ww.sciencedirect.com

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 8

Available online at w

journal homepage: www.elsevier .com/locate/watres

Degradation of Diclofenac by Advanced Oxidationand Reduction Processes: Kinetic Studies,Degradation Pathways and Toxicity Assessments

Hui Yu a, Er Nie b, Jun Xu b, Shuwen Yan a,b, William J. Cooper c, Weihua Song b,*aChinese Research Academy of Environmental Sciences, Beijing, 100012, P. R. ChinabDepartment of Environmental Science & Engineering, Fudan University, Shanghai 200433, P. R. ChinacUrbanWater Research Center, Department of Civil and Environmental Engineering, University of California, Irvine, CA 92697, United States

a r t i c l e i n f o

Article history:

Received 20 November 2012

Received in revised form

31 December 2012

Accepted 9 January 2013

Available online 18 January 2013

Keywords:

Pharmaceuticals

Diclofenac

Advanced Oxidation / Reduction

Processes

Kinetic studies

Degradation mechanism

Toxicity assessments

* Corresponding author. Tel.: þ86 21 6564 20E-mail address: wsong@fudan.edu.cn (W

0043-1354/$ e see front matter ª 2013 Elsevhttp://dx.doi.org/10.1016/j.watres.2013.01.016

a b s t r a c t

Many pharmaceutical compounds and metabolites are found in surface and ground waters

suggesting their ineffective removal by conventional wastewater treatment technologies.

Advanced oxidation/reduction processes (AO/RPs), which utilize free radical reactions to

directly degrade chemical contaminants, are alternatives to traditional water treatment.

This study reports the absolute rate constants for reaction of diclofenac sodium and model

compound (2, 6-dichloraniline) with the two major AO/RP radicals: the hydroxyl radical

(�OH) and hydrated electron (e�aq). The bimolecular reaction rate constants (M�1 s�1) for

diclofenac for �OH was (9.29 � 0.11) � 109, and for e�aq was (1.53 � 0.03) �109. To provide

a better understanding of the decomposition of the intermediate radicals produced by

hydroxyl radical reactions, transient absorption spectra are observed from 1 e 250 ms. In

addition, preliminary degradation mechanisms and major products were elucidated using60Co g-irradiation and LC-MS. The toxicity of products was evaluated using luminescent

bacteria. These data are required for both evaluating the potential use of AO/RPs for the

destruction of these compounds and for studies of their fate and transport in surface

waters where radical chemistry may be important in assessing their lifetime.

ª 2013 Elsevier Ltd. All rights reserved.

1. Introduction (Behera et al. 2011, Matamoros et al. 2009) and as a result they

There is a rising concern with the occurrence and persistence

of Pharmaceutical and Personal Care Products (PPCPs) in the

aquatic environment, due to their potential impacts on the

aqueous ecosystems and human health (Kumar and

Xagoraraki 2010, Schwarzenbach et al. 2006). The worldwide

consumption of medicines provides a continuous release of

these substances or their metabolites to the environment.

Conventionalwastewater treatment systems such as filtration

and activated sludge do not efficiently remove these PPCPs

40.. Song).ier Ltd. All rights reserved

have been found in a wide range of environmental samples

including surface water, groundwater and drinking water

(Benotti et al. 2009, Kim et al. 2007, Kolpin et al. 2002, Makris

and Snyder 2010). Therefore, advanced treatment technol-

ogies need to be evaluated and eventually employed (Yang

et al. 2011), that are capable of either the complete removal

of these chemicals from wastewater or at the very least the

destruction of their biological activity (Snyder et al. 2003).

Recently studies indicate that the nanofiltration and

reverse osmosis processes guarantee the rejection of PPCPs

.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 81910

(Radjenovic et al. 2008). However, biofouling of membrane

elements and disposal of retentate are considered major

problems in these processes (Ben Abdelmelek et al. 2011,

Snyder et al. 2007, Wintgens et al. 2006). Ozonation can

destroy some of PPCPs in raw and/or clarified water; unfor-

tunately, the competition between the PPCPs and organic

material in the raw water may lead to rapid depletion of

ozone, resulting in incomplete oxidation of PPCPs (Ikehata

et al. 2006, Wert et al. 2009), in some cases, more toxic

byproducts formed(Aguinaco et al. 2012). Formation of carci-

nogenic bromate ion is also a general concern during ozone

water treatment where bromide ion is present in the water

(von Gunten 2003).

Advanced oxidation/reduction processes (AO/RPs) are al-

ternatives to traditional treatment and have recently received

considerable attention for PPCPs removal. The formation of

oxidizing hydroxyl radicals (�OH) and the reducing hydrated

electrons (e�aq), can be utilized in the destruction of organic

pollutants present in drinking or wastewater (Deng and

Ezyske 2011). They are effective in the treatment of a variety

of anthropogenic pollutants including PPCPs (Deng 2009, Li

et al. 2012, Song et al. 2009). However, to provide a funda-

mental understanding of the applicability of these processes

to degrade PPCPs, it is necessary to determine the bimolecular

reaction rate constants, the reaction efficiency and degrada-

tion mechanisms, as well as the toxicity of the degradation

products.

This study focused on diclofenac, a common non-

steroidal anti-inflammatory drug (NSAID). It is often found

as a persistent toxic waste and one of the most widely

available drugs in the world. Approximately hundreds of

tons of this prescription drug is sold annually worldwide

(Buser et al. 1998). The average concentrations detected are

in the low mg L�1 range in influents and effluents of

municipal sewage treatment plants and surface waters in

Austria, Pakistan, Germany and the United States (Al-Rifai

et al. 2007, Kolpin et al. 2002, Scheurell et al. 2009, Stulten

et al. 2008). Even at very low concentrations there are

adverse effects in different organisms. In the livers, kidneys

and gills of rainbow trout, the lowest observed effect con-

centration for cytopathology occurred at 1 mg L�1

(Triebskorn et al. 2004). An ecological effect resulted from

diclofenac residues which caused the vulture population

decline in Pakistan (Oaks et al. 2004). Therefore, it is critical

to develop a fundamental understanding of the fate and

oxidative and reductive degradation of diclofenac during

treatment processes.

The objective of this study was to establish the absolute

bimolecular reaction rate constants for reaction of the �OH

and the hydrated electron (e�aq) with diclofenac in aqueous

solution. Transient spectra from the reaction with the �OH

were recorded from 1 e 250 ms to provide a better under-

standing of the nature of the radical intermediate species.

Detailed studies of degradation pathways of diclofenac using

steady-state 60Co g-irradiation were undertaken, and these

suggest that �OH addition to the benzene ring and hydrated

electron reduction of chlorine are responsible for a significant

fraction of the observed degradation. Advanced reduction

process more likely remove toxicity than the advanced oxi-

dation processes.

2. Materials and Methods

2.1. Materials

Diclofenac, 2, 6-dichloraniline and catalase (bovine liver) were

purchased from Sigma-Aldrich and used without any further

purification. Methanol, 2-propanol, and acetic acid (Fisher

Science) were of HPLC grade. All solutionswere prepared in 5.0

mM phosphate buffer and adjusted to pH 7.0 with NaOH or

H3PO4, as necessary.

2.2. Pulse radiolysis and g-radiolysis

Pulse radiolysis experiments were performed at the United

States Department of Energy, Notre Dame Radiation Labora-

tory using an 8-MeV Titan Beta model TBS-8/16-1S linear

accelerator that produced 2 ns electron pulses which generate

radical concentrations of 1-3 mM per pulse. All experimental

data were taken by averaging 8 to 15 replicate pulses using the

continuous flow mode of the instrument. Dosimetry was

performed with N2O-saturated, 1.00 x 10�2 M KSCN solutions

monitored at l ¼ 472 nm.

The radiolysis of water is described in Eq 1:

H2O =n=n=n � > e�aqð0:26Þ þH$ð0:06Þ þ $OHð0:27Þ þH2ð0:05Þ

þH2O2ð0:07Þ þH3Oþð0:27Þ (1)

Where the numbers in parentheses are G values (yields) in

mmol J�1. Reactions with the hydroxyl radical were achieved

by using a nitrous oxide (N2O) pre-saturated solution, which

quantitatively converted solvated electrons and hydrogen

atoms (H�) to the �OH radical. (Buxton et al. 1988)

e�aq þN2OþH2O/N2 þOH� þ $OH k2 ¼ 9:1�109 M�1 s�1 (2)

H$þN2O/$OHþN2 k3 ¼ 2:1� 106 M�1 s�1 (3)

Reactions between the solvated electron and diclofenac

were studied in N2-saturated solutions buffered to pH 7.0.

These solutions contained 0.10 M isopropanol to scavenge the

hydroxyl radicals and hydrogen atoms, to convert them into

relatively inert isopropanol radicals. (Buxton et al. 1988)

ðCH3Þ2CHOHþ $OH/ðCH3Þ2C$OHþH2O k4 ¼ 1:9� 109 M�1 s�1

(4)

ðCH3Þ2CHOHþH$/ðCH3Þ2C$OHþH2 k5 ¼ 7:4� 107 M�1 s�1

(5)

A Shepherd� 109-86 Cobalt-60 source was used for g radi-

olysis with samples of 1.0 mM diclofenac saturated with N2O

or N2 saturated before irradiation. The dose rate was 7.72 krad

min�1, as measured by Fricke dosimetry.

2.3. HPLC and mass spectral analysis

The concentration of diclofenac was determined using an

Agilent 1200 HPLC using the following conditions: column,

Phenomenex Gemini C18 250 � 4.6 mm i.d.; mobile phase

consisting of 15 % CH3OH, 15 % CH3CN and 70 % 10 mM

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 8 1911

phosphate buffer solution (pH 3.0). The LC-MS system was an

Agilent 1100 HPLC Pump and Waters LCT Classic Mass Spec-

trometer with electrospray ionization source and a Phenom-

enex Luna C18 (2) column (2.0� 250mm). The injection volume

of the samples was 10 mL. Themobile phase was A: 98 % H2Oþ2 % CH3CN þ0.2 % formic acid and B: 50 % CH3OH, 50 % CH3CN

and 0.2 % formic acid. Gradient elution was 0 % of B for 5 min

followed by a linear increase to 100 % in 50min, and then held

constant for an additional 10min. Themass spectra data were

obtained in the positive ion mode by scanning from m/z 100

to 350.

2.4. Ion chromatographic analysis

Chloride ion released by the reaction of diclofenac with �OH

and e�aq was quantified by ion chromatography (IC-1010,

Techcomp�) with conductivity detector. Separation was per-

formed on an Ion-Pac AS16 anion column (4 � 250 mm, Dio-

nex) using 35 mM NaOH eluent solution at a flow rate of 1.0

mL min�1.

2.5. Toxicity assay

The bioluminescence of the marine bacterium Vibrio fischeri

was used to assess the toxicity of the diclofenac and its

decomposition products. The assays were performed under

250 300 3500.000

0.005

0.010

0.015

0.020

0.025

Abso

rban

ce

Wavele

(a)

0 100 200 300 400

0.000

0.005

0.010

0.015

0.020

Abso

rban

ce

time (µs)

(330 nm absorbance)fitted in y= 0.0090+0.0091e(-2.1E3x)

(b) (c

Fig. 1 e (a) Transient absorption spectra obtained from electron

diclofenac sodium at pH 7.0 and room temperature. (b) Decay o

the protocol developed by Microtox�. The luminescence was

determined with a Luminometer DXY-2 (Institute of Soil Sci-

ence, Chinese Academy of Sciences). Addition of catalase

(bovine liver) was used to eliminate the effect from H2O2

produced from g irradiation. Toxicity was determined after 30

min incubation. For colored solutions, toxicity was corrected

by subtracting absorbance at 490 nm according to the method

described elsewhere (Corporation 1992).

3. Results and Discussion

3.1. Transient Absorption Spectra of �OH radicalreaction with diclofenac

The absorption spectra of the intermediate transients formed

in the reaction of �OH with diclofenac were monitored over

wavelengths of 250 - 480 nm using pulse radiation, as shown

in Fig. 1a. Two strong absorption peaks appeared at 330 nm

and 370 nm. The transient at 330 nm followed first order

decay, k1 ¼ 2.1 � 103 s�1, as illustrated in Fig. 1b. The decay

curve at 370 nm was different, and fitted to a second order

exponential giving rate constants of k2 ¼ 7.3 �104 s�1 and k3 ¼4.5�103 s�1. The latter was similar to that of the 330 nmdecay,

as illustrated in Fig. 1c. This results suggested that at least two

400 450 500

ngth (nm)

1 µs 5 µs 50 µs 100 µs 250 µs

0 100 200 300 400

0.000

0.005

0.010

0.015

0.020

0.025

Abso

rban

ce

time (µs)

(370 nm absorbance) fitted in y= 0.0056+0.0083e(-4.5E3x)+0.014e(-7.3E4x)

)

pulse radiolysis of N2O saturated aqueous solutions of

f absorption at 330 nm. (c) Decay of absorption at 370 nm.

0.00 0.50 1.00 1.50-0.005

0.000

0.005

0.010

0.015

0.020

0.025

Abso

rban

ce

time (µs)

(a)

0.2 0.4 0.6 0.8 1.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0(b)

Rat

e co

nsta

nt (1

06 s-1)

Conc. of diclofenac (mM)

Fig. 2 e (a) Typical growth kinetics of the transient

absorption at 370 nm in a pulse irradiated solution at pH

7.00 and room temperature of 1.00 (,), 0.412 (B), 0.271 (D),

and 0.184 (V) mM diclofenac. Fitted lines are exponential

growth kinetics, giving pseudo-first-order rate constants of

(8.73 ± 0.45) x 106, (3.84 ± 0.07) x 106, (2.46 ± 0.10) x 106, and

(1.64 ± 0.02) x 106 sL1, respectively. (b) Second-order rate

constant determination for the reaction of hydroxyl

radicals with diclofenac at 330 (,) and 370 (B) nm. Solid

line corresponds to value of overall rate constant of

reaction of k [ (9.29 ± 0.09) 3 109 ML1 sL1.

Table 1e Summary of second-order rate constants in thisstudy (bold) in comparison to values for analogouscompounds from the literature.

Compound kOH/M�1s�1 ke-/M

�1s�1

diclofenac (9.29 ± 0.11) x109 (1.53 ± 0.03) x109

2, 6-dichloroaniline (6.97 ± 0.14) x109 (3.26 ± 0.04) x109

m-dichlorobenzene 5.7 x109 (Kochany

and Bolton 1992)

5.2 x109 (Anbar

and Hart 1964)

Aniline 1.0 x 1010

(Solar et al. 1986)

3.0 x107

(Solar et al. 1986)

Acetic acid 8.5 x107

(Buxton et al. 1988)

1.1 x106

(Buxton et al. 1988)

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 81912

major species exist in the �OH-mediated oxidation of the

diclofenac.

In general hydroxyl radicals are highly reactive with aro-

matic moieties and through addition form the hydrox-

ycyclohexadienyl radicals, as evidenced by the characteristic

transient absorption of 300 to 400 nm (Merga et al. 1994, Merga

et al. 1996, Peller and Kamat 2005). Diclofenac has two distinct

aromatic rings: 2, 6-dichloraniline and phenylacetic acid

group. Therefore two group intermediates could be assigned

according to the �OH radical electrophilic adduction. To verify

this hypothesis, 2, 6-dichloroaniline (DCA)was selected as one

model compound for pulse radiolysis. The transient spectrum

of �OH reaction with DCA (Fig. 1S, Supplemental Material) is

similar to long time scale transient spectrum of diclofenac

after 50 ms. It demonstrated that hydroxycyclohexadienyl

radicals in the DCA ring were relatively stable. Meanwhile,

intermediates from phenylacetic group were unstable and

correspond to the fast decay of the peak of 370 nm.

3.2. Kinetic Measurements

Pseudo-first-order growth rate constants for the reaction of

�OH with diclofenac were determined by fitting exponential

growth curves to the time-dependent absorbance of the

transient monitored at 330 nm and 370 nm over a range of

different diclofenac concentrations (Fig. 2a). At both wave-

lengths, it was observed that the initial growth in absorbance

was followed by a second, smaller, concentration-

independent growth, which was accounted for the data fit-

ting by using the sum of two exponential growths. The hy-

droxyl radical bimolecular rate constant for this reaction, k ¼(9.29� 0.11)� 109M�1 s�1, was determined from a plot of these

pseudo-first-order rate constants as a function of diclofenac

concentration (see Fig. 2b). This rate constant is slightly fast

than the steady-state competition kinetic value of (7.5 � 1.5)�109 M�1 s�1 reported by Huber (Huber et al. 2003) for �OH

generated by g-irradiation. The difference between our stud-

ies and the previous studies may be the result of slightly dif-

ferent reaction conditions (solution pH) and the inherent

errors and uncertainties associated with the different

methods used to measure the rate constants.

To determine the site of �OH attack, the rate constant of

�OH reaction with diclofenac was compared to that of several

model compounds determined in this study or reported else-

where (Table 1). The rate constant of aniline was slightly

higher than DCA and similar to diclofenac indicating that �OH

attacked both aromatic rings. Aniline is somewhat more

reactive that DCA due to the high electron density on the ring.

The �OH reaction rate constant of acetic group was w100

times slower, reflecting the fact that �OH abstraction of

hydrogen from alkyl group played a minor role.

The second order rate constant for the reaction of the

solvated electron (e�aq) with diclofenac was determined by

fitting single exponential decays to the absorbance of e�aq

monitored at 700 nm (Fig. 3a). Plotting these pseudo-first-

order values against diclofenac concentration, a second-

order rate constant of k ¼ (1.53 � 0.03) � 109 M�1 s�1 was

obtained (Fig. 3b). Our diclofenac value is two or three orders

faster than for the reduction of aniline (3 x 107 M�1 s�1) or

acetic acid (1.1 � 106 M�1 s�1), nevertheless similar to rate

0.0 1.0 2.0 3.0 4.0

0.00

0.01

0.02

0.03

0.04

Abso

rban

ce

time (µs)

(a)

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Rat

e C

onst

ant (

106 )

Conc. of diclofenac sodium (mM)

(b)

Fig. 3 e (a): Typical decay kinetics for hydrated electron

reduction at 700 nm for 1.00 (,), 0.621 (D), and 0.363 (A)

mM diclofenac at pH[ 7.0 and room temperature. (b):

Second-order rate constant determination for the reaction

of the hydrated electron with diclofenac. Solid line

corresponds to value of overall rate constant of reaction of

k [ (1.53 ± 0.03) 3 109 ML1 sL1.

0 5 10 15 20

0.0

0.2

0.4

0.6

0.8

1.0

N saturated/ IPA N O saturated

Con

c. o

f Dic

lofe

nac

(mM

)

Irradiation dose (kGy)

(a)

0 5 10 15 20

0.0

0.4

0.8

1.2

1.6

N saturated / IPA N O saturated

x

Con

c. o

f Cl- (m

M)

Irradiation dose (kGy)

(b)

Fig. 4 e (a) Degradation curve of diclofenac sodium at pH

7.0 under g irradiation: N2O saturated conditions to form

�OH (,), N2 saturated IPA solutions to form eLaq (B). (b)

Formation of ClL by g irradiation in N2O saturated

conditions (,); N2 saturated IPA solutions (B).

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 8 1913

constant for m-dichlorobenzene (5.2 � 109 M�1 s�1) or DCA

(3.26 �109 M�1 s�1). From these rate constants, it appears that

the reduction occurs predominately at the DCA moiety in

diclofenac, rather than the phenylacetic acid group. The dif-

ferences of solvated electron reaction rate constants between

diclofenac and DCA can be accounted for based on the elec-

tronic influence of the amino substituent or more likely from

the steric hindrance of the phenylacetic acid group.

3.3. Degradation efficiency and dechlorination

Steady-state g irradiation of diclofenac in N2O saturated (�OH)

and N2 saturated aqueous isopropanol solutions (e�aq) showed

decreasing concentration with increasing dose (Fig. 4a). The

curvature of the plot is consistent with previously reported

irradiation studies for other contaminants in water (Jeong

et al. 2010a, Jeong et al. 2010b, Luo et al. 2012), suggesting

competition for the reactive species (�OH and e�aq) between

diclofenac and the reaction by-products at the higher applied

doses.

These data allow estimation of the efficiencies of initial

�OH oxidation and e�aq reduction of diclofenac (Mezyk et al.

2007). At the lowest dose, an estimation of the initial slope

wasm¼ -3.23� 10�4 M kGy�1 for �OH radical, (the straight line

in Fig. 4a). Using the tangent at t ¼ 0, all of the �OH reacts with

diclofenac and the degradation efficiency computed at 60 %.

The similar assumption was also been applied to estimate the

e�aq reaction efficiency and was 100 %.

Irradiation of diclofenac solution resulted in the release of

chloride ion as measured by ion chromatograph (Fig. 4b). The

dechlorination by �OH (ipso attack at the C-Cl bond) fit to

a straight line gave a formation rate of 9.2 �10�5 M kGy�1,

which reflects only 17 % �OH radical attacked chlorine func-

tional group. Dechlorination resulting from dissociative elec-

tron transfer from the e�aq fitted a pseudo first order growth,

and the initial formation rate is 2.7 �10�4 M kGy�1. This rate,

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 81914

equivalent e�aq G value (2.6 �10�4 M kGy�1), indicated

a dechlorination reaction efficiency of 100 %.

3.4. Degradation mechanism

Diclofenac decomposition products in N2O saturated (�OH

oxidation) solutions or 0.1 M isopropanol N2 saturated (sol-

vated electron reduction) solutions were analyses by LC-MS at

various irradiation doses. The structural assignments of the

breakdown products of diclofenac during g-irradiation were

based on the analysis of the Total Ion Chromatogram (TIC) and

the correspondingmass spectrawith consideration of isotopic

abundance, as shown in Fig. 2S, Supplemental Material. The

masses of the different products were determined from the

peaks corresponding to the protonatedmolecule, [MþH] þ. Forthe purpose of this paper, we will refer to the products by

molecular weight (MW).

The major degradation products produced in the steady-

state g irradiation of diclofenac in N2O saturated solutions

are summarized in Fig. 5. Three separate products withMWof

311 were observed, corresponding to the addition of 16 mass

units to the parent peak. This is consistent with hydroxylation

of the aromatic ring (Hofmann et al. 2007, Homlok et al. 2011).

The addition of the electrophilic hydroxyl radical to the aro-

matic ring forms a resonance-stabilized carbon-centered

radical with subsequent elimination of hydrogen radical,

Fig. 5 e Primary degradation mechanis

yielding the phenolic products. Generally the specificity of

electrophilic aromatic substitution is typically governed by

the nature of the substitute, which may account for our

observation of three different products with the same m/z

ratio. Since the amino group is strong electron donating group

and act as ortho-para directors, three (M þ16) products are

proposed in the Fig. 5.

With further hydroxyl radical oxidation, multi di-

hydroxylation products (MW 327) were formed, which ver-

ifies the assumption of hydroxyl substitution, since the hy-

droxyl group increases the electron density of the aromatic

ring and thus, hydroxyl radical electrophilic adductions pro-

ceed faster. It is also proposed that the MW 309 product was

the result of further oxidation with the above primary phe-

nolic degradation product (MW 311a) to form quinine imine

product, which was also observed in photo-Fenton degrada-

tion (Perez-Estrada et al. 2005). One minor product was

observed with the MW 275, corresponding to the loss of HCl

from the primary phenolic product (MW 311b). This suggests

that phenolic substitute chlorine group subsequently under-

went cyclization to form a six-membered ring. The compound

at MW 177 could be formed following hydroxyl radical ipso-

attack on the primary product MW 311c resulting in bond

cleavage. Another mono-aromatic product of MW 151 was

observed and isotope mass peak indicated that no chlorine

group was present in this product (Fig. 2S, Supplemental

m of �OH oxidation of diclofenac.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 8 1915

Material). Therefore, the structure has also been proposed in

Fig. 5. Ipso-attack of the DCA ring by �OH resulted in aMW 162

product, was not detected under our experimental conditions,

likely due to the low response for the positive ionization.

Isopropanol (0.1 M) diclofenac solutions saturated with N2

were used to study solvated electron degradationmechanism.

Dechlorination is the major degradation pathway, which is in

good agreement with chloride ion mass balance experiments.

As a consequence, carbon centered radicals were the major

intermediates after de-chlorination, which then underwent H

abstraction fromH2O forming the product of MW261, or intra-

molecular reactions forming the product of MW 259, as illus-

trated in Fig. 6.

3.5. Toxicity assessments

While our results demonstrate that �OH and e�aq effectively

degrade diclofenac, it is critical to establish the biological ac-

tivity of the resulting treated solution or the individual

breakdown products. In general, reactions of �OH radical and

e�aq lead to the complex mixtures of products in low overall

yields; it is a daunting task to isolate the individual reaction

byproducts and assess their individual biological activities.

We chose to use the luminescent bacteria assay, V. fischeri, to

assess the biological activity of the treated solutions at various

irradiation doses (Michael et al. 2012). Analysis of the bio-

luminescent inhibition activity helps to assess the eco-

Fig. 6 e Primary degradation mechanism of h

toxicity potential of the treated solutions. A calibration curve

for the bacteria inhibition as a function of the concentration of

diclofenac was constructed as illustrated in Fig. 3S,

Supplemental Material. Eighty percent inhibition of the bio-

luminescence occurred at diclofenac concentrations of 0.23

mM. A lower detection limit of approximately 0.03 mM was

established from the calibration curve. The inhibition curve of

diclofenac showed an EC50 of 0.10 mM. The initial concen-

tration of diclofenac (1.0 mM) in solutions subjected to g

irradiation was outside the linear region of the calibration

curve and hence diluted 5 times accordingly before running

the assays. With the initial �OH radical oxidation, the residual

concentration of diclofenac decreased rapidly, 75 % of diclo-

fenac was destroyed upon 4.0 kGy. However, the biological

activity of the treated samples was constant, implying that

toxic breakdown products are formed to a significant extent,

as shown in Fig. 7. With further oxidation, the toxic products

could be eliminated and the toxicities of treated solution

decrease slowly. While 100 % of diclofenac was removed at

12.0 kGy, residual toxicity remained at 40 % inhibition.

In contrast to the �OH radical oxidation products, the tox-

icity of reductive products resulting from the e�aq reactions

were dissimilar. As illustrated in Fig. 7, the toxicity of the so-

lutions kept constant during the initial 0.5 kGy irradiation,

while 15 % of diclofenac have been removed. During the stage

from1 kGy to 4 kGy, the toxicities kept constant and significant

lower than the oxidative solutions. The results indicated that

ydrated electron reduction of diclofenac.

0

20

40

60

80

100

0 0.2 0.5 1 1.5 4 8 12 20

%

Irradiation dose (kGy)

Inhibition of bacteriam by OH treated

Inhibition of bacterium by e treated

residence of diclofenac by OH treated

residence of dichlofenac by e treated

Fig. 7 e Degradation of diclofenac under g irradiation and

the parallel removal of toxicity measured by inhibition of

V. fischeri after 30 min of exposure: ( ) bioluminescent

inhibition of N2O saturated solutions treated by g

irradiation, (,) the residence of diclofenac in the N2O

saturated solutions measured by HPLC. ( )

bioluminescent inhibition of N2 saturated IPA solutions

treated by g irradiation, (D) the residence of diclofenac in

the N2 saturated IPA solutions measured by HPLC.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 81916

toxic reductive products were formed during the initial deg-

radation, while significantly less toxic than oxidation prod-

ucts.With further irradiation, the toxicities decreased steadily

with increasing solvated electron treatment. The observed

biological activity of the treated samples paralleled the con-

centration of diclofenac from 4 kGy to 20 kGy, implying that

toxic breakdown products are not formed to a significant

extent and/or do not exhibit inhibition of luminescent bacte-

ria. A plausible explanation would be that the toxicity of

diclofenac associated with chlorine, was selectively elimi-

nated by reaction of the e�aq forming the chloride ion. In

comparison, �OH radical is less selective and the initial

products appear similar to the parent compound leading to

a less efficient reduction of toxicity.

4. Conclusion

The absolute bimolecular reaction rate constants for the re-

action of hydroxyl radical and solvated electron with diclofe-

nacweremeasured at (9.29� 0.11)� 109 and (1.53� 0.03)� 109

M�1 s�1, respectively. Destruction pathways of diclofenac

were proposed and the reaction by-products proposed. Prod-

uct identification was consistent with known hydroxylation

mechanism involving attack on the aromatic rings. This

approach provides a firm scientific underpinning for under-

standing the details of the free-radical chemistry involved in

the degradation of diclofenac. Further, by determining the

reaction efficiencies of the two reactive species with diclofe-

nac are 60 % and 100 %, respectively, it is possible to engineer

a more efficient treatment system. Additionally, the toxicity

evaluation of the degradation products and their related in-

termediate species indicated that advanced reduction process

is more suitable for removing the toxicity, while the reduction

processes generally need higher irradiation dose.

Acknowledgments

Pulse radiation performed at the Radiation Laboratory, Uni-

versity of Notre Dame, which is supported by the Office of

Basic Energy Sciences, U.S. Department of Energy. H.Y. ac-

knowledges support for this work from the Major Science and

Technology Program for Water Pollution Control and Treat-

ment (2012ZX07101-001).W.S. thanks partial funding supports

from National Natural Science Foundation of China

(21107016), the Ministry of Science and Technology of China

(2012YQ220113-4), the Science & Technology Commission of

Shanghai Municipality (12PJ1400800). W.J.C. and W.S.

acknowledge support from National Science Foundation

(CBET-1034555). We all thank the reviewers for valuable in-

sights and suggestions. This is contribution 78 of the Urban

Water Research Center, University of California at Irvine.

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.watres.2013.01.016

r e f e r e n c e s

Aguinaco, A., Beltran, F.J., Garcıa-Araya, J.F., Oropesa, A., 2012.Photocatalytic ozonation to remove the pharmaceuticaldiclofenac from water: Influence of variables. ChemicalEngineering Journal 189e190 (0), 275e282.

Al-Rifai, J.H., Gabelish, C.L., Schafer, A.I., 2007. Occurrence ofpharmaceutically active and non-steroidal estrogeniccompounds in three different wastewater recycling schemesin Australia. Chemosphere 69 (5), 803e815.

Anbar, M., Hart, E.J., 1964. The reactivity of aromaticcompounds toward hydrated electrons. J. Am. Chem. Soc. 86,5633e5637.

Behera, S.K., Kim, H.W., Oh, J.-E., Park, H.-S., 2011. Occurrenceand removal of antibiotics, hormones and several otherpharmaceuticals in wastewater treatment plants of the largestindustrial city of Korea. Sci. Total Environ 409 (20), 4351e4360.

Ben Abdelmelek, S., Greaves, J., Ishida, K.P., Cooper, W.J.,Song, W., 2011. Removal of Pharmaceutical and PersonalCare Products from Reverse Osmosis Retentate UsingAdvanced Oxidation Processes. Environ. Sci. Technol 45 (8),3665e3671.

Benotti, M.J., Trenholm, R.A., Vanderford, B.J., Holady, J.C.,Stanford, B.D., Snyder, S.A., 2009. Pharmaceuticals andEndocrine Disrupting Compounds in U.S. Drinking Water.Environ. Sci. Technol 43 (3), 597e603.

Buser, H.-R., Poiger, T., Muller, M.D., 1998. Occurrence and Fate ofthe Pharmaceutical Drug Diclofenac in Surface Waters: RapidPhotodegradation in a Lake. Environ. Sci. Technol 32 (22),3449e3456.

Buxton, G.V., Greenstock, C.L., Helman, W.P., Ross, A.B., 1988.Critical review of rate constants for reactions of hydrated

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 8 1917

electrons, hydrogen atoms and hydroxyl radicals (.OH/.O�) inaqueous solution. J. Phys. Chem. Ref. Data 17 (2), 513e886.

Corporation, M. (1992) MicrotoxR Manual Carlsbad, California.Deng, Y., 2009. Advanced Oxidation Processes (AOPs) for

reduction of organics in mature landfill leachates: A Review.Int. J. Environ. Waste Manag 4 (3-4), 366e384.

Deng, Y., Ezyske, C.M., 2011. Sulfate radical-advanced oxidationprocess (SR-AOP) for simultaneous removal of refractoryorganic contaminants and ammonia in landfill leachate.Water Res. 45 (18), 6189e6194.

Hofmann, J., Freier, U., Wecks, M., Hohmann, S., 2007.Degradation of diclofenac in water by heterogeneous catalyticoxidation with H2O2. Applied Catalysis B: Environmental 70(1e4), 447e451.

Homlok, R., Takacs, E., Wojnarovits, L., 2011. Elimination ofdiclofenac from water using irradiation technology.Chemosphere 85 (4), 603e608.

Huber, M.M., Canonica, S., Park, G.-Y., von Gunten, U., 2003.Oxidation of pharmaceuticals during ozonation andadvanced oxidation processes. Environ. Sci. Technol 37 (5),1016e1024.

Ikehata, K., Naghashkar, N.J., Gamal El-Din, M., 2006.Degradation of aqueous pharmaceuticals by ozonation andadvanced oxidation processes: A review. Ozone: Sci. Eng 28,353e414.

Jeong, J., Jung, J., Cooper, W.J., Song, W., 2010a. Degradationmechanisms and kinetic studies for the treatment of X-raycontrast media compounds by advanced oxidation/reductionprocesses. Water Res. 44 (15), 4391e4398.

Jeong, J., Song, W., Cooper, W.J., Jung, J., Greaves, J., 2010b.Degradation of tetracycline antibiotics: Mechanisms andkinetic studies for advanced oxidation/reduction processes.Chemosphere 78 (5), 533e540.

Kim, S.D., Cho, J., Kim, I.S., Vanderford, B.J., Snyder, S.A., 2007.Occurrence and removal of pharmaceuticals and endocrinedisruptors in South Korean surface, drinking, and wastewaters. Water Res. 41 (5), 1013e1021.

Kochany, J., Bolton, J.R., 1992. Mechanism of photodegradation ofaqueous organic pollutants. 2. Measurement of the primaryrate constants for reaction of $OH radicals with benzene andsome halobenzenes using an EPR spin-trapping methodfollowing the photolysis of H2O2. Environ. Sci.. & Technol 26,262e265.

Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M.,Zaugg, S.D., Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals,Hormones, and Other Organic Wastewater Contaminants inU.S. Streams, 1999-2000: A National Reconnaissance. Environ.Sci. Technol 36 (6), 1202e1211.

Kumar, A., Xagoraraki, I., 2010. Pharmaceuticals, personal careproducts and endocrine-disrupting chemicals in U.S. surfaceand finished drinking waters: A proposed ranking system. Sci.Total Environ 408 (23), 5972e5989.

Li,W.,Nanaboina,V.,Zhou,Q.,Korshin,G.V., 2012.Effectsof Fentontreatment on the properties of effluent organicmatter and theirrelationships with the degradation of pharmaceuticals andpersonal care products. Water Res. 46 (2), 403e412.

Luo, X., Zheng, Z., Greaves, J., Cooper, W.J., Song, W., 2012.Trimethoprim: Kinetic and mechanistic considerations inphotochemical environmental fate and AOP treatment. WaterResearch 46 (4), 1327e1336.

Makris, K.C., Snyder, S.A., 2010. Screening of pharmaceuticalsand endocrine disrupting compounds in water supplies ofCyprus. Water Sci. Technol 62 (11), 2720e2728.

Matamoros, V., Arias, C., Brix, H., Bayona, J.M., 2009. Preliminaryscreening of small-scale domestic wastewater treatmentsystems for removal of pharmaceutical and personal careproducts. Water Res. 43 (1), 55e62.

Merga, G., Rao, B.S.M., Mohan, H., Mittal, J.P., 1994. Reactions ofOH and SO4.bul.- with some halobenzenes andhalotoluenes: A radiation chemical study. J. Phy. Chem. 98(37), 9158e9164.

Merga, G., Schuchmann, H.P., Rao, B.S.M., Von Conntag, C., 1996.OH radical-induced oxidation of chlorobenzene in aqueoussolution in the absence and presence of oxygen. J. Chem. Soc.Farad. T. 2 6, 1097e1103.

Mezyk, S.P., Neubauer, T.J., Cooper, W.J., Peller, J.R., 2007. Free-radical-induced oxidative and reductive degradation of sulfadrugs in water: absolute kinetics and efficiencies of hydroxylradical and hydrated electron reactions. J. Phys. Chem. A 111,9019e9024.

Michael, I., Hapeshi, E., Osorio, V., Perez, S., Petrovic, M.,Zapata, A., Malato, S., Barcelo, D., Fatta-Kassinos, D., 2012.Solar photocatalytic treatment of trimethoprim in fourenvironmental matrices at a pilot scale: Transformationproducts and ecotoxicity evaluation. Sci. Total Environ 430 (0),167e173.

Oaks, L.J., Gilbert, M., Virani, M.Z., Watson, R.T., Meteyer, C.U.,Rideout, B.A., Shivaprasad, H.L., Ahmed, S., Chaudhry, M.,Arshad, M., Mahmood, S., Ali, A., Khan, A.A., 2004. Diclofenacresidues as the cause of vulture population decline inPakistan. Nature 427, 630e633.

Perez-Estrada, L.A., Malato, S., Gernjak, W., Aguera, A.,Thurman, E.M., Ferrer, I., Fernandez-Alba, A.R., 2005. Photo-Fenton Degradation of Diclofenac: Identification of MainIntermediates and Degradation Pathway. Environ. Sci.Technol 39 (21), 8300e8306.

Peller, J., Kamat, P.V., 2005. Radiolytic Transformations ofChlorinated Phenols and Chlorinated Phenoxyacetic Acids.J. Phy. Chem A 109 (42), 9528e9535.

Radjenovic, J., Petrovic, M., Ventura, F., Barcelo, D., 2008. Rejectionof pharmaceuticals in nanofiltration and reverse osmosismembrane drinking water treatment. Water Res. 42 (14),3601e3610.

Scheurell, M., Franke, S., Shah, R.M., Huhnerfuss, H., 2009.Occurrence of diclofenac and its metabolites in surface waterand effluent samples from Karachi, Pakistan. Chemosphere 77(6), 870e876.

Schwarzenbach, R.P., Escher, B.I., Fenner, K., Hofstetter, T.B.,Johnson, C.A., von Gunten, U., Wehrli, B., 2006. The Challengeof Micropollutants in Aquatic Systems. Science 313 (5790),1072e1077.

Snyder, S.A., Adham, S., Redding, A.M., Cannon, F.S.,DeCarolis, J., Oppenheimer, J., Wert, E.C., Yoon, Y., 2007.Role of membranes and activated carbon in the removal ofendocrine disruptors and pharmaceuticals. Desalination202, 156e181.

Snyder, S.A., Westerhoff, P., Yoon, Y., Sedlak, D.L., 2003.Pharmaceuticals, Personal Care Products, and EndocrineDisruptors in Water: Implications for the Water Industry.Environ. Eng. Sci. 20, 449e469.

Solar, S., Solar, W., Getoff, N., 1986. Resovled mutisite OH-attackon aqueous aniline studied by pulse radiolysis. Radiat. Phys.Chem. 28, 229e234.

Song, W., Cooper, W.J., Peake, B.M., Mezyk, S.P., Nickelsen, M.G.,O’Shea, K.E., 2009. Free-radical-induced oxidative andreductive degradation of N, N0-diethyl-m-toluamide (DEET):Kinetic studies and degradation pathway. Water Res. 43 (3),635e642.

Stulten, D., Zuhlke, S., Lamshoft, M., Spiteller, M., 2008.Occurrence of diclofenac and selected metabolites insewage effluents. Sci. Total Environ 405 (1e3), 310e316.

Triebskorn, R., Casper, H., Heyd, A., Eikemper, R., Kohler, H.R.,Schwaiger, J., 2004. Toxic effects of the non-steroidal anti-inflammatory drug diclofenac Part II. Cytological effects in

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 0 9e1 9 1 81918

liver, kidney, gills and intestine of rainbow trout(Oncorhynchus mykiss). Aquat. Toxicol 68 (2), 151e166.

von Gunten, U., 2003. Ozonation of drinking water: Part II.Disinfection and by-product formation in presence ofbromide, iodide or chlorine. Water Res. 37 (7), 1469e1487.

Wert, E.C., Rosario-Ortiz, F.L., Snyder, S.A., 2009. Effect of ozoneexposure on the oxidation of trace organic contaminants inwastewater. Water Res. 43, 1005e1014.

Wintgens, T., Melin, T., Schafer, A., Khan, S., Muston, M., Bixio, D.,Thoeye, C., 2006. The role of membrane processes inmunicipal wastewater reclamation and reuse. Desalination187, 271e282.

Yang, X., Flowers, R.C., Weinberg, H.S., Singer, P.C., 2011.Occurrence and removal of pharmaceuticals and personalcare products (PPCPs) in an advanced wastewater reclamationplant. Water Res. 45 (16), 5218e5228.