A smog chamber and modeling study of the gas phase NOx–air photooxidation of toluene and the...

A Smog Chamber and Modeling Study of the Gas Phase NO,-Air Photooxidation

of Toluene and the Cresols

ROGER ATKINSON Shell Research Ltd., Thornton Research Centre, Chester, United Kingdom, and

Statewide Air Pollution Research Center, University of California, Riverside, California 92521

and

WILLIAM P. L. CARTER,* KAREN R. DARNALL,** ARTHUR M. WINER, and JAMES N. PITTS, Jr.

Statewide Air Pollution Research Center, University of California, Riverside, California 92521

Abstract

An experimental and modeling study of irradiated toluene-NO,-air, toluene-benzaldehyde-NO,-air, and cresol-NO,-air mixtures a t part-per-million concentrations has been carried out. These mixtures were irradiated a t 303 f 1 K in a 5800-liter Teflon-lined, evacuable environmental chamber, with temperature, humidity, light intensity, spectral distribution, and the concentrations of 0 3 , NO, NOz, toluene, PAN, formaldehyde, benzaldehyde, o-cresol, m-nitrotoluene, and methyl nitrate being monitored as a function of time. For the toluene and toluene-benzaldehyde-NO,-air runs a variety of initial reactant concentrations were investigated. Cresol-NO,-air runs were observed to be much less reactive in terms of 0 3 formation and NO to NO2 conversion rates than toluene-NO,-air runs, with the relative reactivity of the cresol isomers being in the order meta >> ortho > para. The addition of benzaldehyde to toluene-NO,-air mixtures decreased the reactivity, in agreement with previous studies. Alternative mechanistic pathways for the NO, photooxidations of aromatic systems in general are discussed, and the effects of varying these mechanistic alternatives on the model predictions for the toluene and o-cresol-NO,-air systems are examined. Fits of the calculations to most of the experimental concentration-time profiles could be obtained to within the experimental uncertainty for two of the mechanistic options considered. In both cases it is assumed that (1) 0 2 adds to the OH-toluene adduct -75% of the time forming, after a further addition of 0 2 , a C7 bicyclic peroxy radical, and (2) this C7 bicyclic peroxy radical reacts with NO -75% of the time to ultimately form a-dicarbonyls and conjugated y-dicarbonyls (e.g., methylglyoxal + 2-butene-1,4-dial) and -25% of the time to form organic nitrates. The major uncertainties in the mechanisms concern (1) the structure of the bicyclic peroxy intermediate, and (2) the y-dicarbonyl photooxidation mechanism. Good fits to the o-cresol concentration-time profiles in the toluene-NO, runs are obtained if it is assumed that o-cresol

* To whom correspondence should be addressed. * * Present address: Sandia National Laboratories, Albuquerque, New Mexico 87185.

International Journal of Chemical Kinetics, Vol. XII, 779-836 (1980) 8 1980 John Wiley & Sons, Inc. 0538-8066/80/0012-0779$05.80

780 ATKINSON ET AL.

reacts rapidly with NO3 radicals. However, it is observed that the model underpredicts nitrotoluene yields by a factor of -10, but this is in any case a minor product. It is concluded that further experimental work will be required to adequately validate the assumptions in- corporated in the aromatic photooxidation mechanisms presented here.

Introduction

Aromatic hydrocarbons are recognized to be an important constituent of anthropogenic emissions in polluted urban atmospheres [l], and, for instance, toluene levels of -250 ppb (parts per billion) have been reported in Japan [2] and levels of -40 ppb are common in the California South Coast Air Basin [3,4].

Hence, in order to develop chemical kinetic computer models of photochemical air pollution as an integral part of urban air shed models for the formulation of emission control strategies, the chemistry involved in the photooxidation of these aromatic hydrocarbons under atmospheric conditions must be reasonably well understood [5,6]. As a first step, such computer kinetic models must, as far as possible, be validated against ac- curate smog chamber data, even though it is recognized that there are significant wall effects associated with smog*chambers [7]. At the present time, while reasonably validated detailed chemical kinetic computer models exist for the n-butane (a representative short chain [IC,] alkane) and propene (a representative a1kene)-NO,-air systems [7,8], the chemistry occurring in the NO, photooxidations of the aromatic hydrocarbons is still very incompletely understood [9].

In recent years laboratory studies have shown that under atmospheric conditions the major initial reaction of the aromatics is with the hydroxyl radical [9-201, and data are now available concerning both the overall OH radical rate constants [9-181 and the relative amounts of OH radical addition and H atom abstraction [9,15,16,19,20] for many of these compounds. Additionally, the reactions of ozone with the aromatic hydrocarbons, leading to the production of a-dicarbonyls [21], have been shown to be negligible under atmospheric conditions [21-241.

Several product studies of irradiated NO,-aromatic-air systems (mainly of the NO,-toluene-air) have been carried out 125-341, although quantitative measurements of a large variety of products and of the reactants under carefully controlled experimental conditions have not yet been made available. Although products retaining the aromatic ring, including benzaldehyde, benzyl nitrate, cresols, nitrotoluene and nitrocresols in the gas phase [25,26,28-341, and dihydroxynitrotoluenes in the aerosol phase [31], are observed, the formation of large yields of products such as PAN, CO, and HCHO in the toluene-NO, irradiations [28,31,32,34] and the recent observations of significant yields of the a-dicarbonyls glyoxal, methylglyoxal, and biacetyl in irradiated NO, -0 -xylene-air mixtures [35,36] indicate the importance of ring cleavage.

NOx-AIR PHOTOOXIDATION OF TOLUENE AND THE CRESOLS 781

Recently, Hendry and co-workers 181 have formulated a detailed chemical computer model for the NO, photooxidation of toluene involving aromatic ring cleavage only after cresol formation, i.e., OH radical addition to toluene yields mainly o-cresol, with the o-cresol reacting further with OH radicals to yield a variety of ring cleavage products. However, Darnall, Atkinson, and Pitts [35] have shown that the rate-determining step for the formation of the ring cleavage product biacetyl from the reaction of OH radicals with o-xylene was that for the initial reaction of OH radicals with o-xylene, and that its production from the reaction of OH radicals (or ozone) with a hy- droxyxylene would have led to an incorrect biacetyl concentration-time profile.

In this work, product studies of toluene-NO,-air, toluene-benzaldehyde-NO,-air, and cresol-NO,-air mixtures irradiated under controlled conditions in an environmental chamber are presented, and a reaction scheme that employs the available kinetic and mechanistic information and is based in part upon a previously developed chemical computer model for the n-butane-propene-NO,-air systems 171 is formulated. Some of the detailed chemistry is, by necessity, speculative, but this reaction scheme fits the data reasonably well over a significant range of initial reactant levels.

Experimental

The apparatus and general techniques used have been described previously 134,371. Irradiations were carried out in a 5800-liter evacuable, Teflon-lined, thermostatted environmental chamber [38] using a 25-kW solar simulator [38,39]. Prior to each run the chamber was evacuated to pressures of torr, and was then filled with purified matrix air [40] humidified to -50% relative humidity (RH) at 303 f 1 K. Toluene or the cresols, and NO and NO2 were injected into the chamber using precision bore syringes, while benzaldehyde was expanded into the evacuated chamber from known benzaldehyde-N2 mixtures prepared in a 5.5-liter bulb. All reactants were allowed to mix for a t least 30 min before t = 0. During a run the temperature of the irradiated mixture was maintained at the desired value by means of the chamber refrigeration or heating system.

The parameters monitored and the methods employed were: 0 3 by UV absorbance (Dasibi model 1212); NO, Nos, and NO, by chemilumines- cencel (Thermo Electron model 14B); CO by gas chromatography (Beck- man 6800); RH by a thin film, bulk effect humidity sensor (Brady Array); light intensity and spectral distribution by a photodiode and an absolute radiometer (EG & G), and by a double monochromator-photomultiplier

NO2 and NO, values are corrected for response to peroxyacetyl nitrate and alkyl nitrates ~411.

782 ATKINSON ET AL.

combination, respectively [34]; and HCHO using an improved chromotropic acid method [34,37,42]. The absolute light intensity was determined pe- riodically using the rate of photolysis of NO2 in N2 to determine k l [43].

Benzaldehyde, benzyl nitrate, cresols, nitrotoluenes, and hydroxynitrotoluenes were monitored by gas chromatography (GC) with flame ion- ization detection (FID) [34] using either (1) a 10-ft X l/*-in. (2-mm inner diameter) borosilicate glass column packed with 60/80 mesh Tenax GC, temperature programmed from 373 to 523 K at 20 K /min, or (2) a 6-ft X Y 4 - h . 0.1% SP 1000/Carbopack C (Supelco) column, temperature programmed from 423 to 523 K at 20 K/min. In both cases 10.8- or 3.6-liter gas samples from the chamber were drawn through a '/4 X 3i-in. borosilicate glass trap packed with @in. Tenax GC 60/80 mesh. The sample was then transferred by the carrier gas a t 528 K from this trap to the column head which was at 373 or 423 K, followed by the temperature programming of the columns as noted above.

Peroxyacetyl nitrate (PAN) and methyl nitrate were monitored by GC using an 18 X l/8-in. Teflon column of 5% Carbowax 400 on Chromosorb G (SO/lOO mesh) operated at 301 K with electron capture detection (ECD) [44,45]. Toluene and the oxygenates were monitored employing a 10-ft X l/s-in. stainless steel (SS) column of 10% Carbowax 600 on C-22 firebrick (100/120 mesh) [34,37,46]. In later runs, toluene and benzaldehyde were monitored employing a 5-ft X l/8-in. SS column of 5% Carbowax 600 on C-22 firebrick (100/115 mesh) for which the sampling trap and injection valve were heated to 368-373 K [34]. Acetylene and background alkanes and alkenes were monitored on Porapak N or 2,4-dimethylsulfolane columns as described previously [34,37,46].

The reliability of the reactant and product analyses employing these techniques is estimated [34,37] to be as follows: 0 3 and NO, f5%; NO2, a t best f5%, but more uncertain because of uncertainties in the measurements of interfering compounds (see footnote 1) and the possibility of in- terferences from unmonitored compounds; CO, f 10%; toluene, &lo%, aromatic products monitored on the SP-1000 or the Tenax GC system, at best f25%; benzaldehyde on the 5 f t Carbowax 600 system, &lo%; PAN, f20%; formaldehyde, +10 to -30%; and methyl nitrate, 3~50%.

In addition, qualitative analyses of the aromatic products (benzaldehyde, benzyl nitrate, cresols, nitrotoluenes, and hydroxynitrotoluenes) were carried out using a Finnigan 3100 combined gas chromatograph-mass spectrometer (GC-MS) with a 6100 data system [47]. A 26-m X 0.29-mm i.d. glass capillary column coated with Ucon-50-HB-5100 polypropylene glycol, temperature programmed from 313 to 492 K at 8 K/min, was employed for these analyses.

Product assignments were made by comparison of the retention times (FID and ECD) and/or mass spectra (GC-MS) [47] with authentic samples, wherever possible.

NO+-AIR PHOTOOXIDATION OF TOLUENE AND THE CRESOLS 783

Computational Methods

Most calculations were carried out at Riverside as discussed previously [7]. Input parameters consisted of the initial concentrations of the reactants, a reaction mechanism consisting of a list of reactions and their rate constants, or photolysis rates, which is given in Appendix A together with the relevant notes and references, and the dilution rate. The set of differential equations resulting from this reaction mechanism were integrated using the Gear algorithm [48,49]. Initial exploratory calculations were carried out on a Univac 1108 computer at Shell Research Ltd., Thornton Research Centre, Chester, U.K., using an implicit one-step method with single precision to solve stiff ordinary differential equations [50].

Reaction Mechanism

The reaction scheme used in our model calculations, along with associated references and notes, is given in Appendix A. The inorganic reactions [reactions (1)-(34)] are essentially identical to those used recently by Carter et al. [7] for the NO,-n-butane-propene-air system, apart from a minor updating of rate constants. Hence, the inorganic reaction scheme will not be discussed in any detail here; [7] and Appendix A should be consulted for further details.

In previous modeling studies [7,8] of alkane- and/or alkene-NO,-air irradiations done in the same chamber as the runs reported here, it was found that model simulations could not fit the data unless the model con- tained a provision for an unknown chamber radical source. This radical source, represented by reaction (34) in our model (see Appendix A), is attributed to photolysis of highly photoreactive contaminants (such as ni- trous acid) offgassing from the chamber wall [7], and its rate has been generally adjusted for each run in order to fit the observed hydrocarbon consumption rates. However, for the runs considered here, the uncertainties in the aromatic photooxidation mechanism (see below) are such that these chamber radical input rates should not be independently adjusted. Consequently, the chamber radical input rates were not treated as adjustable parameters in this study but were estimated based on our previous model simblations [7] of our alkane- and alkene-NO,-air irradiations (see note 8 of Appendix A). The problem of unknown chamber radical sources is discussed in detail elsewhere [7] and is not considered further here.

The organic reaction scheme is given in Appendix A as reactions (35)- (166); reactions (46)-(53) and (70)-(77) represent overall processes whose detailed reactions are given for aromatic systems in general in Appendix B. The general reactions listed in Appendix B were applied to both the toluene-NO,-air and the o-cresol-NO,-air mechanisms, and are intended to be applicable to the NO,-air photooxidations of all substituted benzenes

784 ATKINSON E T AL.

(see notes 14-19,26, and 27 of Appendix A and the discussion below). A number of possible alternative mechanisms, discussed below, were examined, and the appendices list all reactions considered, including those that were subsequently found to give unacceptable fits to the data. Throughout the discussion, unprimed reaction numbers [e.g., (35), (36), etc.] will refer to those in Appendix A, while primed reaction numbers [e.g., (l’), (2’), etc.] will refer to specific examples of the general reactions in Appendix EL

Results

Experimental Results

Initial conditions and selected reactivity parameters for the experiments reported here are given in Table I. In order to check reproducibility, several duplicate irradiations were carried out. In those cases, very similar reactivity parameters were observed (especially for runs carried out with the same solar simulator lamp [34,37], which was changed between runs EC-290 and EC-327), and product yields agreed to within their stated uncertainty limits (compare runs EC-264 to EC-266 and EC-327 and EC-340).

The products observed in the toluene-NO,-air runs for which quantitative results were obtained (in approximate order of yield) were CO, formaldehyde, PAN, benzaldehyde, o-cresol, m-nitrotoluene, acetylene, and methyl nitrate. Also detected 1471 were small amounts of m - and p - cresol, o- and p-nitrotoluene, benzyl nitrate, and 2,3-, 2,5-, 3,2-, 3,4-, and 4,3-hydroxynitrotoluenes. In the cresol-NO,-air runs the products detected were CO, PAN, and 2,3- and 2,5-hydroxynitrotoluenes from o-cresol; CO, PAN, and 3,2- and 3,4-hydroxynitrotoluenes from m-cresol, and CO, PAN, and 4,3-hydroxynitrotoluene from p-cresol. The organic product yields in the cresol runs were very low, and the PAN yields were much lower than those observed in the toluene runs. The maximum yields of the major organic products observed in the toluene-NO,-air and toluene-benzaldehyde-NO,-air irradiations are given in Table 11.

Figure 1 compares the concentration-time profiles observed for 0 3 , NO, NOz, PAN, and the reactant cresol in the o-, m-, and p-cresol-NO,-air runs. (The nonzero initial ozone values observed are believed to be due to a cresol interference on the UV absorbance 0 3 monitor.) In addition, concentration-time profiles for selected species observed in the o -cresol and in selected toluene and toluene-benzaldehyde runs are plotted in Figures 2-7, along with results of model calculations (discussed below). In general, it can be seen that the cresol runs and the added benzaldehyde runs have a lower overall reactivity than comparable toluene runs, and that rn-cresol is by far the most reactive of the cresol isomers. These results can be rationalized in terms of the mechanism which is discussed below.

Detailed data tabulations, giving concentrations of all organic and in-

TA

BL

E I.

Initi

al c

ondi

tions

and

sel

ecte

d re

activ

ity p

aram

eter

s fo

r th

e ir

radi

ated

tolu

ene-

NO

,-air,

to

luen

e-be

nzal

dehy

de-N

O,-a

ir,

and

cres

ol-

NO

,-air

mix

ture

s.

Run

Rea

ctan

ts

(ppm

)

Tol

uene

o-C

reso

l

m-C

reso

l

p-C

reso

l

Ben

zald

ehyd

e

NO

NO2

NO

X

Pn

ysi

cal

Par

amet

ers

kl

(min

-1)

Ave

. T

erp.

(K

) A

ve.

RH (X

)

Rea

ctiv

ity

Par

amet

ers

Max

. O3

(p

pd

Tim

e 0

max

. (m

in.)

a 6-

hour

val

ue g

iven

.

7

:C-2

64 -

1.15

6

0.42

0

0.05

6

0.47

0

0.35

304.

5

42

.0

0.41

9

210 - -

EC

-265

7

1.0

70

0.43

5

0.0

48

0.4

82

0.3

5

302.

6

54.0

0.3

93

210

- -

EC

-266

-

1.19

6

0.43

2

0.05

9

0.49

4

0.3

5

302.

8

47

.5

0.4

03

215

7

I_L

EC

-269

-

0.5

66

0.3

98

0.07

4

0.47

2

0.3

5

302.

9

46.5

a ,0

.298

,360

- E

C-2

71 -

1.14

6

0.18

6

0.0

29

0.21

5

0.37

303.

0

49.5

0.29

6

90

- E

C-2

73 -

0.58

7

0.09

6

0.01

4

0.11

0

0.37

30

3.6

49.5

0.21

5

80

- E

C-2

81 -

0.39

4

0.45

8

0.02

0

0.48

0

0.35

302.

9

48.0

a PO

. 064

>36

0

c_

EC

-289

-

0.29

7

0.40

4

0.0

45

0.44

9

0.39

303.

0

56.5

0.13

7

210

_I

-

EC

-290

-

0.36

4

0.4

13

0.06

7

0.4

81

0.39

30

3.1

52.4

a 20

.068

2360

II

-

EC

-327

0.5

73

0.35

7

0.09

6

0.44

8

0.4

0

303.

4

-60

0.37

6

-360

- -

EC

-337

-

0.95

9

0.17

2

0.32

2

0.12

4

0.44

8

0.39

303.

0

65.6

0.32

5

255

I_

EC

-339

-

0.53

6

0.18

7

0.3

41

0.1

02

0.44

5

0.39

303.

4

64

.0 a

>0.

225

1360

-

3C-3

40 -

0.53

7

0.3

33

0.09

6

0.4

30

0.39

30

3.1

63.2

0.34

6

360

-

TA

BL

E 11.

Obs

erve

d an

d ca

lcul

ated

a max

imum

pro

duct

yie

lds

in th

e to

luen

e an

d to

luen

e-be

nzal

dehy

de-

NO

,-air

runs

.

RUn

Pro

du

cts

(ppb

)

EC-2

66b

EC-2

69

EC

-271

Obs.

Cal

c.

Obs

. C

alc.

O

bs.

Cal

c

Aro

mat

ic

Ben

zald

ehyd

e

o-C

reso

l

m-C

reso

l

p-C

reso

l

o-N

itro

tolu

ene

m-N

itro

tolu

ene

p-N

itro

tolu

ene

C H

yd

rox

yn

itro

tolu

enes

Ben

zyl

nit

rate

'

PBzN

23

18

NCf

NC NC

0.2

NC

NC

14

9 60

22c

9c

2c

2c

OC

3c

4c _-

-- --

gd

7 1' 0.7'

1' lC

0.9'

6.7'

0.8'

--

--

Oth

er O

rgan

ic N

itra

tesg

yh

I --

12

10

NC

NC

NC 0.1

NC

NC

lo

9 La

17

6

13

--

--

0

.1

-- --

-- -- --

--

(e)

11

12

-_

NC NC

-- N

C

2 0.

2

-_

NC

__ N

C 10

39

-_

__ __

18

14

NC

NC

NC

<O.l

NC

NC

8 3 35

(e )

5 8

-- N

C

--

NC

-- NC

0.7

0.1

--

NC

--

NC

--

7

--

35

EC

-273

Obs

. C

alc

16d

10

38

0.2

NC

50.6

NC

0.6

NC

0.3

0

.3

0.3

N

C

NC

4 2 21

-_

__ --

-- -_

4

s

1 -- 33

I --

44

Fra

ymen

tati

on

PAN

Oth

er p

erox

y n

itra

tesi

Gly

oxal

Met

hyl

gly

ox

al

2-B

uten

e-l.

4-di

al

For

mal

dehy

de

Met

hyl

nit

rate

'

Ace

tyle

ne'

coj

75

91

-- 67

18

21

--

75

75

31

1

0.6

500

202

--

NC

-- -_

so

52

56

--

19

lo

39

40

26

1

0.6

-300

21

1 --

NC

__ __ __

41

61

--

46

15

8

__ -- -- 57

65

26

0.6

0.

5

(ef

2.4

NC

24

32

32

--

14

--

4 __ -_

33

48

19

-- 0.

4

(e)

1.0

NC

42

47

51

-- 1

6

__

8 36

__ __ 48

23

(el

0.5

0.5

1.7

NC

a C

alcu

late

d us

ing

mod

el A

. Y

ield

s giv

en a

re m

axim

um c

alcu

late

d at

the

tim

e of

com

plet

e NO

2 co

nsum

ptio

n,

Dup

licat

e ru

ns d

one

gave

eith

er le

ss w

ell c

hara

cter

ized

pro

duct

dat

a, o

r dat

a th

at w

ere

cons

iste

nt w

ith re

sults

Onl

y se

miq

uant

itativ

e G

C-M

S da

ta [

47] a

re a

vaila

ble.

or a

t 6 h

r, w

hich

ever

cam

e fi

rst.

give

n he

re.

d C

hrom

atog

raph

ic c

olum

n em

ploy

ed c

ould

not

sep

arat

e be

nzal

dehy

de a

nd b

enzy

l ni

trat

e, s

o re

port

ed d

ata

are

benz

alde

hyde

+ be

nzyl

nit

rate

. Pr

esen

t as

a re

acta

nt in

this

run.

N

C =

not

cal

cula

ted;

assu

med

to b

e ne

glig

ible

in

the

mod

el.

g R

eact

ions

con

sum

ing

this

pro

duct

wer

e no

t inc

lude

d in

the

mod

el, s

o th

e ac

tual

cal

cula

ted

yiel

ds s

houl

d be

sm

alle

r. Pr

imar

ily C

70N

02 fo

rmed

in r

eact

ions

(46)

-(53

). Pr

imar

ily H

CO

CH

=CH

C03

N02

fo

rmed

in r

eact

ion

(112

). J B

ackg

roun

d le

vels

sub

trac

ted.

788 ATKINSON ET AL.

% OZONE % CRESOL *

A It

* A

Q 0 m

m ' 0 I

120 zuo 360

NITRIC OXIDE =! P A N I

b Q 0

- rn Q Q x I Q i L r n 5 N

u A z 0

Q 0

* 0

A

u o 0

0 360

NITROGEN DIOXIDE

" 0 120 240 360

TIME (MINI Figure 1. Concentration-time profiles observed in cresol-NO,-air runs. (*) o cresol, run EC-281; (0) m-cresol, run EC-289; ( A ) p-cresol, run EC-290.

organic species monitored, and values of measured physical parameters as a function of time, including the spectral distribution of the photolyzing light, are available from the authors upon request [51].

Results of Model Calculations

Calculations were carried out with varying mechanistic pathways and rate constant ratios. The various mechanistic options, which are described in more detail in the discussion, concern (1) the reactions of the species formed by the addition of hydroxyl radicals and then 0 2 to the aromatic ring, (2) the reactions of the predicted conjugated y-dicarbonyl products, (3) the relative importance of the chain-terminating formation of organic

NO,-AIR PHOTOOXIDATION OF TOLUENE AND THE CRESOLS 789

e e

e n e

. .

nitrates from the ROz + NO reaction [e.g., reactions (38) and (4')], (4) the photolysis of benzaldehyde, and (5) the reaction of o-cresol with NO3 radicals. The varying assumptions made concerning these options and the designations given to the calculations using the different sets of assumptions are given in Table 111.

Figures 2 and 3 show the effect of using different alternate mechanisms concerning the reactions of the OH-aromatic-02 adduct and the reactions of the y-dicarbonyl products, compared with the experimental data for a representative toluene run and the o-cresol run; and Figure 4 shows the effect of varying the relative importance of the formation of nitrate from

TA

BL

E 111.

Sets

of m

echa

nist

ic o

ptio

ns e

xam

ined

in m

odel

cal

cula

tions

and

thei

r des

igna

tion

sym

bols

.

Mec

han

isti

c O

pt io

ns

Tri

cy

cli

c

Rad

ical

Cg

HgC

HO

k

(N0

3+

cres

ol)

D

esig

nat

ion

A

rom

OH

-02

Aro

nrO

H-O

2 RC

OCH

=CH

C03

Dec

om

po

siti

on

k(

C70

2+N

0+ c

70N

02)

Ph

oto

lysi

s Sy

mbo

l A

dduc

t Fo

rmed

C

ycl

izat

ion

R

ecy

cliz

atio

n

Cy

cliz

atio

n

Mec

hani

sm

k(C

,02+

NO

) M

echa

nism

(c

c-m

olec

ule-

1 se

c-1)

Non

-Rad

ical

Non

-Rad

ical

Non

-Rad

ical

Non

-Rad

ical

Non

-Rad

ical

Non

-Rad

ical

Non

-Rad

ical

Non

-Rad

ical

Non

-Rad

ical

Non

-Rad

ical

Rad

ical

No

Ph

oto

lysi

s

Non

-Rad

ical

1.5

x

1.5

x

10-l

'

1.5

x

lo-''

1.5

x

1.5

x

LO-''

1.5

x

1.5

x 10-l'

1.5

x

1.5

x

1.5

x

1.5

x

lo-''

1.5

x 0

a 3-

Add

uct:

k4

5 >>

k4

4 (t

olue

ne sy

stem

); k6

9 >>

k6

8 (c

reso

l sys

tem

); 1

-add

uct:

kp

4 >>

k4

5, k

s8 >>

k69

. R

efer

s to

stru

ctur

e of

bic

yclic

ring

syst

em. [

2.2.

2]:

k; >>

k;;

[3.2

.1]:

k; >>

k;.

Y

es:

ki6 >

> k$[

NO

] or k

;7 >>

k,[N

O] o

r k

;8 >>

k;,[

NO

]; N

o:

k;6

, k

;7, k

;8 n

eglig

ible

. Y

es:

kloe

>> (

kllo

[NO

) + kl

~z[N

Oz]

) and

klos

>> (

klll

[NO

] + k1

13[N

0~])

; No:

kl

oa, k

log

negl

igib

le.

NA

= N

ot a

pplic

able

. In

term

edia

tes

invo

lved

not

form

ed in

this

set o

f opt

ions

. N

E =

No

effe

ct.

The

sam

e in

term

edia

tes o

r pro

duct

s ar

e fo

rmed

rega

rdle

ss o

f the

ass

umpt

ions

mad

e

NO,-AIR PHOTOOXIDATION OF

s OZONE , TOLUENE AND THE CRESOLS 791

S mTno CRESOL

2 WllRDGEN OIDXIOE

T I M E I t t l N I

Figure 3. Experimental and calculated concentration-time profiles for reactants and products in o-cresol-NO, run EC-281. Effect of alternate mechanisms for OH-aromatic adduct fragmentations. (*) Experimental data; (-) calculated using mechanisms A-H.

the RO2 + NO reaction. From these figures it can be seen that best fits are obtained if it is assumed that the aromatic-OH-02 adducts undergo cyclization but not recyclization (see Discussion), and that the C7 bicyclic peroxy radicals formed in the toluene and cresol photooxidations react with NO -25% of the time to give nitrates. Both calculations A (assuming exclusively 3-addition of 0 2 to the aromatic-OH adduct, [2.2.2] cyclization of this aromatic-OH-02 adduct, and no cyclization of the RCH=CHCOB radicals formed from the y-dicarbonyl products) and calculations F ([3.2.1] cyclization of either the 1 - 0 2 - or 3-02-aromatic-OH adduct, and rapid cyclization of RCH=CHC03) fit the data reasonably well, and the predictions of these two mechanisms are compared with the data from other toluene and toluene-benzaldehyde runs in Figures 4-7. In general, calculations A predict an initial reactivity which is slightly high, while calcu-

792 ATKINSON E T AL.

O Z O N E O R l M C C R E S U L

T I M E 1MlNl

Figure 4. Experimental and calculated concentration-time profiles for selected reactants and products in toluene-NO, run EC-266. Effect of varying the efficiency of alkyi nitrate formation from ROz + NO from 10% (A10) to 50% ( A ~ o ) . (*) Experimental data; (-, - - -) calculated.

lations F predict a slightly low initial reactivity, although in both cases the calculated initial reactivity can be made to fit experimental results by respectively reducing or increasing the adjustable radical input rate from chamber sources [7] (see note 8 of Appendix A).

Figure 5 shows the effect on model calculations of varying assumptions concerning benzaldehyde photolysis for a toluene-benzaldehyde run. It can be seen that in order to fi t the benzaldehyde concentration-time profiles and the overall reactivity, it is necessary to assume that benzaldehyde undergoes photolysis at a significant rate, but that it forms nonradical products.

Figure 8 shows the effect that a reaction of NO3 with o-cresol has on the 0 3 , Nos, and o-cresol concentration-time profiles in the o-cresol run and


TOLUENE FORnRLDEHlOE

. BENZRLOCHYOE I IETHYL N I l R R l E

_ - - , ,

, 6 0 Bo , i o , b O rba ria 3 i a d o 60 I20 110 2 " l YO0 160

T I M E l l l lN1

Figure 5. Experimental and calculated concentration-time profiles for selected reactants and products in toluene-henzaldehyde-NO, run EC-337. Effect of assuming 100% efficiency of radical production in henzaldehyde photolysis (AR), and effect of neglecting henzaldehyde photolysis (ANP). (*) Experimental data; (- , - - - ) calculated.

in a toluene-benzaldehyde run. (Predicted concentration-time profiles of other monitored reactants and products are less affected by this reaction). It can be seen that the data are much better fit by assuming that the reaction occurs with a rate constant on the order of (1-2) X cm3/molecule sec than by assuming that it is negligible.

Table I1 lists the experimental and calculated (using model A) yields of the major products monitored, and gives the maximum calculated yields of major products predicted by the model which were not experimentally monitored. Among the products monitored, the major discrepancy appears to be that the observed nitrotoluene yields are higher by a factor of -10 than those calculated, despite the fact that the calculations generally fit the cresol

794 ATKINSON ET AL.

RUN 269 R U N 271

OZUNE OZONE

T I H E I M I N I

Figure 6. Experimental and calculated concentration-time profiles for selected reactants and products in toluene-NO, runs EC-269 and EC-271. (*) Experi- mental data; (-, - - -) calculated.

data and used the experimentally determined ratio (k42/k43) for cresol versus nitrotoluene formation obtained by Kenley et al. [19]. In addition, the predicted formaldehyde yields are slightly low, relative to the experimental data. This may be due to the neglect of glyoxal photolysis in our mechanism (see note 29 of Appendix A). For the other major products, the fits of calculated to experimental yields can be considered to be within the range of the experimental uncertainties. Although the model predicts the formation of products which were not experimentally observed [mainly C7 organic nitrates, a-dicarbonyls, 2-butene-1,4-dial, peroxybenzoyl nitrate (PBzN), and nitrophenols], we have no evidence that these species are not formed in the yields predicted. Indeed, the low carbon balance observed for these systems suggests a significant formation of unmonitored products,

NOx-AIR PHOTOOXIDATION OF TOLUENE AND T H E CRESOLS 795

RUN 2 7 3 RUN 339 OZONE O Z O N E

c 0 1 /' . r _ - -

P R N PRN

T I M E I H I N I

Figure 7. Experimental and calculated concentration-time profiles for selected reactants and products in toluene-NO, run EC-273 and in toluene-benzaldehyde-NO, run 339. (*) Experimental data; (-, - - -) calculated.

and thus, except for the nitrotoluene yields (a very minor product), our model simulations can be considered to be generally consistent with the experimental data reported here.

Discussion

Initial Toluene Reactions

The only significant chemical loss process of toluene under simulated polluted atmospheric conditions is now known to be by reaction with the OH radical [8,9,11-15,17,19-211. This reaction proceeds via two pathways: H atom abstraction from the substituent methyl group [reaction (35)], and OH radical addition to the ring [reaction (41)] [8,9,12,15,17,19,20,28-331,

796 ATKINSON ET AL.

OZONE ;1

RUN 281 RUN 337

=! OZONE

N I T R O G E N D I O X I D E x NITROGEN D I O X I D E

% ORTHO CRESOL 0

0 N

O

X 0 0

0 120 zvo 360 0-1

N

0

I c

2 ORTHO CRESOL

;̂1 / ' A N N - - -

/

?Ks, 0 0

0 120 zvo 360 0

T I M E ( M I N I Figure 8. Experimental and calculated concentration-time profiles for 0 3 , NOz, and o-cresol in the o-cresol-NO, run EC-281 and in the toluene-benzaldehyde- NO, run EC-337. Effect of neglecting the NO3 + o-cresol reaction (A"). (*) Experimental data; (-, - - -) calculated.

(35 ) OH + 0 -- @ + H,O

with k35/(k35 + h41) N 0.15 [9,15,19,20] and (1235 + 1241) N 6 X cm3/ molecule sec [9,12,13,15] at room temperature. Under atmospheric conditions the benzyl radical will react to form benzaldehyde and benzyl nitrate


[reactions (36)-(40)], as has been discussed previously [8,9,19,20,28-321. The reasonably good fits of the model calculations to the experimental benzaldehyde yields (see Table 11) indicate that our data are consistent with the reported ratios [9,15,19,20] for OH abstraction versus addition to the aromatic ring.

OH Radical Addition Pathway: Formation of Cresols and Nitrotoluene

The major pathway (-85%) for reaction of OH radicals with toluene is via the initial formation of an OH-toluene adduct [9,15,19,20]. By analogy with the position of O(3P) atom addition to toluene [52-541 and based on the available experimental product data [19,20,28,30-331, the OH radical is presumed to add primarily at the ortho position. The thermalized adduct (I) can react under atmospheric conditions with 0 2 via two pathways [9]: H atom abstraction to form the cresol [reaction (42)], or reversible addition to form the peroxy radical (11) [reactions (44) and (45)]:

CH,

(42)

(44,45)

r & + HO,

(11)

or it can react with NO2 to form the nitrotoluene:

Hendry and co-workers [19,20] concluded from their discharge flow study of the reaction of OH radicals with toluene that the cresol isomers and rn-nitrotoluene [i.e., reactions (42) and (43)] are the sole initial products of the OH radical addition pathway, and that k 4 3 / k 4 2 = (4.4 f 0.5) X lo3. If the discharge flow results [19,20] are applicable to the present experimental study, the only significant initial products of toluene photooxidation would be benzaldehyde and the cresols, with the latter being formed -85% of the time. Because the cresols are consumed more rapidly in the NO,-air irradiations than toluene, owing to their higher rate constant for reaction with the OH radical [9,16,18], this would mean that the chemistry of the toluene-NOx system would be dominated by that of the cresol-NO, system. However, the present smog chamber results indicate that in terms of ozone and PAN formation and NO to NO2 conversion rates, the cresol-

798 ATKINSON ET AL.

NO,-air systems are far less reactive than is the toluene system. The excess reactivity of the toluene system cannot be attributed to benzaldehyde reactions, since, as is discussed below, benzaldehyde acts as .an inhibitor [55-571. Thus, the initial formation of products other than these is necessary to account for the relatively high reactivity of toluene runs, and for the high yields of products such as PAN, as well as for the observation of biacetyl [35,36] and glyoxal and methylglyoxal [36] as initial products in o-xylene-NO,-air irradiations. We find that the o-cresol data are best fit by assuming that it is formed to the extent of only -20% of the total OH + toluene reaction, with products other than cresol, benzaldehyde, benzyl nitrate, and nitrotoluenes being formed -65% of the time.

The apparent inconsistency in terms of predicted cresol and nitrotoluene yields between the results of the discharge flow studies [ 191 and our smog chamber studies may be due to the fact that the former was carried out at a total pressure of -6-15 torr of (Ar + 0 2 ) [19] against a total pressure of 1 atm of air in the smog chamber studies. 0 2 addition to (I) to form (11) and the reaction of (I) with NO2 to form nitrotoluenes both may not have been in the limiting high-pressure second-order kinetic regime under the conditions of the discharge flow study [19], and thus could be significantly faster a t 1-atm total pressure. This is indeed assumed to be the case for the addition reaction of (I) with 0 2 , but the model still uses the low-pressure-derived value for the (I) + NO2 addition reaction [reaction (43)]. However, the fact that the model significantly underpredicts nitrotoluene yields suggests that h 4 3 is pressure dependent as well.

Structure and Reactions of the OH-Aromatic-02 Adducts

be at the 1,3, or 5 position: The structure of the radical (11) is not known; addition of 0 2 to (I) may

(44)

NO,-AIR PHOTOOXIDATION OF TOLUENE AND T H E CRESOLS 799

0 2 addition at the 5 position should be the least thermochemically favored, since the double bonds are nonconjugated. Addition a t the 1 position [reaction (44)] is estimated [58] to be slightly more favored thermochemically over addition to the 3 position [reaction (45)], although the resulting radical (IIa) will also have the most steric hindrance. In the calculations the consequences of assuming exclusively l addition or exclusively 3 addition are both considered.

Based on thermochemical estimates [15,58], 0 2 addition to radical (I) is expected to be only -10 kcal/mol exothermic. This means that significant back decomposition of radicals (11) to radical (I) may be occurring, which could account for the relatively low effective rate constant for 0 2

addition to radical (I) required to fit the cresol data (see note 13, Appen- dix A).

A t the present time, little is known about the other reaction pathways of the OH- aromatic - 0 2 adducts. There are a number of possible alternative pathways, and these are listed for a general aromatic adduct in Ap- pendix B and are discussed below for the case of adduct (IIb) (substituent 6 in Appendix B = CH3, substituents 1-5 = H). Reactions of the other adducts are assumed to be analogous.

If the OH- aromatic - 0 2 adducts react in a manner analogous to other peroxy radicals, the following scheme would appear to be the most reasonable:

( IIh) followed by either reaction with 0,

or ring opening &) P H - C H t -% HCOC(CH,)=CH-CH=CH-CHO + HO,

0.

However, this scheme is inconsistent with the observation of a-dicarbonyls as primary products in the NO,-o-oxylene-air system [35,36], and no reasonable mechanism accounting for those observations [35,36] could be derived without assuming formation of bicyclic radicals. It is thus assumed

800 ATKINSON E T AL

that the OH- aromatic - 0 2 adduct primarily undergoes cyclization reactions such as those shown below:

(7')

(IIIC)

In addition to accounting for the ring cleavage products observed, these reactions are expected to dominate on the basis of thermochemical and kinetic estimates. If we assume that the rate constant for the reaction of the peroxy radical (IIb) with NO is -8 X cm3/molecule sec (see note 9 of Appendix A), then under realistic atmospheric conditions [[NO] < 1 ppm (2 X 1013 molecule/cm3)] cyclization should dominate if its rate constant is >160 sec-'. If we further assume that (1) the A factors for the cyclization reactions are at a minimum sec-l[58], (2) the activation energy for cyclization is [S9 kcal/mol (based on the activation energies of bimolecular radical + double bond addition reactions [59]) + the reaction endothermicity (if any)], and (3) AH (cyclization) (ring strain -14) kcal/mol, based on thermochemical calculations using group additivity principles [58], then cyclization should dominate if the ring strain of the bicyclic radicals is less than -16 kcal/mol. Although such ring strains are not known, this would appear to be likely [58].

Of the radicals (1IIa)-(IIIc), (IIIa) will probably have the least ring strain, while (IIIb) will have some resonance stabilization energy by virtue of its allylic character. Since we cannot determine, a priori, which of these structures is favored, it is again necessary to consider both options in the model calculations, i.e. (IIIa) and (IIIb).

These cyclized aromatic-OH-02 adducts are then expected to add 0 2

to form radicals of the type:


(IV) (Va) ( V h )

which could react with NO to either form an alkoxy radical or yield a stable nitrate (analogous to the >-C4 alkylperoxy cases [7,60], for example:

( VII )

The reactions of the bicyclic peroxy organic nitrates such as (VII) are unknown; they may undergo molecular decomposition to form HN03 or other organic nitrates. In our model we assume that their reactions, if any, have a negligible effect on the overall mechanism and are thus ignored. The bicyclic peroxy alkoxy radicals such as (VI) are expected to undergo a series of highly favorable 0-scission fragmentations, resulting in the formation of a-dicarbonyl and conjugated y-dicarbonyl products. For example, (VI) is assumed to decompose as shown, with the other isomers, resulting from the reaction of NO with (Va) and (Vb), reacting analogously:

HO 0. I I

CH-C--CHJ - CH,COCHCHCH=CHCHO -+

(VI) I ( 5’) H CH ,COCHOH + HCOCH=CHCHO

( 6’) CH,COCHO + HO,

The relative importance assumed for nitrate formation versus alkoxy radical formation has an important effect on model simulations, since the former process is radical terminating and is an effective nitrogen sink, while the latter is radical propagating. This effect is shown in Figure 4, where

802 ATKINSON ET AL

k4fl(k3t + k 4 r ) (and also the analogous rate constant ratios in the cresol system and for the benzyl peroxy radical) is varied from 0.1 to 0.5. It can be seen that toluene-NO,-air data are best fit by assuming that nitrate formation occurs -25% of the time; this is true for the toluene-benzaldehyde-NO,-air and the cresol-NO,-air systems as well. Assuming that nitrate formation is negligible results in a significant overprediction of ozone maxima regardless of the other mechanistic options used.

A possibility, which must also be considered, is that the bicyclic peroxy radicals may further cyclize, such as shown below 0 2 adduct to radical (IIIb)]:

[for radical (IV) and the

If formed, these doubly cyclized radicals should react with 0 2 and then NO, forming alkoxy radicals or alkyl nitrates by reactions analogous to those shown above for (IV). The tricyclic bridged-peroxy alkoxy radicals are also expected to rapidly fragment via consecutive @-scissions:

0 O O H O H 11 HO 0.

I I I

- (CHO), + HCOCHCCHO

CH, OH I

H C ~ H C O C H ~ + HCO “PP’ / HCO~HCCHO

I bHJ ‘m

HCOCHOH + CH,COCHO (VIII) I

k ( C H O ) , + HO,


The relative importance of the two possible modes of decomposition [e.g., reactions (31’) and (33’)] of species such as (VIII) is not known, so both possibilities are considered.

As can be seen from Figures 2 and 3, the data do not clearly indicate which aromatic-OH-02 adduct is formed, or whether it undergoes [2.2.2] or [3.2.1] cyclization, but they do indicate that recyclization probably is negligible. If recyclization occurs, the assumption made concerning the fragmentation of radical (VIII) is important [Assuming that reaction (33’) dominates results in overprediction of reactivity in the toluene and the o -cresol systems, while assuming that reaction (31’) dominates results in the reactivity in the toluene system being underpredicted, but gives reasonably good fits to the o-cresol system.] It is clear, however, that no ad- justment of the k311lk33’ ratio would allow both the toluene-NO, data and the o-cresol-NO, data to be fit by models assuming recyclization. Model A assumes [2.2.2] cyclization and predominant addition of 0 2 to the 1 position of the OH-aromatic adduct, although 3 addition should not be completely negligible. If [3.2.1] cyclization occurs, as assumed in the other model which gives reasonable fits to the data (F), formation of the 1 adduct or the 3 adduct is immaterial, since each adduct cyclizes to form the same intermediate [i.e., radical (IIIb)].

The aromatic ring-opening mechanisms that best fit the toluene and o-cresol data, when applied to the o-xylene system, yield biacetyl after OH radical addition a t the 1 or 2 position; for example,

&H & [ CHBCOCOCHj H_ OH + 0 2 -.+ OH -% $- CHOCH=CHCHO

This is, as noted previously, in accord with the recent observation of significant yields (18 f 4%) of biacetyl from irradiated NO,-o-cresol-air systems [35], and the magnitude of the biacetyl yield is then consistent with the mode of O(3P) atom addition to o-xylene [53].

Benzaldehyde Reactions

The reaction of benzaldehyde with OH radicals must, because of the magnitude of the OH radical rate constant [56], proceed predominately via H atom abstraction from the substituent CHO group [8,9,20] to form the benzoyl radical:

CHO CO I I

(54) OH + @) -+ H,O + 0

804 ATKINSON ET AL.

This CGH5CO radical is then expected to add 0 2 [reaction (56)] to form the peroxybenzoyl radical which can react with NO or NO2. Reaction with NO2 [reaction (57)] will lead to the formation of peroxybenzoylnitrate (PBzN), which was not observed, but which may have been formed and not detected. PBzN thermally decomposes [61] with an analogous mechanism to PAN [62-641, to lead to the ultimate formation of, among other products, o- and p-nitrophenols [reactions (58)-(65)] [57]. Note that in each case the ultimate result of the OH-benzaldehyde reaction is radical termination, which is consistent with these and previously reported observations [55- 571.

While little is known about benzaldehyde photolysis under atmospheric conditions, its photolysis at rates approximately five times greater than that used for the aliphatic aldehydes [7] is included in the mechanism to account for the observed benzaldehyde loss rates in the toluene-benzaldehyde runs (see note 20 of Appendix A). As noted previously (see Fig. 5), the data are best fi t by assuming that benzaldehyde photolysis results in the formation of nonradical products. The most likely pathway would thus appear to be formation of benzene + CO, but although small yields of benzene (-0.5-1 ppb) are indeed observed in the toluene-benzaldehyde runs, the yields are far less than the amounts predicted (50-60 ppb) if it were the major product of benzaldehyde photolysis.

Cresol Reactions

The cresols, represented in this model by o -cresol alone (substantiated to a large extent by the smog chamber data reproduced here), are known to react rapidly with OH radicals [9,16,18], with this reaction being a major cresol loss process under simulated atmospheric conditions [HI. The slow reaction with 0 3 [18,20] has been neglected in the present model since it will contribute 15-10% of the total cresol loss rate under the experimental conditions employed in the NO,-toluene-air and NO,-o -cresol-air irradiations. The irradiated NO,-o- cresol-air data yield little evidence as to the reaction mechanisms-very little PAN is formed, and the major organic products observed are hydroxynitrotoluenes in low (<5%) yields. The reaction of OH radicals with o-cresol can proceed via OH radical addition to the ring or by H atom abstraction [9,16].

The abstraction pathway, which accounts for -8% of the total OH radical reaction [9,16], probably proceeds mainly via H atom abstraction from the substituent-OH group, since the abstraction rate constant for o-cresol is a factor of -4 higher a t 400 K than that for toluene [15,16].

p 3 $a

OH + & -+ H,O + &"


Reactions of NO2 with radical (IX) can, by analogy to the phenoxy radical reaction (64) [57], lead to formation of the observed hydroxynitrotoluene isomers:

This minor H atom abstraction pathway has been neglected in the model calculations since its inclusion would have an insignificant effect on the calculated time-concentration profiles of the major species.

The major OH radical reaction pathway is via addition to the aromatic ring [9,16], and thermochemical calculations [58] show that radical (X) is the most thermochemically favorable adduct. Its formation is thus assumed to be the sole pathway:

(65) OH + 0 (XI

Species such as (X) can also react with NO2 to give rise to the observed hydroxynitrotoluene isomers:

-- -H,O 0 0 2

(IX)

However, the dominant reaction of (X) is expected to be with 02:

The subsequent reactions of radicals (XI) are assumed to be analogous to those of radical (11) formed in the toluene system. The various possible

806 ATKINSON ET AL.

detailed reactions for these species are listed in Appendix B (substituent 1 = OH, 2 = CH3 for the l-adduct; substituent 1 = OH, 6 = CH3 for the 3-adduct), and the overall processes are given in Appendix A as reactions (70)-(77). The best-fit models predict that the major overall processes are as shown below:

HO, + CH,COCO,H + HCOCH=CHCHO or

HO, + 1/2(HCOCO,H + CH,COCH=CHCHO) + 1/2(CH,COCO,H + HCOCH=CHCHO)

I,( There is evidence that there is an additional significant loss process for

cresols in NO,-air systems. It has been independently observed by O'Brien [65] and in our laboratories [66] that although the cresols do not react at significant rates with either 0 3 [18,20,65] or NO2 [65,66], they do appear to rapidly disappear in chambers dosed with NO2 and 0 3 together. Since O3 and NO2 react rapidly to form NO3 and N205 [67],

(5) 0 3 + NO2 + NO3 + 0 2

these observations can be attributed to rapid reaction of cresols with either NO3 or N205. Since NO3 is known to react with alkenes [68], it is probable that it is also NO3 which is reacting with the cresols. We have recently obtained, using a relative rate technique, a rate constant for the reaction of No3 radicals with o-cresol of -1.5 X 1O-I' cm3/molecule sec [66]. As can be seen in Figure 8, this rate constant is sufficiently high that reaction with NOs becomes a significant sink for o-cresol after 0 3 formation begins, and accounts for the rapid decrease in the o-cresol concentrations observed after their initial rise in the toluene-NO,-air runs. The observation that

NOr-AIR PHOTOOXIDATION OF TOLUENE AND THE CRESOLS 807

NO3 radicals react rapidly with phenol and the cresols, but only very slowly with toluene and methoxybenzene [66], indicates that the reaction with phenol and the cresols proceeds via H atom abstraction from the substituent -OH group to form radical (IX) [66]:

Hence, this reaction is another explanation for the formation of the hydroxynitrotoluene isomers observed.

As noted previously, the order of the overall reactivity of the cresol isomers in NO,-air systems is meta >> para > ortho (see Fig. 1). This effect is larger than can be attributed to the relatively small difference in the OH + cresol rate constants [9,18] (meta/ortho = 1.4; para/ortho = l.l), but it can readily be rationalized in terms of the mechanism. On the basis of the mechanisms which best fit the toluene-NO, and the o-cresol-NO, data, m- and p-cresol are expected to react as shown:

1

&OH 4 OH

&; -

OH or

00 &; H -

CH I R

O?, NO -4

HO, + HOCWHO +

HCOC=CHCHO or

CH3COCH= CHCHO

HOCOCHO 9 +oH-HoQoH 4 @to.= + OH HO OH

CHOC02H is expected to be more reactive toward OH radicals than the CH3COCOzH formed in the o-cresol system, since it has the labile -CHO

808 ATKINSON ET AL.

group. In addition, both 2-methyl-2-butene-l,4-dial and 2-pentene- 1,4-dial are expected to give rise to methylglyoxal, a powerful photoinitiator, in their photooxidation mechanism [see Appendix A, reactions (100)-( 141)], which is not formed from the 2-butene-1,4-dial predicted in the o-cresol mechanism [the latter instead forming glyoxal, which is believed to have far lower quantum yields for photodecomposition (see note 29 in Appendix A)]. Furthermore, in the m-cresol system, the OH radical may add to a significant extent at the 2 position, since the radical formed is stabilized by both the OH and the -CH3 groups, unlike analogous adducts for the other cresol isomers. Applying our mechanism, this adduct is expected to give rise to methylglyoxal directly, which may explain the far greater reactivity of m-cresol relative to the 0- and p-isomers.

Subsequent Reactions of Ring-Opening Products

The subsequent reactions of the ring cleavage products formed in the toluene and the cresol photooxidations are given in Appendix A. In many cases the mechanisms are uncertain and rate constants have had to be estimated; these are discussed in the associated notes. The following points concerning the reactions of these products should be noted:

In order for the model to fit the smog chamber data, it must include the formation of radical initiator(s) in the toluene photooxidation mechanism which are not formed in the o-cresol system. The major initiator appears to be methylglyoxal, whose photolysis is also primarily responsible for the observed production of PAN, HCHO, and CH30N02. No other product predicted by our mechanism is believed to photodecompose to give radicals as rapidly as methylglyoxal (see notes 20,29, and 33 in Appendix A for brief discussions of benzaldehyde, glyoxal, and conjugated y-dicar- bony1 photolyses, respectively).

There is a significant uncertainty in the photooxidation mechanism of the conjugated y-dicarbonyls which are predicted to be formed as major products. These products are expected to react predominately with OH radicals, giving rise to unsaturated a-carbonyl peroxy radicals:

(1)

(2)

OH + RCOCH=CHCHO - H20 + RCHOCH=CHCO RCOCH=CHCO + o2 % RCOCH=CHCO~)

(XI11 These radicals could react with NO and NO2 [e.g., reactions (110)-(113)], in a manner analogous to the reactions of acyl peroxy radicals, forming unsaturated PAN analogs such as HCOCH=CHC03NO2. However, these unsaturated peroxy radicals may also undergo cyclization:

RCO 6H

RCOCH=CHCO,* - >C’ ‘CO LO-0’ (XQ


resulting in the ultimate formation of glyoxal (R = H) or PAN + methylglyoxal (R = CHB) (see Appendix A). It is not known at the present time whether this cyclization is rapid enough to compete with reaction of (XII) with NO or NO2. The recent observation of methylvinylketone and methacrolein formation during the early stages (before significant ozone formation) of the NO, photooxidation of isoprene [69] indicates that in the case of the unsaturated alkyl peroxy radicals, (XIII), cyclization is negligible compared to reaction with NO:

C H 2 4 R - E* '--CH,OH (R= H or CH; R' =CH3 or H)

(xm)

However, it is not clear that the unsaturated acylperoxy radicals such as (XII) formed in this system will react analogously, and hence it is necessary to consider both mechanistic options.

As can be seen in Figures 2 and 3, where results of calculations that assume cyclization of (XII) is rapid are compared with those that assume that it is negligible (compare model A with B, C with D, or E with F), the uncertainty concerning the rate of this cyclization has a significant impact on model predictions concerning both the toluene and the cresol systems. It is this uncertainty which does not allow us to distinguish unambiguously whether these data are best fit by assuming l2.2.21 cyclization and predominant 3 addition of 0 2 to the OH-aromatic adduct (calculations A) or by assuming that [3.2.1] cyclization dominates (calculations F), since the former assumes that cyclization of (XII) is negligible, while the latter assumes that it dominates. If the rate of this cyclization were known, one of these mechanisms could probably be eliminated.

Conclusions

The chemical kinetic computer models for irradiated toluene-benzaldehyde-o -cresol-NO,-air mixtures, which are discussed and are consistent with available basic kinetic and mechanistic data and with reasonable thermochemical and kinetic estimates, have been shown to be capable of giving predictions of reactant and major and minor product concentration-time profiles which are in reasonably good agreement with results of well-characterized smog chamber irradiations performed with a variety of initial reactant concentrations. However, with the information available to date, it is not possible to choose between various optional sets of equally reasonable reaction sequences. For example, the experimental data appear to be reasonably well fit by the two sets of mechanistic options discussed. In addition, in our initial studies it was found that most of the concentration-time profiles, except for those of 0-cresol, could be equally well fit by

810 ATKINSON ET AL.

a model giving the same products as A, but assuming that organic nitrate formation from ROe + NO occurs -40% of the time instead of -25%, pro- viding that fragmentation is increased by decreasing the formation of o- cresol from -20 to -8% of the overall reaction, and that the resulting increased production of PAN is compensated by changing the ratio K(CH3C03 + NO)/h(CH3C03 + NOz) from 1.5 to 2.0, which is within its range of uncertainty [7].

These chemical kinetic models predict the formation of products and reactive intermediates in aromatic-NO, -air photooxidations which have not yet been detected in such systems, and whose atmospheric chemistry is a t best poorly characterized. Such products predicted in the toluene system include the a-dicarbonyls glyoxal and methylglyoxal, and the y-unsaturated dicarbonyls RCOCH=CHCHO (R=H, CH3) and possibly the corresponding peroxyacyl nitrates RCOCH=CHCO3N02, together with the alicyclic nitrate which is proposed to be formed from the OH- toluene-02 adduct and NO; for example,

Furthermore, quantitative rate and mechanistic information is required for the photolysis of benzaldehyde and the a-carbonyl carboxylic acids, the fraction of the alkyl nitrate from the peroxy radical intermediates in the aromatic photooxidation mechanism, and the general atmospheric chemistry of a-dicarbonyls and y -unsaturated dicarbonyls.

Acknowledgment

Much of the initial modeling work on the toluene-NO, and o-cresol-NO, systems was carried out while one of the authors (R.A.) was at Shell Re- search Ltd., Thornton Research Centre, Chester, England. The authors wish to thank Dr. A. Glangetas for performing the GC-MS analyses, G. C. Vogelaar and F. R. Burleson for carrying out the gas chromatographic analyses, and W. D. Long for valuable assistance in conducting the chamber experiments. This work was supported by the National Science Founda- tion Research Applied to National Needs under Grant No. ENV73- 02904-A04, the U.S. Environmental Protection Agency under Grant No. 800649-19, and the California Air Resources Board under Contract No. A7 -17 5 -30.

APP

END

IX A

. R

eact

ions

and

rate

con

stan

ts u

sed

in th

e N

O,-t

olue

ne-a

ir m

odel

.

Rat

e co

nsta

nt

cm m

olec

ule

sec

Rea

ctio

n un

itsaz

b N

ote

Ref

.

6.6

X l

o-'

5.7

x 9.

1 x

10-1

2 1.

8 x

10-1

4 3.

7 x

10-1

7 1.

9 x

10-1

1 4.

3 x

10-1

6 1.

9 x

10-8

8 1.

9 x

10-1

2 0.

15

3 x

10-2

1 2.

8 X

7.

4 x

10-2

9.8

X

3.0

X lo

-" 2.

3 x

6.8

X 1

0-l2

1.

1 x

10-1

1 1.

1 x

10-3

(2.4

x 1

0-4)

8.5

x 10

-14

2.1

x 10

-15

8.1

X 1

0-l2

1.

2 x

10-1

2 0.

15

3 x

10-1

2 2.

5 x

10-l

2 6.

5 x

10-3

0

1

2 2,3

Est

imat

ed

4,5 1 1

1 1 2 2 1 2 2 E

stim

ated

6 7

APP

END

IX A

(C

onti

nued

from

pre

viou

s pag

e.)

Rat

e co

nsta

nt

cm m

olec

ule

sec

2

to

Rea

ctio

n un

itsaS

b N

ote

Ref

.

11.

Tol

uene

pho

toox

idat

ion

reac

tions

A

. A

bstr

actio

n O

H t to

luen

e - Cs

H5C

H2 t H

zO

CsH

5CH

zOz

+ N

O - C

BH

SCH

ZO + N

O2

CsH

5CH

zO2 t N

O +

CP,H

SCH

~ON

OZ

CsH

5CH

20 +

02

+ C

sH5C

HO

t H

Oz

CG

H~

CH

~O

+ N

O2 +

CsH

50N

Oz

c~

H~

CH

~

+ o2 - c

~H

~c

H~

O~

M

M

OH

+ b

lw

- &"

&: +

0, -+

&OH

+

HO,

4.7

x 10

-6

Fast

3.

0 x

10-1

3

2.8

x 10

-5

7.0

x 10

-5

(6-2

4) x

107

1 2 [7

01

> 1.

0 x

10-1

2 [W

91

2

Rat

io a

djus

ted

9 0

6 X

10-

l2

2 x

10-1

2

1.2

x 10

-15

6 4

1.6

x 10

-l1

6 r 3 z R

P

1.2

x 10

-15

11

5.2

X

12

[191

Mec

h. A

,B o

nly

Mec

h. C

,D;

(adj

uste

d)

Mec

h. A

,B;

(adj

uste

d)

Mec

h. C

,D o

nly 1

(Slo

w)

3.5

x 10

-15

(44)

(45)

z 0

2,13

3.

5 x

10-1

5

(Slo

w)

a

CH,

(Slo

w)

Fast

M

ech.

A,B

,E-H

M

ech.

C,D

14

,15

0 7N

HO

2 t N

O2 t C

HiC

OC

HO

t 0

25

(.7O

NO

2 &{

. 3 % t H

CO

CH

-CH

CH

O)

H

(47P

Fa

st

(Slo

w)

Mec

h. A

,B

Mec

h. C

-H

15,1

6

(Slo

w)

Fast

M

ech.

A-E

,G,H

M

ech.

E,F

14

,17

t 1

12 H

CO

CH

=CH

CH

O

t 1

12 C

H~C

OC

H-Z

CH

CH

O) t 0

.25

C70

NO

a F z U 4

z m (S

low

) Fa

st

Mec

h. A

-E,G

,H

Mec

h. E

,F

16,1

7 H

t

112

HC

OC

H=C

HC

HO

t 1

12 C

HzC

OC

H=C

HC

HO

i t 0

.26

C70

N02

oo@: !!?%

0.7X

HC

O +

NO

2 t (

CH

0j2

t C

H3C

OC

HO

HC

HO

l (S

low

) Fa

st

Mec

h. A

-F,H

M

ech.

G

2

w

14,1

8 t 0

.2.5

C7O

NO

2

APP

END

IX A

(Con

tinue

d fr

om p

reui

ous

page

.)

m w

Rat

e co

nsta

nt

rp

cm m

olec

ule

sec

Rea

ctio

n un

itsa.

b N

ote

Ref

.

CK

H 00.

+ CH

aCO

CH

OH

CH

O)

0.75

(HC

O +

NO

2 + (

CHO

):!

+ 0.2

5 C

7ON

02

(51)

'

H

(52)

' ~6:

+ CH

3CO

CH

O)

+ 0.2

5 C

70N

02

!%%

0.7X

HO

2 +

NO

2 t

2(C

HO

)2

(Slo

w)

Fast

(Slo

w)

Fast

(Slo

w)

Fast

1.3

X l

o-"

Fast

5 x

10-1

2

8 X

Fa

st

Fast

6

X

2 x

10-1

2

3 x

10-5

3.3

x 10

-4 I

Mec

h. A

-F,H

M

ech.

G

16,1

8

Mec

h. A

-G

Mec

h. H

Mec

h. A

-G

Mec

h. H

14,1

9

16,1

9

[56,

91

Adj

uste

d 20

21

21 6

Rat

io a

djus

ted

9


3 - 2 c- m I Y

w m hl

3

h

P -c

30" z z + +

0

m hl

.- 3 I

X 2

2

-3 hl

s +

m x

4 2

0 2

4 t

h c- W v

h h

Q1 W 00 W

1 v

h

0 c- 1

z A

PPEN

DIX

A (C

ontin

ued f

rom

pre

viou

s pag

e.)

m

Rat

e co

nsta

nt

cm m

olec

ule

sec

units

*.b

Not

e R

ef.

Rea

ctio

n

0.75

(H02

+ N

O2

+ HO

CO

CO

CH

3

+ 0.

25 C

70N

02

oo,

+ HC

OC

H=C

HC

HO

) (7

1Y

Q0:

H

Fast

M

ech.

A,B

(S

low

) M

ech.

C-H

15

,27

&I

CHa

!%%

0.

75(H

Oz +

NO2

+ ?I2

HO

CO

CH

O

(Slo

w)

Mec

h. A

-E,G

,H

(7W

@

: + 1

/2 H

OC

OC

OC

Ha

Fast

M

ech.

E,F

17

,26

+ 112

HC

OC

H=C

HC

HO

+

112

CH

3CO

CH

=CH

CH

O)

+ 0.2

5 C

7ON

Oz

(73Y

!%!&

0.75

(H02

+ N

O2

+ 11

2 H

OC

OC

HO

+ 112

HC

OC

H=C

HC

HO

+ 1

12 C

H&

OC

H=C

HC

HO

) + 0

.25

C70

NO

z

oo.

+ 112

HO

CO

CO

CH

3

H

NL),4

_ 0.

75(H

CO

+ N

O2

+ (C

HO

h + C

H3C

OC

(OH

hCH

O)

(74)

c + 0

.25

~~

0~

0~

(Slo

w)

Mec

h. A

-E,G

,H

Fast

M

ech.

E,F

17

,27

(Slo

w)

Mec

h. A

-F,H

Fa

st

Mec

h. G

18

,26

(Slo

w)

Mec

h. A

-F,H

Fa

st

Mec

h. G

18

,27

oo@

z -

Ny

), 0,

0 75

(H02

f N

O2

+ 3/2

(CH

0)2

f 1

/2 C

HR

CO

CH

O

(CH

0)z + h

u - Pr

oduc

ts

HC

OC

O - H

CO

+ C

O

OH

+ (C

H0)

z + H

zO +

HC

OC

O

HC

O +

02

4 HO

z + C

O

D.

Met

hyl g

lyox

al (a

nd it

s pro

duct

s)

(84)

(86)

(8

7)

CH

3CO

CH

0 + h

u - CH

3C0

+ HC

O

CH

3CO

C0 - C

H3C

0 + C

O

CH

3C0

+ O2 - C

H3C

03

(85)

O

H +

CH

3CO

CH

O +

H20

+ CH

3CO

CO

(88)

(8

9)

CH

3C03

+ N

O - C

HsC

Oi +

NO2

CH3C

Os +

NOz - CH

3C03

N02

(Slo

w)

Fast

(Slo

w)

Fast

(Slo

w)

2 x

10-1

1 Fa

st

5.6

x 10

-l2

1.0

x 10

-3

1.6

X lo

-" Fa

st

Fast

8

x 1O

-lZ

5 x

10-1

2

Mec

h. A

-G

Mec

h. H

Mec

h. A

-G

Mec

h. H

1.5

X lo

-"

1.6

X lo

-"

19,2

6

19,2

7

28

6 29

30

1,6

30

6 6 21

21

APP

END

IX A (C

ontin

ued

from

prev

ious

pag

e.)

2 03

Rat

e co

nsta

nt

cm m

olec

ule

sec

Rea

ctio

n un

itsav

b

(90)

(9

1)

(92)

(9

3)

(94)

(9

5)

(96)

(9

7)

(98)

(9

9)

CH

3C03

N02

- CH

3C03

+ N

O2

CH

3CO

i - CH

; + C

O2

CH; + 0

2 - CH

302

CH@

2 + N

O - C

H30

+ NO

2 C

H30

+ NO

2 +

C

H30

N02

C

H30

+ NO

2 - HC

HO

+ H

ON

O

CH

30 +

02

- HC

HO

+ H

Oz

HC

HO

+ hv

-H

+ HC

O

HC

HO

+ hv +

H2

+ CO

O

H +

HC

HO

+ H

20 +

He

0

E. C

onju

gate

d y-

dica

rbon

yls

(100

) (1

01)

(102

),( 10

3)d

(!04)

,(105

)d

(106

),(10

7)d

RC

OC

H=C

HC

O

+ 02

+RC

OC

H=C

HC

&

(108

),(10

9)d

OH

+ HC

HO

=CH

CH

O - Hz

O +

HC

OC

H=C

HC

O

OH

+ CH

3CO

CH

=CH

CH

O +

HzO

+ 0

3 + R

CO

CH

=CH

CH

O - Pr

oduc

ts

RC

OC

H=C

HC

HO

+ h

v - Pr

oduc

ts

CH

3CO

CH

=CH

CO

M

RCOC

H=CH

CO, +

R C

H ;h0

(110

),(11

1)d

(112

),(11

3)d

RC

OC

H=C

HC

&

+ NO

2 +

RC

OC

H=C

HC

&

+ NO

- NO

2 +

RC

OC

H=

CH

C~

~

M

RC

OC

H=C

HC

03N

02

8.0

x 10

-4

Fast

Fa

st

8 X

10-

l2

1.3

x lo-

" 2.

1 x

10-1

2 1.

2 x

10-1

5

4.1

x 10

-5

9.4

x 10

-12

3 x

10-1

1 1.

6 X

lo-

"

(Slo

w)

(Slo

w)

Fast

(Slo

w)

Fast

8 X

5 x

10-1

2

Not

e R

ef.

163,

641

6

Mec

h. A

,C,E

M

ech.

B,D

,F

30,3

1 30

.31

32

33

34

21

21

m 4

n

n

(1 14

),( 1 1

5)d

( 1 16

),( 1 17

)d

(120

),(12

1)d

0

n

0

0

0

0

II R

C mod

:< iko -. RCO

CH(O

.)CH

O +

CO,

n H

CO

CH

(O.)C

HO

- HC

O +

(CH

0)Z

CH

&O

CH

(O-)

CH

O - 1

/2(C

H&

O +

(CH

0)z)

t 1

/2 (C

H3C

OC

HO

+ H

CO

) R

CO

CH

=C

HC

~~

- RCO

CH

=CH

+ c

o2

M

RC

OC

H=C

H

+ 02

-+ R

CO

CH

=CH

Or

RC

OC

H=C

HO

z + N

O +

NO

2 +

RC

OC

H=C

HO

. - RC

OC

HC

HO

M

R

CO

CH

=CH

Oy

+ NO

+ R

CO

CH

=CH

ON

02

M

RC

OC

HC

HO

+ 02

-+ R

CO

CH

(Oz-

)CH

O

RC

OC

H(0

2*)C

HO

+ NO

- RC

OC

H(O

*)C

HO

+ N

O2

RC

OC

H(0

y)C

HO

t N

O -+

RC

OC

H(O

N02

)CH

O

M

Fast

s x 1

0-1

3

9,35

Fast

Fast

Fa

st

Fast

Fast

7 x

10-1

2

s x 1

0-13

Fast

s x 1

0-13

6 6 6

9,35

9,35

P

2

rj co

W

)-L

APP

END

IX A

(C

onti

nued

from

pre

viou

s pa

ge.)

Rat

e co

nsta

nt

cm m

olec

ule

sec

00

N

0

Rea

ctio

n N

ote

Ref

.

(138

) (1

39)

(140

),( 14

1)d

HC

OC

H(0

-)C

HO

- (CH

0)z +

HCO

C

H&

OC

H(O

-)C

HO

- 1/2

(CH

3CO

+ (C

H0)

z)

RC

OC

H=C

HC

03N

Oz

4 R

CO

CH

=CH

C03

+ N

O:!

+ 1/2

(HC

O +

CH

3CO

CH

O)

F. p

-Hyd

roxy

-P-d

icar

bony

ls

(142

)

(143

)

(145

), (14

6)"

CH

&O

CH

(OH

)CH

O +

OH

- CH3C

OC

(OH

)CH

O +

CH

3CO

C(O

Il)C

HO

+ 0

2 - H

Oz

+ CH

3CO

CO

CH

O

CH

&O

CX

(OH

)CH

O +

OH - C

H&

OC

X(O

H)C

O

CH

3CO

CX

(OH

)C03

+ N

O - C

H~C

OC

X(O

H)C

OZ

HzO

(144

) C

H3C

OC

OC

HO

- CH3C

OC

HO

+ CO

M

(147

),(14

8)"

CH

~CO

CX

(OH

)CO

+ o2 +

CH

~C

OC

X(O

H)C

~~

(1

49),(

150)

" (1

51),(

152)

" C

H3C

OC

X(O

H)C

O3 +

NO2 - CH

3CO

CX

(OH

)-

C03

N02

(1

53),(

154)

' (1

55),(

156)

" (1

57),(

158)

" C

H&

OC

X(O

H)C

03N

Oz +

CH

3CO

CX

(OH

)C&

+

CH

3CO

CX

(OH

)CO

z - CH

3CO

CX

OH

+ CO

z C

H3C

OC

XO

H +

02

- HOz

+ CH

3CO

CO

X

N0z

G

. a-C

arbo

nyl a

cids

(1

59)

(161

)

HO

CO

CH

O +

OH

- HO

CO

CO

+ H

zO

HO

CO

+ 0

2 - HO

z + C

On

(160

)

(162

),(16

3Id

HO

CO

CO

- HOC

O +

CO

HO

CO

CO

R +

hu - Pr

oduc

ts

+ OH -

Plhd

ucts

+ H

Oz

@Lo

*

Fast

Fa

st

8.0

x 10

-4

6 X

10-

l2

Fast

Fa

st

1.6

X lo

-" Fa

st

8 X

10-

l2

5 x

10-1

2

Fast

Fa

st

8.0

x 10

-4

1.6

X 10-l:

Fast

Fa

st

(Slo

w)

-6 X

6 6 36

37

38

30

21

21

6 36

38

30

39

40

IV.

Hyd

rope

roxi

des

(165

) R

Oz‘

+ H

Oz

-+

RO

zH +

02

3 x

10-1

2 6

(166

R

OzH

f + hu

- OH

+ H

Oz

+ Pro

duct

s 5

x 10

-6

6

a R

ate

cons

tant

s ar

e gi

ven

for t

he c

ondi

tions

of t

he sm

og c

ham

ber

runs

list

ed in

Tab

le I

(303

K, 7

35 to

rr, M

= a

ir).

Ph

otol

ysis

rate

con

stan

ts a

re

give

n fo

r the

ligh

t int

ensi

ty a

nd sp

ectr

al d

istr

ibut

ions

ass

ocia

ted

with

runs

EC

-327

-EC

-340

[34

]; pho

toly

sis

rate

s for

runs

EC

-266

-EC

-281

are

som

ewha

t lo

wer

. Firs

t-or

der

reac

tion,

sec

-I;

seco

nd-o

rder

reac

tion,

cm

3/m

olec

ule

sec;

thir

d-or

der

reac

tion,

cm

6/m

olec

ule2

sec;

zer

o-or

der r

eact

ion

[rea

ctio

n (3

4)

only

], m

olec

ule/

cm3

sec.

“F

ast”

mea

ns t

hat t

he re

actio

n is

the

sole

pat

hway

for t

hat p

artic

ular

spe

cies

; the

pre

dict

ions

are

inse

nsiti

ve f

or th

e ac

tual

ra

te c

onst

ant.

“(Sl

ow),”

or r

ate

cons

tant

s giv

en in

par

enth

eses

mea

ns th

at th

e re

actio

n is

ass

umed

to

be n

eglig

ible

or t

hat t

he re

actio

n ha

s a

negl

igib

le

effe

ct o

n th

e pr

edic

tions

of t

he m

odel

for t

he c

once

ntra

tion

regi

mes

stu

died

her

e.

The

reac

tion

as w

ritte

n re

pres

ents

an

over

all p

roce

ss fo

r whi

ch t

he d

etai

led

reac

tions

invo

lved

are

giv

en in

App

endi

x B

. R

eact

ion

num

bers

refe

r to

R =

-H

and

-CH

3,

resp

ectiv

ely.

R

eact

ion

num

bers

refe

r to

X =

-H

and

-OH

, re

spec

tivel

y.

R0z

refe

rs t

o al

l rad

ical

s of

the

form

RO

z or

RC

03 in

the

mec

hani

sm.

RO

zH is

trea

ted

as a

sing

le s

peci

es.

P

Z d 4

T m

APP

EN

DIX

B.

Rea

ctio

ns a

nd r

ate

cons

tant

s use

d in

the

tolu

ene-

NO

,-air

m

odel

: ge

nera

l aro

mat

ic-O

H-0

2 ad

duct

reac

tion

mec

hani

sm.

00 z

Rea

ctio

n R

ate

cons

tant

cm

mol

ecul

e se

c un

itsa.

b N

otes

C

[2.2

.2] C

ycliz

atio

n

[2.2

.2]

Cyc

lizat

ion,

no

recy

cliz

atio

n

(5')

0

OH

I1

I HK), +

6-C-C

-1

sc-c

-1

+ 0, -

Fast

M

ech.

A-D

, (G

,H)d

(S

low

) M

ech.

E,F

Fast

Rat

io

adju

sted

6 x

10-l

2

2 x

10-1

2

Fast

Fas

t

9 38

[3.2

.1] C

ycliz

atio

n

4

(Slo

w)

Mec

h. A

-D

Fast

M

ech.

E,F

,(G,H

)d

on

OH

OH

(8')

&$

u

0,

-+

112

'&&s

+ 11

2 5g

Fast

00'

.OQ

(3

.2.1

1 C

ycliz

atio

n, n

o re

cycl

izat

ion

Rat

io

9 ad

just

ed

(9')

'6

3 +

NO+'&

s +

NO,

6 X

lo-''

00.

0'

(10'

) 2

x 10

-12

00

' O

NQ

,

(12'

) 5 &

OH+

NO

-

00

Rat

io

9 ad

just

ed

6 x

lo-'*

2 x

lo-'?

W

N

w

APP

END

IX B

(Con

tinue

d fr

om pr

evio

us p

age.

) 00

to

rp

Rea

ctio

n R

ate

cons

tant

cm

mol

ecul

e se

c un

itsa,

b N

otes

C

OH 0

I

II - 1-

c--c

-2

+

HO

0

~~

.

I II

(14')

1-C-C

-2 +

0, --f

HQ

, +

Fast

Fast

38

0

OH

Rec

ycliz

atio

n (16'

)

Fast

(Slo

w)

Fast

[Rea

ctio

n (6

'11

Mec

h. A

-F

Mec

h. G

,H


t t

* m a

N 3

l s X W

6 z +

4 6 + m

z

3 O W o .O

4 0" +

t 0 z +

N 3

I

X s N

4 + m

APP

END

IX B

(Con

tinue

d fr

om p

revi

ous p

age.

)

w R

eact

ion

Rat

e co

nsta

nt c

m m

olec

ule

sec

units

arb

Not

esC

a

(24'

)

6 X

1OI2

2 x

10'2

,f

Fast

0

OH

0

0

It I

I II

I1

- 6-

C-C-

C-C-

3 f

12

Rat

io

9 ad

just

ed

NO,-AIR PHOTOOXIDATION OF TOLUENE AND T H E CRESOLS 827

m Ill

g*G 3

0

t

t

LO

t 0

N b m O

@'* 3

111 ""9

b -x II

-cz In N

L PI ""7 +

I "=Y "?

t

3 A

PPEN

DIX

B (

Con

tinue

d fr

om p

revi

ous p

age.

) 00

Rea

ctio

n R

ate

cons

tant

cm

mol

ecul

e se

c un

itsa,

b N

otes

C

0

OH

0

0

II

II

U

II

(3

2')

6-C-C

-C-C

-3 --+

12

0

o o

ono

II

- 5

-c-C-

c-C--2

(2

9')

61

0-

0

0

o on

0 0

II I

II !I

6-C-

C--C

-2

+ 3-C

, I 1

Fast

Fast

Fast

3

a-D

icar

bony

l dec

ompo

sitio

n

Fast

(S

low

) M

ech.

G

Mec

h. H

Fast

(S

low

) M

ech.

G

Mec

h. H

P-H

ydro

xy d

ecom

posi

tion

Qi

o on

6 o

0

OH

II

I I

II I1

I 6-

--c--C

-C-C

-3

-.+

6-C-C

-1 I

1

(34‘

)

ii

no

0

I II

Lt 1-

c-c-

2

3-

2-c-

c-3

If

no

o I1

I I-C

-C-2

+ O

2

,f

I I

(Slo

w)

Fast

M

ech.

G

Mec

h. H

(Slo

w)

Mec

h. G

Fa

st

Mec

h. H

Fast

[R

eact

ion

(14’

)]

Fast

[R

eact

ion

(6‘)

]

a R

ate

cons

tant

s are

giv

en fo

r the

cond

ition

s of

the

smog

cha

mbe

r ru

ns li

sted

in A

ppen

dix

A (T =

303

K, 7

35 to

rr, M

= a

ir).

Ph

otol

ysis

rate

con

stan

ts

are

give

n fo

r the

ligh

t int

ensi

ty a

nd s

pect

ral d

istr

ibut

ions

ass

ocia

ted

with

run

s E

C-3

27-3

40 [

34];

phot

olys

is r

ate

for r

uns

EC

-266

-281

are

som

ewha

t lo

wer

. Firs

t-or

der r

eact

ion,

sec-

I; se

cond

-ord

er re

actio

n, c

m3/

mol

ecul

e se

c.

“Fas

t” m

eans

that

the

reac

tion

is th

e so

le p

athw

ay fo

r tha

t par

ticul

ar sp

ecie

s;

the

pred

ictio

ns a

re in

sens

itive

for

the

actu

al r

ate

cons

tant

. “(

Slow

)” m

eans

that

the

reac

tion

is a

ssum

ed to

he

negl

igib

le, o

r tha

t the

reac

tion

has a

ne

glig

ible

eff

ect o

n th

e pr

edic

tions

of t

he m

odel

for

the

conc

entr

atio

n re

gim

es s

tudi

ed h

ere.

Fo

otno

tes

are

thos

e of

App

endi

x A.

M

echa

nism

s G

and

H pr

edic

t for

mat

ion

of t

he sa

me

prod

ucts

, reg

ardl

ess o

f whe

ther

[2.

2.2]

or [

3.2.

1] cy

cliz

atio

n oc

curs

. Su

bseq

uent

reac

tions

of t

his p

rodu

ct a

re ig

nore

d.

Subs

eque

nt re

actio

ns of

the

se sp

ecie

s or

inte

rmed

iate

s ar

e sh

own

in A

ppen

dix

A.

830 ATKINSON ET AL.

Notes

1. k l was determined from the photolysis of NO2 in Nz [34,37,43]. All other photolysis rate constants were derived as described previously [7] from the absorption coefficients of the photolyzing species, the quantum yields, and the relative spectral distribution of the photolysis light [34].

2. The rate constant is pressure dependent and is given for 1 atm of air.

3. k g was derived from the equilibrium constant of Graham and Johnston [67] combined with the Nz05 decomposition rate constant; k l o was derived by interpolating the data of Connell and Johnston [71], which were reported a t various NZ number densities, to 760 torr of N2.

4. Value given is consistent with that required for the NO,-n-butane-propene-air system [7]. It is also consistent with the gas phase upper limit of k11 5 1.3 X cm3/molecule sec obtained by Morris and Niki ~ 1 .

5. Value given is specific for the SAPRC evacuable chamber. 6. Based on the assignment or estimate of Carter et al. [7] for the same

or for analogous reactions. 7. Pseudo-third-order reaction added to account for the observed ef-

fects of humidity on this reaction [80,86]. sistent with the data of Hamilton and Lii [80] and the HOyHz0 equilibrium constant calculated by Hamilton and Naleway [87]

8. Reaction (34) is required in the model to account for the excess radical initiation associated with all of our smog chamber experiments [7,8]. In modeling our NO,-propene and/or n-butane-air chamber runs [7], we found that the values of k34 required for calculations to fit the data generally depended on initial NO, NO2, and humidity values. Based on these cor- relations [7,88], the following values of k34 have been assigned: k 3 4 2 0.2 ppb/min for run EC-273; 0.3 ppb/min for runs EC-271 and EC-281; and 0.4 ppb/min for runs EC-266, EC-269, EC-293, EC-339, and EC-340.

9. The total RO2 + NO rate constant used was assumed to be equal to that for the CH3Oz + NO rate constant kg3. Recently, Plumb et al. [89] obtained a value of k93 = (8 f 2) x cm3/molecule sec a t room temperature; Adachi and Basco [go] obtained hg3 = (3.0 f 0.2) X cm3/ molecule sec; and Cox and Tyndall [91] obtained (6.5 f 2.0) X cm3/molecule sec. Since Adachi and Basco [go] may have had problems with the product CH30NO in their uv absorption system, and since HOz formed from HCO + 0 2 may have caused problems in the study of Plumb et al. [89], the data to date are ambiguous. We have used a value of k93 of 8 X cm3/molecule sec, identical to that for the reaction of HOz radicals with NO, and consistent with the above-mentioned rate constant studies. The relative rates of alkoxy formation versus alkyl nitrate formation are unknown for these systems, but based on the results of Darnall et al. [60] alkyl nitrate formation probably occurs a t least 10% of the time in Cg+

was calculated to be con

NO*-AIR PHOTOOXIDATION OF TOLUENE AND T H E CRESOLS 831

systems. For these calculations the efficiency of alkyl nitrate formation was assumed to be the same for all Cs and C7 peroxy radicals, and the best fits to the data are obtained when -25% alkyl nitrate formation is assumed. These reactions have also been discussed by Hoshino et al. [30].

10. Ortho addition is assumed to predominate, based on experimental product data [19,30,31], and results of the analogous addition of O(3P) atoms to toluene [52-541.

11. Reaction (42) is assumed to have the same rate constant as RCH20 + 0 2 - RCHO + HO2 [e.g., reaction (96)], since they have similar ex- othermicities [58] and probably have similar transition states.

12. Derived from 1242 and the h42/k43 ratio of Kenley et al. [19]. However, the nitrotoluene yields would be better fit if h43/(h42 + h43 + 1244 + 1245) were a factor of -10 higher.

13. The total rate constant of 0 2 addition to the toluene-OH adduct was the same for all mechanisms considered, and was adjusted to fit the observed maximum cresol yields in the runs modeled. Since the OH-aromatic-02 adduct can back decompose to the 0 2 + the OH-aromatic adduct [9] or cyclize, this rate constant is a composite of h ( 0 2 + OH-aromatic-adduct) X h (cyc1ization)lh (back-decomposition). Since h ( 0 2 + OH-aromatic-adduct) is expected, by analogy with other alkyl radicals, to be cm3/molecule sec [70], then this implies that the ratio of dis- sociation/cyclization for the OH-aromatic-02 adduct is -300. This means that the OH-aromatic adduct and the OH-aromatic-02 adduct are probably essentially in equilibrium under atmospheric conditions.

14. Detailed reactions shown in Appendix B for substituent 2 = CH3, 1,3-6 = H.

15. The overall reaction occurs as shown if 12.2.21 cyclization predominates and recyclization is negligible [reactions (1’)-(6’) in Appendix B]. hy l (h3~ + k4’) is assumed to be 0.75.

16. Detailed reactions shown in Appendix B for substituent 6 = CH3, 1-4 = H.

17. The reaction occurs as shown if [3.2.1] cyclization predominates and recyclization is negligible [reactions (7’)-( 159, (6’) in Appendix B). h g ~ / ( k y + 1210.) = h l l f / ( h l l ~ + h 1 ~ ) is assumed to be 0.75.

18. The reaction occurs as shown if recyclization predominates and if 0-scission of a-carbonyl, /3-hydroxy alkoxy radicals results primarily in formation of carbonyl radicals. The same products are formed regardless of whether [3.2.1] or [2.2.2] cyclization predominates [reactions (l’), (29, (79, (89 , (16’)-(32’)]. k214(&21r + h 2 ~ ) = k234(k231 + 1224’) is assumed to be 0.75.

19. The reaction occurs as shown if recyclization predominates and 0-scission of a-carbonyl, ,8-hydroxy alkoxy radicals results primarily in formation of a-hydroxy radicals. The same products are formed regardless of whether [3.2.1] or [2.2.2] cyclization predominates [reactions (l’), (2’), (7’), (8’1, (16’)-(30’), (33’), (34’1, (6’), (14’)]. hzi~/(k21( + h22O = k23f/(h23’ + h24’) is assumed to be 0.75.

832 ATKINSON ET AL.

20. Estimated from the observed disappearance rates of benzaldehyde in runs EC-337 and EC-339, after correcting for reaction of benzaldehyde with OH [9,56] and for benzaldehyde formation from OH + toluene [9,15,19]. The hydroxyl radical concentrations were estimated using the toluene disappearance rates, and the OH + toluene rate constant [9,15]. It is necessary to assume formation of nonradical products, or the overall reactivities of runs EC-337 and EC-339 are overpredicted.

21. RC03 is assumed to react with NO with the same rate constant as CH3Oz (see note 9). The k(RC03 + NO)/k(RC03 + NO2) ratio assigned by Carter et al. [7] was used.

22. Assumed to occur with the same rate constant as that assigned for ROZ + HO2 and NO3 + HO2 [7].

23. OH addition as shown is estimated [58] to be at least 3-7 kcal/mol more thermochemically favorable than addition at any other position, and is therefore assumed to predominate.

24. Assumed to have same rate constant as assigned for reaction (43) (see note 12).

25. Assumed to be analogous to 0 2 addition to the OH-toluene adduct, i.e., k68 = k44 and k69 = k45 (see note 13).

26. Detailed reaction shown in Appendix B for substituent 1 = OH, 2 = CH3,3,4, and 6 = H.

27. Detailed reactions shown in Appendix B for substituent 1 = OH, 6

28. Included in the mechanism to account for the observed rapid disappearance of cresols in the presence of 0 3 + NO2 [65,66], and the loss of cresol formed in NO,-toluene-air runs following NO consumption. The rate constant was derived from subsidiary experiments [66]. The 2-methyl phenoxy radical assumed to be formed [66] is expected to react primarily with NOz, giving rise to nonradical products (i.e., hydroxynitrotoluenes). Mechanisms which assume that NO3 + cresol gives radicals result in significant overprediction of reactivity in NO,-cresol-air runs.

29. Although glyoxal has large absorption coefficients in the 340-460 wavelength region [92] where the intensity of the photolyzing light is high, it appears to have low photodecomposition quantum yields at atmospheric pressure [92].

30. Assumed to react with a similar rate constant as other aldehydes

31. This mechanism is assumed to be primarily H abstraction as shown since OH radical addition to the double bond would be anticipated, by analogy with the reaction of O(3P) atoms with acrolein and crotonaldehyde [93] (when compared to O(3P) and OH + ethene [9,70]), to have a rate constant of <5 x 10-12 cm3/molecule sec.

32. 0 3 addition to the double bond is anticipated to be slow, by analogy with the reaction of O(3P) atoms with acrolein and crotonaldehyde [93], when compared to ethene [9,70].

= CH3,2-5 = H.

[91.


33. The photochemistry of these species is assumed to be analogous to that of acrolein and crotonaldehyde [92] which are believed to be photo- chemically stable under atmospheric conditions.

34. These initial additions to the double bond are estimated to be exothermic by -3 kcal/mol[58]. If the estimates of Demerjian et al. [94] (assuming an A factor of sec-’ and an activation energy of 11 kcal/mol) are correct, these reactions should dominate over competing processes. However, if their activation energy estimate is low by 2 5 kcal/mol, which is within the uncertainty range of their estimation, the reaction would be expected to be unimportant.

35. Nitrate formation is arbitrarily assumed to occur -10% of the time, based on results for Cq and C5 alkyl radicals [60].

36. Assumed to decompose with the same rate constant as peroxya- cetylnitrate (PAN) [63,64].

37. Assumed t‘o occur with the same rate constant as OH + isopropanol

38. Reaction with 0 2 forming HO2 and carbonyls has been shown to be the primary fate of a-hydroxy radicals in air [95].

39. The rates of photolysis for these acids are unknown, although there is evidence that photodecomposition to C02 + CHsCHO is efficient in the gas phase at 353 K [92]. If they photolyze to form radicals rapidly (analogously to methylglyoxal), model calculations greatly overpredict the reactivity of the o-cresol-NO, system, regardless of which mechanistic option is used.

40. Assumed to react with the same rate constant as OH + toluene. This reaction is only significant in affecting primary nitrotoluene yields.

~91.

Bibliography

[l] W. A. Lonneman, S. L. Kopczynski, P. E. Darley, and F. D. Sutterfield, Enuiron. Sci.

[2] J. Kondo and H. Akimoto, Adu . Enuiron. Sci. Technol., 5 , l (1975). [3] J. M. Heuss, G. J. Nebel, and B. A. D’Alleva, Enuiron. Sci. Technol., 8,641 (1974). [4] H. Mayrsohn, M. Kuramato, J. N. Crabtree, R. D. Sothern, and S. H. Mano, “Atmospheric

Hydrocarbon Concentrations June-September 1974,” California Air Resources Board, April 1975.

Technol., 8,229 (1974).

[5] B. J. Finlayson and J. N. Pitts, Jr., Science, 192,111 (1976). [6] B. J. Finlayson-Pitts and J. N. Pitts, Jr., A d u . Enuiron. Sci. Technol., 7,75 (1977). [7] W. P. L. Carter, A. C. Lloyd, J. L. Sprung, and J . N. Pitts, Jr., Int. J. Chem. Kinet., 11,

45 (1979). [8] D. G. Hendry, A. C. Baldwin, J. R. Barker, and D. M. Golden, “Computer Modeling of

Simulated Photochemical Smog,” EPA-600/3-78-059, U.S. GPO, Washington, DC, June 1978.

191 R. Atkinson, K. R. Darnall, A. C. Lloyd, A. M. Winer, and J. N. Pitts, Jr., Adu . Photo- chem., 11,375 (1979).

[lo] E. D. Morris, Jr., and H. Niki, J. Phys. Chem., 75,3640 (1971). [ l l ) G. J. Doyle, A. C. Lloyd, K. R. Darnall, A. M. Winer, and J. N. Pitts, Jr., Enuiron. Sci.

[12] D. D. Davis, W. Bollinger, and S. Fischer, J. Phys. Chem., 79,293 (1975). Technol., 9,237 (1975).

834 ATKINSON ET AL.

[13] D. A. Hansen, R. Atkinson, and J. N. Pitts, Jr., J. Phys. Chem., 79,1763 (1975). [14] A. C. Lloyd, K. R. Darnall, A. M. Winer, and J. N. Pitts, Jr., J . Phys. Chem., 80, 789

[15] R. A. Perry, R. Atkinson, and J. N. Pitts, Jr., J. Phys. Chem., 81,296 (1977). [16] R. A. Perry, R. Atkinson, and J. N. Pitts, Jr., J . Phys. Chem., 81,1607 (1977). [17] A. R. Ravishankara, S. Wagner, S. Fischer, G. Smith, R. Schiff, R. T. Watson, G. Tesi,

[18] R. Atkinson, K. R. Darnall, and J. N. Pitts, Jr., J. Phys. Chem., 82,2759 (1978). [19] R. A. Kenley, J. E. Davenport, and D. G. Hendry, J. Phys. Chem., 82,1095 (1978). [20] D. G. Hendry, “Chemical Kinetic Data Needs for Modeling the Lower Troposphere,”

Reston, VA, May 15-17,1978, Natl. Bur. Stand. U.S. Spec. Publ. 557, Sept. 1979. [21] C. T. Pate, R. Atkinson, and J. N. Pitts, Jr., J. Enuiron. Sci. Health, A l l , 1 (1976). [22] P. L. Hanst, E. R. Stephens, W. E. Scott, and R. C. Doerr, “Atmospheric Ozone-Olefin

Reactions,” The Franklin Institute, Philadelphia, PA, 1958. [23] J. J. Bufalini and A. P. Altshuller, Can. J. Chem., 43,2243 (1965). [24] D. H. Stedman and H. Niki, Enuiron. Lett., 4,303 (1973). [25] J. M. Heuss and W. A. Glasson, Enuiron. Sci. Technol., 2,1109 (1968). [26] A. P. Altshuller, S. L. Kopczynski, W. A. Lonneman, F. D. Sutterfield, and D. L. Wilson,

[27] W. Schwartz, “Chemical Characterization of Model Aerosols,” Battelle-Columbus Rep.

[28] R. J. O’Brien, Pacific Conf. on Chemistry and Molecular Spectroscopy, Los Angeles,

[29] C. W. Spicer and P. W. Jones, J. Air Pollut. Control Assoc., 27,1122 (1977). [30] M. Hoshino, H. Akimoto, and M. Okuda, Bull. Chem. SOC. Jpn., 51,718 (1978). [31] D. Grosjean, K. Van Cauwenberghe, D. R. Fitz, and J. N. Pitts, Jr., 175th Nat. Am. Chem.

SOC. Meeting, Anaheim, CA, Mar. 12-17,1978. [32] R. J. O’Brien, P. J. Green, R. A. Doty, J. W. Vanderzanden, R. R. Easton, and R. P. Irwin,

in “Nitrogenous Air Pollutants,” D. Grosjean, Ed., Ann Arbor Press, Ann Arbor, MI, 1979.

[33] H. Akimoto, M. Hoshino, G. Inoue, M. Okuda, and N. Washida, Bull. Chem. Soc. Jpn., 5! , 2496 (1978).

[34] J. N. Pitts, Jr., K. R. Darnall, W. P. L. Carter, A. M. Winer, and R. Atkinson, “Mecha- nisms of Photochemical Reactions in Urban Air,” EPA-600/3-79-110, U.S. GPO, Washington, DC, Nov. 1979.

(1976).

and D. D. Davis, Int. J. Chem. Kinet., 10,783 (1978).

Enuiron. Sci. Technol., 4,44 (1970).

to EPA, U.S. GPO, Washington, DC, Aug. 1974.

CA, Oct. 1975.

[35] K. R. Darnall, R. Atkinson, and J. N. Pitts, Jr., J. Phys. Chem., 83,1943 (1979). [36] H. Takagi, N. Washida, H. Akimoto, K. Nagasawa, Y. Usui, and M. Okuda, J. Phys.

Chem., 84,478 (1980). [37] J. N. Pitts, Jr., K. R. Darnall, A. M. Winer, and J. M. McAfee, “Mechanisms of Photo-

chemical Reactions in Urban Air, 11. Smog Chamber Studies,” EPA-600/3-77-014b, U.S. GPO, Washington,DC, Feb. 1977.

[38] A. M. Winer, R. A. Graham, G. J. Doyle, P. J. Bekowies, J . M. McAfee, and J. N. Pitts, Jr., Adu. Enuiron. Sci. Technol., 10,461 (1980).

[39] J. H. Beauchene, P. J . Bekowies, J. M. McAfee, A. M. Winer, L. Zafonte, and J. N. Pitts, Jr., Proceedings ofthe 7th Conference on Space Simulation, paper 66, NASA Spec. Publ. 336 (Nov. 12-14,1973).

[40] G. J. Doyle, P. J. Bekowies, A. M. Winer, and J. N. Pitts, Jr., Enuiron. Sci. Technol., 11, 45 (1977).

[41] A. M. Winer, J. W. Peters, J . P. Smith, and J. N. Pitts, Jr., Enuiron. Sci. Technol., 8,1118 (1974).

[42] R. G. Smith, R. J. Bryan, M. Feldstein, B. Levadie, F. A. Miller, E. R. Stephens, and N. G. White, H.L.S., 7 (1) Suppl., 87 (1970).

[43] J. R. Holmes, R. J. O’Brien, J. H. Crabtree, T. A. Hecht, and J. H. Seinfeld, Enuiron. Sci. Technol., 7,519 (1973).

NO%-AIR PHOTOOXIDATION OF TOLUENE AND T H E CRESOLS 835

[44] E. F. Darley, K. A. Kettner, and E. R. Stephens, Anal. Chem., 35,589 (1963). [45] E. R. Stephens and M. A. Price, J . Chem. Educ., 50,351 (1973). [46] E. R. Stephens, “Hydrocarbons in Polluted Air,” Summary Rep., Coordinating Research

Council, Project CAPA-5-68, NTIS PB 230 993/AS, June 1973. [47] A. Glangetas, unpublished results from this laboratory, 1978. [48] C. W. Gear, Comm. ACM, 14,1976 (1971). [49] A. C. Hindmarsh, “Gear: Ordinary Differential Equations Systems Solver,” Lawrence

[50] A. Prothero and A. Robinson, Math. Comput., 28,145 (1974). [51] Send requests for data tabulations to: W. P. L. Carter, Statewide Air Pollution Research

Livermore Lab., Dep UCID 3001, Rev. 2,1972.

Center, University of California, Riverside, CA 92521. G. R. H. Jones and R. J . Cvetanovic, Can. J . Chem., 39,2444 (1961). E. Grovenstein, Jr., and A. J. Mosher, J. Am. Chem. Soc., 92,3810 (1970). J. S. Gaffney, R. Atkinson, and J. N. Pitts, Jr., J . Am. Chem. Soc., 98,1828 (1976). A. Gitchell, R. Simonaitis, and J. Heicklen, J. Air Pollut. Contr. Assoc., 24. 357 (1974). H. Niki, P. D. Maker, C. M. Savage, and L. P. Breitenbach, J. Phys. Chem., 82, 132 (1978). H. Niki, P. D. Maker, C. M. Savage, and L. P. Breitenbach, in “Nitrogenous Air Pollu- tants,’’ D. Grosjean, Ed., Ann Arbor Press, Ann Arbor, MI, 1979. S. W. Benson, “Thermochemical Kinetics,” 2nd ed., Wiley, New York, 1976. J. A. Kerr and M. J . Parsonage, “Evaluated Kinetic Data on Gas Phase Addition Reac- tions,” CRC Press, Cleveland, OH, 1972. K. R. Darnall, W. P. L. Carter, A. M. Winer, A. C. Lloyd, and J . N. Pitts, Jr., J . Phys. Chem., 80,1948 (1976). D. G. Hendry and R. A. Kenley, ACS Meeting, Washington, Sept. 14-17,1979. C. T. Pate, R. Atkinson, and J . N. Pitts, Jr., J . Enuiron. Sci. Health, A l l , 19 (1976). D. G. Hendry and R. A. Kenley, J . Am. Chem. SOC., 99,3198 (1977). R. A. Cox and M. J. Roffey, Enuiron. Sci. Technol., 11,900 (1977). R. J . O’Brien, private communication, 1979. W. P. L. Carter, A. M. Winer, and J. N. Pitts, Jr., unpublished data, 1979. R. A. Graham and H. S. Johnston, J . Phys. Chem., 82,254 (1978). S. M. Japar and H. Niki, J. Phys. Chem., 79,1629 (1975). R. R. Arnts and B. W. Gay, Jr., “Photochemistry of Some Naturally Emitted Hydro- carbons,” EPA-600/3-79-081, U.S. GPO, Washington, DC, Sept. 1979. R. F. Hampson, Jr., and D. Garvin, Eds. Natl. Bur. Stand. U.S. Spec. Publ. 513, May 1978. P. Connell and H. S. Johnston, Geophys. Res. Lett., 6,553 (1979). G. E. Streit, C. J. Howard, A. L. Schmeltekopf, J. A. Davidson, and H. I. Schiff, J. Chem. Phys., 65,4761 (1976). C. Anastasi and I. W. M. Smith, J. Chem. Soc., Faraday Trans. 2, 74,1056 (1978). R. Overend, G. Paraskevopoulos, and C. Black, J . Chem. Phys., 64,4149 (1976). W. R. Stockwell and J. G. Calvert, J. Photochem., 8,193 (1978). A. R. Ravishankara, P. H. Wine, and A. 0. Langford, J . Chem. Phys., 70,984 (1979). M. Zahniser and C. J. Howard, in R. C. Whitten, W. 3. Borucki, L. A. Capone, and R. P . Turco, Nature, 275,523 (1978). C. J. Howard and K. M. Evenson, Geophys. Res. Lett., 4,437 (1977). R. A. Graham, A. M. Winer, and J. N. Pitts, Jr., Chem. Phys. Lett., 51,215 (1977). E. J. Hamilton, Jr., and R. R. Lii, Znt. J . Chem. Kinet., 9,875 (1977). N. Washida, R. I. Martinez, and K. D. Bayes, 2. Naturforsch., 29A, 251 (1974). K. Shibuya, T. Ebata, K. Obi, and I. Tanaka, J. Phys. Chem., 81,2292 (1977). A. Horowitz and J. G. Calvert, Znt. J . Chem. Kinet., 10,805 (1978). R. Atkinson and J. N. Pitts, Jr., J. Chem. Phys., 68,3581 (1978).

836 ATKINSON ET AL

[85] E. D. Morris, Jr., and H. Niki, J . Phys. Chem., 77,1929 (1973). [86] E. J. Hamilton, Jr., J . Chem. Phys., 63,3682 (1975). [87] E. J. Hamilton, Jr., and C. A. Naleway, J . Phys. Chem., 80,2034 (1976). [88] W. P. L. Carter, unpublished results from this laboratory, 1977-1978. [89] I. C. Plumb, K. R. Ryan. J. R. Steven, and M. F. R. Mulcahy, Chem. Phys. Lett., 63,255

[go] H. Adachi and N. Basco, Chem. Phys. Lett., 63,490 (1979). [91] R. A. Cox and G. S. Tyndall, Chem. Phys. Lett., 65,357 (1979). [92] J. G. Calvert and J. N. Pitts, Jr., “Photochemistry,” Wiley, New York, 1966. [93] J. S. Gaffney, R. Atkinson, and J. N. Pitts, Jr., J. Am. Chem. SOC., 97,5049 (1975), and

[94] K. L. Demerjian, J. A. Kerr, and J. G. Calvert, Ado. Enuiron. Sci. Technol., 4, 1

(951 W. P. L. Carter, K. R. Darnall, R. A. Graham, A.M. Winer, and J. N. Pitts, Jr., J . Phys.

(1979).

references therein.

(1974).

Chem., 83,2305 (1979).

Received December 4,1979 Accepted April 18,1980

A smog chamber and modeling study of the gas phase NOx–air photooxidation of toluene and the...

Documents

Transcript of A smog chamber and modeling study of the gas phase NOx–air photooxidation of toluene and the...