Gas&hyphen;phase products and secondary aerosol yields from the photooxidation of 16 different terpenes

Gas-phase

oducts

and

secondary

aer

osol

yields

the

photooxidation of 16 different terpenes

Anita Lee,

Allen H. Goldstein,

Jesse H. Kroll,

Nga L. Ng,

Varuntida Varutbangkul,

Richard C. Flagan,

and John H. Seinfeld

Received 4 January 2006; revised 18 April 2006; accepted 16 May 2006; published 7 September 2006.

[

]

The photooxidation of isoprene, eight monoterpenes, three oxygenated monoterpenes,

and four sesquiterpenes were conducted individually at the Caltech Indoor Chamber

Facility under atmospherically relevant HC:NO

ratios to monitor the time evolution and

yields of SOA and gas-phase oxidation products using PTR-MS. Several oxidation

products were calibrated in the PTR-MS, including formaldehyde, acetaldehyde, formic

acid, acetone, acetic acid, nopinone, methacrolein + methyl vinyl ketone; other oxidation

products were inferred from known fragmentation patterns, such as pinonaldehyde; and

other products were identified according to their mass to charge (

) ratio. Numerous

unidentified products were formed, and the evolution of first- and second-generation

products was clearly observed. SOA yields from the different terpenes ranged from 1 to

68%, and the total gas- plus particle-phase products accounted for

50–100% of the

reacted carbon. The carbon mass balance was poorest for the sesquiterpenes, suggesting

that the observed products were underestimated or that additional products were formed

but not detected by PTR-MS. Several second-generation products from isoprene

photooxidation, including

113, and ions corresponding to glycolaldehyde,

hydroxyacetone, methylglyoxal, and hydroxycarbonyls, were detected. The detailed time

series and relative yields of identified and unidentified products aid in elucidating

reaction pathways and structures for the unidentified products. Many of the unidentified

products from these experiments were also observed within and above the canopy of a

Ponderosa pine plantation, confirming that many products of terpene oxidation can be

detected in ambient air using PTR-MS, and are indicative of concurrent SOA

formation.

Citation:

Lee, A., A. H. Goldstein, J. H. Kroll, N. L. Ng, V. Varutbangkul, R. C. Flagan, and J. H. Seinfeld (2006), Gas-phase

products and secondary aerosol yields from the photooxidation of 16 different terpenes,

J. Geophys. Res.

111

, D17305,

doi:10.1029/2006JD007050.

1. Introduction

[

] Biogenic emissions of terpene compounds influence

atmospheric chemistry through the formation of tropospheric

ozone (O

) and the production of secondary organic aerosol

(SOA). Terpenoids (or isoprenoids) encompass several wide

classes of compounds, including hemiterpenes (isoprene,

), monoterpenes (C

), sesquiterpenes (C

and oxygenated terpenes (e.g. C

O, C

O). Terpe-

noids are emitted from deciduous and evergreen trees as a

function of temperature, or of both temperature and light

[

Kesselmeier and Staudt

, 1999]. Terpenes, with their un-

saturated carbon bonds, are reactive with OH, O

, and NO

the common atmospheric oxidants, with lifetimes that range

from minutes to hours [

Atkinson and Arey

, 2003]. Several

monoterpene oxidation products are known to undergo gas-

particle partitioning, including nopinone, pinonaldehyde,

pinic acid, and pinonic acid, and have been observed in

ambient air in the gas and particle phases [

Yu et al.

, 1999;

Kavouras et al.

, 1999].

[

] Experiments examining terpene oxidation have been

conducted since F. W. Went first suggested the blue haze in

forested regions were products of these reactions. Early

experiments determined rate constants for the reaction of

many different terpenes with the major atmospheric oxidants

[e.g.,

Atkinson et al.

, 1989;

Shu and Atkinson

, 1994], and

yields of SOA [e.g.,

Pandis et al.

, 1991;

Hoffmann et al.

1997]. Experiments also examined gas-phase products from

the oxidation of isoprene [

Tuazon and Atkinson

, 1990] and

common monoterpene species, such as

-and

-pinene

[

Grosjean et al.

, 1992;

Arey et al.

, 1990;

Hakola et al.

1994], as well as a broader suite of monoterpenes [

Shu et

al.

, 1997;

Reissell et al.

, 1999;

Orlando et al.

, 2000]. SOA

yield from isoprene has been long thought to be negligible

[

Pandis et al.

, 1991]. However, recent observations of tetrol

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 111, D17305, doi:10.1029/2006JD007050, 2006

Department of Environmental Science, Policy, and Management,

University of California, Berkeley, Berkeley, California, USA.

Department of Environmental Science and Engineering and Depart-

ment of Chemical Engineering, California Institute of Technology,

Pasadena, California, USA.

D17305

1of25

compounds with an isoprene skeleton [

Claeys et al.

, 2004;

Kourtchev et al.

, 2005] and isoprene oxidation products

[

Matsunaga et al.

, 2005] in ambient aerosol suggest that

isoprene does contribute to ambient SOA production.

Additionally, the high global isoprene emission rate (500 Tg

year

)[

Guenther et al.

, 1995] and the 1–3% SOA yield

obtained from the photooxidation of isoprene at lower

temperatures and initial isoprene concentrations [

Kroll et al.

2005], also suggest that isoprene oxidation may represent an

important contribution to global SOA production.

[

] The HC:NO

ratio has been shown to impact SOA

production.

Pandis et al.

[1991] showed that increasing the

HC:NO

ratio increased SOA yield when the ratio was

10–15 ppb HC: ppb NO

, but decreased SOA yield at

higher ratios. From isoprene photooxidation under high

, SOA yield decreased with increasing NO

[

Kroll et

al.

, 2006]. Experiments with HC:NO

ratios >8 were

reported to increase SOA production from

-xylene [

Song

et al.

, 2005]. Additionally, higher NO

levels in the chamber

increase O

formation, which may impact SOA yield due to

possible O

reactions. Low HC:NO

ratios (

3–6 ppb HC:

ppb NO

) are generally representative of ambient air ob-

served above Blodgett Forest, a Ponderosa pine plantation

in the Sierra Nevada, California, where typical NO

con-

centrations are

1–2 ppb [

Day et al.

, 2002] and the sum of

mixing ratios of all typically measured biogenic and an-

thropogenic VOCs in ambient air, that have a lifetime on the

order of the time scale of these experiments, is

5–6 ppb

[

Lamanna and Goldstein

, 1999]. Thus, chamber experi-

ments conducted at low HC:NO

ratios (i.e., high NO

) may

be more applicable to the real atmosphere in forested

environments.

[

] Recent studies provide observational support that

terpenes can be oxidized within a forest canopy before

detection by above-canopy flux techniques. In an orange

grove in Spain,

-caryophyllene, a reactive sesquiterpene,

was observed in branch enclosures but not in simultaneous

measurements above the canopy, suggesting that

-caryo-

phyllene was oxidized within the canopy [

Ciccioli et al.

1999]. At Blodgett Forest, chemical O

flux to the ecosys-

tem scaled exponentially with temperature in a similar

manner as monoterpenes [

Kurpius and Goldstein

, 2003],

and from the same site, elevated monoterpene fluxes to the

atmosphere from the mastication of Ponderosa pine trees

[

Schade and Goldstein

, 2003] resulted in increased chem-

ical O

flux to the ecosystem [

Goldstein et al.

, 2004],

suggesting a linkage between chemical O

loss in the canopy

and terpene emissions. In subsequent work,

Holzinger et al.

[2005] observed large quantities of previously unmeasured

oxidation products above the same pine forest canopy. In a

northern Michigan forest, elevated nighttime OH concen-

trations correlated with O

mixing ratios, suggesting that

OH was produced from reactions between O

and unmea-

sured terpenes [

Faloona et al.

, 2001]. From the same

forest, higher than expected OH reactivity was observed

and scaled exponentially with temperature in a similar

manner as biogenic emissions [

Di Carlo et al.

, 2004].

Taken together, these results provide convincing evidence

of chemical loss of reactive compounds within the forest

canopy through oxidation by OH and O

[

] Previously, we reported the yields of SOA and gas-

phase products from the ozonolysis of ten terpene com-

pounds [

Lee et al.

, 2006]. Here, we report the results from a

series of chamber photooxidation experiments conducted on

16 terpene compounds in the presence of OH, NO

, and

secondarily-produced O

. Yields of SOA and gas-phase

products are presented, as well as the time evolution of

selected calibrated gas-phase oxidation products, and gas-

phase oxidation products identified by their mass to charge

ratio (

). A more detailed analysis of the contribution of

first and second-generation oxidation products to SOA

production is presented elsewhere [

Ng et al.

, 2006]. The

goal of the PTR-MS measurements of gas-phase oxidation

products from the ozonolysis and photooxidation of ter-

penes is to confirm if ions observed at Blodgett Forest are

consistent with terpene oxidation, and to provide a guide to

future studies using PTR-MS to monitor secondary gas-

phase compounds in ambient air. Here, we present the

yields of the observed oxidation product ions, and focus

on the two product ions that were dominant at Blodgett

Forest,

113 and 111.

2. Experiment

[

] The photooxidation experiments were conducted at

the Caltech Indoor Chamber Facility, which has been

described in detail elsewhere [

Cocker et al.

, 2001;

Keywood

et al.

, 2004]. Briefly, the facility consists of two suspended

28 m

flexible Teflon chambers that maintain atmospheric

pressure at all times. For the monoterpene, isoprene, and

oxygenated terpene experiments, ammonium sulfate seed

aerosol was added to act as a surface for the condensation of

oxidation products, contributing to the measured yield of

secondary organic aerosol. For the sesquiterpene experi-

ments, no initial seed aerosol was used because nucleation

from sesquiterpene oxidation would result in two aerosol

modes, making data analysis, particularly the wall loss

correction, difficult. The starting concentration of seed

aerosol was about 20,000 particles cm

, with a mean

diameter of 80–100 nm. SOA yields were calculated from

each experiment from the ratio of the amount of SOA

formed (assuming density of 1.25 g cm

) and the amount

of hydrocarbon reacted. The hydrocarbon to NO

ratio

(HC:NO

) for all experiments ranged from 0.7–2.2 ppb

HC: ppb NO

, produced from the photolysis of NO

inside the chamber, was measured using an ambient O

monitor (Horiba APOA-360, Irvine, CA), and calibrated

using an internal O

generator and N

as zero air.

[

] Microliter volumes of individual liquid monoterpenes

and oxygenated terpenes were injected into a 250 mL glass

bulb and gently heated as a stream of clean air passed

through the bulb, vaporizing the terpene and carrying it into

the chamber. The radical precursor used was nitrous acid

(HONO), prepared by dropwise addition of 2 mL of 1%

NaNO

into 15 mL of 10% H

in a glass bulb attached

to the chamber. A stream of dry air was passed through the

bulb, introducing HONO into the chamber. A NO

monitor

(Horiba APNA-360, Irvine, CA) measured NO and NO

which were formed as side products in the preparation of

HONO. HONO was detected by the PTR-MS at

30, the

dehydrated fragment of HONOH

, but not at the parent

ion

48, and also appeared to be detected by the

monitor. However, HONO calibrations were not per-

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

2of25

D17305

formed, so its concentration in the chamber was not well-

constrained.

[

] When the seed, parent hydrocarbon, and NO

con-

centrations stabilized, turning on the blacklights started the

reactions. The indoor chambers were equipped with 276

GE350BL fluorescent blacklights centered at 354 nm,

efficiently photolyzing HONO to OH and NO. To minimize

temperature increases (1–2

C over the course of the exper-

iment), 10% of the lights were used. Between experiments,

the chambers were continuously flushed with clean com-

pressed air that passed through four scrubbing cartridges

containing activated carbon, silica gel, Purafil, and molec-

ular sieve, respectively, and a HEPA filter before entering

the Teflon chambers. The chambers were flushed for at

least 36 h before the start of an experiment, which reduced

and particle concentrations to below 1 ppb and

100 particles cm

, respectively. Hydrocarbon concentra-

tions inside the chamber, before the injection of the terpene,

were below detection limit for all parent terpenes.

2.1. Terpene Compounds and Gas-Phase

Measurements

[

] Sixteen different terpenoid compounds were reacted

in these experiments: isoprene, eight monoterpenes

(

-pinene,

-pinene, 3-carene, terpinolene,

-terpinene,

myrcene, limonene,

-terpinene), four sesquiterpenes (

humulene,

-caryophyllene, longifolene, and aromaden-

drene), and three oxygenated terpenes (methyl chavicol,

also known as 4-allylanisole, linalool, and verbenone).

Table 1 lists the structures and molecular weights of the

parent terpenes used and the rate constants for their reac-

tions with OH (k

). Gas-phase concentrations of the parent

terpene were monitored using two instruments: a Hewlett

Packard gas chromatograph with a flame ionization detector

(GC-FID) using a 60 m

0.32

m DB-5 column (J&W

Scientific, Davis, CA), and a proton transfer reaction mass

spectrometer, or PTR-MS (Ionicon Analytik, Innsbruck,

Austria) [

Lindinger et al.

, 1998]. Air from the reaction

chamber was sampled using SilcoSteel (Restek Corporation,

Bellafonte, PA) tubing, and pulled through a 2

m pore size

PTFE particulate filter (Pall Corportation, East Hills, NY)

before analysis by PTR-MS, which measured parent ter-

penes as well as gas-phase oxidation products. The PTR-

MS is a quadrupole mass spectrometer that uses hydronium

ions (H

) to chemically ionize the compound of interest

through a proton transfer reaction. Thus, any compound that

has a proton affinity higher than that of water can be

detected by the PTR-MS, and is identified by its mass to

charge ratio (

). Because compounds are identified by

their molecular weight plus 1 (H

), the PTR-MS is unable to

distinguish between different compounds with the same

molecular weight. However, because of the controlled

nature of these laboratory chamber experiments, where

one terpene at a time is photooxidized, increasing count

rates of certain ions are indicative of oxidation products

from these reactions. Knowledge of the structures of the

parent terpene allows for the deduction of possible identities

of the oxidation products from reasonable oxidation mech-

anisms. However, structurally different oxidation products

that have the same mass cannot be distinguished from one

another by PTR-MS.

2.2. PTR-MS Calibrations and Concentrations

[

] For calibrations, each pure terpene compound was

diluted in cyclohexane and injected into a Teflon bag filled

to a final volume of 50 L. Cyclohexane, all parent terpene

Table 1.

Parent Terpene Compounds, Listed in Order of Decreasing Aerosol Yield

Compound

Structure

Formula

(

)

molec

Compound

Structure

Formula

(

)

molec

-caryophyllene

(205)

2.0

-pinene

(137)

5.3

-humulene

(205)

3.0

terpinolene

(137)

2.3

longifolene

(205)

4.8

-pinene

(137)

7.7

limonene

(137)

1.7

-terpinene

(137)

1.8

myrcene

(137)

2.1

-terpinene

(137)

3.6

methyl chavicol

(149)

verbenone

(151)

3-carene

(137)

8.7

linalool

(155)

1.6

aromadendrene

(205)

isoprene

(69)

9.9

Rate constants were obtained from

Atkinson and Arey

[2003, and references therein].

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

3of25

D17305

compounds, and nopinone were obtained from Fluka Chem-

icals through Sigma-Aldrich (St. Louis, MO). The mono-

terpenes were calibrated from the same Teflon bag

simultaneously by GC-FID and PTR-MS, and the sesqui-

terpenes and oxygenated terpenes were only calibrated

using PTR-MS. Calibration curves were generated from

measurements at three different terpene concentrations.

Three different cylinder standards (Scott-Marrin Inc., and

Apel-Riemer Environmental Inc.) containing ppm-level

concentrations of acetaldehyde, acetone, isoprene, methyl

vinyl ketone, methacrolein, and 3-methyl furan were diluted

at varying flow rates into the inlet air stream sampling clean

compressed air for calibrations of the PTR-MS at three

concentration levels ranging from

50–150 ppb. Three low

molecular weight oxidation products: formaldehyde, formic

acid, and acetic acid, and nopinone, a higher molecular

weight oxidation product from

-pinene oxidation, were

calibrated in the PTR-MS using Teflon bags. Nopinone was

diluted in cyclohexane before injection into the bag, and the

other three oxidation products were diluted in ultrapure

water prepared by the Millipore Milli-Q system (Billerica,

MA). Two different Teflon bags were used for the calibra-

tions to separate cyclohexane-based and water-based solu-

tions. Cyclohexane and its potential impurities are discussed

in more detail by

Lee et al.

[2006]. However, because

cyclohexane was only used in the standard calibrations,

and not used directly in these photooxidation experiments,

no interferences from cyclohexane + OH oxidation products

are expected. The standard error of the slope of the

calibration curves was <6% for formaldehyde, acetalde-

hyde, 3-methylfuran, 3-carene, and

-pinene, and <3% for

all other terpene compounds and calibrated oxidation prod-

ucts from cylinder standards and Teflon bags. Additional

sources of error include the accuracies of the syringe, the

volumetric flask used for the terpene dilutions, and the flow

controller, together contributing an uncertainty of 3–5%.

[

] The concentrations of compounds for which pure

commercial standards were not available were estimated

based on the rate constant (

) of the proton transfer reaction,

according to the equation [

Lindinger et al.

, 1998]:

½¼

where [

] is the unknown concentration of the compound of

interest, and [

] is the signal of the protonated

compound, [

] is the primary ion signal, and

is the

reaction time in the drift tube. Because the proton transfer

reaction rate constants are not known for all compounds, an

estimated rate constant (

)of2

molecule

was used for those compounds without a measured rate

constant. The rate constants for the proton transfer reaction

of most compounds are generally within ±20% of the

estimated

[

] Because a pinonaldehyde standard was not available,

concentrations were calculated according to Equation (1).

Fragments and isotopes associated with pinonaldehyde

(Table 2) were determined from correlations with

151

(the dominant pinonaldehyde fragment of the unfragmented

169), and were compared with the fragments reported

Wisthaler et al.

[2001]. Because our experiments were

conducted at a higher drift tube pressure than

Wisthaler et

Table 2.

Molecular Weights Associated With Calibrated Oxidation Products and Sesquiterpenes Detected by PTR-MS

Compound

(mass + 1)

Structure

Description

primary signal

isotope of primary ion

primary ion water cluster

oxidation

products

common to all

photooxidation

experiments

formaldehyde

acetaldehyde

formic acid

acetone

acetic acid

-caryophyllene

67, 68,

, 82, 95,

109

121, 137

149, 205

, 206

sesquiterpene, fragments, and isotopes

85, 113

, 141, 147, 153, 165, 167, 175, 177, 179,

181, 183, 189, 191, 197, 201, 207, 209, 217, 219, 223,

235, 237, 253

unidentified oxidation products

-humulene

67,

, 82, 95, 96, 97, 103,

109

110, 121

123,

135,

137,

149,

150,

205

, 206

sesquiterpene, fragments, and

isotopes

57, 71, 73, 75, 83, 85,

, 89, 99,

101

, 115, 125, 127,

129, 133, 139, 141, 151

, 153, 155

, 157, 159

, 161,

169

, 175, 179, 183, 191, 193, 209, 219, 223

unidentified oxidation products

longifolene

81,

82,

, 96, 103,

109

,110

, 121, 123, 124, 135, 136,

149

, 150,

205

, 206, 207

sesquiterpene, fragments, and

isotopes

71, 73, 75, 85, 87, 89, 99, 101, 113

, 115, 129, 201,

203

, 207, 219, 220

, 221, 223, 235

unidentified oxidation products

aromadendrene

95, 96, 103, 109, 110, 121,

123

, 124,

135

, 136,

149

150, 163,

205

, 206, 207

sesquiterpene, fragments, and

isotopes

73, 75, 82, 85, 87, 89, 99, 101, 103, 127, 139, 163,

165,

189

, 190

, 191, 203,

207

, 209, 219, 221, 222

223

unidentified oxidation products

Ions associated with parent terpenes listed in bold font represent ions >20% of the unfragmented parent terpene.

Unidentified oxidation products in bold font are ions with >5% yield (on a mole basis) from the parent terpene.

Unidentified oxidation products with asterisks represent ions that were observed at Blodgett Forest using PTR-MS [

Holzinger et al.

, 2005].

Even mass-to-charge ratio oxidation products that are not carbon isotopes of other oxidation products are likely gas-phase nitrogen containing pro

ducts,

e.g., organic nitrates.

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

4of25

D17305

al.

[2001] (

2.2 mbar versus

2.0 mbar), we expect to

observe more fragmentation than reported by

Wisthaler et

al.

[2001]. The fragment ions we observed were similar to

Wisthaler et al.

[2001], but

151, 152, 169, and 170

represented a smaller proportion of the total pinonaldehyde

signal, and

43, 72, 99, 108, and 123 represented a larger

fraction of the total pinonaldehyde signal. The contribution

of the ions

71 and 107 were similar to

Wisthaler et al.

[2001]. We did not include the small contribution from

81 reported by

Wisthaler et al.

[2001], which was 10% of

151, or 3.6% of the total pinonaldehyde signal, because

of the interference from

-pinene fragmentation. Addition-

ally,

43, 71, and 72 showed time evolutions that

deviated from

151 towards the end of the experiment,

when pinonaldehyde concentrations were decreasing due to

further oxidation, suggesting that additional, non-pinonalde-

hyde compounds were interfering on those mass to charge

ratios. Because of the difficulty in determining the pinonal-

dehyde concentration given the number of fragments and the

possibility of interfering compounds, we calculated concen-

trations in three different ways to estimate a range of likely

pinonaldehyde concentrations. The upper limit mixing ratio,

which is likely overestimated, is calculated from the sum of

concentrations of all pinonaldehyde-associated ions listed in

Table 2. The midrange mixing ratio excludes a contribution

from

43, 71, and 72, which exhibit time evolutions

that significantly deviated from

151 as pinonaldehyde

decreased. The lower-limit mixing ratio, which is probably

underestimated, only includes ions (

107, 151, 152, 169,

170) that showed the same rate of decrease in signal as

151 during the period when pinonaldehyde was undergoing

further oxidation. Thus, pinonaldehyde mixing ratios and

yields contain a great deal of uncertainty, which is evidenced

in the upper and lower-limit range presented in Tables 6 and

9b. Yields of keto-aldehyde compounds from 3-carene,

limonene,

-terpinene, and

-terpinene were determined

similarly, from correlations of other product ions with

169 and 151. Similar fragment ions were observed (Tables 3a

and 3b).

[

] High background counts from the blank chamber air

were observed for the low molecular weight oxidation

products, resulting in higher detection limits for those

Table 3a.

Molecular Weights Associated With the Monoterpenes Detected by PTR-MS

Compound

(mass + 1)

Structure

Description

limonene

69,

82, 95, 96,

137

138

monoterpene, fragments and

isotopes

107

, 108,

123

,124, 133,

151

, 152, 169

, 170

uncalibrated oxidation product:

limononaldehyde, fragments, and

isotopes

139, 140

uncalibrated oxidation product:

limonaketone and isotope

57, 69, 70, 71, 73,

, 77, 83, 85, 87, 89, 93, 97, 99,

101, 103, 109, 111

,113

, 115, 121, 125, 127, 129,

131, 135, 141

, 143, 149, 153, 155

, 156, 157, 165,

167, 171, 181, 183, 185, 187, 200

, 218

unidentified oxidation products

myrcene

82,

137

138

monoterpene, fragments and

isotopes

57, 64

, 65,

, 73,

, 87, 89,

, 97, 99,

101, 103,

111

113

, 115, 125, 127, 129,

139

, 141,

143, 145, 153, 155

, 159

unidentified oxidation products

3-carene

82,

137

138

monoterpene, fragments and

isotopes

107

, 108,

123

,124, 133,

151

, 152, 169

, 170

uncalibrated oxidation product:

caronaldehyde, fragments, and

isotopes

71, 73, 75, 77, 79, 83, 85, 87, 89,

, 94, 97, 99, 101,

103,

107

, 109, 111*,

113*,

115, 121,

123

, 124, 125,

127, 129, 133,

135

139

141

, 142,

151

153

, 155

157, 167, 169

, 181, 183

unidentified oxidation products

-pinene

82,

137

138

monoterpene, fragments and

isotopes

, 72, 99,

107

108,

109

, 123

151

152, 169, 170

uncalibrated oxidation product:

pinonaldehyde, fragments, and

isotopes

69, 73, 75, 83, 85, 87, 89, 94

, 97, 101, 103, 111

113

, 115, 121, 125, 127, 129, 131, 139, 141

, 153,

155

, 157, 165, 167, 171, 183, 185, 200

unidentified oxidation products

Ions associated with parent terpenes listed in bold font represent ions >20% of the unfragmented parent terpene.

All monoterpenes occur predominantly on the parent

137 and the fragment

81 ions, with a small fraction (<20%) occurring on the

C isotope

with

(138, 82).

Unidentified oxidation products in bold font are ions with >5% yield (on a mole basis) from the parent terpene.

Unidentified oxidation products with asterisks represent ions that were observed in the canopy air of a coniferous forest in California by PTR-MS

[

Holzinger et al.

, 2005].

Even mass-to-charge ratio oxidation products that are not carbon isotopes of other oxidation products are likely gas-phase nitrogen containing pro

ducts,

e.g., organic nitrates.

Fragments were determined from correlations with

151 (the dominant pinonaldehyde ion) that also agreed well with fragment ions reported

elsewhere for pinonaldehyde [

Wisthaler et al.

, 2001]. Additional fragment ions reported by

Wisthaler et al.

[2001],

43 and 81, were excluded from the

pinonaldehyde concentration due to interference by other compounds.

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

5of25

D17305

compounds. The detection limit (1

background counts

/sensitiv-

ity) for the monoterpenes was

70 ppt, for the sesquiter-

penes:

50 ppt, for the oxygenated terpenes: <50 ppt, for

nopinone: 20 ppt, for acetaldehyde and acetone: <0.5 ppb,

for formic acid: 1.5 ppb, acetic acid: <1 ppb, and for

formaldehyde: 6 ppb. Starting concentrations for formalde-

hyde and formic acid,

31 and 47, respectively, were

<10 ppb, calculated from count rate divided by the sensi-

tivity of the PTR-MS to those compounds. The background

count rate on

31 does not result from formaldehyde, but

rather from interference from other ions, such as the

isotope of NO

. Additionally, back reactions between pro-

tonated formaldehyde (CH

) and H

O make PTR-MS

measurements of formaldehyde in ambient air difficult due

to low formaldehyde mixing ratios and high water content of

ambient air. Formaldehyde was calibrated up to a concen-

tration of

80 ppb, which was only exceeded in the myrcene

(formaldehyde

110 ppb) and isoprene (formaldehyde

>300 ppb) experiments. Thus, our calculated yields of

formaldehyde from myrcene and isoprene are subject to

significant errors. Mixing ratios of all other oxidation

products, from all experiments, were within the calibrated

range. Starting concentrations for acetaldehyde, acetone, and

acetic acid were <5 ppb. Background counts were subtracted

from the signal for ions monitored. Concentrations of the

calibrated oxidation products should be considered upper-

limit values, as other products with the same molecular

weight may occur on the

of the calibrated products.

Thus, the formation of, e.g., glycolaldehyde, is indistin-

guishable from acetic acid, as they both occur at

61.

3. Results and Discussion

[

] Table 1 shows the structures and OH rate constants

for the 16 terpenes used. The ions (

) associated with

each parent terpene and their oxidation products are detailed

in Tables 2–4. Initial conditions are listed for each exper-

iment in Table 5. The HC:NO

ratios are reported in Table 5

as ppb C: ppb NO

, and in Table 6 as ppb HC: ppb NO

Hydrocarbon oxidation was initiated primarily by OH,

formed from the photodissociation of HONO. Other oxi-

dants, O

and NO

, were also formed, but only later in the

experiment, after much of the initial NO is depleted (from

reactions with peroxy radicals). This typically occurred after

almost all (

80%) of the parent hydrocarbon was depleted,

so for most experiments O

and NO

did not react appre-

ciably with the parent hydrocarbon. Possible exceptions

were the sesquiterpenes,

-caryophyllene and

-humulene,

which react very rapidly with O

[

Shu and Atkinson

, 1994].

Figure 1 shows the time series for O

, NO, NO

, and NO

for a typical experiment (myrcene), where O

increases

when most of the initial terpene has been depleted, and an

Table 3b.

Molecular Weights Associated With the Monoterpenes Detected by PTR-MS

Compound

(mass + 1)

Structure

Description

terpinolene

82,

137

138

monoterpene, fragments and

isotopes

65, 71, 73, 75, 77, 83, 85, 86

, 89,

, 97, 98

101, 103, 107, 109,

111

,113

, 115, 123, 125,

126, 127, 129, 135, 141

, 143, 149, 159, 167, 169, 183

unidentified oxidation products

-pinene

82,

137

138

monoterpene, fragments and

isotopes

83, 93, 97, 103,

121

122,

139

140, 141

calibrated oxidation product:

nopinone, isotopes and fragments

69, 71, 73, 79, 83, 85, 89, 93, 99, 101, 107, 108

, 109,

111

,113

, 115, 123

, 125, 127, 129, 135, 142, 149,

151

, 153, 155

, 157, 165, 167, 169

, 171, 183, 184

185, 198

unidentified oxidation products

-terpinene

82,

137

138

monoterpene, fragments and

isotopes

107

123

, 124,

151

, 152,

169

, 170

uncalibrated oxidation product:

-terpinaldehyde, fragments, and

isotopes

57, 69, 71, 73, 74, 75, 77, 83, 85,

, 89,

, 97, 99, 101,

102, 103, 105, 107, 109, 111

,113

115

, 125, 127, 129,

131, 133,

135

, 139, 141

, 143, 149, 153, 155

, 157,

159

, 165, 167, 171, 181, 183, 185, 199, 230

unidentified oxidation products

-terpinene

93, 95,

82, 135,

137

138

monoterpene, fragments and

isotopes

107

123

, 124,

151

, 152,

169

,170

uncalibrated oxidation product:

-terpinaldehyde, fragments,

and isotopes

65, 69, 71, 73, 75, 77, 87, 89, 97, 99, 101, 103, 109,

111

,113

, 115, 125, 127, 129, 131, 135,

139

, 141

143

, 145, 153, 155

, 157, 158, 159

, 165, 167, 171,

178

, 181, 183, 184

, 185, 199, 201, 202

, 229, 230

unidentified oxidation products

Ions associated with parent terpenes listed in bold font represent ions >20% of the unfragmented parent terpene.

Unidentified oxidation products in bold font are ions with >5% yield (on a mole basis) from the parent terpene.

Unidentified oxidation products with asterisks represent ions that were observed in the canopy air of a coniferous forest in California by PTR-MS

[

Holzinger et al.

, 2005].

Even mass-to-charge ratio oxidation products that are not carbon isotopes of other oxidation products are likely gas-phase nitrogen containing pro

ducts,

e.g., organic nitrates.

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

6of25

D17305

atypical experiment (

-caryophyllene), where O

increases

after all of the initial terpene has been depleted. The two

reactive sesquiterpenes,

-caryophyllene and

-humulene,

both show the atypical profile for O

. This may suggest that

the atypical case of delayed O

production resulted from

titration of O

by the parent sesquiterpene and/or first

generation oxidation products, which still contain C = C

double bonds. For all VOCs, O

and NO

are expected to

react only with unsaturated hydrocarbon oxidation products.

Because these products have generally received little study,

it is difficult to estimate the relative contributions of OH,

, and NO

to such chemistry. However, it will be shown

for selected terpenes that the formation of most of the major

observed oxidation products can be explained using OH

chemistry, suggesting that the roles of O

and NO

reactions

are minimal.

[

] Numerous oxidation product ions were produced,

many of which have been observed in ambient air within

and above Blodgett Forest [

Holzinger et al.

, 2005]. The

yields of the oxidation product ions observed in these

photooxidation experiments that were also observed by

Holzinger et al.

[2005] are listed in Table 7. In ambient

air,

113 and 111 were dominant ions, and these ions

were observed at low concentrations from all terpenes (

113), and from all monoterpenes and a few oxygenated

terpenes (

111). Yield of

113 was highest from

myrcene (32%), and yield of

111 was highest from

terpinolene (29%) and linalool (20%), while aromadendrene

did not produce any ions that were observed at Blodgett

Forest. The number of oxidation ions observed from these

photooxidation experiments is significantly larger than the

number of oxidation ions observed from the ozonolysis

experiments we conducted previously [

Lee et al.

, 2006].

The mean number of oxidation product ions observed from

ozonolysis reactions was 14 ± 5 (mean ± SD) and the mean

number of ions observed from photooxidation reactions was

33 ± 6. The PTR-MS was operated under similar conditions

for both sets of oxidation experiments, with a drift tube

pressure

2.2 mbar, suggesting that the difference in the

number of ions observed is not a result of increased

fragmentation of products in the photooxidation experi-

ments. Given the greater versatility of OH attack on the

parent terpenes, a greater number of different products

formed by photooxidation compared to ozonolysis are

expected.

[

] Yields of SOA and gas-phase oxidation products will

be discussed according to terpene classification, i.e., sesqui-

terpenes, monoterpenes, and other terpenes (oxygenated

terpenes and isoprene). The gas-phase yields reported in

Tables 6–10 were calculated as the slope of the linear least-

squares fit of the regression between the product and the

parent hydrocarbon. This method was used for both first-

and second-generation products. For first-generation prod-

ucts that were further oxidized, only the linear, increasing

portion of the regression was considered. Yields of second-

generation oxidation products were calculated from a re-

gression line based on an individually-selected time period

that most appropriately reflected the production of the

individual ion. Although stoichiometric yields of second-

generation products are more precisely quantified against

their parent compound (the first-generation product oxi-

dized to produce the second-generation product), because

first-generation products are generally unidentified, and

because the parent terpenes are more commonly and easily

measured than first-generation products in the real atmo-

sphere, we report yields of second-generation products as a

function of the parent terpenes, rather than their first-

generation precursors. The carbon balance was calculated

for each experiment as a function of time, and is represented

in Figure 2 as an average for a 30–60 minute snapshot at

the middle or end of each experiment for the purpose of

carbon counting, and does not represent overall the stoi-

Table 4.

Molecular Weights Associated With Other Terpenes Detected by PTR-MS

Compound

(mass + 1)

Structure

Description

methyl chavicol

149

, 150

oxygenated terpene and isotope

65, 71, 73, 75, 85, 89, 97, 99, 101, 103,

109

, 113,

121

123

, 124

, 125, 129, 131,

137

151

, 153, 163, 165, 168

170

, 182

unidentified oxidation products

verbenone

109

, 110, 123, 133,

151

, 152

oxygenated terpene, fragments,

and isotopes

69,

, 73, 75, 77, 83, 85, 87, 89, 93, 94, 97, 99, 100, 101,

107, 111

113

, 115, 123

, 124,

125

, 127, 129, 130, 141

, 143,

144, 145, 151

, 152, 153, 159

, 167, 169

, 185, 230

unidentified oxidation products

linalool

82,

137

138, 155, 156

oxygenated terpene, fragments,

and isotopes

60,

69,

71, 72, 73, 74,

, 77, 79,

, 85, 87, 89 ,

, 97,

99,

101

, 107,

111

,113

, 115, 123

, 124, 125,

127

129

141

, 143, 145, 151

, 153, 159

, 167, 169

, 185, 187, 188

unidentified oxidation products

isoprene

isoprene and isotope

,72

MVK + MACR

83, 84

3-methyl furan

57, 63, 65, 73, 75, 77, 79, 85, 86, 87, 89, 97, 99, 101, 103,

105, 113

, 115, 117, 129, 133

unidentified oxidation products

Ions associated with parent terpenes listed in bold font represent ions >20% of the unfragmented parent terpene.

Unidentified oxidation products in bold font are ions with >5% yield (on a mole basis) from the parent terpene.

Unidentified oxidation products with asterisks represent ions that were observed in the canopy air of a coniferous forest in California by PTR-MS

[

Holzinger et al.

, 2005].

Even mass-to-charge ratio oxidation products that are not carbon isotopes of other oxidation products are likely gas-phase nitrogen containing pro

ducts,

e.g., organic nitrates.

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

7of25

D17305

chiometric yields (Tables 6–10). The estimated carbon

balance from all experiments ranged from 52 ± 5% to

138 ± 55%, with considerable uncertainties associated with

the assignment of carbon number for each unidentified ion,

as well as uncertainty from the calibration of parent terpenes

and oxidation products, and the collisional rate constant.

Although we assumed SOA was 60% carbon (based on the

%C of pinic acid), this number is uncertain and may vary

for each experiment, and likely affected the total carbon

balance by elevating it, resulting in carbon balances for

some experiments >100%. The carbon balances for each

experiment are shown in Figure 2, and because of the large

uncertainty associated with the mass balance, reflect a

qualitative assessment of the ability of the PTR-MS to

detect the unidentified compounds according to their

Clearly, the products formed from sesquiterpene photooxi-

dation are not as readily detected by PTR-MS as the

products from the photooxidation of other terpenes.

[

] Product yields from other studies are also compared

with the results obtained from these experiments (Tables 8

and 9a–9c), however, other experiments often use different

radical precursors, such as CH

ONO, NO

,orH

, which

may impact gas-phase and SOA yields. Thus, these compar-

isons are included to simply present the range of yields

observed elsewhere under different experimental conditions.

In addition to carbonyls and acids, these photooxidation

experiments in the presence of NO

should be expected to

produce organic nitrates. Theoretical calculations have sug-

Table 5.

Comparison of SOA Yields Obtained From This Study With Results From Other Photooxidation Experiments

Terpene

Temp, K

RH, %

HC,

ppb

HC:NO

ppbC/ppb

SOA Mass

Yield,

Reference

-caryophyllene

295

37 ± 3

212 ± 2

68 ± 7

This work

319

845, 998

103, 125

[

Hoffmann et al.

, 1997]

308, 309

17–82

3, 8

37–79

[

Griffin et al.

, 1999]

-humulene

294

46 ± 1

254 ± 2

65 ± 1

This Work

308, 307

13–59

32–85

[

Griffin et al.

, 1999]

longifolene

293

34 ± 2

186 ± 2

65 ± 1

This work

limonene

294

120 ± 2

394 ± 4

58 ± 1

This Work

313, 309

10–120

2, 5

8.7–34

[

Griffin et al.

, 1999]

4.3

[

Larsen et al.

, 2001]

Myrcene

294

112 ± 2

272 ± 3

43 ± 1

This Work

311, 312

3.5, 57.5

2, 4

5.6–6.8

[

Griffin et al.

, 1999]

methyl chavicol

294

79 ± 2

194

± 2

40 ± 1

This work

3-carene

294

109 ± 2

236 ± 3

38 ± 1

This work

319, 312

143, 161

23, 27

[

Hoffmann et al.

, 1997]

310, 312

2.5–99.7

3, 7

2–18

0.5

[

Griffin et al.

, 1999]

0.5

[

Larsen et al.

, 2001]

aromadendrene

294

34 ± 1

107 ± 2

37 ± 2

This work

-pinene

293

109 ± 1

199 ± 3

32 ± 0.1

This work

298

not given

4–40

[

Noziere et al.

, 1999]

[

Hatakeyama et al.

, 1991]

309–324

1–96

7–15

43 1–12

[

Hoffmann et al.

, 1997]

32–45

[

Jaoui and Kamens

, 2001]

1.8

[

Larsen et al.

, 2001]

terpinolene

294

110 ± 5

190 ± 4

31 ± 2

This Work

313, 316

25–133

3, 4

1.5–4.1

[

Griffin et al.

, 1999]

-pinene

293

170 ± 5

293 ± 4

31 ± 1

This work

307–317

153

[

Hoffmann et al.

, 1997]

297–304

21.1

20.9

[

Jaoui and Kamens

, 2003b]

316, 313

7–141

7, 6

3–27

[

Griffin et al.

, 1999]

1.8

[

Larsen et al.

, 2001]

-caryophyllene

294

118 ± 6

193 ± 3

29 ± 2

This Work

312, 311

21, 66

5, 4

9.8–16

[

Griffin et al.

, 1999]

-terpinene

293

103 ± 1

145 ± 1

25 ± 0.4

This Work

316, 313

20, 74

3, 5

8.2–17.5

[

Griffin et al.

, 1999]

Verbenone

294

105 ± 2

127 ± 2

19 ± 0.4

This work

linalool

295

124 ± 2

104 ± 4

13 ± 0.3

This work

320, 318

19, 137

11, 9

4.2, 9.8

[

Hoffmann et al.

, 1997]

312

26.7

5.6

[

Griffin et al.

, 1999]

isoprene

294

506 ± 33

26 ± 1

2 ± 0.1

This work

293

25–500

0.5–30

0.1–1.8

1–3

[

Kroll et al.

, 2005]

The reported hydrocarbon to NO

ratio includes propene, used as the photochemical initiator.

The range reported here represent the lowest and highest yields from a series of experiments, with the corresponding temperatures,

, and

hydrocarbon to NO

ratios.

Assumed aerosol density of 1 g cm

. The range of aerosol yield results from several different experiments with higher yields observed from higher

initial

-pinene concentrations.

From nine different experiments. Average of reported range of temperatures.

Aerosol yields are reported as the maximum carbon yield (19–27%). To convert to total SOA mass yield for comparison with other experiments, we

assumed the aerosol was 60% carbon.

Percent molar yield of identified carboxylic acids and carbonyls in the aerosol phase.

SOA yields are expressed on a percent mass basis, using the ratio of

g organic aerosol m

g parent terpene m

, assuming an aerosol density of

1.25 g cm

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

8of25

D17305

gested that OH oxidation of

-pinene produces a 19% yield

of organic nitrates in chamber experiments, and a 13%

yield in the real atmosphere [

Peeters et al.

, 2001]. Labo-

ratory experiments have suggested that abstraction from

aldehydic H on pinonaldehyde produces high yields (81%)

of a PAN analogue under high NO

conditions [

Noziere

and Barnes

, 1998]. Table 10 lists the molar yields of the

even

product ions (characteristic of a N-containing

compound), which were produced from ten of the terpene

compounds at extremely low yields. The PTR-MS observed

a 1% yield of organic nitrates from

-pinene, suggesting

that perhaps these organic nitrates were thermally unstable

in the PTR-MS, fragmented in the PTR-MS on odd

which are not typically associated with N-containing com-

pounds, were lost to tubing walls, or were not formed at

high yields in these experiments. If these organic nitrates

are high carbon number compounds, such as the PAN

analogue from pinonaldehyde oxidation, then the poor

detection of these compounds by PTR-MS should affect

the carbon mass balance of the experiments. Poor carbon

balances were obtained for

-pinene and the sesquiterpenes,

and undetected organic nitrates may be responsible. How-

ever, due to the high degree of uncertainty in the carbon

mass balance, a 20% yield of undetected organic nitrates

are within the uncertainty estimates of many of the experi-

ments, and may not represent a significant and detectable

carbon loss.

[

] Time series graphs and production mechanisms for

selected ions from selected terpenes will be presented, and

will focus primarily on

113 and 111. The production

mechanisms for many of the product ions are very straight-

forward, however, in order to generate compounds that

correspond with the observed product ions and time series,

many of the production mechanisms represent pathways

that seem less likely or straight-forward. In the following

sections, we suggest possible mechanisms and products for

a limited number of terpenes and their observed ions that

correspond best to the observed data, focusing primarily on

the ions observed in a ponderosa pine forest canopy

[

Holzinger et al.

, 2005], while also addressing the incon-

sistencies between expected and observed ions. Aerosol

yields and comparisons with other studies are presented in

Table 5, but a more detailed analysis of the contribution of

first- and second-generation oxidation products to SOA

production from these photooxidation experiments, and

ozonolysis experiments performed in 2003 [

Lee et al.

2006], is presented elsewhere [

Ng et al.

, 2006].

3.1. Sesquiterpenes

[

] SOA yields were highest from

-caryophyllene

(68%),

-humulene (65%), and longifolene (65%), with

an intermediate SOA yield of 37% from aromadendrene

(Table 5). The relatively few experiments conducted on

sesquiterpene oxidation have focused primarily on

caryophyllene and

-humulene, thus no data was found

in the existing literature on longifolene and aromaden-

drene photooxidation. High SOA yields for

-caryophyllene

and

-humulene have been observed by other studies [e.g.,

Table 6.

Gas-Phase Yields From Terpene Photooxidation Experiments

Terpene

% Yield

Yield

UnID

Yield

Total C

Balance,

-caryophyllene

1.3

42 ± 10

0.6 ± 0.2

6.2 ± 2

1.5 ± 0.4

8.7 ± 2

22 ± 1

62 ± 5

-humulene

1.4

—

0.2 ± 0.1

6.1 ± 0.8

2.4 ± 0.3

7.5 ± 0.7

54 ± 3

77 ± 9

longifolene

0.8

25 ± 3

3.7 ± 0.4

31 ± 3

3.8 ± 0.3

15 ± 1

30 ± 2

62 ± 7

limonene

1.1

43 ± 5

0.7 ± 0.1

5.0 ± 0.6

0.4 ± 0.1

3.2 ± 0.4

61 ± 3

113 ± 17

myrcene

0.9

74 ± 8

0.7 ± 0.1

4.5 ± 0.5

22 ± 2

3.9 ± 0.4

134 ± 10

100 ± 17

methyl chavicol

0.8

52 ± 6

0.7 ± 0.1

7.9 ± 1

0.7 ± 0.1

25 ± 2

121 ± 11

101 ± 20

3-carene

0.8

35 ± 4

1.6 ± 0.2

5.4 ± 0.6

9.3 ± 0.9

5.1 ± 0.6

95 ± 5

138 ± 55

aromadendrene

1.4

65 ± 8

2.9 ± 0.3

12 ± 2

4.8 ± 0.5

4.1 ± 0.4

69 ± 7

74 ± 14

-pinene

1.1

16 ± 2

1.4 ± 0.2

4.5 ± 0.5

6.3 ± 0.6

3.4 ± 0.4

91 ± 14

100 ± 7

terpinolene

1.1

23 ± 3

0.7 ± 0.1

3.5 ± 0.7

20 ± 2

1 ± 0.2

109 ± 8

87 ± 18

-pinene

2.1

49 ± 6

0.6 ± 0.1

8.2 ± 0.7

7.9 ± 0.7

1.4 ± 0.1

26 ± 1

52 ± 5

-terpinene

1.1

17 ± 2

1.2 ± 0.2

8.3 ± 0.8

5.3 ± 0.5

4.5 ± 0.5

130 ± 10

112 ± 28

-terpinene

1.0

7.8 ± 2

0.7 ± 0.1

6.1 ± 1

3.1 ± 0.4

2 ± 0.3

61 ± 4

123 ± 40

verbenone

1.1

32 ± 2

2.1 ± 0.3

8 ± 0.8

13 ± 1

6.9 ± 0.6

92 ± 4

79 ± 17

linalool

1.0

43 ± 5

1.1 ± 0.1

3.2 ± 0.4

25 ± 2

17 ± 2

144 ± 11

104 ± 25

isoprene

1.8

160 ± 19

1.9 ± 0.3

4.6 ± 0.6

0.4 ± 0.1

5.7 ± 0.7

15 ± 1

101–114 ± 22

Initial hydrocarbon (ppb parent terpene) to NO

(ppb NO + NO

) ratio of the experiment.

Gas-phase yields of CH

O (formaldehyde), C

O (acetaldehyde), CH

(formic acid), C

O (acetone), and C

(acetic acid), are expressed on

a percent mole basis. Product yields >20% are listed in bold.

Unidentified molar yields does not include the identified but uncalibrated ions associated with the dominant aldehydes formed from limonene, 3-car

ene,

-pinene,

-terpinene, and

-terpinene (Table 2).

Total carbon balance calculated from the sum of the gas-phase products and SOA, assuming aerosol is 60% carbon. See text for explanation of

uncertainty estimates.

Yield of limononaldehyde was 68 ± 7%.

Subject to significant error due to concentrations beyond range of calibrations.

Likely includes a contribution from glycolaldehyde (C

, MW = 60).

Molar yield of caronaldehyde was 77 ± 8%.

The midrange pinonaldehyde yield was used in the calculation.

The nopinone yield from

-pinene photooxidation was 17 ± 2%.

Yield of

-terpinaldehyde was 57 ± 6%.

Yield of

-terpinaldehyde was 19 ± 2%.

MVK + MACR yield was 60–87 ± 11% and 3-methylfuran yield was 3.6 ± 0.4%.

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

9of25

D17305

Hoffmann et al.

, 1997;

Griffin et al.

, 1999]. Gas-phase yields

of low molecular weight oxidation products (Table 6) varied

for the sesquiterpenes, with high formaldehyde yields from

aromadendrene,

-caryophyllene, and longifolene, all with

an external double bond, and relatively low yields of

acetaldehyde, acetone, and acetic acid. High formic acid

yields were observed from longifolene, but not from

caryophyllene or aromadendrene, all with external double

bonds. The carbon mass balances for all four sesquiterpenes

are relatively low compared to most other monoterpenes,

ranging from 62–77%. The poor mass balances for the

sesquiterpenes suggests that the observed unidentified prod-

ucts are underestimated or that additional gas-phase prod-

ucts were formed that were not measured by PTR-MS.

3.1.1. Aromadendrene and Longifolene

[

] The photooxidation of aromadendrene produced high

yields of

207 and 189 (47 ± 7%), which are likely the

dominant product ions of the ketone formed from OH

addition at the terminal carbon of the external double bond.

The formation of this product would also produce formal-

dehyde, which also accounts for the high formaldehyde

yield observed (65 ± 8%). In contrast, the photooxidation of

longifolene, which is similar in structure to aromadendrene,

produced low yields of

207 (2 ± 0.4%), and

189

was not observed. Yields of formic acid were greater than

formaldehyde, and the dominant unidentified ion produced

was

203, at a yield of only 5 ± 1%, suggesting that the

photooxidation of longifolene is more complex than that of

aromadendrene, with several different pathways that pro-

duce lower gas-phase yields of many different products and

higher yields of SOA, which is surprising given that the

similar, and comparatively simple structures of these two

sesquiterpenes would suggest similar oxidation pathways

and yields.

3.1.2.

-Humulene

[

] The photooxidation of

-humulene, with no external

double bonds, did not produce formaldehyde, and produced

low yields of the other calibrated oxidation products. Initial

oxidation of any one of the three internal double bonds

would produce a high molecular weight C

multifunctional

species (e.g., a C

keto-aldehyde occurring at

Figure 1.

Time series plots of O

, NO, NO

, and NO

for

myrcene, which was fairly representative of all experiments,

and

-caryophyllene, which was similar to

-humulene, but

atypical compared to all other experiments. Red dotted lines

represent O

. Blue dashed lines represent NO. Green dotted

and dashed lines represent NO

. Orange solid lines

represent NO

Table 7.

Percent Yield (Slope of Regression ± Standard Error of Slope) on a Mole Basis, of Product Ions From Terpene Photooxidation

That Were Also Observed in the Ambient Air Within a Ponderosa Pine Canopy in California

Terpene

111

113

123

141

151

155

159

169

-caryophyllene

0.4 ± 0.1

-humulene

1.1 ± 0.3

1.1 ± 0.2

1.5 ± 0.3

0.3 ± 0.1

0.5 ± 0.2

longifolene

0.7 ± 0.1

limonene

1 ± 0.2

21 ± 4

1.3 ± 0.3

13 ± 3

0.7 ± 0.2

0.1 ± 0.02

4.6 ± 1

myrcene

9.3 ± 2

32 ± 7

5.5 ± 1

0.6 ± 0.2

1.1 ± 0.3

0.1 ± 0.02

3-carene

1.7 ± 0.4

2 ± 0.4

26 ± 6

4.5 ± 1

19 ± 4

0.8 ± 0.2

6.9 ± 2

methyl chavicol

0.9 ± 0.1

0.5 ± 0.04

23 ± 0.4

-pinene

1 ± 0.2

0.9 ± 0.2

3.6 ± 0.7

1 ± 0.2

24 ± 5

0.8 ± 0.2

2.4 ± 0.5

terpinolene

29 ± 6

2.8 ± 0.6

1.7 ± 0.4

-pinene

0.7 ± 0.2

0.4 ± 0.1

1.1 ± 0.2

1.4 ± 0.3

0.8 ± 0.2

-terpinene

0.4 ± 0.1

1 ± 0.2

17 ± 4

2.2 ± 0.5

14 ± 3

0.7 ± 0.1

0.1 ± 0.02

14 ± 3

-terpinene

0.6 ± 0.1

0.6 ± 0.2

7.2 ± 2

0.9 ± 0.4

4.6 ± 1

3.4 ± 1

0.1 ± 0.03

2.7 ± 1

verbenone

3.2 ± 0.7

6.5 ± 1

1.8 ± 0.4

1.5 ± 0.3

0.2 ± 0.04

0.4 ± 0.1

linalool

20 ±

2.5 ± 0.5

0.7 ± 0.1

0.5 ± 0.1

0.4 ± 0.1

0.04 ± 0.02

0.3 ± 0.1

isoprene

0.6 ± 0.1

The terpenes listed in italic font have not been measured in the air above the Blodgett Forest canopy. The short lifetime of the two sesquiterpenes,

caryophyllene and

-humulene likely affect their detection in ambient air above the canopy, and linalool, typically associated with flowering citrus plants,

may not be emitted at all from Blodgett Forest. Verbenone, an oxygenated terpene associated with plant-insect interactions, has not been specifical

targeted for measurements at Blodgett Forest.

Yields listed in bold are yields greater than 10%, and all yields of unidentified product ions represent a lower limit due to lack of knowledge about

fragmentation.

Yield of

111. The sum of yields of

111 and 93 is 41 ± 7% (see text).

Yield of

111. Sum of yields of

111 and 93, the dehydrated fragment of

111, was 43 ± 7%.

Yield of

111. The sum of yields of

129, 111 and 93 is 75 ± 10%.

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

10 of 25

D17305

237), however, this ion was not observed from

-humulene

photooxidation, suggesting that this species was not

detected because it was rapidly oxidized, partitioned directly

into the particle phase, or fragmented completely in the

PTR-MS. Additionally, higher yields of lower molecular

weight unidentified ions, e.g.,

87, 99, 101, and 125 were

observed, indicating ring-opening did occur with cleavage

at multiple internal double bonds. Despite the higher yields

of comparatively low molecular weight ions, SOA yield

from

-humulene was still high. The gas- and particle-phase

products from the photooxidation of

-humulene was

studied elsewhere [

Jaoui and Kamens

, 2003a] using

denuder and filter sampling with sample derivatization

and analysis by GC-MS and HPLC. By accounting for nine

different oxidation products in the gas and particle phases,

Jaoui and Kamens

[2003a] obtained a carbon balance of

44%, suggesting that the use of surrogate standards

underestimated yields, or that additional products may be

produced. In the gas-phase, carbon yields of 3-seco-

humulone aldehyde, 7-seco-

-humulone aldehyde, and

humulal aldehyde (all MW = 236) dominated, and summed

to 9.5% [

Jaoui and Kamens

, 2003a]. Although we did not

observe a product ion at

237, we did observe

219,

which is expected to be the dehydrated fragment of

237.

Our observed yield of

219 (3 ± 0.5%) was lower than

the sum of the three C

products observed by

Jaoui and

Kamens

[2003a], which is expected as these compounds

likely produce additional fragments or are otherwise lost to

tubing walls. Our carbon mass balance for

-humulene was

77 ± 9%, suggesting that the gas-phase compounds ob-

served by PTR-MS as 32 product ions, though underesti-

mated, likely include additional products besides the nine

identified by

Jaoui and Kamens

[2003a].

3.1.3.

-Caryophyllene

[

] Figure 3 shows the time series of

-caryophyllene

photooxidation, indicating that the production of

207

occurs quickly, corresponding to OH addition to the termi-

nal C of the external double bond (Figure 4). Oxidation

products with MW = 236 and 252 have been proposed for

-caryophyllene ozonolysis [

Calogirou et al.

, 1997], and

represent first-generation oxidation products from O

attack

at the internal double bond. Although OH attack at the

internal double bond would also likely result in first-

generation products occurring at

237 and 253

(Figure 4), the observed time evolution and mixing ratio

237 suggest that this ion corresponds to a non-

dominant, late-forming (perhaps third-generation) product.

It is possible that the first-generation keto-aldehyde product

corresponding to

237 was formed but not observed in

the gas-phase because it partitioned directly into the particle

phase. The formation of oxidation products at

253 and

235 is observed but delayed compared to

207 and 189

(Figure 3). While

189 may be the dehydrated fragment

207, the time series of the two ions are not identical

(Figure 2) and the correlation coefficient is low (R

= 0.3),

suggesting that

189 is a different product, or that the

very low mixing ratios observed for

189 and 207 makes

a correlation difficult to detect. Figure 4 shows partial

mechanisms for the initial production and further oxidation

207. For comparison, the first-generation products

corresponding to MW = 236 and 252, proposed by

Calogirou et al.

[1997], that are not likely the products

observed in these photooxidation experiments, are illus-

trated in grey. The further oxidation of

207 may

produce the observed oxidation ions

219, 237, the

dehydrated fragments of

255 (Figure 4). Although the

Table 8.

Comparison of Gas-Phase Oxidation Product Yields From Sesquiterpenes, Isoprene, and Oxygenated Terpenes, on a Percent

Mole (ppb/ppb) Basis, Observed in This Study With Results From Other Terpene Photooxidation Experiments

Terpene

Product

Product Yield, % This Work

Product Yield, % Other Studies

Reference

-humulene

219

3 ± 0.5

9.5

[

Jaoui and Kamens

, 2003a]

linalool

acetone

[

Shu et al.

, 1997]

127

8.1 ± 2

6.8

[

Shu et al.

, 1997]

129, 111, 93

75 ± 10

[

Shu et al.

, 1997]

isoprene

formaldehyde

160

63 ± 10

[

Tuazon and Atkinson

, 1990]

57 ± 6

[

Miyoshi et al.

, 1994]

MVK + MACR

60–87 ± 11

[

Paulson et al.

, 1992]

55 ± 6

[

Zhao et al.

, 2004]

[

Tuazon and Atkinson

, 1990]

[

Miyoshi et al.

, 1994]

3-methylfuran

3.6 ± 0.4

4.4 ± 0.6

[

Atkinson et al.

, 1989]

4±2

[

Paulson et al.

, 1992]

[

Sprengnether et al.

, 2002]

1.5 ± 0.3

8.4 ± 2.4

[

Zhao et al.

, 2004]

1.3 ± 0.3

3.3 ± 1.6

[

Zhao et al.

, 2004]

101

0.5 ± 0.1

15%

[

Baker et al.

, 2005]

19.3 ± 6.1

[

Zhao et al.

, 2004]

Sum of yields of 3-seco-

-humulone aldehyde and 7-seco-

-humulone aldehyde, and

-humulal aldehyde, all with MW = 236. We did not observe

237, but did observe

m/z

219, its dehydrated fragment.

Yield of 6-methyl-5-hepten-2-one (C

O = 126 amu).

Yield of 4-hydroxy-2-methyl-5-hexen-1-al and its cyclized isomer (C

= 128 amu).

Likely overestimated due to calibration issue.

Yield of

85 measured by PTR-MS and attributed to a C

carbonyl (C

O = 84 amu).

Yield of

87 measured by PTR-MS and attributed to a C

hydroxycarbonyl (C

= 86 amu).

Estimated yield of two C

hydroxycarbonyls.

Yield of

101 measured by PTR-MS and attributed to a C

hydroxycarbonyl (C

= 100 amu).

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

11 of 25

D17305

production of a keto-keto-aldehyde from further oxidation

207 at the remaining internal double bond, with a

molecular weight of 238 (

239) seems likely, the ions

239 and 221, which are shown for reference in

Figure 4, were not observed from these experiments. It

should also be noted that many products, particularly the

acids, may also be produced from ozonolysis reactions.

[

] We conducted dark ozonolysis experiments of

-caryophyllene and

-humulene in 2003 [

Lee et al.

, 2006],

and in comparison, the photooxidation of these two sesqui-

terpenes produced higher SOA yields than ozonolysis.

Because the lifetime of these two sesquiterpenes from O

reaction is considerably shorter compared to OH oxidation,

the lower SOA yields of

45% from ozonolysis may be

atmospherically relevant.

3.2. Monoterpenes

[

] SOA yields from monoterpenes ranged from 25% to

58% (Table 5). Yields are generally expected to be higher

from terpenes with internal double bonds, because oxidation

does not necessarily result in carbon loss. This trend is

generally observed, with several terpenes with internal

double bonds, such as limonene, 3-carene,

-pinene, and

terpinolene, producing higher SOA yields than

-pinene, the

only single external double bond monoterpene tested. How-

ever, high SOA yields were observed from myrcene, which

is acyclic, and low SOA yields were observed from

- and

-terpinene, both with two internal double bonds.

3.2.1. Limonene

[

] The 58% SOA yield from limonene was highest of

all monoterpenes (Table 5). High yields of formaldehyde

were also observed from limonene, whereas yields of

acetaldehyde, formic acid, acetone, and acetic acid were

low. Limononaldehyde (C

) has been reported as a

limonene oxidation product [

Hakola et al.

, 1994], and the

PTR-MS observed high yields of the ions we determined to

be associated with limononaldehyde (Table 3a), with a total

yield of 68 ± 7%. In comparison to other studies (Table 9a),

we observed a very high yield of limononaldehyde and low

yields of limonaketone, formaldehyde, and formic acid.

The time series of limonene photooxidation is shown in

Figure 5. Limononaldehyde and limonaketone were

observed simultaneously, but higher yields of limononalde-

hyde suggest OH preferentially attacks limonene at the

internal double bond. Other ions are subsequently produced,

and are likely oxidation products of limononaldehyde, such

Table 9a.

Comparison of Gas-Phase Oxidation Product Yields From Monoterpenes, on a Percent Mole (ppb/ppb) Basis, Observed in

This Study With Results From Other Terpene Photooxidation Experiments

Terpene

Product

Product Yield, %

This Work

Product Yield, %

Other Studies

Reference

limonene

Formaldehyde

[

Librando and Tringali

, 2005]

36 ± 5

[

Larsen et al.

, 2001]

formic acid

54 ± 10

[

Larsen et al.

, 2001]

[

Librando and Tringali

, 2005]

Acetone

0.4

<0.03

[

Reissell et al.

, 1999]

not detected

[

Larsen et al.

, 2001]

139

4.9 ± 1

39 ± 15

[

Larsen et al.

, 2001]

20 ± 3

[

Hakola et al.

, 1994]

17 ± 3

[

Arey et al.

, 1990]

169

68 ± 7

29 ± 6

[

Hakola et al.

, 1994]

[

Arey et al.

, 1990]

not detected

[

Larsen et al.

, 2001]

myrcene

Formaldehyde

30 ± 6

[

Orlando et al.

, 2000]

formic acid

5 ± 2

[

Orlando et al.

, 2000]

Acetone

36 ± 5

[

Orlando et al.

, 2000]

[

Reissell et al.

, 1999]

111

9.3 ± 2

19 ± 4

[

Reissell et al.

, 2002]

3-carene

Formaldehyde

21 ± 4

[

Orlando et al.

, 2000]

12 ± 3

[

Larsen et al.

, 2001]

[

Librando and Tringali

, 2005]

formic acid

30 ± 3 11

[

Larsen et al.

, 2001]

[

Librando and Tringali

, 2005]

8±2

[

Orlando et al.

, 2000]

Acetone

15 ± 3

[

Orlando et al.

, 2000]

15 ± 3

[

Reissell et al.

, 1999]

Acetone

15 ± 5

[

Larsen et al.

, 2001]

m/z 169

77 ± 8

34 ± 8

[

Hakola et al.

, 1994]

[

Arey et al.

, 1990]

14 ± 5

[

Larsen et al.

, 2001]

From experiments where the photolysis of H

was used as the OH source.

Sum of

139 and 121.

Yield of limona ketone.

Sum of yields of

169, 170, 151 (the dehydrated fragment of

169), 152, 133, 123, 124, 107, and 108.

Yield of endolim, also called limononaldehyde.

Subject to significant error due to concentrations beyond range of calibrations.

Yield of 4-vinyl-4-pentenal.

Sum of

169, 170, 151 (the dehydrated fragment of

169), 152, 133, 123, 124, 107, and 108.

Yield of caronaldehyde.

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

12 of 25

D17305

171, 153, and 143. A partial mechanism for the

formation of these ions is shown in Figure 6. Other ions are

formed even later, such as

218, which is likely an

organic nitrate compound produced at a low observed yield

of <1% (Table 10), and

113, which was the dominant

ion observed at Blodgett Forest. Molar yields of other

unidentified ions totaled 61%, and the carbon mass balance

at the end of the experiment was 113 ± 17%, suggesting that

the concentrations of unidentified oxidation products to the

carbon balance are overestimated and may have resulted

from incorrect assignments of carbon number to the product

ions or from overestimating the %C in the SOA formed.

3.2.2. Myrcene

[

] The photooxidation of myrcene, an acyclic terpene,

produced relatively high yields of SOA (43%) compared to

other monoterpenes, in contrast to the low yields (

12%)

obtained from myrcene ozonolysis [

Lee et al.

, 2006]. The

lifetimes of myrcene with respect to 50 ppb O

or 2

molecules OH cm

are similar (28 minutes and 39 minutes,

respectively), suggesting that ozonolysis and photooxida-

Table 9b.

Comparison of Gas-Phase Oxidation Product Yields From Monoterpenes, on a Percent Mole (ppb/ppb) Basis, Observed in

This Study With Results From Other Terpene Photooxidation Experiments

Terpene

Product

Product Yield, % This Work

Product Yield, % Other Studies

Reference

-pinene

formaldehyde

23 ± 9

[

Noziere et al.

, 1999]

19 ± 5

[

Orlando et al.

, 2000]

[

Librando and Tringali

, 2005]

8±1

[

Noziere et al.

, 1999]

8±1

[

Larsen et al.

, 2001]

formic acid

28 ± 3

[

Larsen et al.

, 2001]

[

Librando and Tringali

, 2005]

7±2

[

Orlando et al.

, 2000]

acetone

15 ± 2

[

Wisthaler et al.

, 2001]

11 ± 3

[

Reissell et al.

, 1999]

11 ± 3

[

Larsen et al.

, 2001]

11 ± 3

[

Aschmann et al.

, 1998]

9±6

[

Noziere et al.

, 1999]

7±2

[

Noziere et al.

, 1999]

5±2

[

Orlando et al.

, 2000]

pinonaldehyde

63 (range 47–83)

87 ± 20

[

Noziere et al.

, 1999]

30 ± 0.3

56 ± 4

[

Hatakeyama et al.

, 1991]

37 ± 7

[

Noziere et al.

, 1999]

34 ± 9

[

Wisthaler et al.

, 2001]

31 ± 15

[

Vinckier et al.

, 1998]

31–34

[

Jaoui and Kamens

, 2001]

[

Arey et al.

, 1990]

[

Hakola et al.

, 1994]

28 ± 5

[

Aschmann et al.

, 2002]

6±2

[

Larsen et al.

, 2001]

153

2.9 ± 0.6

1.4–1.8

f, g

[

Jaoui and Kamens

, 2001]

155

0.8 ± 0.2

4–6.4

f, h

[

Jaoui and Kamens

, 2001]

171

0.2 ± 0.05

3.9–4.2

f, i

[

Jaoui and Kamens

, 2001]

185

0.3 ± 0.1

6.1–16.9

f, j

[

Jaoui and Kamens

, 2001]

[

Aschmann et al.

, 2002]

terpinolene

formaldehyde

29 ± 6

[

Orlando et al.

, 2000]

formic acid

8 ± 2

[

Orlando et al.

, 2000]

acetone

39 ± 5

[

Orlando et al.

, 2000]

36–45

[

Reissell et al.

, 1999]

111 and 93

43 ± 7

26 ± 6

[

Hakola et al.

, 1994]

26 ± 5

[

Reissell et al.

, 1999]

[

Arey et al.

, 1990]

19 ± 4

[

Reissell et al.

, 2002]

169

—

[

Arey et al.

, 1990]

From experiments where the photolysis of H

was used as the OH source.

Conducted in the absence of NO

See Experiment section for description of the pinonaldehyde yield calculation.

Sum of

169, 170, 151 (the dehydrated fragment of

169), and 152.

A ll yields reported by

Jaoui and Kamens

[2001] are % carbon yields. The yields listed above are converted from % carbon to % total mass yields for

comparison with yields obtained from this and other studies.

All yields reported by

Jaoui and Kamens

[2001] are % carbon yields. The yields listed above are converted from % carbon to % total mass yields for

comparison with yields obtained from this and other studies.

Yield of

-campholenal (MW = 152).

Yield of norpinonaldehyde (MW = 154).

Sum of norpinonic acid (MW = 170) and pinalic4-acid (MW = 170) yields, with yields of pinalic-4-acid roughly 12 times larger than yields of

norpinonic acid.

Sum of pinonic acid (MW = 184) and 1-hydroxypinonaldehyde (MW = 184) yields, with pinonic acid yields roughly 4–8 times larger than 1-

hydroxypinonldehyde.

Yield of 4-methyl-3-cyclohexen-1one (MW = 110).

Yield of 4-vinyl-4-pentanal (MW = 110).

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

13 of 25

D17305

tion are both important loss processes in the real atmo-

sphere. High yields of formaldehyde (74%) and acetone

(22%), and low yields of formic and acetic acid (<5%) were

observed. Our yield of acetone is slightly lower than the

36% reported by

Orlando et al.

[2000] and the 41% yield

reported by

Reissell et al.

[1999]. While our yield of formic

acid is the same as the yield obtained elsewhere [

Orlando et

al.

, 2000], our yield of formaldehyde is much higher than

the 30% yield of

Orlando et al.

[2000]. Because the

formaldehyde mixing ratios in the chamber (

110 ppb)

were beyond the range of the calibrated linear response, the

formaldehyde yield from myrcene is subject to considerable

error.

[

] Numerous compounds can be produced from myr-

cene due to the many possible sites for OH attack. Many

product ions were detected, with highest molar yields

from

93 (32 ± 6.8%) and 113 (32 ± 6.7%). The product

4-vinyl-4-pentenal (MW = 110) was identified from other

studies [

Reissell et al.

, 2002], and was observed from

myrcene ozonolysis at

111 and 93 [

Lee et al.

, 2006].

However, from these photooxidation studies,

111 and

93 were not well correlated (R

= 0.5) and exhibited

different loss rates with time (Figure 7) where

93 was

oxidized completely, as expected given that 2 two double

bonds remain in the structure, whereas

111 was not,

suggesting that other compounds likely interfered on

111. Additionally,

111 correlated well with other ions,

e.g.,

71 (R

= 0.96), 83 (R

= 0.97), 97, 113, and 139

(all R

= 0.93). These correlations were not observed from

the ozonolysis studies [

Lee et al.

, 2006], and may indicate

that these ions are fragments of the same compound, or that

these ions represent different compounds that are produced

and consumed at similar rates. 4-vinyl-4-pentenal is likely

detected at both

111 and 93, but another product or

Table 9c.

Comparison of Gas-Phase Oxidation Product Yields From Monoterpenes, on a Percent Mole (ppb/ppb) Basis, Observed in

This Study With Results From Other Terpene Photooxidation Experiments

Terpene

Product

Product Yield, % This Work

Product Yield, % Other Studies

Reference

-pinene

formaldehyde

[

Hatakeyama et al.

, 1991]

45 ± 8

[

Orlando et al.

, 2000]

23 ± 2

[

Larsen et al.

, 2001]

[

Librando and Tringali

, 2005]

30.3

[

Jaoui and Kamens

, 2003b]

formic acid

8.2

38 ± 4

[

Larsen et al.

, 2001]

[

Librando and Tringali

, 2005]

2±1

[

Orlando et al.

, 2000]

acetone

7.9

16 ± 2

[

Wisthaler et al.

, 2001]

11 ± 3

[

Larsen et al.

, 2001]

8.5 ± 2

[

Aschmann et al.

, 1998]

6.5

[

Jaoui and Kamens

, 2003b]

2±2

[

Orlando et al.

, 2000]

nopinone

79 ± 8

[

Hatakeyama et al.

, 1991]

27 ± 4

[

Hakola et al.

, 1994]

30 ± 5

[

Arey et al.

, 1990]

25 ± 5

[

Larsen et al.

, 2001]

25 ± 3

[

Wisthaler et al.

, 2001]

15.2

[

Jaoui and Kamens

, 2003b]

151

1.4 ± 0.3

0.9

[

Jaoui and Kamens

, 2003b]

153

2.8 ± 0.6

1.9

[

Jaoui and Kamens

, 2003b]

155

0.8 ± 0.2

5.3

[

Jaoui and Kamens

, 2003b]

171

0.1 ± 0.03

<0.1

[

Jaoui and Kamens

, 2003b]

185

0.1 ± 0.02

0.5

[

Jaoui and Kamens

, 2003b]

-terpinene

acetone

10 ± 3

[

Reissell et al.

, 1999]

-terpinene

acetone

[

Reissell et al.

, 1999]

From experiments where the photolysis of H

was used as the OH source.

Sum of yields of

-pinene oxide (1%) and 3-oxonopinone (0.9%). A trace amount of myrtenol was reported but not quantified and is not included in the

sum.

Sum of yields of 1-hydroxynopinone (3.1%) and 3-hydroxynopinone (2.2%).

Yield of 3,7-dihydroxynopinone.

Yield of pinonic acid.

Table 10.

Yields of Organic Nitrate Oxidation Products

Terpene

Yield, %

longifolene

220

0.8 ± 0.2

limonene

200

0.1 ± 0.02

218

0.2 ± 0.04

myrcene

0.1 ± 0.03

methyl chavicol

124

0.4 ± 0.1

168

0.5 ± 0.1

170

0.3 ± 0.1

180

0.5 ± 0.1

182

0.1 ± 0.03

-pinene

0.9 ± 0.2

200

0.2 ± 0.04

-pinene

108

1.9 ± 0.4

184

0.1 ± 0.05

198

0.2 ± 0.04

-terpinene

230

0.1 ± 0.02

-terpinene

178

0.05 ± 0.02

184

0.1 ± 0.05

202

0.1 ± 0.03

230

0.2 ± 0.04

verbenone

230

0.1 ± 0.02

linalool

188

0.2 ± 0.05

Reported yields are limited to those oxidation products with even mass-

to-charge ratios that are unambiguously not carbon isotopes of the

preceding

D17305

LEE ET AL.: TERPENE PHOTOOXIDATION PRODUCTS, YIELDS

14 of 25

D17305