Article pubs.acs.org/est
A General Scavenging Rate Constant for Reaction of Hydroxyl Radical with Organic Carbon in Atmospheric Waters Takemitsu Arakaki,*,†,⊥ Cort Anastasio,*,‡,⊥ Yukiko Kuroki,§ Hitomi Nakajima,§ Kouichirou Okada,§,∥ Yuji Kotani,§ Daishi Handa,§ Sotaro Azechi,§ Taro Kimura,† Ai Tsuhako,§ and Youichi Miyagi§ †
Department of Chemistry, Biology and Marine Science, Faculty of Science, and §Graduate School of Engineering and Science, University of the Ryukyus, 1 Senbaru Nishihara-cho, Okinawa 903-0213, Japan ‡ Department of Land, Air and Water Resources, University of California at Davis, One Shields Avenue, Davis, California 95616, United States ∥ Okinawa Environmental Research & Technology Center, 720 Kyozuka, Urasoe-city, Okinawa, 901-2111, Japan S Supporting Information *
ABSTRACT: Hydroxyl radical (OH) is an important oxidant in atmospheric aqueous phases such as cloud and fog drops and water-containing aerosol particles. We find that numerical models nearly always overestimate aqueous hydroxyl radical concentrations because they overpredict its rate of formation and, more significantly, underpredict its sinks. To address this latter point, we examined OH sinks in atmospheric drops and aqueous particles using both new samples and an analysis of published data. Although the molecular composition of organic carbon, the dominant sink of OH, is extremely complex and poorly constrained, this sink behaves very similarly in different atmospheric waters and even in surface waters. Thus, the sink for aqueous OH can be estimated as the concentration of dissolved organic carbon multiplied by a general scavenging rate constant [kC,OH = (3.8 ± 1.9) × 108 L (mol C)−1 s−1], a simple process that should significantly improve estimates of OH concentrations in atmospheric drops and aqueous particles.
■
INTRODUCTION The fate of many trace gases in the atmosphere is controlled by reactions with hydroxyl radical (OH), which is the most potent environmental oxidant.1,2 Hydroxyl radical is also an important oxidant in atmospheric water drops and aqueous particles, where it reacts with both organic and inorganic species. For example, the OH-mediated oxidation of aqueous glyoxal, glyoxylic acid, and other small, multifunctional organic compounds produces low-volatility, secondary organic aerosol species such as oxalic acid and oligomers.3,4 The lifetime of an aqueous species “S” with respect to reaction with hydroxyl radical is inversely proportional to the steady-state concentration of OH τS =
1 k S,OH[OH]
where POH is the rate of production of OH and k′OH is the apparent first-order rate constant for loss of OH (also called the “OH reactivity”), i.e., the inverse of the OH lifetime. Since there are many scavengers for OH, the rate constant for OH loss is the sum of the individual scavenger contributions, i.e., the product of the bimolecular rate constant and the scavenger concentration for each species S: k′OH =
(1)
POH k′OH
Received: Revised: Accepted: Published:
(2) © 2013 American Chemical Society
(3)
Accurately calculating k′OH in this way is difficult because the major scavengers of OH in atmospheric drops and particles are dissolved organic compounds,5,6 which are incredibly diverse, complex, and poorly characterized. Here we show that numerical models of atmospheric aqueous chemistry generally underestimate k′OH, which contributes to overestimating [OH]. To improve this situation we measure the OH sink in a series of marine particle extracts from Japan and compare the results to previously published data on k′OH in rain, cloud, and fog waters. We find that the
where kS,OH is the bimolecular rate constant between S and OH. Thus, the steady-state concentration of OH is a crucial quantity for understanding the fates of atmospheric pollutants. This concentration is determined by the balance of the OH sources and sinks,
[OH] =
∑ kS,OH[S]
8196
April 30, 2013 June 28, 2013 July 3, 2013 July 3, 2013 dx.doi.org/10.1021/es401927b | Environ. Sci. Technol. 2013, 47, 8196−8203
Environmental Science & Technology
Article
OH sink can be simply estimated by a general carbon rate constant that is applicable both for atmospheric waters as well as surface waters, which allows k′OH to be estimated by using readily measured (or estimated) organic carbon concentrations.
(eq 3) and kBA,OH is the second-order rate constant for BA with OH. We calculated values of kBA,OH for the pH of each particle extract using
MATERIALS AND METHODS Reagents and Aerosol Samples. All chemicals used were reagent grade or the highest purity available and were purchased from Kanto Chemical Co. Milli-Q water (≥18.2 MΩ cm; Millipore) was used in all experiments. Bulk marine aerosol samples (n = 61) were collected at the Cape Hedo Aerosol and Atmosphere Monitoring Station (CHAAMS: 128°14′E, 26°52′N) for 17 months (Aug 2005 to Dec 2006) using a high-volume air sampler (0.5 m3 min−1; HV-500F, SIBATA) with 110-mm diameter quartz filters (Model 2500 QAT-UP, Pallflex). Each sample was collected for 1 week. Aqueous extracts of the aerosols were prepared by shaking a quarter of the sampled filter in 300 mL of Milli-Q water for 3 h on a rotary shaker (100 rounds min−1; RS-2, AS ONE) and filtering (0.45 μm, JHWPO4700, Millipore).7 Extracts were analyzed for anions (Cl−, SO42−, NO3−, NO2−, Br−) by ion chromatography (D-7000 and L-7470 conductivity detector; Hitachi and ERIS 1000 HP autosuppressor, Alltech), cations (Na+, Ca2+, Mg2+) by atomic absorption spectrometry (SOLAAR 969 AA; Thermo Elemental), and dissolved organic carbon (DOC) by a total organic carbon analyzer (TOC-VCSH; Shimadzu). Determination of OH Scavenging Rate Constant, k′OH. The scavenging rate constant, k′OH, of each CHAAMS marine particle extract was determined by using competition kinetics5,8−10 with 55 μM H2O2 added as an additional source of OH; the contribution of the added H2O2 to k′OH (i.e., the scavenging of OH by H2O2) was subtracted from each result (mean contribution = 3.1%). In this procedure, different concentrations of benzoate/benzoic acid (collectively referred to as “BA”) were added to several aliquots of the same aerosol extract, and each solution was illuminated with 313 nm light from a monochromatic illumination system (CT-25T; JASCO) with a 500 W super-high-pressure mercury lamp (Ushio Optical Modulex H500; Ushio Inc.).7 Photoformed OH reacts with both the natural scavengers (S) initially present in the sample and the added BA probe:
where αBAH and αBA− are the mole fractions of benzoic acid (BAH) and benzoate (BA−), respectively, and kBAH,OH (1.8 × 109 M−1 s−1) and kBA−,OH (6.0 × 109 M−1 s−1) are the bimolecular rate constants for OH with benzoic acid and benzoate, respectively.11,12 Taking the inverse of the equation made by combining eqs 6 and 7 yields
■
OH + natural scavengers (S) → products
(4)
OH + BA → o‐hydroxybenzoate + other products
(5)
kBA,OH = αBAHkBAH,OH + αBA−k BA−,OH
1/Po‐HBA =
slope =
k′OH 1 1 POH Yo‐HBA kBA,OH
y‐intercept =
1 1 POH Yo‐HBA
■
RESULTS AND DISCUSSION Comparing Measured and Modeled OH Kinetics and Concentrations. We first compare the relatively few measurements of OH kinetics in atmospheric drops and aqueous particles with model estimates, starting with the production rate of OH. The major sources of OH in atmospheric waters include the reaction between Fe(II) and H2O2; photolysis of nitrate, nitrite, and H2O2; mass transport of OH from the gas phase; and, in less acidic drops, reaction of ozone (O3) with superoxide (O2−).14,15 As shown by the bars in Figure 1a, measured rates of OH production in atmospheric waters span approximately 4 orders of magnitude: POH is lowest in rain (where aqueous sources of OH are most dilute and mass transport from the gas phase is relatively slow); intermediate in dew, cloud, and fog drops; and highest in aqueous particles (where some of the aqueous OH sources are highly concentrated and mass transport is rapid). As described in Table S1 Supporting Information, each measured rate of OH production includes both the formation rate measured in the natural samples in the laboratory as well as the best model estimates of the rates of mass transport of OH(g) to the drops and of the aqueous reaction of ozone with superoxide. For the cloud and fog drops, mass transport and aqueous chemistry are approximately equal sources of aqueous OH, while aqueous chemistry is the dominant source for other sample types (Table S1, Supporting Information). The open circles in Figure 1a are estimates of OH production rates from numerical models; these data are available only for the cloud and fog cases (Table S2, Supporting Information). Models generally overestimate POH in cloud drops, by average factors of 6 for remote continental clouds and 5 for marine clouds. There is better agreementto within a factor of 2between the one study that measured OH in fog waters9 and a modeling study under comparable conditions.16 In some cases modeled values of POH are too high because of overly rapid mass transport of OH(g) (for example, in models that assume a monodispersed cloud with a small drop radius) and/or because the O3 + O2− reaction appears to be overestimated. However, in many cases the
(6)
kBA,OH[BA] k′OH + kBA,OH[BA]
(9)
Thus k′OH is determined by dividing the slope from this plot by the y-intercept and multiplying by kBA,OH. On the basis of our recent work, the average relative standard deviation (RSD) for k′OH from this procedure is approximately 10%.13
where POH is the formation rate of OH in the illuminated sample, FBA,OH is the fraction of OH that reacts with the BA probe, and Yo‑HBA is the molar yield of o-HBA from the reaction of BA with OH.7 The fraction of OH that reacted with the probe in a given sample aliquot is a function of the BA concentration FBA,OH =
k′OH 1 1 1 1 1 + POH Yo − HBA kBA,OH [BA] POH Yo‐HBA
Plotting 1/Po‑HBA versus 1/[BA] yields a straight line defined by
The rate of formation of o-hydroxybenzoate (Po‑HBA), which was measured using HPLC, is related to the rate of OH production by Po‐HBA = POHFBA,OHYo‐HBA
(8)
(7)
where k′OH (=∑kS,OH[S]) is the overall OH scavenging rate constant of the aerosol extract in the absence of added benzoate 8197
dx.doi.org/10.1021/es401927b | Environ. Sci. Technol. 2013, 47, 8196−8203
Environmental Science & Technology
Article
case of fog waters, the model value is 10 times lower than the average measured value. This underprediction of k′OH by models is not surprising, since they use a molecular, bottom-up approach to determine k′OH, i.e., by tracking the concentrations of individual organic compounds in the drop and using eq 3. Determining the rate constant for OH loss in this manner is a herculean task since atmospheric drops and particles can contain on the order of 104 individual organic compounds, as well as significant concentrations of poorly characterized, large molecular weight species such as humic-like substances.19,20 In contrast, the most sophisticated numerical models of atmospheric waters track the reactions of fewer than 100 individual organic compounds, with none larger than four carbon atoms (e.g., ref 21). The discrepancy we find between modeled and measured values of k′OH in atmospheric waters is analogous to the gas phase of forested environments, where measured values of k′OH are typically larger than calculated, bottom-up, values by factors of 2−10 because of uncharacterized organic sinks for OH.17,18 Dividing the OH production rate by the rate constant for OH loss gives the steady-state concentration for hydroxyl radical (eq 2). While measured values of POH and k′OH each span a range of nearly 4 orders of magnitude for different atmospheric water types, they tend to move in concert (Figure 1a,b). Because of this correlation, measured OH concentrations generally fall within a relatively narrow range across all sample types, with average values of (0.5 − 7) × 10−15 M (Figure 1c). This figure also shows that modeled OH concentrations tend to be much higher than measured values, on average by factors of 70 for remote continental clouds, 1000 for marine clouds, and 7 for California fog drops. (In contrast, for deliquesced sea-salt particles the single model prediction is similar to the measured values.) Since the lifetime of an organic compound is inversely proportional to the OH concentration (eq 1), the general disagreement in [OH] between measurements and models implies that models typically overestimate the significance of hydroxyl radical as a sink for individual aqueous species. While there are discrepancies between measurements and models for both the OH production rate and the OH sink, our focus in the rest of this paper is on k′OH. We have focused on the OH sink because it generally shows a larger measured− modeled discrepancy than POH and, as we describe below, we have found a way to significantly improve its estimates. Determining a Carbon-Based Rate Constant for OH Loss, kC,OH. To address the large uncertainties in aqueous OH sinks, we start by examining the relationship between k′OH and dissolved organic carbon (DOC) concentrations in atmospheric waters. Because only three studies,5,9,22 representing a total of 20 samples, report both k′OH and DOC in atmospheric samples, we first made measurements of these quantities in water extracts of 61 marine aerosol samples collected over a period of 17 months at Okinawa, Japan. As shown in Table 1, values of k′OH in the Okinawan particle extracts were in the range of (340−33 000) × 105 s−1, with faster OH loss in spring and summer. This pattern in k′OH follows the seasonal variation in concentrations of DOC, which was the main OH sink, accounting for an average of 88% of OH loss. (The main inorganic sinks for OH were nitrite and bromide (Table S3, Supporting Information); we did not measure aqueous S(IV) but expect that this was a minor sink for OH because the particles were acidic and thus S(IV) concentrations should be low.) Despite the seasonal variations in chemistry, the ratio between k′OH and DOC concentrations was relatively constant,
Figure 1. Measured and modeled kinetic parameters for hydroxyl radical in atmospheric drops and particles from midlatitudes of the northern hemisphere. Each bar represents the average measured value (±1σ) from a single study, while circles represent modeled values for conditions similar to the experimental measurements. Panels a−c show the rate of OH formation (POH), the rate constant for OH loss (k′OH), and the steady-state concentration of OH ([OH]), respectively. Measured values for POH include aqueous-phase photoreactions and our estimates for mass transport of gas-phase OH to the condensed phase. Measured values of k′OH shown here include all OH sinks, i.e., both organic and inorganic species. Key to samples: a and b = rain waters from Italy (spring)22 and Japan (winter and spring),36 respectively; c = dew waters from Japan (fall and winter);36 d = remote continental clouds (summer);5,14,15,21,37−44 e = remote marine clouds (summer);3,9,14,38,41,42,45 f = polluted/urban clouds (summer);14,21,38,40−42,46 g = fog waters from California (winter);9,16 h, i, and j = marine particles from the eastern Pacific Ocean (mostly summer), Atlantic Ocean (spring and fall), and western Pacific Ocean (fall and winter; this study), respectively.3,6,47 Details on the measurement and modeling studies in this figure are in the Supporting Information.
model reports do not provide sufficient detail to allow an assessment of why the modeled POH is greater than measured values. Measured apparent first-order rate constants for loss of OH (i.e., the inverse of the OH lifetime) show a similar trend to POH: values of k′OH are lowest in rain drops; intermediate in dew, cloud, and fog drops; and highest in aqueous particles (Figure 1b). These rate constants correspond to very short OH lifetimes, ranging from approximately 10−9 s in the particles to 10−5 s in rain drops; in contrast, the OH lifetime in the gas phase is much longer, generally 0.01−1 s (refs 17, 18). Compared to their ability to predict POH, models generally do a poorer job of predicting k′OH in atmospheric drops, underestimating measured values on average by factors from 5 (for remote continental clouds) to 500 (for marine clouds). In the 8198
dx.doi.org/10.1021/es401927b | Environ. Sci. Technol. 2013, 47, 8196−8203
Environmental Science & Technology
Article
Table 1. Measured Hydroxyl Radical Sinks, and the Calculated Mole-Carbon-Based Scavenging Rate Constants, for Atmospheric and Terrestrial Waters and Aquatic Humic Substancesa k′OHb (105 s−1) sample type (location, number of samples)
range
k′OH/[DOC] [108 L (mol C)−1 s−1]
[DOC] [(μmol C) L−1]
mean ± σ
range
range
mean ± σ
ref
Atmospheric Waters aerosol extractc (Japan, n = 61) rain waters (Italy, n = 6) cloud waters (USA, n = 7) fog waters (USA, n = 7)
340−33 000
3300 ± 4700
220 000−3 400 000
0.30−12
3.0 ± 2.2
this study
0.053−0.24
0.15 ± 0.062
35−198
0.64−6.9
2.4 ± 2.4
22
0.88−4.1
2.5 ± 1.2
200−2230
1.8−4.4
3.1 ± 0.98
5
6.1−17
12 ± 5.0
680−4100
1.8−11
6.6 ± 3.6
9
3.8 ± 1.9
mean (±1σ) Terrestrial Waters lake waters (Italy, n = 7) (USA, n = 3) (USA, n = 2) (Switz., n = 1) (Switz. & Norway, n = 5) reservoir waters (USA, n = 5) river waters (USA, n = 2) (USA, n = 2) ground waters (Italy, n = 2)
0.30−9.4 0.61−1.12 1.8−7.1 1 0.35−0.65
2.7 ± 3.2 0.93 ± 0.27 4.5 ± 3.7 1 0.53 ± 0.16
33−430 200−620 470−580 300 110−260
1.8−24 1.7−3.1 3.9−12 3 2.4−3.2
11 ± 8.4 2.3 ± 0.73 8.1 ± 5.5 3 2.7 ± 0.46
10 26 28 8 27
0.45−1.4
0.66 ± 0.40
120−310
2.3−4.4
3.2 ± 1.0
28
1.2−2.3 0.86−1.5
1.7 ± 0.80 1.2 ± 0.47
349−581 280−400
3.4−4.0 3.1−3.8
3.7 ± 0.42 3.5 ± 0.47
26 28
2.5−5.7
4.1 ± 2.3
276−462
9.0−12
11 ± 2.3
10
5.4 ± 3.6
mean (±1σ) Aquatic Humic Substances Fulvic acids (Suwannee River) (Suwannee River) (Suwannee River) (Pony Lake) Humic acids (Suwannee River) (Aldrich)
− − − −
− − − −
− − − −
1.39−1.87 − − −
1.6 ± 0.24 3.24 ± 0.06 2.06 ± 0.09 6.9 ± 0.53
24 48 49 49
− −
− −
− −
− −
2.28 ± 0.06 4.8
48 28
3.5 ± 2.0
mean (±1σ) a
b
Dashes indicate data that are not available. Listed values of k′OH are the measured or reported values minus the contributions from inorganic species (e.g., nitrite and bicarbonate), which were calculated by multiplying the bimolecular rate constant and the concentration of the inorganic species, in order to isolate the organic carbon sink of OH. Organic compounds were the dominant sinks for OH in all of these studies, although nitrite was a major sink in rain (accounting for an average of 48% of k′OH) and bicarbonate/carbonate was significant in the terrestrial waters. In the terrestrial water samples the bicarbonate/carbonate sink for OH was removed when k′OH was reported in the original studies. cMarine particles were collected and extracted into water and then illuminated to measure k′OH (see the Materials and Methods section). The measured kinetic data (and DOC concentrations) for the aqueous extracts were normalized to aqueous sea-salt particle conditions at RH = 88% (i.e., an aerosol sodium concentration of 3.1 M) as described by Anastasio and Newberg.6
with a mean (±1σ) value of (3.0 ± 2.2) × 108 L (mol C)−1 s−1 (Table 1). Surprisingly, the three previously published studies on OH sinks in rain, cloud, and fog waters have very similar average values for the ratio k′OH/[DOC], despite their very different carbon concentrations and locations (Figure 2). This similarity indicates that OH scavenging rate constants in different atmospheric waters are primarily controlled by the organic carbon concentration and are relatively insensitive to differences in the complex mixtures of organic compounds in different samples. Thus, OH scavenging rate constants can be
estimated as the product of the organic carbon concentration [in units of (mol C) L−1] and a general bimolecular rate constant between OH and DOC, denoted as kC,OH [in units of L (mol C)−1 s−1] k′OH ≈ [DOC]k C,OH
(10)
The average (±1σ) value for kC,OH determined from the four sets of atmospheric samples in Figure 2 is (3.8 ± 1.9) × 108 L (mol C)−1 s−1 (Table 1). While the idea of a general scavenging rate constant for OH by organic compounds is new for atmospheric waters, this 8199
dx.doi.org/10.1021/es401927b | Environ. Sci. Technol. 2013, 47, 8196−8203
Environmental Science & Technology
Article
Figure 2. Relationship between the dissolved organic carbon (DOC) concentration-normalized rate constant for OH loss [L (mol C)−1 s−1] and DOC concentrations [(mol C) L−1] for atmospheric waters: marine particles collected in Okinawa, Japan (red circles, this work), and calculated values for previously published studies of rain (blue circles),22 cloud (green triangles),5 and fog (yellow squares).9 Results for the Japanese marine particles are normalized to a relative humidity of 88% (i.e., a particle sodium concentration of 3.1 M). Values of k′OH have been adjusted to remove contributions from inorganic sinks for OH; see Table 1. The average value (±1σ) of k′OH/[DOC] from each study is shown by the black diamonds, while the overall average, kC,OH = (3.8 ± 1.9) × 108 L (mol C)−1 s−1, is shown by the horizontal line. Figure 3. Bimolecular rate constants for OH with individual organic compounds (n = 1415) as a function of carbon number. Panel a shows the rate constants expressed in terms of moles of compound,11,12 while values in panel b are expressed in terms of moles of carbon in each compound. Open circles represent values for individual compounds, while red triangles are the average value for a given carbon number; horizontal lines show the overall averages for each plot determined as the grand average of the individual points.
concept has been around for decades in surface water and wastewater studies. For example, the general rate constant for DOC with OH is typically in the range of (2−14) × 108 L (mol C)−1 s−1 for organic scavengers in surface waters and effluents from wastewater treatment plants, with generally higher rate constants for OC fractions of higher polarity or lower molecular weight.23−25 To test the generality of the carbon rate constant for OH loss in atmospheric waters, we compare it to past results from natural terrestrial waters such as lakes.8,10,24,26−28 As shown in Table 1, values of kC,OH for terrestrial waters are generally very similar to those for atmospheric waters, with an average value of (5.4 ± 3.6) × 108 L (mol C)−1 s−1, which is statistically no different from the atmospheric water average (p = 0.32). This is additional evidence that very different aqueous reservoirs of carbon in the natural environment have similar reactivities with hydroxyl radical. Values of kS,OH for Individual Compounds and Humic Substances. To further examine the robustness of the OH− DOC rate constant for different conditions, we also analyzed more than 1400 bimolecular rate constants reported for the reaction of OH with a wide range of individual organic compounds.11,12 Figure 3a shows the relationship between the bimolecular rate constants of organic compounds with OH [kS,OH, in units of L (mol compound)−1 s−1] and the number of carbon atoms in each compound. For small organic molecules, values of kS,OH at a given carbon number vary over 3 orders of magnitude, but as the carbon number increases kS,OH reaches a well-known, diffusion-controlled rate constant of approximately 1010 L (mol compound)−1 s−1. Values of kS,OH in Figure 3a range from 7.9 × 105 (for urea) to 8.5 × 1010 (for apothionein) L (mol compound)−1 s−1 (refs 11, 12), with an overall average of (6.4 ± 7.0) × 109 L (mol compound)−1 s−1. As illustrated in Figure 3b, this same data set can also be examined with rate constants expressed using carbon (rather than molecular) concentrations, i.e., in the L (mol C)−1 s−1 units used above for atmospheric and terrestrial waters. As with the molecule-based bimolecular rate constants (Figure 3a), the carbon-based rate constants still cover a wide range of values
[from 106 to 1010 L (mol C)−1 s−1], but now the average value for compounds of a given carbon number is similar across all carbon numbers (Figure 3b). The average value for kC,OH from all of the individual compounds in Figure 3b is (9.6 ± 12.6) × 108 L (mol C)−1 s−1, which is 2.5 times greater than the average value for atmospheric waters but within the range of individual values (Figure 1). The higher average value for the database bimolecular rate constants indicates that the compounds included in the database are, on average, more reactive than the typical mixture of organic species in atmospheric waters. To explore this further, we calculated the OH reactivity of the mixtures of individual organic compounds reported in five studies of atmospheric drops and particles. The most abundant compounds reported in these studies include small monoand dicarboxylic acid anions (e.g., formate, acetate, oxalate, and lactate) and aldehydes (e.g., aquated formaldehyde and acetaldehyde); altogether, the identified compounds account for 14−63% of DOC.29−33 Carbon-based rate constants for the identified organics range from approximately 107 L (mol C)−1 s−1 (e.g., oxalate) to 109 L (mol C)−1 s−1 (e.g., formaldehyde and formate).11,12 For each study we used these rate constants, along with the reported abundance of each individual organic species, to calculate the value of kC,OH for the mixture of identified organics. Values of kC,OH for the five different studies are in the range of (4−13) × 108 L (mol C)−1 s−1, with a mean (±1σ) of (7.7 ± 3.4) × 108 L (mol C)−1 s−1 (see Table S4, Supporting Information). Formate is generally the dominant OH sink among the identified organics, accounting for an average of nearly 60% of the calculated kC,OH. The average value of kC,OH from the mixtures of identified organics is very similar 8200
dx.doi.org/10.1021/es401927b | Environ. Sci. Technol. 2013, 47, 8196−8203
Environmental Science & Technology
Article
hydroxyl radical such as halogen radicals (e.g., Br2−, which is initiated from the reaction of OH with bromide) and sulfate radical (from the OH oxidation of sulfate) The situation we find for aqueous OHwhere models underestimate the OH sink and overestimate its steady-state concentrationlikely also applies to other aqueous oxidants whose major sinks are organic compounds. This group of oxidants probably includes nitrate radical (NO3) and halogen radicals (e.g., Br and Br2−): if models underestimate the sinks for these oxidants, they will correspondingly overestimate the steady-state concentrations of the oxidants and their impacts on the transformation of specific compounds. It is possible that there are also general rate constants that describe reactions of these oxidants with atmospheric organic compounds, but this remains to be explored.
to the average value from the 1400 individual compounds in Figure 3. It is also within the range of values for individual atmospheric samples (Figure 1), but it is 2.0 times higher than the average atmospheric value. This latter comparison suggests that the identified organic species in fog and cloud drops are, overall, more reactive than the unidentified organics. Humic-like substances (HULIS) probably represent a significant portion of the unidentified dissolved organic carbon in the fog and particle studies described above. HULIS are large, oligomeric organic compounds that can contribute up to approximately half of the water-soluble organic carbon in atmospheric particles.34,35 To estimate the HULIS sink for OH, we compiled reported values of kC,OH for aquatic humic and fulvic acids, which are structurally similar to HULIS20 and thus likely have similar reactivities with OH. As shown in Table 1, the average value for kC,OH for humic and fulvic acids from surface waters is (3.5 ± 2.0) × 108 L (mol C)−1 s−1. This average is only 10% lower than the average value for atmospheric waters and is 2.4 times lower than the average value of kC,OH calculated from the identified organic compound mixtures in past studies. The abundance of HULIS in atmospheric condensed phases and the fact that the OH reactivity of humics and fulvics is similar to the reactivity of bulk DOC in atmospheric waters suggest that HULIS can contribute significantly to the OH sink in atmospheric waters and aqueous particles. However, since the reactivities of the identified organic compounds in the five studies cited above (Table S4, Supporting Information) are greater than the average reactivity of bulk DOC in atmospheric waters (Table 1), this suggests that atmospheric drops and particles contain additional, unidentified organic compounds other than HULIS that are less efficient OH sinks. Implications. Our results show that the scavenging rate constant of OH by organic species in atmospheric waters can be simply estimated as the product of a robust general rate constant (kC,OH) multiplied by the dissolved organic carbon concentration of the sample. In cases where aqueous concentrations of reactive inorganic OH sinks (e.g., nitrite) are high, OH scavenging by these species should be added to the dominant organic term to get a more accurate estimate. The use of kC,OH greatly simplifies estimating the scavenging rate constant of OH (k′OH) in atmospheric drops and aqueous particles, since it requires only measuring (or estimating) the DOC concentration. This procedure should significantly improve values of k′OH in atmospheric models, which typically underestimate the sinks for OH and, therefore, overestimate the steady-state concentration of OH (Figure 1). Improving (i.e., increasing) the sinks for OH in models of atmospheric drop and particle chemistry will decrease the OH steady-state concentration and, therefore, reduce the importance of hydroxyl radical in the processing of individual organic compounds. For organic compounds where OH is a major sink, reducing [OH] in the model will lower the oxidation rate of the compound and increase its lifetime. In turn, this will decrease the modeled formation rates of specific products, such as oxalate and other SOA species from the OH-mediated oxidation of glyoxal and multifunctional organics. Similarly, OH-mediated rates of SOA transformation (including mineralization to CO2) in models of atmospheric waters will slow with decreasing [OH], leading to a less vigorous cycling of carbon in cases where OH is a major oxidant. Decreases in modeled OH concentrations will also decrease the steady-state concentrations of secondary radicals that are formed via
■
ASSOCIATED CONTENT
S Supporting Information *
Data for measured and modeled OH kinetics in hydrometeors, details of aerosol extract compositions and chemistry, and calculations of k′OH for fog waters with detailed organic speciation from the literature. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*T.A.: e-mail,
[email protected]; tel, +81-98-8958553; fax, +81-98-895-8565. C.A.: e-mail, canastasio@ucdavis. edu; tel, 530-754-6095; fax, 530-752-1552. Author Contributions ⊥
These authors contributed equally to this work.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank S. Hatakeyama of Tokyo University of Agriculture and Technology and A. Takami of NIES, Japan, for the use of CHAAMS; A. Tanahara and Y. Nakama of University of the Ryukyus, for assistance in collecting aerosol samples; H. Wang of the University of Maryland University College in Okinawa, for valuable comments on the manuscript; and three anonymous reviewers, for their thought-provoking comments. T.A. acknowledges funding from a Grant-in-Aid for Scientific Research (KAKENHI, #20310013) and for Scientific Research on Innovative Areas (No. 4003) of JSPS and MEXT, Japan. C.A. acknowledges funding from the U.S. National Science Foundation (Grant Numbers AGS-1036675 and AGS1105049) and the California Agricultural Experiment Station (Project CA-D*-LAW-6403-RR).
■
REFERENCES
(1) Thompson, A. M. The oxidizing capacity of the Earth’s atmosphereProbable past and future changes. Science 1992, 256 (5060), 1157−1165. (2) Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments, and Applications; Academic Press: San Diego, 2000; p 969. (3) Warneck, P. Multi-phase chemistry of C-2 and C-3 organic compounds in the marine atmosphere. J. Atmos. Chem. 2005, 51 (2), 119−159. (4) Lim, Y. B.; Tan, Y.; Perri, M. J.; Seitzinger, S. P.; Turpin, B. J. Aqueous chemistry and its role in secondary organic aerosol (SOA) formation. Atmos. Chem. Phys. 2010, 10 (21), 10521−10539.
8201
dx.doi.org/10.1021/es401927b | Environ. Sci. Technol. 2013, 47, 8196−8203
Environmental Science & Technology
Article
(5) Arakaki, T.; Faust, B. C. Sources, sinks, and mechanisms of hydroxyl radical (OH) photoproduction and consumption in authentic acidic continental cloud waters from Whiteface Mountain, New York: The role of the Fe(r) (r = II, III) photochemical cycle. J. Geophys. Res. Atmos. 1998, 103 (D3), 3487−3504. (6) Anastasio, C.; Newberg, J. T. Sources and sinks of hydroxyl radical in sea-salt particles. J. Geophys. Res. Atmos. 2007, 112, D10306 DOI: 10.1029/2006JD008061. (7) Arakaki, T.; Kuroki, Y.; Okada, K.; Nakama, Y.; Ikota, H.; Kinjo, M.; Higuchi, T.; Uehara, M.; Tanahara, A. Chemical composition and photochemical formation of hydroxyl radicals in aqueous extracts of aerosol particles collected in Okinawa, Japan. Atmos. Environ. 2006, 40, 4764−4774. (8) Haag, W. R.; Hoigne, J. Photo-sensitized oxidation in natural water via OH radicals. Chemosphere 1985, 14 (11−12), 1659−1671. (9) Anastasio, C.; McGregor, K. G. Chemistry of fog water in California’s Central Valley: 1. In situ photoformation of hydroxyl radical and singlet molecular oxygen. Atmos. Environ. 2001, 35 (6), 1079−1089. (10) Vione, D.; Falletti, G.; Maurino, V.; Minero, C.; Pelizzetti, E.; Malandrino, M.; Ajassa, R.; Olariu, R. I.; Arsene, C. Sources and sinks of hydroxyl radicals upon irradiation of natural water samples. Environ. Sci. Technol. 2006, 40 (12), 3775−3781. (11) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (OH/O−) in aqueous solution. J. Phys. Chem. Ref. Data 1988, 17 (2), 513−886. (12) Ross, A. B.; Mallard, W. G.; Helman, W. P.; Buxton, G. V.; Huie, R. E.; Neta, P. NDRL/NIST Solution Kinetics Database, Version 3.0, http://kinetics.nist.gov/solution/. (13) Teraji, T.; Arakaki, T. Bimolecular rate constants between levoglucosan and hydroxyl radical: Effects of pH and temperature. Chem. Lett. 2010, 39, 900−901. (14) Herrmann, H.; Ervens, B.; Jacobi, H. W.; Wolke, R.; Nowacki, P.; Zellner, R. CAPRAM2.3: A chemical aqueous phase radical mechanism for tropospheric chemistry. J. Atmos. Chem. 2000, 36 (3), 231−284. (15) Warneck, P. The relative importance of various pathways for the oxidation of sulfur dioxide and nitrogen dioxide in sunlit continental fair weather clouds. Phys. Chem. Chem. Phys. 1999, 1 (24), 5471−5483. (16) Jacob, D. J.; Gottlieb, E. W.; Prather, M. J. Chemistry of a polluted cloudy boundary layer. J. Geophys. Res. Atmos. 1989, 94 (D10), 12975−13002. (17) Di Carlo, P.; Brune, W. H.; Martinez, M.; Harder, H.; Lesher, R.; Ren, X. R.; Thornberry, T.; Carroll, M. A.; Young, V.; Shepson, P. B.; Riemer, D.; Apel, E.; Campbell, C. Missing OH reactivity in a forest: Evidence for unknown reactive biogenic VOCs. Science 2004, 304 (5671), 722−725. (18) Noelscher, A. C.; Williams, J.; Sinha, V.; Custer, T.; Song, W.; Johnson, A. M.; Axinte, R.; Bozem, H.; Fischer, H.; Pouvesle, N.; Phillips, G.; Crowley, J. N.; Rantala, P.; Rinne, J.; Kulmala, M.; Gonzales, D.; Valverde-Canossa, J.; Vogel, A.; Hoffmann, T.; Ouwersloot, H. G.; de Arellano, J. V.-G.; Lelieveld, J. Summertime total OH reactivity measurements from boreal forest during HUMPPA-COPEC 2010. Atmos. Chem. Phys. 2012, 12 (17), 8257− 8270. (19) Hamilton, J. F.; Webb, P. J.; Lewis, A. C.; Hopkins, J. R.; Smith, S.; Davy, P. Partially oxidised organic components in urban aerosol using GCXGC−TOF/MS. Atmos. Chem. Phys. 2004, 4, 1279−1290. (20) Graber, E. R.; Rudich, Y. Atmospheric HULIS: How humic-like are they? A comprehensive and critical review. Atmos. Chem. Phys. 2006, 6, 729−753. (21) Tilgner, A.; Herrmann, H. Radical-driven carbonyl-to-acid conversion and acid degradation in tropospheric aqueous systems studied by CAPRAM. Atmos. Environ. 2010, 44 (40), 5415−5422. (22) Albinet, A.; Minero, C.; Vione, D. Photochemical generation of reactive species upon irradiation of rainwater: Negligible photoactivity of dissolved organic matter. Sci. Total Environ. 2010, 408 (16), 3367− 3373.
(23) Hoigné, J., Chemistry of aqueous ozone and transformation of pollutants by ozonation and advanced oxidation processes. In The Handbook of Environmental Chemistry, Hrubec, J., Ed.; Springer-Verlag: Berlin, 1998; Vol. 5: Water Pollution, Part C, Quality and Treatment of Drinking Water, pp 83−141. (24) Westerhoff, P.; Mezyk, S. P.; Cooper, W. J.; Minakata, D. Electron pulse radiolysis determination of hydroxyl radical rate constants with Suwannee river fulvic acid and other dissolved organic matter isolates. Environ. Sci. Technol. 2007, 41 (13), 4640−4646. (25) Dong, M. M.; Mezyk, S. P.; Rosario-Ortiz, F. L. Reactivity of effluent organic matter (EfOM) with hydroxyl radical as a function of molecular weight. Environ. Sci. Technol. 2010, 44, 5714−5720. (26) Brezonik, P. L.; Fulkersonbrekken, J. Nitrate-induced photolysis in natural watersControls on concentrations of hydroxyl radical photo-intermediates by natural scavenging agents. Environ. Sci. Technol. 1998, 32 (19), 3004−3010. (27) Katsoyiannis, I. A.; Canonica, S.; von Gunten, U. Efficiency and energy requirements for the transformation of organic micropollutants by ozone, O3/H2O2 and UV/H2O2. Water Res. 2011, 45 (13), 3811− 3822. (28) Haag, W. R.; Yao, C. C. D. Ozonation of U.S. drinking water sources: HO concentration and oxidation-competition values. In Proceedings of the 11th Ozone World Congress, 1993, Vol. 2, S-17-119− 126. (29) Herckes, P.; Hannigan, M. P.; Trenary, L.; Lee, T.; Collett, J. L. Organic compounds in radiation fogs in Davis (California). Atmos. Res. 2002, 64 (1−4), 99−108. (30) Mader, B. T.; Yu, J. Z.; Xu, J. H.; Li, Q. F.; Wu, W. S.; Flagan, R. C.; Seinfeld, J. H. Molecular composition of the water-soluble fraction of atmospheric carbonaceous aerosols collected during ACE-Asia. J. Geophys. Res. Atmos. 2004, 109 (D6), D06206. (31) Collett, J. L., Jr.; Herckes, P.; Youngster, S.; Lee, T. Processing of atmospheric organic matter by California radiation fogs. Atmos. Res. 2008, 87 (3−4), 232−241. (32) Raja, S.; Raghunathan, R.; Yu, X.-Y.; Lee, T.; Chen, J.; Kommalapati, R. R.; Murugesan, K.; Shen, X.; Qingzhong, Y.; Valsaraj, K. T.; Collett, J. L., Jr. Fog chemistry in the Texas−Louisiana Gulf Coast corridor. Atmos. Environ. 2008, 42 (9), 2048−2061. (33) Raja, S.; Raghunathan, R.; Kommalapati, R. R.; Shen, X.; Collett, J. L., Jr.; Valsaraj, K. T. Organic composition of fogwater in the Texas− Louisiana gulf coast corridor. Atmos. Environ. 2009, 43 (27), 4214− 4222. (34) Feczko, T.; Puxbaum, H.; Kasper-Giebl, A.; Handler, M.; Limbeck, A.; Gelencser, A.; Pio, C.; Preunkert, S.; Legrand, M. Determination of water and alkaline extractable atmospheric humiclike substances with the TU Vienna HULIS analyzer in samples from six background sites in Europe. J. Geophys. Res. Atmos. 2007, 112 (D23), D23S10. (35) Krivácsy, Z.; Kiss, G.; Ceburnis, D.; Jennings, G.; Maenhaut, W.; Salma, I.; Shooter, D. Study of water-soluble atmospheric humic matter in urban and marine environments. Atmos. Res. 2008, 87 (1), 1−12. (36) Arakaki, T.; Miyake, T.; Shibata, M.; Sakugawa, H. Photochemical formation and scavenging of hydroxyl radical in rain and dew waters. Nippon Kagaku Kaishi 1999, 5, 335−340. (37) Faust, B. C.; Allen, J. M. Aqueous-phase photochemical formation of hydroxyl radical in authentic cloudwaters and fogwaters. Environ. Sci. Technol. 1993, 27 (6), 1221−1224. (38) Matthijsen, J.; Builtjes, P. J. H.; Sedlak, D. L. Cloud model experiments of the effect of iron and copper on tropospheric ozone under marine and continental conditions. Met. Atmos. Phys. 1995, 57 (1−4), 43−60. (39) Monod, A.; Carlier, P. Impact of clouds on the tropospheric ozone budget: Direct effect of multiphase photochemistry of soluble organic compounds. Atmos. Environ. 1999, 33 (27), 4431−4446. (40) Ervens, B.; Feingold, G.; Frost, G. J.; Kreidenweis, S. M. A modeling study of aqueous production of dicarboxylic acids: 1. Chemical pathways and speciated organic mass production. J. Geophys. Res. Atmos. 2004, 109 (D15), D15205. 8202
dx.doi.org/10.1021/es401927b | Environ. Sci. Technol. 2013, 47, 8196−8203
Environmental Science & Technology
Article
(41) Herrmann, H.; Tilgner, A.; Barzaghi, P.; Majdik, Z.; Gligorovski, S.; Poulain, L.; Monod, A. Towards a more detailed description of tropospheric aqueous phase organic chemistry: CAPRAM 3.0. Atmos. Environ. 2005, 39 (23−24), 4351−4363. (42) Deguillaume, L.; Leriche, M.; Monod, A.; Chaumerliac, N. The role of transition metal ions on HOx radicals in clouds: A numerical evaluation of its impact on multiphase chemistry. Atmos. Chem. Phys. 2004, 4, 95−110. (43) Pandis, S. N.; Seinfeld, J. H. Sensitivity analysis of a chemical mechanism for aqueous-phase atmospheric chemistry. J. Geophys. Res. Atmos. 1989, 94 (D1), 1105−1126. (44) Barth, M. C.; Sillman, S.; Hudman, R.; Jacobson, M. Z.; Kim, C. H.; Monod, A.; Liang, J. Summary of the cloud chemistry modeling intercomparison: Photochemical box model simulation. J. Geophys. Res. Atmos. 2003, 108 (D7), 4214. (45) Warneck, P. In-cloud chemistry opens pathway to the formation of oxalic acid in the marine atmosphere. Atmos. Environ. 2003, 37 (17), 2423−2427. (46) Herrmann, H.; Ervens, B.; Nowacki, P.; Wolke, R.; Zellner, R. A chemical aqueous phase radical mechanism for tropospheric chemistry. Chemosphere 1999, 38 (6), 1223−1232. (47) Zhou, X.; Davis, A. J.; Kieber, D. J.; Keene, W. C.; Maben, J. R.; Maring, H.; Dahl, E. E.; Izaguirre, M. A.; Sander, R.; Smoydzyn, L. Photochemical production of hydroxyl radical and hydroperoxides in water extracts of nascent marine aerosols produced by bursting bubbles from Sargasso seawater. Geophys. Res. Lett. 2008, 35 (20), L20803. (48) Goldstone, J. V.; Pullin, M. J.; Bertilsson, S.; Voelker, B. M. Reactions of hydroxyl radical with humic substances: Bleaching, mineralization, and production of bioavailable carbon substrates. Environ. Sci. Technol. 2002, 36 (3), 364−372. (49) McKay, G.; Dong, M. M.; Kleinman, J. L.; Mezyk, S. P.; RosarioOrtiz, F. L. Temperature dependence of the reaction between the hydroxyl radical and organic matter. Environ. Sci. Technol. 2011, 45 (16), 6932−6937.
8203
dx.doi.org/10.1021/es401927b | Environ. Sci. Technol. 2013, 47, 8196−8203