Secondary Organic Aerosol from Ozonolysis of Biogenic Volatile

Dec 1, 2010 - Assessment of variability of particle-bound reactive oxygen species ... Citation data is made available by participants in Crossref's Ci...
0 downloads 0 Views 1MB Size
Download PDF
Environ. Sci. Technol. 2011, 45, 276–282

Secondary Organic Aerosol from Ozonolysis of Biogenic Volatile Organic Compounds: Chamber Studies of Particle and Reactive Oxygen Species Formation X I C H E N , † P H I L I P K . H O P K E , †,* A N D WILLIAM P. L. CARTER‡ Center for Air Resources Engineering and Science (CARES), Clarkson University, Potsdam, New York 13699-5708, United States and College of Engineering, Center for Environmental Research and Technology (CERT), University of California, Riverside, California 92521, United States

Received June 27, 2010. Revised manuscript received November 11, 2010. Accepted November 15, 2010.

The formation of secondary organic aerosol (SOA) produced from R-pinene, linalool, and limonene by ozonolysis was examined using a dynamic chamber system that allowed the simulation of ventilated indoor environments. Experiments were conducted at typical room temperatures and air exchange rates. Limonene ozonolysis produced the highest SOA mass concentrations and linalool the lowest with R-pinene being intermediate. Simplified empirical modeling simulations were conducted to provide insights into reaction chemistry. Assessment of variability of particle-bound reactive oxygen species (ROS) may be important in the understanding of health effects associated with particulate matter. The ROS intensities defined as ROS/SOA mass were found to be moderately correlated with the SOA densities. Greater ROS intensities were observed for the cases where ozone was in excess. ROS intensities approached a relatively constant value in the region where ozone was in deficit. The estimated initial ROS half-life time was approximately 6.5 h at room temperature suggesting the time sensitivity of ROS measurements. The ROS formed from terpenoid ozonolysis could be separated into three categories: short-lived/high reactive/volatile, semivolatile/relatively stable and nonvolatile/ low reactive species based on ROS measurements under various conditions. Such physical characterization of the ROS in terms of reactivity and volatility provides some insights into the nature of ROS.

1. Introduction The understanding of indoor air pollutants that are regularly inhaled has increased as the emission, transport, formation, and removal pathways have been intensively explored. Indoor oxidation chemistry has been given substantial attention (1, 2). Reactions of compounds such as terpenes that react with ozone have been under extensive study. Such reactions have been suggested to produce a series of products including free radicals, and oxygenated species. Some of the product * Corresponding author phone: 1 315 268 3861; fax: 1 315 268 4410; e-mail: [email protected]. † CARES, Clarkson University. ‡ CERT, UC-Riverside. 276

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 1, 2011

compounds have sufficiently low vapor pressures to nucleate, condense, and contribute to the secondary organic aerosol mass (3-8). The growing use of terpenoid-based solvents driven by the trend toward greener cleaning agents and the use of terpenoid compounds as active ingredients in air fresheners leads to the potential for elevated indoor concentrations of ozone-reactive compounds. Increased use of these household products has been indicated by their increased sales. It is reported that around 70% of U.S homes use air fresheners (1). Moreover, certain so-called “green” or “natural” paints that use terpenoids as a solvent were found to be effective sources of these ozone-reactive chemicals (9). Given the minimal photochemical activity in indoor environments, O3/terpenoid reactions are suggested to be an important pathway to indoor H2O2 formation (10) as well as producing organic peroxide species (11). Reactive oxygen species (ROS) consist of species including free radicals such as OH, HO2, and organic peroxyl radicals, molecules like H2O2 and organic peroxides as well as ions like superoxide anion (12). ROS can be found in both the gas and particle phases. Most of the gas phase ROS have high water solubility and molecular diffusivity that results in their absorption by mucus coating of the upper respiratory tract (13). ROS may produce adverse health effects when they partition into the particulate phase and can then penetrate deeper into the lung than if present in the gas phase (2). ROS-induced pulmonary and systemic oxidative stress has been implicated as an important molecular mechanism of particulate matter-mediated toxicity (14). Studies have implicated the role of endogenously produced ROS in the pathogenesis of pulmonary disease (15). Measurements of particle-bound ROS concentrations are limited (12, 16-19). Moreover, exposure to ROS associated with SOA generated from indoor chemistry under prevailing indoor conditions is largely unknown and not yet characterized. Most collected samples are kept refrigerated until shipped to the laboratory to be analyzed. The time interval between sample collection and analysis conducted may allow samples to undergo decay or re-equilibrium. The bias introduced from such time delays is expected to produce substantial underestimation of the true oxidant concentrations of collected samples. The comparison between fresh and 24 h aged samples in this study provides information to assess the time effect on ROS concentrations. Although many investigators have studied the chemistry of terpenoid/ozone systems, given the complexity of reactive chemistry, there is still limited knowledge regarding exposure to SOA formation at concentrations relevant to the indoor atmosphere. Three terpenoids have been examined under prevailing indoor conditions using Clarkson’s chamber system: R-pinene, linalool, and limonene. Detailed discussions of the individual terpenoid ozonolysis studies were given elsewhere (20-22). While building on previous studies, the aim of this paper was to provide an integrative summary and make previously unexplored comparisons on SOA yields and ROS concentrations as well as evaluating the potential exposure to particles produced under such environments in terms of ROS. In addition, the nature of the SOA will be explored in perspective of ROS in hopes of improving the understanding of ROS constitutes and characteristics. Simplified empirical modeling is included as well to provide useful insights into the chemistry of the reaction systems. 10.1021/es102166c

 2011 American Chemical Society

Published on Web 12/01/2010

2. Experimental Methods 3

Terpenoid ozonolysis was studied in a 2.5 m stainless steel chamber under stirred flow conditions in a temperature controlled room. Details regarding the chamber and experimental protocols were given elsewhere (20) so only a brief description is presented here. Ozone-loaded air from a UV ozone generator (UV550, Ozone Solutions Inc., Sioux Center, IA) was first introduced into chamber and sufficient time was allowed for the concentration inside the chamber to reach steady state. The experiments were conducted under low relative humidity (RH) conditions with the RH around 7%. The low humidity prevented any hygroscopic growth of the resulting SOA species and permitted the study of behavior of the undiluted SOA. The selected terpenoid was delivered either from a modified impinger or a diffusion cell at an approximately constant concentration. The flow rates and temperature were controlled to get specific biogenic VOC (BVOC) concentrations to the chamber. The initial VOC concentrations in the chamber were determined based on GC-FID measurements of VOC at a check point before introducing the flow into the chamber system, along with the VOC flow to dilution air ratios. Muffin fans inside the chamber were used to ensure complete mixing. Differences in particle concentrations were measured to be within 10% for the various positions in the chamber indicating the chamber system was reasonably well-mixed. Thorough cleaning of the chamber was performed for at least 36 h in the presence of ozone prior to each experiment to remove any residue from the prior experiment. The resultant background concentration measured less than 100 particles/ cm3, the volume concentrations was less than 0.1 µm3/cm3 and negligible NOx was determined with concentrations below 5 ppb. The ozone concentration inside the chamber was measured with a UV absorption ozone analyzer (49i, Thermo Fisher Scientific Inc., MA) that sampled air at a rate of 1.5 L/min. Steady state gas phase VOC samples were collected from the sampling port at the face plate planted in the center and analyzed by GC-FID. Gas tight syringes (A2, VICI Precision sampling In. IL) were used to collect BVOC samples. The BVOC samples were analyzed immediately (within 5 min) after collection. Given that the reaction time scale for the BVOC ozonolysis is of the order of hours for the concentration ranges in this study, decomposition of BVOC in the syringe should be negligible. Particle number and size distributions were measured with a Scanning Mobility Particle Sizer (SMPS). A TSI model 3071 DMA coupled with TSI model 3775 condensation particle counter (CPC) operating at 3 L/min and 0.3 L/min for sheath and aerosol flows measured a particle size range from 14 to 700 nm. Size distribution data were recorded and analyzed using TSI AIM software v 8.1. The determination of ROS was performed by using a fluorogenic probe, dichlorofluorescin (DCFH) (12, 18, 19). To catalyze the reaction between DCFH and ROS species, horseradish peroxidase (HRP) was added to the DCFH solution. Fluorescent intensities were converted to equivalent H2O2 concentrations by conducting an assay of standard calibration. Fluorescence intensities were measured with a Turner Quantech Digital Filter Fluorometer (No. FM109535, Barnstead Thermolyne Corp, Dubuque, IA). The excitation and emission wavelengths were 485 and 530 nm, respectively. In general, ROS is an operationally defined quantity determined by the conversion of a nonfluorescent chemical to one that fluoresces. Cho et al. (23) developed an assay for PM redox activity, utilizing the reduction of oxygen by dithiothreitol (DTT) that serves as an electron source. Hasson and Paulson (24) used the peroxidase enzyme catalyzed reaction of hydroperoxides with p-hydroxyphenylacetic acid

(POHPAA) to produce a dimer that fluoresces. Venkatachari and Hopke (25) evaluated a series of commonly used compounds for both their response to specific oxidants and the linearity of response. DCFH had the broadest response to potential oxidants and thus, at this time, appears to be a good nonspecific indicator of oxidation. It was also found that there were losses in ROS when integrated filter sampling was performed relative to continuous monitoring of the ROS (25). Thus, any particular protocol for ROS determinations will produce a result that is dependent on the specifics of the protocol employed. Particle wall deposition rates were estimated as described in detail elsewhere (20). To simplify the calculations, a single wall loss coefficient of 0.08 ( 0.01 h-1 at an air exchange rate (AER) of 0.67 ( 0.01 h-1 was used. All of the resulting SOA mass concentrations were corrected for wall loss. The dominant mechanism for the removal of particles was air exchange under the conditions in this study. Samples for ROS and SOA mass measurements were collected on Teflon filters (Pall, Teflo 25 mm, 3 µm) over 30 min and 1 h intervals, respectively, after the system reached steady state using a flow rate of 23 LPM. The steady state condition was defined in terms of particles, which means stable particle size distributions as well as stable ozone and terpenoid concentrations. It took additional time to attain stable particle size distributions after ozone and terpenoid had achieved steady state. More detailed illustrations of the particle development and evolution are given in the Supporting Information (SI). Samples for SOA mass measurements were collected on preweighed filters and then weighed again after sampling in a clean room to estimate mass concentrations of steadystate SOA. Backup filters were used to determine particle collection efficiency; the measured efficiency was approximately 95% ( 2%. Particle density was determined from SMPS measurements and mass collected onto filters to obtain volume and mass concentrations, respectively. To investigate the effect of a time delay on the measured ROS concentrations, several samples were collected, extracted and measured as quickly as possible for ROS concentrations (within 5 min). Several other samples were stored in sealed Petri dishes that were frozen (around -20 °C) for 24 h. Finally, some of the samples were stored at room temperature (22 ( 2 °C) for 24 h following collection. The initial concentration product of BVOC and ozone for each experiment was set to be approximately 3600 [ppb O3 × ppb BVOC] (the differences are within 5%) for all three terpenoids. Keeping the concentration product constant while varying the concentrations of the two reagents helped to examine the effect of changing the limiting reagent with regards to the number concentrations and SOA mass produced. Experiments were conducted at room temperature (21-24 °C) and an air exchange rate of 0.67 ( 0.01 h-1.

3. Results and Discussion 3.1. Consumed BVOCs at Steady State. SI Table S2 tabulates the reaction constants for the three BVOC species for their reactions with ozone (kO3) and OH radicals (kOH) as well as molar yields of OH radicals from ozonolysis (yOH). Model calculations of the steady state BVOC and SOA values were conducted using the approach described in the SI. Reactions of terpenoids with either ozone or OH radicals are second order reactions. The concentration product of the initial ozone and BVOC concentrations was held constant over each series of experiments. This approach explored the conditions for which each of the reactants was the limiting reagent. Figure 1 shows the consumed biogenic VOC fractions (∆BVOC/BVOC) at steady state from the experimental measurements and simple model calculations. Estimations of the consumed BVOC at steady state using simple model calculations were made both including and excluding the VOL. 45, NO. 1, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

277

FIGURE 1. (a) Plot of experimental consumed BVOC fractions (∆BVOC/BVOC) at steady state with initial BVOC concentrations (initial concentration product of BVOC and ozone has been maintained relatively constant at 3600 ppb × ppb.; (b) comparison of experimental and calculated consumed r-pinene fractions; (c) comparison of experimental and calculated consumed linalool fractions; (d) comparison of experimental and calculated consumed limonene fractions. [OH] term in SI equation S4 because the estimated [OH] values may be highly uncertain. The calculations agree better with measurements in the high BVOC regions, while the model calculations seem to overestimate the concentrations in the low BVOC regions for R-pinene and linalool. Linalool/ozone produced the highest consumed BVOC fraction, R-pinene the lowest and limonene had an intermediate value. This order follows from the second order reaction rate constants for these three compounds. However, when O3/BVOC is higher than 2 (or BVOC < 42 ppb), the curve for ∆limonene fraction is slightly above that of linalool. This difference may be attributed to the higher reaction rate of limonene with OH as well as the greater OH yield with limonene compared to linalool. Considering that the O3/ BVOC >2 region represents low BVOC concentrations, experimental uncertainties are expected to be higher, which is another possibility of observed change in order in BVOC consumptions. 3.2. Steady State SOA. Particle Number, Mass Concentrations. Particle number concentrations at steady state are shown in Figure 2a. Measured particle number concentrations for R-pinene and limonene were relatively constant (differences within 10%) for individual experimental conditions within each experimental group. However, particle numbers were observed to be higher in O3/BVOC < 2 region for linalool. Limonene produced the highest particles number concentrations. These differences among R-pinene, limonene, and linalool can be attributed to presence of endocyclic and exocyclic double bonds. R-Pinene and limonene both possess an endocyclic double bond. The attack by ozone on an endocyclic double bond results in ringopening without changing in length of carbon skeleton. While the attack on exocyclic double bonds leads to severing of the carbon skeleton resulting in reduced carbon number in the largest remaining chain. Attacks on both endo and exo double bonds eventually added polar groups. Linalool is an openchain terpenoid. Thus, the dominant attack on one double bond results in compounds with seven carbons, while if both double bonds are consumed, the major products will have only six carbon atoms (26). Although adding polar groups lowers the vapor pressure, the reduced length of the carbon skeleton leads to an increased vapor pressure. These two 278

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 1, 2011

competing processes play a major role in the amount of SOA produced for the different O3/BVOC combinations and are likely to explain the observed gap between the two clusters in the linalool experiments. Thus, the attack on the second unsaturated bond and the position of the two double bonds play important roles in determining reaction product compositions that then nucleate and/or condense. The measured steady state SOA mass concentrations for the three terpenoids demonstrated that limonene ozonolysis resulted in highest SOA formation. Linalool produced the lowest mass and R-pinene had an intermediate value (Figure 2). SOA mass yields showed the same sequence, detailed calculation and illustration are given in SI Figure S3. To better understand the experimental observations, a simple empirical model was developed to simulate SOA production under various conditions. Descriptions of the model as well as parameters used for different simulation scenarios are presented in SI Tables S1 and S2. Figure 2b-d show the steady state SOA masses from the observations and the empirical model simulations. Observed produced SOA mass patterns and shapes are consistent with the model simulations over the experimental ranges. Four simulations were used to estimate the SOA (see the SI for details): Simulation 1: SOA yields assumed to be the same for both O3 and OH reaction; Simulation 2: SOA assumed to be formed only from O3 reaction; Simulation 3: SOA assumed to be formed only from OH reaction; Simulation 4: same as Simulation 1 except no volatile product formation. For simulations 2 and 3, overall yields were kept same as experiments for R-pinene and linalool except limonene simulation 2. The SOA mass concentrations increase to a maximum with increasing initial BVOC, then level off, and gradually decline in the higher BVOC regions. However, the empirical model appears to be overestimating values for R-pinene and linalool in the high BVOC regions, while for limonene, most of the observations fall within range of the simulation results. As discussed in the SI, the empirical model simulated various scenarios with varying assumptions concerning the relative importance of the O3 and OH reactions for SOA formation as well as for the formation of volatile reactive products that react with OH. However, as shown in Figure 2, the effects of

FIGURE 2. Steady state (a) measured SOA number concentrations; (b) estimated and measured SOA mass concentrations for r-pinene experiments; (c) estimated and measured SOA mass concentrations for linalool experiments; (d) estimated and measured SOA mass concentrations for limonene experiments; SOA mass concentrations are reported as wall-loss corrected; parameters for different simulation scenarios 1-4 are given in SI Table S1. these assumptions were relatively small compared to the differences between experimental and calculated results. As discussed in the prior section, limonene has both endocyclic and exocyclic double bonds; R-pinene possesses only endocyclic double bond, and linalool has an open-chain structure. Leungsakul et al. (4) estimated a ratio of 85%:15% for initial attack of ozone on the endo over the exo double bonds for limonene. A similar preference was suggested by Kwok et al. (27) who found that approximately 63% of OH radical attack is at the endo double bond. Hakola et al. (28) reported e4% molar yield of 4-acetyl-1-methylcyclohexene (product from attack on exocyclic double bond) from limonene ozonolysis. While for the limonene reactions with OH radicals, the molar yields of the two main gas phase products identified as 4-acetyl-1-methylcyclohexene and 3-isopropenyl-6-one-heptanal (product from attack on endocyclic double bond) were estimated to be 0.20 and 0.29, respectively. The dominant attack of ozone on limonene and R-pinene takes place at the endo double bond. It is expected that more SOA mass would be formed in these reactions than from linalool where the attack is at an open-chain double bond. The additional double bond in limonene as compared with R-pinene leads to the higher SOA yields. Further oxidation can take place at the second double bond of limonene leading to more oxidized compounds and formation of second generation ozonolysis products. Further reactions of ozone or OH radicals with the first-generation product containing the additional double bond will preferably occur under conditions with high [O3]/[limonene] ratios (or low initial limonene region, SI Figure S4). This same region is where the measured data points are shifted to agree with simulations representing elevated contributions of SOA mass from OH reactions (Figure 2d). The calculations of fractions of OH reacted with primary reactive volatile products (Figure S4) are given in the SI. Figure 2d also shows that the high

ozone region is where the highest SOA masses were observed. The empirical model simulations are inconsistent with observations in both the point at which the maximum SOA is formed and the slope of declining at high BVOC region. Such inconsistencies may have been introduced in these results because the mechanisms do not include further oxidation of the second double bonds. Another reason for greater SOA mass production from limonene is that the reaction rates of limonene with ozone and OH radicals are significantly higher than those for R-pinene. Ng et al. (29) observed continuous further growth of SOA even after the limonene was completely consumed, suggesting the contribution of second generation products to final total SOA. Hoffmann et al. (30) estimated for limonene photooxidation experiments that approximately 70% of SOA resulted from primary oxidation, and 30% from the second generation products. The observed results suggest that under the conditions where the product of the two reagent initial concentrations was constant, these relative proportions play an important role in determining the total SOA formed. The trend observed in Figure 2d shows that a combination with O3/limonene ratios within the range of 1-2 produces the highest SOA mass concentrations. While for linalool and R-pinene, maximum combinations lie within the region where O3/BVOC is lower than 1. The observed decrease in the low O3 region suggests that for O3/limonene ratios higher than 1, ozone along with OH radicals produced by the reaction system were attacking the two carbon double bonds. This suggestion is supported by observations that were much closer to simulations that accounted for the OH contributions to SOA in the low BVOC region compared to those in the high BVOC region (Figure 2d). These further reactions resulted in second generation products that contributed extra SOA mass. VOL. 45, NO. 1, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

279

FIGURE 4. Correlation of ROS intensities with SOA densities.

FIGURE 3. (a) Fresh steady state ROS intensities and (b) ROS yields (fresh ROS concentration/∆BVOC (mol/m3/mol/m3)) with steady state [OH]. For those experiments with O3/limonene less than 1, ozone and OH radical concentrations would generally be insufficient to attack more than one double bond. Thus, with limited oxidant concentrations, the second generation products account for a much smaller portion of SOA formed. However, the opposite was the case for linalool where the SOA mass curve shows higher mass in low ozone region and lower mass in high ozone region. Such observed behavior implies that further reactions of the second double bond resulted in higher volatility products. The volatility of the shortened carbon skeleton products might not be compensated by the addition of a polar group. 3.3. ROS Intensities and Physical Characterizations. The fresh steady state ROS intensity is defined as ROS intensity )

ROS Concentration SOA Mass Concentration

The fresh ROS intensities (denoted as ROS0/SOA mass concentration, ROS0 refers to the ROS concentrations measured immediately after collection) observed for the three terpenoid species are shown in Figure 3a. The fresh ROS intensities were elevated at the low BVOC conditions (i.e., at the lower SOA mass concentrations). For each chemical system, the ROS intensities gradually approach a similar constant value for the SOA mass formed. The observed shapes of ROS intensities are found to be similar as the variation in densities as a function of SOA mass concentrations (20). The relatively constant ROS intensities suggest a similar composition in terms of the portions of contribution from various functional groups to ROS. ROS yield is defined in the same manner as SOA mass yield, which is the ratio of produced amount of ROS per amount of VOC consumed (nmol/m3/ nmol/m3). Steady state fresh ROS yields were compared to [OH]s.s in Figure 3b. Limonene had the highest ROS yields while linalool produced the lowest. Estimation of steady state [OH] is described in the SI. Examination of ROS yields as a function of [OH]s.s suggest a variable role for free radical reactions in the formation of ROS. It appears that the contribution of hydroxyl radical reactions to ROS formation is lower when higher values of 280

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 1, 2011

SOA are formed. Limonene ROS yields in these experiments appear to be similar and show no relationship with [OH]s.s, except for the case of the lowest SOA mass value (35 µg/m3) that is also the lowest ROS yield point. The relationship of fresh ROS intensities with SOA densities was examined (Figure 4). The correlation coefficient r of 0.42 suggests a moderate correlation. Given the nature of the ROS constituents formed in terpenoid ozonolysis, organic peroxides and hydroperoxides are suspected to be dominant contributors to the ROS intensities (31). The partitioning of semivolatile species between gas and particle phases is described by Pankow (32, 33). As the amount of absorbing material (SOA in the present case) increases, compounds of higher volatility (lower partition coefficient Kp) will increasingly partition into the condensed phase (29, 30, 34). Thus, low volatility species and semivolatile compounds can account for different fractions of the SOA mass for low and high SOA mass production. The more oxidized species would exhibit lower volatility, and higher proportions of such compounds contribute to the increased density. With a higher percentage of more oxidized species, ROS intensities should increase as well. This correlation among the ROS intensities, the fraction of nonvolatile species, and the SOA densities supports the interpretation of the observed moderate connection between SOA densities and ROS intensities. Alfarra et al. (35) found that gas/particle partitioning for SOA was precursor concentration dependent. Their results from SOA produced from photooxidation of R-pinene using different initial concentrations in the presence of NOx showed that low initial precursor concentrations led to SOA dominated by m/z 44 that they attribute to oxidized organic species. For higher initial precursor concentrations, the SOA appeared to be relatively more volatile including species with m/z 43 that is commonly attributed to less oxidized species. At the higher initial precursor concentration, there is more organic matter in the particle phase as well as higher yields. Absorption onto existing particle surfaces can extend to species with higher volatility that would normally be primarily in the gas phase. The variation in ratios for species of different volatility present at different precursor concentrations may lead to the changing densities of the SOA. Their findings were consistent with results from this study. The potential of ROS loss over time and the importance of freshness on sample results were investigated by storing fresh samples in the freezer and at room temperature for 24 h. The resulting losses in measured ROS could be the result of unimolecular decomposition as well as reactions of short-lived radical species with other collected species, and the loss of volatile species over the time interval between sample collection and analysis. The differences in ROS between the samples kept at freezer and at room temperature for 24 h are more likely to result from volatilization losses of semivolatile species although reactions rates would be reduced at the lower temperature. Residual ROS concentra-

FIGURE 5. Chemical and physical characterization of SOA properties with respect to ROS reactivity and volatility. tions at room temperature were expected to arise from low volatility or highly oxygenated species. Thus, the ROS formed can be divided into three lifetime/volatility categories: shortlived/high reactive/volatile (the fraction of ROS that disappeared when samples were stored for 24 h in a freezer as opposed to analyzed immediately after collection), semivolatile/relatively stable (the fraction of ROS that disappeared when samples were stored for 24 h at room temperature as opposed to 24 h in a freezer), and nonvolatile/low reactive species (the residual fraction of ROS when samples were stored at room temperature for 24 h). The percentages of each classification category are shown in Figure 5. Nonvolatile/low reactive and semivolatile/ relatively stable ROS from limonene ozonolysis were estimated to be 24% ( 4% and 66% ( 7%, respectively. The residual ROS fractions were estimated to range from 15% to 69% of the fresh samples, which represents the nonvolatile/ low reactive ROS portion from linalool ozonolysis (21). The fractions of species within each volatility/reactivity category are relatively constant for the limonene ozonolysis SOA compared to those from linalool (21, 22). In Figure 3a, the fresh ROS intensities converge to similar values in the ozonelimited region. Figure 3a shows that in the ozone-limited region, fresh ROS intensities are lower than those in the linalool-limited region. Thus, it would be expected that the lower volatility fractions would be higher in the linalool-limited region. However, low volatility fractions were high at both ends of the concentration range. A possible explanation for the high fraction in ozone-limited region is related to linalool’s properties. It has the lowest vapor pressure of the three compounds studied and can form hydroperoxides by oxidation in air at normal temperatures and pressure. These hydroperoxides are allergenic and behave as sensitizers as contact allergens (36). Storing the collected filter samples exposed to air for 24 h provides the opportunity for autooxidation of linalool, particularly for the high linalool concentrations (ozone-limited region). Venkatachari and Hopke (37) developed a ROS generator that can deliver stable particle-bound ROS concentrations using the R-pinene ozonolysis system. Based on the evaluation of the duration of sampling intervals from their study, 1 and 2 h aged samples lost 15% and 20% of ROS, respectively. The sampling was at room temperature, and the losses of ROS could be due to both reactions of short-lived radicals as well as the loss of volatile species. Short-lived/high reactive/ volatile species comprised about 10% ( 5% of the SOA in the limonene ozonolysis samples. These observations provide insights into the nature of the ROS constituents as well as clearly showing that the ROS characteristics are time sensitive. Ambient samples collected

FIGURE 6. Estimation of initial ROS half-life time from SOA formed from terpenoid ozonolysis. from a roadside and an underpass sampling sites in Taipei showed ROS decay with residual fractions ranging from 27% to 38% based on 1 h fresh samples relative to 115 h aged samples in different size intervals (18). Figure 6 provides a rough estimation of the ROS half-life time with SOA generated from terpenoid ozonolysis. ROS entirely from SOA showed an initial half-life of approximately 6.5 h and a longer lifetime after the initial decay period at room temperatures. Antonini et al. (16) investigated fresh and aged welding fumes and found the ROS of welding fumes decreasing exponentially with a half-life of 10 days. Compared with welding fumes, ROS from SOA exhibits a much shorter lifetime, suggesting that reactions of short-lived radical species along with loss of volatile species over the time were significant. It is not surprising given welding fumes are mostly metal oxides, releasing of internal radicals that contribute to ROS is a relatively slow process. The ROS time effect with respect to loss of short-lived and volatile species demonstrates the importance of the freshness of the samples for accurate ROS concentration measurements. The time interval between sample collection and analysis is critical for accurate ROS measurements.

4. Implications Indoor use of any terpenoid-containing household products may have desirable benefits. However, there are potential problems because of the significant SOA formation initiated by terpenoid ozonolysis. This study of terpenoid ozonolysis under realistic indoor conditions supports the need for concern. ROS may be an important particulate matter species with respect to health effects. However, very little information is currently available on the chemical characteristics of ROS formed in such reactions. The physical characterization of the fractions possessing different volatility/reactivity helps elucidate the nature of the ROS constituents. A rough VOL. 45, NO. 1, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

281

estimation of ROS initial half-life time yields a value of approximately 6.5 h. The measured short half-lives and volatility of these species demonstrates the importance of rapid sample analysis for ROS concentrations.

Acknowledgments This work was supported by U.S. Environmental Protection Agency under cooperative agreement X-83269001-0 through a subcontract from Syracuse University. Although the research described in this article has been funded wholly or in part by the United States Environmental Protection Agency, it has not been subjected to the Agency’s required peer and policy review and therefore, does not necessarily reflect the views of the Agency and no official endorsement should be inferred.

Supporting Information Available Figures of time series of particle development, SOA mass yields, and fractions of OH reacted with primary volatile products. Detailed illustration on steady state [OH] calculations and simple empirical modeling to estimate SOA produced. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Weschler, C. J. Changes in indoor pollutants since the 1950s. Atmos. Environ. 2009, 43, 153–169. (2) Weschler, C. J. Ozone’s Impact on Public Health: Contributions from indoor exposures to ozone and products of ozone-initiated chemistry. Environ. Health Perspect. 2006, 114, 1489–1496. (3) Fan, Z.; Lioy, P.; Weschler, C.; Fiedler, N.; Kipen, H.; Zhang, J. Ozone initiated reactions with mixtures of volatile organic compounds under simulated indoor conditions. Environ. Sci. Technol. 2003, 37, 1811–1821. (4) Leungskul, S.; Jaoui, M.; Kamens, R. M. Kinetic mechanism for predicting secondary organic aerosol formation from the reaction of d-limonene with ozone. Environ. Sci. Technol. 2005, 39, 9583–9594. (5) Sarwar, G.; Corsi, R. The effects of ozone/limonene reactions on indoor secondary organic aerosols. Atmos. Environ. 2007, 41, 959–973. (6) Weschler, C. J.; Shields, H. C. Experimental probing the influence of air exchange rates on secondary organic aerosols derived from indoor chemistry. Atmos. Environ. 2003, 37, 5621–5631. (7) Singer, B. C.; Coleman, B. K.; Destaillats, H.; Hodgson, A. T.; Lunden, M. M.; Weschler, C. J.; Nazaroff, W. W. Indoor secondary pollutants from cleaning product and air freshener use in the presence of ozone. Atmos. Environ. 2006, 40, 6696–6710. (8) Destaillats, H.; Lunden, M. M.; Singer, B. C.; Coleman, B. K.; Hodgson, A. T.; Weschler, C. J.; Nazaroff, W. W. Indoor secondary pollutants from household product emissions in the presence of ozone: A bench-scale chamber study. Environ. Sci. Technol. 2006, 40, 4421–4428. (9) Toftum, J.; Freund, S.; Salthammer, T.; Weschler, C. J. Secondary organic aerosols from ozone-initiated reactions with emissions from wood-based materials and a “green” paint. Atmos. Environ. 2008, 42, 7632–7640. (10) Fan, Z.; Weschler, C. J.; Han, I.; Zhang, J. Co-formation of hydroperoxides and ultra-fine particles during the reactions of ozone with a complex VOC mixture under simulated indoor conditions. Atmos. Environ. 2005, 39, 6171–5182. (11) Docherty, K. S.; Wu, W.; Lim, Y. B.; Ziemann, P. J. Contributions of organic peroxides to secondary aerosol formed from reactions of monoterpenes with O3. Environ. Sci. Technol. 2005, 39, 4049– 4059. (12) Venkatachari, P.; Hopke, P. K.; Brune, W. H.; Ren, X.; Lesher, R.; Mao, J.; Mitchell, M. Characterization of wintertime reactive oxygen species concentrations in Flushing, New York. Aerosol Sci. Technol. 2007, 41, 97–111. (13) Friedlander, S. K.; Yeh, E. K. The submicron atmospheric aerosol as a carrier of reactive chemical species: Case of peroxides. Appl. Occup. Environ. Hyg. 1998, 13, 416–420. (14) Gurgueira, S. A.; Lawrence, J.; Coull, B.; Murthy, G. G. K.; Gonzalez-Flecha, B. Rapid increases in the steady state concentration of reactive oxygen species in the lungs and heart after particulate air pollution inhalation. Environ. Health Perspect. 2002, 110, 749–755.

282

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 1, 2011

(15) Kehrer, J. P. Free radicals as mediators of tissue injury and disease. Crit. Rev. Toxicol. 1993, 23, 21–48. (16) Antonini, J. M.; Clarke, R. W.; Krishna Murthy, G. G.; Sreekanthan, P.; Jenkins, N.; Eagar, T. W.; Brain, J. D. Freshly generated stainless steel welding fume induces greater lung inflammation in rats as compared to aged fume. Toxicol. Lett. 1998, 98, 77–86. (17) Hasson, A. S.; Paulson, S. E. An investigation of the relationship between gas-phase and aerosol-bourne hydroperoxides in urban air. J. Aerosol Sci. 2003, 34, 459–468. (18) Hung, H.-F.; Wang, C.-S. Experimental determination of reactive oxygen species in Taipei aerosols. J. Aerosol Sci. 2001, 32, 1201– 1211. (19) Venkatachari, P.; Hopke, P. K.; Grover, B. D.; Eatough, D. J. Measurement of particle-bound reactive oxygen species in Rubidoux aerosols. J. Atmos. Chem. 2005, 50, 49–58. (20) Chen, X.; Hopke, P. K. Secondary organic aerosol from a´-pinene ozonolysis in dynamic chamber system. Indoor Air 2009, 19, 335–345. (21) Chen, X.; Hopke, P. K. A chamber study of secondary organic aerosol formation by linalool ozonolysis. Atmos. Environ. 2009, 43, 3935–3940. (22) Chen, X.; Hopke, P. K. A chamber study of secondary organic aerosol formation by limonene ozonolysis. Indoor Air 2010, 20, 320–328. (23) Cho, A. K.; Sioutas, C.; Miguel, A. H.; Kumagai, Y.; Schmitz, D. A.; Singh, M.; Eiguren-Fernandez, A.; Froines, J. R. Redox activity of airborne particulate matter at different sites in the Los Angeles Basin. Environ. Res. 2005, 99, 40–47. (24) Hasson, A. S.; Paulson, S. E. An investigation of the relationship between gas-phase and aerosol-bourne hydroperoxides in urban air. J. Aerosol Sci. 2003, 34, 459–468. (25) Venkatachari, P.; Hopke, P. K. Development and Laboratory Testing of an Automated Monitor for the Measurement of Atmospheric Particle-bound Reactive Oxygen Species (ROS). Aerosol Sci. Technol. 2008, 42, 629–635. (26) Shu, Y.; Kwok, E. S. C.; Tuazon, E. C.; Atkinson, R.; Arey, J. Products of the gas- phase reactions of linalool with OH radicals, NO3 radicals, and O3. Environ. Sci. Technol. 1997, 31, 896–904. (27) Kwok, E. S.; Atkinson, R. Estimation of hydroxyl radical reaction rate constants for gas-phase organic compounds using a structure-reactivity relationship: an update. Atmos. Environ. 1995, 29, 1685–1695. (28) Hakola, H.; Arey, J.; Aschmann, S. M.; Atkinson, R. Products formation from the gas-phase reactions of OH radicals and O3 with a series of monoterpenes. J. Atmos. Chem. 1994, 18, 75– 102. (29) Ng, N. L.; Kroll, J. H.; Keywood, M. D.; Bahreini, R.; Varubangkul, V.; Flagan, R. C.; Seinfeld, J. H. Contribution of first versus second generation products to secondary organic aerosols formed in the oxidation of biogenic hydrocarbons. Environ. Sci. Technol. 2006, 40, 2283–2297. (30) Hoffmann, T.; Odum, J. R.; Bowman, F.; Collins, D.; Klockow, D.; Flagan, R. C.; Seinfeld, J. H. Formation of organic aerosols from the oxidation of biogenic hydrocarbons. J. Atmos. Chem. 1997, 26, 189–222. (31) Venkatachari, P.; Hopke, P. K. Characterization of products formed in the reaction of ozone with R-pinene: Case for organic peroxides. J. Environ. Monit. 2008, 10, 966–974. (32) Pankow, J. F. An absorption model of gas-particle partitioning of organiccompounds in the atmosphere. Atmos. Environ. 1994, 28, 185–188. (33) Pankow, J. F. An absorption model of gas aerosol partitioning involved in the formation of secondary organic aerosol. Atmos. Environ. 1994, 28, 189–193. (34) Kroll, J. H.; Seinfeld, J. H. Chemistry of secondary organic aerosol: Formation and evolution of low-volatility organics in the atmosphere. Atmos. Environ. 2008, 42, 3593–3624. (35) Alfarra, M. R., Prevot, A. S. H., Duplissy, J., Metzger, A., Lanz, V. A., Hueglin, C., Dommen, J., Weingartner, E., Baltensperger, U. Oxygenated organic aerosols: bridging field and smog chamber observations Proceedings of European Aerosol Conference, T01A026, Salzburg, September 9-14, 2007. (36) Ba¨cktorp, G.; Tobias Johanson Wass, J. R.; Panas, I.; Sko¨ld, M.; Bo¨rje, A.; Nyman, G. Theoretical investigation of linalool oxidation. J. Phys. Chem. A 2006, 110, 12204–12212. (37) Venkatachari, P.; Hopke, P. K. Development and evaluation of a particle-bound reactive oxygen species generator. J. Aerosol Sci. 2008, 39, 168–174.

ES102166C