PCCP View Article Online

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

PAPER

Cite this: Phys. Chem. Chem. Phys., 2015, 17, 10227

View Journal | View Issue

Aqueous benzene-diols react with an organic triplet excited state and hydroxyl radical to form secondary organic aerosol† Jeremy D. Smith,ab Haley Kinneya and Cort Anastasio*ab Chemical processing in atmospheric aqueous phases, such as cloud and fog drops, can play a significant role in the production and evolution of secondary organic aerosol (SOA). In this work we examine aqueous SOA production via the oxidation of benzene-diols (dihydroxy-benzenes) by the triplet excited state of 3,4-dimethoxybenzaldehyde, 3DMB*, and by hydroxyl radical,  OH. Reactions of the three benzene-diols (catechol (CAT), resorcinol (RES) and hydroquinone (HQ)) with 3DMB* or  OH proceed rapidly, with rate constants near diffusion-controlled values. The two oxidants exhibit different behaviors with pH, with rate constants for 3DMB* increasing as pH decreases from pH 5 to 2, while rate constants with  OH decrease in more acidic solutions. Mass yields of SOA were near 100% for all three benzene-diols with both oxidants. We also examined the reactivity of atmospherically relevant mixtures of phenols and benzene-diols in the presence of 3DMB*. We find that the kinetics of phenol and benzene-diol loss, and the production of SOA

Received 29th December 2014, Accepted 16th March 2015 DOI: 10.1039/c4cp06095d

mass, in mixtures are generally consistent with rate constants determined in experiments containing a single phenol or benzene-diol. Combining our aqueous kinetic and SOA mass yield data with previously published gas-phase data, we estimate a total SOA production rate from benzene-diol oxidation in a foggy area with significant wood combustion to be nearly 0.6 mg mair3 h1, with approximately half from the

www.rsc.org/pccp

aqueous oxidation of resorcinol and hydroquinone, and half from the gas-phase oxidation of catechol.

1. Introduction Organic aerosol (OA) is often a dominant component of particulate mass in the atmosphere, with much of the OA resulting from secondary production.1 Traditional secondary organic aerosol (SOA) models often fail to accurately describe observed mass of SOA in the atmosphere.2–4 In part, this might be because these models have typically only included gas-phase pathways to make low-volatility products.5 Recent model observations and laboratory work have shown that the aqueous processing of organics (i.e., reactions in cloud and fog drops and aqueous aerosol particles) can help bridge the gap between observed and modeled global SOA levels.5–11 That is, aqueous reactions of organic compounds by oxidants such as hydroxyl radical ( OH) and triplet excited states of organic chromophores (3C*) can be important sources of SOA and can increase the carbon oxidation state of organic aerosol.7,12–16 Many previous aqueous SOA studies have focused on small aliphatic compounds, such as glyoxal oxidation by  OH and the a

Department of Land, Air and Water Resources, University of California, Davis, 1 Shields Ave., Davis, CA 95616, USA. E-mail: [email protected]; Fax: +1-530-752-1552; Tel: +1-530-754-6095 b Agricultural and Environmental Chemistry Graduate Group, University of California, Davis, 1 Shields Ave., Davis, CA, USA † Electronic supplementary information (ESI) available. See DOI: 10.1039/c4cp06095d

This journal is © the Owner Societies 2015

oxidation products of isoprene and a-pinene.7,13,15,17–19 However, the aqueous reactions of aromatic SOA precursors are less well understood. While  OH is generally considered the dominant oxidant in atmospheric aqueous phases,20 models generally overestimate concentrations of aqueous  OH compared to measurements, and thus are likely overestimating the rates of  OH-mediated oxidations in aqueous drops.11 Another class of oxidants are the triplet excited states of light-absorbing organic compounds (3C*), which are important for the aqueous oxidation of phenols.16,21 These triplet states can rapidly abstract the phenolic hydrogen21,22 to form phenoxy radicals, which oligomerize to efficiently form low volatility, high molecular weight products with SOA mass yields near 100%.16,23,24 In addition to these aqueous reactions with phenols, triplet excited states also appear to be important in heterogeneous aerosol chemistry. For example, illuminated humic acid particles in the presence of gas-phase terpenes show significant increases in size and mass, likely from 3C* oxidation of the terpenes on the particle surfaces.25 Similarly, particles containing imidazol-2-carboxaldehyde in the presence of gaseous limonene show photosensitized aerosol growth via 3C*-mediated oligomerization.26 We are particularly interested in the aqueous reactions of phenols, i.e., molecules containing an aromatic ring with at least one hydroxyl group (–OH) substituent. A major source of

Phys. Chem. Chem. Phys., 2015, 17, 10227--10237 | 10227

View Article Online

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

Paper

atmospheric phenols is biomass combustion, both from residential burning and wildfires.27 The phenol family includes several main classes, including those based on the namesake ‘‘phenol’’ (i.e., C6H5OH (PhOH)), guaiacols (based on 2-methoxyphenol, i.e., guaiacol (GUA)), syringols (based on 2,6-dimethoxyphenol, i.e., syringol (SYR)), and benzene-diols (which contain two –OH substituents on the aromatic ring). We recently discussed the aqueous reactions of PhOH, GUA, and SYR,16 so our focus here is on benzene-diols. There are three isomers of the simplest benzene-diols (i.e., dihydroxybenzene with no other substituents): catechol (1,2-dihydroxybenzene; CAT), resorcinol (1,3-dihydroxybenzene; RES), and hydroquinone (1,4-dihydroxybenzene; HQ). All three of these isomers, along with other, more substituted benzene-diols, are emitted (primarily as gases) by biomass burning28–30 and have modest to high Henry’s Law constants, with values of approximately 103, 105, and 106 M atm1, for CAT, RES, and HQ, respectively.31 In addition to phenols, biomass burning also emits non-phenolic aromatic carbonyls, such as 3,4-dimethoxybenzaldehyde,28,32 which absorb significant amounts of solar radiation to form 3C*. Our goal in this work is to investigate the aqueous reactions of three benzene-diols (catechol, resorcinol, and hydroquinone) with the triplet excited stated of 3,4-dimethoxybenzaldehyde (3DMB*) and with the hydroxyl radical ( OH). For each combination of diol and oxidant we measured the second-order rate constant for reaction and the mass yield of SOA formed. We also investigated the kinetics and SOA product masses resulting from more complicated laboratory mixtures of phenols and benzene-diols oxidized by 3DMB*.

PCCP

crossing efficiency (FISC) of DMB were illuminated with 313 nm light in a temperature-controlled (20 1C) chamber in a monochromatic illumination system (MIS; Spectral Energy) with a 1000 W Hg/Xe lamp.36 Our dark control samples were wrapped in aluminum foil and also kept at 20 1C. We illuminated samples until approximately 50% of the initial amount of phenol had reacted. 2.3

Small aliquots of solution were periodically removed from the illuminated and dark cells to measure the concentrations of phenol and DMB using HPLC (Shimadzu LC-10AT pump, ThermoScientific BetaBasic-18 C18 column, Shimadzu-10AT UV-Vis detector (using detection wavelengths of 270, 276, 268, 281, 286, 289, and 307 nm for PhOH, GUA, SYR, CAT, RES, HQ, and DMB, respectively), degassed eluent of 20 : 80 acetonitrile : water, and a flow rate of 0.70 mL min1). The screw cap on each cell was taken off during aliquot removal to maintain air saturation in the reaction cell. The concentration of hydrogen peroxide was determined daily by absorbance at 240 nm (e240 = 38.1 M1 cm1)37 using a Shimadzu UV-2501PC spectrophotometer. We measured the photon flux on each experiment by determining the photolysis rate constant ( j2NB,exp) of aqueous 10 mM 2-nitrobenzaldehyde (2NB) in the same cell used for the sample illumination.38 Solution pH was measured using an Orion model 420A pH meter. The vast majority of experiments followed a first-order decay during illumination (497%), and experiments were not used in further analysis if first-order decay was not observed. 2.4

2. Methods 2.1

Chemicals and solutions

All chemicals were used as received. Acetonitrile (99.9%; HPLC grade), catechol (99+%), and hydroquinone (99.5%) were from Acros; resorcinol (99%), 3,4-dimethoxybenzaldehyde (DMB) (99%) were from Sigma-Aldrich; hydrogen peroxide (30%) and sulfuric acid (trace metal grade) were from Fisher. All solutions were made using air-saturated purified water (Milli-Q) from a Milli-Q Plus system (Millipore; Z18.2 MO cm) with an upstream Barnstead activated carbon cartridge (Part # D63112). Solution pH was adjusted using sulfuric acid for pH r 5. Solutions contained 5–100 mM of diol (CAT, RES, or HQ) and either 5 mM of 3C* precursor (DMB) or 100 mM HOOH (as a source of  OH). We used diol, DMB, and HOOH concentrations based on atmospheric levels of these species.11,21,28,32–34 In this work we use the abbreviation ‘‘PhOH’’ to refer specifically to phenol (i.e., the compound C6H5OH) and the terms ‘‘phenol’’ and ‘‘ArOH’’ to refer generically to any of the three phenol species (PhOH, guaiacol (GUA), and syringol (SYR)), and all three isomers of benzene-diols. 2.2

Illumination

Air-saturated samples were illuminated in an airtight quartz cell (1 or 5 cm path length) (Spectrocell) and continuously stirred at 20 1C using simulated sunlight from a 1000 W Xe arc lamp with downstream optical filters.16,35 Solutions to determine the intersystem

10228 | Phys. Chem. Chem. Phys., 2015, 17, 10227--10237

Chemical analysis

Hydroxyl radical relative rate experiments

The bimolecular rate constant for each phenol with  OH was determined using a relative rate method.39,40 To do this, we made a solution containing the test phenol (10 mM), a probe compound with a known second-order rate constant (Table S1, ESI,† 10 mM), and 1.0 mM HOOH as a photochemical source of  OH. The desired bimolecular rate constant (kArOH+OH) is calculated from:     ½Probe0 kProbeþOH ½ArOH0 ¼ (1)  ln ln ½Probet kArOHþOH ½ArOHt where kProbe+OH and kArOH+OH are the 2nd-order rate constants for the reaction of hydroxyl radical with the probe and phenol of interest, respectively, and [ArOH/Probe] is the concentration of probe or ArOH at time zero or time t. The ratio kProbe+OH/kArOH+OH is calculated as the slope of a plot of the natural logarithm of the ratios of the concentrations of probe and phenol during illumination. We used several probe compounds (benzoic acid, phenol, guaiacol, and benzene; Table S1, ESI†) to span the range of reactivities of the benzene-diols across our studied pH range (see below). 2.5

Triplet excited state oxidation kinetic analysis

We give a full description of our kinetic analysis in Smith et al.16 and an abbreviated version here. The measured, apparent first 0  order rate constant for phenol loss k Light was determined as the negative of the slope of a plot of ln([ArOH]t/[ArOH]0) versus illumination time, where [ArOH] is the concentration of benzene-diol

This journal is © the Owner Societies 2015

View Article Online

PCCP

Paper

state of DMB, [DMB] is the molar concentration of DMB, kQ is the second-order rate constant for the non-reactive interaction of ArOH with 3DMB* (which doesn’t result in diol loss), and kArOH+3DMB* is the second-order rate constant for reaction of 3 DMB* with ArOH (which results in diol loss). The dominant sink of 3C* in air-saturated natural waters is molecular oxygen (O2) rather than non-reactive decay to the ground state.22,43–45 The apparent second-order rate constant for reaction of triplet and diol, kArOH+3DMB*, is a mole-fraction-weighted combination of the second-order diol rate constants with the protonated and neutral triplet states (kArOH+HT and kArOH+T, respectively):

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

(at times t and zero). We used an analogous procedure to determine the DMB loss rate. There was no significant loss of diol or DMB in 0 any dark control ( p 4 0.05). Values of k Light were normalized to sunlight conditions at midday on the winter solstice at Davis (solar zenith angle = 621; j2NB,win = 0.0070 s1)35 and were corrected for the small amount of internal light screening due to DMB light absorption: " # 0 k Light 0 (2) k ArOH ¼  j2NB;win Sl  j2NB;exp 0

In this equation k ArOH is the normalized first-order decay constant, and Sl is the internal light screening factor, with a value of 1 indicating no screening.41,42 Values of Sl were determined for wavelengths around the peak in the light absorption action spectrum for DMB (310–335 nm); for 5 mM DMB, values of Sl ranged from 0.85 for a 5 cm cell to 0.97 for a 1 cm cell. The DMB triplet state can be protonated and the protonated form has a pKa of approximately 3.3.16,21 Following recent work,16 we performed our diol kinetic experiments at pH 2 and 5 to characterize the reactivity with the protonated and neutral triplet state of DMB, respectively. At pH 2 the mole fractions of protonated triplet (aHT) and neutral triplet (aT) are 0.95 and 0.05, respectively; at pH 5, values of aHT and aT are 0.02 and 0.98, respectively. The protonated triplet is significantly more reactive (i.e., has higher rate constants with phenols) than the neutral triplet,16,21 and small adjustments must be made to the pH 5 experiments to correct for this large reactivity differ0

kArOH+3DMB* = aHT  kArOH+HT + aT  kArOH+T

As a simplified form of eqn (3), we fit our experimental data to the form of eqn (5) (Sigmaplot Version 11): 0

k ArOH ¼

2.6

0

0

3.1 Oxidation of benzene-diols by the triplet excited state of DMB To determine second-order rate constants for the diols with 3 DMB*, we examined the kinetics for loss of the three benzenediols as a function of concentration in illuminated pH 2 and 5

!

where kO2+3DMB* is the second-order rate constant for the 0

reaction of 3DMB* with oxygen, k 3 DMB is the first-order rate constant for decay of 3DMB* to the ground state, jhn,DMB is the rate constant for light absorption by DMB, FISC is the intersystem crossing efficiency from the singlet to triplet excited

This journal is © the Owner Societies 2015

ðMass of illuminated sample  Mass of dark sampleÞ Mass of phenol reacted (6)

3. Results and discussion

0

kO2 þ 3 DMB ½O2  þ k 3 DMB jhn;DMB FISC ½DMB  kArOHþ 3 DMB

(5)

Secondary organic aerosol mass measurements

YSOA ¼

diol loss, k ArOH increases. A kinetic model of this system for DMB as the triplet pre-cursor yields eqn (3):16 0

1 a þ b½ArOH

SOA mass yields were determined for pH 5 solutions containing 100 mM diol and either 5 mM DMB or 100 mM HOOH. Samples were illuminated until half of the initial diol had reacted. Then we placed 8 or 10 mL of solution in an aluminum cup and blew it down with specialty-grade nitrogen gas (99.997%, Praxair) to dryness. (Aluminum cups were made by pressing a 6 cm diameter piece of foil into a mold, baking at 500 1C for 8–12 hours to remove organics, and weighing the empty cup with a CAHN 29 electrobalance (precision  1 mg)).16 This is a vigorous evaporative method and recent work has shown that 16–38% of volatile and semivolatile organic mass is lost to evaporation.24 After blow down we reweighed the cups and calculated the SOA mass yield using:

ence. The rate constant k ArOH at pH 2 is essentially equal to k HT ; however, at pH 5 the small fraction of protonated triplet can affect the observed rate constant. Therefore, all data shown at pH 5 has been corrected for the contribution of the protonated triplet; this correction ranged from 1–30% for the three diols (ESI,† Section S1). We performed SOA mass experiments at only pH 5; these determinations do not work at pH 2 because of the large amount of sulfuric acid, which is very hygroscopic. The aqueous kinetics of phenols with 3C* has been explained in detail elsewhere,16 and we will describe it only briefly here. In solution there is a competition between the three major fates of 3C*: reaction with dissolved molecular oxygen, relaxation to the ground state, and oxidation of phenol. As the amount of diol decreases, the steady-state concentration of 3C* increases, thus the apparent first-order rate constant for

k ArOH ¼

(4)

 þ

1

kArOHþ 3 DMB þ kQ jhn;DMB FISC ½DMB  kArOHþ 3 DMB

(3)

  ½ArOH

solutions containing 3,4-dimethoxybenzaldehyde (DMB). There is no significant loss of the diols or DMB in dark control samples. In illuminated solutions, each diol follows firstorder decay (as illustrated in Fig. S1, ESI†), as we saw in our previous work for PhOH, GUA, and SYR.16,21 Also similar to our

Phys. Chem. Chem. Phys., 2015, 17, 10227--10237 | 10229

View Article Online

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

Paper

PCCP

For CAT and RES there is no significant direct photodegradation (i.e., no loss of benzene-diol in illuminated solutions not containing DMB), but there is for HQ. The direct photodegradation of HQ is significantly slower than its loss in the presence of 3DMB*. The rate constant for HQ direct photodegradation ( jHQ) depends on HQ concentration and ranges from 9.5  104 min1 at 10 mM HQ to 5.3  104 min1 at 100 mM HQ, corresponding to aqueous lifetimes of 1.1  103 and 1.9  103 min, respectively (Fig. S2, ESI†). There is no pH dependence of the direct photodegradation. The concentration dependence is most likely due to reaction of the HQ triplet excited state with a ground state HQ molecule, which is in competition with an unknown first-order loss process that also depends on [HQ]. Based on our experimental results and past literature,46 we have constructed a model of HQ photochemistry, which is shown in Fig. S3 (ESI†). We used these results to correct for the loss of HQ due to direct photodegradation in our subsequent experiments with 3DMB* and  OH. We next examined the effects of pH and diol concentration on the kinetics of diol loss by the DMB triplet state. At pH 2, 3 DMB* oxidation of HQ is slightly slower than that of CAT and RES, which have similar kinetics (Fig. 1A). At pH 5, the observed first-order rate constants are not statistically different between the three diols, except at the lowest concentrations, where the rate of HQ loss is higher (Fig. 1B). As shown previously,16,21 the reactivity of the triplet excited state of DMB with phenols is pH dependent: the protonated form of 3DMB* (i.e., HT) has a pKa of 0

3.3 and first-order rate constants with phenols (k HT ) that are 0

Fig. 1 First-order rate constants for benzene-diol loss due to reaction with the triplet excited state of 3,4-dimethoxybenzaldehyde (3DMB*) as a function of initial benzene-diol concentration. Panel A shows experiments done at pH 2 (k 0 HT), where oxidation is by the protonated triplet state. 0 Panel B represents experiments done at pH 5 (k T ), where values are corrected for the contribution of the protonated triplet, and only show oxidation by the neutral form of the DMB triplet. Lines are non-linear regression fits to the data using eqn (5). Error bars represent 1 SE, propagated from 0

uncertainties in k Light and j2NB. Rate constants are normalized to Davis, CA winter solstice sunlight.

previous work,16 we observe no significant loss of the sensitizer, DMB, in the majority (approximately 70%) of our illuminated samples. In experiments with DMB loss, the loss is less than 20% of the diol loss.

Table 1

Regression parameters for apparent rate constants for loss of diols by the protonated (k 0 HT) and neutral (k 0 T) triplet excited states of DMB 0

0

k HT (pH 2) a

Catechol Resorcinol Hydroquinone a

higher than those of the neutral triplet (T; rate constant k T ).16 As shown in Fig. 1, the oxidations of CAT, RES, and HQ by 3 DMB* are 1.7 to 5.8 times faster at pH 2 than pH 5 (for diol concentrations of 5–100 mM), consistent with a more reactive protonated triplet state. For all three benzene-diols the observed first-order rate constant for loss increases with decreasing diol concentration (Fig. 1). We use the regression parameters derived from the fits of eqn (5) to these data (Table 1) to extract fundamental kinetic data for the triplet-diol reactions. First, we derive the secondorder rate constant, kArOH+3DMB*, from regression parameter ‘a’ in eqn (5). In our previous work16 we used literature data to estimate that the quantum yield for intersystem crossing, FISC, is 0.06  0.04.21 To better constrain this parameter, in this current work we performed new experiments to determine that FISC for DMB is 0.10  0.03 (Section S3, ESI†). Using this value and the molar absorptivities of DMB,16 we estimate that the rate of triplet formation (R3DMB*,F = jhn,DMB FISC[DMB]) in

1

k T (pH 5) a

1

a

1

fReaction a

1

a (min )

b (min mM )

a (min )

b (min mM )

pH 2b

pH 5b

98 (23) 87 (4) 95 (11)

1.0 (0.7) 1.5 (0.2) 3.0 (0.8)

432 (42) 353 (25) 167 (13)

11 (2) 6.0 (1.0) 14 (1.3)

0.45 (0.36) 0.29 (0.10) 0.14 (0.06)

0.04 (0.01) 0.07 (0.03) 0.03 (0.01)

Parameters were found by fitting eqn (5) to data in Fig. 1. Errors are (1 SE) calculated from the regression fit.

10230 | Phys. Chem. Chem. Phys., 2015, 17, 10227--10237

b

Calculated using eqn (7).

This journal is © the Owner Societies 2015

View Article Online

PCCP Table 2

Paper Second-order rate constants for reactions of phenols with hydroxyl radical ( OH) or the triplet excited state (3C*) of DMB at 293 K 

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

Phenol Guaiacol Syringol

3

OH

This work kArOH+ OH (109 M1 s1)

Prior literature kArOH+ OH (109 M1 s1)

pH 2

pH 2

1.8 (0.3) 6.8 (3.1) 15 (7) 

pH 5

a

15 (5) 16 (5) 20 (4)

1.99 — —

pH 2

13 (9) 20 26

c

2.0 (0.5) 3.2 (0.7)d 6.7 (1.5)d 3

1

This work kArOH+ OH (10 M

1

s )

9

1

Prior literature kArOH+ OH (10 M c

pH 5 d

OH 9

Catechol Resorcinol Hydroquinone

pH 6b

C*

This work kArOH+3DMB* (109 M1 s1)

1

s )

0.29 (0.09)d,e 2.5 (0.6)d 3.5 (0.8)d

C*

This work kArOH+3DMB* (109 M1 s1)

pH 2

pH 5

pH 2

pH 9

pH 2

pH 5

2.5 (0.3) 1.6 (0.1) 14 (2)

6.9 (2.4) 5.8 (1.3) 11 (8)

— — —

11 12 29

2.5 (1.0) 2.9 (0.9) 2.6 (0.9)

0.58 (0.20) 0.71 (0.24) 1.5 (0.5)

a Taken from Herrman et al.69 b Taken from Buxton et al.51 c Average (1s) of two literature values presented in Buxton et al.51 d Recalculated using data from Smith et al.16 in conjunction with the new FISC determined in this work. e Changes in the treatment of the PhOH experimental data and accompanying regression since the Smith et al.16 report are summarized in Fig. S4 (ESI).

our experiments (normalized to midday, winter solstice sunlight in Davis) is 2.3 (0.5) mM min1. Previous measurements have determined that O2 is the dominant sink of 3C* in air-saturated aqueous solutions, with kO2+3C* = 2.0  109 M1 s1.43–45,47,48 Based on this rate constant, and using an O2 concentration in air-saturated solution29 of 291 mM at 20 1C, we used eqn (3) and (5) with the information above to calculate kArOH+3DMB* for the three diols. At pH 2, values of kArOH+HT, (2.0–6.7)  109 M1 s1, are approximately four times faster than rate constants at pH 5 (i.e., kArOH+T), (0.58–1.5)  109 M1 s1. Using our updated intersystem crossing yield, we also recalculated values of kArOH+3DMB* for the reactions of 3DMB* with phenol (PhOH; C6H5OH), guaiacol (GUA), and syringol (SYR) (Table 2). Because our new value for FISC is 1.7 times larger than the previously estimated value, the recalculated rate constants for GUA, and SYR are 1.7 times lower than previously listed.16 The value with phenol is 2.2 times higher than previously reported16 due to incorporation of the new FISC value as well as a reassessment of the original kinetic data (Fig. S4, ESI†). This new kPhOH+3DMB* is within experimental error of the value determined in a recent relative rate study comparing PhOH and SYR.49 Overall, rate constants for diol oxidation by 3DMB* are intermediate between values for PhOH and guaiacol (Table 2).16 As the second step in our kinetics determination, we used the fitted regression parameter ‘b’ in eqn (5) to determine the fraction of benzene-diol interaction with 3DMB* that lead to diol loss rather than unreactive quenching of the triplet state with no resulting diol loss. While energy transfer from 3DMB* to the diols is insignificant due to a large triplet energy barrier,22 there may still be significant physical quenching of a 3C* without a transfer of triplet energy. Similarly, the phenoxyl radical formed from oxidation of a phenol by 3C* can be reduced back to the parent phenol,22,50 which would appear to be a path of no reaction in our system since we monitor the loss of diol. Therefore, the rate constant for unreactive quenching, kQ, includes both physical quenching of 3DMB* by a diol, and reactive pathways followed by reduction of the phenoxy radical back to the parent diol. We can rearrange the regression

This journal is © the Owner Societies 2015

parameter ‘b’ to calculate the fraction of phenol that undergoes transformative reaction with 3DMB*: fReaction ¼ 

kArOHþ 3 DMB 1 ¼ (7) b  j kArOHþ 3 DMB þ kQ hn;DMB FISC ½DMB

As shown in Table 1, the major interaction of diols with 3DMB* in our solutions is not irreversible reaction, but instead is quenching of 3DMB* with no observed loss of diol (Table 1): over 90% of the diol–3DMB* interactions at pH 5 (and 55–86% of the interactions at pH 2) result in no loss of the diol. Values for kQ range from (3.1–45)  109 M1 s1 and are summarized in Table S3 (ESI†). This agrees with previous work where only 10% of PhOH–3C* interactions lead to loss of PhOH at pH 5.22 As with PhOH, it is likely that oxidation of the benzene-diols to make a phenoxy radical is generally followed by reduction of the radical (e.g., by HO2 ) to reform the parent diol rather than producing new products.22 We calculate the steady state concentrations of 3DMB* in solution by combining kQ and kArOH+3DMB* (eqn (S21), ESI†) at pH 5 as 7.1  1015 to 6.7  1014 M for HQ at 100 mM and RES at 1 mM, respectively. 3.2

Oxidation of methoxyphenols and benzene-diols by  OH

There are some bimolecular rate constants for hydroxyl radical ( OH) oxidation of phenols and benzene-diols in the literature, but most are at relatively high pH values (pH 6 and 9) or in organic solvents.20,50,51 To better understand the fate of aqueous phenols in atmospheric drops and deliquesced particles, we measured the second-order rate constants for reaction of  OH with six phenols and benzene-diols at pH 2 and 5 (Fig. S5, ESI†). As shown in Table 2, our derived rate constants agree well with the available literature values for PhOH at pH 2 and pH 6. We could only find literature values for the diol rate constants at pH 9;51 because this is close to the pKa of the diols there will be significant amounts of the more reactive phenolate ion, and so the pH 9 results cannot be compared to values at pH 2 or 5. All of our measured apparent second-order rate constants with  OH are very fast, in the range of 109 to 1010 M1 s1. For PhOH, GUA and SYR, rate constants are 8.3, 2.4 and

Phys. Chem. Chem. Phys., 2015, 17, 10227--10237 | 10231

View Article Online

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

Paper

PCCP

1.3 times faster at pH 5 than at pH 2, respectively. For the diols these ratios are 2.8, 3.6 and 0.8 (with the latter not statistically different from 1) for CAT, RES, and HQ, respectively. The observed increase in the phenol +  OH rate constants with increasing pH has been seen in previous work, where oxidation of PhOH by  OH at pH 2 was found to be very similar to our value.50 While we did not experimentally explore the mechanism for this effect, the literature suggests that the acidity or high ionic strength at pH 2 might either slow the  OH attack on phenols or significantly decrease the lifetime of the hydroxycylcohexadienyl radical intermediates formed from  OH addition, thus slowing irreversible transformation of the diols.52–54 3.3

Secondary organic aerosol mass yields

We performed SOA mass experiments using illuminated solutions containing 100 mM diol and either 5 mM DMB or 100 mM hydrogen peroxide. After illuminating until half of the diol reacted, we blew down the illuminated and dark solutions and determined the SOA mass yield from the difference (eqn (6)).16 The diols used here are relatively volatile (with saturation concentrations of 2.2  104, 2.9  104, and 1.4  102 mg m3 for CAT, RES, and HQ, respectively, at 298 K55) and should evaporate completely during our nitrogen blowdown.28,32 However, after blowdown we found a significant fraction of the unreacted diols remaining in the sample cups (Fig. S7, ESI†). To correct for this, we reconstituted each blown-down sample in water to one-half the original volume, quantified the remaining diol using HPLC, and corrected the illuminated sample mass for the amount of remaining reactants, as well as for the mass of low-volatility material in the dark blowndown sample (eqn (6) and Fig. S6, ESI†). The illuminated diol + oxidant solutions formed significant amounts of low-volatility products, while the corresponding dark samples had significantly lower masses (Fig. S7, ESI†). Average SOA mass yields (1s) from the 3DMB* oxidation of aqueous CAT, RES and HQ are of 72 (25)%, 112 (35)%, and 113 (31)%, respectively (Fig. 2). These are similar to our previously measured yields of near 100% for the 3DMB* oxidation of PhOH, GUA and SYR.16 In comparison, SOA mass yields are generally somewhat lower for the  OH oxidation of aqueous RES and HQ, with yields (n = 1) of 59%, and 82%, respectively, but are similar for CAT (80%, n = 1); we expect that the uncertainty in these yields is similar to the 3C* experiments, which have an average RSD of 31% (Fig. 2). The direct photodegradation of HQ also produces SOA very efficiently, with a mass yield of 87 (8)% (n = 3; Fig. 2). Our recent work,24 as well as a recent investigation of SOA from direct photodegradation of vanillin (4-hydroxy-3-methoxybenzaldehyde),56 suggest that different aqueous reaction pathways ( OH vs. direct photodegradation) result in the different products. Our results suggest a possible difference in reaction mechanisms between the oxidants, with 3C* oxidation producing higher molecular weight products, possibly with less fragmentation.16 There have been several recent studies of the formation of SOA from catechol in the gas phase and condensed phase,57–60 but to the best of our knowledge there have been no studies quantifying aqueous SOA from RES or HQ with  OH or 3C*.

10232 | Phys. Chem. Chem. Phys., 2015, 17, 10227--10237

Fig. 2 Average SOA mass yields (YSOA) from oxidation by the DMB triplet state (n = 3 or 4; darker bars at left for each compound), by  OH (n = 1; lighter bars), and – for HQ only – direct photodegradation (n = 2; white bar). Error bars are 1s, calculated from replicate experiments except for  OH, where we used the average relative standard deviation (31%) from the other experiments.

Coeur-Tourneur et al. found that gas-phase ozonolysis of CAT and methylcatechols results in SOA mass yields ranging from 17–86%,57,58 while Ofner et al. found that ozonolysis under humid conditions increased the mass yields and the oxygen levels of the products.59 Several recent studies on gas-phase  OH oxidation (of PhOH, GUA, SYR, and CAT) under atmospherically relevant conditions have reported SOA mass yields (10–86%) that are generally lower than our observed aqueous oxidation mass yields, with the exception of CAT (Fig. 2).59–64 However, our diol results are similar to aqueous SOA formation yields reported from phenol oxidation by  OH23,65 and 3C*.16 3.4

Phenol kinetics and SOA formation in mixtures

In this and previous work we investigated the reaction of a single phenol in each solution. To better mimic the complexity of atmospheric drops, in this section we study aqueous SOA production in mixtures of phenols, using DMB as a 3C* precursor. The total aqueous phenol concentration in each solution was 100 mM (representative of an area with significant biomass burning), with the contributions of individual phenols varying to represent differences between softwood and hardwood burning (Section S4, ESI;† Fig. 3A).34 Because the oxidation of PhOH produces CAT and HQ,64 we cannot accurately follow the kinetics of benzene-diols in mixtures containing PhOH. Thus we examined the hardwood (HW) and softwood (SW) mixtures using two solutions for each: one with PhOH but no diols, and one with diols but no PhOH. RES was excluded from these mixtures due to the difficulty in separating it from CAT and HQ in our isocratic HPLC. Fig. 3A shows the composition of each of the four mixtures: Hardwood mixtures are dominated by SYR, while softwood mixtures are dominated by PhOH (in the No Diols solution) and CAT (in the No PhOH solution). Each solution was illuminated until half of the initial total phenol concentration had reacted. For each mixture we calculated the expected initial rate of loss for each phenol, RArOH,L,

This journal is © the Owner Societies 2015

View Article Online

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

PCCP

based on a calculated steady-state concentration of the DMB triplet (eqn (S18) and (S19), Section S4, ESI†). These predicted rates of phenol loss, along with our measured rates, are shown in Fig. 3B. We start by examining the kinetics in the ‘‘No Diols’’ solutions. In the hardwood mixture (HW no diols) conditions the predicted total RArOH,L is approximately 1% lower than the measured loss rate for total phenols (Fig. 3B). Examining the individual phenols in this mix, predicted rates of GUA and SYR loss are close (within 1 SE) to measured values (29 and o1% lower, respectively), but the predicted loss rate of PhOH is approximately 74% slower than measured. We suspect that the larger difference for PhOH is due to production of triplet precursors in the illuminated solution: SYR and GUA react much faster than PhOH16 (Table S4, ESI†) and form carbonyl products that may form 3C* during illumination,24 thereby increasing PhOH loss. Another possible explanation for the higher-than-expected measured value of RArOH,L in the ‘‘HW no Diols’’ mixture is the production of HOOH, a precursor for  OH, from the oxidation of phenols by 3C*;21 due to the large fraction of SYR and GUA reacted, there may be an increase in  OH that may contribute to PhOH loss that is not accounted for in our prediction. Overall, SYR loss is the best predicted of this mixture, and is also the largest contributor to the group, resulting in an overall accurate prediction (within 1%) relative to the measured values. In the softwood (SW) mixture with no diols (SW no diols) we see excellent agreement between calculated and measured phenol loss rates, with our predicted value of RArOH,L only 10% higher than measured. Within this total, PhOH predictions are approximately 10% lower than measured, while GUA predictions are 14% higher than measured (Fig. 3B). In the HW mixture containing diols (and no PhOH) there is relatively poor agreement, with the calculated total rate 59% higher than the measured value. Part of this disagreement is caused by our inability to measure CAT loss in this experiment due to a coeluting peak. However, if we assume the ‘‘measured’’ CAT rate is equal to the calculated value, the calculated total rate of phenol loss would still be 48% higher than the measured value, largely driven by the overestimate of SYR loss (Fig. 3B). Since the ‘‘No Diols’’ HW case showed the opposite problem (i.e., the measured rate of SYR loss is greater than calculated rate), it does not appear to be a problem with the SYR rate constant. There is evidence to suggest under aqueous and gas-phase conditions that demethoxylation of methoxy-phenols can occur,64,66,67 which would result in the conversion of SYR to GUA, and may be contributing to the smaller than predicted rates loss of GUA in the HW mixtures with and without diols. In addition, our estimates tend to have increasing errors with increasing illumination times for the two diol (No PhOH) experiments, suggesting that in longer experiments secondary chemistry is occurring that we are not capturing in our rate calculations. 3.5

Secondary organic aerosol in phenol mixtures

In addition to examining the 3DMB*-mediated kinetics in phenol mixtures (Fig. 3B), we also examined SOA production

This journal is © the Owner Societies 2015

Paper

Fig. 3 Composition and kinetics in pH 5 aqueous phenol mixtures containing 5 mM DMB as the triplet precursor. Panel A describes the contributions of individual phenols to the total phenol concentration of 100 mM in the hardwood (HW) and softwood (SW) mixtures with PhOH (No Diols) and with diols (No PhOH). These fractions were calculated using wood combustion emissions data,28 literature Henry’s law constants (see main text), a temperature of 5 1C and a liquid water content of 1.0  107 Laq Lg1 (Section S3, ESI†). Panel B shows measured and calculated initial rates of phenol loss (RArOH,L) for each mixture, normalized to Davis, CA winter solstice sunlight. RArOH,L was calculated 0 using k ArOH determined from eqn (S18) (ESI†), calculated steady-state concentrations of 3C* (eqn (S19) and Section S5, ESI†) and initial [ArOH] (Table S4, ESI†). Error bars represent 1 SE of the total RArOH,L from all species, 0

propagated from values of kArOH+3DMB*, k Light , and j2NB. We were unable to quantify CAT kinetics in the ‘‘HW No PhOH’’ experiment due a co-eluting SYR oxidation peak. Panel C shows predicted (colored bars) and measured (grey bars) SOA mass concentrations in each mixture after illumination (until half of the initial total phenol had reacted) and then blow-down with N2. Predicted masses were calculated using the SOA mass yields of individual phenols (eqn (8); Fig. 2 and Smith et al.16) and the amount of each phenol lost in the experiment. Error bars of measured values are the analytical uncertainties of mass measurements (1 SE), while errors on predicted values are 1 SE, propagated from YSOA and the mass uncertainty from the change in phenol during oxidation.

Phys. Chem. Chem. Phys., 2015, 17, 10227--10237 | 10233

View Article Online

Paper

PCCP

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

in these same solutions. There was no measureable loss of ArOH in the dark during these mixture experiments, however, low-volatility mass was formed during blowdown of dark solutions containing diols, as mentioned in Section 3.3. For each mixture solution we predicted the total low-volatility mass (mPred) produced using: P (8) mPred = (D[ArOH]  YSOA) where D[ArOH] is the amount of a phenol that reacted during illumination (mg mL1; Table S4, ESI†) and YSOA is the SOA mass yield of that individual phenol (mg SOA formed/mg phenol reacted; Fig. 2 and ref. 16). As an initial test, we measured the SOA mass concentration in the HW (no PhOH) mixture after oxidation by OH: as shown in Fig. S10 (ESI†), the measured SOA mass concentration is in good agreement (within 13%) of the calculated value. Fig. 4C shows that SOA mass production in mixtures of phenols oxidized by 3C* is essentially additive. For the HW mixtures (i.e., with and without diols) the measured and predicted SOA masses after blowdown are in excellent agreement, within 6% of each other. SOA masses in these experiments are dominated by syringol oxidation due to the large amount of SYR in HW burning emissions (Fig. 3A) and its fast reaction rates. For the softwood mixtures the agreement between the measured and predicted SOA masses is good (within 30% for the two cases), though not as close as for the HW mixtures. The slight under-prediction may come from secondary reactions during the longer illumination times in the SW mixtures (Table S4, ESI†) producing low-volatility products not accounted for in our calculation. Overall, our results indicate that the SOA masses from different phenols are additive and that the yields are independent of phenol concentration. While our results indicate that the aqueous SOA masses from phenols are additive, our dataset is relatively small and our system does not contain the many other chemical components typical of fogs and aqueous aerosols. To more thoroughly test our additive result, future research should include more chemically diverse solutions (e.g., added ions and organic components typical of fog) over a wider range of conditions. 3.6

Benzene-diol reactions in a foggy/cloudy atmosphere

Our goal in this section is to explore the behavior and importance of benzene-diol oxidation in a foggy or cloudy atmosphere. We start by examining the relative importance of  OH and 3C* as oxidants for benzene-diols in cloud or fog drops, assuming that our bulk kinetics and SOA yields apply to individual drops. As shown in Fig. 4A, for intermediate oxidant concentrations (i.e., the middle portion of each bar) in pH 5 drops affected by wood combustion, 3C* and  OH are similarly important, with the 3 C*/ OH ratio of rate constants equal to 0.8, 1.0, and 0.7 for CAT, RES, and HQ, respectively. In contrast, at pH 2 and intermediate oxidant levels 3C* is the dominant oxidant for all three diols, reacting 8, 13, and 1.2 times faster than  OH with CAT, RES, and HQ, respectively. These results suggest that in areas of biomass combustion, the aqueous-phase oxidation of diols is dominated by 3C* at low pH, with  OH becoming similarly important at

10234 | Phys. Chem. Chem. Phys., 2015, 17, 10227--10237

Fig. 4 Panel A compares the calculated pseudo-first-order rate constants for the aqueous oxidation of benzene-diols by  OH and 3C* at pH 2 (lighter colored, left bar in each pair) and pH 5 (darker bar on right in each pair). Boxes 0 show values of k ArOH at 10 mM phenol based on a likely range of aqueous oxidant concentrations for fog waters in Davis, CA at midday on the winter solstice.  OH concentrations range from 5  1015 (bottom line in  OH boxes) to 7.5  1015 (middle line) to 10  1015 M (top line).11 For excited triplet states we use DMB as a proxy, with initial concentrations of 1, 5, and 10 mM for the bottom, middle, and top lines, corresponding to triplet excited state concentrations of approximately 9  1015, 5  1014, and 9  1014, respectively in our hypothetical fog/cloud drops; these are in the range of the 3C* concentration of 9  1013 M that we recently estimated for a Davis fog water.49 For comparison, direct photodegradation of 10 mM HQ has a rate constant of 1.6  105 s1 (Fig. S2, ESI†). Panel B compares the ratio of the aqueous SOA  0  rate constants from 3C* oxidation k ArOH;3 C  YSOA;3 C and  OH oxidation  0  k ArOH; OH  YSOA;OH . Values were calculated for intermediate levels of 3C* and OH from Panel A, OH rate constants as described in Section S4 (ESI†), and SOA mass yields (YSOA) from Fig. 2. Panel C compares the rates of SOA formation in the aqueous phase (considering reactions with  OH, 3C*, and (for HQ) direct photodegradation), and the gas phase (considering  OH) in a foggy air parcel (L = 107 Laq Lg1, T = 5 1C) assuming Henry’s law partitioning for the diols.

This journal is © the Owner Societies 2015

View Article Online

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

PCCP

higher pH values. Considering intermediate concentrations of  OH and 3C*, the aqueous lifetimes of CAT, RES, and HQ are 3, 3 and 2 hours, respectively, at pH 5 and 2, 2, and 1 hours, respectively, at pH 2. We next investigate the relative importance of 3C* and  OH in making aqueous SOA from the diols. To do this we examine the ratio of ‘‘rate constants’’ for SOA formation, i.e.,  0 . 0  k ArOH; OH  YSOA; OH using secondk ArOH;3 C  YSOA;3 C order rate constants from Table 2 and intermediate aqueous oxidant levels. As shown in Fig. 4B, triplet excited states dominate aqueous SOA production for the diols below pH 4, while 3C* and OH reactions are comparable in less acidic drops. SOA production via 3C* is approximately 7, 27 and 2 times higher than the  OH pathway for CAT, RES, and HQ, respectively, at pH 2; values at pH 5 and above are approximately 0.7, 2, and 1.2, respectively. In the last panel of Fig. 4 we compare the relative importance of aqueous- and gas-phase reactions for SOA production from the diols. In the gas phase we use literature values of the second-order rate constants with  OH (CAT: 1.04  1010 cm3 mlc1 s1; RES: assumed same as CAT; HQ: 4.60  1012 cm3 mlc1 s1),68 assume [ OH(g)] = 1  106 mlc cm3, and assume SOA yields are the same as for guaiacol (YSOA = 0.48),64 another mono-substituted phenol. We use a liquid water content of 1.0  107 Laq Lgas1 and literature Henry’s Law constants (3.6  103, 2.3  107, and 8.3  106 M atm1 for CAT, RES and HQ at 5 1C)31 to determine the gas-aqueous partitioning of the phenols (Section S5, ESI†). Fig. 4C shows that the relative importance of the aqueous- and gas-phase pathways for making SOA varies strongly with diol identity and weakly with pH. For HQ and RES, the rate of SOA formation in the aqueous phase is 100–2000 times higher than the rate in the gas phase, a result of high Henry’s Law constants and fast aqueous reaction rates (Fig. 4A). In contrast, SOA formation from CAT is 40 to 70 times faster in the gas-phase, a result of its significantly smaller Henry’s Law constant. As the final step in assessing the importance of the diol pathways as a source of SOA in a cloudy/foggy atmosphere, we estimate the rates of SOA formation expected during wintertime in the Central Valley of California. For this calculation, we estimate the gas-phase diol concentrations using emission factors reported by Schauer et al.28 and apply these emission factors to the measured organic mass loading during a large biomass burning event in Fresno, CA34 (Section S4, ESI†). We find gas-phase concentrations of the diols to be: 1.44, 0.45, and 0.31 mg m3 for HQ, RES, and CAT, respectively.34 (For reference, the OA and levoglucosan concentrations in this episode were 26.2 and 7.6 mg m3, respectively.34) Putting these concentrations into the fog conditions in Fig. 4B at pH 4 gives a total (gas + aqueous) initial rate of SOA formation from diols of 0.6 mg mair3 h1. (In this calculation we consider only reactions within the drop and not any possible contributions from reactions at the gas–liquid interface.) Of this total rate of SOA production, approximately half is from the aqueous-phase oxidation of RES and HQ (accounting for 21% and 32% of

This journal is © the Owner Societies 2015

Paper

the total diol-derived SOA, respectively), while approximately half (47%) is from the gas-phase oxidation of CAT. Thus, the oxidation of the three diols is overall a modest source of SOA in wood-burning regions, with a significant contribution from aqueous reactions.

4. Conclusions and implications Aqueous benzene-diols react rapidly with both the triplet excited state of an organic chromophore (3DMB*) and  OH, but the two oxidants have the opposite pH behavior, with the triplet kinetics increasing at lower pH. Because of this, in atmospheric aqueous phases impacted by wood combustion, triplet excited states (3C*) appear to dominate benzene-diol oxidation at low pH, but 3C* and  OH are comparable at pH values above 4. Reactions of the benzene-diols with 3DMB* or  OH efficiently produce SOA, with mass yields generally higher as a result of triplet oxidation (72–113%) compared to OH oxidation (59–82%). Using our measured second-order rate constants for the reactions of phenols and benzene-diols with 3 DMB*, we can reasonably predict the reaction kinetics of phenols and diols in mixtures that mimic fog drops in areas of hardwood or softwood burning. Similarly, the SOA mass produced in these mixtures is essentially the sum of the masses predicted from individual phenols. Simple calculations indicate that reactions of benzene-diols can be a significant source of SOA in regions with significant wood combustion. Applying ambient concentration data from a wintertime stagnation event in Fresno, CA28,34 to a hypothetical fog event, we estimate that diol oxidation produces 0.6 mg m3 h1 of SOA, with approximately half from the aqueous oxidation of RES and HQ and half from the gas-phase oxidation of catechol.

Acknowledgements Funding for this research was provided by the National Science Foundation (Grant Number AGS-1036675), the University of California Toxic Substances Research and Teaching Program (TSR&TP) through the Atmospheric Aerosols and Health Lead Campus Program, and the California Agricultural Experiment Station (Project CA-D-LAW-6403-RR).

References 1 Q. Zhang, J. L. Jimenez, M. R. Canagaratna, J. D. Allan, H. Coe, I. Ulbrich, M. R. Alfarra, A. Takami, A. M. Middlebrook, Y. L. Sun, K. Dzepina, E. Dunlea, K. Docherty, P. F. DeCarlo, D. Salcedo, T. Onasch, J. T. Jayne, T. Miyoshi, A. Shimono, S. Hatakeyama, N. Takegawa, Y. Kondo, J. Schneider, F. Drewnick, S. Borrmann, S. Weimer, K. Demerjian, P. Williams, K. Bower, R. Bahreini, L. Cottrell, R. J. Griffin, J. Rautiainen, J. Y. Sun, Y. M. Zhang and D. R. Worsnop, Geophys. Res. Lett., 2007, 34, L13801. 2 M. Hallquist, J. C. Wenger, U. Baltensperger, Y. Rudich, D. Simpson, M. Claeys, J. Dommen, N. M. Donahue,

Phys. Chem. Chem. Phys., 2015, 17, 10227--10237 | 10235

View Article Online

Paper

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

3

4

5 6 7 8 9 10 11

12 13 14 15 16 17

18 19 20 21 22 23 24

C. George, A. H. Goldstein, J. F. Hamilton, H. Herrmann, T. Hoffmann, Y. Iinuma, M. Jang, M. E. Jenkin, J. L. Jimenez, A. Kiendler-Scharr, W. Maenhaut, G. McFiggans, T. F. Mentel, A. Monod, A. S. H. Prevot, J. H. Seinfeld, J. D. Surratt, R. Szmigielski and J. Wildt, Atmos. Chem. Phys., 2009, 9, 5155–5236. R. Volkamer, J. L. Jimenez, F. San Martini, K. Dzepina, Q. Zhang, D. Salcedo, L. T. Molina, D. R. Worsnop and M. Molina, Geophys. Res. Lett., 2006, 33, L17811. C. L. Heald, H. Coe, J. L. Jimenez, R. J. Weber, R. Bahreini, A. Middlebrook, L. M. Russell, M. Jolleys, T.-M. Fu, J. D. Allan, K. N. Bower, G. Capes, J. Crosier, W. T. Morgan, N. H. Robinson, P. I. Williams, M. J. Cubison, P. F. DeCarlo and E. Dunlea, Atmos. Chem. Phys., 2011, 11, 12673–12696. C. He, J. Liu, A. G. Carlton, S. Fran, L. W. Horowitz, H. Levy II and S. Tao, Atmos. Chem. Phys., 2013, 13, 1913–1926. X. Ge, Q. Zhang, Y. Sun, C. R. Ruehl and A. Setyan, Environ. Chem., 2012, 9, 221–235. Y. B. Lim, Y. Tan, M. J. Perri, S. P. Seitzinger and B. J. Turpin, Atmos. Chem. Phys., 2010, 10, 10521–10539. B. Ervens, B. J. Turpin and R. J. Weber, Atmos. Chem. Phys., 2011, 11, 11069–11102. H. Lim, A. G. Carlton and B. J. Turpin, Environ. Sci. Technol., 2005, 39, 4441–4446. D. S. Kaul, T. Gupta, S. N. Tripathi, V. Tare and J. L. Collett Jr., Environ. Sci. Technol., 2011, 45, 7307–7313. T. Arakaki, C. Anastasio, Y. Kuroki, H. Nakajima, K. Okada, D. Handa, S. Azechi, T. Kimura, A. Tsuhako and Y. Miyagi, Environ. Sci. Technol., 2013, 47, 8196–8203. J. D. Blando and B. J. Turpin, Atmos. Environ., 2000, 34, 1623–1632. R. Volkamer, P. J. Ziemann and M. Molina, Atmos. Chem. Phys., 2009, 9, 1907–1928. X. H. Huang, H. S. S. Ip and J. Z. Yu, J. Geophys. Res., 2011, 116, D03206. A. K. Y. Lee, P. Herckes, W. R. Leaitch, A. M. Macdonald and J. P. D. Abbatt, Geophys. Res. Lett., 2011, 38, L11805. J. D. Smith, V. Sio, L. Yu, Q. Zhang and C. Anastasio, Environ. Sci. Technol., 2014, 48, 1049–1057. A. Lee, A. H. Goldstein, J. H. Kroll, N. L. Ng, V. Varutbangkul, R. C. Flagan and J. H. Seinfeld, J. Geophys. Res., 2006, 111, D17305. M. J. Perri, S. P. Seitzinger and B. J. Turpin, Atmos. Environ., 2009, 43, 1487–1497. Y. Tan, M. J. Perri, S. P. Seitzinger and B. J. Turpin, Environ. Sci. Technol., 2009, 43, 8105–8112. H. Herrmann, D. Hoffmann, T. Schaefer, P. Brauer and A. Tilgner, ChemPhysChem, 2010, 11, 3796–3822. C. Anastasio, B. C. Faust and C. J. Rao, Environ. Sci. Technol., 1997, 31, 218–232. S. Canonica, B. Hellrung and J. Wirz, J. Phys. Chem. A, 2000, 104, 1226–1232. Y. L. Sun, Q. Zhang, C. Anastasio and J. Sun, Atmos. Chem. Phys., 2010, 10, 4809–4822. L. Yu, J. D. Smith, A. Laskin, C. Anastasio, J. Laskin and Q. Zhang, Atmos. Chem. Phys. Discuss., 2014, 14, 21149–21187.

10236 | Phys. Chem. Chem. Phys., 2015, 17, 10227--10237

PCCP

¨ller, G. Adler, 25 M. E. Monge, T. Rosenørn, O. Favez, M. Mu A. A. Riziq, Y. Rudich, H. Herrmann, C. George and B. D’Anna, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 6840–6844. ´re 26 S. Rossignol, K. Z. Aregahegn, L. Tinel, L. Fine, B. Nozie and C. George, Environ. Sci. Technol., 2014, 48, 3218–3227. 27 T. C. Bond, D. G. Streets, K. F. Yarber, S. M. Nelson, J. Woo and Z. Klimont, J. Geophys. Res., 2004, 109, D14203. 28 J. J. Schauer, M. J. Kleeman, G. R. Cass and B. R. T. Simoneit, Environ. Sci. Technol., 2001, 35, 1716–1728. 29 S. B. Hawthorne, M. S. Krieger, D. J. Miller and M. B. Mathiason, Environ. Sci. Technol., 1989, 23, 470–475. 30 D. R. Oros and B. R. T. Simoneit, Appl. Geochem., 2001, 16, 1545–1565. 31 R. Sander, http://www.henrys-law.org/, 1999. 32 C. G. Nolte, J. J. Schauer, G. R. Cass and B. R. T. Simoneit, Environ. Sci. Technol., 2001, 35, 1912–1919. 33 W. F. Rogge, L. M. Hildemann, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit, Environ. Sci. Technol., 1998, 32, 13–22. 34 J. J. Schauer and G. R. Cass, Environ. Sci. Technol., 2000, 34, 1821–1832. 35 C. Anastasio and K. G. McGregor, Atmos. Environ., 2001, 35, 1079–1089. 36 L. Chu and C. Anastasio, J. Phys. Chem. A, 2005, 105, 6264–6271. 37 W. L. Miller and D. R. Kester, Anal. Chem., 1988, 60, 2711–2715. 38 E. S. Galbavy, K. Ram and C. Anastasio, J. Photochem. Photobiol., A, 2010, 209, 186–192. 39 N. Borduas, J. P. D. Abbatt and J. G. Murphy, Environ. Sci. Technol., 2013, 47, 6377–6383. 40 B. J. Finlayson-Pitts and J. N. Pitts, Chemistry of the Upper and Lower Atmosphere: Theory, Experiments and Applications, Academic Press, San Diego, 2000. 41 A. Leifer, The Kinetics of Environmental Aquatic Photochemistry: Theory and practice, American Chemical Society, Washington, DC, 1988. 42 Q. Zhang and C. Anastasio, Environ. Sci. Technol., 2003, 37, 3522–3530. 43 C. M. Sharpless, Environ. Sci. Technol., 2012, 46, 3912–3920. 44 R. G. Zepp, P. F. Schlotzhauer and R. M. Sink, Environ. Sci. Technol., 1985, 19, 74–81. 45 S. Halladja, A. Ter Halle, J.-P. Aguer, A. Boulkamh and C. Richard, Environ. Sci. Technol., 2007, 41, 6066–6073. 46 G. F. Vesley, J. Phys. Chem., 1971, 75, 1775–1781. 47 E. De Laurentiis, B. Sur, M. Pazzi, V. Maurino, C. Minero, G. Mailhot, M. Brigante and D. Vione, Atmos. Environ., 2013, 70, 318–327. 48 K. S. Golanoski, S. Fang, R. Del Vecchio and N. V. Blough, Environ. Sci. Technol., 2012, 46, 3912–3920. 49 N. K. Richards-Henderson, A. T. Pham, B. B. Kirk and C. Anastasio, Environ. Sci. Technol., 2015, DOI: 10.1021/ es503656m. 50 A. Monod, L. Poulain, S. Grubert, D. Voisin and H. Wortham, Atmos. Environ., 2005, 39, 7667–7688. 51 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross, J. Phys. Chem. Ref. Data, 1988, 17, 513–886.

This journal is © the Owner Societies 2015

View Article Online

Published on 16 March 2015. Downloaded by Ryerson Polytechnic University on 14/09/2015 20:01:58.

PCCP

52 E. J. Land and M. Ebert, Trans. Faraday Soc., 1967, 63, 1181–1190. 53 L. Ashton, G. V. Buxton and C. R. Stuart, J. Chem. Soc., Faraday Trans., 1995, 91, 1631–1633. 54 A. A. Heath, F. S. Ehrenhauser and K. T. Valsaraj, J. Environ. Chem. Eng., 2013, 1, 822–830. 55 USEPA, Estimation Programs Interface Suite for Microsoft Windows V. 4.1, United States Environmental Protection Agency, Washington, DC, 2014. 56 Y. J. Li, D. Huang, H. Y. Cheung, A. K. Y. Lee and C. K. Chan, Atmos. Chem. Phys. Discuss., 2014, 14, 2871–2885. 57 C. Coeur-Tourneur, V. Foulon and M. Lareal, Atmos. Environ., 2010, 44, 852–857. 58 C. Coeur-Tourneur, A. Tomas, A. Guilloteau, F. Henry, F. Ledoux, N. Visez, V. Riffault, J. C. Wenger and Y. Bedjanian, Atmos. Environ., 2009, 43, 2360–2365. 59 J. Ofner, H. U. Kruger, H. Grothe, P. Schmitt-Kopplin, K. Whitmore and C. Zetzsch, Atmos. Chem. Phys., 2011, 11, 1–15. 60 S. Nakao, C. Clark, P. Tang, K. Sato and D. Cocker III, Atmos. Chem. Phys., 2011, 11, 10649–10660.

This journal is © the Owner Societies 2015

Paper

61 A. Lauraguais, C. Coeur-Tourneur, A. Cassez, K. Deboudt, ¨l, Atmos. Environ., 2014, 86, M. Fourmentin and M. Choe 155–163. 62 A. Lauraguais, C. Coeur-Tourneur, A. Cassez and A. Seydi, Atmos. Environ., 2012, 55, 43–48. 63 P. S. Chhabra, N. L. Ng, M. R. Canagaratna, A. L. Corrigan, L. M. Russell, D. R. Worsnop, R. C. Flagan and J. H. Seinfeld, Atmos. Chem. Phys., 2011, 11, 8827–8845. 64 L. D. Yee, K. E. Kautzman, C. L. Loza, K. A. Schilling, M. M. Coggon, P. S. Chhabra, M. N. Chan, A. W. H. Chan, S. P. Hersey, J. D. Crounse, P. O. Wennberg, R. C. Flagan and J. H. Seinfeld, Atmos. Chem. Phys., 2013, 13, 3485–3532. 65 C. Anastasio and J. Sun, 2015, to be submitted for publication. 66 J. Dec, K. Haider and J.-M. Bollag, Soil Sci., 2001, 166, 660–671. 67 R. Pollard, S. Wu, G. Zhang and P. Wan, J. Org. Chem., 1993, 58, 2605–2613. 68 R. Olariu, I. Barnes, K. H. Becker and B. Klotz, Int. J. Chem. Kinet., 2000, 32, 696–702. 69 H. Herrmann, Chem. Rev., 2003, 103, 4691–4716.

Phys. Chem. Chem. Phys., 2015, 17, 10227--10237 | 10237