Search

The dithiothreitol (DTT) assay is widely used to measure the oxidative potential of particulate matter. Results are typically presented in mass-normalized units (e.g., pmols DTT lost per minute per microgram PM) to allow for comparison among samples. Use of this unit assumes that the mass-normalized DTT response is constant and independent of the mass concentration of PM added to the DTT assay. However, based on previous work that identified non-linear DTT responses for copper and manganese, this basic assumption (that the mass-normalized DTT response is independent of the concentration of PM added to the assay) should not be true for samples where Cu and Mn contribute significantly to the DTT signal. To test this we measured the DTT response at multiple PM concentrations for eight ambient particulate samples collected at two locations in California. The results confirm that for samples with significant contributions from Cu and Mn, the mass-normalized DTT response can strongly depend on the concentration of PM added to the assay, varying by up to an order of magnitude for PM concentrations between 2 and 34 μg mL-1. This mass dependence confounds useful interpretation of DTT assay data in samples with significant contributions from Cu and Mn, requiring additional quality control steps to check for this bias. To minimize this problem, we discuss two methods to correct the mass-normalized DTT result and we apply those methods to our samples. We find that it is possible to correct the mass-normalized DTT result, although the correction methods have some drawbacks and add uncertainty to DTT analyses. More broadly, other DTT-active species might also have non-linear concentration-responses in the assay and cause a bias. In addition, the same problem of Cu- and Mn-mediated bias in mass-normalized DTT results might affect other measures of acellular redox activity in PM and needs to be addressed.

We investigated the aqueous photochemistry of six phenolic carbonyls - vanillin, acetovanillone, guaiacyl acetone, syringaldehyde, acetosyringone, and coniferyl aldehyde – that are emitted from wood combustion. The phenolic carbonyls absorb significant amounts of solar radiation and decay rapidly via direct photodegradation, with lifetimes (τ) of 13–140 min under Davis, CA winter solstice sunlight at midday (solar zenith angle = 62°). The one exception is guaiacyl acetone, where the carbonyl group is not directly connected to the aromatic ring: This species absorbs very little sunlight and undergoes direct photodegradation very slowly (τ > 103 min). We also found that the triplet excited states (3C*) of the phenolic carbonyls rapidly oxidize syringol (a methoxyphenol without a carbonyl group), on timescales of 1–5 h for solutions containing 5 μM phenolic carbonyl. The direct photodegradation of the phenolic carbonyls, and the oxidation of syringol by 3C*, both efficiently produce low volatility products, with SOA mass yields ranging from 80 to 140%. Contrary to most aliphatic carbonyls, under typical fog conditions we find that the primary sink for the aromatic phenolic carbonyls is direct photodegradation in the aqueous phase. In areas of significant wood combustion, phenolic carbonyls appear to be small but significant sources of aqueous SOA: over the course of a few hours, nearly all of the phenolic carbonyls will be converted to SOA via direct photodegradation, enhancing the POA mass from wood combustion by approximately 3–5%.

Green leaf volatiles (GLVs) are a group of biogenic volatile organic compounds (BVOCs) released into the atmosphere by vegetation. BVOCs produce secondary organic aerosol (SOA) via gas-phase reactions, but little is known of their aqueous-phase oxidation as a source of SOA. GLVs can partition into atmospheric water phases, e.g., fog, mist, dew or rain, and be oxidized by hydroxyl radicals (˙OH). These reactions in the liquid phase also lead to products that have higher molecular weights, increased polarity, and lower vapor pressures, ultimately forming SOA after evaporation of the droplet. To examine this process, we investigated the aqueous, ˙OH-mediated oxidation of methyl jasmonate (MeJa) and methyl salicylate (MeSa), two GLVs that produce aqueous-phase SOA. High performance liquid chromatography/electrospray ionization mass spectrometry (HPLC-ESI-MS) was used to monitor product formation. The oxidation products identified exhibit higher molecular mass than their parent GLV due to either dimerization or the addition of oxygen and hydroxyl functional groups. The proposed structures of potential products are based on mechanistic considerations combined with the HPLC/ESI-MS data. Based on the structures, the vapor pressure and the Henry's law constant were estimated with multiple methods (SPARC, SIMPOL, MPBPVP, Bond and Group Estimations). The estimated vapor pressures of the products identified are significantly (up to 7 orders of magnitude) lower than those of the associated parent compounds, and therefore, the GLV oxidation products may remain as SOA after evaporation of the water droplet. The contribution of the identified oxidation products to SOA formation is estimated based on measured HPLC-ESI/MS responses relative to previous aqueous SOA mass yield measurements.

Recent findings suggest that secondary organic aerosols (SOA) formed from aqueous-phase reactions of some organic species, including phenols, contribute significantly to particulate mass in the atmosphere. In this study, we employ a Fourier transform infrared (FT-IR) spectroscopic technique to identify and quantify the functional group makeup of phenolic SOA. Solutions containing an oxidant (hydroxyl radical or 3,4-dimethoxybenzaldehyde) and either one phenol (phenol, guaiacol, or syringol) or a mixture of phenols mimicking softwood or hardwood emissions were illuminated to make SOA, atomized, and collected on a filter. We produced laboratory standards of relevant organic compounds in order to develop calibrations for four functional groups: carbonyls (CO), saturated C–H, unsaturated C–H and O–H. We analyzed the SOA samples with transmission FT-IR to identify and determine the amounts of the four functional groups. The carbonyl functional group accounts for 3–12% of the SOA sample mass in single phenolic SOA samples and 9–14% of the SOA sample mass in mixture samples. No carbonyl functional groups are present in the initial reactants. Varying amounts of each of the other functional groups are observed. Comparing carbonyls measured by FT-IR (which could include aldehydes, ketones, esters, and carboxylic acids) with eight small carboxylic acids measured by ion chromatography indicates that the acids only account for an average of 20% of the total carbonyl reported by FT-IR.

Green leaf volatiles (GLVs) are a class of oxygenated hydrocarbons released from vegetation, especially during mechanical stress or damage. The potential for GLVs to form secondary organic aerosol (SOA) via aqueous-phase reactions is not known. Fog events over vegetation will lead to the uptake of GLVs into water droplets, followed by aqueous-phase reactions with photooxidants such as the hydroxyl radical (OH). In order to determine if the aqueous oxidation of GLVs by OH can be a significant source of secondary organic aerosol, we studied the partitioning and reaction of five GLVs: cis-3-hexen-1-ol, cis-3-hexenyl acetate, methyl salicylate, methyl jasmonate, and 2-methyl-3-butene-2-ol. For each GLV we measured the kinetics of aqueous oxidation by OH, and the corresponding SOA mass yield. The second-order rate constants for GLVs with OH were all near diffusion controlled, (5.4–8.6) × 109 M−1 s−1 at 298 K, and showed a small temperature dependence, with an average activation energy of 9.3 kJ mol−1 Aqueous-phase SOA mass yields ranged from 10 to 88%, although some of the smaller values were not statistically different from zero. Methyl jasmonate was the most effective aqueous-phase SOA precursor due to its larger Henry's law constant and high SOA mass yield (68 ± 8%). While we calculate that the aqueous-phase SOA formation from the five GLVs is a minor source of aqueous-phase SOA, the availability of other GLVs, other oxidants, and interfacial reactions suggest that GLVs overall might be a significant source of SOA via aqueous reactions.

Much attention has focused on the highly reactive hydroxyl radical in the oxidation of trace organic compounds on snow and ice (and subsequent release of volatile organics to the atmospheric boundary layer) but other oxidants are likely also important in this processing. Here we examine the ice chemistry of singlet molecular oxygen (1O2*), which can be significant in atmospheric water drops but has not been examined in ice or snow. To examine 1O2* on ice we illuminate laboratory ices containing Rose Bengal (RB) as the source of 1O2*, furfuryl alcohol (FFA) as the probe, and Na2SO4 to control the total solute concentration. We find that the 1O2*-mediated loss of FFA (and, thus, the 1O2* concentration) is up to 11,000 times greater on ice than in the equivalent liquid sample at the same photon flux. We attribute this large increase in the 1O2* steady-state concentration to the freeze-concentration of solutes into liquid-like regions (LLRs) in/on ice: compared to the initial solution, in the LLRs of ice the sources for 1O2* are highly concentrated, while the concentration of the dominant sink for 1O2* (i.e., water) remains largely unchanged. Similar to results expected in liquid solution, rates of FFA loss in ice depend on both the initial sensitizer concentration and temperature, providing evidence that these reactions occur in LLRs. However, we find that the enhancement in 1O2* concentrations on ice does not follow predictions from freezing-point depression, likely because experiments were conducted below the eutectic temperature for sodium sulfate, where all of the salt should have precipitated. We also explore a method for separating 1O2* and OH contributions to FFA oxidation in laboratory ices and show its application to two natural snow samples. We find that 1O2* concentrations in these snows are approximately 100 times higher than observed in polluted, mid-latitude fog waters, showing that the enhancement of 1O2* on ice is environmentally relevant and that 1O2* could be a significant sink for electron-rich organic compounds in snow.

Hydrogen peroxide (HOOH) is an important trace constituent in snow and ice, including in Arctic and Antarctic ice cores. To better understand the budget of snowpack HOOH, here we examine its production in illuminated snow and ice. To evaluate what types of compounds might be important photochemical sources of HOOH, we first illuminated laboratory ice samples containing 10 different model organic compounds: guaiacol, phenol, syringol, benzoate, formate, octanal, octanoic acid, octanedioic acid, phenylalanine, and mixtures of oxalate with iron (III). Half of these compounds produced little or no HOOH during illumination, but two classes of compounds were very reactive: phenolic compounds (with rates of HOOH of 6–62 nM-HOOH h −1 μM −1 -phenolic) and mixtures of Fe(III) with a stoichiometric excess of oxalate (with rates of HOOH production as high as 2,000,000 nM h −1 per μM iron). To quantify rates of HOOH production in the environment we also illuminated snow samples collected from the Arctic and Antarctic. The average (±1 σ ) HOOH production rate in these samples was low, 5.3 ± 5.0 nM h −1 and replicate measurements showed high variability. In some natural samples there was an initial burst of HOOH production (with a rate approximately 10 times higher than the average production rate), followed by reduced rates at subsequent time points. Although our laboratory ice samples reveal that illuminated organics and metal-organic complexes can form HOOH, the low rates of HOOH formation in the Arctic and Antarctic snow samples suggest this process has only a modest impact on the HOOH budget in the snowpack.

Although polycyclic aromatic hydrocarbons (PAHs) are common pollutants in snow, there is little quantitative data about their rates of photodegradation in this environment. To begin to address this gap, we have measured the degradation kinetics of phenanthrene, pyrene, and fluoranthene on ice, as these are the most abundant PAHs in arctic snow. Frozen aqueous solutions of individual PAHs, with and without added hydrogen peroxide (HOOH) as a source of hydroxyl radical ( OH), were illuminated with simulated sunlight. For all three PAHs, direct photodecay is the main mechanism of degradation, while OH-initiated indirect photodegradation is a minor sink. Rate constants (±1 SE) for direct photodegradation extrapolated to midday, surface snow conditions at Summit, Greenland on the summer solstice are 3.8 (±0.8) × 10 −5 , 28 (±3) × 10 −5 , and 1.4 (±0.7) × 10 −5 s −1 for phenanthrene, pyrene, and fluoranthene, respectively. Apparent quantum efficiencies for photodegradation with simulated sunlight were 3.8 (±0.8) × 10 −3 , 4.3 (±0.5) × 10 −4 , and 2 (±1) × 10 −5 , respectively. Calculated PAH lifetimes in surface snow under Summit conditions are 1–19 h during mid-summer, but increase to >100 days in the dark winter. While the short photodegradation lifetimes in the summer suggest that there should be no appreciable PAH levels in this season, past measurements at Summit sometimes show significant levels of these PAHs in summer surface snow. This discrepancy is likely due to differences in PAH location between lab samples (where the PAHs are probably in quasi-liquid layers) and real snow (where PAHs are likely primarily associated with particulate matter).

Epidemiological research has linked exposure to atmospheric particulate matter (PM) to several adverse health effects, including cardiovascular and pulmonary morbidity and mortality. Despite these links, the mechanisms by which PM causes adverse health effects are poorly understood. The generation of hydroxyl radical ( OH) and other reactive oxygen species (ROS) through transition-metal-mediated pathways is one of the main hypotheses for PM toxicity. In order to better understand the ability of particulate transition metals to produce ROS, we have quantified the amounts of OH produced from dissolved iron and copper in a cell-free, surrogate lung fluid (SLF). We also examined how two important biological molecules, citrate and ascorbate, affect the generation of OH by these metals. We have found that Fe(II) and Fe(III) produce little OH in the absence of ascorbate and citrate, but that they efficiently make OH in the presence of ascorbate and this is further enhanced when citrate is also added. In the presence of ascorbate, with or without citrate, the oxidation state of iron makes little difference on the amount of OH formed after 24 h. In the case of Cu(II), the production of OH is greatly enhanced in the presence of ascorbate, but is inhibited by the addition of citrate. The mechanism for this effect is unclear, but appears to involve formation of a citrate–copper complex that is apparently less reactive than free, aquated copper in either the generation of hydrogen peroxide (HOOH) or in the Fenton-like reaction of copper with HOOH to make OH. By quantifying the amount of OH that Fe and Cu can produce in surrogate lung fluid, we have provided a first step into being able to predict the amounts of OH that can be produced in the human lung from exposure to PM containing known amounts of transition metals.

We measured the photoformation of hydroxyl radical ( ⋅ OH ) on snow grains at Summit, Greenland during the spring and summer. Midday rates of ⋅ OH formation in the snow phase in the summer range from 130 to 610 nmol L - 1 h - 1 , expressed relative to the liquid equivalent volume of snow. Calculated formation rates of snow-grain ⋅ OH based on the photolysis of hydrogen peroxide and nitrate agree well with our measured rates during summer, indicating that there are probably not other major sources of ⋅ OH under these conditions. Throughout both the spring and summer, HOOH is by far the dominant source of snow-grain ⋅ OH ; on average, HOOH produces approximately 100 times more ⋅ OH than does NO 3 - . Rates of ⋅ OH photoformation have a strong seasonal dependence and increase by approximately a factor of 10 between early spring and summer at midday. The rate of ⋅ OH photoformation on snow grains decreases rapidly with depth in the snowpack, with approximately 90% of photoformation occurring within the top 10 cm, although ⋅ OH formation occurs to depths below 20 cm. The formation of ⋅ OH on snow grains likely initiates a suite of reactions in the snowpack, including the transformation of organic carbon (OC) and oxidation of halides. The reaction of ⋅ OH with OC probably forms a number of volatile organic compounds (VOCs) that are potentially emitted into the atmospheric boundary layer. Indeed, our measured rates of ⋅ OH photoformation on snow grains are large enough that they could account for previously reported fluxes of VOCs from the snowpack at Summit, although the relative importance of thermal desorption and photochemical production for most of these VOCs still needs to be determined.

Author(s): Dibb, JE; Albert, M; Anastasio, C; Atlas, E; Beyersdorf, AJ; Blake, NJ; Blake, DR; Bocquet, F; Burkhart, JF; Chen, G; Cohen, L; Conway, TJ; Courville, Z; Frey, MM; Friel, DK; Galbavy, ES; Hall, S; Hastings, MG; Helmig, D; Greg Huey, L; Hutterli, MA; Jarvis, JC; Lefer, BL; Meinardi, S; Neff, W; Oltmans, SJ; Sherwood Rowland, F; Sjostedt, SJ; Steig, EJ; Swanson, AL; Tanner, DJ

Rates of photochemical reactions within the snowpack, both on snow grains and in the firn air, depend on how actinic flux is attenuated as a function of depth. This snowpack photon flux can either be measured directly (e.g., with spectral radiometers in the snow) or indirectly (e.g., by chemical actinometry where the rate of a photochemical reaction is measured). In this work we use both techniques to measure the rate constant for nitrate photolysis on water–ice, j ( NO 3 - ) , on the snowpack surface and beneath at Summit, Greenland during spring and summer. The surface measurements from these two methods are generally similar ((1−2)×10–7 s−1 during midday near the summer solstice) and follow expected diurnal and annual trends. In addition, both methods show a similar effect of snow albedo on photolysis at the surface: rate constants measured on the surface snow were approximately 45% higher than values predicted for the surface based on regressions of in-snow rate constants. Average (± σ ) j ( NO 3 - ) e-folding depths (the depth along which the rate constant decreases by a factor of e) during the 2 field seasons are 10.3(±2.8) cm from actinometry tubes and 8.4(±2.4) cm from in-snow spectral radiometers.

We have carried out a series of laboratory experiments to investigate the oxidation of bromide (Br−) by hydroxyl radical ( OH) in solutions used to mimic sea-salt particles. Aqueous halide solutions with nitrate or hydrogen peroxide (HOOH) as a photochemical source of OH were illuminated with 313 nm light and the resulting gaseous bromine (Br*(g)) was collected. While illumination of these solutions nearly always formed gaseous bromine (predominantly Br2 based on modeling results), there was no evidence for the release of gaseous chlorine. The rate of Br*(g) release increased (up to a plateau value) with increasing concentrations of bromide and was enhanced at lower pH values for both nitrate and HOOH solutions. Increased ionic strength in nitrate solutions inhibited Br*(g) release and the extent of inhibition was dependent upon the salt used. In HOOH solutions, however, no ionic strength effects were observed and the presence of Cl− strongly enhanced Br*(g) release. Overall, for conditions typical of aged, deliquesced, sea-salt particles, the efficiencies of gaseous bromine release, expressed as mole of Br*(g) released per mole of OH photochemically formed, were typically 20–30%. Using these reaction efficiencies, we calculated the Br2(g) release rate from aged, ambient sea-salt particles due to OH oxidation to be approximately 0.07 pptv h−1 with the main contributions from nitrate photolysis and partitioning of gas-phase OH into the particle. While our solution conditions are simplified compared to ambient particles, this estimated rate of Br2 release is high enough to suggest that OH-mediated reactions in sea-salt particles could be a significant source of reactive bromine to the marine boundary layer.

Epidemiological and toxicological studies have shown a relation between the inhalation of atmospheric particles and adverse cardiopulmonary health effects. The generation of reactive oxygen species (ROS) by particles is one current hypothesis for their toxic effects. Thus a quantitative measurement of ROS is important since that will be an index to assess the oxidative stress that particles may cause in the lung. We have developed a technique to quantitatively and specifically measure OH (the strongest biological ROS) in an aqueous, buffered extract solution as a surrogate lung fluid (SLF). Using this technique we quantitatively measured OH formation in SLF containing hydrogen peroxide (HOOH) for samples of flame soot particles, carbon black, and ambient fine particles (PM2.5). We have found that OH is formed by flame soot, independent of transition metals, with a dose-dependent linear response that depends upon HOOH concentration. Experiments with carbon black revealed that its mass-normalized OH generation was ∼10 times lower than that of flame soot, suggesting that carbon black is not a good surrogate for soot particles in health effect studies, at least in terms of oxidative stress. Mass-normalized OH generation by ambient PM2.5 was 6–30 times larger than that of flame soot. While much of the PM2.5 reactivity was suppressed by pretreating samples with a transition metal chelator, there was a significant fraction of reactivity which was not affected. Our results suggest that the in vivo generation of free radicals, specifically OH, by inhalation of PM2.5 is partially due to carbonaceous soot as well as transition metals.

We have studied oxidant photoformation in aqueous extracts of aerosol particles collected from Alert, Nunavut, Canada during springtime. Absorption spectra of the extracts reveal that Alert particles have tremendous light absorption coefficients (e.g., α 300 ∼100 cm −1 ) as a result of unidentified, pH-dependent, water-soluble chromophores. Illumination of the extracts leads to the rapid formation of both hydroxyl radical ( OH) and hydrogen peroxide (HOOH). Based on our laboratory results, the calculated rate of OH photoformation in Alert particles is very rapid, ∼1 mM h −1 (midday, 1 April, 248 K), with nitrate photolysis contributing only ∼10% of the total rate. Deposition of these aerosol particles, in conjunction with smaller contributions from gaseous chromophores, leads to estimated rates of OH photoformation in the quasi-liquid layer of surface snow of 20–40 μM h −1 ; approximately a third of this reactivity is from nitrate photolysis. The estimated 24-h-average rate of HOOH photoformation in Alert particles (∼9 mM h −1 on 1 April) is large enough to be a major source of HOOH to both the particles and the gas phase. In contrast, particle-derived reactions in the snow appear to be a minor source of HOOH to the surface snowpack. The effects of OH and HOOH photoformation in particles and snowpack likely include the oxidation of organic carbon, halides, and S(IV) species to yield products such as volatile aldehydes and carboxylic acids, photoactive halogens, and sulfuric acid. In addition, OH and HOOH photoformation within the snowpack might significantly alter snow and ice core records of HOOH and other trace gases.

Atmospheric fine particles (PM2.5) collected during August 1997–July 1998 and wintertime fog waters collected during 1997–1999 at Davis, California were analyzed for free and combined amino compounds. In both PM2.5 and fog waters, the average concentrations of combined amino compounds (CAC, e.g., proteins and peptides) were generally 4–5 times higher than those of free amino compounds (FAC, i.e., amino acids and alkyl amines). Concentrations of total amino compounds (TAC=FAC+CAC) ranged from 1260 to 3650 pmol m−3 air in PM2.5, and from 1620 to 5880 pmol m−3 air in fog waters. Average values (±1σ) were 2500±879 and 3400±1430 pmol m−3 air, respectively. Concentrations of amino compounds in PM2.5 varied seasonally, with a peak during late winter and early spring. Ornithine was a major FAC component in both PM2.5 and fog waters (typically accounting for ∼20% of FAC), but these sample types otherwise had fairly different FAC distributions. FAC in PM2.5 were enriched in protein-type amino species such as glycine/threonine, serine and alanine, while fog water FAC had significantly higher levels of non-protein species such as methylamine, γ-aminobutyric acid and ethanolamine. The compositions of CAC in PM2.5 and fogs were fairly similar and were mainly protein-type. Mass concentrations of TAC in PM2.5 and fog waters were, on average, 302 and 399 ng m−3 air, respectively. Amino compounds were an important component of the organic carbon pool for both fog and particles, with TAC accounting for an average of 13% of the dissolved organic carbon in fog waters and ∼10% of the water-soluble organic carbon in PM2.5. At these levels amino compounds likely play important roles in the chemistry of fog drops and fine particles, for example by influencing their buffering capacity and basicity.

Although organic nitrogen (ON) has been found to be a ubiquitous and significant component in wet and dry deposition, almost nothing is known about its concentration or composition in fog waters. To address this gap, we have investigated the concentration and composition of ON in fog waters collected in Davis, in California's Central Valley. Significant quantities of dissolved organic nitrogen (DON) were found in these samples, with a median concentration of 303 μM N (range=120–1630 μM N). DON typically represented approximately 16% of the total dissolved nitrogen (inorganic+organic) in Davis fog waters. The median concentration of nitrogen in free amino acids and alkyl amines was 16 μM N (range=3.8–120 μM N), which accounted for 3.4% of the DON in Davis fogs. Thus, although the absolute concentrations of free amino compounds were significant, they were only a minor component of the DON pool. Combined amino nitrogen (e.g., proteins and peptides) was present at higher concentrations and accounted for 6.1–29% (median=16%) of DON. Overall, free and combined amino compounds typically accounted for a median value of 22% of DON in the fog waters. The high concentrations of DON found, and the fact that amino and other N-containing organic compounds can serve as nitrogen sources for microorganisms and plants, indicate that atmospheric ON compounds likely play an important role in nitrogen cycling in the Central Valley. In addition, due to the basicity of some N functional groups, ON compounds likely contribute to the previously observed acid buffering capacity of Central Valley fog waters. Finally, a comparison of fog waters with fine particles (PM2.5) collected from the same site during the same period of time indicated that the median concentrations (mol N m−3-air) of total water-soluble ON, free amino nitrogen and total amino nitrogen were very similar in the fog water and PM2.5. Given the high water solubility of many organic N compounds, this result suggests that ON might contribute to the hygroscopic properties of atmospheric particles.

Although amino compounds are seemingly ubiquitous in atmospheric particles and deposition, little is known of their fate in the troposphere. We report here on the fate of 21 amino acids and alkyl amines in fog waters from Davis, California, illuminated with simulated sunlight or monochromatic light. In all experiments four amino acids – histidine (His), methionine (Met), tryptophan (Trp), and tyrosine (Tyr) – consistently decayed, with half-lives that ranged from ∼1 h (Met) to ∼23 h (Tyr) in midday, winter-solstice sunlight at Davis, CA ( solar zenith angle =62°) . Half-lives for the remaining amino compounds examined were typically >45 h in our experiments. Reactions with photoformed hydroxyl radical ( · OH ) and singlet molecular oxygen (O2(1Δg) or 1 O 2 ∗ ) accounted for essentially all of the loss of His and Tyr, the less reactive of the four amino acids that consistently decayed, but were minor sinks for the more reactive compounds (Met and Trp). Additional experiments revealed that methionine sulfoxide (MetSO) was formed with a yield of 58–88% during the oxidation of methionine, suggesting that the ratio of MetSO to Met might be a useful chemical marker for the age of atmospheric particles and drops. Other products expected from the transformation of amino compounds include ammonia, organic acids, and possibly mutagenic nitrosoaromatics. To complement our laboratory experiments, we also calculated rates of transformations of amino acids in near-neutral pH fog drops under ambient conditions. These calculations reveal that ozone should be a major sink for amino acids and that half-lives for many amino acids in ambient fog drops will be much shorter than those determined in our photochemistry experiments. Overall, our results indicate that reactions in atmospheric condensed phases will transform amino nitrogen compounds (including free amino acids as well as proteins and peptides) and, consequently, increase the bioavailability of nitrogen in atmospheric deposition.

The aqueous-phase photoformation of hydroxyl radical ( ⋅ OH ) and singlet molecular oxygen (O2(1Δg) or 1 O ∗ 2 ) was characterized in winter fog waters collected in Davis, California. All of the samples studied formed ⋅ OH and 1 O ∗ 2 upon illumination with simulated sunlight. Nitrite photolysis was a major source of ⋅ OH in these samples, accounting for 47–100% of ⋅ OH photoformation. Compared to calculated rates of gas-to-drop partitioning, in situ photoformation was a significant source of ⋅ OH to all but the smallest fog drops, and was the dominant source of aqueous-phase 1 O ∗ 2 . Measured lifetimes of ⋅ OH in the fog drops ranged from 0.38 to 1.5 μs. These values are significantly shorter than those predicted based on known drop constituents, indicating that uncharacterized compounds – likely organic – are significant sinks for ⋅ OH in fog waters. Based on measured steady-state concentrations, both ⋅ OH and 1 O ∗ 2 are likely to play significant roles in the transformations of trace species in fog drops. Hydroxyl radical appears to be a relatively significant sink for refractory compounds and a minor sink for reactive trace species. Conversely, 1 O ∗ 2 will be a minor sink for refractory compounds but a significant sink for certain electron-rich reactive trace species.

Aqueous-phase photolysis of biacetyl represents a heretofore uncharacterized sink for biacetyl and source of organic acids and peroxides to aqueous aerosols, and fog and cloud drops. The photolysis half-life of aqueous-phase biacetyl is approximately 1.0–1.6 h for a solar zenith angle of 36° (midday equinox sunlight in Durham, NC). Major products of aqueous biacetyl photolysis are acetic acid, peroxyacetic acid, and hydrogen peroxide. Pyruvic acid and methylhydroperoxide are minor photoproducts. Common atmospheric reductants (H-atom donors), such as formate, formaldehyde, glyoxal, phenol (as a model for phenolic compounds) and α- d -glucose (as a model for carbohydrates), substantially increase the quantum yields of peroxyacetic acid Formate also significantly increases the quantum yields of hydrogen peroxide. The highly oxidizing acetylperoxyl radical is proposed as a key intermediate in the photolysis of aqueous biacetyl. The sources and reactions of acetylperoxyl radicals in aqueous aerosols, fogs, and clouds should be investigated in future studies of atmospheric water-drop chemistry.

Quantum yields (Φλ) of hydrogen peroxide (HOOH) photoproduction are reported at 313, 334, and 366 nm for authentic cloud waters collected over a three-year period (1991–1993) at Whiteface Mountain, New York. Quantum yields of HOOH production (based on the total sample absorbance) in these cloud waters ranged up to 0.014. For each sample the HOOH quantum yield decreased with increasing wavelength: Φ313 > Φ334 > Φ366. Action spectra for HOOH photoproduction in these cloud waters indicate that light between 290 and 380 nm is primarily responsible for the aqueous-phase HOOH photoproduction in tropospheric water drops. Quantum yields at 334 and 366 nm were each highly correlated with quantum yields at 313 nm. The cloud water absorbances per centimeter at 334 and 313 nm were also highly correlated. This suggests that similar classes; of chromophores were responsible for the aqueous-phase HOOH photoformation in the cloud waters from different events over a three-year period. The chromophores have not been identified, thus absolute quantum yields for HOOH formation from specific chromophores in an authentic sample could not be calculated. Comparisons of wavelength-dependent relative quantum yields in a given authentic cloud water sample (assuming Fe(oxalate)2− was the primary source of HOOH) with those for Fe(oxalate)2− in well-defined aqueous systems indicate that photolysis of Fe(oxalate)2− was not the primary source of HOOH photoproduction in any of the nine authentic cloud water samples studied. Filtration (0.5 μm Teflon) caused only small differences (⩽15% increase) in rates of HOOH photoformation versus unfiltered controls. Filtration through a 0.05 μm polycarbonate membrane filter did not cause distinguishable differences in the ultraviolet-visible absorption spectra of a cloud water sample previously filtered with a 0.5 μm Teflon filter. These observations suggest that particles were neither the dominant source nor the dominant sink of the photoproduced HOOH. Freezing (up to 403 d) caused only minor changes in HOOH photoproduction rates (−9 to + 12%), and in the ultraviolet-visible absorbance spectra and pH values of the cloud water studied. These observations indicate that freeze-thaw cycles in clouds will have only minor influences on the aqueous-phase photoformation of HOOH.