Atmospheric Oxidation Mechanism of Isoprene

CSIRO PUBLISHINGResearch Paper

J. Fan and R. Zhang, Environ. Chem. 2004, 1, 140–149. doi:10.1071/EN04045 www.publish.csiro.au/journals/env

Atmospheric Oxidation Mechanism of Isoprene

Jiwen FanA and Renyi ZhangA,B

A Department of Atmospheric Sciences, Texas A&M University, College Station, TX 77843, USA.B Corresponding author. Email: [email protected]

Environmental Context. Many plant species biosynthesize and emit the volatile hydrocarbon isoprene.Once in the atmosphere, isoprene is susceptible to a range of reactions involving potentially hundred ofproducts and intermediate compounds. The products of these reactions in turn may pose a risk to humanand plant health and impact the climate through the generation of acids, ozone, and atmospheric aerosols.

Abstract. The atmospheric oxidation mechanism of isoprene initiated by OH, O3, NO3, and Cl, which incorpo-rates the most recent laboratory and theoretical studies, is described. A box model intercomparison between thenew mechanism and previous available isoprene oxidation mechanisms has been performed. Ozone and OH con-centrations are compared with predictions by the previous mechanisms in high and low NOx scenarios. The O3 andOH sensitivities to the chlorine–isoprene reactions have also been investigated by comparing the box model resultswith and without the chlorine–isoprene reactions, showing that the ozone production rate and OH concentrationsare slightly impacted. The new mechanism facilitates more accurate modelling of isoprene photochemistry in theatmosphere.

Keywords. atmospheric chemistry — hydrocarbons — isoprene — oxidation

Manuscript received: 19 May 2004.Final version: 9 October 2004.

Introduction

Isoprene (2-methyl-1,3-butdiene; C5H8) is one of the mostimportant nonmethane hydrocarbons in the atmosphere.Isoprene emissions from vegetation constitute ∼40% of thetotal mass of nonmethane hydrocarbon compounds emittedto the atmosphere.[1] Atmospheric reactions of isoprene areinitiated by OH, O3, NO3, or halogen radicals and involveseveral thousand subsequent reactions and hundreds of inter-mediate species.[2] Photochemical oxidation of isoprene hasimportant implications for local and regional air quality, thegreenhouse effect including ozone, and acid production.[3]Isoprene oxidation represents the dominant source of sum-mertime ozone in North America and a major source ofHOOH in the lower atmosphere.[4]

Several mechanisms of atmospheric isoprene oxidationhave been developed and evaluated by air quality simulationmodels. Paulson and Seinfeld developed a mechanism foratmospheric photooxidation of isoprene that included exper-imental results of explicit identification of product yields suchas methacrolein (CH2 C(CH3)CHO; MACR), methyl vinylketone (CH2 CHCOCH3; MVK), and 3-methyl furan.[5]Another isoprene oxidation mechanism, developed by Carterand Atkinson, was evaluated for prediction of ozone for-mation based on smog chamber simulations.[6] Carter laterdeveloped two condensed mechanisms—a four-product anda one-product mechanism for airshed model applications

based on the detailed mechanism, which represented themost up-to-date isoprene mechanism developed before thiswork.[7]

The isoprene photooxidation mechanism developed byPaulson and Seinfeld incorporated the laboratory results circa1992.[5] For the OH–isoprene reactions, six peroxy radicalswere formed and their branching ratios were assumed. Anitrate yield of 0.12 per NO–peroxy radical reaction wasused in the mechanism. The ‘self’ reaction kinetics betweenthe peroxy radicals were included, taken from the work ofMadronich and Calvert.[8] A 4% yield of 3-methyl furanwas assumed in this mechanism. The O3–isoprene reactionmechanism was simply represented and most of the productyields were identified from environmental chamber experi-ments. This mechanism did not include the NO3–isoprenereaction, since the NO3–isoprene reactions were rather poorlyunderstood at that time.

The mechanism for the atmospheric reactions of isopreneand its major oxidation products developed by Carter andAtkinson reflected the knowledge of isoprene atmosphericchemistry circa 1996.[6] There were no explicit intermediateprocesses in this mechanism. For each isoprene reaction withOH, O3, or NO3, only one reaction was presented. The mech-anism lumped together several rapid consecutive reactionsand represented them as a single overall process. The four-product mechanism was less condensed and was intended for

© CSIRO 2004 140 1448-2517/04/030140


applications where explicit representation of isoprene’s majoridentified products, MACR and MVK, were desired.[7] Theone-product mechanism was the most condensed isoprenemechanism. It used only a single species, ISOPROD, to rep-resent the reactions of isoprene’s unique products. MACRand MVK were thus lumped with the other species repre-sented by ISOPROD. Both condensed mechanisms yieldedpredictions close to those of the detailed mechanism for O3

and OH. The O3 simulations were close to the environmentalchamber experimental data.

Recently, there has been substantial improvement inunderstanding of isoprene atmospheric chemistry by exper-imental and theoretical studies. Experimental studies haveinvestigated the kinetics and product yields of the OH–isoprene reactions.[9–15] Theoretical calculations employingquantum chemical methods and kinetic rate theories havepredicted the reaction rate constants and isomeric branchratios of the OH–isoprene reaction and several intermedi-ate steps.[16–25] The peroxy radical kinetics, especially the‘self’reactions between the peroxy radicals, were experimen-tally studied by Jenkin et al.[26] The work by Zhang et al.investigated the detailed mechanism of isoprene ozonoly-sis employing theoretical calculations.[27,28] Recent studiesinvestigated the NO3–isoprene reaction by combining exper-imental and theoretical studies.[29–31] In addition, the chlo-rine atom reaction with isoprene was also studied with theexperimental and theoretical methods.[32–38]

In this study, we provide a detailed and complete iso-prene oxidation mechanism by incorporating the most recentprogress in understanding of isoprene atmospheric chem-istry. The new mechanism is evaluated by comparing theresults with those of the Carter’s four-product condensedmechanism (CM), Paulson and Seinfeld’s mechanism (PSM),and a NCAR box model mechanism (MM2.1) on O3 andOH formation.[39] Also, the mechanism of the Cl–isoprenereactions is presented and compared using the box modelresults.

Description of Isoprene Oxidation Mechanism

The mechanism has been developed on the basis of themost up-to-date advances on the isoprene oxidation reac-tions initiated by OH, O3, NO3, and Cl. Table 1 summarizesthe relevant chemical species included in this mechanism,and Table 2 lists the reaction rate constants and isomericbranching ratios.There are no chemical operators and lumpedspecies in this mechanism. All major oxidation products arerepresented explicitly. The detailed descriptions of the newisoprene mechanism are presented as follows.

OH–Isoprene Reactions

The oxidation reaction of isoprene initiated by OH occursalmost entirely by OH addition to the two double bondsof isoprene, yielding four possible hydroxyalkyl radicals(Fig. 1). The reaction rate constant of 1.0 × 10−10 cm3

molec−1 s−1 has been measured experimentally and calcu-lated theoretically,[9,10,17] and the relative isomeric branch-ing ratios have been predicted as 0.56:0.02:0.05:0.37 for

ISOA:ISOB:ISOC:ISOD (Table 2).[17] Under atmosphericconditions, ISOA and ISOD react primarily with O2, eachforming a β- and a δ-hydroxyperoxy radical, becauseof the presence of two addition centres (β or δ to theOH position) in those two hydroxyalkyl radicals.[18] Theresulting δ-hydroxyperoxy radicals possess (E ) and (Z )configurations.[21,25] A ratio of 50:50 is assumed for thepartitioning between the (E ) and (Z ) configurations. There-fore, ISOA and ISOD forms three hydroxyperoxy radicals byreacting with O2, which are ISO2A, ISO2E, ISO2G, ISO2D,ISO2F, and ISO2H, respectively (Fig. 1). On the basis of therecent work by Park et al.,[23] ISOB primarily undergoes acyclic isomerization pathway to form α-hydroxy isopreneradicals, and then through hydrogen abstraction to produceC5 carbonyls. About 40% of ISOC forms α-hydroxy isopreneradicals and 60% of ISOC reacts with O2 to form hydrox-yalkyl peroxy radicals, ISO2C. The subsequent reaction ofhydroxyalkyl peroxy radicals with NO leads to the formationof organic nitrates or hydroxyalkoxy radicals. A yield of8% organic nitrate is assumed for NO–peroxy radical reac-tions based on the literature values, which range from 4 to14%.[13,40,41] The hydroxyalkoxy radicals from ISO2A andISO2D undergo decomposition to form MACR or MVK.[19]Likewise, the hydroxyalkoxy radicals from ISO2C decom-pose to form formaldehyde and the corresponding radicals;the radicals further react with O2 to produce MACR.[19]The hydroxyalkoxy radicals from ISO2E and ISO2F are(Z) configurations.[15,21] The study by Zhao et al. indicatedthat the (Z ) isomers primarily undergo 1,5-H shifts to form C5

hydroxycarbonyls (Mw 100).[15,25] The hydroxyalkoxy radi-cals from ISO2G ((E) configuration) undergo a 1,5-H shift,and then react with O2 and NO, yielding C4 hydroxycarbonyls(Mw 86) by undergoing C–C fission.[15] The hydroxyalkoxyradicals from ISO2H only form C5 hydroxycarbonyls after1,5-H shift.[15]

The hydroxyalkyl peroxy radicals also undergo self- orcross-reactions. The reaction rate constants and products ofthe peroxy self- or cross-reactions are updated based on thework of Madronich et al. and Jenkin et al.[26,39]

O3–Isoprene Reactions

The mechanism of isoprene ozonolysis is also shown inTable 2. Fig. 2 contains a schematic representation of themajor reaction pathways. On the basis of the theoreti-cal work by Zhang et al.,[27,28] two primary ozonides areformed from addition of O3 to isoprene. The decomposi-tions of the primary ozonides yield chemically activatedcarbonyl oxides, the so-called Criegee intermediates, becausethe reactions are strongly exothermic. In total, five car-bonyl oxides are formed (CI1–CI5). The relative branchingratios are 0.33:0.26:0.205:0.21:0.02.[27,28] The overall iso-prene ozonolysis rate constant is 1.6 × 10−17 cm3 molec−1

sec−1.[27,28] The excited carbonyl oxides will undergo stabi-lization and/or unimolecular reactions to form dioxirane andOH promptly. Under atmospheric conditions, the competi-tion between stabilization and unimolecular reactions of thecarbonyl oxides depends on their structures. For CI1, 95% is

141

J. Fan and R. Zhang

Table 1. Species in the New Isoprene Mechanism

C5 Compounds

ISOP isoprene ISN1 CH2=CHC(CH3)(O•)CH2(ONO2)ISOA CH2=CHC•(CH3)CH2OH ISN3 CH2=CHC(CH3)(ONO2)CH2(O•)ISOB CH2=CHC(CH3)OHCH•

2 ISN5 CH2(ONO2)C(CH3)=CHCH2(O•) (Z)ISOC CH2=C(CH3)CH(OH)CH•

2 ISN6 CH2(ONO2)C(CH3)=CHCH2(O•) (Z)ISOD CH2=C(CH3)CH•CH2OH ISN7 CH2(ONO2)C(CH3)=CHCH2(O•) (E)ISO2A CH2=CHC(CH3)(OO•)CH2OH ISN8 CH2(ONO2)C(CH3)=CHCH2(O•) (E)ISO2C CH2=C(CH3)CH(OH)CH2OO• ISOCl1 CH2=CHC(CH3)(OO•)CH2ClISO2D CH2=C(CH3)CH(OO•)CH2OH ISOCl2 CH2=CHC(CH3)ClCH2OO•

ISO2E •OOCH2CH=C(CH3)CH2OH (Z) ISOCl3 CH2=C(CH3)CH(Cl)CH2OO•

ISO2F HOCH2CH=C(CH3)CH2OO• (Z) ISOCl4 CH2=C(CH3)CH(OO•)CH2ClISO2G •OOCH2CH=C(CH3)CH2OH (E) ISOCl5 •OOCH2CH=C(CH3)CH2ClISO2H HOCH2CH=C(CH3)CH2OO• (E) ISOCl6 ClCH2CH=C(CH3)CH2OO•

ISON1 CH2=CHC(CH3)(OO•)CH2(ONO2) ISCl1 CH2=CHC(CH3)(O•)CH2ClISON2 CH2=C(CH3)CH(OO•)CH2(ONO2) ISCl2 CH2=CHC(CH3)ClCH2O•

ISON3 CH2=CHC(CH3)(ONO2)CH2(OO•) ISCl3 CH2=C(CH3)CH(Cl)CH2O•

ISON4 CH2=C(CH3)CH(ONO2)CH2(OO•) ISCl4 CH2=C(CH3)CH(O•)CH2ClISON5 CH2(ONO2)C(CH3)=CHCH2(OO•) (Z) ISCl5 •OCH2CH=C(CH3)CH2ClISON6 CH2(ONO2)C(CH3)=CHCH2(OO•) (Z) ISCl6 ClCH2CH=C(CH3)CH2O•

ISON7 CH2(ONO2)C(CH3)=CHCH2(OO•) (E) ISON Hydroxyalkylnitrates and alkylnitratesISON8 CH2(ONO2)C(CH3)=CHCH2(OO•) (E) IALD1 C5 hydroxycarbonyls,

e.g. CHOC(CH3)=CHCH2OHCO51 C5 carbonyls

C4 Compounds

CHC C4 hydroxycarbonyls and dihydroxy carbonylsMACR methacroleinMVK methyl vinyl ketone

CI2 CH2=CHC(CH3)(-O-O) OO

CI3 CH2=CHC(CH3)(-O-O) OO

CI4 CH2=C(CH3)CH(-O-O) OO

CI5 CH2=C(CH3)CH(-O-O)O

O

Table 2. The New Isoprene Oxidation Mechanism

Rate constantA Ref.

OH–Isoprene Reactions

ISOP + OH = 0.56 ISOA + 0.02 ISOB + 0.05 ISOC + 0.37 ISOD 1.00 × 10−10 [9,17]ISOA + O2 = 0.60 ISO2A + 0.20 ISO2E + 0.20 ISO2G 2.20 × 10−12 [18,21]ISOB + O2 = CO51 + HO2 2.50 × 10−11 [23]ISOC + O2 = 0.40 CO51 + 0.40 HO2 + 0.60 ISO2C 1.40 × 10−12 [18,23]ISOD + O2 = 0.78 ISO2D + 0.11 ISO2F + 0.11 ISO2H 5.90 × 10−13 [18,23]ISO2A + NO = 0.08 ISON + 0.92 MVK + 0.92 CH2O + 0.92 HO2 + 0.92 NO2 1.00 × 10−11 [5,19,40,41]ISO2C + NO = 0.08 ISON + 0.92 MACR + 0.92 CH2O + 0.92 HO2 + 0.88 NO2 1.00 × 10−11 [5,19,40,41]ISO2D + NO = 0.08 ISON + 0.92 MACR + 0.92 CH2O + 0.92 HO2 + 0.88 NO2 1.00 × 10−11 [5,19,40,41]ISO2E + NO = 0.08 ISON + 0.92 CHC + 0.92 CH2O + 0.92 HO2 + 0.92 NO2 1.00 × 10−11 [15,21,40,41]ISO2F + NO = 0.08 ISON + 0.92 IALD1 + 0.92 HO2 + 0.92 NO2 1.00 × 10−11 [15,21,40,41]ISO2G + NO = 0.08 ISON + 0.92 IALD1 + 0.92 HO2 + 0.92 NO2 1.00 × 10−11 [15,21,40,41]ISO2H + NO = 0.08 ISON + 0.92 IALD1 + 0.92 HO2 + 0.92 NO2 1.00 × 10−11 [15,21,40,41]ISO2A + ISO2E = 0.9 MVK + 1.1 CHC + 1.8 CH2O + 1.8 HO2 2.50 × 10−12 [26,39]ISO2A + ISO2F = 0.9 MVK + 1.1 CHC + 1.8 CH2O + 1.8 HO2 4.70 × 10−12 [26,39]ISO2C + ISO2A = 0.9 MACR + 0.2 CHC + 1.8 CH2O + 1.8 HO2 + 0.9 MVK 2.60 × 10−12 [26,39]ISO2C + ISO2C = 1.2 MACR + 0.8 CHC + 1.2 CH2O + 1.2 HO2 4.80 × 10−12 [26,39]ISO2C + ISO2D = 1.2 MACR + 0.8 CHC + 1.2 CH2O + 1.2 HO2 5.20 × 10−12 [26,39]ISO2C + ISO2E = 0.6 MACR + 1.4 CHC + 0.6 CH2O + 1.2 HO2 4.70 × 10−12 [26,39]ISO2C + ISO2F = 0.6 MACR + 1.4 CHC + 0.6 CH2O + 1.2 HO2 3.70 × 10−12 [26,39]

continued

142


Table 2. continued

Rate constantA Ref.

ISO2D + ISO2A = 0.9 MACR + 0.3 CHC + 1.8 CH2O + 1.8 HO2 + 0.9 MVK 3.90 × 10−12 [26,39]ISO2E + ISO2E = 2 CHC + 1.2 CH2O + 1.2 HO2 3.90 × 10−12 [26,39]ISO2F + ISO2F = 2 CHC + 1.2 CH2O + 1.2 HO2 2.80 × 10−12 [26,39]

O3–Isoprene Reactions

ISOP + O3 = 0.26 CI2 + 0.205 CI3 + 0.21 CI4 + 0.02 CI5 + 0.12 MVK + 1.60 × 10−7 [27,28]0.21 MACR + 0.33 CH2-O-O + 0.67 CH2O

CH2-O-O = 0.05 CH•2(OO•) + 0.95 CH2-O-O†B 9.62 [27,28]

CH2-O-O† + H2O = HCOOH + HO2 + OH 4.00 × 10−16 [39]CH2-O-O† + NO (NO2) = CH2O + NO2 (NO3) 4.00 × 10−16 [39]CH•

2(OO•) = HCOOH 2.00 × 10−1 [42,43]CI2 = 0.08 CH2=CHC•(CH3)(OO•) + 0.38 MVK + 0.38 OH + 0.54 CI2† 1.50 × 10−1 [27,28]CI2† + NO (NO2, etc.) = MVK + NO2 (NO3, etc.) 4.00 × 10−16 [39]CI3 = 0.60 CH2=CHC•(CH3)(OO•) + 0.06 MVK + 0.06 OH + 0.34 CI3† 9.40 × 10−1 [27,28]CI3† + NO (NO2, etc.) = MVK + NO2 (NO3, etc.) 4.00 × 10−16 [39]CHC•(CH3)(OO•) = MVK + O 2.00 × 10−1 [42,43]CI4 = 0.80 CH2=C(CH3)CH•(OO•) + 0.20 CI4† 4.5 [27,28]CI4† + H2O=CH2=C(CH3)COOH + HO2 + OH 4.00 × 10−16 [39]CI4† + NO (NO2, etc.) = MACR + NO2 (NO3, etc.) 4.00 × 10−16 [39]CI5 = 0.63 CH2=C(CH3)CH•(OO•) + 0.37 CI5† 3.5 [27,28]CI5† + H2O = CH2=C(CH3)COOH + HO2 + OH 4.00 × 10−16 [39]CI5† + NO (NO2, etc.) = MACR + NO2 (NO3, etc.) 4.00 × 10−16 [39]CH2=C(CH3)CH•(OO•) = MACR + O 2.00 × 10−1 [42,43]

NO3–Isoprene Reactions

ISOP + NO3 = 0.285 ISON1 + 0.044 ISON2 + 0.005 ISON3 + 0.01 ISON4 + 7.20 × 10−13 [29–31]0.32 ISON5 + 0.032 ISON6 + 0.074 ISON7 + 0.23 ISON8

ISON1 + NO = ISN1 + NO2 6.80 × 10−12 [39]ISON2 + NO = ISN2 + NO2 6.80 × 10−12 [39]ISON3 + NO = ISN3 + NO2 6.80 × 10−12 [39]ISON4 + NO = ISN4 + NO2 6.80 × 10−12 [39]ISON5 + NO = ISN5 + NO2 6.80 × 10−12 [39]ISON6 + NO = ISN6 + NO2 6.80 × 10−12 [39]ISON7 + NO = ISN7 + NO2 6.80 × 10−12 [39]ISON8 + NO = ISN8 + NO2 6.80 × 10−12 [39]ISN1 = 0.74 CH•

3 + 0.74 CH2=CHCOCH2(ONO2) + 0.26 MVK + 0.26 CH2O + 0.26 NO2 2.80 × 106 [31]ISN2 = MACR + CH2O + NO2 8.00 × 106 [31]ISN3 = MVK + CH2O + NO2 0.14 [31]ISN4 = MACR + CH2O + NO2 6.10 × 107 [31]ISN5 = CH2(ONO2)C(CH3)=CH• + CH2O 1.46 [31]ISN6 = CH2(ONO2)CH=C•(CH3) + CH2O 16.60 [31]ISN7 = CH2(ONO2)CH=C•(CH3) + CH2O 32.90 [31]ISN8 = CH2(ONO2)C(CH3)=CH• + CH2O 7.00 × 10−3 [31]

Cl–Isoprene Reactions

ISOP + Cl = 0.15 ISOCl1 + 0.02 ISOCl2 + 0.07 ISOCl3 + 0.25 ISOCl4 + 4.27 × 10−10 [32–34,38]0.19 ISOCl5 + 0.17 ISOCl6 + 0.15 CH2=C(CH•

2)CH=CH2 + 0.15 HClCH2=C(CH•

2)CH=CH2 + O2 = CH2=C(CHO)CH=CH2 + HO2 2.50 × 10−13 [32,38]ISOCln + NO = ISCln + NO2 (n = 1–6) 6.80 × 10−12 [39]ISCl1 = 0.02 CH•

3 + 0.15 CH2=CHCOCH2Cl + 0.98 MVK + 0.98 ClCH•2 9.00 × 103 [32,37]

ISCl2 + O2 = HO2 + CH2=CHC(CH3)(Cl)CHO 1.00 × 10−15 [32,37]ISCl3 + O2 = HO2 + CH2=C(CH3)CH(Cl)CHO 1.00 × 10−15 [32,37]ISCl4 + O2 = HO2 + CH2=C(CH3)COCH2(Cl) 1.00 × 10−15 [32,37]ISCl5 + O2 = HO2 + CH2(Cl)C(CH3)=CCH2O 1.00 × 10−15 [32,37]ISCl6 + O2 = HO2 + CH2(Cl)CH=C•CH3 1.00 × 10−15 [32,37]

A At 300 K. Units are s−1 for unimolecular reactions and cm3 molec−1 s−1 for bimolecular reactions.B The dagger (†) represents the stabilized state.

stabilized and 5% undergoes ring closure to form dioxiranes.For CI2 and CI3, 54% and 34% undergo stabilization and 8%and 60% undergoes ring closure to form dioxiranes respec-tively. The OH yields are 38% for CI2 and 6% for CI3. For

CI4 and CI5, 20% and 37% are stabilized, respectively. Theremaining undergoes ring closure to form dioxiranes. Thedioxiranes formed from CI1 further decompose to produceformic acid. Other dioxiranes also undergo decompositions

143

J. Fan and R. Zhang

OO.

OH

OH

OO.

OO.

OH

OO. OH

OH

O2

OH OO.

OH

OO.

O2

OH O

OO.

OH

OH

OH

OH

OOH

�

0.56

0.02

0.05

0.37

0.60

0.20

O2

0.40

0.60

OH(ISOA)

(ISOB)

(ISOC)

(ISOD)

0.20

O. OH OH OHO2

O2

�

Organic nitrateNO0.08

0.92

O.

OH

O

OH

HO2O2

�

Organic nitrateNO

0.08

0.92

O.

OH

O .CH2OH�

Organic nitrateNO0.08

0.92

O HO2CH2OO2

��

Organic nitrate

OH

NO0.08

0.92

OH O. O OH

HO2O2 �

Organic nitrateNO

0.08

0.92

O

OH

O.

OH

.CH2OH�

Organic nitrateNO

0.08

0.92

O.CH2OH

O.

OH

�

Organic nitrateNO

0.08

0.92

0.78

0.11

0.11

(ISO2G)

(ISO2E)

(ISO2C)

(ISO2D)

(ISO2H)

(ISO2F)

(ISO2A)

Fig. 1. Mechanistic representation of the reaction pathways of isoprene initiated by OH.

CH3O

H

OO

CH3

OH O

CH3

CH2O

CH2O

CH2O

CH2O

OH

H O

H

CH3

OO

OH

H O

H

CH3

OO

CH3O

O

O O

OCH3 O

OH O

CH3

O3

CH2O

HCOOH

OO

O

MVK

MVK

MACR

MACR

Stabilization

Stabilization

Stabilization

Stabilization

Stabilization

0.125

0.26

0.205

0.18

0.21

0.02

0.95

0.05

0.54

0.08

0.38

0.34

0.60

0.06

0.20

0.80

0.37

0.63

�

�

OMVK �

HCOOH HO2 OH��

�

O OH�

OMVK ��

CH2�C(CH3)COOH HO2 OH��

CH2�C(CH3)COOH HO2 OH��

�

OHC O �

�

�

OMACR �

OMACR �

NO, etc.

NO, etc.

(CI1)

(CI3)

(CI2)

(CI4)

(CI5)

(CI1)

NO, etc.

H2O

NO, etc.

H2O

NO, etc.

H2O

Fig. 2. Mechanistic representation of the reaction pathways of isoprene initiated by O3.

to produce MACR (MVK) and oxygen.[42,43] The stabilizedCI1, CI4, and CI5 have two major pathways, the reactionswith H2O and NO (NO2, etc.).[39] The former pathway pro-duces organic acid and OH; the latter produces aldehydes

or ketones. The stabilized CI2 and CI3 only react with NO(NO2, etc.) to form MVK.

The isoprene ozonolysis mechanism presented here ismuch more explicit than the previous mechanisms,[5–7] in

144


OO.

NO3

OO.

NO3

OO.

NO3

NO3

OO.

O2

NO3

NO3

O3NCH2O.

CH3 O

CH2NO3

ONO2

NO3

NO

NO

NO

NO

NO

NO

NO

NO

O2

O2

NO3 OO.

NO3

OO.

NO3

OO.

OO. NO3

O2

O.

NO3

NO3 O.

NO3

NO2O

NO3

O.

CH2O.

NO3

NO2O

CH2NO3

O.O NO2

NO3

O.

O. NO3

NO3

�

�

� �

�

�

�

� �

�

�

�

0.835

0.005

0.01

0.15

0.74

0.26 CH2O

0.285

0.32

0.23

0.044

0.032

0.074

(ISON1)

(ISON5)

(ISON8)

(ISON3)

(ISON4)

(ISON2)

(ISON6)

(ISON7)

CH2O

CH2O

CH2O

CH2O

NO3

.CH2O

NO3

.

Fig. 3. Mechanistic representation of the reaction pathways of isoprene initiated by NO3.

which only the final products were presented based on theenvironmental chamber experiments.

NO3–Isoprene Reactions

The detailed NO3–isoprene reaction mechanism is also pro-vided inTable 2. Fig. 3 depicts the schematic representation ofthe NO3–isoprene reaction mechanism. Suh et al. obtained anoverall rate constant of the NO3–isoprene reaction using com-bined experimental and theoretical approaches, with a valueof 7.2 × 10−13 cm3 molec−1 s−1.[29] The NO3–isoprene reac-tion proceeds by NO3 addition to the >C C< bonds,forming four NO3–isoprene adduct radicals. Under atmo-spheric conditions, the nitrooxyalkyl radicals react primar-ily with oxygen molecules to form nitrooxyalkyl peroxyradicals. Similar to the OH–isoprene reaction, there aretwo addition centres (β or δ to the NO3 position) for theterminal addition adducts and two of the nitrooxyalkylperoxy radicals possess (E) and (Z) configurations. There-fore, a total of eight nitrooxyalkyl peroxy radicals areformed (ISON1–ISON8). The relative branching ratiosare 0.285:0.044:0.005:0.01:0.32:0.032:0.074:0.23.[31] In thepresence of NO, the subsequent reaction of the nitrooxyalkylperoxy radicals leads to eight nitrooxyalkoxy radicals (ISN1–ISN8). Zhao and Zhang[31] investigated the decompositionpathways of those nitrooxyalkoxy radicals by a theoreti-cal approach. ISN1 has two major decomposition pathways:one forms 1-nitroxy-2-butan-3-one and methyl radical, andthe other forms MVK, CH2O, and NO2. Decompositionof ISN2 and ISN4 is very fast, with the rate constants of8.0 × 106 and 6.1 × 107 s−1, respectively, to produce MACR,CH2O, and NO2. The decomposition of ISN3 is relativelyslow (0.14 s−1), yielding MVK, CH2O, and NO2. Both ISN5and ISN8 decompose to CH2O and the relevant radicals, butthe rate constant for ISN5 (1.46 s−1) is much faster than that

for ISN8 (7.0 × 10−3 s−1). ISN6 and ISN7 also decomposeto form CH2O and the corresponding radicals, with compa-rable decomposition rates.[31] The work on the nitrooxyalkylperoxy radicals and the decompositions of nitrooxyalkoxyradicals is unpublished from our research group.

Cl–Isoprene Reactions

Previous isoprene oxidation mechanisms[5–7] have notincluded the Cl–isoprene reactions. A recent study hasindicated the importance of natural and anthropogenicsources of chlorine and the related ozone formation in theatmosphere.[44] Because of the high reactivity of chorineatoms, the Cl–isoprene reaction is considerably faster thanthe reactions of isoprene with OH and O3, with a rate con-stant of 4.27 × 10−10 cm3 molec−1 sec−1 (considering bothCl addition and H abstraction).[35,38] The initial reactionbetween isoprene and Cl proceeds mostly by Cl addition,forming the Cl–isoprene adduct radicals. Alternatively, iso-prene also undergoes methyl-hydrogen abstraction to formHCl with ∼15% yield.[32–34] Under atmospheric conditions,the Cl–isoprene adducts react with oxygen molecules toform the chloroalkenyl peroxy radicals. Six chloroalkenylperoxy radicals are formed (ISOCl1–ISOCl6, Fig. 4), withbranching ratios of 0.15, 0.02, 0.07, 0.25, 0.19, and 0.17,respectively.[36] In the presence of NO, the chloroalkenylperoxy radicals react with NO, leading to the formationof chloroalkenyl alkoxy radicals (ISCl1–ISCl6). The dom-inant reaction pathway of ISCl1 is to undergo C–C fission,and the major products are MVK and CH2Cl radicals.[32,37]The product yield of MVK is ∼14%, consistent with themeasurements.[32] Zhang et al. predicted ∼25% 1-chloro-3-methyl-3-butene-2-one (CMBO) formation, which is formedfrom ISCl4 reaction with O2.[37] The most likely reactionpathways for ISCl2, ISCl3, ISCl5, and ISCl6 are also to

145

J. Fan and R. Zhang

OO.

Cl

.OO

Cl

OO.

Cl

Cl

O2

O2

O2

O2

Cl

OO.

NO

NO O.Cl

NO

NO

NO

NO

Cl

.OOCl

.O

ClO.

O.Cl

ClO.

CH3.

OClCH2.

HO2

HO2

NO CHO

O2

O2O2

O

CH2Cl

Cl

O

Cl

O

Cl

O

HO2

HO2

HO2

HO2

ClO

Cl

O.Cl

OO.

O2

O2

O2

O2

Cl

CHO

CH2. CH2OO. CH2O.

Cl

.

Cl

.

Cl

.�

�

�

�

�

�

�

�

�

�

0.34

0.02

0.07

0.42

0.15

0.19

0.25

0.17

0.02

0.98

0.15

HCl

(ISOCl1)

(ISOCl5)

(ISOCl2)

(ISOCl3)

(ISOCl4)

(ISOCl6)

Cl

.

Fig. 4. Mechanistic representation of the reaction pathways of isoprene initiated by Cl.

react with O2 and form the CMBO isomers. The relativeyield of CMBO to its other isomers is ∼36%.[37] ISCl1reacts differently because of the lack of an abstractablehydrogen atom. The methyl-hydrogen abstraction in the Cl–isoprene reactions yields CH2 C(CH•

2)CH CH2 radicals,which react with O2 then NO to form alkoxy radicals. Thealkoxy radicals undergo hydrogen abstraction by O2, yieldingCH2 C(CHO)CH CH2.[32,38]

Results and Discussion

The new isoprene oxidation mechanism provides detailedkinetics and mechanisms of the isoprene reactions with OH,O3, NO3, or Cl. To evaluate the impact of the updated mecha-nism on O3 and OH formation, we employed the new isoprenemechanism in a box model and compared the results fromthose of the previous mechanisms.[5–7] The model results ofthe various mechanisms with integration for a five-day periodwere compared. We also examined the differences betweenthe new mechanism and a NCAR box model (MM2.1).[39]

A Fortran 90 program incorporating the Gear algorithmwas operated on an SGI ‘Octane’ workstation running underIRIX 6.5.[45] The algorithm by Madronich was employed tocalculate the photolysis frequencies.[46] The model calcu-lations for high and low NOx (NO + NO2) scenarios wereperformed. The initial NOx concentration was 10 ppb for thehigh NOx case and 50 ppt for the low NOx case. For both sce-narios, the initial isoprene concentration was 5 ppb. The H2Ovolume mixing ratio was 0.6%.Temperature was at 288 K andthe total pressure was 1013 hPa. The scenarios were assumedto be static, meaning that isoprene and NOx were present atthe beginning of the simulation and allowed to react with-out subsequent emissions. The model runs were initiated atmidnight.

Since the Cl–isoprene reactions were not included inthe previous mechanisms, we compared the model differ-ences with and without the Cl–isoprene reactions in thenew mechanism. Also, we focussed on comparisons of thesimulated O3 and OH formation predicted by the variousmechanisms.

High NOx Scenarios

Fig. 5 shows the model results in the high NOx scenario.The simulations of the accumulated ozone concentrationsduring the five-day period are shown in Fig. 5a. Ozoneis produced at daytime and the trend of the accumulatedconcentrations increases with time in the high NOx initial-ization scenario, consistent for the different mechanisms.Initially O3 concentrations increase very fast, because iso-prene is substantially consumed. The maximum ozone pro-duction rate is ∼1.3 × 1011 molec cm−3 h−1 in the CM and∼6.5 × 1010 molec cm−3 h−1 in the new mechanism. For thePSM and MM2.1, the maximum ozone production ratesare ∼6.0 × 1010 and 2.8 × 1010 molec cm−3 h−1, respec-tively. The model calculations from the CM have the largestincreases in the ozone concentration because of its overpre-diction of the initial rate of NO oxidation.[7] Also, this mech-anism predicts the highest final ozone concentration, whichis ∼12% higher than that by the new mechanism and 21%higher than that by the PSM.This is probably because of muchhigher peroxy radical yields in the CM. The new isoprenemechanism predicts relatively high O3 formation as well,especially in days three and four. Relative to the predictions bythe CM, the deviations are small in the last three days (within5% on the average). Compared to the calculations from thePSM, the calculated O3 levels from the new mechanismare ∼8% higher. Apparently, the NCAR MM2.1 mechanism

146


(a)

0.0E�00

1.0E�12

2.0E�12

3.0E�12

0.0 2.0 4.0 6.0Time (days)

Con

cent

ratio

n (m

olec

ule

cm�

3 )

NewPSMCMMM2.1

0.0 2.0 4.0 6.0Time (days)

(b)

0.0E�00

4.0E�06

8.0E�06

1.2E�07

1.6E�07

Con

cent

ratio

n (m

olec

ule

cm�

3 )

NewPSMCMMM2.1

Fig. 5. Comparisons of (a) O3 and (b) OH concentrations in the high NOx scenario predicted by the various mechanisms.

(a)

2.0E�11

4.0E�11

6.0E�11

8.0E�11

0.0 2.0 4.0 6.0Time (days)

Con

cent

ratio

n (m

olec

ule

cm�

3 )

New

PSM

CM

MM2.1

Con

cent

ratio

n (m

olec

ule

cm�

3 )

(b)

0.0E�00

1.0E�06

2.0E�06

3.0E�06

0.0 2.0 4.0 6.0Time (days)

New

PSM

CM

MM2.1

Fig. 6. Comparisons of (a) O3 and (b) OH concentrations in the low NOx scenario predicted by the various mechanisms.

(a)

0.0E�00

1.0E�12

2.0E�12

3.0E�12

0.0 2.0 4.0 6.0Time (days)

Con

cent

ratio

n (m

olec

ule

cm�

3 )

New ClNew

Con

cent

ratio

n (m

olec

ule

cm�

3 )

(b)

0.0E�00

4.0E�06

8.0E�06

1.2E�07

1.6E�07

0.0 2.0 4.0 6.0Time (days)

New ClNew

Fig. 7. Comparisons of (a) O3 and (b) OH concentrations with (New Cl) and without (New) the Cl–isoprene reactions.

predicts much lower ozone formation than any other mech-anisms, with a 20% difference from the new mechanismthroughout the five-day period. The reasons for the differenceare primarily from the hydroperoxy self- and cross-reactions

and much higher decomposition rates of nitrooxyalkoxy rad-icals in the new mechanism: Both producing more O3 byincreasing NO2 concentration. Since the CM was improvedto fit the environmental chamber data, it is believed to produce

147

J. Fan and R. Zhang

better predictions in ozone formation.[6,7] The new mecha-nism produce a good agreement with the CM in the predictionof ozone formation, except that the CM overpredicts theinitial O3 formation under the high NOx conditions.[7]

The model simulations of OH exhibit diurnal cycles(Fig. 5b), which are consistent with the actual OH varia-tion in the atmosphere. On the first day, the new mechanismpredicts the highest OH concentration, ∼20% higher thanthe PSM and 50% higher than MM2.1. Overall, the newisoprene mechanism predicts OH concentrations in betteragreement with the PSM mechanism during the simulationperiod. MM2.1 produces relatively lower OH predictions thanthe new mechanism and the PSM, with a 25% difference onaverage in the daytimes. This appears to correspond to the20% lower O3 predictions by MM2.1. The CM predicts verylow peak OH formation in the first four days, which is con-sistent with the simulations performed by Carter, that is, thiscondensed mechanism predicts about half of the maximumOH level as its detailed mechanism in the two-day simulationunder the high NOx conditions.[7] Note that there is a narrowOH peak when the sun rises on the first day, because of therapid production of O3.

Low NOx Scenarios

Fig. 6 shows the O3 and OH simulations during the five-daymodel period in the low NOx scenario. Ozone concentra-tions decrease during the model period in this case. Becausethe NO–peroxy reactions are of little importance under thelow NOx conditions, the production of ozone is inhibited andthe ozone formed is consumed by reacting with other com-pounds in the system.Throughout the simulated time, the newisoprene mechanism shows a good agreement with the CMand PSM in the ozone simulations, and produces the high-est O3 predictions over most of the simulated time (Fig. 6a).The maximum difference between O3 predictions by thosethree mechanisms is less than 4%. However, the MM2.1mechanism shows relatively larger deviations from the othermechanisms. Especially after the first day, the differencesbecome greater with elapsed time. The largest deviation isover 20% relative to the new mechanism. The reasons for thedeviation have been stated in the above high NOx scenarios.Therefore, except for MM2.1, there are no significant dif-ferences between the other three mechanisms in predictingozone formation under the low NOx conditions.

The OH simulations in the new mechanism are in goodagreement with CM and PSM (Fig. 6b).The noticeable devia-tions take place at the OH peaks.The peak OH concentrationspredicted by the new mechanism and the CM agree within20% over the first two days and within 10% over the followingthree days. The PSM predicts ∼10% higher peak OH concen-trations than the CM and the new mechanism. As shown inFig. 6b, MM2.1 predicts much higher OH concentrations,especially in the first three days. It is likely that the decom-position rates of carbonyl oxides are much higher in MM2.1and the decomposition reactions are dominant for carbonyloxides under the low NOx conditions, which are the directsources of OH in the isoprene atmospheric reactions.

Impact of the Cl–Isoprene Reaction

The new mechanism including the Cl–isoprene reactions isreferred to as ‘New Cl’ mechanism (Fig. 7). The New Clmechanism was considered in the previous high NOx sce-nario.The initial chlorine atom concentration was assumed tobe 1 × 105 molec cm−3. The comparisons of O3 and OH for-mation due to the Cl–isoprene reactions are shown in Fig. 7.

Compared with the calculations without the Cl–isoprenereactions, ozone production rate is slightly faster for theNew Cl mechanism during the daytime of day one (Fig. 7a).The ozone production rate with the New Cl mechanism is3.7 × 1010 molec cm−3 h−1 from 12:00 to 18:00, comparedwith 2.9 × 1010 molec cm−3 h−1 without the Cl–isoprenereactions. Because of the higher reactivity of chlorine atom,the Cl–isoprene reaction is faster than other isoprene oxi-dation reactions. It accelerates the ozone production rateat the early stage of the model period. With the O3 reac-tions included, the differences become less, and mostly fadetowards the end of the five-day simulation period. As shownin Fig. 7b, OH concentrations are more sensitive to the Cl–isoprene reaction than O3 concentrations. The New Cl mech-anism yields a lower OH prediction on the first simulated day,with the largest difference of 10%. On the other hand, the NewCl mechanism predicts a higher peak OH concentration (5%)than the mechanism without the Cl–isoprene reactions dur-ing the last three days. This most likely occurs because thereactions of Cl and Cl-related radicals produce more HO2,yielding more OH. We also examined the differences in theO3 and OH formation due to the Cl–isoprene reactions underthe low NOx conditions. Similar O3 and OH performanceswere observed, except that the differences were smaller.

Conclusions

In this study the isoprene oxidation mechanism has beenupdated to accommodate most recent experimental and theo-retical developments. The new isoprene mechanism providesthe explicit reaction steps and detailed intermediates for iso-prene oxidation reactions with OH, O3, NO3, and Cl. We havecompared the box model performances of the new mechanismwith Carter’s four-product condensed mechanism, Paulsonand Seinfeld’s mechanism, and the NCAR MM2.1 under bothhigh and low NOx conditions. The O3 and OH sensitivitiesto the Cl–isoprene reaction have also been investigated.

Under high NOx conditions, the new mechanism pro-duces O3 formation close to Carter’s condensed mechanismthroughout most of the simulated time but gives better O3

predictions at the initial stage. Compared with Paulson andSeinfeld’s mechanism and MM2.1, it predicts ∼8% and20% higher O3 concentrations, respectively. The OH con-centrations predicted by the new mechanism are close to thesimulation with the Paulson and Seinfeld’s mechanism, butmuch better than that with Carter’s condensed mechanism.The peak OH concentrations predicted by the new mechanismare ∼25% averagely higher than those predicted by MM2.1.Therefore, the new mechanism gives better O3 and OH sim-ulations than other mechanisms under high NOx conditions.In the low NOx scenarios, the new isoprene chemistry shows

148


a good agreement with Carter’s condensed mechanism andPaulson and Seinfeld’s mechanism in the ozone simulationsthroughout the simulated time, and produces the highest O3

predictions during most of the model run, ∼20% higher thanthe simulations of MM2.1. OH simulations of the new mech-anism under low NOx conditions are also in good agreementwith those of Carter’s condensed mechanism and Paulson andSeinfeld’s mechanism.The peak OH concentrations predictedby the new mechanism is ∼10% lower than those predictedby Paulson and Seinfeld’s mechanism and less than 5% lowerthan those predicted by Carter’s mechanism. The ozone pro-duction is slightly promoted by including the Cl–isoprenereaction, which also has an impact on the OH concentrations.

The new mechanism presented in this paper representsthe current state-of-art isoprene atmospheric chemistry.Compared with other isoprene mechanisms, OH– and O3–isoprene reactions have been updated, and detailed NO3– andCl–isoprene mechanisms have been added. Although consid-erable progress has been made recently in understanding ofthe isoprene atmospheric reactions, some uncertainties suchas the chemistry under very low NOx conditions remain,which may affect model predictions. Since the isoprenechemistry has crucial implications on regional air quality,acid deposition, and climate due to ozone and potentialsecondary organic aerosol formation, an accurate represen-tation of the isoprene chemistry in atmospheric models is ofsubstantial importance.

Acknowledgements

This work was supported by the Robert A. Welch Founda-tion (A-1417) and the USA Environmental ProtectionAgency(R03–0132).

References

[1] A. Guenther, C. N. Hewitt, D. Erickson, R. Fall, C. Geron,T. Graedel, P. Harley, L. Klinger, et al., J. Geophys. Res. 1995,100, 8873. doi:10.1029/94JD02950

[2] U. Pöschl, R. V. Kuhlmann, N. Poisson, P. J. Crutzen, J. Atmos.Chem. 2000, 37, 29. doi:10.1023/A:1006391009798

[3] IPCC Scientific Assessment Climate Change 1992 (Ed.J. T. Houghton) 1992 (Cambridge University Press: Cambridge).

[4] M. Trainer, E. Y. Hsie, S. A. McKeen, R. Tallamraju,D. D. Parrish, F. C. Fehsenfeld, S. C. Liu, J. Geophys. Res.1987, 92, 11879.

[5] S. E. Paulson, J. H. Seinfeld, J. Geophys. Res. 1992, 97, 20703.[6] W. P. L. Carter, R. Atkinson, Int. J. Chem. Kinet. 1996, 28, 497.[7] W. P. L. Carter, Atmos. Environ. 1996, 30, 4275. doi:10.1016/

1352-2310(96)00088-X[8] S. Madronich, J. G. Calvert, J. Geophys. Res. 1990, 95, 5697.[9] R. Zhang, I. Suh, A. D. Clinkenbeard, W. Lei, S. W. North,

J. Geophys. Res. 2000, 105, 24627. doi:10.1029/2000JD900330[10] W. S. McGivern, I. Suh, A. D. Clinkenbeard, R. Zhang,

S. W. North, J. Phys. Chem. A 2000, 104, 6609. doi:10.1021/JP001163C

[11] R. Zhang, W. Lei, J. Chem. Phys. 2000, 113, 8574. doi:10.1063/1.1318736

[12] D. Zhang, R. Zhang, C. Church, S. W. North, Chem. Phys. Lett.2001, 343, 49. doi:10.1016/S0009-2614(01)00654-6

[13] M. Sprengnether, K. L. Demerjian, N. M. Donahue,J. G. Anderson, J. Geophys. Res. 2002, 107, D15,4269.doi:10.1029/2001JD000716

[14] D. Zhang, R. Zhang, S. W. North, J. Phys. Chem. A 2003, 107,11013. doi:10.1021/JP0360016

[15] J. Zhao, R. Zhang, E. C. Fortner, S. W. North, J. Am. Chem.Soc. 2004, 126, 2686. doi:10.1021/JA0386391

[16] W. Lei, A. Derecskei-Kovacs, R. Zhang, J. Chem. Phys. 2000,113, 5354. doi:10.1063/1.1290020

[17] W. Lei, R. Zhang, W. S. McGivern, A. Derecskei-Kovacs,S. W. North, Chem. Phys. Lett. 2000, 326, 109. doi:10.1016/S0009-2614(00)00736-3

[18] W. Lei, R. Zhang, W. S. McGivern, A. Derecskei-Kovacs,S. W. North, J. Phys. Chem. A 2001, 105, 471. doi:10.1021/JP0027039

[19] W. Lei, R. Zhang, J. Phys. Chem A 2001, 105, 3808. doi:10.1021/JP0041353

[20] T. S. Dibble, J. Phys. Chem. A 1999, 103, 8559. doi:10.1021/JP991879B

[21] T. S. Dibble, J. Phys. Chem. A 2002, 106, 6643. doi:10.1021/JP025682M

[22] D. Zhang, R. Zhang, J. Park, S. W. North, J. Am. Chem. Soc.2002, 124, 9600. doi:10.1021/JA0255195

[23] J. Park, C. G. Jongsma, R. Zhang, S. W. North, Phys. Chem.Chem. Phys. 2003, 5, 3638. doi:10.1039/B306580D

[24] J. Park, J. C. Stephens, R. Zhang, S. W. North, J. Phys. Chem. A2003, 107, 6408. doi:10.1021/JP0303321

[25] J. Zhao, R. Zhang, S. W. North, Chem. Phys. Lett. 2003, 369,204. doi:10.1016/S0009-2614(02)02006-7

[26] M. E. Jenkin, A. A. Boyd, R. Lesclaux, J. Atmos. Chem. 1998,29, 267. doi:10.1023/A:1005940332441

[27] D. Zhang, R. Zhang, J. Am. Chem. Soc. 2002, 124, 2692.doi:10.1021/JA011518L

[28] D. Zhang, W. Lei, R. Zhang, Chem. Phys. Lett. 2002, 358, 171.doi:10.1016/S0009-2614(02)00260-9

[29] I. Suh, W. Lei, R. Zhang, J. Phys. Chem. A 2001, 105, 6471.doi:10.1021/JP0105950

[30] D. Zhang, R. Zhang, J. Chem. Phys. 2002, 116, 9721.doi:10.1063/1.1476695

[31] J. Zhao, R. Zhang, unpublished results.[32] M. L. Ragains, B. J. Finlayson-Pitts, J. Phys. Chem. A 1997,

101, 1509. doi:10.1021/JP962786M[33] W. Lei, R. Zhang, J. Chem. Phys. 2000, 113, 153. doi:10.1063/

1.481782[34] I. Suh, R. Zhang, J. Phys. Chem. A 2000, 104, 6590.

doi:10.1021/JP000605H[35] W. Lei, D. Zhang, R. Zhang, T. L. Molina, J. M. Molina, Chem.

Phys. Lett. 2002, 357, 45. doi:10.1016/S0009-2614(02)00437-2[36] W. Lei, R. Zhang, T. L. Molina, J. M. Molina, J. Phys. Chem. A

2002, 106, 6415. doi:10.1021/JP025799A[37] D. Zhang, R. Zhang, T. D. Allen, J. Chem. Phys. 2003, 118,

1794. doi:10.1063/1.1531660[38] J. J. Orlando, G. S. Tyndall, E. C. Apel, D. D. Riemer,

S. E. Paulson, Int. J. Chem. Kinet. 2003, 35, 334. doi:10.1002/KIN.10135

[39] S. Madronich, J. G. Calvert, P. Gregoire, NCAR Technical NoteTN-333+STR 1994 (UCAR: Boulder, CO).

[40] E. C. Tuazon, R. Atkinson, Int. J. Chem. Kinet. 1990, 22, 1221.[41] X. H. Chen, D. Hulbert, P. B. Shepson, J. Geophys. Res. 1998,

103, 25563. doi:10.1029/98JD01483[42] P. Aplincourt, M. F. Ruiz-López, J. Am. Chem. Soc. 2000, 122,

8990. doi:10.1021/JA000731Z[43] H. Niki, P. D. Maker, C. M. Savage, L. P. Breitenbach,

M. D. Hurley, J. Phys. Chem. 1987, 91, 941.[44] K. W. Oum, M. J. Lakin, D. O. DeHaan, T. Brauers,

B. J. Finlayson-Pitts, Science 1998, 279, 74. doi:10.1126/SCIENCE.279.5347.74

[45] C. W. Gear, Prentice-Hall Series in Automatic Computation no.17 1971 (Prentice-Hall: Eaglewood Cliffs, NJ).

[46] S. J. Madronich, J. Geophys. Res. 1987, 92, 9740.

149

Atmospheric Oxidation Mechanism of Isoprene

Documents

Transcript of Atmospheric Oxidation Mechanism of Isoprene