. Mackay D (2001) Multimedia environmental models: The fugacity approach, 2nd edn.

Lewis Publishers, Boca Raton, FA21. Gruber L, Wolz G, Piringer O (1998) Deutsche Lebensmittel-Rundschau 94:17722. Cousins I, Mackay D (2000) Chemosphere 41:138923. Berding V, Schwartz S, Matthies M (1999) Environ Sci Pollut Res 6 :3724. European Chemical Bureau (EC) (1997) EUSES documentation – The European Union sys-

tem for the evaluation of substances. National Institute of Public Health and Environment(RIVM), the Netherlands, available from European Chemical Bureau (EC/DGXI), Ispra

Observed Concentrations in the Environment 177

© Springer-Verlag Berlin Heidelberg 2003

Multimedia Mass Balance Modelling of Two PhthalateEsters by the Regional Population-Based Model (RPM)

Ian T. Cousins · Donald Mackay

Canadian Environmental Modelling Centre, Environmental and Resource Studies,Trent University, Peterborough, Ontario, K9J 7B8, Canada. E-mail: dmackay@trentu.ca

Achieving an adequate understanding of the fate of commercial chemicals in the environmentis best demonstrated by assembling a comprehensive multimedia mass balance model of thechemical’s behaviour in a specified region. This approach is demonstrated for two phthalate es-ters, di-n-butyl phthalate (DBP) and di-2-ethylhexyl phthalate (DEHP), in an industrialized re-gion with similar characteristics (including population density) to the Netherlands and theEastern United States. To accomplish this task the relevant physical-chemical properties ofthese substances are compiled and emission rates are estimated on a per capita basis. The “re-gional population-based model” (RPM) is described and concentrations of the phthalate estersare estimated and compared with available monitoring data from Europe.A sensitivity and un-certainty analysis is also included. Comparison of model predictions and monitoring data sug-gests that the major uncertainties in the multimedia fate assessment are the degradation half-lives and emission rates, but despite this, most reported concentration ranges are within a factorof four of the median predicted concentrations. Improved agreement between predicted andobserved water, sediment and fish concentrations of DEHP is obtained by using a measuredvalue of KOC for the suspended particles in the water column. The overall persistence or resi-dence time of DBP and DEHP attributable to reaction only (not including advection) are esti-mated to be 25 and 47 days, respectively; thus historical accumulation of these chemicals overperiods of years or decades is unlikely. It is concluded that, in general, the model captures thekey processes that control the behaviour of these substances and predicts environmental con-centrations that are of the correct order of magnitude. Consequently, this analysis appears toprovide a reasonable quantitative description of the environmental fate of these chemicals.

Keywords. Phthalate ester, DBP, DEHP, Multimedia, Model, Mass balance

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

2 Regional Population-Based Model (RPM) . . . . . . . . . . . . . 181

3 Chemical Input Data . . . . . . . . . . . . . . . . . . . . . . . . . 185

4 Environmental Emissions . . . . . . . . . . . . . . . . . . . . . . 186

5 Monitoring Data Used for Model Evaluation . . . . . . . . . . . . 190

6 Model Results and Discussion . . . . . . . . . . . . . . . . . . . . 191

7 Conclusions and Recommendations . . . . . . . . . . . . . . . . 199

8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 179–200DOI 10.1007/b11466

1Introduction

The objective of this chapter is to demonstrate how a regional multimedia massbalance model can be successfully used to reconcile observed environmental con-centrations of two commercial phthalate esters with reported emission rates. Asatisfactory reconciliation has several benefits indicating that emissions are fullyaccounted for, and that the dominant fate processes (including degradation rates)are in accord with model predictions. Furthermore, once a reasonable modelevaluation has been demonstrated, future changes in emission scenarios can betranslated into projected changes in environmental concentrations. It can also,with appropriate modification, be applied to other phthalate esters, and indeedto other chemicals. In general, the availability of a validated mass balance modelprovides confidence that the fate of chemicals in the environment can be pre-dicted, given sufficiently accurate data on their properties and emissions. To ac-complish this objective, a model is developed that is regarded as intermediate be-tween the use of a region-specific model such as CalTOX [1] or ChemCAN 4.0 [2]applied to a specific region, and a purely evaluative model such as EQC [3]. Thisapproach is similar to that taken by the SimpleBox 1.0 regional model [4], whichis designed to represent a typical region of the European Union (EU). Simple-Box 1.0 is used in the EUSES (European Union System for the Evaluation of Sub-stances) software, which is the preferred modelling tool for risk assessments ofnew and existing substances, performed by EU Member States [5]. The modelpresented here, the “regional population-based model” (RPM), is parameterisedto represent an industrially developed region, typical of Western Europe or theEastern United States. The aim is not to validate the model against actual con-centration data from a specific geographical region, rather it is to reconcile modelpredicted concentrations with observed values from regions of similar popula-tion density. This approach has the advantage that it permits a larger database ofconcentrations to be used. This model is regarded as being particularly suitablefor substances which are consumed on a fairly constant per capita basis such ascomponents of plastics and “down the drain”chemicals. The model was evaluatedby using European emission and monitoring data, mainly because of data avail-ability.

This chapter is structured into five sections. First, the RPM model is described.Second, the required physical-chemical property and reactivity model inputs aresummarized, drawing on information presented in the physical-chemical prop-erty chapter. Model calculations are presented for two phthalate esters, di-n-butylphthalate (DBP) and di(2-ethylhexyl) phthalate (DEHP), for which we have themost extensive and highest quality chemical input data available. Moreover, forthese two substances, observed environmental concentrations are consequentlyabove detection limits thereby allowing a thorough model evaluation. Third, percapita emission estimates taken from Parkerton et al. [6] are used to calculateemissions to the model region. Fourth, the monitoring data used for comparisonwith the model predictions are reviewed. Finally, RPM-predicted concentrationsare compared with environmental measurements, and sensitivity and uncer-tainty analysis is used to identify the most sensitive and important input data.

180 I.T. Cousins and D. Mackay

2Regional Population-Based Model (RPM)

RPM is used to evaluate the environmental fate of DBP and DEHP by relatingtheir emissions on a per capita basis to observed environmental concentrations.The region modelled is designed to be representative of densely populated, in-dustrialized areas. Therefore, the first task of model development was to identifya variety of industrialized regions and collect data for surface area, population,population density and surface water coverage (Table 1). Experience with de-signing model environments has shown that a region should not be too large,since this results in excessive “dilution”, or too small, since losses by advection inair leaving the region become relatively too important. The EQC model [3], de-veloped as a tool for evaluating chemical fate, has a surface area of 100,000 km2,which is about the size of the US state of Ohio or England, Greece or Portugal.Wedecided that this was a reasonable geographic area for the RPM model. Popula-tion densities in the selected industrialized regions are 101–419 people km–2

(Table 1). Japan, New Jersey and the heartlands of industrial Europe have thehighest population densities of 300–450 people km–2. It is noteworthy that withinthese larger regions there will be population “hotspots” in cities in which popu-lation densities can be as high as 1000–5000 people km–2, and low populationsin rural areas of less than 10 people km–2. The population density selected islargely dependent on the region from which the monitoring data used for modelevaluation are collected. We decided to use a population density of 400 peoplekm–2 (comparable to the Netherlands or New Jersey) for the RPM, which equatesto a population for the model environment of 40 million people. This populationdensity is higher than average, but the most reliable and comprehensive fieldmonitoring data that are presently available for comparison to model predictionsare available for this characteristic region.

RPM is a Level III, steady-state model that treats five bulk compartments: air,surface water, bottom sediments, surface soil and terrestrial vegetation. Fig-ure 1 summarizes the model compartments, transport processes and reactionlosses.

Multimedia Mass Balance Modelling of Two Phthalate Esters 181

Table 1. Surface area, population and percent water cover of industrialized regions

Country/Province/ Area Population Pop. density % WaterState (km2) (per km2)

New Jersey 20,000 8,314,000 412 2.8New York 128,000 17,783,000 138 0.7Ohio 107,000 10,744,000 101 1.4Pennsylvania 117,000 11,853,000 101 2.5UK 245,000 57,592,000 235 1.3France 547,000 58,609,000 107 0.3Germany 357,000 82,072,000 230 2.1Italy 301,000 56,831,000 189 2.4Netherlands 37,000 15,650,000 419 9.1Belgium 31,000 10,165,000 333 0.9Japan 378,000 125,733,000 333 0.8

A mass balance equation is written for each compartment and the set of equa-tions is solved algebraically. The equations are written in fugacity format andsolved for the fugacities in each of the bulk media (Table 2). In fugacity format,the chemical concentration in a medium (C, mol m–3) equals the fugacity ofchemical in the medium (f, Pa) multiplied by the fugacity capacity or Z value (Z,mol m–3 Pa–1) of the medium for that chemical; thus, C = Zf. The Z value quanti-fies the relative affinity of the chemical for a particular medium. The processesthat govern chemical movement and reaction are quantified by D values (D, molh–1 Pa–1). The D values, in conjunction with the chemical fugacity in a medium,are used to calculate the chemical flux (N, mol h–1) that is either transported orreacted. A full description of the formulation of the Z and D values is availablein a series of articles by Mackay et al. [3, 7–9]. The method of including vegeta-tion as an additional compartment in the mass balance is discussed in detail byCousins and Mackay [10]. It was deemed necessary to include a vegetation com-partment because it is expected that phthalate esters will partition appreciablyto foliage because of their high octanol-air partition coefficients (see the chap-ter discussing physical-chemical properties). Environmental characteristics ofthe RPM model region are given in Table 3. Mass transfer coefficients are listedin Table 4. A notable difference between RPM and the EQC model [3] is the

182 I.T. Cousins and D. Mackay

Fig. 1. Schematic diagram of the RPM model showing transport and reaction processes

assumed percentage of surface area covered by freshwater. For RPM it is 3%,chosen to be representative of an area drained by rivers, and not containing largelakes (see Table 1). The EQC model has a large water surface area of 10%, whichis chosen to account for the effect of large lakes and coastal waters on chemicalfate. The 3% surface area selected for RPM is the same as in SimpleBox [4].

Surface water predictions obtained from the RPM model were used to estimateconcentrations in fish using the equation:

CFISH (mg/kg lipid) = BCF ¥ CWATER (mg/L) ¥ FDISSOLVED ¥ frLIPID (a)

where BCF is the measured bioconcentration factor for fish, CWATER is the total wa-ter concentration, FDISSOLVED is the fraction of chemical in the dissolved phase,which is calculated (as explained previously in the chapter discussing physical-chemical properties), and frLIPID is the mass fraction of lipid in the fish (kg lipidkg–1 fish), which is assumed to be 0.05 (or 5%).

In addition to choosing single ‘preferred’ values for the model input parame-ters, a range of values has been selected to give a more realistic expression of un-certainty in the output. Uncertainty and sensitivity analysis of the model outputto all model input parameters (most of which are listed in Tables 3–6) were car-ried out by simultaneously varying them by using Crystal Ball 4.0* (Deci-

Multimedia Mass Balance Modelling of Two Phthalate Esters 183

Table 2. Steady-state mass balance equations used in RPM and their algebraic solution

Air f1 DT1 = I1 + f2 D2,1 + f4 D4,1 + f5 D5,1Water f2 DT2 = I2 + f1 D1,2 + f3 D3,2 + f5 D5,2Sediment f3 DT3 = I3 + f2 D2,3Vegetation f4 DT4 = I4 + f1 D1,4 + f5 D5,4Soil compartment f5 DT5 = I5 + f1 D1,5 + f4 D4,5

where f1 to f5 are the fugacities for each compartment (Pa), DT1 to DT5 are the sum of all theD-processes that transport chemical out of each compartment (reaction and advection losses)(mol Pa–1 h–1), I1 to I5 are the inputs (emissions and chemical flowing into the model in back-ground air or water) to each compartment (mol h–1) and the D values on the right hand sideof the equations are intercompartment D values (mol Pa–1 h–1)

Solutionf1 = (IAEV + f2 D2,1)/DTC1

f2 = IT/(DT2 –D3,2 D2,3/DT3 – D2,1/DTC1 (D1,2 + DTC2D5,2/DTC3))f3 = I3/DT3 + f2D2,3/DT3

f4 = (I4 + I5 D5,4/DT5 + f1(D1,4 + D5,4D1,5/DT5))/(DT4 – D4,5D5,4/DT5)f5 = (I5 + I4 D4,5/DT4 + f1 (D1,5 + D4,5 D1,4/DT4))/(DT5 – D5,4D4,5/DT4)

where

IT = I2 + I3 D3,2/DT3 + I4 D4,5 D5,2/(DT4DTC5) + I5 D5,2/DTC5 + IAEV DTC2/DTC1

IAEV = I1 + I4 D4,1/DTC4 + I4 D4,5 D5,1/(DTC5DT4) + I5 D5,1/DTC5 + I5 D5,4 D4,1/(DTC4DT5)DTC1 = DT1 –D4,1 (D1,4 + D5,4 D1,5/DT5)/DTC4 – D5,1 (D1,5 + D4,5 D1,4/DT4)/DTC5

DTC2 = D1,2 + (D5,2/DTC5) (D1,5 + D4,5 D1,4 DT4)DTC4 = DT4 – D4,5 D5,4/DT5

DTC5 = DT5 – D5,4 D4,5/DT4

sioneering, Colorado, USA). Lognormal distributions were assumed for all inputvariables to reduce skewness and were parameterised by the median (m) andstandard deviation (s) of the corresponding normal distribution on a logscale. 95% confidence factors (Cf) were used as convenient expressions ofvariance where s is equal to 0.5 ln Cf for a lognormal distribution. A confidencefactor of n implies that 95% of the data will be between n times and 1/n of themedian. Input variables were sampled from their 95% confidence factors shownin these tables for 1000 simulations. With this many simulations, the standard error of estimate for the median was always less than 5% of the standard devia-tion, which was considered adequate.

184 I.T. Cousins and D. Mackay

Table 3. Environmental input parameters used in RPM

Parameter Median Unit Confidence value factor a

Total area of region 1011 m2 –Water surface area 3 ¥ 109 m2 5Environmental temperature 12 °C 1.5Water depth 3 m 3Air mixing height 1000 m 3Soil mixing depth 0.05 m 3Sediment mixing depth 0.03 m 3Percent terrestrial surface covered by vegetation 80 % 3Leaf area index 4.0 m2 m–2 3Plant biomass 1.0 kg m–3 3Fraction of rain intercepted by foliage 0.1 – 3Rain scavenging ratio 200,000 – 3Volume fraction of particles in air 2 ¥ 10–11 – 3Volume fraction of particles in water 10–5 – 3Volume fraction of fish in water 10–6 – 3Volume fraction of air in soil 0.2 – 1.5Volume fraction of water in soil 0.3 – 1.5Volume fraction of soil solids 0.5 – 1.5Volume fraction of sediment pore water 0.8 – 1.5Volume fraction of sediment solids 0.2 – 1.5Volume fraction of water in vegetation 0.8 – 1.5Organic carbon fraction of particles in water 0.1 – 3Organic carbon fraction of soil solids 0.02 – 3Organic carbon fraction of sediment solids 0.05 – 3Lipid fraction of vegetation 0.01 – 3Residence time of air 4.17 days 3Residence time of water 41.67 days 3Residence time of annual vegetation cycle 180 days 3Vegetation density 950 kg m–3 1.5Water and aquatic biota density 1000 kg m–3 1.5Soil solids, sediment solids, atmospheric aerosol 1500 kg m–3 1.5

and water particle density

a 95% confidence factors (Cf) were used as convenient expressions of variance where the stan-dard deviation (s) is equal to 0.5 ln Cf (or Cf = e2 s) for a lognormal distribution. A Cf of nimplies that 95% of the data will be between n times and 1/n of the median value.

Crystal Ball™ was used to provide both a sensitivity analysis and a con-tribution to variance analysis. In a classical sensitivity analysis the objective is to determine the change in an output parameter that results from a fixedchange in each input parameter, which can be illustrated with the following equation:

S = (DO/O)/(DI/I) (b)

where S is the sensitivity, O and I are the best estimates for input and output pa-rameters and DO and DI are changes in the output and input parameters. Thecontribution to variance analysis also assesses the change in an output param-eter that results from changes in the input parameters. However, the change ineach input parameter is not fixed, since it varies according to the confidence factors that have been assigned to each input parameter. The application ofsensitivity and uncertainty analysis to multimedia models is discussed in detailby MacLeod et al. [11].

3Chemical Input Data

The key chemical input data are vapour pressures, solubilities in water, octanol-water partition coefficients (KOW) and reaction half-lives in air, water, soil,sediment and vegetation. Phthalates are in the liquid state at environmental temperature ranges so correction of solid vapour pressures to sub-cooled

Multimedia Mass Balance Modelling of Two Phthalate Esters 185

Table 4. Mass transfer coefficients used in RPM

Mass Transfer Coefficient (MTC) Median value Confidence(m h–1) factor a

Air side air-water MTC 5.0 3Water side air-water MTC 0.05 3Rain rate 1 ¥ 10–4 3Aerosol deposition 10.8 3Soil-air phase diffusion MTC 0.02 3Soil water phase diffusion MTC 1 ¥ 10–5 3Soil air boundary layer MTC 5.0 3Sediment-water MTC 1.0 ¥ 10–4 3Sediment deposition 5.0 ¥ 10–7 3Sediment resuspension 2.0 ¥ 10–7 3Soil water runoff 5.0 ¥ 10–5 3Soil solids runoff 1.0 ¥ 10–8 3Sediment burial 3.0 ¥ 10–7 3Diffusion to stratosphere 0.01 3Leaching from soil 1.0 ¥ 10–5 3Air side air-vegetation MTC 5 3Vegetation water uptake velocity 8 ¥ 10–4 3

a 95% confidence factors (Cf) were used as convenient expressions of variance where the stan-dard deviation (s) is equal to 0.5lnCf (or Cf = e2s) for a lognormal distribution. A Cf of nimplies that 95% of the data will be between n times and 1/n of the median value.

liquid vapour pressures is not required. Physical-chemical properties and de-gradation half-lives have been recently reviewed by Staples et al. [12], Cousinsand Mackay [13] and in the chapter of this handbook discussing environ-mental degradation rates of phthalates. The recommended values presented in the last of these sources are used as inputs to RPM (see Table 5). An adjust-ment of vapour pressure and aqueous solubility to a mean annual tem-perature of a temperate geographical region of 12 °C has been included by usingClausius-Clapeyron equations [9]. The enthalpies of phase change for solution(DHSOL) and vaporization (DHVAP) were assumed to be 10,000 and 50,000 J mol–1,respectively. No temperature correction has been applied to KOW or the reac-tion half-lives.

There is a paucity of plant metabolism data on chemical reaction rates onplant surfaces and in plant tissues in the literature and certainly no data avail-able for phthalate esters. It would appear, however, that metabolism is viable in plant tissues because plant cells contain a complement of active enzymes [14]. Furthermore, photodegradation of compounds sorbed to the surface ofplants may be rapid as a result of the tendency of leaves to align to achieve maximum exposure to sunlight [14]. Here, it is assumed that the half-life of achemical in vegetation is intermediate in value – the same as the water half-life.Vegetation reaction rates are seldom measured, but for certain chemicals forwhich partitioning to vegetation is known to be important (e.g. the phthalate esters), measurement of degradation rates is a research need.

For the higher molecular weight phthalates, measured values of KOC are oftenmuch higher than predicted from KOW (see the physical-chemical property chap-ter). For example, the relationship proposed by Seth [15] (KOC = 0.35 KOW) cal-culates a KOC value for DEHP of 1.9¥107 L kg–1 for KOC, whereas measured val-ues of KOC for DEHP are much lower and of the order of 104 –106 L kg–1 [12]. Thisoverestimation of KOC is usually observed for suspended particles in the watercolumn. Suspended particles may collide inducing desorption, and equilibriumbetween the particles and the dissolved-phase may not have been achieved [16].The DEHP simulation has been run with both a measured (105 L kg–1) and a pre-dicted KOC (1.9 ¥ 107 L kg–1) for the suspended sediments in the water column.The KOC for the bottom sediments is fixed at the default predicted value of1.9 ¥ 107 L kg–1.

Mean fish bioconcentration factors for DBP and DEHP of 167 and 280 L kg–1

(wet fish) were taken from Staples et al. [12] and used to estimate regional con-centrations in fish, as described earlier.

4Environmental Emissions

Worldwide “production” and “consumption” data have recently been compiled by Parkerton et al. [6] in a comprehensive study on the environmental emissionsof phthalates during their life cycle (see Table 6). This study was used as the basis for estimating the environmental emissions used in the modelling cal-culations.“Production” indicates the amount of a substance that is annually man-ufactured within a region, whereas “consumption” reflects the amount of the

186 I.T. Cousins and D. Mackay

Multimedia Mass Balance Modelling of Two Phthalate Esters 187

Tabl

e5.

Esti

mat

ed c

hem

ical

pro

pert

ies

ofD

BP a

nd D

EHP

at 1

2°C

Phth

alat

e Ph

ysic

al-c

hem

ical

pro

pert

y A

ssum

ed r

eact

ion

half-

lives

(h)

este

r

S Wa

VP

aLo

gK

OW

DHSO

LDH

VAP

Air

W

ater

So

il Se

dim

ent

Vege

tati

on

(mg

L–1)

(Pa)

(J

mol

–1)

(Jm

ol–1

)

DBP

m

edia

n va

lue

8.24

1.

88¥

10–3

4.27

10

,000

50

,000

55

17

0 17

00

5500

170

conf

iden

ce fa

ctor

b2

21.

11.

51.

55

55

55

DEH

Pm

edia

n va

lue

2.07

¥10

–31.

00¥

10–5

7.73

10,0

0050

,000

1755

055

0055

0055

0co

nfid

ence

fact

orb

22

1.1

1.5

1.5

55

55

5

aS W

is th

e so

lubi

lity

ofth

e liq

uid

in w

ater

,VP

is th

e liq

uid

vapo

ur p

ress

ure.

b95

% c

onfid

ence

fact

ors

(Cf)

wer

e us

ed a

s co

nven

ient

exp

ress

ions

ofv

aria

nce

whe

re th

e st

anda

rd d

evia

tion

(s)

is e

qual

to 0

.5 ln

Cf(

or C

f=e2s

) fo

r a

logn

orm

al d

istr

ibut

ion.

A C

fofn

impl

ies

that

95%

oft

he d

ata

will

be

betw

een

nti

mes

and

1/n

ofth

e m

edia

n va

lue.

substance that is converted into end-use products. Differences between total production tonnage and consumption tonnage estimates reflect import/export of these commodity substances into or out of a region to meet worldwide demands. Phthalates can be released to the environment industrially from facil-ities that produce or process phthalates, and during consumption, as a result ofevaporation and possible leaching of phthalates from products during use.Despite the low volatility of the heavier phthalates, their vapour pressures are sufficient to cause evaporation over the long term. This is commonly observed asplastic materials become less flexible with age. In the report by Parkerton et al.[6], the estimated percentage losses of phthalate esters from each stage of theirlife cycle, from manufacture to waste disposal, have been estimated (reproducedin Table 6).

Table 7 gives estimated environmental emissions for DBP and DEHP to theRPM model region from different stages of the phthalate life cycle. First, the emis-sions to the whole EU were calculated by multiplying the estimated percentagelosses from each life stage reported [6] by the EU production tonnage (for lossesfrom production and industrial use and losses from transportation) or EU con-sumption tonnage (for losses from product end use and product disposal). Thetotal annual production in metric tonnages of DBP and DEHP in the EU are es-timated to be 37,000 and 595,000, respectively, whereas the consumption ton-nages are estimated to be 21,000 and 476,000, respectively [6]. Confidence factorsof 1.1 were applied to these production and consumption data. The estimated total emissions to the EU on a per capita basis were calculated by dividing the total estimated emissions for the region by the total population of the region (370 million for the EU, assumed to be accurate). The emissions to RPM were estimated by multiplying the per capita emissions for the EU by the assumedpopulation of RPM (40 million).

188 I.T. Cousins and D. Mackay

Table 6. Summary of emission factors to different environmental compartments at each stepin the life cycle for DBP and DEHP (these unitless fractions are multiplied by production andconsumption tonnages to estimate emissions)

Phthalate Life stage Air Water Soil Confidence ester factor a

DBP production and industrial use 0.5 0.25 0.05 3transportation 0 0.01 0 3product end use 23.5 0.095 0.132 3product disposal 0 0.008 0 3total 24.0 0.363 0.182 –

DEHP production and industrial use 0.25 0.015 0.005 3transportation 0 0.01 0 3product end use 1.0 0.029 0.065 3product disposal 0 0.002 0 3total 1.25 0.056 0.07 –

a 95% confidence factors (Cf) were used as convenient expressions of variance where the standard deviation (s) is equal to 0.5 ln Cf (or Cf = e2s) for a lognormal distribution. A Cf ofn implies that 95% of the data will be between n times and 1/n of the median value.

Multimedia Mass Balance Modelling of Two Phthalate Esters 189

Tabl

e7.

Esti

mat

ed e

nvir

onm

enta

l em

issi

ons

to th

e R

PM m

odel

reg

ion

(bas

ed o

n EU

em

issi

on d

ata)

Phth

alat

e es

ter

Life

sta

geEm

issi

on to

air

Em

issi

on to

wat

er

Emis

sion

to s

oil

Tota

l(k

g ye

ar–1

)(k

g ye

ar–1

)(k

g ye

ar–1

)(k

g ye

ar–1

)

DBP

prod

ucti

on a

nd in

dust

rial

use

20,9

5010

,475

2095

33,5

20tr

ansp

orta

tion

041

90

419

prod

uct e

nd u

se55

8,85

522

5931

3956

4,25

3pr

oduc

t dis

posa

l0

190

019

0to

tal e

mis

sion

579,

805

12,9

2452

3459

7,96

3ad

vect

ing

in fr

om o

utsi

de r

egio

n28

,009

710

–28

,700

emis

sion

and

adv

ecti

on

607,

814

13,6

3452

3462

6,68

2

DEH

Ppr

oduc

tion

and

indu

stri

al u

se16

8,44

910

,107

3369

181,

925

tran

spor

tati

on0

6738

067

38pr

oduc

t end

use

539,

038

15,6

3235

,037

589,

707

prod

uct d

ispo

sal

010

780

1078

tota

l em

issi

on70

7,48

733

,555

38,4

0677

9,44

8ad

vect

ing

in fr

om o

utsi

de r

egio

n11

,380

1420

–12

,800

emis

sion

and

adv

ecti

on

718,

867

34,9

7538

,406

792,

248

5Monitoring Data Used for Model Evaluation

Over the last few years, a regional field monitoring program has been conductedin the Netherlands to obtain concentration measurements of selected phthalatesin different environmental media for comparison to multimedia fate model pre-dictions. Samples of outdoor air, soil, sediment, vegetation and fish were collectedby the RIVM and analysed with state-of-the-art methodologies. These resultshave been reported by the two contract laboratories that performed these analy-ses (Alcontrol [17] and the Research Institute of Chromatography [18]). Regionalmonitoring data for surface water concentrations of DBP and DEHP were takenfrom the analyses of German rivers reported by Furtmann [19] and Alberti [20].In each of these studies, considerable precautions were taken to minimize theconfounding problems associated with laboratory contamination.

Statistical summaries of these data including median concentrations, and 10thand 90th percentile concentrations are given in Table 8.

The comprehensive monitoring database developed for the American Chem-istry Council (described in detail in the chapter of this handbook discussing en-vironmental concentrations) was used to estimate background concentrations inair and water that are advected into a region (see Table 5 for advection estimates).It was decided to take the 10th percentile concentrations for air and water fromthis monitoring database as boundary conditions for this modelling exercise (seeTable 3 for values). A confidence factor of three was applied to the air and waterbackground concentrations.

190 I.T. Cousins and D. Mackay

Table 8. Summary of monitoring data used for model evaluation

Phthalate Medium Regional European monitoring dataester

10th Median 90th percentile percentile

DBP air (ng m–3) 5 10.3 33.9water (µg L–1) 0.052 0.12 0.25fish (µg kg–1 fat) <50 449 1778soil (µg kg–1 dry) 5 12 30sediment (µg kg–1 dry) 34 95 160vegetation (µg kg–1 dry) 5 5 36

DEHP air (ng m–3) 1.0 16.7 85.2water (µg L–1) 0.087 0.46 1.3fish (µg kg–1 fat) <50 449 5566soil (µg kg–1 dry) 13 28 61sediment (µg kg–1 dry) 66 250 730vegetation (µg kg–1 dry) 82 224 484

6Model Results and Discussion

The model was run for steady-state conditions for DBP and DEHP; mass balancediagrams are given in Figs. 2 and 3.

The majority of environmental releases for all phthalates, including the rela-tively higher molecular weight DEHP, are to the air compartment. The diagramshows that DEHP partitions from air to vegetation, air to soil and water to sedi-ment and this is likely to be the pattern for other phthalates with long alkylchains. Furthermore, the lighter DBP also tends to partition out of the air andinto the soil, but does not partition as appreciably to bottom sediments. For DBPthe uptake by vegetation from the atmosphere is balanced by volatilisation fromthe plant surface and thus accumulation in vegetation is not as significant as for DEHP.

The predicted environmental concentrations are given in Table 9 with error orvariation limits (10th and 90th percentiles) corresponding to model uncertainty.Comparisons between fugacities calculated from monitoring data and fugacitiescalculated from model predicted concentrations are shown in Figs. 4 and 5. Forconsistency it was decided to use the same methods for converting concentra-tions into fugacities that were used in the chapter of this handbook discussingenvironmental concentrations of phthalates. Direct comparison between the con-

Multimedia Mass Balance Modelling of Two Phthalate Esters 191

Fig. 2. RPM steady-state mass balance for DBP (all fluxes in mol h–1)

centration values can be made by comparing Tables 8 and 9. The 10th and 90thpercentiles of the monitoring data are approximately a factor of three greater orless than the median predicted concentration. In general, predicted environ-mental concentrations (or fugacities) compare favourably with the monitoringdata, with most of the predictions being within an order of magnitude ofobserved data. The reduction in fugacity of DEHP from abiotic media, such as air and water, to biotic media, such as vegetation and fish, is thought to be due primarily to metabolism in the biotic media and possibly, in the case of fish,trophic dilution. These findings are consistent with observations made in thechapter of this handbook discussing environmental concentrations of phthalates.

For DBP, predicted concentrations in all media are within a factor of four ofobserved concentrations. Agreement within a factor of four is considered satis-factory in a regional mass balance modelling exercise in which there are oftenlarge uncertainties in chemical and environmental model input parameters. ForDEHP, air, soil, vegetation and sediment predicted concentrations are within afactor of three of the median concentrations from the two monitoring datasets.Water concentrations of DEHP, are underestimated by an order of magnitude ifan estimated KOC is used for the suspended particles in the water column. This suggests that the model does not accurately describe the partitioning and/ortransport between bottom sediments and the overlying water. However, the

192 I.T. Cousins and D. Mackay

Fig. 3. RPM steady-state mass balance for DEHP (all fluxes in mol h–1). (In this simulation ameasured KOC of 105 was used for the suspended particles in the water column)

Multimedia Mass Balance Modelling of Two Phthalate Esters 193

Tabl

e9.

RPM

pre

dict

ed e

nvir

onm

enta

l con

cent

rati

ons

Phth

alat

e es

ter

Stat

isti

cA

ir

Wat

er

Fish

So

il Se

dim

ent

Vege

tati

on

(ng

m–3

)(µ

g L–1

)(µ

g kg

–1fa

t)(µ

g kg

–1dr

y)(µ

g kg

–1dr

y)(µ

g kg

–1dr

y)

DBP

10th

per

cent

ile8.

40.

017

551.

14.

03.

4m

edia

n25

0.05

919

04.

231

1490

th p

erce

ntile

640.

2167

017

170

73

DEH

Pa

10th

per

cent

ile4.

60.

011

5.0

3.1

9523

med

ian

150.

037

1812

440

110

90th

per

cent

ile44

0.13

6344

2000

490

DEH

Pb

10th

per

cent

ile4.

50.

0418

02.

928

20m

edia

n15

0.12

650

1215

012

090

th p

erce

ntile

450.

4121

0048

950

810

aU

sing

an

esti

mat

ed K

OC

from

the

equa

tion

KO

C=

0.41

KO

W(0

.41

¥10

7.73

=2.

2¥

107

L kg

–1).

bU

sing

a m

easu

red

KO

Cof

105

L kg

–1.

reconciliation between observed and measured water concentrations improvesconsiderable, to within a factor of four, if a measured KOC value is used for suspended particles. The agreement between the predicted and observed fish and sediment concentrations is also much improved if the measured KOC valueis used.

A validation issue that arises with these, and indeed with all mass balancemodels is that there may be some cancellation of errors. For example, a concen-tration may be correctly predicted by using an over-estimate of emission rate andan over-estimate of reactivity (or correspondingly an under-estimate of half-life).The model provides a constraint on the relationship between these variables, forexample, in this case the product of emission rate and half-life is known. Thereare also constraints of “reasonableness” for each variable that tend to narrow therange of possible values. In the present situation the various parameter values se-lected are judged to be both reasonable and in accord with monitoring data. Wedo, however, preclude the possibility of some error cancellation.

As discussed earlier, sensitivity and contribution of variance analyses havebeen undertaken by using Crystal Ball™ (Figs. 6 and 7). For DBP, KOW is the mostsensitive model input parameter controlling the total environmental inventory,followed by the estimated consumption of DBP in Europe, the emission to airfrom product end use and the soil reaction half-life. It is noteworthy that the most sensitive model input parameters are chemical parameters rather than en-vironmental parameters. The contribution of variance analysis, which relies on

194 I.T. Cousins and D. Mackay

Fig. 4. Comparison of RPM predicted and observed median concentrations for DBP (error barsrepresent 10th and 90th percentiles)

our estimated confidence factors ascribed to input parameters, finds that the fourmost important input parameters controlling output variance of the total envi-ronmental inventory of DBP are: soil reaction rate, the emission factor to air fromproduct end use, the average air height and the air reaction rate. In both analy-ses for DBP, reaction rates and parameters associated with environmental emis-sions are shown to be the key input parameters controlling the predicted envi-ronmental inventory. Similar results were obtained for DEHP. The highsensitivity of KOW, solubility in water, vapour pressure and the enthalpy of phasechange for vaporization illustrate the importance of obtaining accurate physical-chemical properties.

It is worth noting that given the same model input data other environmentalmodels such as SimpleBox [1], ChemCAN [2] or CalTOX [4] will give similar pre-dictions as RPM. Recently, an intercomparison exercise between SimpleBox 1.0(used in EUSES) and a Level III fugacity model [21] yielded near-identical pre-dicted environmental concentrations of DEHP from the same set of input param-eters. The Level III fugacity model was calibrated by changing the bulk proper-ties of the environmental compartments and mass-transfer coefficients in themodel to those used in the EUSES regional environment. Equations used to de-scribe environmental partitioning, however, were not altered in the fugacitymodel. This agreement between different models is not surprising, since a widerange of multimedia models has previously been shown to give very similar pre-

Multimedia Mass Balance Modelling of Two Phthalate Esters 195

Fig. 5. Comparison of RPM predicted and observed median concentrations for DEHP (errorbars represent 10th and 90th percentiles)

196 I.T. Cousins and D. Mackay

Fig.

6.Se

nsit

ivit

y an

alys

is fo

r D

BP s

how

ing

the

perc

enta

ge c

ontr

ibut

ion

to o

utpu

t var

ianc

e re

sulti

ng fr

om a

fixe

d va

rian

ce in

eac

hin

put p

aram

eter

Multimedia Mass Balance Modelling of Two Phthalate Esters 197

Fig.

7.C

ontr

ibut

ion

to v

aria

nce

anal

ysis

for

DBP

sho

win

g th

e pe

rcen

tage

con

trib

utio

n to

out

put v

aria

nce

resu

lting

from

est

imat

edva

rian

ce in

eac

h in

put p

aram

eter

dictions when using the same chemical input data and a standardized environ-ment [22]. It is noteworthy that these models are based on a similar set of equa-tions describing partitioning and transport.

RPM predicts that when total emissions of DEHP to the environment are792,000 kg year–1 the total inventory in the environment is 83,000 kg.As a result,the residence time is 83,000/792,000 or 0.10 years (38 days). Likewise, the over-all residence time of DBP in the RPM model is 26,000/627,000 or 0.16 years(15 days). Losses due to reaction dominate over losses due to advection out of themodel environment. The estimated residence times of DBP and DEHP attribut-able to reaction only are 25 and 47 days, respectively. DEHP is estimated to residelonger in the environment than DBP because it has longer degradation half-livesin most environmental media and a greater proportion of the more hydropho-bic DEHP partitions to soils and sediments in which degradation tends to beslower. These results are consistent with the results of the evaluative fate model-ling presented in the physical-chemical property chapter.

Mathematically, the calculated residence time in the model is a “characteris-tic time” describing chemical dynamics in the system. A dynamic model underinitial conditions of zero concentrations, followed by sustained constant emis-sions would display an approach to a steady state that would be essentially 97%complete after three characteristic times. Thus, steady state would be achieved forDBP and DEHP after 45 days (0.12 years) and 114 days (0.31 years), respectively.Similarly, if emissions were stopped the system would be clear of the chemicalsalmost entirely in the same periods of time. The implication is that, since phtha-lates have been in use for many decades, they must have reached a steady-statecondition in the environment; thus, the use of a steady-state model is justified.The fact that emission rates and observed concentrations were reconciled witha steady-state model is further support for this assertion. Fears that conditionsmay be getting progressively worse as a result of accumulation of phthalates frompast discharges are unfounded. This hypothesis is supported by published mon-itoring data that have shown that concentrations of phthalate esters in surfacewater and sediments over the last decade (1990–2000) have remained constantand have reached a steady-state condition (see the chapter discussing environ-mental concentrations of phthalates for more details).

The validity of scaling emissions on a per capita basis should be better testedby applying the RPM model to a number of regions with different populationdensities. The predicted concentrations are linearly related to population; thus,if the population density is doubled, then so will the predicted concentrations.There are no other datasets that compare in quality to the multimedia moni-toring dataset used here for the model evaluation, but it is possible to make a few general observations from the data contained in the database described in the chapter of this handbook discussing the environmental concentrations ofphthalates. For example, concentrations of phthalates in Western Europe are2–3 times higher than concentrations in Northern Europe (i.e. the Scandinaviancountries). National average population densities, however, are greater than tentimes higher in Western European countries than in the Scandinavian countries.The problem with applying the RPM model to a region such as Scandinavia isthat the population is not evenly distributed. People are concentrated in urban

198 I.T. Cousins and D. Mackay

areas, and rural populations densities are low. The low transport potential ofphthalates is likely to create pronounced urban-rural gradients in environmen-tal concentrations and exposures in such regions. Environmental concentrationsin cities should be 2-3 times higher than in the RPM region because of the highpopulation densities found in urban areas (>1000 km–2). Alberta Environmentmeasured surface water concentrations of phthalates in urban and rural areas ofAlberta and they found that environmental concentrations in urban areas wereon average higher than in rural areas, but usually less than two times higher [23].Urban areas may act as sources to surrounding rural areas as the air and waterin cities will have fairly short residence times. This discussion suggests that although there is certainly a qualitative relationship between environmental concentrations of phthalate esters and population density, this relationship maynot be perfectly linear and requires further detailed investigation as more highquality multimedia datasets become available. It would appear that the RPMmodel is best applied to homogenously and densely populated regions in West-ern Europe and the Eastern United States.

7Conclusions and Recommendations

The RPM model has successfully reconciled the known properties of two phtha-late esters with the latest emission rate estimates and monitoring data. Reportedconcentration ranges are within a factor of four of the median predicted con-centrations. There is a caveat that there has been some cancellation of errorswithin the model to give a good agreement between predictions and monitoringdata, but such errors are believed to be relatively small in magnitude given theconstraints between the various parameters. The sensitivity and uncertaintyanalysis suggests that the major uncertainties are the degradation half-lives andemission rates. Improved agreement between predicted and observed water, sed-iment and fish concentrations of DEHP is obtained by using a measured value ofKOC for the suspended particles in the water column. The residence time of DBPand DEHP including advective losses are 15 and 38 days, respectively; thus,accumulation of these chemicals over periods of years and decades is unlikely.The residence times attributable to reaction only are 25 and 47 days, respectively.It is concluded that, in general, the model captures the key processes that controltheir behaviour and predicts concentrations of the correct order of magnitude.We are thus close to having an adequate quantitative description of the environ-mental fate of these chemicals.

The following are recommendations for further research:

– It would be advisable to experimentally determine the degradation rate inplants because a significant fraction of DEHP partitions to vegetation and po-tentially degrades there.

– RPM should be applied to other regions with different population densities.The further application of the RPM model requires that further high qualitymultimedia monitoring surveys are undertaken.

– More phthalate esters and other chemicals used on a per capita basis shouldbe modelled by the RPM model.

Multimedia Mass Balance Modelling of Two Phthalate Esters 199

Acknowledgement. We are grateful to the Phthalate Ester Panel of the American ChemistryCouncil (ACC) for funding this research, and to NSERC and the consortium of chemical com-panies that support the Canadian Environmental Modelling Centre.

8References

1. McKone TE (1993) CalTOX,A multi-media total-exposure model for hazardous waste sites,Part II: the dynamic multi-media transport and transformation model. Lawrence Liver-more National Laboratory, Livermore, CA, No UCRL-CR-111456 PtII

2. Mackay D, Paterson S, Tam DD (1991) Assessments of chemical fate in Canada: continueddevelopment of a fugacity model. A report prepared for Health and Welfare Canada

3. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Environ Toxicol Chem 15:16184. Meent D Van De (1993) SimpleBox: a generic multimedia fate evaluation model.

Bilthoven, National Institute of Public Health and Environmental Protection. Report No 672720001

5. European Chemical Bureau (ECB) (1997) EUSES documentation – the European Union Sys-tem for the Evaluation of Substances. National Institute of Public Health and Environment(RIVM), the Netherlands, available from European Chemicals Bureau (EC/DGXI), Ispra

6. Parkerton T, Konkel W (2000) Evaluation of the production, consumption, end use and potential emissions of phthalate esters. Prepared for the American Chemistry Council(Draft of November 2000) by Exxon Mobil Biomedical Sciences Inc (EMBSI), East Mill-stone, NJ, USA

7. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Environ Toxicol Chem 15:16278. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Environ Toxicol Chem 15:16389. Mackay D (2001) Multimedia models: the fugacity approach, 2nd edn. Lewis, Boca Raton

10. Cousins IT, Mackay D (2001) Chemosphere 44:64311. MacLeod M, Fraser A, Mackay D (2002) Environ Toxicol Chem 21:700–70912. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:66713. Cousins IT, Mackay D (2000) Chemosphere 41:138914. McFarlane JC (1995) In: Trapp S, McFarlane JC (eds) Plant contamination – modeling and

simulation of organic chemical processes. Lewis, Boca Raton, FL, pp 13–3415. Seth R, Mackay D, Muncke J (1999) Environ Sci Technol 33:239016. Williams MD, Adams WJ, Parkerton TF, Biddinger GR, Robillard KA (1995) Environ Toxi-

col Chem 14:147717. Alberti J, Brull U, Furtmann K, Braun G (2000) Occurrence of phthalates in German

surface and wastewater. Poster presentation at the 10th annual SETAC Europe meeting,Brighton, UK, 21–25 May

18. Alcontrol Biochem Laboratoria (1999) The analysis of phthalates in soil and sediment.Hoogvliet, The Netherlands

19. Furtmann K (1993) Phthalates in the aquatic environment. PhD dissertation, Regional Water and Waste Water Authority, Nordrhein-Westfalen, Düsseldorf, Germany

20. Research Institute of Chromatography (2001) Final report on analysis of selected phthalates in the Dutch environment, Report ECPI-2001-XX. Kortrik, Belgium

21. Cousins IT, Mackay D (2000) Review of EUSES modelling for di-2-ethylhexyl phthalate(DEHP). Final report prepared for the European Chemical Industry Council (CEFIC),April2001, CEMC Report No 200101

22. Cowan CE, Mackay D, Feijtel TCJ, van de Meent D, Di Guardo A, Davies J, Mackay N (1995)The multi-media model: a vital tool for predicting the fate of chemicals. Proceedings of aworkshop organized by the Society of Environmental Toxicology and Chemistry (SETAC).Based on an international task force which addressed the application of multi-media fatemodels to regulatory decision-making, held at Leuven, Belgium, April 14–16, 1994 andDenver, Colorado, November 4–5, 1994

23. Alberta Environment (1999) Data analysed for phthalates in surface water in Alberta.Water Sciences Branch, Water Data Management Section. Edmonton, AB, Canada

200 I.T. Cousins and D. Mackay

© Springer-Verlag Berlin Heidelberg 2003

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs

Frank A.P. C. Gobas 1 · Cheryl E. Mackintosh 1 · Glenys Webster 1

Michael Ikonomou 2 · Thomas F. Parkerton 3 · Kenneth Robillard 4

1 School of Resource and Environmental Management, Simon Fraser University, Burnaby,British Columbia, V5A 1S6, Canada. E-mail: gobas@sfu.ca

2 Department of Fisheries and Oceans, Contaminants Science Section, Institute of Ocean Sciences, Sidney, British Columbia, V8L 4B2, Canada

3 Exxon Mobil Biomedical Sciences, Inc., Machelen, Belgium4 Eastman Kodak Company, 1100 Ridgeway Ave., Rochester, NY 14652-6276, USA

This chapter explores the bioaccumulation behavior of several phthalate esters in aquatic food-webs. It includes: (i) a compilation of bioconcentration data from reported laboratory studiesin the literature, (ii) an overview and discussion of the results from a recently completed food-web bioaccumulation field study, and (iii) an analysis of the results of a bioaccumulation mod-eling study. The study concludes that laboratory and field studies indicate that phthalate estersdo not biomagnify in aquatic food-webs. Higher molecular weight phthalate esters (DEHP,DnOP, and DnNP) show evidence of trophic dilution in aquatic food-webs, which is consistentwith findings from laboratory and modeling studies which indicate that metabolic transfor-mation is a key mitigating factor. Bioaccumulation patterns of DBP, DiBP, and BBP indicate nosignificant relationship with trophic position consistent with a lipid-water partitioning model.The lowest molecular weight phthalate esters (DMP and DEP) show bioaccumulation factorsin laboratory and field studies that are greater than predicted from a lipid-water partitioningmodel. The considerable variability in the field-derived bioaccumulation factors (BAFs) forlower molecular weight phthalate esters across aquatic species suggests that species-specificdifferences in metabolic transformation can have a significant effect on observed bioaccumu-lation.With some exceptions discussed below, the bioconcentration and bioaccumulation fac-tors of the phthalate esters discussed in this paper are below the UNEP bioaccumulation cri-terion of 5000. The low bioavailability of the high-molecular weight phthalate esters in naturalwaters is the main reason why the BAFs of the higher molecular weight phthalate esters are be-low the UNEP bioaccumulation criterion. Since the intention of the bioaccumulation criteriais to identify substances as being “bioaccumulative”, if they (like PCBs) biomagnify in the food-web then current evidence supports the conclusion that phthalate esters do not appear to be“bioaccumulative”.

Keywords. Bioaccumulation, Biomagnification, Phthalate Esters, Aquatic Food-Webs, Fish

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202

2 Bioaccumulation Nomenclature . . . . . . . . . . . . . . . . . . . 203

3 Phthalate Ester Nomenclature . . . . . . . . . . . . . . . . . . . . 204

4 Bioconcentration Studies . . . . . . . . . . . . . . . . . . . . . . 204

5 Dietary Bioaccumulation Studies . . . . . . . . . . . . . . . . . . 207

6 Bioaccumulation Studies from Sediments . . . . . . . . . . . . . 208

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 201–225DOI 10.1007/b11467

7 Food-Web Bioaccumulation Studies . . . . . . . . . . . . . . . . . 208

7.1 Biota-Sediment-Accumulation . . . . . . . . . . . . . . . . . . . . 2087.2 Food-Web Bioaccumulation . . . . . . . . . . . . . . . . . . . . . 2127.3 BAFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

8 Bioaccumulation Models . . . . . . . . . . . . . . . . . . . . . . . 217

9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

Abbreviations

BAFs Bioaccumulation FactorsBBP Butylbenzyl PhthalateBCFs Bioconcentration FactorsBSAF Biota-Sediment Accumulation FactorDBP Di-n-Butyl PhthalateDEHP Di-2(Ethylhexyl) PhthalateDEP Diethyl PhthalateDiBP Diisobutyl PhthalateDMP Dimethyl PhthalateDnNP Di-n-Nonyl PhthalateDnOP Di-n-Octyl PhthalateLRTAP Long-Range Transboundary Air Pollution ProtocolMS-MS detection Mass Spectrometry-Mass Spectrometry detectionPCBs Polychlorinated BiphenylsPOPs Persistent Organic PollutantsQA/QC Quality Assurance/Quality ControlUNEP United Nations Environmental Program

1Introduction

Dialkyl phthalate esters are hydrophobic substances with octanol-water partitioncoefficients ranging between 101.61 for dimethyl phthalate esters to values ex-ceeding 108 for congeners like diundecyl phthalate ester and ditridecyl phthalateester [1]. Due to their hydrophobicity, phthalate esters are often believed to havea high potential to bioconcentrate and bioaccumulate in aquatic organisms.

The degree of bioaccumulation and the mechanism by which phthalate estersare absorbed and retained by aquatic organisms is of considerable importanceas the United Nations Environmental Program (UNEP) long-range transbound-ary air pollution protocol (LRTAP) on POPs as well as domestic legislation inCanada (Canadian Environmental Protection Act, 1999) and several countries [2]aim to eliminate substances from commerce that are bioaccumulative, persistent,and toxic. The bioaccumulation criterion identifies chemicals as “bioaccumula-

202 F.A.P.C. Gobas et al.

tive” if they exhibit bioaccumulation or bioconcentration factors (BAFs or BCFs)greater than 5000 in aquatic organisms. In absence of BAF or BCF data “bio-accumulative”substances are defined as compounds with octanol-water partitioncoefficients (KOWs) greater than 105. The intent of these legislative efforts is toidentify substances that biomagnify in aquatic food-webs. Biomagnification isthe process in which the lipid-normalized concentration of the chemical in-creases with each step in the food-web. The significance of biomagnification isthat organisms at the top of the food-web are exposed to chemical concentrationsthat are greater than those at lower trophic levels. The scientific rationale for thebioaccumulation criterion is based on findings for persistent organochlorines,such as PCBs and chlorobenzenes, which indicate that persistent substances bio-magnify in aquatic food-webs if laboratory-derived bioconcentration factors exceed approximately 5000 or the octanol-water partition coefficient of the sub-stance exceeds approximately 105. Several authors have suggested that the bioac-cumulation behavior of phthalate esters is not comparable to that of persistentorganochlorines such as PCBs. Laboratory studies, which in most cases involvedone particular phthalate ester (i.e., diethylhexyl phthalate ester), have pointed outthat the bioaccumulation factors of phthalate esters are typically less than ex-pected from their lipid-water partitioning properties [3]. Metabolism and a re-duced bioavailability of phthalate esters have been proposed to be the main fac-tors causing the lower than expected bioaccumulation factors of phthalate esters[3–8]. However, field studies to confirm this hypothesis have not previously beenreported.

In this chapter, we will explore the bioaccumulation behavior of several phthalate esters in aquatic food-webs. We will present a compilation of bioac-cumulation data from reported laboratory studies in the literature and from a recently completed bioaccumulation field study that we conducted in a marinefood-web. The objective of our analysis is to gain insights into the mechanismsof phthalate ester uptake, elimination, and bioaccumulation in aquatic food-webs. This information can be useful in assessing the bioaccumulative potentialof this group of ubiquitous and widely used substances relative to other chemi-cal classes.

2Bioaccumulation Nomenclature

Bioconcentration is defined as the process in which an aquatic organism achievesa concentration level that exceeds that in the surrounding water as a result of ex-posure of the organism via the respiratory surface and the skin [9]. Bioconcen-tration refers to a condition, usually achieved under laboratory conditions, inwhich the organism is exposed to a chemical substance in the water, but not inits diet. The underlying mechanism of this process is the lipid-water partition-ing property of the substance [10]. A number of depuration processes includingegestion in fecal matter, deposition in eggs, growth, and metabolic transforma-tion can interfere with the lipid-water partitioning behavior of the chemical sub-stance, typically resulting in bioconcentration factors that are less than the cor-responding lipid-water partition coefficient [9].

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 203

Bioaccumulation refers to the process by which the chemical concentration inan aquatic organism achieves a level that exceeds that in the water as a result ofchemical uptake through all routes of chemical exposure (e.g., dietary absorp-tion, transport across the respiratory surface, dermal absorption). Bioaccumu-lation takes place under field conditions. It is a combination of chemical bio-concentration and biomagnification [9].

Biomagnification refers to the process by which the chemical concentration inthe predator exceeds that in the prey organisms it consumes. In this chapter bio-magnification refers to field conditions where the predator and prey organismsare simultaneously exposed to chemical via both water and diet. Biomagnifica-tion has been observed in aquatic and terrestrial food-webs in the field [11, 12]and under laboratory conditions and mechanisms for this process have been postulated [13, 14].

Food-web bioaccumulation is the process by which the chemical concentra-tion in organisms increases with increasing trophic level. It is the result of a se-quential series of biomagnification events.

Trophic dilution is the process by which the chemical concentration in or-ganisms drops with increasing trophic level. It occurs when the chemical con-centration in the predator remains below the concentration in the prey, typicallyas a result of metabolic transformation of the chemical in the predator.

3Phthalate Ester Nomenclature

Phthalate esters include a large number of substances that share a commonchemical core structure. The phthalate esters discussed in this document andtheir chemical structures and acronyms are listed in Table 1.

4Bioconcentration Studies

Several laboratory studies have investigated the bioconcentration of phthalate es-ters in various fish species, algae, macrophytes, polychaetes, molluscs, crustacean,aquatic insects, and other organisms. The data reported in these studies havebeen compiled and reviewed in Staples et al. [3].

When evaluating the results from laboratory bioconcentration studies, it is im-portant to recognize some of the characteristics and experimental artifacts ofbioconcentration studies for phthalate esters. First, the majority of reported bio-concentration studies involve only one particular phthalate ester, that is, DEHP(Table 1). Bioconcentration data for other phthalate esters are scarce, causingmuch of the experimental evidence on the bioaccumulation of phthalate estersto rely on observations for a single congener. Secondly, the majority of the re-ported studies use radiolabeled phthalate ester congeners. Because phthalate es-ters can be subject to metabolic transformation in organisms, BCFs based on to-tal radioactivity (i.e., radioactivity from the parent substance and its metabolites)can be greater than BCFs based on radioactivity of the parent (i.e., unmetabo-lized) compound alone. BCFs determined with the use of radioactive test sub-

204 F.A.P.C. Gobas et al.

Table 1. Chemical structure and acronyms of phthalate esters

Phthalate Ester Abbreviation Chemical Structure

Dimethyl Phthalate DMP

Diethyl Phthalate DEP

Di-iso-butyl Phthalate DiBP

Di-n-Butyl Phthalate DBP

Butyl Benzyl Phthalate BBP

Di(2-Ethylhexyl) Phthalate DEHP

Di-n-Octyl Phthalate DnOP

Di-n-Nonyl Phthalate DnNP

stances may therefore overestimate the BCF of the parent substance. Third, theaqueous solubility of especially the higher molecular weight phthalate esters islow and the water concentrations used in some of the bioconcentration tests sig-nificantly exceed the aqueous solubility. Results from these bioconcentration testsare difficult to interpret. On one hand, water concentrations above the aqueoussolubility indicate that a considerable fraction of the total chemical concentrationin the water is not available for uptake via the respiratory surface. On the otherhand, phthalate esters can form emulsions when concentrations are in excess ofthe water solubility due to their surface-active properties. These emulsions cancreate micelles that may adhere to the outer surface of the organism and whichmay be also ingested by organisms. As a result, it is unclear to what degree BCFsdetermined at concentrations above the water solubility are representative offield conditions. Fourth, water concentrations of phthalate ester that are constantover the exposure duration are typically not achieved in the bioconcentrationtests due to the low phthalate ester concentrations in the water, rapid absorptionby fish, and degradation in the water phase. Ignoring this experimental artifactin deriving the BCF can lead to an underestimate of the BCF, especially whennominal or initial water concentrations are used to derive the BCF or uptake rateconstants [15]. Fifth, the exposure duration in most bioconcentration tests is rel-atively short and typically much shorter than exposure conditions in the field.For example, a number of bioconcentration tests have been conducted over a period of one day or less while a 28 day period is generally recommended for bioconcentration studies [16]. Experiments that use short exposure times havea tendency to underestimate the actual BCF, since steady-state conditions maynot be achieved.

To eliminate some of the experimental artifacts of laboratory bioconcentrationtests in the analysis of reported bioconcentration data, we plotted BCFs in aquaticmacrophytes, algae, benthic invertebrates, and fish (Fig. 1), and then eliminatedBCF data determined under conditions in which (i) the water concentration ex-ceeded the water solubility and (ii) the exposure time was less than three days.The remaining BCF data are presented as a function the chemical’s octanol-water partition coefficient in Fig. 1. Figure 1 also shows the BCF expected ifphthalate esters simply partition between the water and lipids of the organisms.A 5% lipid content is assumed. Figure 1 illustrates a number of characteristics ofthe bioaccumulation behavior of phthalate esters. First, despite the availability ofa large number of experimental BCF data, there are few data that meet basic dataquality criteria. This illustrates the experimental difficulties of measuring BCFsfor phthalate esters. Secondly, reported BCFs for individual phthalate esters exhibit a large variability. This variability has also been noticed by other authors.For example, Karara and Hayton [17–18] report BCFs for DEHP in 1-5 gSheepshead Minnow (Cyprinodon variegatus) of 6–637 L kg-1 wet weight withina temperature range of 10–23 °C. Thirdly, the reported BCFs of the higher mol-ecular weight phthalate esters are below those expected from lipid-water parti-tioning. This has been explained by metabolic transformation of phthalate esters[4] and by a reduced bioavailability of phthalate esters in the bioconcentrationtests [5–8, 20, 21]. Fourth, the BCFs for DMP and DEP are approximately an or-der of magnitude greater than expected from simple lipid-to-water partitioning.

206 F.A.P.C. Gobas et al.

Fifth, when bioconcentration factors of individual congeners are compared between different taxa, it appears that bioconcentration factors for the highermolecular weight phthalate esters in benthic organisms are greater than those infish. The latter has also been observed by Staples et al. [3] and Wofford et al. [21]who explained these observations by inter-species differences in metabolic trans-formation rates. Finally, the bioconcentration factors are generally less than 5000.

5Dietary Bioaccumulation Studies

While the bioconcentration of phthalate esters has been investigated in manystudies, the dietary bioaccumulation of phthalate esters has received little atten-tion. Macek et al. [23] examined the dietary transfer of DEHP from daphnids tobluegills (Lepomis macrochirus) and concluded that the contribution of the di-etary route to the equilibrium body burden in the bluegill may be small. Glossand Biddinger [24] investigated dietary transfer in daphnids that were feeding ondihexyl phthalate ester-contaminated algae. Perez et al. [25] suggested that di-etary exposure was responsible for seasonal differences in the accumulation ofDEHP in marine biota in microcosm studies. Staples et al. [3] conducted theo-retical calculations to show that as much as 60% of the DEHP exposure in preda-tors could be derived from the diet. They further argued that the general increase

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 207

Fig. 1. Laboratory-derived bioconcentration factors (BCFs) of parent (�) and total (i.e.,parent phthalate ester and metabolites), (+) phthalate esters in various aquatic organisms.The solid line represents the lipid-water partitioning assuming a 5% lipid content (i.e.,BCF= 0.05 KOW)

in the rate of metabolic transformation with increasing trophic level may resultin trophic dilution in which organisms at the top of the food-web contain lowerconcentrations than those in organisms at lower levels.

6Bioaccumulation Studies from Sediments

Uptake and bioaccumulation of sediment-associated phthalate esters has beeninvestigated by Woin and Larson [26] and Brown et al. [27] in dragon flies andchironomid larvae. Based on these data, Staples et al. [3] estimated that the lipidand organic carbon-normalized biota-sediment-accumulation factor (BSAF) forDEHP is 0.1 kg organic carbon (OC) kg–1 lipid in dragonflies and 0.5 kg organiccarbon kg–1 lipid for DEHP and diisodecyl phthalate ester in chironomids.They concluded that these values are lower than the theoretical BSAFs based onan equilibrium partitioning value of 1.0 kg organic carbon kg–1 lipid [28], and suggested that metabolic transformation is a plausible explanation for this dis-crepancy.

7Food-Web Bioaccumulation Studies

A field study to assess the food-web bioaccumulation of a range of phthalate es-ters was recently carried out by our research group. The details of the study canbe found in Mackintosh [29]. The study involved the collection and subsequentchemical analysis of phthalate esters in water, sediments, algae, plankton, filterfeeders (mussels), deposit feeders, forage fish, benthic feeding fish and predatoryfish, and carnivorous water fowl in a marine embayment referred to as FalseCreek. Table 2 lists the species included in the study. With the exception of thedogfish, all species selected can be considered resident species. Environmentalmedia were collected from three different stations in the embayment with a sam-pling frequency of three or four samples per site. Since inter-site variability inconcentration was not a significant factor, concentration data were reported forall stations combined, representing a sample size of 12 for sediment and waterand nine for the biota investigated. The trophic status of the organisms (Table 2)was identified by applying the trophic positioning model by Vander Zanden andRasmussen [30] to dietary composition data from various studies [31–38]. Aconceptual diagram of the food-web is presented in Fig. 2. The study focused oneight individual phthalate esters, that is, DMP, DEP, DiBP, DBP, BBP, DEHP, DnOP,and DnNP. Water concentration measurements identified dissolved and partic-ulate-bound phthalate ester fractions in the water [29].

7.1Biota-Sediment-Accumulation

Figure 3 shows the biota-sediment-accumulation factors (BSAFs) of phthalate esters in a range of benthic invertebrate species as a function of the seawater-corrected octanol-water partition coefficient. Octanol-water partition coeffi-

208 F.A.P.C. Gobas et al.

cients of phthalate esters in seawater were derived from those measured in fresh-water [1] following Xie [39]. The observed BSAFs are shown in relation to theBSAF based on simple organic carbon-lipid partitioning (i.e., BSAF = 1/0.35 =2.86 kg OC kg–1 lipid) [28, 40–43]. It shows that among the various benthicspecies, the BSAFs in geoduck clams are the highest. The BSAFs in geoduck clamsare the lowest for DMP and then appear to increase with increasing log KOW tovalues that, with the exception of DEHP, are not statistically different from 2.86.BSAFs for the other benthic species are significantly lower than those in geoduckclams and show a parabolic relationship with KOW with maximum BSAFs forDBP, DiBP, and BBP. Figure 3 illustrates that there is a substantial variability in theBSAFs among benthic organisms. Sediment burying invertebrates like the geo-duck clams and Manila clams appear to exhibit higher BSAFs than the inverte-brates (e.g., Dungeness crabs) inhabiting surficial sediment. Filter feeding ben-thic invertebrates such as the mussels and oysters exhibit intermediary BSAFs.

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 209

Table 2. Names of species and their trophic position included in the bioaccumulation fieldstudy

Species common name Species Latin name Trophic position

Green algae Enteromorpha intestinalis 1.00Brown algae Nereocystis luetkeana & Fucus gardneri 1.00Phytoplankton 1.00Minnows 2.33Shiner perch Cymatogaster aggregataPacific staghorn sculpin Leptocottus armatusCutthroat trout Salmo clarki clarkiThree spine stickleback Gasterosteus aculeatusWhitespotted greenling Hexogrammos stelleriStarry flounder Platichthys stellatusManila clams Tapes philippinarum 2.40Blue mussels Mytilus edulis 2.48Pacific oysters Crassostrea gigas 2.48Geoduck clams Panope abrupta 2.53Pile perch Rhacochilus vacca 3.05Striped seaperch Embiotoca lateralisForage fish 3.25Pacific herring Clupea harengus pallasiSurf smelt Hypomesus pretiosos pretiosusNorthern anchovy Engraulis mordax mordaxPurple starfish Pisaster ochraccus 3.47Surf scoters Melanitta perspicilata 3.49Pacific staghorn sculpin Leptocottus armatus 3.51Dungeness crabs Cancer magister 3.55Flatfish 3.64English sole Pleuronectes ventulusStarry flounder Platichthys stellatusWhitespotted greenling Hexogrammos stelleri 3.81Spiny dogfish Squalus Acanthias 4.07

One of the contributing factors to the differences in the observed BSAFs be-tween the species is the chemical disequilibrium that appears to exist between thesediments and the overlying water. This disequilibrium is illustrated in Fig. 4,which shows the observed sediment-water distribution coefficients (expressed interms of L kg–1 organic carbon) in relation to the sediment-seawater partition co-efficients (L kg–1 organic carbon), derived from the seawater corrected octanol-water partition coefficients according to Seth et al. [44]. A sediment-water dise-quilibrium occurs if the sediment-water distribution coefficient exceeds thechemical’s sediment-water partition coefficient. It can be expressed by the degree

210 F.A.P.C. Gobas et al.

Fig.

2.A

sim

plifi

ed c

once

ptua

l dia

gram

oft

he fo

od-w

eb in

tera

ctio

ns in

the

Fals

e C

reek

food

-web

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 211

Fig. 3. Biota-sediment-accumulation factors (BSAFs) in units of kg organic carbon kg–1 lipidof phthalate esters in a range of benthic invertebrate species as a function of the octanol-sea-water partition coefficient. The solid line represents the sediment-organism equilibrium par-tition coefficient (BSAF=2.86)

Fig. 4. Sediment-water distribution coefficients in units of L kg–1 organic carbon in relation tosediment-seawater partition coefficients (L kg–1 organic carbon), derived as 0.35 KOW accord-ing to Seth et al. [44]

to which the observed sediment-water distribution coefficient (KSW) exceeds thesediment-water equilibrium partition coefficient (KSW, EQ), that is, KSW/KSW, EQ. Itrepresents a situation in which the sediments are at a higher concentration thansediment-water partitioning thermodynamics dictates. Figure 4 illustrates thatsediment-water disequilibria fall with increasing KOW from values as high as229,000 for DMP to 41 for DEP and reach a constant value between approximatelytwo and ten for the higher molecular weight phthalate esters. A disequilibriumof ten indicates that the sediment pore water concentration is an order of mag-nitude greater than the concentration in the overlying water.A value above unityalso suggests that the sediments are serving as an exposure source to the watercolumn. From a bioaccumulation perspective, the significance of the apparentdisequilibrium between sediments and overlying water is that the degree ofdirect exposure of the organism to sediments and associated interstitial waterversus exposure to the overlying water will have a significant effect on the bodyburden of the exposed organism.A sediment burying invertebrate (such as geo-duck clams) with greater contact to sediments is therefore expected to be exposedto a higher effective concentration than epibenthic organisms that inhabit theepilimnion (e.g., the Dungeness crab), where they are exposed to the overlyingwater. Differences in an organism’s habitat utilization are therefore expected tobe partly responsible for the differences in the BSAFs that are observed. Otherfactors, such as metabolic transformation, growth dilution, and low dietary up-take efficiencies of phthalate esters may also play a role.

7.2Food-Web Bioaccumulation

Figure 5 illustrates the relationship between the lipid equivalent concentration(CL in ng g–1 lipid) of phthalate esters and trophic position for the organisms inthe False Creek food-web. For fish and shellfish, lipid equivalent concentrationswere derived by dividing the wet weight-based concentration CB (ng g–1 wetweight) by the lipid content L (kg lipid kg–1 organism or tissue on a wet weightbasis):

CL = CB/L (1)

For algae and plankton, the calculation of the lipid equivalent concentration wasconducted as:

CL = CB*/(LB*+ 0.35 fOC) (2)

where CB* is the chemical concentration on a dry weight basis, LB* is the lipid con-tent on a dry weight basis (kg lipid kg–1 sample, dry weight), fOC is the organiccarbon content (kg OC kg–1 sample, dry weight), and 0.35 is a proportionalityconstant reflecting the differences in the sorptive capacities between organic car-bon and octanol [44]. The reason for the difference in the methodology for lipidnormalization between algae, plankton, fish, and shellfish is that due to the lowlipid content of algae and plankton (i.e., 0.1–0.4%) but high organic carbon con-tent (i.e., fOC = 33–40%) lipids are not the main site for chemical accumulation[45]. The purpose of the lipid normalization is to remove the effect of differences

212 F.A.P.C. Gobas et al.

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 213

in lipid content among organisms of different trophic levels on phthalate esterconcentrations.

Figure 5 illustrates that there are no statistically significant relationships be-tween the lipid equivalent concentrations and trophic position for DMP, DEP,DiBP, DBP, and BBP. Analysis of covariance shows that lipid equivalent concen-trations do not appear to increase or drop significantly (i.e., P >0.05) betweentrophic levels. This indicates that these phthalate esters do not biomagnify in thefood-web. Biomagnification in the food-web is defined as the increase in the lipidequivalent concentration with increasing trophic level. The apparent constancyof the lipid equivalent concentrations with increasing trophic position suggeststhat the accumulation of these phthalate esters is due to simple water-to-lipidpartitioning, which produces approximately equal lipid equivalent concentra-tions in the various organisms of the food-web.

For the higher molecular weight phthalate esters (i.e., DEHP, DnOP, andDnNP) there appears to be a statistically significant drop (i.e., P <0.05) in thelipid equivalent concentration with increasing trophic position. The latter indi-cates trophic dilution, in which lipid equivalent concentrations in organisms de-cline with increasing trophic level. The observations indicate that higher trophiclevel organisms are exposed to lower concentrations of these higher molecularweight phthalate esters than organisms of lower trophic levels.

7.3BAFs

Figure 6 illustrates the observed relationships between the BAF (expressed inunits of L kg–1 equivalent lipid) and KOW for all the species included in the fieldbioaccumulation study. To simplify Fig. 6, the species are grouped into trophicguilds. For the purpose of this analysis, we distinguished between algae, plank-ton, benthic invertebrates, small forage fish, predatory fish, and aquatic birds. Toprovide a basis for comparison, Fig. 6 also presents the expected BAFs assumingthat only simple lipid-water partitioning of the chemicals between the organismsand the water controls bioaccumulation [10], that is, BCF (L kg–1 equivalentlipid) = KOW. This simple model ignores the potential role of dietary uptake, bio-magnification, metabolism, growth dilution, and the reduction of the chemicalbioavailability due to sorption in the water phase.

Figure 6 illustrates a number of characteristics of the bioaccumulation be-havior of phthalate esters in the field. Firstly, BAFs of individual phthalate estersexhibit a considerable variability. The variability in BAFs ranges from a factor ofapproximately 30 for the lower molecular weight phthalate esters to a factor of1000 for DEHP, DnOP, and DnNP. There appears to be no apparent relationshipbetween the BAF and the trophic position of the organism for the lower molec-ular weight phthalate esters. This indicates that the observed variability in theBAF is not due to differences in trophic position among the organisms sampled.However, for the higher molecular weight phthalate esters, there appears to be atrend for the BAFs to drop with increasing trophic position. This trend is themain reason that the BAFs of the higher molecular weight phthalate esters showa greater variability than the BAFs of the lower molecular weight phthalate esters.

214 F.A.P.C. Gobas et al.

Fig. 5. Relationship between the lipid equivalent concentrations of phthalate esters and trophicposition for a range of organisms in a marine food-web

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 215

Fig. 6. Relationships between the observed BAF, expressed relative to the total concentrationin the water, in units of L kg–1 equivalent lipid in algae, plankton, benthos, small fish, predatoryfish, and birds, and the octanol-seawater partition coefficient for phthalate esters. The dashedline represents the Canadian Environmental Protection Act’s bioaccumulation criterion ex-pressed on a lipid-normalized basis assuming a 5% lipid content

Secondly, the comparison of the observed BAFs with the BCFs derived by sim-ple lipid-water partitioning shows that the BAFs for DMP and DEP are greaterthan expected from simple lipid-to-water partitioning. In particular, the BAF ofDMP is much greater (i.e., on average a factor of 100) than that expected basedon lipid-water partitioning. Considering that (i) all biota and water concentrationexceed the method detection limits by many fold, (ii) DMP showed high extrac-tion recoveries, negligible degradation in and evaporative losses from the watersamples (due to immediate analysis at low temperature), and (iii) positive MS-MSconfirmation in water and biota samples, it is unlikely that the much higher thanexpected BAFs are due to analytical error. Also, due to the mass-specific analy-sis, metabolites can be ruled out as a factor. A possible explanation for the highBAFs for DMP is the large disequilibrium between sediments and overlying water. Contact of organisms with the sediments (e.g., sculpins burying in sedi-ments) may elevate the body burden of DMP in organisms over that absorbedfrom the overlying water.

Thirdly, the BAFs of DBP, DiBP, and BBP are generally in reasonable agreementwith the BCFs based on simple lipid-water partitioning. This suggests that thebioaccumulation of these substances is mainly the result of chemical exchangebetween the organism and the water via the respiratory surface of the organisms.Dietary uptake, metabolic transformation, and growth dilution appear to play a

secondary role, but may contribute to the variability in the observed BAFs amongthe different organisms. The BAFs of the higher molecular weight phthalate es-ters (i.e., DEHP, DOP, DnNP) are lower than anticipated based on lipid-water par-titioning. The low bioavailability of the total water concentration is a key factorcausing the BAFs of the higher molecular weight phthalate esters to fall below thelipid-water partition coefficients. For example, our study suggests that approxi-mately 0.1% of the total water concentration of DEHP is freely dissolved andhence assumed to be available for uptake via the respiratory surface. The freelydissolved water concentration is believed to represent the phthalate ester con-centration that can be absorbed by organisms via the respiratory surface area asa result of lipid-water partitioning. Figure 7 illustrates the BAF based on freelydissolved concentrations BAFfd (L kg–1 lipid) normalized to the lipid-water par-tition coefficient, that is, BAFfd/KOW. It shows that the BAFs in plankton and green algae approach the lipid-water partition coefficients, that is, BAFfd/KOW isapproximately 1.0. This result appears to be reasonable as algae can be expectedto lack a metabolic transformation capability for lipid-like molecules such as

216 F.A.P.C. Gobas et al.

Fig. 7. Relationship between the observed BAFs, expressed in terms of the freely dissolved wa-ter concentration, in units of L kg–1 equivalent lipid, divided by the octanol-seawater partitioncoefficient (KOW) in algae (shaded triangles), plankton (�), benthos (+), small fish (shaded cir-cles), predatory fish (�), and birds (–) for phthalate esters. The solid line represents the or-ganism-water equilibrium partition coefficient (BAFL = KOW)

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 217

phthalate esters. The BAFs of the higher molecular weight phthalate esters in themajority of benthic invertebrates, forage fish, predatory fish, and birds are lessthan the lipid-water partition coefficients (i.e., BAFfd/KOW < 1), indicating thatbioaccumulation factors for the freely dissolved phthalate ester are significantlyless than their octanol-water partition coefficients. This suggests that process-es other than lipid-water partitioning (e.g., metabolic transformation, growth,fecal excretion) have a significant effect on the BAF.

Fourthly, Fig. 6 further shows that when the BAFs are compared to the Cana-dian Environmental Protection Act’s cut-off value of 5000 (if the BAF is expressedon a wet weight) or a 100,000 (if the BAF is expressed on a lipid weight basis), theBAFs of all the phthalate esters generally fall below the cut-off value. The only ex-ceptions are the BAFs of BBP for green algae, plankton, geoduck clams, stripedseaperch, pile perch, staghorn sculpins, and surfscoters.

8Bioaccumulation Models

Bioaccumulation models can be useful tools in the investigation of the mecha-nism of phthalate ester bioaccumulation. The merit of such models is in investi-gating the role that uptake and elimination processes contribute to the observedBAFs. For example, these models can be used to estimate whether a chemical sub-stance is predominantly absorbed by an organism from the water via the respi-ratory surface or through the dietary route. These models can also be applied toassess to what degree chemical substances can be expected to be eliminated viawater or feces and to what degree metabolic transformation and growth affect tis-sue concentrations. Most importantly, the model can be used to test hypothesesregarding the mechanisms contributing to the bioaccumulation process. The hy-pothesis can be tested by comparing model predictions to observed data. In thissection, we will discuss a general bioaccumulation model [46] for fish and test themodel against the data from the field bioaccumulation study. The model esti-

Table 3. Model equations, parameters, and their units of the fish bioaccumulation model ofGobas [13]. BAF = CF/CW = [k1 fDW + kD CD/CW)]/(k2 + kE + kM + kG)

Parameter Units Definition

fDW fraction Fraction of the water concentration that is freely dissolved CD g kg–1 wet weight Chemical concentration in diet CF g kg–1 wet weight Chemical concentration in fish CW g L–1 Total chemical concentration in overlying water k1 L water kg–1 organ- Uptake clearance rate from water

ism wet weight d–1

k2 d–1 Elimination rate constant from fishkD kg food kg–1 organ- Dietary uptake clearance rate

ism wet weight d–1

kE d–1 Fecal egestion elimination rate constant from fishkG d–1 Growth dilution rate constant kM d–1 Metabolic transformation rate constant from fish

218 F.A.P.C. Gobas et al.

Tabl

e4.

Bioa

ccum

ulat

ion

mod

el in

put p

aram

eter

s an

d th

e re

sults

oft

he m

odel

cal

cula

tion

s of

the

lipid

-nor

mal

ized

BA

F Lfo

r se

vera

l pht

hala

te e

ster

s in

a fie

ld e

xpos

ed 0

.1kg

sta

ghor

n sc

ulpi

n (l

ipid

con

tent

is 5

.0%

) and

a 3

kg d

ogfis

h (l

ipid

con

tent

is 1

5%) i

n re

lati

on to

the

obse

rved

BA

Fsa

Scul

pins

Ph

thal

ate

este

r Lo

gK

OW

CW

CD

f DW

(%)

k 1k 2

k Dk E

k Mk G

BAF L

,PBA

F L,O

DM

P 1.

78

3.51

2000

100

8327

.70.

0246

0.00

616

00.

0021

70.4

19,0

00D

EP2.

7412

711

,500

9918

86.

830.

0246

0.00

616

00.

0021

552

3900

DBP

4.52

110

32,8

0075

.422

10.

237

0.02

460.

0061

60

0.00

2114

,200

22,0

00D

iBP

4.52

5.15

2750

75.4

221

0.13

40.

0246

0.00

615

00.

0021

25,3

0028

,000

BBP

4.98

3.48

11,4

0053

.322

20.

0464

0.02

460.

0061

40

0.00

2172

,700

204,

000

DEH

P8.

0822

813

1,00

00.

1122

20.

0000

369

0.00

653

0.00

163

0.00

110.

0021

16,0

0016

,000

DnO

P8.

0812

.887

000.

1122

20.

0000

369

0.00

653

0.00

163

0.00

520.

0021

10,0

0010

,000

DnN

P8.

9877

.726

,900

0.01

222

0.00

0004

650.

0010

70.

0002

680.

0001

0.00

2131

0031

00

Dog

fish

Phth

alat

e es

ter

Log

KO

WC

WC

Df D

W (%

)k 1

k 2k D

k Ek M

k GBA

F L,P

BAF L

,O

DM

P1.

783.

5124

3010

021

.42.

370.

0148

0.00

370

00.

0011

8990

40D

EP2.

7412

716

,900

9948

.10.

584

0.01

480.

0037

00

0.00

1156

089

1D

BP4.

5211

034

8075

.456

.70.

0114

0.01

480.

0036

70.

170.

0011

189

189

DiB

P4.

525.

1582

,700

75.4

56.7

0.01

140.

0148

0.00

369

0.09

0.00

1132

4032

40BB

P4.

983.

4812

,800

53.3

56.8

0.00

397

0.01

470.

0036

90.

040.

0011

11,8

0011

,800

DEH

P8.

0822

811

4,00

00.

1156

.90.

0000

0316

0.00

392

0.00

0980

0.01

70.

0011

580

580

DnO

P8.

0812

.898

000.

1156

.90.

0000

0316

0.00

392

0.00

0980

0.05

20.

0011

374

374

DnN

P8.

9877

.715

,100

0.01

56.9

0.00

0000

397

0.00

0642

0.00

0161

0.03

0.00

1128

28

aT

he m

odel

inpu

t par

amet

ers

are

the

octa

nol-

seaw

ater

par

titi

on c

oeff

icie

nt K

OW

,the

obs

erve

d to

tal w

ater

con

cent

rati

on C

W(n

g L–1

),th

e di

etar

y co

n-ce

ntra

tion

CD

(ng

kg–1

wet

wei

ght)

,the

free

ly d

isso

lved

frac

tion

oft

he w

ater

con

cent

rati

on f

DW

(%).

The

mod

el o

utpu

t par

amet

ers

are

the

gill

upta

kera

te c

onst

ant k

1(L

kg–1

wet

wei

ght)

,the

gill

elim

inat

ion

rate

con

stan

t k2

(day

–1),

the

diet

ary

upta

ke r

ate

cons

tant

kD

(kg

food

kg–1

wet

wei

ghtd

ay–1

),th

e fe

cal e

gest

ion

rate

con

stan

t (da

y–1),

the

met

abol

ic tr

ansf

orm

atio

n ra

te c

onst

ant k

M(d

ay–1

),th

e gr

owth

rate

con

stan

t kG

(day

–1),

the

mod

el-p

redi

cted

lipid

equ

ival

ent b

ioac

cum

ulat

ion

fact

or e

xpre

ssed

bas

ed o

n th

e to

tal w

ater

con

cent

rati

on B

AF L

,P(L

kg–1

lipid

) and

the

obse

rved

lipi

d eq

uiva

lent

bio

ac-

cum

ulat

ion

fact

or e

xpre

ssed

bas

ed o

n th

e to

tal w

ater

con

cent

rati

on B

AF L

,O(L

kg–1

lipid

).

mates a whole organism BAF in units of L kg–1 wet weight based on the total wa-ter concentration as:

BAF = CF/CW = [k1 fDW +kD CD/CW]/(k2 + kE + kM +kG) (3)

The corresponding lipid equivalent BAFL (L kg–1 equivalent lipid) is BAF/L, whereL is the lipid content of the fish (kg lipid kg–1 wet weight organism). The modelparameters are explained in Table 3. The methods for the calculation of the up-take and elimination rate constants can be found in Gobas [46]. The model cal-culations are illustrated for several phthalate esters in a 0.1 kg staghorn sculpin(lipid content of 5.0%) and a 3 kg dogfish (lipid content is 15%) at a water tem-perature of 10 °C. The weight, lipid content, and temperature of the species cor-respond to the animals that were sampled as part of the field bioaccumulationstudy. The model input parameters and results are listed in Table 4 and a com-parison of model-predicted and observed BAFLs are given in Fig. 8. The modelcalculations were conducted by using (i) a metabolic transformation rate con-stant kM of 0 and (ii) a metabolic transformation rate constants that was fitted toproduce a perfect agreement between model-predicted and field-observedBAFLs. The latter method is essentially an approximation of the potential mag-nitude of the metabolic transformation rate constant kM.

The model calculations illustrate that the lower molecular weight phthalate es-ters (i.e., DMP, DEP, DiBP, DBP, and BBP) in sculpin and DMP, DEP, DiBP, and DBPin dogfish are almost exclusively absorbed from the water via the gills. Dietaryuptake of these phthalate esters is small compared to the uptake from the water,and model-predicted biomagnification factors are less than 1.0, indicating thatbiomagnification is not expected to occur. The latter is supported by the resultsof the bioaccumulation field study, which shows that lipid equivalent concentra-tions as well as lipid equivalent BAFs do not increase with increasing trophic po-sition. The model calculations further show that in sculpins the lower molecularweight phthalate esters are virtually completely depurated through gill elimina-tion. Growth and fecal egestion do not have a significant effect on the BAF insculpins. Metabolic transformation rate is not required in the model to explainthe observed BAFs in sculpins in the field. This does not mean that metabolictransformation does not occur; only that its rate may be too low to have a sig-nificant effect on the BAF.

The model calculations show that with the exception of DMP and DEP, thelower molecular weight phthalate esters in dogfish require a substantial meta-bolic transformation rate to explain the observed BAFs. The model results sug-gest that metabolic transformation is the main route of depuration of the DBP,DiBP, and BBP in the upper-trophic level dogfish. As a result, the BAFLs of thesesubstances are lower than their KOW.

The model calculations show that the exposure of sculpins and dogfish tohigher molecular weight phthalate esters (DEHP, DnOP, DnNP) is a combinationof both direct exposure to the water and dietary uptake. Dietary uptake appearsto be more important than direct uptake from the water. However, the uncertaintyin the determination of the freely dissolved water concentrations prevents a moreconclusive assessment of whether the diet or the water is the main source of up-take in these fish species. Staples et al. [3] predicted that as much as 60% of the

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 219

220 F.A.P.C. Gobas et al.

Fig. 8. Model-predicted lipid-normalized BAFs, expressed relative to the total concentrationin the water, in units of L kg–1 lipid in relation to the observed values for staghorn sculpins andspiny dogfish. Grey bars represent model predictions assuming metabolism (see Table 4). Blackbars represent model predictions assuming no (kM = 0) metabolism. White bars represent theobserved values

DEHP exposure in predators could be derived from the diet. The model showsthat the depuration rates for the higher molecular weight phthalate esters are sub-stantially smaller than those for the lower molecular weight phthalate esters. Thismeans that even low rates of metabolic transformation can have a significant ef-fect on the BAF. The model calculations show that for sculpins only small meta-bolic transformation rates need to be invoked to explain the observed BAFs insculpins. Considering the error in the model parameterization and in the ob-served BAFs, it is unclear whether metabolic transformation has a significant ef-fect on the BAF. Assuming no metabolic transformation, model predicted BAFsin sculpins are in good agreement with the observed values. In dogfish, however,high rates of metabolic transformation need to be used to explain the observedBAFs. The latter suggests that metabolic transformation is an important depu-ration process in dogfish causing the BAFs to be much lower than the lipid-wa-ter partition coefficients of these phthalate esters. Metabolic transformation ratesof the higher molecular weight phthalate esters in dogfish may be between0.02 and 0.05 d–1. These rates are on average about an order of magnitude lowerthan depuration rates derived from laboratory bioconcentration studies with fish[3]. The lower rate of metabolic transformation may be due to (i) the larger sizeand higher lipid content of the dogfish compared to the fish used in the labora-tory experiments and (ii) the much longer exposure duration in the field com-pared to that in laboratory tests. The high lipid content and size of the organismmay generate a large lipid storage compartment for phthalate esters and reducethe fraction of phthalate in the fish that is available for metabolic transformationcompared to that in smaller, less lipid-rich fish. The longer exposure duration inthe field-exposed fish is likely to increase the fraction of the total amount of ph-thalate ester in less accessible “slow” storage compartments. The greater fractionof phthalate ester stored in these less accessible compartments (such as the lipids)may cause a smaller fraction of the total amount of phthalate ester in the fish tobe available for metabolic transformation.

For DMP and DEP the model underestimates the observed BAFs, which maybe explained by the apparent sediment-water disequilibria. The sediment-waterdisequilibria may cause the exposure concentration of this benthic fish speciesto exceed that measured in the overlying water. Use of the overlying water con-centration can therefore be expected to underestimate the actual BAF in this fishspecies.

9Conclusions

Currently, there exists considerable information on the bioaccumulation behav-ior of phthalate esters in aquatic systems. Laboratory experiments, field studies,and mathematical modeling studies have all been carried out.A number of con-clusions can be drawn from the information available.

Firstly, there is no evidence from laboratory and field bioaccumulation stud-ies to support the hypothesis that phthalate esters biomagnify in aquatic food-webs. Dietary bioaccumulation studies, sediment bioaccumulation studies in thelab and the field as well as the food-web bioaccumulation study discussed in this

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 221

chapter all indicate that food-web bioaccumulation (i.e., the increase of the lipidequivalent concentration with increasing trophic level) does not occur. This in-dicates that despite their high octanol-water partition coefficients, phthalate es-ters do not appear to biomagnify in aquatic food-webs.

Secondly, it is interesting that the lowest molecular weight phthalate estersDMP and DEP exhibit BAFs that are greater than their lipid-to-water partitioncoefficients. The results from laboratory bioconcentration experiments showsimilar results (Fig. 1). These observations suggest that DMP and DEP have agreater bioaccumulation and bioconcentration potential than indicated by theiroctanol-water partition coefficient. The laboratory observations may be ex-plained by the possible formation of metabolic products of DMP and DEP that,due to the method of detection used, were indistinguishable from the parentcompounds. However, due to the more specific detection methodology in thebioaccumulation field study (i.e., MS-MS detection), the formation of metabolicproducts of DMP and DEP is unlikely to explain the higher-than-expected BAFs.While analytical error of the water concentration of DMP and DEP is a likelycause for the higher-than-expected BAFs, the QA/QC procedures applied indicatethat analytical error cannot explain the observations either. The model calcula-tions indicate that DMP and DEP are almost exclusively absorbed from the wa-ter via the respiratory surface. The large apparent sediment-to-water disequilib-ria for DMP and DEP is the most likely explanation of the higher-than-expectedBAFs in the field.

Thirdly, the lipid equivalent BAFs of DBP, DiBP, and BBP, appear to be fairlyuniform among the organisms of the marine food-web investigated in this study(Fig. 5). This suggests that the organisms are exposed to a common source andthat dietary uptake of phthalate esters has little effect on the BAF of these phthalate esters. Model calculations support this, by demonstrating that directuptake of these phthalate esters from the water via the respiratory surface of theorganisms can be expected to be the main exposure route and that dietary up-take is less important. The general agreement between lipid-normalized BAFsand lipid-water partition coefficients (Fig. 6) also support this conclusion andfurther indicates that the bioaccumulation of the lower molecular weight ph-thalate esters generally follows the lipid-water partitioning model. Laboratorybioconcentration studies suggest that the BCFs can reach values up to the lipid-water partition coefficients of these phthalate esters. However, several observedBCFs appear to be lower than the lipid-water partition coefficients. An interest-ing observation from the laboratory bioconcentration tests and the bioaccumu-lation field study is that there is a substantial variability in the observed biocon-centration and bioaccumulation factors. While experimental artifacts can beexpected to be an important cause of the variability in the observed BCFs, theyare an unlikely source of the variability in the BAFs. One potential cause of theobserved variability is metabolic transformation. Several authors have implicatedmetabolic transformation as an important factor controlling the bioaccumula-tion factors of phthalate esters. The model calculations illustrate that for themetabolic transformation to have a significant effect on the BAFs of the lowermolecular weight phthalate esters, the rates of metabolic transformation have tobe relatively high. It is possible that metabolic transformation differ among or-

222 F.A.P.C. Gobas et al.

ganisms and that in certain organisms, metabolic transformation rates are suf-ficiently large to affect the BAF and cause some of the variability in the observedBAFs among organisms of the food-web. A second cause of the variability inBAFs may relate to the concentration gradients between sediment and water anddifferences among organisms in their interaction with the sediments and water.Water and sediment concentrations indicate that the sediments may providehigher phthalate esters exposure concentrations than the overlying water. Hence,the interaction of the organisms with the sediments and its pore water is likelyto be responsible for some of the variability in the observed BAFs.

Fourthly, the BAFs of the higher molecular weight phthalate esters (i.e., DEHP,DnOP, and DnNP) show a tendency to decrease with increasing trophic position.This suggests that organisms at higher trophic levels are exposed to lower phthalate ester concentrations via prey.A similar apparent relationship betweenBCF and trophic status has been found in laboratory experiments in which BCFswere highest for algae and lowest for fish with invertebrates exhibiting interme-diate values [3].Assessment of the freely dissolved concentrations indicates thatthe higher molecular weight phthalate esters exhibit a very low bioavailability,that is, only a very small fraction of these phthalate esters in natural waters canbe absorbed via the respiratory surface of aquatic organisms. When expressedrelative to the freely dissolved concentration in the water, the BAFs in algae andplankton appear to be within an order of magnitude of the lipid-water partitioncoefficients. This suggests that partitioning is likely an important mechanism forbioaccumulation in algae and plankton. In higher trophic level organisms suchas fish, model calculations indicate that dietary uptake is likely to be an impor-tant route of exposure as bioavailable concentrations in the water are expectedto be very low. The inability of dietary uptake to cause biomagnification is there-fore an interesting characteristic of the high-molecular weight phthalate estersin particular and phthalate esters in general. The model calculations provide twopossible explanations for this phenomenon. First, it is possible that after these phthalate esters have been absorbed, the phthalate esters are metabolized in thefish. This explanation has been proposed by several authors and supported by thedetection of some phthalate ester metabolites [3–8]. A greater rate of metabolictransformation has been suggested to explain the drop in BCFs with increasingtrophic level. The other possible explanation is that phthalate esters ingested withthe diet are very effectively metabolized in the gastro-intestinal tract even beforethey are absorbed (i.e., effectively decreasing kD in Eq. (3)). This first-pass effectessentially prevents a significant rate of dietary uptake of the parent phthalate es-ters. The structural similarity between lipids and phthalate esters may favor sucha process as pH and enzymatic conditions in the gastro-intestinal tract are tai-lor-made for the hydrolysis of lipids and perhaps phthalate esters. The uptakethat would still occur is directly from the water. Model calculations illustrate thatin the absence of dietary uptake the BAF can be expected to drop with increas-ing organism size (which correlates well with trophic level) as has been observedin the field study. This is due to the fact that with increasing organism size (andreducing area-to-volume ratio), the gill elimination and fecal egestion rates dropand become negligible compared to growth rates or even small metabolic trans-formation rates. This results in smaller BCFs for larger organisms. This second

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 223

hypothesis does not require the occurrence of a high rate of metabolism in thefish. The bioaccumulation behavior of the lower molecular weight phthalate es-ters is not consistent with a high rate of metabolism. It is therefore possible thatphthalate esters are fairly slowly metabolized after they have been absorbed, butthey are effectively metabolized in the gastro-intestinal tract before they are ab-sorbed.We are currently carrying out laboratory experiments to distinguish be-tween these two possible explanations. The toxicological significance of these dif-ferent mechanisms is that metabolic transformation in organisms has thepotential to create metabolic products, while an effective first-pass effect mayprevent dietary uptake and the formation of potentially reactive metabolic prod-ucts within the organism.

The majority of observed BAFs for phthalate esters did not exceed the bioac-cumulation criterion of 5000 L kg–1 wet weight or 100,000 L kg–1 lipid if expressedon a lipid equivalent basis. Only BAFs of BBP in green algae, plankton, geoduckclams, striped seaperch, pile perch, staghorn sculpins, and surfscoters exceededthe bioaccumulation criterion The results of the field study also confirmed thehypothesis that these substances do not appear to biomagnify in the food-web.

Since the intention of the bioaccumulation criteria is to identify substancesthat, like PCBs, exhibit biomagnification, current evidence in the literature andfrom our study support the conclusion that phthalate esters do not appear to bebioaccumulative.

Acknowledgement. The authors wish to acknowledge the Natural Sciences and Engineering Research of Canada and the American Chemistry Council for sponsoring this research. We further thank the contributions of Audrey Chong, Judy Carlow, Zhongping Lin, Jing Hongwu,and Natatsha Hoover for their contributions to the research.

10References

1. Cousins I, Mackay D (2000) Chemosphere 41:13892. OECD (1998) Harmonized integrated hazard classification system for human health and

environmental effects of chemical substances. Organization for Economic Cooperation andDevelopment, Paris

3. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:6674. Barron MG, Schultz IR, Hayton WL (1988) Toxol Appl Pharmacol 98:495. Hogan JW (1977) In: Johnson BT, Stalling DL, Hogan JW, Schoettger RA (eds) Pollutants in

the air and water environments. Wiley, New York, p 2926. Metcalf RL, Booth GM, Schuth CK, Hansen DJ, Lu PY (1973) Environ Health Perspect

June:277. Carr KH, Coyle GT, Kimerle RA (1992) 13th annual Society of Environmental Toxicology

& Chemistry meeting. Seattle, Washington8. Barron MG, Albro PW, Hayton WL (1995) Environ Toxicol Chem 14:8739. Gobas FAPC, Morrison HA (1999) Bioconcentration & bioaccumulation in the aquatic

environment. In: Boethling R, Mackay D (eds) Handbook of property estimation methodsfor chemicals: environmental and health sciences. CRC Press, Boca Raton, p 139

10. Mackay D (1982) Environ Sci Technol 16:27411. Connolly JP, Pedersen CJ (1988) Environ Sci Technol 22:9912. Kelly BC, Gobas FAPC (2001) Environ Sci Technol 35:32513. Gobas FAPC, Zhang X, Wells R (1993) Environ Sci Technol 27:2855

224 F.A.P.C. Gobas et al.

14. Gobas FAPC, Wilcockson JWB, Russell RW, Haffner GD (1999) Environ Sci Technol 33:133

15. Gobas FAPC, Zhang X (1995) Chemosphere 25:196116. Organization for Economic Co-operation and Development (1996) Bioaccumulation: flow-

through fish test, 305 E. OECD guideline for testing chemicals17. Karara AH, Hayton WL (1984) Aquat Toxicol 5 :18118. Karara AH, Hayton WL (1989) Aquat Toxicol 15:2719. Gobas FAPC, Clark KE, Shiu WY (1989) Environ Toxicol Chem 8:23120. Mayer FL (1976) Fish Res Board Can 33:261021. Boese BL (1984) Can J Fish Aquat Sci 41:171322. Wofford H, Wilsey CD, Neff GS, Giam CS, Neff JM (1981) Ecotoxicol Environ Safety 5 :20223. Macek KJ, Petrocelli SR, Sleight BH (1979) Considerations in assessing the potential for and

significance of biomagnification of chemical residues. In: Marking LL, Kimerle RA (eds)Aquatic food chains, aquatic toxicology.American Society for Testing and Materials, p 251

24. Gloss SP, Biddinger GR (1985) Comparison and system design and reproducibility to estimate bioconcentration of di-n-hexylphthalate by Daphnia magna. In: Cardwell RD,Purdy R, Bahner RC (eds) Aquatic toxicology and hazard assessment: seventh symposium.American Society for Testing and Materials, Philadelphia, p 202

25. Perez KT, Davey EW, Lackie NF, Morrison GE, Murphy PG, Soper AE, Winslow DL (1985)Environmental assessment of phthalate ester di-(2-ethylhexyl) phthalate derived from amarine microcosm. US Environmental Protection Agency Report, Naragannsett RI, EPA600/D-85/070

26. Woin P, Larsson P (1987) Bull Environ Contam Toxicol 38:22027. Brown D, Thompson RS, Stewart KM, Croudace CP, Gillings E (1996) Chemosphere 32:217728. DiToro DM, Zarba CS, Hansen DJ, Berry WJ, Swartz RC, Cowan CE, Pavlou SP, Allen HE,

Thomas NA, Paquin PR (1991) Environ Toxicol Chem 10:154129. Mackintosh CE (2001) MSci thesis, Simon Fraser University30. Vander Zanden MJ, Rasmussen JB (1996) Ecological Monographs 66:45131. Butler TH (1980) Shrimps of the Pacific coast of Canada. Canadian Bulletin of Fisheries and

Aquatic Sciences no 202. Department of Fisheries and Oceans32. Forrester CR (1969) Life history information on some groundfish species. Fisheries Re-

search Board of Canada. Technical Report no 10533. Hart JL (1973) Pacific fishes of Canada. Fisheries Research Board of Canada. Bulletin 18034. Jamieson GS, Francis K (eds) (1986) Invertebrate and marine plant fishery resources of

British Columbia. Canadian Special Publication of Fisheries and Aquatic Sciences 91.Department of Fisheries and Oceans, Ottawa, Ontario

35. Jones BC (1976) PhD thesis, Simon Fraser University36. Miller BS (1967) J Fish Res Board Can 24:251537. Ricketts EF, Calvin J, Hedgpeth JW, Phillips DW (1985) Between Pacific tides, 5th edn. Stan-

ford University Press, Stanford38. Vermeer K, Ydenburg RC (1989) Feeding ecology of marine birds in the Strait of Georgia.

In: Vermeer K, Butler RW (eds) The ecology and status of marine and shoreline birds in theStrait of Georgia, British Columbia. Canadian Wildlife Service, Ottawa, p 62

39. Mallhot H (1987) Environ Sci Technol 21:100940. Shea D (1988) Environ Sci Technol 22:125641. Gobas FAPC, Bedard DC, Ciborowski C, Jan JH (1989) J Great Lakes Res 15:58142. Bierman VJ Jr (1990) Environ Sci Technol 24:140743. Parkerton TF (1993) PhD thesis, Rutgers University44. Seth R, Mackay D, Muncke (1999) Environ Sci Technol 33:239045. Swackhamer DL, Skoglund RS (1993) Environ Toxicol Chem 12:83146. Gobas FAPC (1993) Ecol Modell 69 :1

Bioaccumulation of Phthalate Esters in Aquatic Food-Webs 225

© Springer-Verlag Berlin Heidelberg 2003

Assessment of Critical Exposure Pathways

Kathryn Clark 1 · Ian T. Cousins 2 · Donald Mackay 2

1 BEC Technologies Inc., 61 Catherine Avenue, Aurora, Ontario, L4G 1K6, CanadaE-mail: kclark@pathcom.com

2 Canadian Environmental Modelling Centre, Environmental and Resource Studies,Trent University, Peterborough, Ontario, K9J 7B8, Canada

Human exposure to phthalate esters for five different age classes is evaluated for the followingroutes of exposure: inhalation of air (indoors and outdoors), ingestion of drinking water,incidental ingestion of soil, ingestion of dust (indoors), and ingestion of food. Exposure is estimated for: dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP),butylbenzyl phthalate (BBP), and bis(2-ethylhexyl) phthalate (DEHP).

For the five phthalate esters evaluated, the median estimated daily intake is highest for toddlers and lowest for infants. For all five phthalates evaluated (except BBP exposure for for-mula-fed infants), food represents the most important source of exposure. The food categoriescontributing most to exposure depend upon the phthalate ester and the age group evaluated.Ingestion of dust and inhalation of indoor air represent the most important non-food sourcesof exposure to phthalate esters. Detection limits have a large influence on the estimated intakes.

A comparison of the results of the present study with studies that back-calculate phthalateester intake from urinary metabolite data suggests that exposure in the present study may beoverestimated for DEHP, BBP, and DBP due to changes in food processing over time (many ofthe measured concentrations of phthalates in food are not recent), loss of phthalates due tocooking has not been accounted for in the present study, and some measured concentrationsin food may be elevated due to background contamination. Conversely, exposure to DEP is underestimated in the present study because direct exposure to personal care products is notincluded. The overestimate of exposure to BBP and DBP from food, referred to above, may bepartially cancelled by the lack of inclusion of personal care products.

Keywords. Phthalate ester, Human exposure, Probabilistic analysis

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

1.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2291.2 Receptor Characteristics . . . . . . . . . . . . . . . . . . . . . . . 2291.2.1 Body Weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2301.2.2 Inhalation Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2301.2.3 Food Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . 2301.2.4 Drinking Water Consumption . . . . . . . . . . . . . . . . . . . . 2321.2.5 Soil Ingestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2331.2.6 Dust Ingestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2331.2.7 Time Spent Indoors . . . . . . . . . . . . . . . . . . . . . . . . . . 2331.3 Chemical Concentrations . . . . . . . . . . . . . . . . . . . . . . . 233

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 227–262DOI 10.1007/b11468

2 Dimethyl Phthalate (DMP) . . . . . . . . . . . . . . . . . . . . . . 234

2.1 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 2352.2 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 236

3 Diethyl Phthalate (DEP) . . . . . . . . . . . . . . . . . . . . . . . 236

3.1 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 2383.2 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 239

4 Dibutyl Phthalate (DBP) . . . . . . . . . . . . . . . . . . . . . . . 239

4.1 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 2424.2 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 243

5 Butylbenzyl Phthalate (BBP) . . . . . . . . . . . . . . . . . . . . . 246

5.1 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 2465.2 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 248

6 Bis(2-Ethylhexyl) Phthalate (DEHP) . . . . . . . . . . . . . . . . . 252

6.1 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 2526.2 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 254

7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258

8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

Abbreviations

BBP Butylbenzyl phthalateDBP Dibutyl phthalateDEHP Bis(2-ethylhexyl) phthalateDEP Diethyl phthalateDMP Dimethyl phthalate

1Introduction

Human exposure to five phthalate esters is evaluated for the following routes ofexposure: inhalation of air (indoors and outdoors), ingestion of drinking water,incidental ingestion of soil, ingestion of dust (indoors), and ingestion of food.Food is separated into the following categories: beverages excluding water, milk,cereals, dairy products excluding milk, eggs, fats and oils, fish, fruit products,grains, meats, nuts and beans, poultry, processed meats, vegetable products,“other” foods (includes soup, desserts, snacks), and, for infants, infant formulaand breast milk. Exposure to phthalate esters contained in children’s products or other consumer products is not evaluated in this chapter. The following phthalate esters are evaluated: DMP, DEP, DBP, BBP, and DEHP.

228 K. Clark et al.

1.1Methodology

Probabilistic analysis is used to estimate human exposure to each phthalate ester. The physical and behavioral characteristics of receptors in various ageclasses are first specified. The concentrations of each phthalate ester in eachmedium are also specified. A spreadsheet containing the appropriate receptorcharacteristics and chemical concentrations is then prepared and used to cal-culate the intake of each phthalate in micrograms per kilogram of body weightper day (µg kg–1 d–1). Excel (Microsoft Corporation) and Crystal Ball (Deci-sioneering Inc.) are then used to generate probability distributions for each exposure estimate, based on a simulation comprising 3000 iterations, and to produce a summary of the results.A sensitivity analysis is performed by CrystalBall, to identify the input variables contributing most to the variance in the ex-posure estimates.

The following three probability distributions are used in the present study:

– the lognormal distribution– the triangular distribution– the uniform distribution

Where the lognormal distribution is used, it is positively-skewed (i.e., most val-ues lie closer to the minimum value than the maximum). This distribution is usedto define most of the parameters in the present study, including: body weight andinhalation rate of the receptors; rates of food, drinking water, soil, and dust in-gestion; and the concentration of each phthalate ester in drinking water and inthe various food groups. In most cases, the standard deviation is set equal to 65%of the mean value. This results in a shape with a skewness coefficient of approx-imately 2.2.

A triangular probability distribution is typically used when only the mini-mum, most likely, and maximum values of a variable are known. The distributionis described by a triangle, with the minimum, most likely, and maximum valuesforming the vertices. Use of a triangular distribution tends to maximize the vari-ability in the parameter and, in comparison to the lognormal distribution, resultsin more frequent selection of values in the extremes of the distribution [1].Vari-ables for which a triangular distribution is used include the concentrations ofsome phthalate esters in indoor and outdoor air and soil.

The uniform distribution describes a situation in which any value between aspecified minimum and a maximum is equally likely. This distribution is used todescribe the amount of time spent indoors (any value between 20 and 24 h d–1 isconsidered equally probable).

1.2Receptor Characteristics

The population is divided into five age groups, which is consistent with previousevaluations undertaken by Environment Canada and Health Canada (EC&HC)[2, 3] and a previous evaluation of exposure to DEHP [4]:

Assessment of Critical Exposure Pathways 229

Adults – 20–70 years of ageTeens – 12–19 years of ageChildren – 5–11 years of ageToddlers – 7 months to 4 years of ageInfants – 0–6 months of age

Information describing the physical characteristics and habits of the Canadianpopulation is obtained primarily from Health and Welfare Canada (HWC) [5]. Inmost cases this document summarizes only mean or median values, so it is necessary to make assumptions regarding the shape of the probability distribu-tions for a given characteristic. The assumed probability distributions assignedto each characteristic are summarized in Table 1 and are described in the fol-lowing sections.Although some of the receptor characteristics may be correlated,correlations are not used in the present study. Finley et al. [1] found that corre-lating interdependent variables had little effect on lifetime risk estimates pro-vided that age-specific distributions were used.

1.2.1Body Weight

For the adult, teen, and child age groups, mean body weights (males and femalescombined) are derived from Stephens and Craig [6] who reported 1988 data.Stephens and Craig [6] did not report data for toddlers or infants. For these lat-ter age groups, mean body weights are modified from 1970 Nutrition CanadaSurvey data (Health Canada, unpublished), adjusting for a 2.9% increase in meanbody weight in children between 1970 (Health Canada, unpublished) and 1988[6]. Standard deviations for toddler and infant body weight are assumed to beequivalent to those measured in 1970.

1.2.2Inhalation Rate

Lognormal probability distributions are used to describe ranges of inhalationrates for the Canadian population. Recommended air intake rates reported toHealth Canada [7] are used for each age group.

1.2.3Food Consumption

HWC [5] used the results of a Nutrition Canada survey, conducted during 1970 to1972, as the basis for estimating the rate of ingestion of various individual foods,for the different age groups. The survey involved detailed dietary surveys of over13,000 individuals, based on a 24-hour recall method. HWC [5] also used infor-mation from Statistics Canada to comment on changes in food consumption pat-terns since the Nutrition Canada survey, in order that appropriate adjustmentsmay be made to the assumed ingestion rates. Important changes in food con-sumption patterns between 1970 and 1989 include decreases in consumption ofwhole milk and eggs, and increases in consumption of 2% milk, poultry, and fish.

230 K. Clark et al.

Assessment of Critical Exposure Pathways 231Ta

ble1

.R

ecep

tor

char

acte

rist

icsa

Inpu

t par

amet

erU

nits

Adu

ltTe

enC

hild

Todd

ler

Infa

nt20

–70

y12

–19

y5–

11y

0.5-

4y

0–0.

5y

Gen

eral

rece

ptor

cha

ract

eris

tics

Body

wei

ght

kgLN

(71;

14)

LN (6

0;14

)LN

(27;

7.3)

LN (1

5;3.

8)LN

(7.5

;3.2

)In

hala

tion

rat

em

3d–1

LN (1

6;3.

9)LN

(16;

4.0)

LN (1

5;3.

2)LN

(9.3

;2.6

)LN

(2.1

;0.5

7)R

ecep

tor

inge

stio

n ra

tesb

Tap

wat

erL

d–1LN

(0.8

0;0.

52)

LN (1

.0;0

.67)

LN (1

.1;0

.70)

LN (0

.70;

0.46

)LN

(0.8

;0.5

2)B

ever

ages

L d–1

LN (0

.96;

0.62

)LN

(0.4

3;0.

28)

LN (0

.23;

0.15

)LN

(0.1

2;0.

08)

NA

Cer

eals

g d–1

LN (2

7;16

)LN

(24;

15)

LN (3

4;22

)LN

(42;

27)

NA

Dai

ry p

rodu

cts

(exc

l.m

ilk)

g d–1

LN (5

3;34

)LN

(50;

33)

LN (4

5;29

)LN

(38;

25)

NA

Eggs

g d–1

LN (3

2;21

)LN

(22;

14)

LN (2

1;14

)LN

(24;

16)

NA

Fats

and

oils

g d–1

LN (2

5;16

)LN

(29;

19)

LN (2

1;14

)LN

(11;

7.1)

NA

Fish

g d–1

LN (1

4;9.

0)LN

(11;

7.3)

LN (8

.4;5

.5)

LN (3

.4;2

.2)

NA

Frui

tsg

d–1LN

(190

;120

)LN

(160

;100

)LN

(200

;130

)LN

(190

;120

)N

AG

rain

sg

d–1LN

(160

;100

)LN

(210

;130

)LN

(190

;120

)LN

(90;

58)

NA

Mea

tsg

d–1LN

(95;

61)

LN (9

3;60

)LN

(55;

36)

LN (3

8;25

)N

AM

ilkL

d–1LN

(0.2

30;0

.15)

LN (0

.523

;0.3

4)LN

(0.5

64;0

.37)

LN (0

.632

;0.4

1)N

AN

uts

and

bean

sg

d–1LN

(28;

18)

LN (3

1;20

)LN

(24;

15)

LN (1

5;9.

7)N

AO

ther

food

sg

d–1LN

(220

;144

)LN

(25

0;16

0)LN

(210

;140

)LN

(270

;180

)N

APo

ultr

yg

d–1LN

(21;

14)

LN (2

0;13

)LN

(17;

11)

LN (1

3;8.

6)N

APr

oces

sed

mea

tsg

d–1LN

(22;

14)

LN (2

3;15

)LN

(19;

12)

LN (1

1;7.

0)N

AVe

geta

bles

g d–1

LN (2

30;1

50)

LN (2

40;1

50)

LN (1

90;1

20)

LN (1

20;7

6)N

AIn

fant

form

ula

(pow

der)

g d–1

NA

NA

NA

NA

LN (1

30;8

5)Br

east

milk

L d–1

NA

NA

NA

NA

LN (0

.750

;0.4

9)To

tal m

ax.f

ood

cons

umpt

ion

cg

d–123

0021

0018

0015

0082

0In

cide

ntal

soi

lm

g d–1

LN (4

0;10

0)LN

(40;

100)

LN (4

0;10

0)LN

(40;

100)

LN (4

0;10

0)In

cide

ntal

dus

tm

g d–1

LN (4

0;10

0)LN

(40;

100)

LN (4

0;10

0)LN

(40;

100)

LN (4

0;10

0)Ex

posu

re fr

eque

ncy

Tim

e sp

ent i

ndoo

rsh

d–1U

(20;

24)

U (2

0;24

)U

(20;

24)

U (2

0;24

)U

(20;

24)

aSe

e te

xt fo

r re

fere

nces

;NA

not

app

licab

le;U

(min

imum

;max

imum

) uni

form

dis

trib

utio

n;LN

(mea

n;st

anda

rd d

evia

tion

) log

norm

al d

istr

ibut

ion.

bSt

anda

rd d

evia

tion

ass

umed

to b

e 65

% o

fthe

mea

n va

lue.

cLo

gnor

mal

dis

trib

utio

ns fo

r foo

d gr

oup

cons

umpt

ion

and

beve

rage

con

sum

ptio

n ha

ve b

een

trun

cate

d w

ith

an u

pper

lim

it o

fthe

tota

l max

imum

con

-su

mpt

ion

offo

od.

Although the type of milk consumed has changed, the overall rate of milk con-sumption has not changed substantially [8]. Similarly, although there have beenchanges in the type of meat consumed, overall meat consumption has notchanged markedly.

Infant feeding practices have also changed since the Nutrition Canada survey.In the period from 1965 to 1971, HWC [5] reports that only 25% of mothersbreast-fed their babies and more than three quarters of this group had discon-tinued by one to two months. Data as of 1990 indicate that breast-feeding initia-tion rates are close to 80%, with 30% still continuing to breast feed at six months[5]. HWC reports that the Nutrition Canada survey likely overestimates the con-sumption of solid foods by infants due to changes in the age when solid foods areintroduced to infants. Solid foods are introduced to approximately 50% of babiesby four months, and 89.5% of babies by six months, in contrast to the results ofthe Nutrition Canada survey in which solid foods are generally introduced to in-fants’ diets at a much earlier age. HWC recommends that a typical infant shouldbe considered to be exclusively breast-fed up to six months of age and to consumeapproximately 750 mL of breast milk per day. Alternatively, the typical infantcould be considered to be exclusively formula-fed for the first six months. Such isthe approach taken in this exposure assessment; for infants up to age six months,two types of exposure are calculated – one for infants consuming breast milk exclusively and the other for infants consuming formula exclusively. Measured phthalate concentrations are available for both ready-to-feed liquid formula andpowdered formula. Consistent with a study of phthalate exposure to children, per-formed by Zaleski et al. [9], the mass of powdered formula is assumed to repre-sent one-seventh of the total mass of formula ingested. Only very limited data areavailable for “baby food”for DEHP and, as these data fall within the range of avail-able measurements for DEHP in fruit and vegetable products used for all ageclasses, the data for “baby food” are not used in the exposure assessment.

For the present study, composite food groups are developed for all age groups,other than infants. The composite food groups are: beverages excluding water,milk, cereals, dairy products excluding milk, eggs, fats and oils, fish, fruit prod-ucts, grains, meats, nuts and beans, other foods (includes soups, desserts, snacks),poultry, processed meats, and vegetable products. Mean consumption rates foreach age group (summarized in Table 1) are calculated by adding the consump-tion rates for all foods obtained from [5], in each composite food group. Log-normal distributions are used to define food consumption rates for each agegroup. The standard deviations are defined as being 65% of the mean values toobtain positively skewed distributions. Maximum values are stipulated for eachdistribution; maxima are set equal to the age-appropriate total daily food consumption (2300 g d–1 for adults, 2100 g d–1 for teens, 1800 g d–1 for children,1500 g d–1 for toddlers, and 820 g d–1 for infants) [5,10].

1.2.4Drinking Water Consumption

Tap water consumption rates for each age group are determined by calculatingage-weighted averages from the data in HWC [5] and subtracting the tea and cof-

232 K. Clark et al.

fee consumption rates from the same document (because tea and coffee con-sumption are included in the “other beverages” food group). The resulting meanvalues (summarized in Table 1) are used to define positively skewed lognormaldistributions, with standard deviations assumed equal to 65% of the mean val-ues. For infants, the mean tap water ingestion rate is assumed to be 800 mL d–1

for infants consuming powdered infant formula [5].

1.2.5Soil Ingestion

The distribution for assumed soil ingestion is derived from Stanek and Calabrese[11]. A lognormal distribution is created with an arithmetic average intake of40 mg d–1 and a standard deviation of 100, which gives a 95th percentile intakerate of approximately 200 mg d–1. There are insufficient age-related data to definedifferent rates of soil ingestion for different age groups [11]. Most quantitativesoil ingestion data have been collected for children [12]. Therefore, in the absenceof any quantitative data suggesting otherwise, the same assumption is applied toall age groups.

1.2.6Dust Ingestion

Due to a lack of available data, the average intake and standard deviation for dustingestion are assumed to be the same as that for soil ingestion. Both the dust andthe soil ingestion rates are highly uncertain due to difficulties in measuring theserates.Also, there is a large degree of variability in these ingestion rates, from oneindividual to another.

1.2.7Time Spent Indoors

The number of hours a person spends indoors each day is represented by a uni-form distribution. It is assumed that the total time spent indoors (at all locations)is 20–24 hours per day, with any value in the range considered equally likely. Thisdistribution is used for all age groups. The amount of time spent indoors is usedto calculate exposure from breathing indoor air and the remainder is used to cal-culate exposure from breathing outdoor air.

1.3Chemical Concentrations

Distributions of concentrations of each phthalate ester, in the various exposuremedia, are assigned by using the compilation of data reported elsewhere [13] andsummarized in Chapter 5. These concentrations, and the assigned distributioncharacteristics, are further described below. Note that, due to limited data, thedatasets from different regions (Canada, United States, Europe, and Japan/Asia)are combined. Exposure estimates are not made for individual regions.

Assessment of Critical Exposure Pathways 233

2Dimethyl Phthalate (DMP)

The assigned distributions for the concentration of DMP in each medium usedfor the exposure assessment are summarized in Table 2. These concentration dis-tributions are obtained from the report prepared for the American ChemistryCouncil [13]. For DMP, few data are available for food. Measurements of DMP infish, milk, and vegetables are all reported as not detected. One of the larger stud-ies of phthalate ester concentrations in food [14] indicated that the presence ofDMP was evaluated, but was not found. Therefore, to obtain an approximate es-timate of DMP exposure, the mean concentration of DMP is assumed to be equalto one half of the reported detection limit for DEP, for the food categories forwhich no data are available.

As shown in Table 2, the median estimated daily intake of DMP for each groupin mg kg–1 d–1 is: adults (0.7), teens (0.7), children (1.4), and toddlers (1.6). The es-timated intake for infants, from non-food sources, is 0.05 µg kg–1 d–1 and 0.01 µgkg–1 d–1, for formula-fed and breast-fed infants, respectively.

Table 2 also presents the estimated percentage intake of DMP for each agegroup via each medium. As shown, for all age groups, food represents the dom-

234 K. Clark et al.

Table 2. Dimethyl phthalate – exposure estimates a

ConcentrationMedium Units Dist. Min. Mean Max. Std. Dev.

Outdoor air µg m–3 LN 0.0026 0.0017Indoor air µg m–3 LN 0.02 0.013Drinking water µg L–1 LN 0.5 0.33Ingested soil µg g–1 C 0Ingested dust µg g–1 LN 1.73 1.12

FoodBeverages excl. water µg L–1 LN 25 b 16Milk µg L–1 LN 1.25 0.8Cereals µg g–1 LN 0.02 b 0.013Dairy products µg g–1 LN 0.05 b 0.03Eggs µg g–1 LN 0.05 b 0.03Fats and oils µg g–1 LN 0.25 b 0.16Fish µg g–1 LN 0.005 0.003Fruit products µg g–1 LN 0.02 b 0.013Grains µg g–1 LN 0.05 b 0.03Meats µg g–1 LN 0.05 b 0.03Nuts and beans µg g–1 LN 0.045 b 0.029Other foods µg g–1 LN 0.005 b 0.003Poultry µg g–1 LN 0.05 b 0.03Processed meats µg g–1 LN 0.05 b 0.03Vegetable products µg g–1 LN 0.005 0.003Infant formula – powder µg g–1 C 0Breast milk µg g–1 C 0

a Dist. distribution type; LN log normal; C constant; NA not available or not applicable.b Concentration based on one half of assumed detection limit.

Assessment of Critical Exposure Pathways 235

Table 2 (continued)

Intake Adult Teen Child Toddler Infant

Formula- Breast-fed fed

Total daily in- mg kg–1 d–1 0.7 0.7 1.4 1.6 0.05 0.01take (median)

Total daily intake (% of total)Outdoor Air 0.0 0.0 0.0 0.0 0.1 0.4Indoor Air 0.5 0.6 0.7 0.7 8.2 36.0Drinking water 0.7 1.0 1.2 1.3 77.3 0.0Ingested soil 0.0 0.0 0.0 0.0 0.0 0.0Ingested dust 0.1 0.1 0.2 0.3 14.4 63.5

FoodBeverages excl. water 39.0 22.4 15.3 11.7 0.0 0.0Cereals 1.0 1.0 1.8 3.4 0.0 0.0Dairy products (excl. milk) 4.7 5.1 5.7 7.0 0.0 0.0Eggs 2.9 2.3 2.8 4.9 0.0 0.0Fats and oils 11.0 15.3 14.1 11.0 0.0 0.0Fish 0.2 0.1 0.1 0.1 0.0 0.0Fruit products 6.6 6.7 10.7 15.3 0.0 0.0Grains 14.2 21.6 25.3 18.0 0.0 0.0Meats 8.4 9.8 7.3 7.8 0.0 0.0Milk 0.5 1.4 1.9 3.2 0.0 0.0Nuts and beans 2.2 2.9 2.8 2.7 0.0 0.0Other foods 2.0 2.6 2.8 5.5 0.0 0.0Poultry 1.9 2.1 2.2 2.7 0.0 0.0Processed meats 1.9 2.4 2.5 2.2 0.0 0.0Vegetable products 2.1 2.5 2.5 2.3 0.0 0.0Infant formula/breast milk NA NA NA NA NA NA

Total food 98.7 98.3 97.9 97.8 NA NA

Total 100 100 100 100 100 100

inant source of exposure, accounting for more than 97% of exposure. It is not appropriate to identify any particular food group as more important than an-other, as the relative importance of the various foods is controlled by the detec-tion limits assumed.

2.1Sensitivity Analysis

The input parameters contributing most to the variation in the exposure esti-mates are identified by Crystal Ball and are shown below. The values in paren-theses indicate the percentage contribution to variance. Note that only para-meters contributing 5% or more to the variance are listed:

– Adult – concentration of DMP in beverages (26.9%), body weight (24.6%),ingestion rate of beverages (24.2%), ingestion rate of grains (5.2%).

– Teen – body weight (37.3%), concentration of DMP in beverages (11.8%),ingestion rate of beverages (9.3%), ingestion rate of grains (8.7%), concen-tration in grains (8.6%).

– Child – body weight (47.4%), ingestion rate of grains (9.9%), concentration ingrains (9.7%), ingestion rate of fats and oils (5.4%).

– Toddler – body weight (47.4%), ingestion rate of grains (7.3%), concentrationin fruit (7.1%), concentration in grains (6.6%), ingestion rate of fruit (5.7%).

– Formula-fed infant – concentration in drinking water (31.3%), ingestion rateof drinking water (31.0%), body weight (27.4%), ingestion rate of dust (5.2%).

– Breast-fed infant – ingestion rate of dust (47.3%), body weight (22.2%), con-centration in indoor air (15.4%), and concentration in dust (7.3%).

The parameters listed above are used to quantify exposure for the pathwaysfound to be the dominant sources of DMP exposure but, as described above, therelative importance of the various pathways is largely controlled by the detectionlimits assumed for each food group/pathway. Reduction in the variability and/orthe uncertainty in the above parameters would reduce the uncertainty in the exposure estimates.

2.2Comparison to Other Studies

No other studies that evaluated human exposure to DMP have been identified.

3Diethyl Phthalate (DEP)

The assigned distributions for the concentration of DEP in each medium used forthe exposure assessment are summarized in Table 3. These concentration distri-butions are obtained from the report prepared for the American ChemistryCouncil [13]. For DEP, data are not available for infant formula or breast milk.

As shown in Table 3, the median estimated daily intake of DEP for each groupin mg kg–1 d–1 is: adults (2.5), teens (3.0), children (5.7), and toddlers (10.6). Theestimated intake for infants, from non-food sources, is 0.2 µg kg–1d–1.

Table 3 also presents the estimated percentage intake of DEP for each agegroup via each medium. As shown, for all age groups, food represents the dom-inant source of exposure, accounting for more than 95% of exposure. The otherfoods category is the largest source of exposure, representing 61–78% of expo-sure.

For the non-infant, inhalation of indoor air represents the most importantsource of non-food exposure, accounting for 4.4% of the exposure to DEP foradults. Inhalation of outdoor air, ingestion of drinking water, soil, and dust, com-bined, represent less than 1% of exposure. For the infant, where food data are notavailable, inhalation of indoor air is the most important pathway of exposure,followed by ingestion of dust.

236 K. Clark et al.

Assessment of Critical Exposure Pathways 237

Table 3. Diethyl phthalate – exposure estimates a

ConcentrationMedium Units Dist. Min. Mean Max. Std. Dev.

Outdoor air µg m–3 LN 0.039 0.025Indoor air µg m–3 LN 0.621 0.40Drinking water µg L–1 LN 0.5 0.33Ingested soil µg g–1 LN 0.62 0.40Ingested dust µg g–1 LN 2.7 1.8

FoodBeverages excl. water µg L–1 LN 27 18Milk µg L–1 LN 2 1.3Cereals µg g–1 LN 0.11 0.07Dairy products µg g–1 LN 0.05 0.03Eggs µg g–1 LN 0.05 0.03Fats and oils µg g–1 LN 0.25 0.16Fish µg g–1 LN 0.059 0.038Fruit products µg g–1 LN 0.076 0.049Grains µg g–1 LN 0.05 0.03Meats µg g–1 LN 0.05 0.03Nuts and beans µg g–1 LN 0.045 0.029Other foods µg g–1 LN 0.58 0.38Poultry µg g–1 LN 0.05 0.03Processed meats µgg–1 LN 0.05 0.03Vegetable products µg g–1 LN 0.005 0.003Infant formula – powder µg g–1 C 0Breast milk µg g–1 C 0

Intake Adult Teen Child Toddler Infant

Formula-fed Breast-fed

Total daily in- µg kg–1 d–1 2.5 3.0 5.7 10.6 0.2 0.2take (median)

Total daily intake (% of total)Outdoor air 0.0 0.0 0.0 0.0 0.4 0.5Indoor air 4.4 4.2 4.5 2.6 70.6 89.9Drinking water 0.2 0.2 0.2 0.2 21.4 0.0Ingested soil 0.0 0.0 0.0 0.0 1.4 1.8Ingested dust 0.1 0.1 0.1 0.1 6.2 7.8

FoodBeverages excl. water 11.4 5.4 3.4 1.6 0.0 0.0Cereals 1.4 1.2 2.1 2.3 0.0 0.0Dairy products (excl. milk) 1.3 1.2 1.2 0.9 0.0 0.0Eggs 0.8 0.5 0.6 0.6 0.0 0.0Fats and oils 3.0 3.4 2.9 1.3 0.0 0.0Fish 0.6 0.3 0.3 0.1 0.0 0.0Fruit products 6.8 5.7 8.3 7.1 0.0 0.0Grains 3.8 4.8 5.2 2.2 0.0 0.0Meats 2.3 2.2 1.5 0.9 0.0 0.0

a Dist. distribution type; LN log normal; C constant; NA not available or not applicable.

3.1Sensitivity Analysis

The input parameters contributing most to the variation in the exposure esti-mates are identified by Crystal Ball and are shown below. The values in paren-theses indicate the percentage contribution to variance. Note that only parame-ters contributing 5% or more to the variance are listed:

– Adult – ingestion rate of other foods (36.9%), concentration of DEP in otherfoods (36.9%), body weight (15.2%).

– Teen – concentration in other foods (39.3%), ingestion rate of other foods(37.9%), body weight (15.1%).

– Child – concentration in other foods (36.4%), ingestion rate of other foods(34.9%), body weight (20.8%).

– Toddler – concentration in other foods (40.2%), ingestion rate of other foods(38.6%), body weight (15.6%).

– Formula-fed infant – concentration in indoor air (38.9%), body weight(38.0%), inhalation rate (9.6%).

– Breast-fed infant – concentration in indoor air (52.7%), body weight (29.2%),inhalation rate (11.9%).

The parameters listed above are used to quantify exposure for the pathwaysfound to be the dominant sources of DEP exposure (i.e., ingestion of other foods,inhalation of indoor air). Body weight is also an important parameter, as the in-takes are expressed on a “per body weight” basis. Reduction in the variabilityand/or the uncertainty in the above parameters would reduce the uncertainty inthe exposure estimates.

238 K. Clark et al.

Table 3 (continued)

Intake Adult Teen Child Toddler Infant

Formula-fed Breast-fed

Milk 0.2 0.5 0.6 0.6 0.0 0.0Nuts and beans 0.6 0.7 0.6 0.3 0.0 0.0Other foods 61.5 68.1 67.2 78.3 0.0 0.0Poultry 0.5 0.5 0.5 0.3 0.0 0.0Processed meats 0.5 0.5 0.5 0.3 0.0 0.0Vegetable products 0.6 0.6 0.5 0.3 0.0 0.0Infant formula/breast milk NA NA NA NA NA NA

Total food 95.3 95.5 95.2 97.1 NA NA

Total 100 100 100 100 100 100

3.2Comparison to Other Studies

Several other studies have evaluated human exposure to DEP. Table 4 presents acomparison of the results of these studies with those of the present study.

The Centers for Disease Control and Prevention (CDC) measured monoestermetabolites of phthalate esters, including DEP, in urine from a population of289 adult humans [15]. These measurements integrate exposure from all routes(oral, inhalation, dermal, and intravenous) and all sources (environmental me-dia, food, and consumer products). David [16] and Kohn et al. [17] used the CDCurinary metabolite data to back-calculate the intake of various phthalate esters.As shown in Table 4, the calculated geometric mean intake from David [16] andthe median intake from Kohn et al. [17] are approximately a factor of five greaterthan the estimated median intake in the present study.As described below, theseresults are in contrast to those for DEHP, BBP, and DBP, for which the presentstudy overestimates the intake relative to the intake estimated from urinarymetabolite data. The most likely reason for the discrepancy, for DEP, is that DEPis used in consumer products and direct exposure to consumer products is notincluded in the present study. In two more recent studies, the CDC reports mea-surements of monoester metabolites of phthalate esters in urine from 1000 in-dividuals, age 6 through adult [18] and in the urine of toddlers aged 12 to18 months [19]. David has back-calculated the intake of DEP for these studies[20, 21] by using the same method as in David [16] and the results are presentedin Table 4. As shown in Table 4, the intakes back-calculated for the CDC studies[18, 19] are comparable to the intakes estimated in the present study.

4Dibutyl Phthalate (DBP)

The assigned distributions for the concentration of DBP in each medium used forthe exposure assessment are summarized in Table 5. These concentration distri-butions are obtained from the report prepared for the American ChemistryCouncil [13]. For DBP, data are available for all media and food categories.

As shown in Table 5, the median estimated daily intake of DBP for each groupin mg kg–1 d–1 is: adults (5.6), teens (6.4), children (11), toddlers (14), formula-fedinfants (1.5), and breast-fed infants (2.9). These results are compared to the re-sults of other studies in Sect. 4.2.

Table 5 also presents the estimated percentage intake of DBP for each agegroup via each medium. As shown, for all age groups, food represents the dom-inant source of exposure, accounting for 63–96% of exposure. The food categorycontributing most to total exposure depends upon the age group considered but,in general, the most important sources of exposure are: meats, other foods, bev-erages excluding water, fats and oils, grains, vegetables, and infant formula andbreast milk.

Inhalation of indoor air and ingestion of dust represent the most importantsources of non-food exposure, accounting for between 4% and 34% of exposure,depending upon the age group. Inhalation of outdoor air, ingestion of drinking

Assessment of Critical Exposure Pathways 239

240 K. Clark et al.

Tabl

e4.

Com

pari

son

ofD

EP e

xpos

ure

esti

mat

es w

ith

esti

mat

es fr

om o

ther

stu

dies

Stud

yD

EP in

take

(µg

kg–1

d–1)

Adu

ltTe

enC

hild

Todd

ler

Infa

nt

Pres

ent s

tudy

2.5

3.0

5.7

10.6

Non

-foo

d on

ly:

Prob

abili

stic

ana

lysi

s;m

edia

n in

take

0.

2 (f

orm

ula-

fed)

(all

expo

sure

pat

hway

s ex

clud

ing

0.2

(bre

ast-

fed)

child

ren’

s an

d ot

her

cons

umer

pro

duct

s)Fo

od o

nly

2.4

2.9

5.4

10.3

NA

Dav

id [1

6]12

.34

(geo

met

ric

mea

n)N

Aa

NA

NA

NA

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

93

.33

(95t

h pe

rcen

tile

)da

ta (B

loun

t et a

l.[1

5])

242.

81 (h

ighe

st v

alue

)

Koh

n et

al.

[17]

12 (m

edia

n)N

AN

AN

AN

AIn

take

cal

cula

ted

from

uri

nary

met

abol

ite

110

(95t

h pe

rcen

tile

)da

ta (B

loun

t et a

l.[1

5])

320

(hig

hest

val

ue)

AC

C [2

0]A

ge 6

thro

ugh

adul

t:In

take

cal

cula

ted

from

uri

nary

met

abol

ite

5.42

(geo

met

ric

mea

n)da

ta [

18]

31.9

0 (9

0th

perc

enti

le)

50.0

6 (9

5th

perc

enti

le)

Dav

id [2

1]6.

29 (

geom

etri

c m

ean)

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

36

.72

(95t

h pe

rcen

tile

)da

ta [1

9]42

.95

(hig

hest

val

ue)

aN

A n

ot a

vaila

ble.

Assessment of Critical Exposure Pathways 241

Table 5. Dibutyl phthalate – exposure estimates

ConcentrationMedium Units Dist. Min. Mean Max. Std. Dev.

Outdoor air µg m–3 T 0.00008 0.016 0.38Indoor air µg m–3 LN 1.0Drinking water µg L–1 LN 0.5 0.33Ingested soil µg g–1 LN 0.18Ingested dust µg g–1 LN 56.7

FoodBeverages excl. water µg L–1 LN 100 65Milk µg L–1 LN 12 7.8Cereals µg g–1 LN 0.3 0.20Dairy products (excl. milk) µg g–1 LN 0.04 0.03Eggs µg g–1 LN 0.09 0.06Fats and oils µg g–1 LN 2.5 1.6Fish µg g–1 LN 0.23 0.15Fruit products µg g–1 LN 0.033 0.02Grains µg g–1 LN 0.26 0.17Meats µg g–1 LN 0.92 0.60Nuts and beans µg g–1 LN 0.18 0.12Other foods µg g–1 LN 0.16 0.10Poultry µg g–1 LN 0.13 0.08Processed meats µg g–1 LN 0.54 0.35Vegetable products µg g–1 LN 0.17 0.111Infant formula – powder µg g–1 LN 0.07 0.05Breast milk µg g–1 LN 0.03 0.02

Intake Adult Teen Child Toddler Infant

Formula- Breast-fed fed

Total daily in- µg kg–1 d–1 5.6 6.4 11 14 1.5 2.9take (median)

Total daily intake (% of total)Outdoor air 0.0 0.0 0.1 0.0 0.2 0.1Indoor air 3.6 3.6 4.4 4.0 15.2 7.7Drinking water 0.1 0.1 0.1 0.1 2.9 0.0Ingested soil 0.0 0.0 0.0 0.0 0.1 0.0Ingested dust 0.5 0.6 0.7 1.1 19.1 9.7

FoodBeverages excl. water 21.0 10.7 7.6 5.5 0.0 0.0Cereals 1.9 1.8 3.4 5.9 0.0 0.0Dairy products (excl. milk) 0.5 0.5 0.6 0.7 0.0 0.0Eggs 0.7 0.5 0.6 1.0 0.0 0.0Fats and oils 14.8 18.3 17.4 12.8 0.0 0.0Fish 1.2 0.6 0.6 0.4 0.0 0.0Fruit products 1.5 1.3 2.2 2.9 0.0 0.0Grains 10.0 13.5 16.3 10.9 0.0 0.0Meats 20.8 21.5 16.6 16.6 0.0 0.0

water, and ingestion of soil, combined, represent less than 1% of exposure (except for the formula-fed infant where ingestion of drinking water accounts for2.9% of exposure).

4.1Sensitivity Analysis

The input parameters contributing most to the variation in the exposure estimates are identified by Crystal Ball and are shown below. The values in paren-theses indicate the percentage contribution to variance. Note that only para-meters contributing 5% or more to the variance are listed:

– Adult – body weight (31.0%), ingestion rate of meats (10.6%), concentrationof DBP in beverages (10.5%), ingestion rate of beverages (9.0%), concentrationin meats (8.9%), concentration in fats and oils (6.8%).

– Teen – body weight (36.1%), concentration of DBP in meats (10.1%), ingestionrate of meats (10.0%), ingestion rate of fats and oils (7.3%), concentration infats and oils (6.9%), and concentration in grains (5.0%).

– Child – body weight (46.3%), ingestion rate of fats and oils (6.5%), ingestionrate of meats (6.0%), concentration of DBP in grains (5.8%), ingestion rate ofgrains (5.0%), concentration in fats and oils (5.0%).

– Toddler – body weight (45.9%), concentration of DBP in other foods (8.7%),ingestion rate of other foods (8.4%), concentration in meats (6.3%), ingestionrate of meats (5.5%).

– Formula-fed infant – body weight (33.7%), concentration in formula (25.8%),ingestion rate of infant formula (22.6%), ingestion rate of dust (7.8%).

– Breast-fed infant – concentration in breast milk (37.5%), ingestion rate ofbreast milk (29.0%), body weight (25.7%).

242 K. Clark et al.

Table 5 (continued)

Intake Adult Teen Child Toddler Infant

Formula- Breast-fed fed

FoodMilk 0.7 1.6 2.2 3.6 0.0 0.0Nuts and beans 1.2 1.4 1.4 1.3 0.0 0.0Other foods 8.5 10.0 11.2 20.5 0.0 0.0Poultry 0.7 0.7 0.7 0.8 0.0 0.0Processed meats 2.8 3.2 3.4 2.7 0.0 0.0Vegetable products 9.5 10.1 10.5 9.3 0.0 0.0Infant formula/breast milk NA NA NA NA 62.7 82.6

Total food 95.8 95.6 94.7 94.7 62.7 82.6

Total 100 100 100 100 100 100

a Dist. distribution type; T triangular; LN log normal; NA not available or not applicable.

The parameters listed above are used to quantify exposure for the pathwaysfound to be the dominant sources of DBP exposure (i.e., ingestion of meats, otherfoods, beverages excluding water, fats and oils, and infant formula and breastmilk). Body weight is also an important parameter, as the intakes are expressedon a “per body weight” basis. Reduction in the variability and/or the uncertaintyin the above parameters would reduce the uncertainty in the exposure estimates.

4.2Comparison to Other Studies

Several other studies have evaluated human exposure to DBP [3, 16, 17, 20–24].Table 6 presents a comparison of the results of these studies with those of the present study.

As described in Sect. 3.2, David [16] and Kohn et al. [17] used the CDC urinarymetabolite data to back-calculate the intake of various phthalate esters.As shownin Table 6, the calculated geometric mean intake from David [16] is a factor of3.6 less than the estimated median intake in the present study and the 95th per-centile intake is slightly larger than the median intake in the present study. Themedian and 95th percentile intakes from Kohn et al. [17] are comparable to thegeometric mean and 95th percentile intakes in David [16]. Similarly, the medianintakes in the present study are larger than the intakes back-calculated [20, 21]from the more recent CDC studies [18, 19]. Comparison of the results of the present study with the studies using urinary metabolite data suggests that humanexposure to DBP may be overestimated in the present study in terms of exposurevia food, but may be underestimated in terms of exposure to consumer products.Possible sources of the overestimation include:

– Changes in food processing technology with time have not been evaluated.Allavailable data on measurements of DBP in food have been pooled and a largeportion of the data was obtained in the late 1980s. The data may not reflectcurrent concentrations in food.

– The reported concentrations in food are generally prior to any food prepara-tion. Cooking of the food, and removal of fats during cooking, may reduce theconcentration of DBP in the food that is ultimately consumed.

– Background contamination in analysis of food samples. Background con-centrations may be determined in the analysis, but often the measured con-centrations are not corrected for the blank concentrations.

MAFF [22] estimated the intake of DBP to infants at birth and at six months ofage, due to ingestion of infant formula. MAFF found that the estimated intake decreased by a factor of about six from 1996 to 1998. The estimated intake for asix month old, using the 1998 data, is comparable to the estimated intake for theinfant in the present study.

The International Program on Chemical Safety (IPCS) [24] estimated adult ex-posure to DBP using the 1986 Canadian market basket survey data (the same dataused by EC&HC [3]) and found food to be the dominant source of exposure. Asshown in Table 6, the results of the present study are comparable to those of IPCS[24].

Assessment of Critical Exposure Pathways 243

244 K. Clark et al.

Tabl

e6.

Com

pari

son

ofD

BP e

xpos

ure

esti

mat

es w

ith

esti

mat

es fr

om o

ther

stu

dies

Stud

yD

BP in

take

(µg

kg–1

d–1)

Adu

ltTe

enC

hild

Todd

ler

Infa

nt

Pres

ent s

tudy

5.6

6.4

1114

1.5

(for

mul

a-fe

d)Pr

obab

ilist

ic a

naly

sis;

med

ian

inta

ke (a

ll2.

9 (b

reas

t-fe

d)ex

posu

re p

athw

ays

excl

udin

g ch

ildre

n’s

and

othe

r co

nsum

er p

rodu

cts)

Food

onl

y5.

46.

110

130.

9 (f

orm

ula-

fed)

2.4

(bre

ast-

fed)

Dav

id [1

6]1.

56 (g

eom

etri

c m

ean)

NA

aN

AN

AN

AIn

take

cal

cula

ted

from

uri

nary

met

abol

ite

6.87

(95t

h pe

rcen

tile

)da

ta (B

loun

t et a

l.[1

5])

116.

96 (h

ighe

st v

alue

)K

ohn

et a

l.[1

7]0.

084

(min

imum

)N

AN

AN

AN

AIn

take

cal

cula

ted

from

uri

nary

met

abol

ite

1.5

(med

ian)

data

(Bl

ount

et a

l.[1

5])

7.2

(95t

h pe

rcen

tile

)11

0 (h

ighe

st v

alue

)A

CC

[20]

Age

6th

roug

h ad

ult:

0.90

(geo

met

ric

mea

n)In

take

cal

cula

ted

from

uri

nary

met

abol

ite

2.70

(90

th p

erce

ntile

);3.

64 (

95th

per

cent

ile)

data

[18]

Dav

id [2

1]2.

45 (

geom

etri

c m

ean)

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

16

.57

(95t

h pe

rcen

tile

)da

ta [1

9]18

0.7

(hig

hest

val

ue)

MA

FF [2

2]N

AN

AN

AN

A19

98da

ta:2

.4 (a

t bir

th)

Esti

mat

ed in

take

due

to in

gest

ion

of1.

4 (a

t age

6m

onth

sin

fant

form

ula)

1996

data

:14

(at b

irth

)9.

3 (a

t age

6m

onth

s)IP

CS

[24]

7.1

NA

NA

NA

NA

Use

d da

ta fo

r co

ncen

trat

ions

in fo

od

as in

EC

&H

C [3

]Fo

od o

nly

7N

AN

AN

AN

A

Assessment of Critical Exposure Pathways 245

Tabl

e6

(con

tinu

ed)

Stud

yD

BP in

take

(µg

kg–1

d–1)

Adu

ltTe

enC

hild

Todd

ler

Infa

nt

MA

FF [2

3]0.

19 (m

ean)

NA

NA

NA

NA

Esti

mat

ed in

take

due

to in

gest

ion

ofca

r-0.

44 (9

7.5t

h pe

rcen

tile

)ca

ss m

eat,

poul

try,

eggs

and

milk

.C

onve

rted

inta

ke in

mg/

pers

on/d

ay b

y as

sum

ing

a bo

dy w

eigh

t of7

0kg

.

EC&

HC

[3]b

1.9

2.3

4.3

5.0

2.4

Det

erm

inis

tic

anal

ysis

(all

expo

sure

pat

h-w

ays

excl

udin

g ch

ildre

n’s

and

othe

r co

n-su

mer

pro

duct

s)Fo

od o

nly

1.1

1.4

3.2

4.1

1.6

aN

A n

ot a

vaila

ble.

bEC

&H

C [3

] est

imat

e ob

tain

ed u

sing

con

cent

rati

ons i

n in

divi

dual

food

item

s,ra

ther

than

food

cat

egor

ies,

as w

as d

one

in th

e pr

esen

t stu

dy (s

ee S

ect.

4.2)

.

MAFF [23] estimated the intake of DBP to adults from ingestion of carcassmeat, poultry, eggs, and milk. The mean estimated intake is approximately a factor of 30 lower than the estimated median total intake in the present study,because not all sources of exposure are included.

As shown in Table 6, the estimated intakes in the present study are approxi-mately three times greater than the estimated intakes in the EC&HC [3] study forall ages except the infant. In both studies, food is the most important source ofexposure; however, because the estimated intake from food is lower in theEC&HC study, the relative contribution from indoor air to total intake is greaterin the EC&HC study. Examination of the supporting documentation for theEC&HC study [25] shows that intake of DBP via food is calculated for individ-ual foods (rather than for food categories as is done in the present study). Thismethodology is consistent with that used by EC&HC [26] for BBP (explained inSect. 5.2). In cases where measurements are limited to only a few food items, thismethod may result in an underestimate of the total intake, as concentrations inmany food items are assigned a zero or laboratory detection limit rather than theconcentration measured in a related food item. For the infant, the estimated in-take is comparable between the present study and EC&HC [3].

5Butylbenzyl Phthalate (BBP)

The assigned distributions for the concentration of BBP in each medium used forthe exposure assessment are summarized in Table 7. These concentration distri-butions are obtained from the report prepared for the American ChemistryCouncil [13]. For BBP, data are not available for soil or breast milk.

As shown in Table 7, the median estimated daily intake of BBP for each groupin mg kg–1 d–1 is: adults (3.7), teens (5.7), children (7.9), toddlers (9.3), and for-mula-fed infants (1.5). These results are compared to the results of other studiesin Sect. 5.2.

Table 7 also presents the estimated percentage intake of BBP for each agegroup via each medium. As shown, for the non-infants, food represents between91% and 96% of exposure. The food category contributing most to total exposureis fats and oils, representing 51–67% of exposure depending upon the age group.

For the formula-fed infant, ingestion of formula represents 27% of exposure,while ingestion of dust accounts for 70% of exposure, and ingestion of drinkingwater accounts for 2% of exposure. Ingested dust represents the most importantsource of non-food exposure for the other age groups, accounting for 4–9% ofthe exposure to BBP depending upon the age group. Inhalation of air and inges-tion of drinking water, combined, represent less than 1% of exposure. As men-tioned above, no data are available with which to evaluate exposure to soil.

5.1Sensitivity Analysis

The input parameters contributing most to the variation in the exposure esti-mates are identified by Crystal Ball and are shown below. The values in paren-

246 K. Clark et al.

Assessment of Critical Exposure Pathways 247

Table 7. Butyl benzyl phthalate – exposure estimates a

ConcentrationMedium Units Dist. Min. Mean Max. Std. Dev.

Outdoor air µg m–3 T 0.0005 0.0017 0.02Indoor air µg m–3 LN 0.035 0.023Drinking water µg L–1 LN 0.5 0.33Ingested soil µg g–1 C 0Ingested dust µg g–1 LN 333 216

FoodBeverages excl. water µg L–1 LN 39 25Milk µg L–1 LN 1.2 0.8Cereals µg g–1 LN 0.05 0.03Dairy products (excl. milk) µg g–1 LN 0.4 0.26Eggs µg g–1 LN 0.08 0.05Fats and oils µg g–1 LN 7.4 4.8Fish µg g–1 LN 0.01 0.007Fruit products µg g–1 LN 0.02 0.013Grains µg g–1 LN 0.05 0.03Meats µg g–1 LN 0.13 0.08Nuts and beans µg g–1 LN 0.045 0.029Other foods µg g–1 LN 0.09 0.06Poultry µg g–1 LN 0.04 0.03Processed meats µg g–1 LN 0.05 0.03Vegetable products µg g–1 LN 0.005 0.003Infant formula – powder µg g–1 LN 0.044 0.029Breast milk µg g–1 C 0

Intake Adult Teen Child Toddler Infant

Formula-fed

Total daily intake (median) µg kg–1 d–1 3.7 5.7 7.9 9.3 1.5

Total daily intake (% of total)Outdoor air 0.0 0.0 0.0 0.0 0.0Indoor air 0.2 0.2 0.2 0.2 0.4Drinking water 0.1 0.1 0.2 0.2 2.0Ingested soil NA NA NA NA NAIngested dust 4.2 4.2 5.3 8.5 70.2

FoodBeverages excl. water 11.2 5.2 3.7 2.9 0.0Cereals 0.4 0.4 0.7 1.3 0.0Dairy products (excl. milk) 6.9 6.1 7.0 8.8 0.0Eggs 0.8 0.5 0.7 1.2 0.0Fats and oils 60.1 66.7 64.2 51.0 0.0Fish 0.1 0.0 0.0 0.0 0.0Fruit products 1.2 1.0 1.6 2.4 0.0Grains 2.6 3.2 3.9 2.8 0.0Meats 4.0 3.8 2.9 3.2 0.0

a Dist. distribution type; T triangular; LN log normal; C constant; NA not available or not applicable.

theses indicate the percentage contribution to variance. Note that only para-meters contributing 5% or more to the variance are listed:

– Adult – concentration of BBP in fats and oils (35.3%), ingestion rate of fats andoils (34.6%), body weight (16.7%).

– Teen – concentration of BBP in fats and oils (37.2%), ingestion rate of fats andoils (36.1%), body weight (17.6%).

– Child – ingestion rate of fats and oils (35.3%), concentration in fats and oils(34.2%), body weight (20.2%).

– Toddler – concentration in fats and oils (28.7%), body weight (25.7%), inges-tion rate of fats and oils (24.8%).

– Formula-fed infant – ingestion rate of dust (51.9%), body weight (17.5%),concentration in dust (9.8%), concentration in formula (8.9%), ingestion rateof formula (7.7%).

The parameters listed above are used to quantify exposure for the pathwaysfound to be the dominant sources of BBP exposure (i.e., ingestion of fats and oils,infant formula, and dust). Body weight is also an important parameter, as the in-takes are expressed on a “per body weight” basis. Reduction in the variabilityand/or the uncertainty in the above parameters would reduce the uncertainty inthe exposure estimates.

5.2Comparison to Other Studies

Several other studies have evaluated human exposure to BBP [16, 17, 20–23,26, 27]. Table 8 presents a comparison of the results of these studies with thoseof the present study.

As shown in Table 8, for all age groups, the estimated median intake of BBP isa factor of three or more greater in the present study compared to the “average

248 K. Clark et al.

Table 7 (continued)

Intake Adult Teen Child Toddler Infant

Formula-fed

FoodMilk 0.1 0.2 0.3 0.5 0.0Nuts and beans 0.4 0.4 0.4 0.4 0.0Other foods 6.5 7.0 7.8 15.5 0.0Poultry 0.3 0.3 0.3 0.3 0.0Processed meats 0.4 0.4 0.4 0.3 0.0Vegetable products 0.4 0.4 0.4 0.4 0.0Infant formula NA NA NA NA 27.4

Total food 95.5 95.5 94.3 91.1 27.4

Total 100.0 100.0 100.0 100.0 100.0

Assessment of Critical Exposure Pathways 249

Tabl

e8.

Com

pari

son

ofBB

P ex

posu

re e

stim

ates

wit

h es

tim

ates

from

oth

er s

tudi

es

Stud

yBB

P In

take

(µg

kg–1

d–1)

Adu

ltTe

enC

hild

Todd

ler

Infa

nt

Pres

ent s

tudy

3.7

5.7

7.9

9.3

1.5

(for

mul

a-fe

d)Pr

obab

ilist

ic a

naly

sis;

med

ian

inta

ke

(all

expo

sure

pat

hway

s ex

clud

ing

child

ren’

s an

d ot

her

cons

umer

pro

duct

s)Fo

od o

nly

3.5

5.4

7.4

8.5

0.4

(for

mul

a-fe

d)

EC&

HC

[26]

0.64

–2.

05 (2

0–59

year

old

s);

0.74

–2.

330.

98–

3.77

1.29

–4.

960.

01–

0.14

(for

mul

a-fe

d)0.

42–

1.58

(60+

yea

rs)

0.29

–3.

21 (n

ot fo

rmul

a-fe

d)D

eter

min

isti

c an

alys

is;a

vera

ge in

take

a

(all

expo

sure

pat

hway

s ex

clud

ing

child

ren’

s an

d ot

her

cons

umer

pro

duct

s)Fo

od o

nly

0.63

–2.

01 (2

0–59

year

old

s);

0.72

–2.

290.

97–

3.70

1.27

–4.

880

(for

mul

a-fe

d)0.

40–

1.55

(60+

year

s)0.

28–

3.15

(not

form

ula-

fed)

Rea

sona

ble

wor

st c

aseb

(all

expo

sure

pat

h-11

–14

(20–

59ye

ar o

lds)

;21

–25

40–

4771

–82

0.27

–0.

38 (f

orm

ula-

fed)

way

s ex

clud

ing

child

ren’

s an

d ot

her

9–12

(60+

year

s)12

9–14

5 (n

ot fo

rmul

a-fe

d)co

nsum

er p

rodu

cts)

Food

onl

y11

–14

(20–

59ye

ar o

lds)

;21

–25

40–

4771

–82

0 (f

orm

ula-

fed)

9–12

(60+

year

s)12

9–14

5 (n

ot fo

rmul

a-fe

d)

Dav

id [1

6]0.

73 (g

eom

etri

c m

ean)

NA

cN

AN

AN

AIn

take

cal

cula

ted

from

uri

nary

met

abol

ite

3.34

(95t

h pe

rcen

tile

)da

ta (B

loun

t et a

l.[1

5])

19.7

9 (h

ighe

st v

alue

)

aEC

&H

C [2

6],e

stim

ate

of“a

vera

ge in

take

”us

ed c

once

ntra

tion

s in

indi

vidu

al fo

od it

ems,

rath

er th

an fo

od c

ateg

orie

s (s

ee S

ect.

5.2)

.b

Rea

sona

ble

wor

st c

ase”

esti

mat

e us

ed fo

od c

ateg

orie

s.c N

A n

ot a

vaila

ble.

250 K. Clark et al.

Tabl

e8

(con

tinu

ed)

Stud

yBB

P In

take

(µg

kg–1

d–1)

Adu

ltTe

enC

hild

Todd

ler

Infa

nt

Koh

n et

al.

[17]

0.09

4 (m

inim

um)

NA

NA

NA

NA

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

0.

88 (m

edia

n)da

ta (

Blou

nt e

t al.

[15]

)4.

0 (9

5th

perc

enti

le)

29 (h

ighe

st v

alue

)

AC

C [2

0]A

ge 6

thro

ugh

adul

t:0.

66 (g

eom

etri

c m

ean)

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

2.

52 (

90th

per

cent

ile);

3.18

(95

th p

erce

ntile

)da

ta [1

8]

Dav

id [2

1]1.

51 (

geom

etri

c m

ean)

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

6.

42 (

95th

per

cent

ile)

data

[19]

7.75

(hig

hest

val

ue)

IPC

S [2

7]2

NA

App

rox.

6N

AN

AIn

take

bas

ed o

n in

gest

ion

ofBB

P m

easu

red

in fo

ur fo

ods

(but

ter,

ched

dar

chee

se,

crac

kers

,yog

hurt

)

MA

FF [2

2]N

AN

AN

AN

A19

98da

ta:0

.2 (a

t bir

th)

Esti

mat

ed in

take

due

to in

gest

ion

of0.

1 (a

t age

6m

onth

s)in

fant

form

ula

1996

data

:8.7

(at b

irth

)5.

6 (a

t age

6m

onth

s)

MA

FF [2

3]0.

1 (m

ean)

NA

NA

NA

NA

Esti

mat

ed in

take

due

to in

gest

ion

ofca

rcas

s 0.

3 (9

7.5t

h pe

rcen

tile

)m

eat,

poul

try,

eggs

and

milk

.Con

vert

ed

inta

ke in

mg/

pers

on/d

ay b

y as

sum

ing

a bo

dy w

eigh

t of7

0kg

intake” in the EC&HC [26] study. The “reasonable worst-case” intake in theEC&HC [26] study is larger than the median intake in the present study. In bothstudies, food is the dominant source of exposure. The EC&HC study is a deter-ministic analysis, using receptor characteristics similar to the mean receptorcharacteristics employed in the present study. There are some important differ-ences between the two studies in terms of the concentrations of BBP in food andtheir treatment. To determine the average intake of BBP, EC&HC [26] calculatedthe intake from ingestion of only 8 of 181 food items. The calculations are basedon the results of measurements of BBP in six foods (yogurt, cheddar cheese, but-ter, pork, mixed vegetable juice, and crackers). For the remaining 173 food items,a concentration of zero or an assigned laboratory detection limit is used to cal-culate a range of intakes. In the present study, the 181 food items are grouped into15 food categories. If a phthalate is measured in one or more of the items in thatgroup, then that concentration is used for the entire group. This is done to over-come the fact that phthalates have not been tested in every food item. EC&HCemployed a similar strategy in their estimate of “reasonable worst-case” intakes.The measured concentrations of BBP from four food items are assumed to rep-resent the average concentration of BBP in each of four food groups. A concen-tration of zero or a laboratory detection limit is assumed for the remaining eightfood groups.Another difference between the two studies is that EC&HC [26] as-sumed a concentration of 0.64 µg g–1 BBP in butter based on Page and Lacroix[14]. In an earlier paper [28], the concentration of BBP in butter and margarineis reported as 3.1–47.8 µg g–1, based on measurements in 20 samples. By using thedata from Page and Lacroix [14,28], a mean concentration of 7.4 µg g–1 is calcu-lated and is used in the present study. This results in a large portion of the intakeof BBP (51–67%, depending upon the age category) coming from ingestion offats and oils.

As described in Sect. 3.2, David [16] and Kohn et al. [17] used the CDC urinarymetabolite data to back-calculate the intake of various phthalate esters. As shown in Table 8, the geometric mean intake of BBP, calculated by David [16],is a factor of five less than the estimated median intake in the present study and the 95th percentile intake is comparable to the median intake in the presentstudy. The median and 95th percentile intakes from Kohn et al. [17] are com-parable to the geometric mean and 95th percentile intakes in David [16].Similarly, the median intakes in the present study are larger than the intakesback-calculated [20, 21] from the more recent CDC studies [18, 19]. Comparisonof the results of the present study with the studies using urinary metabolite datasuggests that human exposure to BBP may be overestimated in the present study.Possible sources of the overestimation include changes in food processing tech-nology with time, a reduction in the concentration of BBP in food followingcooking (especially since fats and oils are estimated to be the largest source ofBBP exposure in the present study), and measured concentrations in food maynot be corrected for background contamination. Conversely, total human ex-posure to BBP may be underestimated in the present study because BBP is usedin personal care products [16] and such exposures have not been included in the present study, but would be accounted for in studies using urinary meta-bolite data.

Assessment of Critical Exposure Pathways 251

The International Program on Chemical Safety [24] estimated adult exposureto BBP using the 1986 Canadian market basket survey data (the data used in[26]). The estimated intake for an adult is less than that in the present study andcomparable to the estimated “average intake” by EC&HC [26].

MAFF [22] estimated the intake of BBP to infants at birth and at six monthsof age, due to ingestion of infant formula. MAFF found that the estimated intakedecreased by a factor of approximately 50 from 1996 to 1998. The estimated intake for the infant in the present study is greater than the MAFF estimate using 1998 data, but less than the intake using the 1996 data.

MAFF [23] estimated the intake of BBP to adults from ingestion of carcassmeat, poultry, eggs, and milk. The mean estimated intake is a factor of 37 lowerthan the estimated median total intake in the present study, because not allsources of exposure are included.

6Bis(2-Ethylhexyl) Phthalate (DEHP)

The assigned distributions for the concentration of DEHP in each medium usedfor the exposure assessment are summarized in Table 9. These concentration dis-tributions are obtained from the report prepared for the American ChemistryCouncil [13]. For DEHP, data are available for all required media.

As shown in Table 9, the median estimated daily intake of DEHP for eachgroup in mg kg–1 d–1 is: adults (8.2), teens (10), children (19), toddlers (26), for-mula-fed infants (5.0), and breast-fed infants (7.3). These results are compared tothe results of other studies in Sect. 6.2.

Table 9 also presents the estimated percentage intake of DEHP for each agegroup via each medium. As shown, for the non-infants, food represents between92% and 95% of the total exposure. The food category contributing most to total exposure depends upon the age group considered but, in general, the mostimportant sources of exposure are: beverages excluding water, dairy products,fats and oils, grains, milk, and other foods.

For both the formula-fed and breast-fed infants, food (infant formula or breastmilk) represents approximately one half of the exposure, with ingestion of dustaccounting for the remainder. Ingested dust represents the most importantsource of non-food exposure for the other age groups, accounting for between4.2% and 6.6% of exposure to DEHP. Inhalation of indoor air represents about1% of exposure for all age groups, while inhalation of outdoor air, ingestion ofdrinking water, and ingestion of soil, combined, represent less than 1% of expo-sure.

6.1Sensitivity Analysis

The input parameters contributing most to the variation in the exposure estimates are identified by Crystal Ball™ and are shown below. The values inparentheses indicate the percentage contribution to variance. Note that only parameters contributing 5% or more to the variance are listed:

252 K. Clark et al.

Assessment of Critical Exposure Pathways 253

Table 9. Bis(2-ethylhexyl) phthalate – exposure estimates a

ConcentrationMedium Units Dist. Min. Mean Max. Std. Dev.

Outdoor air µg m-3 T 0.0003 0.02 0.5Indoor air µg m–3 T 0.02 0.2 1.0Drinking water µg L–1 LN 0.5 0.325Ingested soil µg g–1 T 0.00003 0.048 1.0Ingested dust µg g–1 LN 662 430

FoodBeverages excl. water µg L–1 LN 77 50Milk µg L–1 LN 80 52Cereals µg g–1 LN 0.53 0.34Dairy products (excl. milk) µg g–1 LN 1.5 0.98Eggs µg g–1 LN 0.21 0.14Fats and oils µg g–1 LN 4.1 2.7Fish µg g–1 LN 0.46 0.30Fruit products µg g–1 LN 0.03 0.02Grains µg g–1 LN 0.5 0.33Meats µg g–1 LN 0.35 0.23Nuts and beans µg g–1 LN 0.21 0.14Other foods µg g–1 LN 0.28 0.18Poultry µg g–1 LN 1.1 0.7Processed meats µg g–1 LN 0.94 0.61Vegetable products µg g–1 LN 0.17 0.11Infant formula – powder µg g–1 LN 0.2 0.13– ready-to-feed liquid µg g–1 LN 0.007 0.005Breast milk µg g–1 LN 0.062 0.040

Intake Adult Teen Child Toddler Infant

Formula- Breast-fed fed

Total daily in- µg kg–1 d–1 8.2 10.0 18.9 25.8 5.0 7.3take (median)

Total daily intake (% of Total)Outdoor air 0.0 0.0 0.0 0.0 0.1 0.0Indoor air 1.0 0.9 1.0 0.9 1.5 1.1Drinking water 0.1 0.1 0.1 0.1 0.7 0.0Ingested soil 0.0 0.0 0.0 0.0 0.0 0.0Ingested dust 4.3 4.2 5.0 6.6 54.1 39.3

FoodBeverages excl. water 11.2 5.2 3.3 2.2 0.0 0.0Cereals 2.4 2.0 3.5 5.5 0.0 0.0Dairy products (excl. milk) 13.2 11.7 12.2 12.9 0.0 0.0Eggs 1.1 0.7 0.8 1.3 0.0 0.0Fats and oils 16.9 19.1 16.5 11.1 0.0 0.0Fish 1.6 0.8 0.7 0.4 0.0 0.0Fruit products 0.9 0.8 1.1 1.4 0.0 0.0Grains 13.4 16.6 18.1 11.1 0.0 0.0

a Dist. distribution type; T triangular; LN log normal; NA not available or not applicable.

– Adult – body weight (35.9%), concentration of DEHP in fats and oils (9.1%),ingestion rate of fats and oils (6.2%), ingestion rate of grains (6.0%), and con-centration of DEHP in beverages (5.5%).

– Teen – body weight (40.9%), ingestion rate of fats and oils (9.1%), concentra-tion of DEHP in fats and oils (8.3%), concentration in grains (7.6%), ingestionrate of grains (6.4%).

– Child – body weight (48.0%), concentration in grains (7.8%), ingestion rate ofgrains (6.6%), ingestion rate of fats and oils (6.5%).

– Toddler – body weight (43.9%), ingestion rate of other foods (8.2%), concen-tration in other foods (7.7%), ingestion rate of milk (5.1%).

– Formula-fed infant – ingestion rate of dust (34.1%), body weight (24.7%),concentration in infant formula (18.6%), ingestion rate of formula (12.9%),concentration in dust (5.5%).

– Breast-fed infant – concentration in breast milk (24.9%), body weight (23.6%),ingestion rate of breast milk (22.5%), ingestion rate of dust (20.7%).

As expected, the parameters listed above are used to quantify exposure for thepathways found to be the dominant sources of DEHP exposure (i.e., ingestion ofbeverages excluding water, infant formula and breast milk, fats and oils, andgrains). Body weight is also an important parameter, as the intakes are expressedon a “per body weight” basis. Reduction in the variability and/or the uncertaintyin the above parameters would reduce the uncertainty in the exposure estimates.

6.2Comparison to Other Studies

Several other studies have evaluated human exposure to DEHP [2, 4, 9, 16, 17,20–23]. Table 10 presents a comparison of the results of these studies with thoseof the present study.

254 K. Clark et al.

Table 9 (continued)

Intake Adult Teen Child Toddler Infant

Formula- Breast-fed fed

FoodMeats 5.5 5.2 3.7 3.3 0.0 0.0Milk 3.1 6.7 8.6 12.6 0.0 0.0Nuts and beans 1.0 1.0 0.9 0.8 0.0 0.0Other foods 10.3 11.2 11.3 18.9 0.0 0.0Poultry 3.9 3.6 3.5 3.6 0.0 0.0Processed meats 3.4 3.5 3.4 2.5 0.0 0.0Vegetable products 6.6 6.5 6.1 4.9 0.0 0.0Infant formula/breast milk NA NA NA NA 43.7 59.6

Total food 94.6 94.7 93.8 92.4 43.7 59.6

Total 100 100 100 100 100 100

Assessment of Critical Exposure Pathways 255

David [16] used the CDC data to back-calculate the intake of each of the phthalate esters and the results are shown in Table 10 for DEHP. As shown inTable 10, the calculated geometric mean intake is approximately an order of mag-nitude less than the estimated median intake in the present study and the 95thpercentile intake is 2.7 times less than the median intake in the present study.Kohn et al. [17] also back-calculated the intake of DEHP using the CDC data,using an alternate model. As shown in Table 10, the median and 95th percentileintakes from Kohn et al. [17] are less than those in the present study and are com-parable to those in David [16]. Similarly, the median intakes in the present studyare larger than the intakes back-calculated [20, 21] from the more recent CDCstudies [18, 19]. These results suggest that human exposure to DEHP may beoverestimated in the present study. As for DBP and BBP, possible sources of theoverestimation include changes in food processing technology with time, a re-duction in the concentration of DEHP in food, following cooking, and measuredconcentrations in food may not be corrected for background contamination.

Zaleski et al. [9] estimated the intake of DEHP for toddlers and infants. Boththe Zaleski et al. study and the present study are probabilistic and use a similarconcentration database; however, the age groupings employed are slightly dif-ferent. For the present study, the toddler is defined as being 0.5-4 years old, whilein the Zaleski et al. study the toddler is 1.5–4.5 years old. Despite the age differ-ence, the total estimated intake of DEHP, and the estimated intake from food, forthe toddler is comparable in both studies. The infant is defined in the presentstudy as being 0-6 months of age and is either exclusively formula-fed or exclu-sively breast-fed. In contrast, in the Zaleski et al. study, the infant is defined as0–1 year in age and consumes other foods in addition to either infant formula orbreast milk. This is the reason why the estimated intake for the infant in the Zaleski et al. study is about a factor of four greater than the estimated intake forthe infant in the present study.

MAFF [22] estimated the intake of DEHP to infants at birth and at six monthsof age, due to ingestion of infant formula. MAFF found that the estimated intakedecreased by a factor of about three from 1996 to 1998. The estimated intake fora six month old, using the 1998 data, is comparable to the estimated intake for theinfant in the present study.

MAFF [23] estimated the intake of DEHP to adults from ingestion of carcassmeat, poultry, eggs, and milk. The mean estimated intake is about a factor of fourlower than the estimated median total intake in the present study, because not allsources of exposure are included.

Huber et al. [29] reviewed several references that evaluated exposure of adultsto DEHP in food. The estimated daily intake ranged between 0.3 µg kg–1 d–1 and30 µg kg–1 d–1, which is in agreement with the estimate in the present study. Huber et al. [29] also presented an estimated “worst-case” intake of485 µg kg–1 d–1.

For all age groups, the estimated intake of DEHP is lower in the present studycompared with the estimated intake in Health Canada [4]. The present study usedthe same distributions for receptor characteristics as in Health Canada [4], butused updated distributions for concentrations based on considerably more (recent) data.

256 K. Clark et al.

Tabl

e10

.C

ompa

riso

n of

DEH

P ex

posu

re e

stim

ates

wit

h es

tim

ates

from

oth

er s

tudi

es

Stud

yD

EHP

Inta

ke (µ

g kg

–1d–1

)

Adu

ltTe

enC

hild

Todd

ler

Infa

nt

Pres

ent s

tudy

Prob

abili

stic

ana

lysi

s;m

edia

n in

take

(all

ex-

8.2

1019

265.

0 (f

orm

ula-

fed)

posu

re p

athw

ays

excl

udin

g ch

ildre

n’s

and

7.3

(bre

ast-

fed)

othe

r co

nsum

er p

rodu

cts)

Food

onl

y7.

89.

518

242.

2 (f

orm

ula-

fed)

4.4

(bre

ast-

fed)

Dav

id [1

6]0.

60 (g

eom

etri

c m

ean)

NA

dN

AN

AN

AIn

take

cal

cula

ted

from

uri

nary

met

abol

ite

3.05

(95t

h pe

rcen

tile

)da

ta (B

loun

t et a

l.[1

5])

38.4

8 (h

ighe

st v

alue

)

Koh

n et

al.

[17]

0.71

(med

ian)

NA

NA

NA

NA

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

3.

6 (9

5th

perc

enti

le)

data

(Blo

unt e

t al.

[15]

)46

(hig

hest

val

ue)

AC

C [2

0]A

ge 6

thro

ugh

adul

t:0.

65 (5

0th

perc

enti

le)

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

2.

13 (

90th

per

cent

ile);

2.62

(95

th p

erce

ntile

)da

ta [1

8]

Dav

id [2

1]2.

76 (

geom

etri

c m

ean)

Inta

ke c

alcu

late

d fr

om u

rina

ry m

etab

olite

12

.91

(95t

h pe

rcen

tile

)da

ta [1

9]21

.38

(hig

hest

val

ue)

Zal

eski

et a

l.[9

]aN

AN

AN

A22

–32

17–

26 (

form

ula-

fed)

Prob

abili

stic

ana

lysi

s;m

edia

n in

take

(all

ex-

23–

31 (b

reas

t-fe

d)po

sure

pat

hway

s ex

clud

ing

child

ren’

s an

d ot

her

cons

umer

pro

duct

s)Fo

od o

nly

NA

NA

NA

18–

2013

(fo

rmul

a-fe

d)18

–19

(bre

ast-

fed)

Assessment of Critical Exposure Pathways 257

Tabl

e10

(con

tinu

ed)

Stud

yD

EHP

Inta

ke (µ

g kg

–1d–1

)

Adu

ltTe

enC

hild

Todd

ler

Infa

nt

MA

FF [2

2]N

AN

AN

AN

A19

98da

ta:1

3.8

(at b

irth

)Es

tim

ated

inta

ke d

ue to

inge

stio

n of

7.7

(at a

ge 6

mon

ths)

infa

nt fo

rmul

a.19

96da

ta:3

5 (a

t bir

th)

23 (a

t age

6m

onth

s)

MA

FF [2

3]2.

1 (m

ean)

NA

NA

NA

NA

Esti

mat

ed in

take

due

to in

gest

ion

ofca

rcas

s 4.

3 (9

7.5t

h pe

rcen

tile

)m

eat,

poul

try,

eggs

and

milk

.Con

vert

ed in

-ta

ke in

mg/

pers

on/d

ay b

y as

sum

ing

a bo

dy

wei

ght o

f70

kg.

Hea

lth C

anad

a [4

]b27

3879

134

153

Prob

abili

stic

ana

lysi

s;m

edia

n in

take

(all

ex-

posu

re p

athw

ays

excl

udin

g ch

ildre

n’s

and

othe

r co

nsum

er p

rodu

cts)

Food

onl

y26

3878

133

152

EC&

HC

[2]c

5.78

–5.

828.

15–

8.20

14.0

3–14

.10

18.8

6–18

.98

8.87

–9.

12D

eter

min

isti

c an

alys

is (a

ll ex

posu

re p

ath-

way

s ex

clud

ing

child

ren’

s an

d ot

her

cons

umer

pro

duct

s)Fo

od o

nly

4.91

7.18

12.8

517

.81

7.88

aIn

fant

in Z

ales

ki e

t al.

[9] i

s 0–

1ye

ars

old

and

cons

umes

a v

arie

ty o

ffoo

ds in

add

itio

n to

eit

her i

nfan

t for

mul

a or

bre

ast m

ilk.I

n th

e pr

esen

t stu

dy,t

hein

fant

is 0

–6

mon

ths

old

and

cons

umes

onl

y in

fant

form

ula

or b

reas

t milk

.b

Hea

lth C

anad

a [4

] use

d sa

me

rece

ptor

cha

ract

eris

tics

as

pres

ent s

tudy

,but

con

cent

rati

on d

atab

ase

was

exp

ande

d fo

r pr

esen

t stu

dy.

cEC

&H

C [2

] use

d lo

wer

con

cent

rati

ons

for

DEH

P in

man

y fo

ods

com

pare

d to

the

pres

ent s

tudy

(see

Sec

t.6.

2).

dN

A n

ot a

vaila

ble.

The estimated intake of DEHP for all age groups, except the infant, is slightlylarger in the present study compared to EC&HC [2]. In addition to being a de-terministic analysis rather than a probabilistic analysis, the EC&HC [2] studyused lower concentrations than are used in the present study, for many foods,based upon the data that were available at that time. Differences between theEC&HC study and the present study include: the concentrations of DEHP indairy products, fats and oils, vegetables, and meat, are a factor of 3–12 lower inEC&HC [2] than the mean concentrations used in the present study, and expo-sure to DEHP in beverages other than drinking water is not evaluated in EC&HC[2]. These differences in concentrations account for the differences in the estimated intake of DEHP.

7Discussion

For the five phthalate esters evaluated in this chapter, the median estimated dailyintake is highest for toddlers (ranging from 1.6 µg kg–1 d–1 for DMP to 26 µgkg–1 d–1 for DEHP) and lowest for infants (ranging from 1.5 µg kg–1 d–1 for DBPand BBP to 7.3 µg kg–1 d–1 for DEHP); exposure to infants from food could not beevaluated for DEP or DMP. For adults, the median estimated intake ranges from0.7 µg kg–1 d–1 for DMP to 8.2 µg kg–1 d–1 for DEHP. Table 11 presents a summaryof estimated total daily intake for each of the five phthalate esters for each agegroup.

For all five phthalate esters evaluated (except BBP exposure for formula-fed infants), food represents the most important source of exposure. The food categories contributing most to exposure depend upon the phthalate ester andthe age group evaluated. For all phthalate esters, except DMP, eggs, fish, and nutsand beans each represent less than 2% of exposure for all age groups.

The exposure estimates presented are considered to be representative oftypical population exposures, as the estimates employed “average” exposure factors [5]. Individuals who consume large quantities of fish are sometimes atgreater risk of exposure to chemicals than are other members of the population.

258 K. Clark et al.

Table 11. Estimated total daily intake by age group for a series of phthalate esters

Phthalate Estimated total daily intake (median value) in mg kg–1 d–1

esterAdult Teen Child Toddler Infant Infant

(formula-fed) (breast-fed)

DMP 0.7 0.7 1.4 1.6 0.05 a 0.01 a

DEP 2.5 3.0 5.7 10.6 0.2 a 0.2 a

DBP 5.6 6.4 11 14 1.5 2.9BBP 3.7 5.7 7.9 9.3 1.5 NA b

DEHP 8.2 10.0 18.9 25.8 5.0 7.3

a Non-food exposure only.b NA not available.

Richardson [8] reported a lognormal distribution for consumption of fish bymembers of the Amerindian and Inuit communities. The mean ingestion rate(excluding individuals who reported no fish consumption) is 220 g d–1 for adults.By using this mean fish ingestion rate, the estimated intake of each phthalate ester in mg kg–1 d–1 due only to fish consumption is: 0.015 (DMP), 0.18 (DEP),0.7 (DBP), 0.03 (BBP), and 1.4 (DEHP). This estimated intake from high fish con-sumption represents between 0.8% and 17% of the total estimated intake for eachphthalate (reported in Tables 2, 4, 6, 8, and 10).As high fish consumers would havelower ingestion rates for some foods compared to the general population (e.g.,meat, processed meat), the estimated intake of phthalates for the high fish consumers is not likely to be higher than the estimated intake for the general population.

By using data obtained in 1996 and 1998, MAFF [22] estimated the intake ofDBP, BBP, and DEHP in infants due to ingestion of infant formula. MAFF foundthat the estimated intake decreased by factors of 2.5–56 over the two-year period(see Tables 6, 8, and 10). The influences of changes in food processing method-ologies and storage have not been evaluated in the present study, as recent mea-surements are combined with older measurements to obtain a more completedataset. Many of the measurements of phthalate esters in food were obtained inthe late 1980s. Estimates of phthalate ester exposure would be improved by ob-taining recent measurements of phthalate esters in a variety of foods.

Ingestion of dust and inhalation of indoor air represent the most importantnon-food sources of exposure to phthalate esters. These estimates rely heavily onEuropean data, due to a lack of North American data.

Detection limits have a large influence on the estimated intake of phthalate esters, particularly for DMP. The estimated intake for DMP should be regardedas approximate and, quite possibly, an overestimate, as it is based primarily onconcentrations equal to one half of a detection limit.

The results of the present study are compared to results from other studies forDEP, DBP, BBP, and DEHP. The estimated intakes in the present study are higherthan those in EC&HC [2, 3, 26] due primarily to methodological differences inthe treatment of food items versus food categories. EC&HC calculated intakes forindividual food items (up to 181 items). Where data are lacking for a particularfood item, the concentration is assigned a zero or a laboratory detection limit. Incontrast, in the present study, intake is evaluated for food categories. Measuredconcentrations of one or more food items in a category are assigned to the en-tire category. For chemicals with a large database of measured concentrations(i.e., concentrations have been measured in most food items), the difference between the two methods may not be large but, if little data are available, ananalysis by individual items may underestimate exposure and analysis by foodcategories may overestimate exposure. This observation leads to the recommen-dation for additional testing of a variety of food items.

A comparison of the results of the present study with studies that back-calculate phthalate ester intake from urinary metabolite data suggests that exposure in the present study may be overestimated for DEHP, BBP, and DBP dueto changes in food processing over time (many of the measured concentrationsof phthalates in food are not recent), loss of phthalates due to cooking has not

Assessment of Critical Exposure Pathways 259

been accounted for in the present study, and some measured concentrations infood may be elevated due to background contamination. The comparison alsosuggests that the significant sources of exposure to DEHP have been accountedfor in the present study. Conversely, exposure to DEP (and possibly BBP and DBP)is underestimated in the present study because direct exposure to personal careproducts has not been included. The overestimate of exposure to BBP and DBPfrom food, referred to above, may be partially cancelled by the lack of inclusionof personal care products as a source of exposure. Table 12 presents a summaryof the comparison of the intakes estimated in the present study with the back-calculated intakes.

A link between human exposure and multimedia modeling of phthalate estershas not yet been established. This link has been difficult to estimate, for severalreasons, including:

– Phthalate ester exposure may arise from industrial or consumer use in addi-tion to environmental sources;

– As discussed in Chapter 5, phthalate esters appear to be metabolized as theyprogress up the food chain (i.e., they biodilute), making it difficult to estimateconcentrations in food from concentrations in abiotic media;

– Some foods that contribute significantly to human exposure are importedfrom regions outside the region being modeled. As a result, human exposureestimates will not be substantially improved by measuring phthalate ester con-centrations in foods from different regions. Regional differences in exposureare more likely to be due to dietary differences (i.e., differences in ingestionrates of different foods) than in the concentrations of phthalate esters.

Acknowledgement. We are grateful to the Environmental Research Task Group (ERTG) of thePhthalate Ester Panel of the American Chemistry Council (ACC) for funding this research.

260 K. Clark et al.

Table 12. Comparison of estimated total daily intake in present study with studies that back-calculate intake from metabolite data

Phthalate Estimated total daily intake (µg kg–1 d–1)ester

Present study (median) Intake calculated from urinary metabolite data(mean or median)

Adult Toddler Adult Age 6 to Adult Toddler [15–17] [18, 20] [19, 21]

DMP 0.7 1.6 NAa NA NADEP 2.5 10.6 12–12.34 5.42 6.29DBP 5.6 14 1.5–1.56 0.90 2.45BBP 3.7 9.3 0.73–0.88 0.66 1.51DEHP 8.2 25.8 0.6–0.71 0.65 2.76

a NA not available.

8References

1. Finley B, Proctor D, Scott P, Harrington N, Paustenbach D, Price P (1994) Risk Anal 14:5332. Environment Canada and Health Canada (EC&HC) (1994) Canadian Environmental

Protection Act, Priority Substances List assessment report – bis(2-ethylhexyl) phthalate.Ottawa, ON

3. Environment Canada and Health Canada (EC&HC) (1994) Canadian Environmental Protection Act, Priority Substances List assessment report – dibutyl phthalate. Ottawa, ON

4. Health Canada (1996) Environmental pathways analysis for diethylhexyl phthalate(DEHP) and other phthalate esters. Prepared by O’Connor Associates Environmental Inc.for the Health Protection Branch, Ottawa, ON

5. Health and Welfare Canada (HWC) (1993) Reference values for Canadian populations.Prepared by the Environmental Health Directorate Working Group on Reference Values.July 1988; updated May 1993

6. Stephens T, Craig CL (1990) The well-being of Canadians: highlights of the 1988 Campbell’ssurvey. Canadian Fitness and Lifestyle Research Institute, Ottawa, p 95, appendices, data

7. Health Canada (1995) Probabilistic assessment of 24-hour breathing rates. Prepared byCornerstone Engineering and Consulting Inc. for the Health Protection Branch, Ottawa,ON. October

8. Richardson GM (1997) Compendium of Canadian human exposure factors for risk assessment. O’Connor Associates Environmental, Ottawa, ON

9. Zaleski RT, Parkerton TF, Konkel WJ (2000) An assessment of di-2-ethylhexyl phthalate exposure for children. Report prepared for the European Council of Plasticisers & Inter-mediates (ECPI), a Sector Group of CEFIC, Brussels, Belgium, by Exxon Biomedical Sci-ences, Inc, Annandale, NJ, p 35, appendices

10. Health Canada (1994) Probabilistic exposure assessments for twenty contaminants in theCanadian environment. Prepared by O’Connor Associates Environmental Inc. for theHealth Protection Branch, Ottawa, ON. October

11. Stanek EJ III, Calabrese EJ (1995) Hum Ecol Risk Assess J 1 :13312. Calabrese EJ, Stanek EJ, Gilbert CE, Barnes RM (1990) Regul Toxicol Pharmacol 12: 8813. Clark K, Cousins I, Mackay D (2001) Multimedia modelling and exposure assessment for

phthalate esters – observed concentrations in the environment. Prepared for AmericanChemistry Council. December

14. Page BD, Lacroix GM (1995) Food Addit Contam 12:12915. Blount BC, Silva MJ, Caudill SP, Needham LL, Pirkle JL, Sampson EJ, Lucier GW, Jackson RJ,

Brock JW (2000) Environ Health Perspect 108:97916. David RM (2000) Environ Health Perspect 108:A44017. Kohn MC, Parham F, Masten SA, Portier CJ, Shelby MD, Brock JW, Needham LL (2000)

Environ Health Perspect 108:A44018. Centers for Disease Control and Prevention (CDC) (2001) National report on human

exposure to environmental chemicals. National Center for Environmental Health.Availableonline at http://www.cdc.gov/nceh/dls/report/contact.htm

19. Brock JW, Caudill SP, Silva MJ, Needham LL, Hilborn ED (2002) Bull Environ Contam Tox-icol 68:309

20. American Chemistry Council (ACC) (2001) What the CDC national report says about phthalate exposures. Phthalate Esters Panel.Available online at http://www.phthalates.org/mediacenter/pep_2001-6-7.html

21. David RM (2002) Personal communication22. Ministry of Agriculture, Fisheries and Food (MAFF) (1998) MAFF UK – phthalates in

infant formulae – follow-up survey. Joint Food Safety and Standards Group, Food Surveil-lance Information Sheet #168, December

23. Ministry of Agriculture, Fisheries and Food (MAFF) (1996) MAFF UK – phthalates in food.Joint Food Safety and Standards Group, Food Surveillance Information Sheet #82, March

Assessment of Critical Exposure Pathways 261

24. International Program on Chemical Safety (IPCS) (1997) Environmental health criteria –189 – di-n-butyl phthalate.World Health Organization (WHO), Geneva, Switzerland. ISBN92 4 157189 6, as cited in National Toxicology Program – Center for the Evaluation of Risksto Human Reproduction (NTP-CERHR) (2000) NTP-CERHR expert panel report on di-n-butyl phthalate. US Department of Health and Human Services. NTP-CERHR-DBP-00.October

25. Health Canada (1994) Canadian Environmental Protection Act, Priority Substances List –supporting documentation health-related sections – di-n-butyl phthalate. Ottawa, ON

26. Environment Canada and Health Canada (EC&HC) (2000) Canadian Environmental Pro-tection Act, Priority Substances List assessment report – butylbenzyl phthalate. Ottawa, ON

27. International Program on Chemical Safety (IPCS) (1999) Concise international chemicalassessment document 17 – butyl benzyl phthalate. World Health Organization (WHO),Geneva, Switzerland, as cited in National toxicology program – Center for the Evaluationof Risks to Human Reproduction (NTP-CERHR) (2000) NTP-CERHR expert panel reporton butyl benzyl phthalate. US Department of Health and Human Services. NTP-CERHR-BBP-00. October.

28. Page BD, Lacroix GM (1992) Food Addit Contam 9:19729. Huber WW, Grasl-Kraupp B, Schulte-Hermann R (1996) Crit Rev Toxicol 26:365

262 K. Clark et al.: Assessment of Critical Exposure Pathways

© Springer-Verlag Berlin Heidelberg 2003

Aquatic Toxicity of Phthalate Esters

Christopher A. Bradlee 1 · Paul Thomas 2

1 BASF Corporation, Corporate Ecology & Safety, Wyandotte, MI, USA2 Akzo Nobel, Chemicals Research Arnhem, Arnhem, The Netherlands

The aquatic toxicity of Phthalate esters is discussed in this chapter. This chapter begins withan examination of the physico-chemical properties of phthalate esters that have a significant in-fluence on toxicity.Acute and chronic toxicity are then reviewed with a focus on chronic effects.This chapter concludes with a discussion of studies that have examined the potential for endocrine modulating effects of phthalate esters in aquatic organisms. Perhaps what is most notable about phthalate esters with regard to their physico-chemical properties is that theydemonstrate a solubility cut-point (threshold) so that phthalate esters with alkyl chain lengthof C6 or greater have water solubilities less than 1 mg/L. The acute and chronic toxicity datashow that while the lower phthalates (<C6) demonstrate toxicity, the higher phthalates (≥C6)are not toxic to aquatic organism (fish, algae, invertebrates); even at concentrations up to thelimit of solubility. It has been suggested that the lack of toxicity for the higher phthalates is re-lated to their relative insolubility in water and their ready metabolism by aquatic organisms,so that the critical body burden for toxicity is not reached. The evidence for endocrine mod-ulation effects caused by phthalate esters is equivocal. Studies from a few in-vitro and in-vivo assays have suggested that DBP and BBP, but no other phthalate ester, were capable ofinteracting with estrogenic receptors. Testis-ova in male gonads and induction of hepatic vitel-logenin have occasionally been reported for some phthalate esters at very high doses. Thesefindings, occurring at concentrations already considered to be toxic based on conventionalstudies, were not supported upon repetition of the original study or were reported in studiesthat employed injection of test material into individual fish. Inter-peritoneal injection is an in-appropriate route of administration for phthalates because they are known to under undergometabolism before they enter the systemic circulatory system.

Keywords. Phthalates, Chronic toxicity, Endocrine disruption, Aquatic

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

1.1 Solubility Influence on Toxicity . . . . . . . . . . . . . . . . . . . 264

2 Acute Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265

3 Chronic Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . 266

3.1 Algae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2663.2 Invertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2763.3 Fish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 263–298DOI 10.1007/b11469

3.4 Endocrine Modulating Effects . . . . . . . . . . . . . . . . . . . . 2843.4.1 Lower Phthalates . . . . . . . . . . . . . . . . . . . . . . . . . . . 2843.4.2 Higher Phthalates . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294

7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295

1Introduction

The aquatic toxicity of phthalate esters is discussed in this chapter. A compre-hensive review on the aquatic toxicity of phthalate esters was previously pre-sented in Staples et al. [1] and this chapter provides a summary of that informa-tion and reviews the literature either not included in, or published after, theStaples et al. [1] review. This chapter begins with an examination of the physico-chemical properties of phthalate esters that have a significant influence on toxi-city.Acute and chronic toxicity are then reviewed with a focus on chronic effects.This chapter concludes with a discussion of studies that have examined the po-tential for endocrine modulating effects of phthalate esters in aquatic organisms.

1.1Solubility Influence on Toxicity

For an in-depth discussion of the physico-chemical properties of phthalate estersthe reader is referred to Chapter 3 of this handbook. Perhaps most notable aboutphthalate esters in regards to their physico-chemical properties is that theydemonstrate a solubility cut-point, and like many organic chemicals the aquatictoxicity of phthalate esters is strongly influenced by water solubility.As shown inTable 1, phthalate esters have solubilities ranging from 5200 mg/L to less than1¥10–7 mg/L, with solubility decreasing with increasing molecular weight. Thereis a solubility cut-point (threshold) for phthalate esters with an alkyl chain lengthof C6 or greater where the water solubility decreases to less than 1 mg/L. The sol-ubility of BBP is 3.8 mg/L while for DHP it is only 0.159 mg/L. It is appropriateto treat phthalate esters as two groups when discussed in terms of their aquatictoxicological properties; the “lower phthalates” with side chains of less thanC6 and “higher phthalates” of C6 or greater side chains.

The data show that higher phthalate esters with alkyl chain lengths ≥C6 do notpose intrinsic toxicity to aquatic organisms. For the higher phthalates, in manyaquatic toxicity studies neither acute nor chronic effects are found even when thetest concentrations are 2–3 times higher than the true water solubility. Parker-ton and Konkel [2] in their paper on the application of quantitative structure-ac-tivity relationships for assessing the toxicity of phthalate esters have provided atheoretical basis for understanding the lack of acute or chronic aquatic toxicityeffects from the higher phthalates. It is well established that phthalate esters arerapidly metabolized in biota and thus the actual bioconcentration factor for theseproducts is considerably lower than would be predicted from their octanol/wa-

264 C.A. Bradlee and P. Thomas

ter partition coefficients. This fact, together with the low water solubility, suggeststhat the critical body burden for toxicity is not reached.

Another important physico-chemical property of phthalate esters, in relationto aquatic toxicity, is that the higher phthalates appear to form stable emulsionsin water due to a self-dispersing property. The higher phthalates dissolve in wa-ter up to the limit of solubility and anything beyond this point forms a stableemulsion. This property may account for the considerable range of publisheddata for both water solubility and octanol/water partition coefficient values. Moreimportantly, the stable emulsions may lead to artifactual toxicity in laboratory ex-periments when, for example, a Daphnia dies after it becomes entrapped in thewater-phthalate-air interface.Almost all aquatic toxicity testing conducted on thehigher phthalates has been carried out a concentrations far above the water sol-ubility where artifactual toxicity from emulsions is relevant.

2Acute Toxicity

Acute aquatic test results on eighteen different phthalate esters are reviewed inStaples et al. [1], which includes data on both freshwater and saltwater species of

Aquatic Toxicity of Phthalate Esters 265

Table 1. Solubility of phthalate esters

Name CAS no. Abbreviation Aqueous solu-bility (mg/L)

Dimethyl phthalate 131-11-3 DMP 5220Diethyl phthalate 84-66-2 DEP 591Dibutyl phthalate 84-74-2 DBP 9.9Butyl benzyl phthalate 85-68-7 BBP 3.8Dihexyl phthalate 84-75-3 DHP 0.159Butyl(2-ethylhexyl) phthalate 85-69-8 BOP 0.385

68151-50-4Di-(n-hexyl,n-octyl,n-decyl) phthalate 25724-58-7 610P 0.0009

68515-51-5Di(2-ethylhexyl) phthalate 117-81-7 DEHP 0.0025Di-iso-octyl phthalate 27554-26-3 DIOP 0.0025Di-iso-nonyl phthalate 26761-40-0 DINP 0.0003

28553-12-0Di-iso-decyl phthalate 26761-40-0 DIDP 0.00004

68515-49-1Di(heptyl,nonyl,undecyl) phthalate 111381-89-6 711P 0.0003

111381-90-9111381-91-03648-20-268515-44-668515-45-7

Di-undecyl phthalate 3648-20-2 DUP 0.000004Di-tridecyl phthalate 119-06-2 DTDP 0.00000007

68515-47-9

Data regarding physical properties of phthalate esters from Cousins and Mackay, this volume.

algae, invertebrates and fish.A discussion on studies with microorganisms is alsoincluded. Of the eighteen phthalate esters for which acute data are available, onlysix esters (DMP, DEP, DAP, DBP, DIBP, BBP) have acute effects consistently acrosstest type and species. Therefore, acute toxicity is demonstrated only in the lowerphthalate esters (<C6), and phthalate esters with alkyl chain lengths equal to orgreater than C6 are not acutely toxic to aquatic organisms; even at concentrationsat or above their solubility limit [1].The authors examined the data from the higherquality studies (well documented procedures and measured test concentrations)with algae, invertebrates and fish and concluded that acute toxicity results withDMP ranged from 29 to 377 mg/L,results with DEP ranged from 10.3 to 131 mg/L,results with DBP ranged from 0.35 to 6.29 mg/L and results with BBP ranged from0.21 to 5.3 mg/L. The reader is referred to Staples et al. [1] for a complete discus-sion of acute aquatic toxicity; however, Table 2 provides a summary of selectEC/LC50 values for algae, fish and invertebrates based mostly on data from stud-ies with well documented test procedures and measured test concentrations.

For the higher phthalates two studies on DEHP did demonstrate effects inDaphnia magna at concentrations exceeding its solubility limit [3, 4]; however,the weight-of-evidence overwhelmingly suggests that the toxicity was attributedto physical effects (surface entrapment) and not intrinsic toxicity. First, the ma-jority of studies on DEHP and the other higher phthalates with daphnids showedEC/LC50s higher than the solubility limits [5–11]. Secondly, Brown et al. [12] re-ports no acute toxicity for DEHP at concentrations 100 times higher than the solubility limit in a study that used a non-toxic solubilizer to disperse the higherphthalate esters. The purpose of the solubilizer was to prevent the formation ofa surface layer of the phthalate, which could lead to entrapment of the Daphnids.

In summary, the data show that lower phthalates are acutely toxic at concen-trations below their solubility and that toxicity increases with increasing alkylchain length up to and including four carbon atoms. In studies with the higherphthalates (≥C6), acute toxicity occurred only at concentrations significantlyabove their water solubilities and, therefore, effects were likely due to physical effects and not intrinsic toxicity.

3Chronic Toxicity

The results of chronic toxicity studies with fourteen phthalate esters are shownin Tables 3–7. The data in Table 3 are sorted by individual phthalate ester, pre-sented in order of increasing molecular weight and are divided by test species.Tables 4–7 are sorted by tests species and are divided by individual phthalate es-ter. Whether the exposure concentrations were measured or nominal, test dura-tion, endpoint(s) or criteria and NOEC/LOEC are also presented in Tables 3–7.

3.1Algae

Chronic toxicity test results on algae are shown in Table 4.Algal assays that haveat least a 72-h duration are considered chronic assays [13] because the duration

266 C.A. Bradlee and P. Thomas

Aquatic Toxicity of Phthalate Esters 267

Tabl

e2.

Sele

ct E

C/L

C50

valu

es fo

r th

e ac

ute

aqua

tic

toxi

city

ofp

htha

late

est

ers,

as p

rese

nted

in s

tapl

es e

t al.

[1]

Test

spe

cies

Phth

alat

e Fr

esh/

salt

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t EC

or

LC50

R

ef.

este

r(F

/S)

nom

inal

or

cri

teri

a(m

g/L)

(M/N

)

Alg

aePh

thal

ate

este

rs <

C6

Sele

nast

rum

cap

rico

rnut

umD

MP

FM

6d

stat

icC

ell n

umbe

r14

2*[5

,43]

S.ca

pric

ornu

tum

DEP

FM

96h

Cel

l num

ber

85.6

[44]

Scen

edes

mus

sub

spic

atus

DA

PF

N96

hC

ell m

ulti

plic

atio

n 4.

5[4

5]in

hibi

tion

S.su

bspi

catu

sD

BPF

M72

hC

ell g

row

th1.

2[1

5]S.

capr

icor

nutu

mBB

PF

M96

hC

ell n

umbe

r0.

4[1

6]

Phth

alat

e es

ters

≥C

6S.

capr

icor

nutu

m,

DH

P,B

OP,

610P

,F

M

7d

stat

ic

Cel

l num

ber

no e

ffec

t at o

r [5

,43,

DEH

P,D

IOP,

DIN

P,ab

ove

solu

bilit

y46

,47]

DID

P,71

1P D

UP

limit

Inve

rteb

rate

s Ph

thal

ate

este

rs <

C6

Wat

er fl

ea D

aphn

ia m

agna

D

MP

F M

48

h Su

rviv

al

45.9

[5

,6]

Wat

er fl

ea D

aphn

ia m

agna

D

EP

F M

48

h Su

rviv

al

37.5

[4

,5]

Wat

er fl

ea D

aphn

ia m

agna

D

AP

F N

24

h Su

rviv

al

26

[48]

Wat

er fl

ea D

aphn

ia m

agna

D

BP

F M

48

h Su

rviv

al

3.0

[5,6

]W

ater

flea

Dap

hnia

mag

na

BBP

F M

48

h Su

rviv

al

1.8

[3]

Phth

alat

e es

ters

≥C

6W

ater

flea

Dap

hnia

mag

naD

HP,

BO

P,61

0P,

FM

48h

Surv

ival

no e

ffec

t at o

r [5

,6]

DEH

P,D

IOP,

DIN

P,ab

ove

solu

bilit

y D

IDP,

711P

DU

Plim

it

268 C.A. Bradlee and P. Thomas

Tabl

e2

(con

tinu

ed)

Test

spe

cies

Phth

alat

e Fr

esh/

salt

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t EC

or

LC50

R

ef.

este

r(F

/S)

nom

inal

or

cri

teri

a(m

g/L)

(M/N

)

Fish

Phth

alat

e es

ters

<C

62

Fath

ead

min

now

P.p

rom

elas

DM

PF

M96

h st

atic

Surv

ival

39[5

,49]

Fath

ead

min

now

P.p

rom

elas

DEP

FM

96h

stat

icSu

rviv

al16

.8[5

,50]

Fath

ead

min

now

P.p

rom

elas

DBP

FM

96h

stat

icSu

rviv

al1.

54[5

,50]

Fath

ead

min

now

P.p

rom

elas

DIB

PF

M96

h flo

w-t

hrou

ghSu

rviv

al0.

9[4

8]Fa

thea

d m

inno

w P

.pro

mel

asBB

PF

M96

h st

atic

Surv

ival

>0.

78[5

,50]

Phth

alat

e es

ters

≥C

6Fa

thea

d m

inno

w P

.pro

mel

asD

HP,

BO

P,61

0P,

F M

96

h st

atic

Su

rviv

al

no e

ffec

t at o

r [5

,21,

DEH

P,D

IOP,

DIN

P,ab

ove

solu

b-48

,50]

DID

P,71

1P D

UP

ility

lim

it

Not

e:EC

50=

Effe

ct c

once

ntra

tion

50%

,LC

50=

Leth

al c

once

ntra

tion

50%

,M=

Mea

sure

d ex

posu

re c

once

ntra

tion

s use

d to

cal

cula

te re

sults

,N=

Nom

inal

conc

entr

atio

ns u

sed

to c

alcu

late

res

ults

(w

heth

er o

r no

t con

cent

rati

ons

wer

e an

alyt

ical

ly c

onfir

med

),F

=Fr

eshw

ater

,S=

Saltw

ater

.*

Valu

es fr

om re

fere

nce

[5] w

ere

reca

lcul

ated

from

ori

gina

l con

trac

t lab

orat

ory

repo

rts p

er c

urre

nt U

SEPA

pra

ctic

e.O

rigi

nal v

alue

s wer

e ba

sed

on m

ea-

sure

d in

itia

l tes

t con

cent

rati

on;c

urre

nt v

alue

s ar

e ba

sed

on a

vera

ging

init

ial a

nd fi

nal m

easu

red

test

con

cent

rati

ons.

Aquatic Toxicity of Phthalate Esters 269

Tabl

e3.

Chr

onic

toxi

city

ofp

htha

late

est

ers

to a

quat

ic o

rgan

ism

s

Test

spe

cies

Fres

h/sa

lt M

easu

red/

Test

dur

atio

nTe

st e

ndpo

int

NO

EC (L

OEC

) R

ef.

(F/S

)no

min

al

or c

rite

ria

(mg/

L)(M

/N)

Dim

ethy

l pht

hala

te (D

MP)

Alg

aeSe

lena

stru

m c

apri

corn

utum

FM

6d

stat

icC

ell n

umbe

r(6

4.7)

*[5

,43]

Inve

rteb

rate

sW

ater

flea

Dap

hnia

mag

naF

M21

dSu

rviv

al9.

6 (2

3.0)

[7,8

]G

rass

shr

imp

Pala

emon

etes

pug

ioS

M30

dLa

rval

mor

talit

y10

(100

)[1

8]

Fish

Rai

nbow

trou

t Onc

orhy

nchu

s m

ykis

sF

M60

d po

st-h

atch

Gro

wth

/sur

viva

l11

(24)

[8]

Die

thyl

pht

hala

te (D

EP)

Alg

aeS.

capr

icor

nutu

mF

M8

d st

atic

Cel

l num

ber

3.65

[5,4

3]

Inve

rteb

rate

s W

ater

flea

D.m

agna

FM

21d

Surv

ival

/rep

rodu

ctio

n25

(59)

[7,8

]Fi

sh–

–N

o in

form

atio

n

Dib

utyl

pht

hala

te (D

BP)

Alg

aeS.

capr

icor

nutu

mF

M10

d st

atic

Cel

l num

ber

0.21

[5,4

3]Sc

ened

esm

us s

ubsp

icat

usF

M7

dG

row

th r

ate

6.1

[14]

S.su

bspi

catu

sF

M72

hC

ell g

row

th0.

5[1

5]S.

subs

pica

tus

FM

72h

Gro

wth

rat

e0.

5In

vert

ebra

tes

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

al0.

96 (2

.5)

[7,8

]W

ater

flea

D.m

agna

FM

21d

Surv

ival

/rep

rodu

ctio

n0.

11 (0

.20)

[21]

270 C.A. Bradlee and P. Thomas

Tabl

e3

(con

tinu

ed)

Test

spe

cies

Fres

h/sa

lt M

easu

red/

Test

dur

atio

nTe

st e

ndpo

int

NO

EC (L

OEC

) R

ef.

(F/S

)no

min

al

or c

rite

ria

(mg/

L)(M

/N)

Inve

rteb

rate

sW

ater

flea

D.m

agna

FM

21d

Surv

ival

/rep

rodu

ctio

n1.

05 (1

.91)

[21]

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

al/r

epro

duct

ion

0.16

(0.6

4)[2

1]G

rass

shr

imp

P.pu

gio

SM

30d

Larv

al m

orta

lity

10.0

(50.

0)[1

8]Es

tuar

ine

mic

roco

smS

M2

wk

Abu

ndan

ce a

nd d

iver

sity

0.04

(0.3

4)[1

9,51

]

Fish

Rai

nbow

trou

t O.m

ykis

sF

M60

d po

st-h

atch

Gro

wth

/sur

viva

l 0.

1 (0

.19)

[8]

Fath

ead

min

now

P.p

rom

elas

FN

20d

flow

-thr

uH

atch

abili

ty0.

56 (1

.0)

[9]

Fath

ead

min

now

P.p

rom

elas

FM

144

h flo

w-t

hru

Mor

talit

y0.

32[4

8]

Buty

lben

zyl p

htha

late

(BBP

)A

lgae

Nav

icul

a pe

llicu

losa

FM

96h

Cel

l num

ber

0.3

[16]

S.ca

pric

ornu

tum

FM

96h

Cel

l num

ber

0.1

[16]

S.ca

pric

ornu

tum

FM

6d

stat

icC

ell n

umbe

r(0

.10)

[5,4

3]D

unal

iella

tert

iole

cta

SM

96h

Cel

l num

ber

0.3

[16]

Skel

eton

ema

cost

atum

SM

96h

Cel

l num

ber

0.1

[16]

Inve

rteb

rate

sW

ater

flea

D.m

agna

FM

42d

Surv

ival

/rep

rodu

ctio

n0.

26 (0

.76)

[16]

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

al/r

epro

duct

ion

0.28

(1.4

)[7

,8]

Wat

er fl

ea D

.mag

naF

M21

d st

atic

Gro

wth

/rep

rodu

ctio

n0.

35 (0

.70)

[3]

Wat

er fl

ea D

.mag

naF

M21

d flo

w-t

hru

Rep

rodu

ctio

n0.

26 (0

.76)

[3]

Gro

wth

0.76

Mys

id s

hrim

p M

ysid

opsi

s ba

hia

SM

28d

flow

-thr

uR

epro

duct

ion/

grow

th0.

075

(0.1

7)[5

2]Fi

shR

ainb

ow tr

out O

.myk

iss

FM

109

dSu

rviv

al/g

row

th0.

20[8

]Fa

thea

d m

inno

w P

.pro

mel

asF

M30

dSu

rviv

al/g

row

th0.

14 (0

.36)

[16]

Aquatic Toxicity of Phthalate Esters 271Ta

ble

3(c

onti

nued

)

Test

spe

cies

Fres

h/

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t N

OEC

(LO

EC)

Ref

.sa

ltno

min

al

or c

rite

ria

(mg/

L)(F

/S)

(M/N

)

Dih

exyl

pht

hala

te (D

HP)

Alg

aeS.

capr

icor

nutu

mF

M7

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]In

vert

ebra

tes

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[7,8

]Fi

shR

ainb

ow tr

out O

.myk

iss

FM

111

dSu

rviv

al /g

row

thno

eff

ect a

t sol

ubili

ty[8

]

Buty

l(2-

ethy

lhex

yl)p

htha

late

(BO

P)A

lgae

S.ca

pric

ornu

tum

FM

6d

stat

icC

ell n

umbe

rno

eff

ect a

t sol

ubili

ty[5

,43]

Inve

rteb

rate

sW

ater

flea

D.m

agna

FM

21d

Surv

ival

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[7

,8]

Fish

––

No

info

rmat

ion

Di-

(n-h

exyl

,n-o

ctyl

,n-d

ecyl

)pht

hala

te(6

10P)

Alg

aeS.

capr

icor

nutu

mF

M6

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]In

vert

ebra

tes

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[7,8

]Fi

sh–

–N

o in

form

atio

n

Di-

(2-e

thyl

hexy

l) P

htha

late

(DEH

P)A

lgae

S.ca

pric

ornu

tum

FM

6d

stat

icC

ell n

umbe

rno

eff

ect a

t sol

ubili

ty[5

,43]

Inve

rteb

rate

sW

ater

flea

D.m

agna

FM

21d

stat

icSu

rviv

al/ g

row

th/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[3]

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[53]

272 C.A. Bradlee and P. ThomasTa

ble

3(c

onti

nued

)

Test

spe

cies

Fres

h/

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t N

OEC

(LO

EC)

Ref

.sa

ltno

min

al

or c

rite

ria

(mg/

L)(F

/S)

(M/N

)

Inve

rteb

rate

sW

ater

flea

D.m

agna

FM

21d

Surv

ival

no e

ffec

t at s

olub

ility

[7,8

]W

ater

flea

D.m

agna

FM

21d

Surv

ival

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

1]W

ater

flea

D.m

agna

FM

21d

sem

i-st

atic

Rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

7]W

ater

flea

D.m

agna

FM

21d

Surv

ival

/gro

wth

/rep

rod.

no e

ffec

t at s

olub

ility

[12]

Gra

ss S

hrim

p P.

pugi

oS

M28

dLa

rval

mor

talit

yno

eff

ect a

t sol

ubili

ty[1

8]M

usse

l Myt

ilus

edul

isS

M28

dM

orta

lity,

byss

al th

read

att

achm

ent

no e

ffec

t at s

olub

ility

[54]

Fish

Rai

nbow

trou

t O.m

ykis

sF

N90

d flo

w-t

hru

Gro

wth

no e

ffec

t at s

olub

ility

[23]

Rai

nbow

trou

t O.m

ykis

sF

M90

d flo

w-t

hru

Hat

chab

ility

,sur

viva

l,gr

owth

no e

ffec

t at s

olub

ility

[21]

Broo

k tr

out S

alve

linus

font

inal

isF

N15

0d

flow

-thr

uG

row

thno

eff

ect a

t sol

ubili

ty[2

3]Fa

thea

d m

inno

w P

.pro

mel

asF

N12

7d

flow

-thr

uG

row

th in

hibi

tion

no e

ffec

t at s

olub

ility

[23]

Med

aka

Ori

zias

lati

pes

FM

168

d flo

w-t

hru

Gro

wth

no e

ffec

t at s

olub

ility

[21]

Stic

kleb

ack

Gas

tero

steu

s ac

ulea

tus

FN

28d

mor

talit

y,gr

owth

,sub

leth

al e

ffec

tsno

eff

ect a

t sol

ubili

ty[2

2]Z

ebra

fish

Brac

hyda

nio

reri

oF

N28

dm

orta

lity,

subl

etha

l eff

ects

,gro

wth

no e

ffec

t at s

olub

ility

[22]

Med

aka

O.l

atip

esF

N28

dm

orta

lity,

subl

etha

l eff

ects

,gro

wth

no e

ffec

t at s

olub

ility

[22]

Flag

fish

Jord

anel

la fl

orid

aeF

N28

dm

orta

lity,

subl

etha

l eff

ects

,gro

wth

no e

ffec

t at s

olub

ility

[22]

Gup

py P

oeci

lia r

etic

ulat

aF

N28

dm

orta

lity,

subl

etha

l eff

ects

,gro

wth

no e

ffec

t at s

olub

ility

[22]

Dii

sooc

tyl p

htha

late

(DIO

P)A

lgae

S.ca

pric

ornu

tum

FM

6d

stat

icC

ell n

umbe

rno

eff

ect a

t sol

ubili

ty[5

,43]

Inve

rteb

rate

sW

ater

flea

D.m

agna

FM

21d

Surv

ival

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[7

,8]

Fish

––

No

info

rmat

ion

Dii

sono

nyl p

htha

late

(DIN

P)A

lgae

Sele

nast

rum

cap

rico

rnut

umF

M5

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]

Aquatic Toxicity of Phthalate Esters 273

Tabl

e3

(con

tinu

ed)

Test

spe

cies

Fres

h/

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t N

OEC

(LO

EC)

Ref

.sa

ltno

min

al

or c

rite

ria

(mg/

L)(F

/S)

(M/N

)

Inve

rteb

rate

sW

ater

flea

Dap

hnia

mag

naF

M21

dSu

rviv

alno

eff

ect a

t sol

ubili

ty[7

,8]

Wat

er fl

ea D

aphi

a m

agna

FM

21d

Surv

ival

/gro

wth

/rep

rod.

no e

ffec

t at s

olub

ility

[12]

Fish

––

No

info

rmat

ion

Dii

sode

cyl p

htha

late

(DID

P)

Alg

aeS.

capr

icor

nutu

mF

M8

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]In

vert

ebra

tes

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

alno

eff

ect a

t sol

ubili

ty[7

,8]

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[11]

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[12]

Mus

sel M

ytilu

s ed

ulis

SM

28d

Mor

talit

y,by

ssal

thre

ad a

ttac

hmen

tno

eff

ect a

t sol

ubili

ty[5

4]

Fish

––

No

info

rmat

ion

Di(

hept

yl,n

onyl

,und

ecyl

)pht

hala

te (7

11P)

Alg

aeS.

capr

icor

nutu

mF

M7

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]In

vert

ebra

tes

Wat

er fl

ea D

.mag

naF

M21

dSu

rviv

alno

eff

ect a

t sol

ubili

ty[7

,8]

Fish

Rai

nbow

trou

t O.m

ykis

sF

M12

0d

Surv

ival

/gro

wth

no e

ffec

t at s

olub

ility

[8]

Di-

unde

cyl p

htha

late

(DU

P)A

lgae

S.ca

pric

ornu

tum

FM

8d

stat

icC

ell n

umbe

rno

eff

ect a

t sol

ubili

ty[5

,43]

274 C.A. Bradlee and P. Thomas

Tabl

e3

(con

tinu

ed)

Test

spe

cies

Fres

h/

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t N

OEC

(LO

EC)

Ref

.sa

ltno

min

al

or c

rite

ria

(mg/

L)(F

/S)

(M/N

)

Inve

rteb

rate

sW

ater

flea

D.m

agna

FM

21d

Surv

ival

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[7

,8]

Fish

Rai

nbow

trou

t Onc

orhy

nchu

s m

ykis

sF

M12

0d

Surv

ival

/gro

wth

no e

ffec

t at s

olub

ility

[8]

Dit

ride

cyl p

htha

late

(DTD

P)A

lgae

Sele

nast

rum

cap

rico

rnut

umF

M8

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]In

vert

ebra

tes

Wat

er fl

ea D

aphn

ia m

agna

FM

21d

Surv

ival

no e

ffec

t at s

olub

ility

[7,8

]Fi

sh–

–N

o in

form

atio

n

Not

e:N

OEC

=N

o ob

serv

ed e

ffec

t con

cent

rati

on,L

OEC

=Lo

wes

t obs

erve

d ef

fect

con

cent

rati

on,M

=M

easu

red

expo

sure

con

cent

rati

ons

used

to c

alcu

-la

te r

esul

ts,N

=N

omin

al c

once

ntra

tion

s us

ed t

o ca

lcul

ate

resu

lts (

whe

ther

or

not

conc

entr

atio

ns w

ere

anal

ytic

ally

con

firm

ed),

F=

Fres

hwat

er,

S=Sa

ltwat

er.

*Va

lues

from

refe

renc

e [5

] wer

e re

calc

ulat

ed fr

om o

rigi

nal c

ontr

act l

abor

ator

y re

port

s per

cur

rent

USE

PA p

ract

ice.

Ori

gina

l val

ues w

ere

base

d on

mea

-su

red

init

ial t

est c

once

ntra

tion

;cur

rent

val

ues

are

base

d on

ave

ragi

ng in

itia

l and

fina

l mea

sure

d te

st c

once

ntra

tion

s.

Aquatic Toxicity of Phthalate Esters 275Ta

ble

4.C

hron

ic to

xici

ty o

fpht

hala

te e

ster

s to

alg

ae

Test

spe

cies

Phth

alat

e Fr

esh/

salt

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t N

OEC

(LO

EC)

Ref

.es

ter

(F

/S)

nom

inal

or

cri

teri

a(m

g/L)

(M/N

)

Phth

alat

e es

ters

<C

6A

lgae

Sele

nast

rum

cap

rico

rnut

umD

MP

FM

6d

stat

icC

ell n

umbe

r(6

4.7)

*[5

,43]

S.ca

pric

ornu

tum

DEP

FM

8d

stat

icC

ell n

umbe

r3.

65[5

,43]

S.ca

pric

ornu

tum

DBP

FM

10d

stat

icC

ell n

umbe

r0.

21[5

,43]

Scen

edes

mus

sub

spic

atus

DBP

FM

7d

Gro

wth

rat

e6.

1[1

4]S.

subs

pica

tus

DBP

FM

72h

Cel

l gro

wth

0.5

[15]

S.su

bspi

catu

sD

BPF

M72

hG

row

th r

ate

0.5

[15]

Nav

icul

a pe

llicu

losa

BBP

FM

96h

Cel

l num

ber

0.3

[16]

Sele

nast

rum

cap

rico

rnut

umBB

PF

M96

hC

ell n

umbe

r0.

1[1

6]S.

capr

icor

nutu

mBB

PF

M6

d st

atic

Cel

l num

ber

(0.1

0)[5

,43]

Dun

alie

lla te

rtio

lect

aBB

PS

M96

hC

ell n

umbe

r 0.

3 [1

6]Sk

elet

onem

a co

stat

um

BBP

S M

96

h C

ell n

umbe

r 0.

1 [1

6]

Phth

alat

e es

ters

≥C

6A

lgae

Sele

nast

rum

cap

rico

rnut

um

DH

P F

M

7d

stat

ic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5

,43]

S.ca

pric

ornu

tum

B

OP

F M

6

d st

atic

C

ell n

umbe

r no

eff

ect a

t sol

ubili

ty[5

,43]

S.ca

pric

ornu

tum

610P

FM

6d

stat

icC

ell n

umbe

rno

eff

ect a

t sol

ubili

ty[5

,43]

S.ca

pric

ornu

tum

DEH

PF

M6

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]S.

capr

icor

nutu

mD

IOP

FM

6d

stat

icC

ell n

umbe

rno

eff

ect a

t sol

ubili

ty[5

,43]

S.ca

pric

ornu

tum

DIN

PF

M5

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]S.

capr

icor

nutu

mD

IDP

FM

8d

stat

icC

ell n

umbe

rno

eff

ect a

t sol

ubili

ty[5

,43]

S.ca

pric

ornu

tum

711P

FM

7d

stat

icC

ell n

umbe

rno

eff

ect a

t sol

ubili

ty[5

,43]

S.ca

pric

ornu

tum

DU

PF

M8

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]S.

capr

icor

nutu

mD

TD

PF

M8

d st

atic

Cel

l num

ber

no e

ffec

t at s

olub

ility

[5,4

3]

Not

e:N

OEC

=N

o ob

serv

ed e

ffec

t con

cent

rati

on,L

OEC

= L

owes

t obs

erve

d ef

fect

con

cent

rati

on,M

=M

easu

red

expo

sure

con

cent

rati

ons

used

to c

alcu

-la

te r

esul

ts,N

=N

omin

al c

once

ntra

tion

s us

ed t

o ca

lcul

ate

resu

lts (

whe

ther

or

not

conc

entr

atio

ns w

ere

anal

ytic

ally

con

firm

ed),

F=

Fres

hwat

er,

S=Sa

ltwat

er

*Va

lues

from

refe

renc

e [5

] wer

e re

calc

ulat

ed fr

om o

rigi

nal c

ontr

act l

abor

ator

y re

port

s per

cur

rent

USE

PA p

ract

ice.

Ori

gina

l val

ues w

ere

base

d on

mea

-su

red

init

ial t

est c

once

ntra

tion

;cur

rent

val

ues

are

base

d on

ave

ragi

ng in

itia

l and

fina

l mea

sure

d te

st c

once

ntra

tion

s.

of the assay covers several generations of cell growth and, therefore, a no ob-servable effect concentrations (NOEC) or lowest observable effect concentration(LOEC) is reported. Considering ≥72-h duration assays as chronic is consistentwith the monitored endpoints of population growth rates, cell number andchlorophyll A content, whereas the typical acute endpoint of mortality is not mea-surable in algal assays. There is chronic toxicity data on fourteen phthalate esterswith five different species of algae, including both freshwater and saltwaterspecies. While the lower phthalates have chronic toxicity, the data clearly showthat phthalate esters with side chains ≥C6 are not chronically toxic to algae; evenat concentrations up to the limit of solubility [5, 14–16].

In the Adams et al. [5] study, fourteen phthalate esters were tested with thefreshwater alga Selenastrum capricornutum and only four phthalate esters werechronically toxic (DMP, DEP, DBP and BBP). Studies by Huels [14], Scholz [15]and Gledhill et al. [16] also showed chronic algae toxicity for DBP and BBP.Overall, for the phthalate esters that were chronically toxic to algae the re-ported NOECs ranged from 0.1 mg/L for BBP to 3.65 mg/L for DEP, which sug-gests that BBP has the highest chronic toxicity to algae. BBP was also found tohave the highest acute toxicity of the <C6 phthalate esters reported in Staples et al. [1].

In summary, DEHP, BOP, 610P, DEHP, DIOP, DINP, DIDP, 711P DUP and DTDPare all expected to be non-toxic chronically to both saltwater and freshwater algae,while DMP, DEP, DBP and BBP have shown both acute and chronic algal toxicity.

3.2Invertebrates

Toxicity test results on invertebrates are shown in Table 5. There is chronic toxi-city data on twenty-two phthalate esters with five different species of inverte-brates, including both freshwater and saltwater species. While the lower phtha-lates have chronic toxicity, the data clearly show that phthalate esters with sidechains ≥C6 are not chronically toxic to invertebrates; even at concentrations up to the limit of solubility [3, 7, 8, 11, 12, 17, 18]. For phthalate esters <C6the chronic toxicity data support a NOEC for saltwater species of 10 mg/L [18] for survival and a reproduction/growth NOEC of 0.075 mg/L. For freshwaterspecies the lowest NOEC reported for <C6 phthalate esters is 0.11 mg/L forgrowth/reproduction.

A lower NOEC of 0.04 mg/L from a study by Tagatz et al. [19] for saltwaterspecies exists but the results are from a mesocosm study with a potentially sig-nificant study design limitation. In this study the authors examined the effects ofDBP on a laboratory- and field-colonized estuarine benthic communities. Sixteenaquaria made of acrylic plastic were filled with clean silica sand (particle size,0.2 to 0.8 mm) taken from Santa Rosa Sound, Florida. Half the aquaria were col-onized by macrobenthic organisms in the laboratory by settling of planktonic lar-vae entrained in continuously-supplied unfiltered seawater from Santa RosaSound. The other eight aquaria were covered with a hardware cloth to excludelarge predators and were placed by SCUBA divers in Santa Rosa Sound so thattheir surfaces were level with the surrounding bottom.

276 C.A. Bradlee and P. Thomas

Aquatic Toxicity of Phthalate Esters 277

Tabl

e5.

Chr

onic

toxi

city

ofp

htha

late

est

ers

to in

vert

ebra

tes

Test

spe

cies

Phth

alat

e Fr

esh/

Mea

sure

d/Te

st

Test

end

poin

t or

crite

ria

NO

EC (L

OEC

) R

ef.

Este

r sa

lt no

min

al

dura

tion

(mg/

L)(F

/S)

(M/N

)

Phth

alat

e es

ters

<C

6In

vert

ebra

tes

Wat

er fl

ea D

aphn

ia m

agna

DM

PF

M21

dSu

rviv

al9.

6 (2

3.0)

[7,8

]G

rass

shr

imp

Pala

emon

etes

pug

ioD

MP

SM

30d

Larv

al m

orta

lity

10(1

00)

[18]

Wat

er fl

ea D

.mag

naD

EPF

M21

dSu

rviv

al/r

epro

duct

ion

25 (5

9)[7

,8]

Wat

er fl

ea D

.mag

naD

BPF

M21

dSu

rviv

al0.

96 (2

.5)

[7,8

]W

ater

flea

D.m

agna

DBP

FM

21d

Surv

ival

/rep

rodu

ctio

n0.

11 (0

.20)

[21]

Wat

er fl

ea D

.mag

naD

BPF

M21

dSu

rviv

al/r

epro

duct

ion

1.05

(1.9

1)[2

1]W

ater

flea

D.m

agna

DBP

FM

21d

Surv

ival

/rep

rodu

ctio

n0.

16 (0

.64)

[21]

Gra

ss s

hrim

p P.

pugi

oD

BPS

M30

dLa

rval

mor

talit

y10

.0 (5

0.0)

[18]

Estu

arin

e m

icro

cosm

DBP

SM

2w

kA

bund

ance

and

div

ersi

ty0.

04 (0

.34)

[19,

51]

Wat

er fl

ea D

.mag

naBB

PF

M42

dSu

rviv

al/r

epro

duct

ion

0.26

(0.7

6)[1

6]W

ater

flea

D.m

agna

BBP

FM

21d

Surv

ival

/rep

rodu

ctio

n0.

28 (1

.4)

[7,8

]W

ater

flea

D.m

agna

BBP

FM

21d

stat

icG

row

th/r

epro

duct

ion

0.35

(0.7

0)[3

]W

ater

flea

D.m

agna

BBP

FM

21d

flow

-thr

uR

epro

duct

ion

0.26

(0.7

6)[3

]G

row

th0.

76[3

]M

ysid

shr

imp

Mys

idop

sis

bahi

aBB

PS

M28

d flo

w-t

hru

Rep

rodu

ctio

n/gr

owth

0.07

5 (0

.17)

[52]

Phth

alat

e es

ters

≥C

6 In

vert

ebra

tes

Wat

er fl

ea D

.mag

naD

HP

F M

21

d Su

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[7

,8]

Wat

er fl

ea D

.mag

naB

OP

F M

21

d Su

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[7

,8]

Wat

er fl

ea D

.mag

na61

0P

F M

21

d Su

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[7,8

]W

ater

flea

D.m

agna

DEH

PF

M21

d st

atic

Surv

ival

/ gro

wth

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[3

]W

ater

flea

D.m

agna

DEH

PF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[53]

278 C.A. Bradlee and P. Thomas

Tabl

e5

(con

tinu

ed)

Test

spe

cies

Phth

alat

e Fr

esh/

Mea

sure

d/Te

st

Test

end

poin

t or

crite

ria

NO

EC (L

OEC

) R

ef.

Este

r sa

lt no

min

al

dura

tion

(mg/

L)(F

/S)

(M/N

)

Wat

er fl

ea D

.mag

naD

EHP

FM

21d

Surv

ival

no e

ffec

t at s

olub

ility

[7,8

]W

ater

flea

D.m

agna

DEH

PF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[11]

Wat

er fl

ea D

.mag

naD

EHP

FM

21d

sem

i-st

atic

Rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

7]W

ater

flea

D.m

agna

DEH

PF

M21

dSu

rviv

al/g

row

th/r

epro

d.no

eff

ect a

t sol

ubili

ty[1

2]G

rass

Shr

imp

P.pu

gio

DEH

PS

M28

dLa

rval

mor

talit

yno

eff

ect a

t sol

ubili

ty[1

8]M

usse

l Myt

ilus

edul

isD

EHP

SM

28d

Mor

talit

y,by

ssal

thre

ad a

ttac

hmen

tno

eff

ect a

t sol

ubili

ty[5

4]W

ater

flea

D.m

agna

DIH

PF

M21

d st

atic

Surv

ival

/gro

wth

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty

[12]

Wat

er fl

ea D

.mag

naD

IOP

FM

21d

Surv

ival

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[7

,8]

Wat

er fl

ea D

.mag

naD

INP

FM

21d

Surv

ival

no e

ffec

t at s

olub

ility

[7,8

]W

ater

flea

D.m

agna

DIN

PF

M21

dSu

rviv

al/g

row

th/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[12]

Wat

er fl

ea D

.mag

naD

IDP

FM

21d

Surv

ival

no e

ffec

t at s

olub

ility

[7,8

]W

ater

flea

D.m

agna

DID

PF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[11]

Wat

er fl

ea D

.mag

naD

IDP

FM

21d

Surv

ival

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

2]M

usse

l Myt

ilus

edul

isD

IDP

SM

28d

Mor

talit

y,by

ssal

thre

ad a

ttac

hmen

tno

eff

ect a

t sol

ubili

ty[5

4]W

ater

flea

D.m

agna

711P

FM

21d

Surv

ival

no e

ffec

t at s

olub

ility

[7,8

]W

ater

flea

D.m

agna

DU

PF

M21

dSu

rviv

al/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[7,8

]W

ater

flea

D.m

agna

DT

DP

FM

21d

Surv

ival

no e

ffec

t at s

olub

ility

[7,8

]W

ater

flea

D.m

agna

L911

PF

M21

d st

atic

Surv

ival

/gro

wth

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

2]W

ater

flea

D.m

agna

L810

PF

M21

d st

atic

Surv

ival

/gro

wth

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

2]W

ater

flea

D.m

agna

DIH

PF

M21

d st

atic

Surv

ival

/gro

wth

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

2]W

ater

flea

D.m

agna

DPH

PF

M21

d st

atic

Surv

ival

/gro

wth

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

2]W

ater

flea

D.m

agna

L79P

FM

21d

stat

icSu

rviv

al/g

row

th/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[12]

Wat

er fl

ea D

.mag

naD

IUP

FM

21d

stat

icSu

rviv

al/g

row

th/r

epro

duct

ion

no e

ffec

t at s

olub

ility

[12]

Wat

er fl

ea D

.mag

naD

TD

PF

M21

d st

atic

Surv

ival

/gro

wth

/rep

rodu

ctio

nno

eff

ect a

t sol

ubili

ty[1

2]

Not

e:N

OEC

=N

o ob

serv

ed e

ffec

t con

cent

rati

on,L

OEC

=Lo

wes

t obs

erve

d ef

fect

con

cent

rati

on,M

=M

easu

red

expo

sure

con

cent

rati

ons

used

to c

alcu

-la

te r

esul

ts,N

=N

omin

al c

once

ntra

tion

s us

ed t

o ca

lcul

ate

resu

lts (

whe

ther

or

not

conc

entr

atio

ns w

ere

anal

ytic

ally

con

firm

ed),

F=

Fres

hwat

er,

S=Sa

ltwat

er.

The laboratory and field aquaria were allowed to be colonized for 8 weeks (Oc-tober 5 – November 30, 1981). After colonization, the field tanks were collectedand both the field tanks and laboratory tanks were exposed to DBP for 2 weeks(December 1–15, 1981). DBP dissolved into a carrier solution of 60% acetone and40% distilled water was dosed at nominal concentrations of 0.05, 0.5 and 5 mg/Linto a continuous supply of seawater that was pumped to the aquaria at 1.5 L/min.An equal amount of carrier solution (0.5 ml/L) was metered in to the water thatentered the control aquaria.

Water samples were taken three times weekly and analyzed for DBP content.The measured average and standard deviation of DBP concentrations in the laboratory-colonized aquaria were 0.044 mg/L ±0.014, 0.34 mg/L ±0.12, and3.7 mg/L ±0.70 and field-colonized aquaria were 0.036 mg/L ±0.005, 0.45 mg/L±0.14 and 3.8 ±0.62. After a 2-week exposure to DBP, organisms were collectedusing a 1-mm mesh sieve and were preserved and identified.

In the laboratory-colonized communities a total of 2331 animals representing29 species in 6 phyla were collected, including Chordata, Mollusca, Anthropoda,Annelida, Echinodermata, and Coelenterates. Community structure was clearlyreduced at the highest concentration (3.7 mg/L), which showed statistically sig-nificant (p=0.05) decreases in the average density and average number of speciesfor all phyla. At the mid-dose (0.34 mg/L) the only phylum affected was Anthro-poda which had a reduced average density. The average number of species wasnot significantly reduced and, moreover, none of the other phyla demonstratedany toxicity. In several instances the density and number of species in the DBPtreatments were higher than the control.Arthropods were identified as the mostsensitive phyla with a NOEC and LOEC of 0.04 mg/L and 0.34 mg/L, respectively.For the other phylum the NOEC and LOEC are 0.34 mg/L and 3.7 mg/L, respec-tively. In the field study, significantly fewer organisms were present. Only 181 an-imals representing 46 species and 7 phyla were collected, which included; Mol-lusca, Annelida, Rhynchoceia, Anthropoda, Cordata, Echinodermata, andCoelenterates. However, mollusks and annelids dominated in number and rela-tively few animals from other phyla were collected. The results of this part of thestudy showed that DBP exposure resulted in a statistically significant reductionin the average density and average number of species only in the phylum Mol-lusca, and only in the highest concentration. Therefore, the field experiment didnot reflect the observations from the laboratory experiment.

It should be noted that there was a study design limitation that had the po-tential to significantly confound the findings. The authors used acetone as a sol-vent but did not use a solvent-free control, so that any effects of acetone in the di-lution water (0.3 mL/L) on the macrobenthic communities were not determined.Moreover, the authors acknowledge that acetone may have effected the bioavail-ability of DBP thereby artificially altering the toxicity of DBP to the benthic com-munities.

A more relevant study with Daphnia magna further confirms that ≥C6phthalate esters are not chronically toxic to invertebrates. Brown et al. [12] pub-lished results from a 21-day reproduction test performed on Daphnia magna.Twelve phthalate esters were tested, which included: DEHP, L911P, DINP, L810P,DIDP, DIHP, DPHP, 610P, DUP, L79P, DIUP and DTDP. The test was conducted

Aquatic Toxicity of Phthalate Esters 279

using OECD 202 part ii methods that were modified to provide one daphnid pertest vessel, as recommended in the more recently modified OECD 211 guideline.Each phthalate ester was tested at a single nominal concentration of 1 mg/L, ex-cept DEHP which was tested at 0.25 mg/L.

The endpoints for this test were survival of parental organisms (Po), the num-ber of offspring (F1) and the mean body length of surviving Po. Previous at-tempts to perform chronic studies on daphnids resulted in “floaters” when theconcentrations of ≥C6 phthalates exceeded the stable colloidal suspension con-centration of the phthalate (approximately between 200 µg/L for phthalatesgreater than C6). In order to avoid this problem, the authors used a dispersant,Marlowet R40 (a castor oil 40 mole ethoxylate), with a ratio of phthalate to dis-persant of 1 :10. The test was conducted in M4 Elendt Schneider medium, usinga concentration of 1 mg/L of DEHP in 10 mg/L dispersant and a solvent controlat 10 mg/L Marlowet R40 was run beside the negative control. Daphnids were feddaily with a specified quantity of suspension of Chlorella vulgaris cells and Frip-pak booster. Water quality parameters (O2 concentration, pH, etc) were main-tained within the required limits throughout the study. With only one exception(L911P), the concentrations of the phthalates were maintained to plus or minus20% of the nominal concentration as confirmed by analysis.

The results of this study showed there was no significant difference betweenboth controls and any the test concentrations for survival of parental organisms(Po), the number of offspring (F1) and the mean body length of surviving Po. Infact the number of offspring were almost always higher in the phthalate expo-sures than in the control. Therefore a NOEC >1 mg/L is established for the testedphthalate esters except DEHP that had a NOEC >0.25 mg/L, which was the high-est concentration tested. Furthermore, these results demonstrate that mortalityin previous studies carried out on D. magna, at concentrations exceeding col-loidal stability in water, was likely related to physical effects of phthalate in sus-pension or as a surface film and not to any toxicological properties of the phthalates.

Typically, the route of administration for the phthalate esters in invertebratestudies is via the water; however, there has been a study in which Penaeid shrimpwere exposed via food. Hobson et al. [20] conducted a study in which Penaeidshrimp (Penaeus vannamei) were fed diets containing 40 ppm to 50,000 ppmDEHP for 14 days at 4% body weight per day. The results of this study showed noincrease in mortality or histopathological alterations in any of the doses. More-over, whole-body DEHP residues in the shrimp were 18 ppm at the highest dose,and bioconcentration factors were inversely proportional to dose. These findings,as well as findings of other dietary exposure studies, are summarized in Table 6.

In conclusion, phthalate esters with alkyl chain lengths equal to or greater thanC6 were not toxic to invertebrates at concentrations at or above their solubilitylimit. For the lower phthalates (<C6), in terms of toxic potency, the reportedNOECs support the finding that BBP >DBP >DMP >DEP and that the saltwaterspecies Mysid shrimp (Mysidopsis bahia) may be a more sensitive species thanthe freshwater Daphnia magana.

280 C.A. Bradlee and P. Thomas

Aquatic Toxicity of Phthalate Esters 281Ta

ble

6.Fe

edin

g st

udie

s us

ing

phth

alat

e es

ters

wit

h aq

uati

c or

gani

sms

Test

spe

cies

Phth

alat

e Fr

esh/

M

easu

red/

Test

dur

atio

nTe

st e

ndpo

int o

r cr

iteri

aN

OEC

(LO

EC)

Ref

.Es

ter

salt

nom

inal

(m

g/kg

food

)(F

/S)

(M/N

)

Phth

alat

e es

ters

<C

6Fi

shM

edak

a O

ryzi

as la

tipe

sD

BPF

M2

gene

rati

onM

orta

lity,

hist

opat

holo

gy,g

row

th,g

onad

al:

5 (5

0)

[28]

som

atic

inde

x,se

xual

dev

elop

men

t,fe

cund

ity,

embr

yona

l dev

elop

men

t,vi

tello

geni

n in

duc-

tion

,hep

atic

mic

roso

mal

test

oste

rone

m

etab

olis

mPh

thal

ate

este

rs ≥

C6

Inve

rteb

rate

s Pe

naei

d sh

rim

p Pe

naeu

s va

nnam

ei

DEH

P S

N

14d

Mor

talit

y,m

olti

ng>

50,0

00*

[ 20]

Fish

A

tlant

ic s

alm

on

DEH

P F

N

4w

ks

Sex

rati

o,liv

er:s

omat

ic in

dex

300

(150

0)[3

4]Sa

lmo

sala

rA

tlant

ic s

alm

on

DEH

PF

N4

wks

pos

t-

Gro

wth

,sur

viva

l,se

x ra

tio,

liver

:som

atic

>

1500

*fo

od[2

9]Sa

lmo

sala

r (R

epea

t of

yolk

-sac

inde

xm

g/kg

stud

y R

ef.1

20)

reso

rpti

onR

ainb

ow tr

out

DEH

PF

M7

wks

Pero

xiso

me

prol

ifera

tion

>20

,000

*[2

4]O

ncor

hync

hus

myk

iss

Med

aka

O.l

atip

esD

INP

FM

3ge

nera

tion

Mor

talit

y,hi

stop

atho

logy

,gro

wth

,gon

adal

:>

20*

[33]

som

atic

inde

x,se

xual

dev

elop

men

t,fe

cund

ity,

embr

yona

l dev

elop

men

t,m

icro

som

al te

st-

oste

rone

met

abol

ism

Med

aka

O.l

atip

esD

IDP

FM

3ge

nera

tion

Mor

talit

y,hi

stop

atho

logy

,gro

wth

,gon

adal

:>

20*

[33]

som

atic

inde

x,se

xual

dev

elop

men

t,fe

cund

ity,

embr

yona

l dev

elop

men

t,m

icro

som

al te

st-

oste

rone

met

abol

ism

Not

e:N

OEC

=N

o ob

serv

ed e

ffec

t con

cent

rati

on,L

OEC

=Lo

wes

t obs

erve

d ef

fect

con

cent

rati

on,M

=M

easu

red

expo

sure

con

cent

rati

ons

used

to c

alcu

-la

te r

esul

ts,N

=N

omin

al c

once

ntra

tion

s us

ed t

o ca

lcul

ate

resu

lts (

whe

ther

or

not

conc

entr

atio

ns w

ere

anal

ytic

ally

con

firm

ed),

F=

Fres

hwat

er,

S=Sa

ltwat

er.

*N

o tr

eatm

ent r

elat

ed e

ffec

ts a

t hig

hest

feed

con

cent

rati

on te

sted

.

3.3Fish

Toxicity test results on fish are shown in Table 7. Data are presented on seven phthalate esters with eight different species of fish. While the lower phthalatesshow chronic toxicity, the data clearly demonstrate that phthalate esters with sidechains of ≥C6 are not chronically toxic to fish, even at concentrations up to thelimit of solubility. For phthalates with <C6 side chains the NOECs ranged from0.1 to 11 mg/L. The lowest reported NOEC was 0.1 mg/L for growth in a 60 daypost-hatch early life-stase (ELS) study on Rainbow trout (Oncorhynchus mykiss)with DBP [8]. In Rhodes et al. [8], ELS studies with Rainbow trout were conductedon six different phthalate esters. Chronic effects were observed for DMP (sur-vival) at 24 mg/L and DBP (growth) at 0.19 mg/L, resulting in NOECS for DMPand DBP of 11 and 0.1 mg/L, respectively. The other phthalates tested in this studyincluded BBP, DHP, 711P and DUP, and their NOECS were all greater than thehighest concentration tested. For DHP, 711P and DUP the high concentrationtested was at or above their solubility and, therefore, these phthalates demon-strated no chronic toxicity in the ELS study with Rainbow trout at or above thelimit of solubility. No effects at the limit of solubility have been demonstrated inother species of fish for phthalate esters with alkyl chain lengths >C6. In chronicstudies with Brook trout, Fathead minnow, Medaka, Stickleback, Zebrafish, andGuppy no effects on mortality or growth were seen with exposure to DEHP at orabove the solubility limit [21–23].

The results summarized in Table 7 also show that survival can be equally assensitive and sometimes a more sensitive endpoint than growth; however, otherendpoints such has lipid metabolism in fish have been studied also. As shown inTable 6 a dietary study by Henderson and Sargent [24] was performed on DEHPto determine its effects on lipid metabolism in Rainbow trout. The authors carriedout the study as DEHP is a known peroxisome proliferator in rat liver and theywished to examine potential effects of this kind in the aquatic environment fromexposure via contaminated food. The trout were fed on a diet of freeze dried zoo-plankton mixed with DEHP at a concentration of 20,000 ppm for a period of sevenweeks before sacrifice. Liver homogenates were prepared and assayed for mito-chondria and peroxisomes. Lipids were extracted from specific fish tissues and thelipid profile analyzed. Overall percent lipid content was significantly lower thanthe control in the adipose tissue and the liver; however, the authors did not observeany significant difference in body weight between controls and the exposed groupor in liver weight (peroxisome proliferation in rats classically increases liverweight). Prepared liver peroxisomes from DEHP exposed fish did not influenceoxidising activity in-vitro compared to the control. The authors concluded thatthis finding makes it is unlikely that DEHP pose a significant threat to lipid ca-tabolism in fish in the natural environment.

In summary, the results of the studies on fish show those phthalate esters withalkyl chain lengths greater than C6 were not toxic to fish at concentrations at orabove their solubility limit. This includes mortality, growth and other sub-lethaleffects. For the lower phthalates (<C6), in terms of toxic potency, the reportedNOECs generally support the finding that BBP >DBP >DMP >DEP.

282 C.A. Bradlee and P. Thomas

Aquatic Toxicity of Phthalate Esters 283Ta

ble

7.C

hron

ic to

xici

ty o

fpht

hala

te e

ster

s to

fish

Test

spe

cies

Phth

alat

eFr

esh/

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t or

crite

ria

NO

EC (L

OEC

) R

ef.

este

rsa

lt no

min

al

(mg/

L)(F

/S)

(M/N

)

Phth

alat

e es

ters

<C

6Fi

shR

ainb

ow tr

out O

ncor

hync

hus

myk

iss

DM

PF

M60

d po

st-h

atch

Surv

ival

11 (2

4)[8

]R

ainb

ow tr

out O

.myk

iss

DBP

FM

60d

post

-hat

chG

row

th

0.1

(0.1

9)[8

]Fa

thea

d m

inno

w P

imep

hale

pro

mel

asD

BPF

N20

d flo

w-t

hru

Hat

chab

ility

0.56

(1.0

) [9

]Fa

thea

d m

inno

w P

.pro

mel

asD

BPF

M14

4 h

flow

-thr

u M

orta

lity

0.32

[46]

Rai

nbow

trou

t O.m

ykis

sBB

PF

M10

9d

Surv

ival

/gro

wth

>0.

20[8

]Fa

thea

d m

inno

w P

.pro

mel

asBB

PF

M30

dSu

rviv

al/g

row

th0.

14 (0

.36)

[1

6]

Phth

alat

e es

ters

≥C

6Fi

shR

ainb

ow tr

out O

.myk

iss

DH

PF

M11

1d

Surv

ival

/gro

wth

no e

ffec

t at s

olub

ility

[8]

Rai

nbow

trou

t O.m

ykis

sD

EHP

FN

90d

flow

-thr

uG

row

th

no e

ffec

t at s

olub

ility

[23]

Rai

nbow

trou

t O.m

ykis

sD

EHP

FM

90d

flow

-thr

uH

atch

abili

ty,s

urvi

val,

grow

thno

eff

ect a

t sol

ubili

ty[2

1]Br

ook

trou

t Sal

velin

us fo

ntin

alis

DEH

PF

N15

0d

flow

-thr

uG

row

th

no e

ffec

t at s

olub

ility

[23]

Fath

ead

min

now

P.p

rom

elas

DEH

PF

N12

7d

flow

-thr

uG

row

thno

eff

ect a

t sol

ubili

ty[2

3]M

edak

a O

rizi

as la

tipe

sD

EHP

FM

168

d flo

w-t

hru

Gro

wth

no e

ffec

t at s

olub

ility

[21]

Stic

kleb

ack

Gas

tero

steu

s ac

ulea

tus

DEH

PF

N28

dm

orta

lity,

grow

th,

no e

ffec

t at s

olub

ility

[22]

subl

etha

l eff

ects

Zeb

rafis

h Br

achy

dani

o re

rio

DEH

PF

N28

dm

orta

lity,

grow

th,

no e

ffec

t at s

olub

ility

[22]

subl

etha

l eff

ects

Med

aka

O.l

atip

esD

EHP

FN

28d

mor

talit

y,gr

owth

,no

eff

ect a

t sol

ubili

ty[2

2]su

blet

hal e

ffec

ts

Flag

fish

Jord

anel

la fl

orid

aeD

EHP

FN

28d

mor

talit

y,gr

owth

,no

eff

ect a

t sol

ubili

ty[2

2]su

blet

hal e

ffec

ts

Gup

py P

oeci

lia r

etic

ulat

aD

EHP

FN

28d

mor

talit

y,gr

owth

,no

eff

ect a

t sol

ubili

ty[2

2]su

blet

hal e

ffec

tsR

ainb

ow tr

out O

.myk

iss

711P

FM

120

dSu

rviv

al/g

row

thno

eff

ect a

t sol

ubili

ty[8

]R

ainb

ow tr

out O

.myk

iss

DU

PF

M12

0d

Surv

ival

/gro

wth

no e

ffec

t at s

olub

ility

[8]

Not

e:N

OEC

=N

o ob

serv

ed e

ffec

t con

cent

rati

on,L

OEC

=Lo

wes

t obs

erve

d ef

fect

con

cent

rati

on,M

=M

easu

red

expo

sure

con

cent

rati

ons

used

to c

alcu

-la

te r

esul

ts,N

=N

omin

al c

once

ntra

tion

s us

ed t

o ca

lcul

ate

resu

lts (

whe

ther

or

not

conc

entr

atio

ns w

ere

anal

ytic

ally

con

firm

ed),

F=

Fres

hwat

er,

S=Sa

ltwat

er.

3.4Endocrine Modulating Effects

Recently, much attention has been given to the potential for endocrine modulat-ing effects of phthalate esters on aquatic organisms. A number of studies havebeen published on phthalates since 1996 that have shifted focus to estrogenicproperties and away from standard chronic endpoints such as growth. The fol-lowing section summarizes the findings of studies with endocrine endpoints forboth the higher and lower phthalates, with the results summarized in Table 8.

3.4.1Lower Phthalates

For DBP a number of in-vitro and in-vivo studies have been performed sug-gesting equivocal estrogenic effects of this substance.An in-vitro receptor bind-ing trout hepatocytes assay resulted in an IC50 of 1 mM (0.28 mg/L), a reportergene assay on yeast was positive above 1 µM but on trout hepatocytes DBP wasinactive up to 100 µM (highest concentration tested) [25]. An in-vivo study ontrout, DBP intra-peritoneally injected did not increase yolk protein precursorsynthesis at a very high dose level (50 mg/kg), while an in-vivo frog sex deter-mination study [26] it did result in sex reversal and inter-sex when tadpoles wereexposed to 2.8 mg/L (10 µM) for a 5 day period at the critical life stage.

Ohtani et al. [26] examined the estrogenic potential of DBP by examining phe-notypically male populations of the Japanese wrinkled frog (Rana rugosa). Thesex-determining chromosomes of this species vary from one population to an-other allowing the authors to produce laboratory bred F1 generation, geneticmales (XZ) making these animals particularly suitable for determination of in-tersex and sex-reversal effects due to external sources. Fifty genetically male eggswere hatched and maintained in water until 19 days after fertilization. Test con-centrations of DBP and 17b-Estradiol (E2), as a positive control, were preparedat 0.028, 0.28 and 2.8/L respectively in acetone (100 µl/L) and tadpoles were ex-posed to these solutions from day 19 to 23 at a temperature of approximately25 °C. The concentration of DBP was chosen as one tenth of the five minutesLC100 of the tadpoles. Controls were exposed to 100 µl/L acetone only at thistime. The tadpoles were then replaced and maintained in uncontaminated labo-ratory water until day 40 when the gonads of selected tadpoles (approximately30 per group) were removed for histological inspection. Results from this studyare presented in Table 9.

At the highest concentration of DBP (2.8 mg/L) the structure of one tadpolegonad was found to be entirely ovarian while four were mixed ovarian and tes-ticular. At 0.28 mg/L two tadpole gonads were observed to contain ovarian andtesticular tissue although none were entirely ovarian. For E2, the highest con-centration (0.38 mg/L) resulted in gonadal transformation to ovaries in all tad-poles exposed. Despite the relative increase in number of cases of observationsof meiotic cells, the authors do not consider this to be an objective measurementof feminization as they noted these cells in many of the tadpoles from the naturalpopulation of this species. The authors note that the level of feminization brought

284 C.A. Bradlee and P. Thomas

Aquatic Toxicity of Phthalate Esters 285

Tabl

e8.

Supp

lem

enta

l end

poin

ts a

ddre

ssin

g po

tent

ial m

echa

nism

s of

chro

nic

toxi

city

usi

ng a

quat

ic o

rgan

ism

s

Test

spe

cies

Phth

alat

eFr

esh/

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t or

crite

ria

NO

EC (L

OEC

) R

ef.

este

rsa

lt no

min

al

(mg/

L)(F

/S)

(M/N

)

Phth

alat

e es

ters

<C

6A

mph

ibia

ns

Japa

nese

wri

nkle

d fr

og

DBP

FM

4d

post

-G

onad

al s

ex d

iffer

enti

atio

n 0.

28 (

2.8)

mg/

L[2

6]

Ran

a ru

gosa

fert

iliza

tion

(fem

iniz

atio

n of

mal

es)

Fish

Fath

ead

min

now

BB

PF

N6

wks

Num

ber

ofeg

gs s

paw

ned,

num

ber

ofsp

awni

ngs,

>10

0*m

g/L

[29]

Pim

epha

le p

rom

elas

egg

batc

h si

ze,g

onad

o-so

mat

ic in

dex,

vite

lloge

nn

synt

hesi

s an

d th

e se

cond

ary

sexu

al c

hara

cter

-is

tics

,mal

e fa

tpad

and

tube

rcle

form

atio

n M

edak

a O

ryzi

as la

tipe

sD

BPF

M2

gene

rati

onM

orta

lity,

hist

opat

holo

gy,g

row

th,

5 (5

0) m

g/kg

[2

8]go

nada

l:so

mat

ic in

dex,

sexu

al d

evel

opm

ent,

food

fe

cund

ity,

embr

yona

l dev

elop

men

t,vi

tello

geni

n in

duct

ion,

hepa

tic

mic

roso

mal

test

oste

rone

m

etab

olis

m

Phth

alat

e es

ters

≥C

6Fi

shA

tlant

ic s

alm

on

DEH

PF

N4

wks

pos

t-Se

x ra

tio,

liver

:som

atic

inde

x30

0 (1

500)

[3

4]Sa

lmo

sala

ryo

lk-s

ac

mg/

kg fo

odre

sorp

tion

Atla

ntic

sal

mon

Sal

mo

DEH

PF

N4

wks

pos

t-G

row

th,s

urvi

val,

sex

rati

o,liv

er:s

omat

ic in

dex

>15

00*

[29]

sala

r (R

epea

t ofs

tudy

yo

lk-s

ac

mg/

kg fo

odR

ef.1

20)

reso

rpti

onR

ainb

ow tr

out

DEH

PF

M7

wks

Pero

xiso

me

prol

ifera

tion

>20

,000

*[2

4]O

ncor

hync

hus

myk

iss

mg/

kg fo

odM

edak

a O

ryzi

as la

tipe

sD

EHP

FN

90d

Gon

adal

sex

diff

eren

tiat

ion,

tota

l len

gth,

>5,

000*

mg/

L[3

1]w

et w

eigh

t

286 C.A. Bradlee and P. Thomas

Tabl

e8

(con

tinu

ed)

Test

spe

cies

Phth

alat

eFr

esh/

Mea

sure

d/Te

st d

urat

ion

Test

end

poin

t or

crite

ria

NO

EC (L

OEC

) R

ef.

este

rsa

lt no

min

al

(mg/

L)(F

/S)

(M/N

)

Phth

alat

e es

ters

≥C

6Fi

shM

edak

a O

.lat

ipes

DIN

PF

M3

gene

rati

onM

orta

lity,

hist

opat

holo

gy,g

row

th,

>20

*[3

3]go

nada

l:so

mat

ic in

dex,

sexu

al d

evel

opm

ent,

mg/

kg fo

odfe

cund

ity,

embr

yona

l dev

elop

men

t,m

icro

som

al

test

oste

rone

met

abol

ism

Med

aka

O.l

atip

esD

IDP

FM

3ge

nera

tion

Mor

talit

y,hi

stop

atho

logy

,gro

wth

,>

20*

mg/

kg

[33]

gona

dal:

som

atic

inde

x,se

xual

dev

elop

men

t,fo

odfe

cund

ity,

embr

yona

l dev

elop

men

t,m

icro

som

al

test

oste

rone

met

abol

ism

Not

e:N

OEC

=N

o ob

serv

ed e

ffec

t con

cent

rati

on,L

OEC

=Lo

wes

t obs

erve

d ef

fect

con

cent

rati

on,M

=M

easu

red

expo

sure

con

cent

rati

ons

used

to c

alcu

-la

te r

esul

ts,N

=N

omin

al c

once

ntra

tion

s us

ed t

o ca

lcul

ate

resu

lts (

whe

ther

or

not

conc

entr

atio

ns w

ere

anal

ytic

ally

con

firm

ed),

F=

Fres

hwat

er,

S=Sa

ltwat

er.

*N

o tr

eatm

ent r

elat

ed e

ffec

ts a

t hig

hest

feed

con

cent

rati

on te

sted

.

about by the highest concentration of DBP is similar to that of the lowest con-centration of E2 used, and they conclude that in this study DBP is about1000 times less potent than E2.

Overall the results suggest that DBP does have an estrogenic effect under spe-cific conditions. However, conclusions on environmental and population effectsof such endocrine modulation cannot be drawn from the frog study as the sex reversal observed on frog gonads occurred at concentrations of DBP that areknown to be detrimental to other fauna based on other direct toxicological ef-fects. For example, the high concentration in the frog study was 2.8 mg/L whichis comparable to many LC50s from acute studies on DBP. Moreover, a LOEC of0.28 mg/L is established based on effects in the frog testes (2 intersex out of thirtyfrogs), which is nearly thirty times higher than the recommended maximum per-missible concentration (MPC) of 10 µg/L DBP [27] (based on the 60 d NOEC forgrowth of rainbow trout, divided by 10).

The relevance of the frog study can also be questioned at another level.Monobutyl phthalate is the primary metabolite of DBP and it is not active in in-vitro studies; therefore, the metabolism of DBP may be of primary importancein deactivating the diester to the more estrogenically inactive monoester. The exposure concentrations in the frog study were high enough to saturate the enzymes systems such that the gonads may have been exposed to the unmetab-olized DBP. At lower, environmentally relevant, concentrations the DBP is suffi-ciently metabolized that endocrine effects may not occur.A more pertinent studyon possible endocrine effects of DBP was conducted by Patyna et al. [28].

Medaka was used to examine the multigenerational effects of DBP in a dietaryexposure study by Patyna et al. [28]. Also, a positive control consisting of 17b-Estradiol (E2) was conducted as part of this study. Seven treatment groups werestudied, which included an ethanol control, 0.5, 5, 50 mg DBP/kg food, and 0.05,0.5, 5 mg E2/kg food. An ethanol control was tested because each phthalate was

Aquatic Toxicity of Phthalate Esters 287

Table 9. Histological evaluation of genetically male Rana rugosa gonads following E2 and DBPexposure

Concentration Histological evaluation of gonads (mg/L)

Ovarian Ovarian and Testicular Testicular Totalthroughout testicular (many meiotic (many meiotic

germ cells) germ cells)

DBP 0.028 0 0 1 29 300.28 0 2 8 20 302.8 1 4 14 11 30

17b-Estradiol0.0038 1 4 10 13 280.038 5 14 3 8 300.38 28 0 0 0 28Solvent control 0 0 2 28 30

Data from Ohtani et al. [26]

spiked into the food via an ethanol solution. Each treatment group was dividedinto two replicate tanks of 20 fish each. The F0 generation was first exposed as 14-day old larvae by feeding either DBP or E2 in dry flake food at a daily ration of5% body weight. The F0 and F1 generations were fed each treatment through sex-ual maturation. The endpoints for the study were mortality, histopathology,growth, gonadal:somatic index (GSI), sexual development, fecundity, embryonaldevelopment, vitellogenin induction, and hepatic microsomal testosterone me-tabolism. The authors found that all the fish exposed to 0.05 mg E2/kg food werephenotypic females and that no eggs were produced by the E2 treated fish. Incontrast, DBP had no significant treatment related effects on the F0 generation;however, the authors found that in females 50 mg DBP/kg food decreased F1 gen-eration ovary weight, GSI, and egg production and in males it reduced testesweight and GSI in the F1 generation. DBP at 50 mg DBP/kg food increased mi-crosomal protein levels, liver weight and hepatic somatic index in both F1 gen-eration males and females. No adverse effects were reported in the F1 and F2 eggsfrom the DBP treated groups and, furthermore, they showed normal embryonicdevelopment. The sex ratios in DBP treated groups were similar to the control.

As with DBP, BBP equivocal estrogenicity has been observed between in-vitroand in-vivo studies [25]. A trout hepatocyte receptor binding assay had anIC50 for E2 of 10 µM and a yeast reporter gene assay was positive above 1 µM. Forthe monoesters monobutyl and monobenzyl phthalate, metabolites of BBP wereinactive at concentrations between 500 µM and 1 mM. But while in one in-vivostudy on trout injected intra-peritoneally with 50 mg DEHP/kg an induction ofvitellogenin in one fish out of six was observed, and in a water-borne in-vivo BBPstudy on Pimephales promelas by Harries et al. [29] no reproductive effects wereobserved up to concentrations of 82 µg/L. These data suggest that metabolism ofBBP may occur in-vivo and under environmental conditions the substance doesnot arrive intact at the estrogen receptors.

In a review of estrogenic assays by Moore [25] the relevance of in-vitro testson yeast strains, trout hepatocytes and in-vivo studies on trout by intra-peri-toneal injection to phthalates was discussed. For the lower phthalates, receptorbinding assay data using trout hepatocytes exist for DBP with an IC50 at 1 mMBBP had an IC50 for E2 binding of approximately 10 µM and maximum inhibi-tion of 60%. Relative binding of the phthalates compared to E2 in this study was0.1% supporting the in-vivo results found by Ohtani et al. [26] for DBP.

Results from yeast (Saccharomyces cerevisiae) reporter gene assays transfectedwith human estrogen receptor performed on DEP, DBP, DiBP and BBP werefound to demonstrate estrogenic activity. DBP, DiBP and BBP elicited responsesat concentrations above 1 µM reaching a plateau at about 50% implying they arepartial agonists for E2. Mono-n-butyl phthalate and monobenzyl phthalate werefound to be inactive at concentrations between 500 µM and 1 mM. Trout hepa-tocytes are also used in reporter gene assays and will synthesize vitellogeninwhen exposed to estrogens but DBP did not elicit a response at concentrationsup to 100 µM.

In-vivo studies on the synthesis of fish oocyte proteins, vitellogenin and zonaradiata have been performed on juvenile Rainbow trout using DBP and BBP [25].DBP did not elicit oocyte protein synthesis at a dose of DBP injected intra-peri-

288 C.A. Bradlee and P. Thomas

toneally (i.p.), and BBP induced vitellogenin production above control level inone out of six trout injected. Some doubt exists as to the reason for this effect reported on one fish as neither the age nor the sex of the animals is reported. Ati.p. doses of 5 and 50 mg/kg BBP increased the level of hepatic estrogen receptors(but this was not statistically significant) but significantly decreased the level ofzona radiata in plasma while E2 was found to increase both oocyte proteins.Moore [25] concludes that it is essential that the substance tested should be in theform in which it will be found in the organism considered. Phthalate diesters arehydrolyzed to the monoester form in fish gills, for example, and so the use of in-tra-peritoneal injection of the diester in these in-vivo studies is questionable. Thesame is true for the in-vitro studies which may not be relevant to the situation inthe environment.

Harries et al. [29] measured the reproductive performance of pair breedingfathead minnow (Pimephales promelas) over six weeks using BBP and nonylphe-nol (NP). Number of eggs spawned, number of spawnings, egg batch size,gonado-somatic index, vitellogenin synthesis, secondary sexual characteristics,male fatpad and tubercle formation were assessed prior to and after three weekexposure. The concentrations selected, up to 100 µg/L nominal (69 to 82 µg/Lmeasured), were lower than the aquatic solubility of both BBP and NP. Stock solutions of both substances were made up using methanol as a carrier solventto allow use of a flow-through testing apparatus. The authors stated that BBP hadno significant effect on fecundity although there was a significant decrease innumber of spawnings compared to the control there was also a significant in-crease of egg batch size per spawning. No discernible effects of BBP were ob-served on any of the other reproductive parameters, while NP was found to elicita response on fecundity, GSI and secondary sexual characteristics even at the low-est concentration used (between 1 and 10 µgNP/L measured). It should be notedthat, based on the published bar charts, the results from the solvent control moreclosely resembled results from BBP exposure than those of the negative control.

In conclusion, the studies have demonstrated equivocal findings from in-vitroand in-vivo studies on the endocrine modulating effects of DBP and BBP. In-sufficient information exists on other lower phthalates to draw conclusions aboutthe environmental relevance of assays suggesting estrogenic activity for the di-esters, DEP and DIBP.

3.4.2Higher Phthalates

For the higher phthalates, an in-vitro assay using DEHP did not result in recep-tor binding even at one order of magnitude above the water solubility of the di-ester and was negative in a reporter gene assay. DHP, DIDP and the monoesterMEHP were negative in yeast reporter gene assays while results for DINP were atbest equivocal.

In his review on in-vitro studies on the higher phthalate esters, Moore [25]stated that assays on receptor binding using trout hepatocytes did not result ina response for DEHP at 2 µM but reduced E2 binding at higher concentrationswith a maximum of 25% at 1 mM. However, the solubility limit of this substance

Aquatic Toxicity of Phthalate Esters 289

is 0.008 µM (3 µg/L) and positive results above this level should be treated withcaution, as non-selective binding is likely to occur.

Using yeast reported gene assays, Harris et al. [30] showed that DINP demon-strated an equivocal response; however, the results are confounded by the factthat the DINP used in the study may have contained a supplement of nonyl phe-nol. DHP, DEHP, DIDP and several monoesters, mono(2-ethylhexyl) phthalate(MEHP), and its oxidative metabolites, were all inactive between 500 nM and1 mM. A trout hepatocyte reporter gene assay using DEHP failed to elicit an es-trogenic response at concentrations up to 100 µM.

Metcalfe et al. [31] performed an in-vitro yeast estrogenicity reporter genescreening assay and also an in-vivo “90 day” Japanese medaka (Oryzias latipes)assay. In both cases, the organisms were exposed to 500, 1000 and 5000 µg/L con-centrations of DEHP. Other substances tested were estriol, 17b-estradiol, 17a-ethinylestradiol, NPEO, 4-NP and bisphenol A (all at doses considerably lowerthan for DEHP). Exposure was initiated 1 day after hatch in a static renewal sys-tem. Stock solutions were prepared by dissolving the tests substance in acetone(maximum 5 µl/L). Sixty fish were exposed to the tests substances in aqueous solution. For DEHP concentrations of 0, 500, 1000 and 5000 µg/L were employeddespite the fact that DEHP is only water soluble at concentrations of up to about3 µg/L. While it may be possible to generate emulsions of DEHP in laboratorytests that greatly exceed aqueous solubility, such an experimental design con-founds test interpretation because toxicity due to physical effects may be misin-terpreted as intrinsic toxicity of DEHP. Moreover, concentrations that far exceedaqueous solubility have little environmental relevance. Solutions were replacedevery 48-h over the duration of the assay. Study termination time was based onfish size (1.5 cm) and not on the duration so the length of the assay was between85 and 110 days post hatch.

The authors report that a background contamination level of DEHP was foundin all solutions in which acetone was used and they consider that this came fromthe tubing and joints of the test system which were made of PVC. All measuredconcentrations reported were equivalent to the nominal plus the background.Fish were embedded whole in paraffin and microtomed for light microscopicalexamination of the testes or ova. The authors examined two end points: changein expected sex ratio of the fish, and presence of testis-ova (at least one oogoniain the testes), hypothesized to signal the presence of estrogenic substances (al-though this could not be confirmed). For the yeast assay no responses were ob-served for DEHP or for certain NPEOs while NP and bisphenol A were found tobe positive at 5 orders of magnitude greater than E2. In the medaka in-vivo studythe results were generally consistent with those of the in-vitro assay. DEHP didnot induce a positive response and the authors conclude that this is in line withother in-vitro assays that show that DEHP is not an estrogen agonist.

Other studies examining endocrine effects following aquatic exposure tohigher phthalates have been performed on fish. Shioda and Wakabayashi [32] exposed adult male medaka (Oryzias latipes) to concentrations of DEHP (0.1,0.3 and 1 µM equivalent to 10 to 100 times the water solubility), nonylphenol (up to 0.3 µM) and bisphenol-A (up to 10 µM) for a period of two weeks. Afterthis time each male fish was placed together with two females and resulting

290 C.A. Bradlee and P. Thomas

spawning parameters were observed. The number of eggs per spawn and thenumber of eggs hatched was measured every day for a week. DEHP had no ad-verse effects on the number of eggs spawned or on hatch compared to the con-trol. Bisphenol-A significantly decreased the number of eggs and hatch from10 µM but for NP no significant differences could be observed because groupvariation was so high.

Patyna et al. [33] performed a study on Japanese medaka that shows neitherDIDP nor DINP elicit an effect on reproduction or development after three gen-erations of dietary exposure. Four treatment groups were studied that includedan untreated control, acetone control, 20 mg DIDP/kg food, and 20 mg DINP/kgfood. An acetone control was tested because each phthalate was spiked into thefood via an acetone solution. Each treatment group was divided into five repli-cate tanks of 50 fish each. The F0 generation was first exposed as 14-day old lar-vae by feeding either DIDP or DINP in dry flake food at a daily ration of 5% bodyweight. The F0 and F1 generations were fed each treatment through sexual mat-uration and oviposition (140 days-post-hatch). The test ended during the F2 gen-eration prior to sexual maturation and the F0 adults were terminated at day 123.The endpoints for the study were mortality, histopathology, growth, gonadal:so-matic index (GSI), sexual development, fecundity, embryonal development, andmicrosomal testosterone metabolism. The authors found that there were no sta-tistically significant changes in either fecundity or mortality in both the DIDPand DINP treatment groups. Normal embryonic development of the F1 genera-tion was observed except for a transient decrease in red blood cell pigmentationin the DIDP, DINP and acetone treatment groups. The presence of this effect inthe acetone groups suggests the transient pigmentation effects are not related tophthalate exposure. The only histopathologic change observed in the F0 adultswas a slight alteration in hepatocellular staining around the central vein. The sexratios and microsomal testosterone metabolism in DIDP and DINP treatedgroups were similar to the untreated control. Overall, the authors concluded thatneither of the phthalate esters elicited an effect on reproduction or developmentof Japanese medaka after three generations of dietary exposure.

Two in-vivo dietary studies on Atlantic salmon (Salmo salar) using DEHP havebeen performed. In 1999, Norrgren et al. [34] published a study that indicatedsome increase in the proportion of female fish at the highest dose tested. The sec-ond study, Norman et al. [35], was conducted by the same laboratory in 2001 asa direct repeat of the 1999 study but used an improved methodology. Improve-ment included analyzing the amount of DEHP in the feed and using a broaderrange of doses to better define the NOEC. The primary purpose of these studieswas to evaluate if DEHP exposure would alter the normal sex ratio of males to females, and although a 1 :1 (50%:50%) male:female may be considered the normal sex ratio for Atlantic salmon the authors provide no information on therange of normal variations to the 1:1 sex ratio within populations of Atlanticsalmon. Overall, the findings of the first study in 1999 that showed some higherproportion of females; however, these findings were not confirmed in the repeatstudy of 2001.

In the 1999, study exposure was initiated by providing one replicate per groupof 200 fish with food spiked with DEHP at nominal concentrations of 300 and

Aquatic Toxicity of Phthalate Esters 291

1500 mg DEHP/kg food four weeks after hatch (immediately after yolk-sac resorption), and continued for four weeks. The dietary concentration of DEHPwas not measured in this study and as such is considered to be a study designlimitation because the actual exposure doses of DEHP are not confirmed. Spikedconcentrations of 17b-Estradiol (E2) at 15 and 30 mgE2/kg and 4-n-nonylphenol(NP) at 300 and 1500 mg NP/kg were also run in this first test.

The fish were maintained for a further four months by providing them withuncontaminated food until maturity such that histological results on gonadal de-velopment could be obtained. Between 82 and 184 individuals per group weredissected to remove the liver and gonads and the sex was determined by light mi-croscopic evaluation. In an extension of this study, salmonids were injected in-tra-peritoneally with various concentrations of DEHP, NP and PCB two to fourtimes over a 17-day exposure period.

The authors determined the liver :somatic index (LSI) and found that for thetwo concentrations of E2 and the highest concentration of DEHP, (but not for NP)there was a significant increase in this parameter compared to the control. Thisfinding was contrary to the findings of Henderson and Sargent [24] who exposedtrout to a far greater concentration and for a longer time period, although basedon the weight (30–50 g) of the fish used in the experiment exposure was startedat a later life stage. The environmental significance of increased LSI for fish is un-clear but this is not an indicator of estrogenic activity.

E2 was found to increase female:male sex ratio at 15 and 30 mg E2/kg to 88 and100% females respectively. At the highest concentration of DEHP a statisticallysignificant ratio change was also noted with 64% of the population determinedhistologically as female. In the extension to the 1999 study, intra-peritoneally injected fish were analyzed for vitellogenin at the end of the exposure period. Theresults showed no statistically significant difference in vitellogenin levels betweenthe control and any of the tested substances.

Overall in the 1999 study [34] the authors concluded that for the endpoints of sex ratio, LSI, and vitellogenin the NOEC is considered as 300 mg DEHP/kg.Owing to the large difference between the NOEC and the LOEC and also to thefact that there was no true dose response relationship, Norrgren initiated a repeatthis study.

The 2001 repeat study focused on re-examining the results of the feeding studywith DEHP and its apparent effect on sex ratio; therefore the repeat study did notinclude E2 or NP exposures and interperitoneal injections to examine for vitel-logenin were not conducted. The repeat study utilized a similar dosing regime to the first study but used dietary concentrations of 400, 800 and 1500 mgDEHP/kg food. The concentration of DEHP in the food was measured at the be-ginning and end of the study and was found to have excellent agreement with thetargeted concentrations. The endpoints for the repeat study included survival,length and weight in addition to sex ratio and LSI index. The findings of the re-peat study did not concur with the first study with regards to effects on the LSIor sex ratio.

The first study showed a statistically significant (p<0.05) difference in the sexratio between the control and the 1500 mg DEHP/kg food dose level where therewere 49% and 64% females, respectively. A 50%:50% sex ratio of male:female is

292 C.A. Bradlee and P. Thomas

thought to be normal; however, the authors provide no information on the nor-mal range of variation within populations of Atlantic salmon. A second findingfrom the 1999 study showed a statistically significant (p<0.01) difference be-tween the control and the 1500 mg DEHP/kg food dose level for the LSI, wherethe LSI for the control and DEHP were 1.74 and 2.22, respectively. In the repeatstudy neither the sex ratio nor LSI were statistically different as compared to thecontrol. The number of females in the control and 1500 mg DEHP/kg food were50% and 51%, respectively; while the LSI for the control and the high DEHP dosewere 1.20 and 1.11, respectively. Moreover, no effects on survival or mean lengthand weight were seen at the high DEHP dose level (1500 mg DEHP/kg food).

Another finding of the 2001 study was a slight, reversible, increase in testis-ovo.An examination of the gonads from the 1500 mg/kg exposure concentrationshowed a small (3%), but statistically significant increase in the incidence oftestis-ovo. This occurred only in the highest concentration. This effect was onlytemporary as a second sampling of the exposed fish after a five month recoveryperiod where the fish were maintained in DEHP-free water showed no significantincidence of testis-ovo. There is no clear understanding of the significance of theobserved, reversible, testis-ovo; however, significant long-term adverse impactson populations Atlantic salmon is not expected.

In study by Freeman et al. [36] performed on Atlantic cod testes and head kid-neys in which the organs were incubated in-vitro with equimolar amounts of ra-diolabeled pregnenolone and progesterone with various concentrations of coldor radiolabeled DEHP. They examined the steroid metabolites and analyzed forDEHP metabolic breakdown products. The authors found no differences fromcontrols at any of the concentrations used, up to 1000 mg DEHP/kg tissue. More-over, perhaps unexpectedly, DEHP was not metabolized to MEHP or any otherdegradation product, echoing Moore’s [25] cautionary conclusion on in-vitrostudies when using hydrolyzable test substances.

The authors subsequently tested DEHP on steroid metabolism of Atlantic codin vivo in order to determine effects of potential metabolites. Fish were fed gelatine capsules containing a dietary concentration of 0, 10, 100 or 1000 mgDEHP/kg wet food, twice weekly, over 121 days. The testes, ovaries and corre-sponding head kidneys were prepared and assayed with radiolabeled preg-nenolone and progesterone prior to X-ray auto-radiographical analysis while theremaining fractions were isolated and identified by sequential thin layer and pa-per chromatography followed by scintillation spectrometry. Histological exam-inations were also carried out on certain tissues. No abnormalities were found inmale fish compared to the control in any of the DEHP dosed groups, both interms of the steroid metabolism and histological parameters examined.

In female cod the authors reported a significantly, and dose related, in-creased synthesis of one of the steroid intermediates (designated S) at 100 and1000 mg/kg wet weight, suggesting a change in biosynthetic pathway, but theywere unclear as to the significance of this change and no histological changes infemale fish were reported. No changes were found in male cod.

Van Wezel et al. [27] proposed Environmental Risk Limits (ERL) or MaximumPermissible Concentrations (MPC) for the higher phthalate, DEHP, coming to theconclusion that 0.19 µg/L would be a suitable ERL for this compound based on

Aquatic Toxicity of Phthalate Esters 293

ecotoxicology and environmental chemistry. The ERL (or MPC) value is based onthe original Larsson and Thurén [37] frog egg study using a NOEC of 10 mg/kgfresh weight and an uncertainty factor of 10 to provide a PNEC of 1 mg/kg f.w.The authors consider that, in standard sediment the organic carbon content at1 mg/kg f.w. would be 17.5 mg/kg o.c. (based on a Koc of 87¥103 l/kg o.c.) andthat this would be in equilibrium with 0.2 µg/L in the aquatic phase.

The authors were aware of a second study performed by Wennberg et al. [38],but discounted it due to the difference in study length (60 days for the former asopposed to 29 days for the latter). They also mention problems of counting andbacterial infection with the Wennberg et al. [38] study. They do not mention thatthe reason for the difference in study time was due to the low temperature em-ployed (5 °C in the first study as opposed to 10 °C in the second) which virtuallydoubled egg incubation time. Moreover, they do not mention the numerous prob-lems and unknown quantities (ethanol concentration, spiking method, controlmortality compared to the later study) encountered in the Larrson and Thurén[37] study. Finally, they did not take into account the second repeat frog egg studyperformed at both 5 and 10 °C, at two different organic carbon concentrationsand accepted by the EU commission as the definitive frog egg study on DEHP.Neither the Wennberg et al. [38] nor the IVL study found any effects at any con-centration used up to 450 mg/kg in the former and 1000 mg/kg in the latter study.The EU risk assessment states that no study on DEHP conclusively shows any effect in the aquatic compartment and so the ERL recommended by Van Wezelet al. [27] is not appropriate for DEHP.

6Discussion

The aquatic toxicity of phthalate esters varies widely across the class of com-pounds, with only the lower molecular weight esters (C1 to C4) consistently pos-ing toxicity. Phthalate esters higher than about C6 pose no toxicity up to theiraqueous solubility limits. The lack of toxicity for the higher phthalates is relatedto their relative insolubility in water and their ready metabolism by aquatic or-ganisms, so that the critical body burden for toxicity is not reached [2].

The toxicity of the lower molecular weight (C1 to C4) phthalate esters toaquatic organisms has been widely documented [1, 2]. These publications havefocused on studies that measured ecologically relevant endpoints related to sur-vival, growth and development, and reproductive fitness of populations ofaquatic organisms. Parkerton and Konkel [2] calculated predicted no effect con-centrations for DMP, DEP, DBP and BBP using the abundant acute and chronictoxicity data, acute to chronic ratios and statistical calculation procedures. ThePNECs calculated by the authors using three slightly different procedures were 3109 to 4780 µg/L for DMP, 865 to 1173 µg/L for DEP, 43 to 62 µg/L for DBPand 38 to 60 µg/L for BBP. The range of chronic NOEC values were 9600 to10,000 µg/L for DMP, 3650 to 25,000 µg/L for DEP, 40 to 10,000 µg/L for DBP and75 to 350 µg/L for BBP. Thus the PNEC values calculated by Parkerton and Konkel[2] are demonstrated to be amply protective of the aquatic environment as theyare below the range of available valid chronic data for these compounds.

294 C.A. Bradlee and P. Thomas

Recently, the issue of endocrine disruption in wildlife species has generatedconsiderable scientific, political, and public interest [39–41]. Endocrine-dis-rupting chemicals can, as a consequence of their molecular structure, bind tohormone receptors and may mimic or antagonize the action of the natural hor-monal ligand. Adverse impact should be judged on the basis of environmentallyrelevant endpoints that affect populations and communities of organisms, suchas survival, growth, and reproductive success. Moreover, studies that utilize in-ter-peritoneal injection (i.p.) as the route of administration do not accuratelyevaluate effects from phthalates esters because they do not account for metabo-lism. Inter-peritoneal injection may be a relevant route of exposure for substancesthat do not undergo metabolism before they enter the systemic circulatory sys-tem or those substances otherwise found systemically in the administered form.However, i.p. administration does not accurately represent the plasma metabo-lite profile for phthalates, and consequently the systemic response, resulting fromenvironmental exposure to phthalates, because it bypasses the normal “firstpass”metabolic pathways for the phthalates (esterase activity in the gill), nor does it offer a measure of the inherent activity of the diesters themselves, because theyare metabolized in the plasma and liver.

The evidence for endocrine modulation effects caused by phthalate esters isequivocal at best. Studies from a few in-vitro and in-vivo assays that were de-signed to identify endocrine modulating compounds have suggested that DBPand BBP, but no other phthalate ester, were capable of interacting with estrogenicreceptors. In addition, another study suggested that BBP was also anti-andro-genic [42]. Testis-ova in male gonads and induction of hepatic vitellogenin haveoccasionally been reported for some phthalate esters at very high doses. Thesefindings occurred at concentrations already considered to be toxic based on con-ventional studies, were not supported upon repetition of the original study orwere reported in studies that employed injection of test material into individualfish. The abundant numbers of high quality studies testing a range of phthalateesters that have examined ecologically relevant endpoints have consistently reported some toxicity with the lower esters and no toxicity in the higher esters.It is using these studies measuring ecologically relevant endpoints that should beused for risk assessment purposes and for developing appropriate water qualitycriteria.

7References

1. Staples CA,Adams WJ, Parkerton TF, Gorsuch JW, Biddinger GR, Reinert KH (1997) Aquatictoxicity of eighteen phthalates esters. Environ Toxicol Chem 15:875–981

2. Parkerton TF, Konkel WJ (2000) Application of quantitative structure – activity relation-ships for assessing the aquatic toxicity of phthalate esters. Ecotoxicol Environ Saf 45:61–78

3. Adams WJ, Heidolph BB (1985) Short-cut chronic toxicity estimates using Daphnia magna.Aquatic Toxicology and Hazard Assessment: Seventh Symposium,ASTM STP 854, CardwellRD, Purdy R, Bahner RC (eds) American Society for Testing and Materials, Philadelphia, PA,pp 87–103

4. Passino DRM, Smith SB (1987) Acute bioassays and hazard evaluation of representativecontaminants detected in Great Lakes fish. Environ Toxicol Chem 6(11) :901–907

Aquatic Toxicity of Phthalate Esters 295

5. Adams WJ, Biddinger GR, Robillard KA, Gorsuch JW (1995) A summary of the acute tox-icity of 14 phthalate esters to representative aquatic organisms. Environ Toxicol Chem14:1569–1574

6. Springborn Bionomics, Inc (1984) Acute toxicity of fourteen phthalate esters to (Daphniamagna). Chemical Manufacturers Association, Washington, DC

7. Springborn Bionomics, Inc (1984) Chronic toxicity of fourteen phthalate esters to Daph-nia magna. Chemical Manufacturers Association, Washington, DC

8. Rhodes J, Adams WJ, Biddinger GR, Robillard KA, Gorsuch JW (1995) Chronic toxicity of14 phthalate esters to Daphnia magna and rainbow trout (Oncorhynchus mykiss). EnvironToxicol Chem 14:1967–1976

9. McCarthy JF,Whitmore DK (1985) Chronic toxicity of di-n-butyl and di-n-octyl phthalateto Daphnia magna and the fathead minnow. Environ Toxicol Chem 4:167–179

10. Scholz N (1995) Determination of the effect of Vestinol AH (DEHP) on the swimming behavior of Daphnia magna. Complies with Directive 92/69/EEC. Final Report DK-631,Marl, Germany

11. Brown D, Thompson RS (1982) Phthalates and the aquatic environment: Part I. The effectof di-2-ethylhexyl phthalate (DEHP) and di-isodecyl phthalate (DIDP) on the reproduc-tion of Daphnia magna and observations on their bioconcentration. Chemosphere 11:417–426

12. Brown D, Croudace CP, Williams JJ, Shearing JM, Johnson PA (1998) The effect ofphthalate ester plasticisers tested as surfactant stabilised dispersions on the reproductionof the Daphnia magna. Chemosphere 36:1367–1379

13. Versteeg D.J (1990) Comparison of short- and long-term toxicity test results for the greenalga, Selenastrum capricornutum. Plants for Toxicity Assessment. ASTM STP 1091, WangW, Gorsuch JW, Lower WR (eds) American Society for Testing and Materials, Philadelphia,PA, pp 40–48

14. Huels AG (1991) Untersuchung über den Einfluss von Di-n-butylphthalat auf Scenedesmussubspicatus. Huels AG, unveroffentliche (12.03.91), Marl, Germany

15. Scholz N (1995) Determination of the effect of Vestinol C (DBP) on the growth ofScenedesmus subspicatus 86.81. SAG. Complies with Directive 92/69/EEC. Final Report AW-392, Marl, Germany

16. Gledhill WE, Kaley RG,Adams WJ, Hicks O, Michael PR, Saeger VW, LeBlanc GA (1980) Anenvironmental safety assessment of butyl benzyl phthalate. Environ Sci Tech 14:301–305

17. Scholz, N (1995) Determination of the effects of Vestinol AH (DEHP) on the reproductionrate of Daphnia magna. OECD Guideline 202 Part II. Final Report DL-160, Marl, Germany

18. Laughlin RB Jr, Neff JM, Hrung YC, Goodwin TC, Giam CS (1978) Effects of three phthalate esters on the larval development of the grass shrimp Palaemonetes pugio(Holthuis). Water, Air, Soil Pollut 9 :323–336

19. Tagatz ME, Plaia GR, Deans CH (1986) Toxicity of dibutyl phthalate-contaminated sedimentto laboratory and field-colonized estuarine benthic communities. Bull Environ ContamToxicol 37:141–150

20. Hobson JF, Carter DE, Lighter DV (1984) Toxicity of a phthalate ester in the diet of a pe-naied shrimp. J Toxicol Environ Health 13:959–968

21. DeFoe DL, Holcombe GW, Hammermeister DE, Biesinger KE (1990) Solubility and toxic-ity of eight phthalate esters to four aquatic organisms. Environ Toxicol Chem 9:623–636

22. van den Dikkenberg RP, Canton HH, Mathijssen-Spiekman LAM, Roghair CJ (1989) Usefulness of Gasterosteus aculeatus – the three-spined stickleback – as a test organism inroutine toxicity tests. Rijksinstitut voor de Volksgezondheid en Milieuhygiene, Bilthoven,Netherlands, pp 28

23. Mayer FL, Mehrle PM, Schoettger RA (1977) Collagen metabolism in fish exposed to organic chemicals. In: Taub RA (ed) Recent Advances in Fish Toxicology (ed) EPA600/3–77–085, U.S. Environmental Protection Agency, Corvallis, OR, pp 31–54

24. Henderson RJ, Sargent JR (1983) Studies of the effects of Di-2-ehthylexyl Phthalate on LipidMetabolism in Rainbow Trout Fed Zooplankton Rich Wax Esters. Comp Biochem Physiol74C:325–330 (1983)

296 C.A. Bradlee and P. Thomas

25. Moore NP (2000) The oestrogenic potential of the phthalate esters. Repro Toxicol14:183–192

26. Ohtani H, Miura I, Ichikawa Y (2000) Effects of Dibutyl Phthalate as an environmental endocrine disruptor on gonadal sex differentiation of genetic males of the frog Rana rugosa. Environ Health Perspec 108:1189–1193

27. van Wezel AP, van Vlaardingen P, Posthumus R, Crommentuijn GH, Sijm DT (2000) Envi-ronmental risk limits for two phthalates, with special emphasis on endocrine disruptiveproperties. Ecotoxicol Environ Saf 46:305–321

28. Patyna PJ, Thomas PE, Cooper KR (1999) Multigeneration reproductive effects of di-n-butyl phthalate in Japanese medaka (Oryzias latipes). Toxicologist. 48(1-S) :262

29. Harries JE, Runnalls T, Hill E, Harris C, Maddix S, Sumpter JP, Tyler CR (2000) Develop-ment of a reproductive performance test for endocrine disrupting chemicals using pair-breeding fathead minnow (Pimephales promelas). Environ Sci Technol 34:3003–3011

30. Harris CA, Henttu P, Parker MG, Sumpter JP (1997) The estrogenic activity of phthalate esters in vitro. Environ Health Perspect 105:802–811

31. Metcalfe CD, Metcalfe TL, Kiparissis Y, Koenig BG, Khan C, Hughes RJ, Croley TR, MarchRE, Potter T (2001) Estrogenic potency of chemicals detected in sewage treatment plant effluents as determined by in-vivo assays with Japanese medaka (Oryzias latipes). EnvironToxicol Chem 20:297–308

32. Shioda T,Wakabayashi M (2000) Effect of certain chemicals on the reproduction of medaka(Oryzias latipes). Chemosphere. 40:239–243

33. Patyna PJ, Parkerton TF, Davi RA, Thomas PE, Cooper KR (1998) Evaluation of two phthalate ester mixtures in a three generation reproduction study using Japanese medaka(Oryzias latipes). Toxicologist (1998 42(1-S) :338–339

34. Norrgren L, Blom A, Andersson PL, Borjeson H, Larsson DGJ, Olsson PE (1999) Effects ofpotential xenoestrogens (DEHP, nonylphenol and PCB) on Sexual Differentiation in Juve-nile Atlantic Salmon (Salmo salar). Aquatic Ecosys Health Manag 2 :311–317

35. Norman L, Norrgren L, Borjeson H (2001) Exposure of Atlantic salmon (Salmo salar) toDEHP-contaminated food. Manuscript in Preparation

36. Freeman HC, Sangalang GB, Burns BG, McMenemy M (1980) The Effects of Di-(2-ethyl-hexyl) phthalate (DEHP) on Steroid Metabolism in the Atlantic Cod Gadus morhua.Proceedings of the 7th Annual Aquatic Toxicity Workshop, Montreal, Canada 11/5–7/1980,pp 198–226

37. Larsson P, Thuren A (1987) Di-2-ethylhexylphthalate inhibits the hatching of frog eggs andis bioaccumulated by tadpoles. Environ Toxicol Chem 6:417–422

38. Wennberg L, Parkman H, Remberger M,Viktor T,Williams C (1997) Influence of sediment-associated phthalate esters (DEHP and DIDP) on hatching and survival of the moorfrog,Rana arvalis. Govt Reports Announcements & Index (GRA&I), Issue 19

39. Coburn T, vom Saal FS, Soto AM (1993) Developmental effects of endocrine-disruptingchemicals in wildlife and humans. Environ Health Perspect 101:378–84

40. Kendall RJ, Dickerson RL, Giesy JP et al. (eds) (1998) Principles and Processes for Evalu-ating Endocrine Disruption in Wildlife. SETAC Press, Pensacola, FL, USA

41. Tattersfield L, Matthiessen P, Campbell P et al. (eds) (1997) SETAC – Europe/OECD/EC Expert Workshop on Endocrine Modulators and Wildlife: Assessment and Testing,WMWAT; Veldhoven, The Netherlands, 10–13 April 1997. Brussels: Society of Environ-mental Toxicology and Chemistry-Europe

42. Sohoni P, Sumpter JP (1998) Several environmental oestrogens are also anti-androgens.J Endrocrinol 158:327–339

43. Springborn Bionomics, Inc (1984) Toxicity of fourteen phthalate esters to the freshwatergreen algae (Selenastrum capricornutum). Chemical Manufacturers Association, Wash-ington, DC

44. Suggatt RH, Foote K (1981) Comprehensive review of acute aquatic toxicity data on phthalate esters. Final Report, Contract SRC TR 81–537. Syracuse Research Corp., April1981

Aquatic Toxicity of Phthalate Esters 297

45. Kuhn R, Pattard M (1990) Results of the harmful effects of water pollutants to green algae(Scenedesmus subspicatus) in the cell multiplication inihibition test. Water Res 24:31–38

46. Scholz N (1995) Determination of the effect of Vestinol AH (DEHP) on the growth ofScenedesmus subspicatus 86.81. SAG. Complies with Directive 92/69/EEC. Final Report AW-391, Marl, Germany

47. Wilson WB, Giam CS, Goodwin TE, Aldrich A, Carpenter V, Hrung YC (1978) The toxicityof phthalates to the marine dinoflagellate Gymnodinium breve. Bull Environ Contam Tox-icol 20:149–154

48. Springborn Bionomics, Inc (1983) Acute toxicity of thirteen phthalate esters to fatheadminnow (Pimephales promelas) under flow-through conditions. Chemical ManufacturersAssociation, Washington, DC

49. Springborn Bionomics, Inc (1983) Acute toxicity of fourteen phthalate esters to fatheadminnows (Pimephales promelas). Chemical Manufacturers Association, Washington, DC

50. Geiger DL, Northcott CE, Call DJ, Brooke LT (1985.Acute toxicities of organic chemicals tofathead minnows (Pimephales promelas) vol 2. Center for Lake Superior EnvironmentalStudies, University of Wisconsin, Superior, WI, pp 326

51. Tagatz ME, Deans CH, Moore JC, Plaia GR (1983) Alterations in composition of field- andlaboratory-developed estuarine benthic communities exposed to di-n-butyl phthalate.Aquat Toxicol 3 :239–248

52. Springborn Bionomics, Inc (1986) Chronic toxicity of butylbenzyl phthalate to mysidshrimp (Mysidophsis bahia). Final report, Springborn Bionomics, Inc. Monsanto Company,St. Louis, MO

53. Knowles CO, McKee MJ, Palawski DU (1987) Chronic effects of di-2-ethylhexyl phthalateon biochemical composition, survival and reproduction of Daphnia magna. Environ Tox-icol Chem 63:201–208

54. Brown D, Thompson RS (1982) Phthalates and the aquatic environment: Part II. The bioconcentration and depuration of di-2-ethylhexyl phthalate (DEHP) and di-isodecyl phthalate (DIDP) in mussels (Mytilus edulis). Chemosphere. 11(4) :427–435

298 C.A. Bradlee and P. Thomas: Aquatic Toxicity of Phthalate Esters

© Springer-Verlag Berlin Heidelberg 2003

Introduction

Marian K. Stanley 1 · Kenneth A. Robillard 2 · Charles A. Staples 3

1 American Chemistry Council, 1300 Wilson Blvd., Arlington, VA 22209, USA E-mail: Marian_Stanley@americanchemistry.com

2 Eastman Chemical, Rochester, NY, USA3 Assessment Technologies, Inc. 10201 Lee Highway, Suite 580, Fairfax, VA 22030, USA

The esters of 1,2-benzene dicarboxylic acid, commonly called phthalate esters, are a diversegroup of compounds that have broad use in a wide array of industrial applications. Regulatoryoversight of the manufacture, transport, use and disposal of phthalate esters has produced alarge amount of data regarding the properties, environmental fate, exposure, and toxicity ofthese compounds. Such data are critical for the development of safe and accepted productionpractices, effluent discharge limits, and human exposure limits. The following chapters present,in detail, the information that has been collected regarding these properties.

Keywords. Phthalate, Manufacture, Use, Releases, Regulation

1 General Description . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 Manufacture of Phthalate Esters . . . . . . . . . . . . . . . . . . . 2

3 Nomenclature and Physical/Chemical Properties . . . . . . . . . 3

4 Chemical Interactions with Vinyl . . . . . . . . . . . . . . . . . . 3

5 Uses of Phthalate Esters . . . . . . . . . . . . . . . . . . . . . . . 3

6 Disposal and Releases into the Environment . . . . . . . . . . . . 5

7 Regulations and Phthalate Esters . . . . . . . . . . . . . . . . . . 6

8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

1General Description

Phthalate esters are widely used industrial chemicals. Higher molecular weightphthalate esters act as an additive which imparts flexibility in vinyl resins; this is the highest volume use of phthalates. Both linear and branched phthalate esters are used in the manufacture of vinyl articles. The linear esters provide superior low temperature properties to the finished vinyl products and also

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 1– 7DOI 10.1007/b11460

have lower volatility. The C8–C13 phthalate esters are the dominant vinyl plas-ticizers with di-2-ethylhexyl phthalate, diisononyl phthalate predominant and diisodecyl phthlatate. The lower molecular weight phthalates are used as plasticizers in some non-vinyl resins, including acrylics, urethanes and cellu-losics. The various esters used in commerce have alkyl side chains containingfrom 1 to 13 carbon atoms. Table 1 contains a listing of the most common phthalate esters. Most of the high molecular weight phthalates esters are used in the manufacture of a wide variety vinyl goods, both commercial and con-sumer. The lower molecular weight phthalates, those with alkyl side chains from 1 to 4 carbon atoms, have a very broad use which includes consumer prod-ucts and pharmaceuticals. This will be detailed in the use section.As plasticizers,phthalates are additives which improve the flexibility, processability and soft-ness of vinyl. Phthalates with alkyl side chains lower than C6 are not often usedalone as a plasticizer because of volatility concerns. In general, the factors thatdictate selection of a phthalate or combination of phthalates for a particular application are functionality and economics of use [1]. Overall, phthalate estersare used because they combine qualities such as compatibility, permanence,efficiency and processability at reasonable cost. Compatibility problems with the vinyl resin preclude the use of phthalates esters of molecular weight higherthan ditridecyl. In vinyl, dibutyl phthalate is only used in isolated cases in conjunction with higher molecular weight plasticizers to reduce volatility.Dipropyl and dipentyl phthalate, C3 and C5, are not available commercially in theUnited States.

2Manufacture of Phthalate Esters

The ortho phthalate esters are generally manufactured by the sequential additionof either branched or normal alcohols to phthalic anhydride in the presence ofan acid catalyst. The alcohol manufacturing processes are stable, so although thephthalates produced from branched alcohols are complex substances, they arenot variable.

Phthalate esters are products of simple esterification reactions, which can becarried out readily in heated kettles with agitation and provision for water re-moval.While some plants produce phthalates by the batch method, newer, highlyautomated plants operate continuously, particularly if they emphasize a singleproduct. The purity requirements for commercial plasticizers are very high andphthalate diesters are usually colorless and mostly odorless. The reaction usuallyrequires an excess of alcohol, which is readily recycled.

Diisononyl phthalate (DINP) is a complex substance assigned two differentCAS numbers. CAS number 68515-48-0 is manufactured from polygas branchedolefin that is converted to alcohol moieties consisting mainly of 3,4-, 4,6-, 3,6-,3,5-, 4,5- and 5,6-dimethyl-1-heptanol. The CAS number 28553-12-0 is producedfrom dimerized n-butene that is converted primarily to methyl octanols and di-methyl heptanols. This CAS number also represents DINP which is producedfrom n-butene and isobutene that are converted to alcohols, with 60% consistingof methylethyl hexanols. The two types of DINP are considered commercially

2 M.K. Stanley et al.

interchangeable. Other phthalates that are complex mixtures are diisodecyl phthalate (DIDP) and the linear phthalates D610P and D711P.

3Nomenclature and Physical/Chemical Properties

The physical and chemical properties of phthalate esters are well documentedand will be discussed in detail in this handbook. A summary of some of theirphysical properties is contained in Table 1.

The phthalate esters discussed here are liquids at room temperature. The di-esters derived from the lower molecular weight alcohols such as DMP and DEPare colorless fluids of low viscosity, but phthalate esters become more viscous andoily as the size of the alkyl side-chain increases. They have low freezing points,many well below 0 °C (see Table 1).

Generally, the water solubility of the alkyl phthalate ester varies inversely withthe length of the alkyl side chain. DMP is the most hydrophilic and water solu-ble of the esters. The C10, C11 and C13 esters are the most hydrophobic and leastwater soluble (< 0.001 mg/l). Most of the dialkyl phthalates are soluble in com-mon organic solvents such as benzene, toluene, xylene, diethyl ether, chloroformand petroleum ether [2].

In many cases alternate names are used in the literature and in commerce forthe common phthalate esters. Table 1, while not intending to be exhaustive, liststhese synonyms. Note the DOP (“dioctyl phthalate”) is used as a synonym forDEHP (di(2-ethylhexyl)phthlate).

4Chemical Interactions with Vinyl

Incorporating phthalate esters into a polymeric matrix reduces the glass transi-tion temperature of the polymer [3]. Phthalate esters are not bound to the poly-mer with covalent chemical bonds and are therefore able to migrate to the sur-face of the polymer matrix where they may be lost by a variety of physicalprocesses. Nevertheless, various chemical-physical attractive forces hold the phthalate ester tightly within the vinyl matrix, so that such migration occurs ata very low rate.

Retention in the polymer matrix is one of the main factors in consideringwhich phthalate ester to use. The ester must be sufficiently nonvolatile to remainin the compound during its mixing and formation stages

5Uses of Phthalate Esters

Uses of phthalate esters can be broadly split into three general categories – vinylplasticizers, plasticizers for non-PVC polymers and other minor specialized applications. In the United States, DEHP, DINP and DIDP account for 52.2% ofphthalates consumed. Linear phthalates account for 21.4% of the consumptionand the 26.4% balance of US consumption includes all other phthalates esters [4].

Introduction 3

4 M.K. Stanley et al.Ta

ble

1.Ph

ysic

al p

rope

rtie

s of

phth

alat

e es

ters

Abb

revi

atio

nPh

thal

ate

Form

ula

CA

S N

o.EI

NEC

S N

o.M

olec

ular

M

elti

ng P

oint

Sp

ecifi

c G

ravi

ty

Este

rW

eigh

t(°

C)

(20

°C)

DM

PD

imet

hyl P

htha

late

C10

H10

O4

131-

11-3

205-

011-

619

4.2

5.5

1.19

2D

EPD

ieth

yl P

htha

late

C12

H14

O4

84-6

6-2

201-

550-

622

2.2

–40

1.11

8D

nBP

Di-

n-Bu

tyl P

htha

late

C16

H22

O4

84-7

4-2

201-

557-

427

8.4

–35

1.04

2D

IBP

Diis

obut

yl P

htha

late

C16

H22

O4

84-6

9-5

201-

553-

227

8.4

–58

1.05

0BB

PBu

tylb

enzy

l Pht

hala

teC

19H

20O

485

-68-

720

1-62

2-7

312.

4–

351.

111

DH

PD

ihex

yl P

htha

late

C20

H30

O4

84-7

5-3

201-

559-

533

4.4

–27

.41.

011

6851

5-50

-427

1-09

3-5

DIH

PD

iisoh

epty

l Pht

hala

teC

22H

34O

471

88-8

9-6

276-

15-8

363

–45

1.00

6815

-44-

6D

nOP

Di-

n-O

ctyl

Pht

hala

teC

24H

38O

411

7-84

-020

4-21

4-7

390.

6–

250.

978

D61

0PD

i (n-

Hex

yl,n

-Oct

yl,

C25

H40

O4

2572

4-58

-724

7-21

0-0

404.

6 {3

34–

447}

–4

0.97

n-D

ecyl

) Ph

thal

ate

6851

5-51

-527

1-09

1-4

DEH

PD

i(2-

Ethy

lhex

yl)

C24

H38

O4

117-

81-7

204-

211-

039

0.6

–47

0.98

6Ph

thal

ate

DIN

PD

iison

onyl

Pht

hala

teC

26H

42O

428

553-

12-0

249-

079-

541

8.6

{418

.6–

432.

6}–

480.

9768

515-

48-0

271-

090-

9D

IDP

Diis

odec

yl P

htha

late

C28

H46

O4

2676

1-40

-024

7-97

7-1

446.

7 {4

32.7

–44

6.7}

–46

0.96

168

515-

49-1

271-

091-

4D

711P

Di(

Hep

tyl,

Non

yl,

C26

H42

O4

3648

-20-

222

2-80

4-9

418.

6 {3

62.6

–47

4.7}

<–

500.

97U

ndec

yl) P

htha

late

6851

5-44

-627

1-08

6-7

6851

5-45

-727

1-08

7-2

1113

81-8

9-6

–11

1381

-90-

9–

1113

81-9

1-0

–D

UP

Diu

ndec

yl P

htha

late

C30

H50

O4

3648

-20-

222

2-88

4-9

447.

7 {4

32.7

–47

4.7}

–9

0.96

DT

DP

Dit

ride

cyl P

htha

late

C34

H58

O4

119-

06-2

204-

294-

353

0.8

{506

.8–

544.

8}–

370.

953

6851

5-47

-927

1-08

9-3

The principle phthalate plasticizer application in conjunction with polymersis, by far, the plasticization of vinyl resins. Plasticized vinyl is unique because a wide range of resin/plasticizer ratios permits production of plastics that range from very hard to very flexible. The phthalate plasticizer is usually chosento produce the desired performance characteristics with the greatest possibleeconomy of material costs and the greatest ease and speed of processing [1].High molecular weight phthalates, C6 and above, are generally used as vinyl plasticizers.

Low molecular weight phthalates have a diverse set of uses. The largest use ofDMP is as a stabilizing diluent for the shipping and storage of organic peroxides.The main use of DEP is in the compounding of cellulosic films. Cosmetic-gradeDEP acts as a fixative or carrier for perfumes and fragrances. There is also a phar-maceutical grade used in time-released preparations. DBP is consumed primar-ily in vinyl acetate emulsion adhesives and in cellulose lacquers. BBP is normallyused with other general-purpose plasticizers in PVC applications to improve theperformance of the final product and to aid in its processing.

6Disposal and Releases into the Environment

The releases of phthalates to the environment are discussed in the chapter onMultimedia Modeling. Very little phthalate ester is released to the environmentduring the manufacturing process. Very little is released into the air. Essentiallyall of the phthalate released during production and processing is disposed of inwastewater that is treated in wastewater treatment plants where it is either biode-graded or adsorbed to sludge with very little going to air. The fate of phthalatesduring wastewater treatment is covered in more detail in the chapter on Envi-ronmental Degradation. Phthalate adsorbed to sewage sludge is usually either in-cinerated or landfilled in the United States. In some countries, application ofsewage sludge to agricultural land is also common. The latter two disposal routeswill result in some release of phthalates to soil. Degradation and fate of phthalatesin soil is also treated in subsequent chapters. Poorly operating incinerators mayalso lead to some phthalate emission to air.

The major portion of phthalate esters that are found in the environment arethe result of the slow releases of phthalates from plastics and other phthalate con-taining articles due to weathering. As stated earlier, phthalates within such arti-cles are not covalently bound and under conditions of high surface exposure andwarm temperatures, as in the case of exterior building materials, phthalate esterscan diffuse from the solid surface into the air, despite their rather low vapor pres-sures. Thus, phthalate-containing consumer items during their useful lifetimemay continue to be a source of phthalate esters to the atmosphere. Burial ofphthalate ester containing articles in landfills will preclude further emissions tothe air.As discussed in the Degradation chapter, phthalate esters contained withinplastics buried in soil are degraded at the surface of the plastic by molds and bac-teria. Phthalate diesters themselves show poor mobility in soil but aqueousleachates from landfills may contain trace amounts of more soluble products ofphthalate degradation. The overall concentration, distribution and disposition of

Introduction 5

typical phthalate esters in the environment are topics covered in the MultimediaModeling chapter.

7Regulations and Phthalate Esters

Because of their very large production volumes, phthalate esters are subject toconsiderable regulatory scrutiny world-wide. Regulations on phthalate esterscover all aspects of their production, transportation, use, and disposal. Phthalatesare regulated under the Clean Water Act, so that at certain manufacturing facil-ities in the US, wastewater to be treated in municipal sewage treatment plantsmay be required to undergo pretreatment prior to leaving the facility (Pretreat-ment Standards).When they become waste products, certain phthalates are sub-ject to Resource Conservation and Recovery Act (RCRA) requirements. Drinkingwater standards have been set under the Safe Drinking Water Act for severalDEHP. Releases to the environment of several phthalate esters are required to bepublicly reported in the US, Canada, and Japan.

Like all commercial chemicals, phthalate esters are subject to regulation un-der the U.S. Toxic Substance Control Act. Pursuant to an enforceable consentagreement under that Act, the industry has generated considerable data on phthalate physical-chemical properties, environmental fate characteristics thatgovern distribution and concentration of phthalate esters, and toxicity studies.In addition, the industry has voluntarily conducted many additional such studies.

Phthalate esters have undergone comprehensive risk assessments regardingvirtually all aspects of environmental and human health under “existing sub-stances” regulations in the US, Canada, the EU, and at the Organization for Eco-nomic Cooperation and Development (OECD) level. These risk assessments haveused the extensive data that has been generated for phthalate esters. The resultsof the various risk assessments completed to date have led to varying conclusionsranging from no further information needed and/or no need for further restric-tions on use, to proposed requirements for some use-specific risk reduction measures. Some of the recent reviews include cancer classifications for DEHP and BBP by the International Agency for Research on Cancer, reviews of sevenphthalate esters by an Expert Panel of the U.S. National Toxicology Program Center for the Evaluation for Risk to Human Reproduction, and review of DINPin children’s toys by the U.S. Consumer Product Safety Commission Chronic Hazard Advisory Panel on Diisononyl Phthalate. The European Union is nearingcompletion of comprehensive risk assessments for DBP, BBP, DEHP, DINP andDIDP.

The commercial and regulatory interest in the ortho phthalate esters has re-sulted in an extensive list of health and environmental information on these substances. The remaining chapters in this review organize and present this information. Physical-chemical data are provided for all of the commercially important phthalate esters. The fate and distribution of phthalate esters in the environment, particularly in aquatic systems, is described. The effects ofphthalate esters on plants and aquatic animals are discussed as well as its accu-

6 M.K. Stanley et al.

Introduction 7

mulation in an aquatic food web. Finally, human exposure to phthalate esters andtheir effects on human health are reviewed. All of this information provides anexcellent data set for the continued use of these chemicals in a safe and respon-sible manner.

8References

1. Use Category Document – Plastics Additive; Building Research Establishment Ltd., June1998, pp 44–45

2. K.N. Woodward (ed) (1988) Phthalate Esters: Toxicity and Metabolism, vol I; CRC Press3. Use Category Document – Plastics Additive; Building Research Establishment Ltd., June

1998, pp 44–454. Chemical Economics Handbook – SRI International, January 2000

© Springer-Verlag Berlin Heidelberg 2003

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters

Raymond M. David 1 · Gerhard Gans 2

1 Health and Environment Laboratories, Eastman Kodak Company, 1100 Ridgeway Avenue,Rochester, NY 14652-6272, USA. E-mail: raymond.david@kodak.com

2 BASF Corporation, Washington, DC, 20005, USA

Phthalate esters can be divided into three general categories based on size. These categories arelow-MW esters used as solvents in cosmetics or as plasticizers of cellulose acetate polymers,mid-MW esters used as solvents in some PVC products but only in combination with otherplasticizers in floor coverings, or as a solvent or plasticizer in the cosmetic and pharmaceuti-cal industries, and high-MW esters used as plasticizers of PVC for wire and cable coverings,medical products, and other consumer products. Phthalate esters are data-rich for effects in lab-oratory animals, having been the focus of much research because of their effects on the bio-chemistry of liver cells, the effects on the testes, and the effects on the development of labora-tory animals. All phthalate esters have little or no toxicity following single (acute) exposures.These substances are not dermal sensitizers, but may produce minor skin irritation with pro-longed exposure to the neat chemical. Long-term hazards from short-term exposures are ei-ther minimal or reversible because many long-term effects are observed only following con-tinuous exposure. Long-term effects such as liver cancer only occur in laboratory animalsfollowing life-time or near life-time exposure to doses of >100 mg kg–1 d–1 of high-MW esters.Cancer is thought to occur through a mechanism that involves biochemical changes in the livercells of rats and mice. These biochemical changes are not seen in primates. As a result, scien-tists do not regard humans to be at risk of cancer from exposure to phthalate esters. Repro-ductive toxicity or developmental effects in the offspring of laboratory animals exposed to mid-MW phthalates during gestation have been reported. Reproductive toxicity is the result ofdamage to the testes causing lack of sperm production. It is not known if such effects occur inhumans, but adult primates have been resistant to the effects seen in adult rodents. Develop-mental effects can also occur with high exposure to mid-MW esters. The effects include in-complete skeletal formation in the head, spinal cord, tail, and ribs. In addition, male ratsdemonstrate incomplete formation of the urogenital tract. It is not known if primates or hu-mans are also susceptible to these effects, and the mechanism in laboratory animals is un-known. Exposure to phthalate esters is not thought to cause respiratory diseases such as asthmabecause these substances are hypoallergenic, but some have tried to associate exposure withan increased sensitivity to respiratory allergens. Sufficient data are lacking to make such a cor-relation. Phthalate esters are not neurotoxic. In evaluating the concern for humans, many prin-ciples of toxicology are discussed to provide the reader with sufficient understanding of speciesextrapolation. Primates are not as sensitive to phthalate esters as are rodents. There may be avariety of reasons for this lack of sensitivity, for example, lower absorption and different meta-bolic pathways. There are also intrinsic differences in the responses of primate and human cellsto the biochemical effects of phthalate esters. While this difference relates directly to the like-lihood of cancer, it may also impact the sensitivity to other effects seen in animals. Thus, pre-dicting the effects in humans must convey some level of uncertainty.

Keywords. Phthalate esters, Toxicity

The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 299–316DOI 10.1007/b11470

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

1.1 Assumptions Used in Toxicology . . . . . . . . . . . . . . . . . . 3011.2 Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3021.3 Uncertainties in the Assessment of Risk . . . . . . . . . . . . . . . 3031.4 Uses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305

2 Short-Term Exposure . . . . . . . . . . . . . . . . . . . . . . . . 306

2.1 Oral/Dermal/Inhalation . . . . . . . . . . . . . . . . . . . . . . . 3062.2 Irritation/Sensitization . . . . . . . . . . . . . . . . . . . . . . . . 3062.3 Long-Term Risks from Short-Term Exposure . . . . . . . . . . . . 3072.3.1 Metabolic Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . 3072.3.2 Regenerative Effects . . . . . . . . . . . . . . . . . . . . . . . . . 3082.3.3 Developmental Effects and Cancer . . . . . . . . . . . . . . . . . . 309

3 Prolonged Exposure . . . . . . . . . . . . . . . . . . . . . . . . . 309

3.1 Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3093.2 Reproductive/Developmental Toxicity and Endocrine-Related

Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3113.3 Asthma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3133.4 Neurotoxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314

Abbreviations

NTP National Toxicology ProgramDEHP di(2-ethylhexyl) phthalateDOP dioctyl phthalateDnOP di-n-octyl phthalateDINP diisononyl phthalateDMP dimethyl phthalateDEP diethyl phthalateBBP butylbenzyl phthalateDIDP diisodecyl phthalateDiBP diisobutyl phthalateDUP diundecyl phthalateMW molecular weightPVC poly(vinyl chloride)CDC Centers for Disease Control and PreventionTDI tolerable daily intakeLOEL lowest observed effect levelNOEL no-observable-effect levelBUA Beratergremium für Umweltrelevante Altstoffe (advisor committee for

environmental relevant old materials)DNA deoxyribonucleic acid

300 R.M. David and G. Gans

IARC International Agency for Research on CancerPPAR peroxisome proliferator activated receptorRfD oral reference doseATSDR Agency for Toxic Substances and Disease Registry

1Introduction

1.1Assumptions Used in Toxicology

Toxicology is not an exact science. It deals with probability or the likelihood thatsome event or biological response will occur.A toxicologist predicts biological re-sponses in one species, or in one individual, based on the responses of anotherspecies or based on the response of a population. To predict the effects on humanhealth, a toxicologist relies on laboratory animals as surrogates for humans. Lab-oratory animals are smaller, have a shorter lifespan, and can be bred to a ho-mogenous population. They are inexpensive and relatively easy to maintain. Therat has become a popular test species for toxicological study. Over the past50 years, scientists have come to realize that not all rats are the same, and breed-ing laboratories have purposely bred laboratory animals with slightly differentcharacteristics to be used in different types of studies. In addition, testing labo-ratories have become more aware of animal diseases, diet, and husbandry pro-cedures, which result in better survival with fewer background effects. Becauseof improved survival, fewer background lesions, and better husbandry of ani-mals, data from recent toxicity studies are likely to be more reliable predictors ofhuman health effects than data from 50 years ago.

Although the rat is a cornerstone of toxicology, scientists have recognized inrecent years that responses in rats do not always reflect responses in humans, thatis, rats are not small people [1–3]. This fact was first recognized for phthalate es-ters in the 1980s. After the National Toxicology Program (NTP) reported livercancer in rats and mice exposed to high levels of one phthalate ester, di(2-ethyl-hexyl)phthalate (DEHP), also known in the industry as DOP, which is differentfrom DnOP, researchers began to understand the mechanism of this cancer. It isthought that liver cancer is the result of biochemical changes in the liver cells, aprocess called peroxisome proliferation [4]. Because several therapeutic drugsused to lower cholesterol also produce these biochemical effects in rats, scientistsbegan looking at responses in human cells to determine if patients receivingthese drugs were at risk for cancer. Some of the first studies using cultured hu-man liver cells demonstrated that the biochemical changes that occurred in ratsdid not occur in human cells [5–8]. These studies were followed by others usingmonkeys, which also failed to demonstrate the same biochemical changes ob-served in rats [9–12]. Thus, it appears that humans may not be at risk for cancervia this mechanism because human liver cells do not undergo the kind of bio-chemical changes that rat cells do. Some of the most recent information suggeststhat other effects observed in rats, such as effects on the testes, also do not occurin monkeys [11, 12]. Thus, a rat cannot always predict effects in humans, at least

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters 301

not for phthalate esters. Why that is true is not clear. There are many differencesbetween rats and humans (or monkeys) in how phthalate esters are absorbed,broken down (metabolized) in the body, and excreted [13, 10]. There may also beintrinsic differences between rat cells and human cells in their capacity to re-spond, that is, rat cells may be far more susceptible to the effects of phthalate es-ters than are human cells. Whatever the reason, it is not appropriate to automat-ically assume that effects in laboratory animals accurately predict effects inhumans, at least not for phthalate esters.

1.2Exposure

Exposure is a key element in assessing the potential risk of adverse health effects.Unfortunately, determining the exposure of humans to phthalate esters has al-ways been difficult because, in part, phthalate esters are found in many products,especially products in the laboratory, that lead to false positive or suspicious an-alytical results. In addition, the medium (air, water, food, article) that contains thephthalate ester greatly influences the biological response because phthalate estersare very water-insoluble (especially high-molecular weight (MW) esters) and willbe absorbed only slowly through any membrane (intestinal wall from ingestionof water or food, lung from inhalation of air particles, skin from contact with ar-ticles). These issues are discussed in Chapter 2.

Hydrolysis of the phthalate diester to a monoester enhances absorption. Stud-ies have shown that for high-MW esters, breakdown (metabolism; hydrolysis) ofthe ester bond to liberate one alcohol and a remaining monoester greatly in-creases the absorption, since the monoester and alcohol are absorbed morerapidly than the diester [14–16]. The more hydrolysis occurs, the more mo-noester is available for absorption. Once absorbed, the monoester continues tobe metabolized into substances that are excreted in the urine [17].

The fact that high-MW phthalate esters need to be hydrolyzed to be absorbedis an important factor that relates to the medium (route) of exposure. For exam-ple, if exposure is through the skin from contact with poly(vinyl chloride) (PVC)articles containing phthalate esters, the absorption is very slow because the ester is not hydrolyzed and must be absorbed intact [18]. Experiments with laboratory animals and human skin have demonstrated that the absorption rateof phthalate esters is slow for high-molecular weight esters [19, 20]. Therefore,high levels of exposure of the skin to neat chemical might not result in adversehealth effects because the absorption is so slow and the metabolism of the esteris minimal. Likewise for inhalation exposure, absorption is likely to be slow butfaster than absorption through the skin and slower than absorption from inges-tion. This is because the lungs have some capacity to hydrolyze the ester [21], butit is not as predominant as it is in the intestine [22].

Not only is the medium (or route) of exposure (inhalation, ingestion, dermal)important in assessing the potential human health effects, the form of the esteris also important. For example, air might contain vapors of neat chemical or par-ticles of PVC containing phthalate esters. Inhaling vapors of neat chemical wouldhave a greater chance for absorption than inhaling PVC dust because the ester is

302 R.M. David and G. Gans

released from the PVC matrix very slowly [18]. In addition, particles such as PVCdust may get lodged in the nose or other airways rather than reaching the deeplung, further reducing the possibility of absorption. Laboratory experiments havedetermined that dermal contact with articles that contain phthalate esters resultsin minute quantities transferred to the skin, much less than if a worker is exposedto the neat chemical. Thus, the form of the exposure has a significant impact onany potential health effect.

The route of exposure that results in the most efficient absorption of phtha-late esters is ingestion. Laboratory studies have demonstrated, however, that ratsare far more efficient at hydrolyzing the esters and, subsequently, absorbing themonoester than primates (and presumably humans) [9, 10]. This means thatwhen studies of phthalate esters are conducted in laboratory animals wherehealth effects are observed following very high doses of an ester, it is very diffi-cult to reproduce such effects in primates (and presumably humans) because pri-mates do not absorb phthalate esters as efficiently as other laboratory animals.Primates and humans absorb about seven times less phthalate than do rats (es-pecially for high-MW esters) [9].At low doses, the absorption may be more com-parable.

Exposures to phthalate esters have been estimated by numerous governmentagencies. Based on concentrations of phthalates in the air, water, and food supply, the primary source of exposure is thought to be food. Minor levels oflow- and high-MW phthalates can migrate into food from packaging or inks.These levels and the total exposures from food are generally low (estimated to beless than 30 µg kg–1 d–1). Recent studies in the United States conducted by theCenters for Disease Control and Prevention (CDC) have measured the levels ofexcreted monoester in approximately 300 individuals as a means of evaluating total exposure [23]. The study was extended to about an additional 1000 indi-viduals with similar findings. The CDC results indicate that exposure to high-MW phthalate esters is lower than estimated by consumption of food. Conversely,exposure to low-MW phthalates is higher than estimated from levels in food.This finding should not be startling because low-MW phthalate esters are usedin consumer products such as pharmaceuticals and cosmetics. In any event,the levels are at or below calculated tolerable daily intake (TDI) values in the United States (Table 1). Exposure levels are presented and discussed in detail in Chapters 5 and 8. The concept of TDI and how it is derived is discussedbelow.

1.3Uncertainties in the Assessment of Risk

Risk is the predicted frequency of occurrence of an adverse effect of a chemicalsubstance from a given exposure to humans. Therefore, an indispensable pre-requisite is a dose-response relationship, which correlates an adverse effect (haz-ard) with an internal or external dose (exposure) allowing the identification ofa dose without any effect. The stages of a risk assessment include hazard identi-fication, exposure assessment, and risk estimation. Each of these steps providesa source of uncertainty to the whole process.

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters 303

Hazard identification can be obtained from human case reports, studies of hu-man volunteers, mechanistic information, and animal studies. Each of thesemethods has a drawback and introduces some level of uncertainty into the eval-uation process. For example, the major problem in using case reports is the un-certainty with respect to exposure levels. Case studies rarely provide adequatedata on actual exposure concentrations. Only studies of human volunteers, per-formed under well-defined exposure conditions, are best able to correlate actualexposure with human responses. On the other hand, for studies of human vol-unteers often only subjective parameters are reported. Therefore, these studiesrely heavily on the personal judgment of the test subjects, which can be biased byperception. For example, in a study of a malodorous chemical, subjects exposedto the malodorous substance will likely report more effects than subjects exposedto the same chemical but with an additive that masks the odor.

Uncertainty may also be introduced when relying on the results of mechanis-tic studies. Special attention should be paid to mechanistic studies in order toidentify a relevant end point for the human situation. Mechanistic studies pro-vide a good basis for extrapolation from lower species to higher species, or theymay provide a basis for discounting effects in lower species when considering thepotential to occur in higher species. An example is the effect of some phthalateesters to modify the biochemistry of the liver of rats and mice that results in livercancer. Mechanistic studies demonstrate convincingly that humans are not af-fected in the same way as rats and mice. The use of mechanistic studies to iden-tify hazards in lower species, which pertain to higher species, assumes that all rel-evant interspecies physiologic processes are the same. Uncertainty is introducedwhen we assume that the processes are the same, but we have not demonstratedsuch equivalency.

For animal studies, the most important information to be obtained is the dose-response relationship, identifying effect (lowest observed effect level or LOEL)and no-effect level (no observed effect level or NOEL). Determining these regu-latory values is dependent upon the selection of dose levels used in the experi-ment. Typically, three dose levels are used in animal studies: one at a maximumlevel that can just be tolerated by the animal, one at an intermediate level, and onethat should be at a no-effect level. Selecting the dose levels for any study is sci-

304 R.M. David and G. Gans

Table 1. Exposure levels for common phthalate esters based on urinary metabolites

Phthalate Geo. mean 95th percentile Estimated intake RfD

DEP 12.34 a 93.33 57 b 800DBP 1.45 6.37 7 100BBP 0.73 3.34 6 200DEHP 0.60 3.05 30 c 20DINP 0.21 1.08 10.8 ND d

a All values in mg kg–1 d–1 based on a maximum creatinine clearance of 20 mg kg–1 d–1.b Estimated intake taken from ATSDR, IPCS, or EU draft risk assessments.c From Doull J, Cattley R, Elcombe C, Lake BG, Swenberg J, Wilkinson C, Williams G, Van

Gemert M (1999) Reg Toxicol Pharmacol 29:327–357 using ATSDR estimates.d ND not determined.

entific judgment based on preliminary information, but there is a practice of us-ing multiples of three between dose levels. Not every investigator follows thispractice. Occasionally, investigators will use multiples of ten to establish dose lev-els.When evaluating the data from multiple studies, one should not assume thatthe lowest NOEL is the best value to establish risk.A reasonable NOEL should bebased on all available studies, not on only the lowest dose without any effect froma single study.

Once a reasonable NOEL is determined, the next step in risk assessment is tocompare the NOEL from animal studies to the exposure levels for humans. As-sessment of exposure levels also introduces some uncertainty because exposuresof the general public are only estimated. One way in which exposures can be es-timated is to determine the levels in products/foods/environment and extrapo-late to exposure level based on contact/consumption/physiology. The measure-ment of phthalate esters in a matrix carries with it inherent difficulties. Theapplication of an analytical method, which may liberate matrix-bound phthalateesters by using organic solvents, will overestimate the biologically available pro-portion. Furthermore, there could be significant differences between the exter-nal and internal exposure levels. For example, an air concentration measurementthat includes dust particles will overestimate the exposure because some parti-cles will be large enough that they will be impacted or retained in the nasal cav-ity.Another example is the absorption of a substance from the gut into the blood-stream may not reach 100% and a certain amount will be excreted in the feceswithout being absorbed. These examples need to be taken into account when de-veloping a reasonable and scientifically based risk assessment.

A goal of the comparisons between NOEL values from animal studies and human exposure is to determine the margin of safety (also called the margin ofexposure) to determine if humans are at risk of adverse health effects. In the absence of variable data for human exposure, regulators can establish a TDI orin the case of EPA, a reference dose (RfD) which sets an upper limit for humanexposure. These values take into account the NOEL value and the severity of theresponse, and assign a margin of safety for human exposure by using certain as-sumptions about the variability within a species and variability between species.Thus, a TDI value is typically 100 times lower than the NOEL from animal stud-ies. This assumes that variability among animals can be accounted for by a fac-tor of ten, that is, the most sensitive individual animal is likely no more than tentimes more sensitive than the rest of the population, and variability betweenspecies can be accounted for be a factor of ten, that is, humans are no more thanten times more sensitive than rats.

1.4Uses

Phthalate esters comprise the diesters of phthalic anhydride with alcohol moi-eties ranging from C1 to C13. These esters were developed to serve various technical demands. Table 2 summarizes the main characteristics of the phthalateesters [24]. This list is not comprehensive. Some low-MW phthalates can appearin a variety of products, some of which are consumer products. The manu-

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters 305

facturers of these products incorporate the phthalates because of the propertiesthey impart to the product. Generally, such applications use small quantities ofphthalate esters.

2Short-Term Exposure

2.1Oral/Dermal/Inhalation

The primary, common aim of toxicological testing is to provide information thatcan be used to assess the potential for adverse effects in humans. Short-term testsidentify effects after a single high-dose administration and an observation periodof 14 days. These tests reflect the possible hazards resulting from a single high ex-posure, which can occur in the workplace after an accidental release of phthalatesduring production, maintenance, or transportation.

In general, phthalates show little acute toxicity. A summary of LD50 values forthe most important commercial phthalates is given in the following table (com-piled from producer material safety data sheets and various reports from the German regulatory agency, BUA, Table 3). Because most regulatory bodies andscientists consider LD50 of greater than 5000 mg kg–1 to reflect a lack of tox-icity, these data demonstrate that phthalates are relatively non-toxic followingacute exposure.

2.2Irritation/Sensitization

The ability of a chemical to irritate the skin or eyes after local administration isusually determined in rabbits. For the phthalates, there is only slight or moder-ate irritation reported after administration of the undiluted substance (compiledfrom producer material safety data sheets and various BUA reports). There is also

306 R.M. David and G. Gans

Table 2. Uses of common phthalate esters

Phthalate Ester Use

DMP Specialty plasticizer used in cellulose esters and in cosmeticsDEPDBP

DBP Very good gelling agent for PVC, used in combination with other DiBP plasticizers; in floor coverings; as a plasticizer in the cosmetic and BBP pharmaceutical industries

DEHP Standard plasticizer for PVC, moderate volatility, DEHP used in medical DINP applicationsDIDP

DUP Specialty plasticizer

no evidence for a sensitization potential of the phthalate esters from animal stud-ies. There are a few case reports describing potential skin sensitization after der-mal contact with phthalate-containing products. Based on limited information ofpossible interference with other substances, these reports are not conclusive. Ina comprehensive volunteer study with seven dialkyl (C6-C13) phthalate esters ina human-repeated, insult patch test using the modified Draize procedure, no evidence of dermal irritation or sensitization for any of the seven phthalates wasobserved [25]. These data suggest that phthalate esters are not likely to play a rolein the induction of asthma, or other allergic or irritation-induced response (videinfra).

2.3Long-Term Risks from Short-Term Exposure

Toxicologists evaluate the short-term consequences of single exposures and theeffects of prolonged exposure, up to a lifetime. Because some tissues repair dam-age easily, the question of the long-term consequences of brief, sub-lethal expo-sures may be difficult to answer. Typically, tissues that grow rapidly or have a highrate of turnover, repair damage well. Tissues which repair easily include the bloodsynthesizing system (hematopoietic), gastrointestinal tract, testes (not ovaries),liver, and kidneys [26]. The nervous system repairs itself very poorly. All of thisassumes that the damage to the cell, tissue, or organ is not lethal and does not ir-reversibly alter the architecture of the organ.

2.3.1Metabolic Effects

Phthalate esters are a class of chemicals that in laboratory rodents cause changesin metabolism. Some enzyme systems are induced, while others appear to be suppressed. Such effects occur primarily in the liver (the primary drug-metabo-lizing organ of the body) or in the kidneys [4, 27–29]. There is little evidence thatadverse effects occur in other tissues or organs [30]. The effect can be observedmacroscopically as larger liver or kidneys (or increased organ weight), micro-

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters 307

Table 3. Acute lethal oral doses for common phthalate esters

Phthalate ester Oral LD50 (rat) Dermal LD50 (rabbit)(mg kg–1 bw) (mg kg–1 bw)

DMP > 5,000 >12,000DEP > 9,000 >20,000DBP > 8,000 >20,000DiBP >15,000 >10,000DEHP >30,000 >24,500DINP >10,000 > 3,100DIDP >20,000 > 3,600DUP >15,800 > 7,900

scopically as liver cells that contain more endoplasmic reticulum and increasedsize and number of organelles called peroxisomes, and biochemically as in-creased enzyme activity. Short-term exposure in animals of as little as five daysis adequate to produce such effects. These effects will persist as long as the ex-posure and will subside within one to two weeks following cessation of exposure[4]. Because the effects are transient and metabolic, they are considered to be“adaptive” in that the organs adapt to the presence of the phthalate. Not all phthalate esters produce such metabolic changes. Low-MW phthalates do not induce these metabolic changes [31, 32]; mid-MW phthalates are weak inducers,while high-MW phthalates are the best inducers among phthalates [33, 32].Exceptions to this rule exist [34]. It is known that some of the metabolic changesare unique to laboratory animals, specifically rats and mice [5, 11, 10, 35–38]. The relevance for metabolic changes in humans has been debated for nearly 15 years[3, 6, 8, 39–41].

2.3.2Regenerative Effects

Occasionally, short-term exposure can result in immediate tissue damage, but thenature of the tissue is such that the damage can be repaired. Such regenerationis limited to organs that have a high capacity for cell turnover and growth. Forexample, exposure of very young laboratory animals to high-dose levels(>1000 mg kg–1) of mid-MW phthalate esters (including DEHP) causes a reduc-tion in sperm production or aspermatogenesis.While this can be a severe adverseeffect, over time the testes are able to recover and produce normal amounts ofsperm [42].

For any organ such as the testis, the timing of exposure relative to the life cycle of the organism is an important factor in recovery. Studies comparing theeffect on the testes of young (immature) laboratory animals to old (mature) animals indicate that lower dose levels of transitional phthalates and DEHP haveeffects in young animals, but that older animals are insensitive to such lower lev-els [43]. The reason for the age-related difference in sensitivity is not known, butis likely related to the formation of the blood-testis barrier, which forms prior toadolescence. Following formation of this barrier, less toxicant would likely reachthe testes to cause effects.

Another organ capable of repair is the liver. Studies in laboratory animals using long-term exposure to high-MW phthalates followed by a period of non-exposure indicate that some liver effects are reversible. For example, a lesion associated with the Ito cell of the liver (spongiosis hepatis) or pigmentation ofKupffer cells is reversed during recovery even after prolonged exposure of three-quarters of a lifetime [44]. In addition, some effects considered to be end-stage,such as neoplastic changes in the liver, may be partially reversible [45]. This sug-gests that short-term exposure would not be expected to result in permanentdamage such as these types of cellular changes in later life. Studies with DEHPhave demonstrated that short-term exposure does not cause long-term effects inthe liver or testes [46, 42]. However, not all possible health effects have been in-vestigated.

308 R.M. David and G. Gans

Another organ that has been identified as a target for high-MW phthalate es-ters is the kidney. Effects on the kidneys vary depending on the length of expo-sure. Following short-term exposure, there is evidence of renal tubular degener-ation and proliferation, and cyst formation [47]. Over an extended period ofexposure, mineralization of the renal tubules may occur, as well as chronicnephropathy. These cellular changes also occur in untreated animals and are partof a normal aging process [48, 49]. Thus, exposure to high-MW phthalates maynot be solely responsible for the effects observed. Instead, it is possible that theselarger substances enhance the normal process of aging in laboratory animals.Therefore, the relevance to humans is unclear [50].

2.3.3Developmental Effects and Cancer

Although covered in a later section specifically on developmental toxicity, it bearsnoting that mid-MW phthalates including DEHP can have an effect on the developing embryo/fetus of laboratory animals following short-term exposure.Rats appear to be more sensitive than mice with dose levels of approximately100 mg kg–1 demonstrating no overt effect. Some very subtle effects have beenseen in rats. Low-MW phthalates and high-MW phthalates (excluding DEHP)produce effects on the developing embryo at dose levels at least ten times higherthan the effect dose levels for mid-MW phthalates. This will be discussed furtherin a later section.

The potential of phthalate esters to produce cancer following long-term ex-posure will be discussed below. Occasionally, questions arise about the potentialto develop cancer following short-term exposure. Based on animal studies, livercancer associated with exposure to some high-MW phthalates occurs only afterprolonged exposure [45, 51, 52].

3Prolonged Exposure

3.1Cancer

Carcinogens can be classified into two different groups: DNA-reactive carcino-gens (genotoxic) and epigenetic carcinogens. The genotoxic category comprisescarcinogens that chemically interact with DNA, based on an inherent reactivityor on metabolism-mediated reactivity. The second category is comprised of car-cinogens for which there is evidence of a lack of direct interaction with geneticmaterial, and for which another biologic process has been identified that couldbe the basis for carcinogenicity. This differentiation has major implications forhuman risk assessment. Genotoxic carcinogens may induce a mutagenic or precancerous effect after a single exposure. Conversely, epigenetic carcinogensproduce cancer only after a sustained level of exposure. Thus, for epigenetic car-cinogens, it may be possible to establish a “safe” threshold or even demonstratethat the underlying biochemical mechanism is not relevant for humans.

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters 309

In studies in rats and mice, DEHP and DINP induce liver tumors after high ex-posure [45, 53–55]. These phthalates belong to a diverse group of chemicals andtherapeutics, which induce specific changes called peroxisome proliferation inthe liver of rats and mice [4]. In the last decade, a better understanding of the roleof peroxisome proliferation in the carcinogenic process has been achieved. Thisunderstanding has led to the consensus that cancer induced by peroxisome pro-liferation is not relevant for humans [1, 39].

Recently, the IARC re-assessed the cancer classification of DEHP taking into ac-count the mechanism of carcinogenicity [27, 56, 57]. In summary, they concluded:

– “The weight of evidence for DEHP and its metabolic products demonstratesthat they do not act as direct DNA-damaging agents.

– DEHP produces liver tumors in rats and mice.– Under conditions of the bioassay, DEHP induces peroxisome proliferation and

cell replication in liver that are characteristic of a peroxisome proliferator inmice and rats.

– Rodent peroxisome proliferators exercise their pleiotropic effects in liver dueto activation of PPARa. This process is essential for liver hypertrophy and hy-perplassia and eventual hepatocarcinogenesis in response to peroxisome pro-liferation.

– Hepatic peroxisome proliferation has not been adequately evaluated in stud-ies of human livers following exposure of DEHP in vivo; however, the effect oftreatment of human and mouse hepatocytes with DEHP metabolites which areactive in rat hepatocytes, as well as other peroxisome proliferators, indicatethat humans can reasonably be predicted to be refractory to induction of per-oxisome proliferation and hepatocellular proliferation by DEHP. The evidenceindicates that the mechanism of peroxisome proliferation induced by DEHPin rat hepatocytes does not operate in humans.

– The absence of a significant response of human liver to induction of peroxi-some proliferation and hepatocellular proliferation is explained by several aspects of PPARa-mediated regulation in gene expression.

– Overall these findings indicate that the increased incidence of liver tumors in mice and rats treated with DEHP results from a mechanism that does notoperate in humans.”

This assessment of DEHP applies to all the other phthalate esters that are non-genotoxic but can induce peroxisome proliferation.Yet, there are marked differ-ences in the potency of peroxisome proliferation. Most of the commercial phthalate esters were compared for their potential to induce peroxisome prolif-eration [33]. In general, phthalate esters are weak peroxisome proliferators com-pared with other peroxisome proliferators such as hypolidemic drugs. The low-MW phthalate esters do not show any peroxisome proliferator induction capacityor, at best, minimal capacity. The mid- and high-MW phthalates are peroxisomeproliferators with the highest potency seen with DEHP, DINP, and DIDP [33, 58].In accordance with this ranking, there is no evidence or equivocal evidence ofcarcinogenic potential for DEP, DMP, and BBP [59, 60], but evidence of carcino-genic potential in animals for DEHP and DINP.Again, cancer in animals throughthis mechanism is not considered relevant for humans.

310 R.M. David and G. Gans

3.2Reproductive/Developmental Toxicity and Endocrine-Related Events

The male reproductive tract was identified as a target organ for some phthalateesters in the late 1970s. Investigators found that mid-MW phthalates could decrease sperm production in the testes of rats [61–64]. Studies of low- to high-MW phthalates have not demonstrated these effects, so testicular damage inadults occurs primarily with mid-MW phthalates, including DEHP, rather thanwith other phthalates. Dose levels that cause effects in adult male rats are generally > 250 mg kg–1 d–1. The most prominent effect of DEHP exposure is loss of sperm production. This effect translates into reduced fertility for rodents(rats and mice) at dose levels of >250 mg kg–1 d–1. The exact mechanism for the testicular damage is unknown, but there are data to suggest that the effect is on the Sertoli cell rather than on the Leydig cells or directly on sperma-togonia [65].

The relevance to humans of this effect in rodents has been debated. Two fac-tors modify the assessment. One factor indicates that absorption of at least high-MW phthalates is far less in monkeys (and probably humans) than in rodents [9].This means that much less of the active metabolite can reach the target organ tocause the adverse effect. The other factor is the lack of testicular effects in mon-keys treated with dose levels that cause testicular effects in rodents [11, 12]. Basedon these two factors, the adult human male may not be at risk for the same an-tifertility effects seen in rodents. This same conclusion has been voiced by someexpert groups [57, 66].

There are additional data for pre-adolescent (pre-pubertal) male rodents,which indicate that pre-pubescent males may have a greater susceptibility thanadolescent or adult males [43]. Testicular damage in pre-pubertal males can oc-cur at lower dose levels than testicular damage in adult males. The reason for thisdifference in age susceptibility is not known, but may be related to the formationof the blood-testis barrier during the puberty. This barrier provides some extraprotection to the adult testes. Conversely, very young males that are affected (hav-ing exhibited aspermia, for example) also exhibit the ability to regenerate and re-cover [42] to the point that the testes of treated animals are indistinguishablefrom untreated when the animals are adults. Thus, pre-pubertal male rodents,while being more susceptible to the reproductive effects of some phthalate esters,do not exhibit permanent damage.

Effects on the reproduction of female rodents have been reported based on thereduced fertility following breeding of treated females with untreated males.However, unlike studies in males, there are no cellular changes in the ovaries orreproductive organs of the female rodents that can be associated with these effects [26]. There are data to suggest that DEHP, a high-MW phthalate, given tofemale rats results in a disturbance of the estrous cycle (ovulation and fertility cycle), an effect that might cause reduced fertility in rodents [67]. Only very highdose levels are known to cause such effects, and it is not known if these distur-bances can be related to humans who have a menstrual cycle rather than an estrous cycle. The mechanism suggested for these disturbances is alterations inthe enzymes that form the sex hormones in the ovaries. Whether or not this

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters 311

mechanism is applicable for humans is not known, nor is the dose level at whichthese effects do not occur in rodents (a NOEL).

Toxicity to the developing embryo or neonate has also been investigated formost phthalate esters. In general, disturbances in the normal development of therodent fetus have been observed for mid-MW phthalates and DEHP at dose lev-els of >100 mg kg–1, as stated previously. The nature of the effects observed varywith the timing of the exposure and the dose level. Exposure throughout gesta-tion at dose levels above 500 mg kg–1 results in craniofacial abnormalities (cleftpalate, cleft lip, exposed brain, or incomplete formation of the skull or spine) andskeletal abnormalities (incomplete bone formation) [68–73]. The mechanism(s)for these effects is(are) not known. Recent studies have also indicated that ex-posure of pregnant rodents during late gestation to dose levels of >300 mg kg–1

mid-MW phthalates and DEHP affect the developing male reproductive tract re-sulting in incomplete formation of secondary sex organs or testicular malfor-mation [74–76]. No effects have been identified in the reproductive organs of fe-male rodent offspring. The mechanism for these effects on male offspring isthought to be related to the hormone signals necessary for proper developmentof the urogenital tract in males. There is evidence to suggest that some phthalateesters interfere with the formation of testosterone, although the exact mechanismis not clear. How much this interference is attributable to changes in biochem-istry, which is unique to rodents (as has been observed for the effects in the liver),has been debated. One investigator has demonstrated that the biochemicalchanges in the liver are not totally independent of effects in the testes [27].

It is of interest that the same phthalates that have effects on the male repro-ductive tract also affect development of the embryo. One possible explanation isthat both the male reproductive tract and the developing embryo are rapidly di-viding tissues that can be affected by exposure to some phthalate esters.Anotherexplanation is that there is a common mechanism for both reproductive effectsand developmental effects. At present, there is no information to suggest a com-mon mechanism. Some have attributed the similarity in response to endocrine-related events. Interference with endocrine-associated events in the developingreproductive tract and the pre-pubertal reproductive tract has been suggested.Associations with other developmental effects have not been established. How-ever, the evidence, along with data from tissue culture studies, has suggested tosome that these phthalate esters are endocrine disruptors. Regrettably, the termendocrine disruptor is not well defined. Phthalate esters are not estrogens (or es-trogen mimics). Some diesters have been reported to bind to, or weakly activate,the estrogen receptor in tissue culture [77, 78]. This has been misinterpreted asindicating that phthalate esters are estrogens. Further study using intact animalshas demonstrated that no estrogenic activity can be detected following oral exposure to doses of 20–2000 mg kg–1 d–1 phthalate esters [79, 80]. Neither arephthalate esters androgens [81, 82]. These substances do not either bind to theandrogen receptor, or interfere with, the activation of the androgen receptor bytestosterone.Yet, some of the phthalate esters cause effects that are similar to theeffects caused by substances that interfere with testosterone-triggered events. Forthis reason, the effects have been termed “anti-androgen-like,” albeit through anunknown mechanism.

312 R.M. David and G. Gans

The question: “Are phthalate esters endocrine disruptors?” is continuallyraised. Because these substances do not act on the hormone receptors (either toactivate or deactivate them), it is a difficult question to answer. In the sense thatthe normal function of the endocrine system is somehow altered to produce theeffect observed because there is less testosterone produced, then these substancesmay be endocrine disruptors. How that function is distinguished from othermetabolic changes that occur in rats and mice, and whether such changes can occur in humans, remains to be answered.

3.3Asthma

Recently, a question regarding a possible role of phthalate esters in the patho-genesis of asthma has been raised [83, 84]. The studies in question found phtha-late esters from vinyl products in the air of households with asthmatic children.They speculated on a possible mechanism for DEHP to cause asthma. The hypothesis is that DEHP and its metabolites have some structural similarities to, and can mimic, some prostaglandins and thromboxanes, and the widespreaduse of DEHP in wall coverings, flooring, and other construction material leads to the induction or exacerbation of asthma. The merit of this hypothesis is questionable. There is also no direct evidence that DEHP or MEHP acts likeprostaglandin, the hormone associated with inflammatory responses. Further-more, there is no indication of any sensitization potential from exposure toDEHP, based on animal experiments or human patch tests [25]. Phthalates are notthe large, complex molecules which are more commonly associated with allergy,and are not among the substances such as insect parts, animal dander, and pollen,which authorities have associated with the induction of asthma.Vinyl flooring iseasy to clean; it is often recommended for use in homes where asthmatics live. Ifan association between phthalate esters used in vinyl exists, it is not clear whetherthe use of vinyl flooring is a cause of the asthma or a remedy to reduce dust andanimal dander that cause asthma. Further work is necessary to evaluate this phe-nomenon.

3.3Neurotoxicity

Only one phthalate ester has been tested in a stand-alone neurotoxicity studies.One study under EPA guidelines was conducted and the results indicated no evidence of neurotoxicity [85]. Other studies using only functional tests ofneurobehavior support this finding even in laboratory animals exposed duringdevelopment [86]. These stand in contrast to a single report of altered neuro-behavioral activity in a single, invalidated beam-walking test following in uteroexposure [87]. The significance of this finding is questionable.

As for the other phthalate esters, none of the recent 13-week and/or chronictoxicity studies of some low-, mid-, and high-MW phthalates have indicated ab-normalities in clinical observations or behavioral assessments. While the focusof these studies was not neurotoxicity, they incorporated detailed daily observa-

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters 313

tion of the animals that should have detected neurotoxicity. Thus, phthalate esters do not appear to be neurotoxic even following short-term or prolonged exposure.

4References

1. IARC (1995) Peroxisome proliferation and its role in carcinogenesis. IARC technical reportno 24. IARC Press, Lyon

2. IPCS (1992) Environmental health criteria document no 131, WHO3. Cattley RC, DeLuca J, Elcombe C, Fenner-Crisp P, Lake BG, Marsman DS, Pastoor TA, Popp

JA, Robinson DE, Schwetz B, Tugwood J, Wahli W (1998) Regul Toxicol Pharmacol 27:474. Moody DE, Reddy JK (1978) Toxicol Appl Pharmacol 45:4975. Benford DJ, Patel S, Reavy HJ, Mitchell A, Sarginson NJ (1986) Food Chem Toxicol 24:

7996. Elcombe CR, Mitchell AM (1986) Environ Health Perspect 70:2117. Bichet N, Cahard D, Fabre G, Remandet B, Gouy D, Cano JP (1990) Toxicol Appl Pharma-

col 106:5098. Elcombe CR, Bell DR, Elias E, Hasmall SC, Plant NJ (1996) Ann NY Acad Sci 804:6289. Rhodes C, Orton TC, Pratt IS, Batten PL, Bratt H, Jackson SJ, Elcombe CR (1986) Environ

Health Perspect 65:29910. Short RD, Robinson EC, Lington AW, Chin AE (1987) Toxicol Ind Health 3 :18511. Kurata Y, Kidachi F,Yokoyama M, Toyota N, Tsuchitani M, Katoh M (1998) Toxicol Sci 42:4912. Pugh G Jr, Isenberg JS, Kamendulis LM, Ackley DC, Clare LJ, Brown R, Lington AW, Smith

JH, Klaunig JE (2000) Toxicol Sci 56:18113. Astill BD (1989) Drug Metab Rev 21:3514. Rowland LR, Cottrell RC, Phillips JC (1974) Food Cosmet Toxicol 12:29315. Rowland LR, Cottrell RC, Phillips JC (1977) Food Cosmet Toxicol 15:1716. White RD, Carter DE, Earnest D, Mueller J (1980) Food Cosmet Toxicol 18:38317. Albro PW, Moore B (1974) J Chromatogr 94:20918. Deisinger PJ, Perry LG, Guest D (1998) Fd Chem Toxic 36:52119. Elsisi AE, Carter DE, Sipes IG (1989) Fundam Appl Toxicol 12:7020. Scott RC, Dugard PH, Ramsey JD, Rhodes C (1987) Environ Health Perspect 74:22321. Bogdanffy MS, Randall HW, Morgan KT (1986) Toxicol Appl Pharmacol 82:56022. Carter JE, Roll DB, Petersen RV (1974) Drug Metab Dispos 2 :34123. Blount BD, Silva MJ, Caudill SP, Neeham LL, Pirkle JL, Sampson EJ, Lucier GW, Jackson RJ,

Brock JW (2000) Environ Health Perspect 108:97924. Meier L (1990) Plasticizers. In: Gächter R, Müller H (eds) Plastics additives, 3rd edn. Hanser

Publishers25. Medeiros AM, Devlin DJ, Keller LH (1999) Cont Dermat 41:28726. Lamb JC IV, Chapin RE, Teague J, Lawton AD, Reel JR (1987) Toxicol Appl Pharmacol

88:25527. Ward JM, Peters JM, Perella CM, Gonzales FJ (1998) Toxicol Pathol 26:24028. Cimini AM, Sulli A, Stefanini S, Serafini B, Moreno S, Rossi L, Giorgi M, Ceru MP (1994) Cell

Mol Biol 40:106329. Ohno S, Fujii Y, Usuda N, Murata F, Nagata T (1982) Acta Histochem Cytochem 15:4030. Bell FP, Gillies PJ (1977) Lipids 12:58131. Bell FP (1982) Environ Health Perspect 45:4132. Gray TJB, Lake BG, Beamand JA, Foster JR, Gangolli SD (1983) Toxicology 28:16733. Barber ED, Astill BD, Moran EJ, Schneider BF, Gray TJB, Lake BG, Evans JG (1987) Toxicol

Ind Health 3 :734. DeAngelo AB, Garrett CT, Manolukas LA, Yario T (1986) Toxicology 41:27935. Butterworth BE, Smith-Oliver T, Earle L, Loury DJ, White RD, Doolittle DJ, Working PK,

Cattley RC, Jirtle R, Michalopoulos G, Strom S (1989) Cancer Res 49:1075

314 R.M. David and G. Gans

36. Deangelo AB, Daniel FB, Mcmillan L, Wernsing P, Savage RE Jr (1989) Toxicol Appl Phar-macol 101:285

37. Dirven HAAM, van den Broek PHH, Peeters MCE, Peters JGP, Mennes WC, Blaauboer BJ,Noordhoek J, Jongeneelen FJ (1993) Biochem Pharmacol 45:2425

38. Hall M, Matthews A, Webley L, Harling R (1999) Toxicol Sci 24:23739. Ashby J, Brady A, Elcombe CR, Elliott BM, Ishmael, J, Odum J, Tugwood, JD, Kettle S,

Purchase LF (1994) Hum Exp Toxicol 13, Suppl 2 :S140. Bentley P, Calder I, Elcombe C, Grasso P, Stringer D, Wiegand HJ (1993) Food Chem Toxi-

col 31:85741. Lake BG (1995) Ann Rev Pharmacol Toxicol 35:48342. Dostal LA, Chapin RE, Stefanski SA, Harris MW, Schwetz BA (1988) Toxicol Appl Pharma-

col 95:10443. Sjöberg P, Lindquist NG, Plöen L (1986) Environ Health Perspect 65:23744. David RM, Moore MR, Finney DC, Guest D (2001) Toxicol Pathol 29:45. David RM, Moore MR, Cifone MA, Finney DC, Guest D (1999) Toxicol Sci 50:19546. Ward JM, Diwan BA, Ohshima M, Hu H, Schuller HM, Rice JM (1986) Environ Health

Perspect 65:27947. Crocker J F, Safe SH, Acott P (1988) J Toxicol Environ Health 23:43348. David RM, Moore MR, Finney DC, Guest D (2000) Toxicol Sci 55:43349. David RM, Moore MR, Finney DC, Guest D (2000) Toxicol Sci 58:37750. Woodward KN (1990) Hum Exp Toxicol 9 :39751. Ward JM, Hagiwara A,Anderson LM, Lindsey K, Diwan BA Toxicol Appl Pharmacol 96:49452. Marsman DS, Cattley RC, Conway JG, Popp JA (1988) Cancer Res 48:673953. Kluwe WM, McConnell EE, Huff JE, Haseman JK, Douglasw JF, Hartwell WV (1982) Envi-

ron Health Perspect 45:12954. Butala JH, Moore MR, Cifone MA, Bankston JR, Astill B (1996) Toxicologist 30:20255. Butala JH, Moore MR, Cifone MA, Bankston JR, Astill B (1996) Toxicologist 36:17356. Peters JM, Cattley RC, Glonzalez FJ (1997) Carcinogenesis 18:202957. IARC Monographs on the evaluation of carcinogeic risks to humans (2000) Some indus-

trial chemicals, vol 77. IARC Press, Lyon58. Huber WW, Grasl-Kraupp B, Schulte-Hermann R (1996) Crit Rev Toxicol 26:36559. National Toxicology Program (NTP) NTP TR429, 199560. National Toxicology Program (NTP) NTP TR458, 199761. Agarwal DK, Maronpot RR, Lamb JC IV, Kluwe WM (1985) Toxicology 35:18962. Agarwal DK, Eustis S, Lamb JC IV, Reel JR, Kluwe WM (1986) Environ Health Perspect

65:34363. Gray TJB, Butterworth KR (1980) Arch Toxicol Suppl 4 :45264. Gray TJB, Beamand JA (1984) Food Chem Toxicol 22:12365. Lloyd SC, Foster PMD (1988) Toxicol Appl Pharmacol 95:48466. Center for the Evaluation of Risks to Human Reproduction (CERHR) NTP-CERHR expert

panel report on di(2-ethylhexyl) phthalate. NTP-CERHR-DEHP-00, 200067. Davis BJ, Maronpot RR, Heindel JJ (1994) Toxicol Appl Pharmacol 128:21668. Ema M, Harazono A, Miyawaki E, Ogawa Y (1997) Bull Environ Contam Toxicol 58:63669. Ema M, Amano H, Ogawa Y (1994) Toxicology 86:163070. Ema M, Miyawaki E, Kawashima K (1998) Toxicology Lett 98:8771. Ema M, Miyawaki E, Kawashima K (2000) Toxicology Lett 111:27172. Ema M, Kurosaka R, Harazono A, Amano H, Ogawa Y (1996) Arch Environ Contam Toxi-

col 31:17073. Ema M, Kurosaka R, Amano H, Ogawa Y (1994) Reprod Toxicol 8 :23174. Wine RN, Li LH, Barnes LH, Gulati DK, Chapin RE (1997) Environ Health Perspect 105:10275. Mylchreest E, Cattley RC, Foster PMD (1998) Toxicol Sci 43:4776. Gray LE, Ostby J, Furr J, Price M, Rao Veeramachaneni DN, Parks L (2000) Toxicol Sci

58:35077. Jobling S, Reynolds T, White R, Parker MG, Sumpter JP (1995) Environ Health Perspect

103:582

Summary of Mammalian Toxicology and Health Effects of Phthalate Esters 315

78. Coldham NG, Dave M, Sivapathasundaram S, McDonnell DP, Connor C, Sauer MJ (1997)Environ Health Perspect 105:734

79. Milligan SR, Balasubramanian AV, Kalita JC (1998) Environ Health Perspect 106:180. Zacharewski TR, Meek MD, Clemons JH, Wu ZF, Fielden MR, Matthews JB (1998) Toxicol

Sci 46:28281. Gray LE Jr, Wolf C, Lambright C, Mann P, Price M, Cooper RL, Ostby J (1999) Toxicol Ind

Health 15:9482. Sohoni P, Sumpter JP (1998) J Endocrinol 158:32783. Oie L, Hersoug L-G, Madsen JO (1997) Environ Health Perspect 105:97284. Jaakola JJ, Oie L, Nafstad P, Botten G, Samuelsen SO, Magnus P (1999) Am J Pub Health

89:18885. Moser VC, Cheek BM, MacPhail RC (1995) J Toxicol Environ Health 45:17386. Merkle J, Klimisch HJ, Jaeckh R (1988) Toxicol Lett 42:21587. Arcadi FA, Costa C, Imperatore C, Marchese A, Rapidisarda A, Salemi M, Trimarchi GR,

Costa G (1998) Food Chem Toxicol 36:963

316 R.M. David and G. Gans: Summary of Mammalian Toxicology and Health Effects