Significant Contributions of Isoprene to Summertime Secondary

Jul 11, 2017 - during TexAQS 200049 and on top of the Moody Tower in the. University of Houston campus during TexAQS 200650 (Figure. 3). At La Porte, ...
0 downloads 0 Views 5MB Size
Download PDF
Article pubs.acs.org/est

Significant Contributions of Isoprene to Summertime Secondary Organic Aerosol in Eastern United States Qi Ying,* Jingyi Li,† and Sri Harsha Kota‡ Zachry Department of Civil Engineering, Texas A & M University, College Station, Texas 77843, United States S Supporting Information *

ABSTRACT: A modified SAPRC-11 (S11) photochemical mechanism with more detailed treatment of isoprene oxidation chemistry and additional secondary organic aerosol (SOA) formation through surface-controlled reactive uptake of dicarbonyls, isoprene epoxydiol and methacrylic acid epoxide was incorporated in the Community Multiscale Air Quality Model (CMAQ) to quantitatively determine contributions of isoprene to summertime ambient SOA concentrations in the eastern United States. The modified model utilizes a precursor-origin resolved approach to determine secondary glyoxal and methylglyoxal produced by oxidation of isoprene and other major volatile organic compounds (VOCs). Predicted OC concentrations show good agreement with field measurements without significant bias (MFB ∼ 0.07 and MFE ∼ 0.50), and predicted SOA reproduces observed day-to-day and diurnal variation of Oxygenated Organic Aerosol (OOA) determined by an aerosol mass spectrometer (AMS) at two locations in Houston, Texas. On average, isoprene SOA accounts for 55.5% of total predicted near-surface SOA in the eastern U.S., followed by aromatic compounds (13.2%), sesquiterpenes (13.0%) and monoterpenes (10.9%). Aerosol surface uptake of isoprene-generated glyoxal, methylglyoxal and epoxydiol accounts for approximately 83% of total isoprene SOA or more than 45% of total SOA. A domain wide reduction of NOx emissions by 40% leads to a slight decrease of domain average SOA by 3.6% and isoprene SOA by approximately 2.6%. Although most of the isoprene SOA component concentrations are decreased, SOA from isoprene epoxydiol is increased by ∼16%.

■

and Lin et al.17 reported that aerosol surface uptake of isoprene epoxydiol (IEPOX) and methacrylic acid epoxide (MAE), which are formed from isoprene photooxidation under lowNOx and high-NOx conditions respectively, followed by acid catalyzed reactions in the aqueous phase can be important isoprene SOA formation pathways. While recent modeling studies have been carried out to evaluate the importance of these additional pathways on predicted SOA concentrations,18−20 a full evaluation of the contributions of isoprene to SOA that includes all these new pathways has not been performed. Karambelas et al.18 and Pye et al.17,20 were focused on IEPOX and MAE, respectively. Lin et al.19 and Li et al.21 studied SOA formation from dicarbonyls and isoprene epoxides. However, the contributions of different precursors to dicarbonyls were not determined, preventing a full evaluation of the contribution of isoprene to overall SOA budget. In this study, we applied a photochemical mechanism that incorporates the most recent updates on the gas phase isoprene oxidation pathways as well as SOA formation from aerosol

INTRODUCTION Isoprene, the most abundant volatile organic compound (VOC) emitted from biosphere into the atmosphere,1 is extremely reactive and produces tropospheric ozone2 and secondary organic aerosol (SOA).3,4 The contribution of isoprene to ambient SOA has been evaluated in a number of studies using regional and global chemical transport models (CTMs).5−7 Most of these models are based on the equilibrium adsorption partitioning theory of Pankow8 and laboratorymeasured isoprene SOA yields3,9 fitted to a two-product parametrization that was initially proposed by Odum et al.10 While the models suggested that isoprene plays an important role in global and regional SOA budget, predicted concentrations of organic carbon (OC) in the summer were systematically biased low.6,11 Recently, heterogeneous pathways for the formation of SOA from isoprene oxidation products have been investigated. Lim et al.12 proposed that in-cloud processing of isoprene oxidation products glyoxal (GLY), methylglyoxal (MGLY) and glycoaldehyde can lead to formation of organic acids. Further, it was found that GLY and MGLY also lead to significant SOA formation in the aqueous phase of aerosols.13,14 Clayes et al.15 suggested that 2-methyltetrols formed in the gas phase photooxidation of isoprene can be a significant source of SOA due to their low saturation vapor pressure. Paulot et al.16 © 2015 American Chemical Society

Received: Revised: Accepted: Published: 7834

December 11, 2014 May 30, 2015 June 1, 2015 June 1, 2015 DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842

Article

Environmental Science & Technology

isoprene epoxide uptake coefficient calculation was calculated using ISORROPIA (v1.7),29 which is the default inorganic thermal dynamics module in AERO5. The original AERO5 in CMAQ treats the oligomerization of semivolatile SOA components using a simple first order decay approach.30 To attribute oligomers to the precursor VOC species, seven oligomer model species were used: AOLGA_A (oligomers from semivolatile long-chain alkane SOA), AOLGA_B (benzene), AOLGA_T (ARO1), AOLGA_X (ARO2), AOLGB_T1 (TERP), AOLGB_T2 (SESQ), and AOLGB_I (ISOP). SOA Formation from Reactive Aerosol Surface Uptake. Reactive uptake of volatile isoprene oxidation products GLY, MGLY, IEPOX, and MAE into the aqueous phase can contribute significantly to SOA formation.17,31−34 In this study, formation of SOA in the aqueous phase of the aerosols due to IEPOX, MAE, GLY, and MGLY are modeled as irreversible surface-controlled uptake processes, as shown in eq 1:

surface uptake processes of GLY, MGLY, IEPOX, and MAE, in addition to the traditional 2-product equilibrium partitioning approach, to study SOA formation. A source-oriented version of the photochemical mechanism was used to track GLY, MGLY and oligomers from different primary precursor species. The model was applied to study isoprene SOA formation in the eastern United States, which is chosen based on the prevalence of acidic aerosols due to significant emissions of NOx and SO2 from fossil fuel combustion sources,22 thus could have significant amount of SOA formation from the acidic-catalyzed pathways. This is the first study to quantitatively determine the contributions to SOA from isoprene and other major biogenic and anthropogenic precursors by accurately tracking their key oxidation products and SOA components using a sourceoriented three-dimensional regional transport model.

■

MATERIALS AND METHODS Modified SAPRC-11 Photochemical Mechanism. A modified SAPRC-11 (S11) photochemical mechanism23 was used in this study to allow predictions of IEPOX and MAE, as well as concentrations and precursor-origin of GLY and MGLY so that SOA formation from isoprene can be determined directly. The S11 photochemical mechanism is an improvement of the widely used SAPRC-07 photochemical mechanism24,25 to provide better predictions of the oxidation of aromatic hydrocarbons. In this study, S11 with standard lumping of VOCs was modified to include more detailed treatment of the isoprene oxidation chemistry as used by Lin et al.17 The detailed isoprene scheme is largely based on the isoprene scheme described by Xie et al.,26 which includes the formation of IEPOX suggested by Paulot et al.16 as well as more detailed treatment of isoprene nitrate chemistry. The Lin et al.17 isoprene scheme also includes a number of reactions that describe the formation of MAE and hydroxymethyl-methyl-αlactone (HMML) from the oxidation of methacyloyl-peroxy nitrate (MPAN). Loss of epoxides from photolysis, such as that reported by Bates et al.,27 was not included in the current mechanism. The modified S11 mechanism was further expanded to track GLY and MGLY from major groups of precursors separately using precursor-tagged species. Six groups of precursors were tracked in this study: (1) aromatics with OH reaction rates less than 2 × 104 ppm−1 min−1 (ARO1), (2) aromatics with OH reaction rates greater than 2 × 104 ppm−1 min−1 (ARO2), (3) isoprene (ISOP), (4) monoterpenes (TERP), (5) sesquiterpenes (SESQ), and (6) all other primary VOC precursors and direct emissions of GLY and MGLY. For example, GLY_I and GLY_T1 are used to represent GLY formed from isoprene and monoterpenes, respectively. In addition to the precursor-tagged GLY and MGLY species, oxidation products that lead to GLY and MGLY formation from different groups of precursors are also represented using tagged species to determine their precursor-origin. For example, acetaldehyde (CCHO) is an oxidation product that can be further oxidized to form glyoxal. In the updated mechanism, CCHO_I1 and CCHO_T1 are used to represent CCHO from ISOP and TERP, respectively. In this way, precursor-origin of second and later generations of GLY and MGLY can be determined. The modified S11 mechanism was implemented in the Community Multiscale Air Quality Model (CMAQ)11,28 version 5.0.1 and linked with the aerosol module version 5 (AERO5) in CMAQ. The molality of hydrogen ion used in

dMg, i dt

1 = − γivAM i g, i 4

(1)

where Mg is the species concentration (μg m−3) in the gas phase, A is the total aerosol surface area (m2), γ is the reactive uptake coefficient, v is the thermal velocity of the gas molecule (m s−1), and subscript i is the index of the species that undergo irreversible heterogeneous reactions. The increase of the aqueous phase SOA concentration is calculated by assuming mass of the products from precursor is conserved and the heterogeneous reaction products are nonvolatile. As the reactive uptake of GLY and MGLY can be reversible35 and reactions of GLY and MGLY with oxidants in the aqueous phase can produce low volatility products,14,36 the simple irreversible mass conservation treatment could lead to biases in the predicted SOA concentrations and should be evaluated in a future study.. The reactive uptake coefficient for IEPOX (γisoepox) is calculated using an aerosol acidity dependent relationship based on several uptake experiments for isoprene epoxides,31,37 as shown in eq 2: ln γisoepox = 0.01446(ln m H+)2 + 0.60394ln m H+ − 7.46325 (2)

where mH+ is the molality of the hydrogen ion in the aqueous phase. More details of this acidity dependent reactive uptake coefficient can be found in Li et al.21 In this study, the predicted γisoepox is approximately 4−6 × 10−4 in most areas and highest values of ∼1.4 × 10−3 occur over the ocean, where mH+ is the highest (see Supporting Information Figure S1). In general, these values are consistent with the γisoepox values reported by Pye et al.,20 which were based on more mechanistic calculations. Although a recent study shows the uptake coefficient of MAE is likely different from that of IEPOX,38 the same γisoepox value was used for MAE in this study due to lacking of additional published data on the uptake coefficient of MAE. Uptake coefficients for GLY and MGLY are taken as 2.9 × 10−3, as used by Fu et al.39 This is slightly lower than the average γglyoxal (8 × 10−3) reported by Ervens and Volkamer for daytime SOA formation on hygroscopic aerosols.14 At night, uptake of GLY and MGLY is assumed to follow a surface area independent reaction with a first order reaction rate constant of 3.33 × 10−4 s−1.14,40 Reactive surface uptake of these species is 7835

DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842

Article

Environmental Science & Technology

Figure 1. Schematic of SOA formation pathways. Black arrows represent SOA formation pathways in the original CMAQ AERO5. Red arrows represent additional pathways17,21 included in the study.

SPECIATE emission profile databases, and processed using the speciation database program provided by Carter.45 The meteorology fields for both episodes were generated by the Texas Commission of Environmental Quality (TCEQ) using the PSU/NCAR mesoscale model (MM5) and were shown to reproduce the observed meteorological conditions.46

allowed to occur on wet particles. Aerosol liquid water is estimated using the ZSR method. The traditional 2-product approach is retained as an estimation of the semivolatile organic-phase isoprene production. The original CMAQ isoprene mechanism also contains an acid enhanced pathway, which generates nonvolatile SOA. This pathway is disabled to avoid double counting of the nonvolatile SOA generated from the aerosol surface uptake described above. A schematic of the SOA formation pathways used in the current study is shown in Figure 1. Model Setup. The modified S11 mechanism with the updated SOA formation pathways were used to simulate regional SOA formation in the eastern U.S. during two Texas Air Quality Study (TexAQS) episodes, from August 16 to September 7, 2000 and August 28 to September 16, 2006 using a 36-km horizontal resolution domain (see Figure 4). Both episodes have been extensively studied and detailed descriptions of the domain setup, and preparation of emissions and meteorological inputs can be found in previous publications41−43 and the references therein. Only a brief summary of the model inputs is provided below. Anthropogenic emissions were based on the 2001 Clean Air Interstate Rules (CAIR) and 2005 National Emission Inventory (NEI), and were processed using the Sparse Matrix Operator Kernel Emission (SMOKE) model. Biogenic emissions (excluding open biomass burning) were generated using Biogenic Emission Inventory Version 3 (BEIS3).44 Speciation of VOC into model species was based on profiles extracted from the

■

RESULTS Model Performance Evaluation. Ozone directly affects SOA formation and it is a good metric for successful gas-phase

Figure 2. Comparison of predicted and observed daily average organic carbon (OC) concentrations at all monitors within the model domain. (a) Points are color-coded by the fraction of secondary organic carbon; (b) Predicted OC based on standard CMAQ AERO5 SOA module (without reactive aerosol surface uptake products). 7836

DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842

Article

Environmental Science & Technology

Figure 3. Comparison of (a) predicted hourly PM2.5 POA with AMS HOA+BBOA, (b) predicted PM2.5 SOA with AMS OOA at La Porte, TX during TexAQS 2000, and (c) predicted hourly PM2.5 POA with AMS HOA and (d) predicted PM2.5 SOA with AMS OOA at the Moody tower. Units are μg m−3. Time is Central Standard Time (CST). IEPOX, MAE, GLY, and MGLY represent SOA from reactive aerosol surface uptake of these species.

1.28) during the 2000 and 2006 model episodes, respectively. All raw observations of OC for 2000 and 2006 were uniformly decreased by 0.79 and 1.17 μgC m−3, respectively, before they were compared with model predictions. As shown in Figure 2(a), the predicted and observed OC concentrations are generally within a factor of 2 in this study. The mean fractional bias (MFB) and mean fractional error (MFE) are 0.09 and 0.53 (for 2000) and 0.06 and 0.48 (for 2006), respectively, based on a total of 81 and 446 prediction-observation pairs for 2000 and 2006, respectively. There are no significant biases in the predicted OC, as can be seen from the small MFB values. The predictions were colored by SOC/OC ratio. In general, higher OC concentrations are associated with higher fractions of SOC. This suggests that SOC concentrations were reasonably predicted, even though they are not directly measured. Figure 2(b) shows that predicted OC concentrations are significantly biased low if contributions to SOA due to surface reactive uptake were not included (MFB = −0.30 and −0.35, and MFE = 0.54 and 0.59 for 2000 and 2006, respectively), suggesting that they account for a significant fraction of the predicted SOA at most locations and times. In addition to total OC, predicted hourly PM2.5 SOA and POA were also compared with hydrocarbon-like organic aerosol (HOA) and oxidized organic aerosol (OOA) measured by an Aerosol Mass Spectrometer (AMS) at La Porte, Texas during TexAQS 200049 and on top of the Moody Tower in the University of Houston campus during TexAQS 200650 (Figure 3). At La Porte, a three-factor PMF solution was obtained, and the third factor was determined to be primary emissions due to biomass burning organic aerosol (BBOA). The BBOA factor was combined with the HOA factor to compare with predicted POA. In general, predicted POA agrees with the temporal variation of AMS HOA during both years (MFB = 0.15 and 0.36, for TexAQS 2000 and 2006, respectively; and MFE = 0.72 and 0.57, respectively). Under-prediction of POA on September 5−6, 2000 is likely due to under-estimation of wildfire emissions as BBOA concentrations are much higher than HOA concentrations in these days.51 Predicted diurnal and day-to-day variations of SOA also generally agree with the AMS OOA values. It should be noted that while the model

mechanism behavior. Observed ozone concentrations were retrieved from the Air Quality System maintained by the U.S. EPA (AQS, available at http://www.epa.gov/ttn/airs/airsaqs/ detaildata/downloadaqsdata.htm). The modified S11 mechanism can successfully capture the diurnal and day-to-day variation of the observed ozone pattern in general. The overall model performance statistics for 1 h and 8 h ozone using data at all stations are given in Supporting Information Table S1. Predicted isoprene concentrations were compared with hourly observations measured at 13 ground based stations using automatic gas chromatography (AutoGC) instruments operated by TCEQ throughout Texas during TexAQS 2006. A list of the AutoGC sites can be found in Supporting Information Table S2. Six AutoGC sites in Houston and 2 sites in Corpus Christi are located in the same 36 km grid cells. The observed concentrations at these sites in the same grid cell were averaged before being used to compare with observations. Supporting Information Figure S2 shows that model predictions generally captured the spatial and diurnal variations, although concentrations at the Dallas Hinton and Danciger sites were under-predicted. Predicted daily PM2.5 OC concentrations in 2000 and 2006 were compared with observations at all monitoring sites within the AQS database. Supporting Information Figure S3 shows the location of the stations with available data. OC concentrations were measured at these sites once every 3 or 6 days. Primary organic carbon (POC) was directly predicted in CMAQ. Emission rates of POC used in the simulations were converted from emission rates of primary organic aerosol (POA) using an organic aerosol (OA) to OC ratio of 1.43.47 Predicted SOA was converted to secondary organic carbon (SOC) using SOA/ SOC ratios used in the CMAQ model (see Supporting Information Table S3), which are based on suggestions by Kleindienst et al.48 The raw OC observations were blankcorrected using an average blank-OC value based on all blank measurements at the OC monitoring sites within the domain during the study periods. As shown in Supporting Information Figure S4, the PM2.5 OC values in blank samples are mostly within 0.5 and 1.0 μgC m−3 (35 samples; mean = 0.79, standard deviation = 0.55, median = 0.77) and 1.0 and 1.4 μgC m−3 (170 samples; mean = 1.17, standard deviation = 0.71, median = 7837

DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842

Article

Environmental Science & Technology

Figure 5. Total isoprene SOA (a) and SOA formation from isoprene generated GLY+MGLY (b), IEPOX+MAE (c), oligomers (d), and semivolatile products (e), between August 28 and September 15, 2006. Units are μg m−3.

contributions to GLY/MGLY SOA from precursors ARO1+ARO2 (4(b,g)), ISOP (4(c,h)), TERP+SESQ (4(d,i)) and other precursors (4(e,j)) for the TexAQS 2006 episode. Results for the TexAQS 2000 are included in the Supporting Information (Figures S5 and S6). Although the spatial distributions are different from those from 2006, the domainwide relative contributions do not differ significantly (see Supporting Information Table S4). The spatial distributions of GLY and MGLY SOA are similar to highest concentrations occur in the southeast U.S., and MGLY SOA (peak concentration ∼2 μg m−3) is approximate two times higher than GLY SOA. The spatial distribution of GLY and MGLY SOA due to aromatic compounds is different than those due to biogenic monoterpenes and sesquiterpenes. Contributions of aromatic compounds to GLY and MGLY SOA are similar, ranging from 0.05 to 0.25 μg m −3 (approximately 15%). Higher concentrations occur near major urban areas where significant emissions of aromatic compounds are expected from anthropogenic combustion sources. Isoprene is a significant contributor to GLY SOA (∼48%) and the dominant contributor to MGLY SOA (∼82%). This agrees well with estimations by Fu et al.53 that isoprene oxidation accounts for 47% of gas phase glyoxal and 79% of methylglyoxal globally, suggesting that the precursor attribution mechanism is working as expected. Contributions of other biogenic terpenes to GLY and MGLY SOA are small (∼0.5%). Other precursors and direct emissions account for approximate 35% of GLY SOA, but are not significant contributors of MGLY SOA (∼4%). With appropriate attribution of GLY and MGLY to isoprene, as shown in Figure 4(c) and 4(h), it is now possible to determine the concentration of total isoprene SOA and contributions to isoprene SOA due to surface uptake of GLY, MGLY and IEPOX, equilibrium partitioning of semivolatile

Figure 4. Total glyoxal (GLY) (a) and methylglyxoal (MGLY) (f) SOA, and GLY/MGLY SOA formation from aromatic compounds (b,g), isoprene (c,h), terpenes (d,i) and other precursors (e,j), between August 28 and September 15, 2006. Units are μg m−3. Overlaid on panel (a) is the episode average surface wind speed. The spatial distribution is largely determined by the wind field.

predictions are for PM2.5, AMS measurements are mostly sensitive to particles less than 1.0 μm in diameter.52 Precursor Contributions to Isoprene SOA. Figure 3(b) shows that at La Porte the predicted SOA from irreversible uptake of IEPOX, GLY, and MGLY accounts for 20% (hourly contribution ranges from 0 to 30%), 9% (6−22%) and 19% (14−38%) of episode average SOA, whereas semivolatile components and oligomers account for 23% (9−65%) and 27% (3−40%) of episode average SOA, respectively. Contributions of the SOA components to episode average SOA at the Moody Tower during TexAQS 2006, as shown in Figure 3(d) are similar: IEPOX 19% (2−35%), GLY 11% (8−22%), MGLY 23% (16−40%), semivolatile 29% (14−51%) and oligomers 17% (4−26%). Contributions of MAE to total SOA are low, with an average of 1% at both locations and range from 0 to 3%. Figure 4(a)−(e) and (f)−(j) show episode-average regional distribution of GLY (4(a)) and MGLY SOA (4(b)) concentrations from reactive aerosol surface uptake, and 7838

DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842

Article

Environmental Science & Technology

Figure 6. Predicted (a) average SOA concentrations (maximum = 7.40 μg m−3) and fractional contributions to predicted SOA due to (b) isoprene (65.6%), (c) monoterpenes (20.6%), (d) sesquiterpene (20.2%), (e) aromatics (37.4%), and (f) other precursors, between August 28 and September 15, 2006.

terpene and sesquiterpene SOA are different and generally follow the spatial distribution of the emissions of the precursors in the southeastern U.S. Contributions of aromatic compounds can be as high as 40% in the urban areas, especially in the northeast and in part of Florida. On average, they account for approximately 15% of total SOA. Contributions of other precursors (through formation of GLY and MGLY) account for ∼10% of predicted SOA. Sensitivity of Isoprene SOA to NOx/SO2 Emission Reductions. Formation of biogenic SOA is affected by anthropogenic emissions. A sensitivity simulation with 40% NOx emission reduction was performed for the August− September 2006 episode. In most areas in the southeastern U.S., this led to a net reduction of isoprene SOA by 10−15%, although some increases were predicted in the north and west part of the domain (see Figure 7(a)). Domain average isoprene SOA decreases by approximately 3%. Figure 7(b)−(f) show the regional changes of each SOA component. SOA from isoprene generated GLY and MGLY is expected to reduce throughout the domain, with higher reduction of up to 0.45 μg m−3 in the southeastern U.S. Interestingly, yet not unexpected, IEPOX SOA increased throughout the domain by up to 0.2 μg m−3 (or ∼12%), as lower NOx favors the formation of gas phase IEPOX. Similarly, reduction of NOx led to decrease of MAE, and thus MAE SOA throughout the domain. Semivolatile SOA components and oligomers decreased slightly. These results are in general agreement with the 25% NOx reduction simulation reported by Pye et al.20 An additional simulation with 40% SO2 reduction predicted a 20% reduction of IEPOX SOA and 17% reduction of MAE SOA but negligible amount of semivolatile SOA. These results are also in general agreement with Pye et al.,20 which predicted a 30−35% reduction of IEPOX SOA and 35−40% reduction of MAE SOA due to a 25% reduction of SO2 emissions.

Figure 7. Changes in the predicted episode average (a) total isoprene SOA and isoprene SOA components, (b) GLY+MGLY SOA, (c) IEPOX SOA, (d) MAE SOA, (e) semivolatile SOA, and (f) oligomers, between August 28 and September 15, 2006 for a sensitivity simulation case with 40% NOx emission reduction. Units are μg m−3. Difference is calculated by sensitivity case minus base case.

components as well as oligomers. Episode average concentrations of isoprene SOA were high in the southeastern U.S., with a peak concentration of ∼4 μg m−3 (Figure 5(a)). SOA from GLY+MGLY aerosol surface uptake pathway (Figure 5(b)) accounts for approximately 50% of the total isoprene SOA, ranging from 0.5 to 2.5 μg m−3. SOA from IEPOX+MAE (Figure 5(c)) also accounts for a significant fraction of isoprene SOA, with a peak concentration of 1.4 μg m−3 and a domain average contribution of 34% (MAE 2.5%). Semivolatile components and oligomers account for approximately 7% and 9% of total isoprene SOA, respectively. Contributions of SOA Components to Total SOA. Figure 6 shows that the predicted episode average SOA concentration is approximately 2−6 μg m−3 in the southeastern U.S. with a peak concentrations of 7.4 μg m−3. Isoprene accounts for approximately 55% of the total SOA with a rather uniform spatial distribution, even in areas with low direct isoprene emissions. This uniform spatial distribution is due to multigeneration of glyoxal and methylglyoxal as isoprene is gradually oxidized downwind. Contributions of monoterpenes and sesquiterpenes are mostly located in southeast with maximum relative contributions of 25%, and a domain average of approximately 10%. The spatial distributions of mono-

■

DISCUSSION Isoprene is predicted to be a significant contributor of total SOA in the eastern U.S.. Although the predicted SOA and OC generally agree with observations, significant uncertainties still exist. For example, a previous regional modeling study using CMAQ with the Master Chemical Mechanism for the TexAQS 2006 episode showed higher contributions of IEPOX to total 7839

DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842

Article

Environmental Science & Technology SOA at the Moody Tower. 21 As the photochemical mechanisms were mostly tuned for reproducing ozone concentrations, predictions of later generation oxidation products can be significantly different. This would lead to different estimation of the importance of the aerosol surface uptake pathways. Future studies should be directed to evaluate the photochemical mechanisms for oxidation products such as glyoxal and methylglyoxal. In addition to comparison with observations, intercomparison of photochemical mechanisms needs to be conducted to understand the limitation of the mechanisms. Uncertainty in the model results could also arise from uncertainty in biogenic emissions. The Model of Emissions of Gases and Aerosols from Nature (MEGAN) model is another popular model for biogenic emission estimations. Several studies have shown that MEGAN generally predicts higher isoprene and other biogenic VOC emissions in the U.S. than BEIS 3,54−57 which could lead to higher concentrations of isoprene SOA. Although the fractional contributions of isoprene SOA might not change much because emissions of other biogenic species from MEGAN are also higher, future studies are needed to further reduce the uncertainty in biogenic emission estimations.

■

(4) Carlton, A. G.; Wiedinmyer, C.; Kroll, J. H. A review of secondary organic aerosol (SOA) formation from isoprene. Atmos. Chem. Phys. 2009, 9 (14), 4987−5005. (5) Henze, D. K.; Seinfeld, J. H. Global secondary organic aerosol from isoprene oxidation. Geophys. Res. Lett. 2006, 33 (9), L09812. (6) Liao, H.; Henze, D. K.; Seinfeld, J. H.; Wu, S. L.; Mickley, L. J. Biogenic secondary organic aerosol over the United States: Comparison of climatological simulations with observations. J. Geophys. Res.: Atmos. 2007, 112 (D6), D06201. (7) Zhang, Y.; Huang, J. P.; Henze, D. K.; Seinfeld, J. H. Role of isoprene in secondary organic aerosol formation on a regional scale. J. Geophys. Res.: Atmos. 2007, 112, D20. (8) Pankow, J. F. An Absorption-model of the gas aerosol partitioning involved in the formation of secondary organic aerosol. Atmos. Environ. 1994, 28 (2), 189−193. (9) Kroll, J. H.; Ng, N. L.; Murphy, S. M.; Flagan, R. C.; Seinfeld, J. H. Secondary organic aerosol formation from isoprene photooxidation under high-NOx conditions. Geophys. Res. Lett. 2005, 32 (18), L18808. (10) Odum, J. R.; Hoffmann, T.; Bowman, F.; Collins, D.; Flagan, R. C.; Seinfeld, J. H. Gas/particle partitioning and secondary organic aerosol yields. Environ. Sci. Technol. 1996, 30 (8), 2580−2585. (11) Foley, K. M.; Roselle, S. J.; Appel, K. W.; Bhave, P. V.; Pleim, J. E.; Otte, T. L.; Mathur, R.; Sarwar, G.; Young, J. O.; Gilliam, R. C.; Nolte, C. G.; Kelly, J. T.; Gilliland, A. B.; Bash, J. O. Incremental testing of the Community Multiscale Air Quality (CMAQ) modeling system version 4.7. Geosci. Model Dev. 2010, 3 (1), 205−226. (12) Lim, H. J.; Carlton, A. G.; Turpin, B. J. Isoprene forms secondary organic aerosol through cloud processing: Model simulations. Environ. Sci. Technol. 2005, 39 (12), 4441−4446. (13) Volkamer, R.; Martini, F. S.; Molina, L. T.; Salcedo, D.; Jimenez, J. L.; Molina, M. J., A missing sink for gas-phase glyoxal in Mexico City: Formation of secondary organic aerosol. Geophys. Res. Lett. 2007, 34, (19). (14) Ervens, B.; Volkamer, R. Glyoxal processing by aerosol multiphase chemistry: Towards a kinetic modeling framework of secondary organic aerosol formation in aqueous particles. Atmos. Chem. Phys. 2010, 10 (17), 8219−8244. (15) Claeys, M.; Graham, B.; Vas, G.; Wang, W.; Vermeylen, R.; Pashynska, V.; Cafmeyer, J.; Guyon, P.; Andreae, M. O.; Artaxo, P.; Maenhaut, W. Formation of secondary organic aerosols through photooxidation of isoprene. Science 2004, 303 (5661), 1173−1176. (16) Paulot, F.; Crounse, J. D.; Kjaergaard, H. G.; Kürten, A.; St. Clair, J. M.; Seinfeld, J. H.; Wennberg, P. O. Unexpected epoxide formation in the gas-phase photooxidation of isoprene. Science 2009, 325 (5941), 730−733. (17) Lin, Y.-H.; Zhang, H.; Pye, H. O. T.; Zhang, Z.; Marth, W. J.; Park, S.; Arashiro, M.; Cui, T.; Budisulistiorini, S. H.; Sexton, K. G.; Vizuete, W.; Xie, Y.; Luecken, D. J.; Piletic, I. R.; Edney, E. O.; Bartolotti, L. J.; Gold, A.; Surratt, J. D. Epoxide as a precursor to secondary organic aerosol formation from isoprene photooxidation in the presence of nitrogen oxides. Proc. Natl. Acad. Sci. U. S. A. 2013, 110 (17), 6718−6723. (18) Karambelas, A.; Pye, H. O. T.; Budisulistiorini, S. H.; Surratt, J. D.; Pinder, R. W. Contribution of isoprene epoxydiol to urban organic aerosol: Evidence from modeling and measurements. Environ. Sci. Technol. Lett. 2014, 1 (6), 278−283. (19) Lin, G.; Penner, J. E.; Sillman, S.; Taraborrelli, D.; Lelieveld, J. Global modeling of SOA formation from dicarbonyls, epoxides, organic nitrates and peroxides. Atmos. Chem. Phys. 2012, 12 (10), 4743−4774. (20) Pye, H. O. T.; Pinder, R. W.; Piletic, I. R.; Xie, Y.; Capps, S. L.; Lin, Y. H.; Surratt, J. D.; Zhang, Z. F.; Gold, A.; Luecken, D. J.; Hutzell, W. T.; Jaoui, M.; Offenberg, J. H.; Kleindienst, T. E.; Lewandowski, M.; Edney, E. O. Epoxide pathways improve model predictions of isoprene markers and reveal key role of acidity in aerosol formation. Environ. Sci. Technol. 2013, 47 (19), 11056−11064. (21) Li, J.; Cleveland, M.; Ziemba, L. D.; Griffin, R. J.; Barsanti, K. C.; Pankow, J. F.; Ying, Q. Modeling regional secondary organic aerosol

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.5b02514.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: 979-587-3781; ax: 979-862-1542; e-mail: qying@civil. tamu.edu. Present Addresses †

Atmospheric and Ocean Sciences Program, Princeton University. ‡ Department of Civil Engineering, Indian Institute of Technology, Guwahati. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This project has been partially supported with funds from the State of Texas as part of the program of the Texas Air Research Center (Project Number 079ATM0099A, 078ATM2080A and 312ATM0126A). We acknowledge the Texas A&M Supercomputing Facility (http://sc.tamu.edu) and the Texas Advanced Computing Center (TACC) at The University of Texas at Austin (http://www.tacc.utexas.edu) for providing computing resources useful in conducting the research reported in this paper.

■

REFERENCES

(1) Guenther, A.; Karl, T.; Harley, P.; Wiedinmyer, C.; Palmer, P. I.; Geron, C. Estimates of global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from Nature). Atmos. Chem. Phys. 2006, 6 (11), 3181−3210. (2) Atkinson, R.; Arey, J. Atmospheric degradation of volatile organic compounds. Chem. Rev. 2003, 103 (12), 4605−4638. (3) Kroll, J. H.; Ng, N. L.; Murphy, S. M.; Flagan, R. C.; Seinfeld, J. H. Secondary organic aerosol formation from isoprene photooxidation. Environ. Sci. Technol. 2006, 40 (6), 1869−1877. 7840

DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842

Article

Environmental Science & Technology using the master chemical mechanism. Atmos. Environ. 2015, 102, 52− 61. (22) West, J. J.; Ansari, A. S.; Pandis, S. N. Marginal PM2.5: Nonlinear aerosol mass response to sulfate reductions in the Eastern United States. J. Air Waste Manage. Assoc. 1999, 49 (12), 1415−1424. (23) Carter, W. P. L.; Heo, G. Development of revised SAPRC aromatics mechanisms. Final Report to the California Air Resources Board, Contracts No. 07-730 and 08−326, April 12, 2012, 2012. (24) Carter, W. P. L. Development of the SAPRC-07 chemical mechanism. Atmos. Environ. 2010, 44 (3), 5324−5335. (25) Carter, W. P. L. Development of a condensed SAPRC-07 chemical mechanism. Atmos. Environ. 2010, 40 (44), 5336−5345. (26) Xie, Y.; Paulot, F.; Carter, W. P. L.; Nolte, C. G.; Luecken, D. J.; Hutzell, W. T.; Wennberg, P. O.; Cohen, R. C.; Pinder, R. W. Understanding the impact of recent advances in isoprene photooxidation on simulations of regional air quality. Atmos. Chem. Phys. 2013, 13 (16), 8439−8455. (27) Bates, K. H.; Crounse, J. D.; St. Clair, J. M.; Bennett, N. B.; Nguyen, T. B.; Seinfeld, J. H.; Stoltz, B. M.; Wennberg, P. O. Gas phase production and loss of isoprene epoxydiols. J. Phys. Chem. A 2014, 118 (7), 1237−1246. (28) Byun, D.; Schere, K. L. Review of the governing equations, computational algorithms, and other components of the models-3 Community Multiscale Air Quality (CMAQ) modeling system. Appl. Mech. Rev. 2006, 59 (2), 51−77. (29) Nenes, A.; Pandis, S. N.; Pilinis, C. ISORROPIA: A new thermodynamic equilibrium model for multiphase multicomponent inorganic aerosols. Aquat. Geochem. 1998, 4 (1), 123−152. (30) Carlton, A. G.; Bhave, P. V.; Napelenok, S. L.; Edney, E. D.; Sarwar, G.; Pinder, R. W.; Pouliot, G. A.; Houyoux, M. Model representation of secondary organic aerosol in CMAQv4.7. Environ. Sci. Technol. 2010, 44 (22), 8553−8560. (31) Lal, V.; Khalizov, A. F.; Lin, Y.; Galvan, M. D.; Connell, B. T.; Zhang, R. Heterogeneous reactions of epoxides in acidic media. J. Phys. Chem. A 2012, 116 (24), 6078−6090. (32) Volkamer, R.; San Martini, F.; Molina, L. T.; Salcedo, D.; Jimenez, J. L.; Molina, M. J. A missing sink for gas-phase glyoxal in Mexico City: Formation of secondary organic aerosol. Geophys. Res. Lett. 2007, 34 (19), L19807. (33) Carlton, A. G.; Turpin, B. J.; Altieri, K. E.; Seitzinger, S.; Reff, A.; Lim, H.-J.; Ervens, B. Atmospheric oxalic acid and SOA production from glyoxal: Results of aqueous photooxidation experiments. Atmos. Environ. 2007, 41 (35), 7588−7602. (34) Zhao, J.; Levitt, N. P.; Zhang, R.; Chen, J. Heterogeneous reactions of methylglyoxal in acidic media: Implications for secondary organic aerosol formation. Environ. Sci. Technol. 2006, 40 (24), 7682− 7687. (35) Galloway, M. M.; Chhabra, P. S.; Chan, A. W. H.; Surratt, J. D.; Flagan, R. C.; Seinfeld, J. H.; Keutsch, F. N. Glyoxal uptake on ammonium sulphate seed aerosol: Reaction products and reversibility of uptake under dark and irradiated conditions. Atmos. Chem. Phys. 2009, 9 (10), 3331−3345. (36) 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. (37) Wang, T.; Liu, Z.; Wang, W.; Ge, M. Uptake kinetics of three epoxides into sulfuric acid solution. Atmos. Environ. 2012, 56 (0), 58− 64. (38) Birdsall, A. W.; Miner, C. R.; Mael, L. E.; Elrod, M. J. Mechanistic study of secondary organic aerosol components formed from nucleophilic addition reactions of methacrylic acid epoxide. Atmos. Chem. Phys. 2014, 14 (23), 12951−12964. (39) Fu, T. M.; Jacob, D. J.; Heald, C. L. Aqueous-phase reactive uptake of dicarbonyls as a source of organic aerosol over eastern North America. Atmos. Environ. 2009, 43, 1814−1822. (40) Ying, Q.; Cureño, I. V.; Chen, G.; Ali, S.; Zhang, H.; Malloy, M.; Bravo, H. A.; Sosa, R. Impacts of Stabilized Criegee Intermediates, surface uptake processes and higher aromatic secondary organic

aerosol yields on predicted PM2.5 concentrations in the Mexico City Metropolitan Zone. Atmos. Environ. 2014, 94 (0), 438−447. (41) Kota, S. H.; Zhang, H.; Chen, G.; Schade, G. W.; Ying, Q. Evaluation of on-road vehicle CO and NOx National Emission Inventories using an urban-scale source-oriented air quality model. Atmos. Environ. 2014, 85, 99−108. (42) Zhang, H.; Li, J.; Ying, Q.; Guven, B. B.; Olaguer, E. P. Source apportionment of formaldehyde during TexAQS 2006 using a sourceoriented chemical transport model. J. Geophys. Res., Atmos. 2013, 118 (3), 1525−1535. (43) Li, J.; Ying, Q.; Yi, B.; Yang, P. Role of stabilized Criegee Intermediates in the formation of atmospheric sulfate in eastern United States. Atmos. Environ. 2013, 79 (0), 442−447. (44) Vukovich, J. M.; Pierce, T. In The Implementation of BEIS3 within the SMOKE modeling framework; MCNC-Environmental Modeling Center, Research Triangle Park and National Oceanic and Atmospheric Administration, 2002. (45) Carter, W. P. L. Development of a chemical speciation database and software for processing VOC emissions for air quality models. In Proceedings of the 13th International Emission Inventory Conference “Working for Clean Air in Clearwater”, Clearwater, FL, 2004. (46) Ngan, F.; Byun, D.; Kim, H.; Lee, D.; Rappenglück, B.; PourBiazar, A. Performance assessment of retrospective meteorological inputs for use in air quality modeling during TexAQS 2006. Atmos. Environ. 2012, 54, 86−96. (47) Turpin, B. J.; Lim, H. Species contributions to PM2.5 mass concentrations: Revisiting common assumptions for estimating orgainc mass. Aerosol Sci. Technol. 2001, 35, 602−610. (48) Kleindienst, T. E.; Jaoui, M.; Lewandowski, M.; Offenberg, J. H.; Lewis, C. W.; Bhave, P. V.; Edney, E. O. Estimates of the contributions of biogenic and anthropogenic hydrocarbons to secondary organic aerosol at a southeastern US location. Atmos. Environ. 2007, 41 (37), 8288−8300. (49) Wood, E. C.; Canagaratna, M. R.; Herndon, S. C.; Kroll, J. H.; Onasch, T. B.; Kolb, C. E.; Worsnop, D. R.; Knighton, W. B.; Seila, R.; Zavala, M.; Molina, L. T.; DeCarlo, P. F.; Jimenez, J. L.; Weinheimer, A. J.; Knapp, D. J.; Jobson, B. T.; Stutz, J.; Kuster, W. C.; Williams, E. J. Investigation of the correlation between odd oxygen and secondary organic aerosol in Mexico City and Houston. Atmos. Chem. Phys. Discuss. 2010, 10, 3547−3606. (50) Cleveland, M. J.; Ziemba, L. D.; Griffin, R. J.; Dibb, J. E.; Anderson, C. H.; Lefer, B.; Rappenglück, B. Characterization of urban aerosol using aerosol mass spectrometry and proton nuclear magnetic resonance spectroscopy. Atmos. Environ. 2012, 54 (0), 511−518. (51) Wood, E. C.; Canagaratna, M. R.; Herndon, S. C.; Onasch, T. B.; Kolb, C. E.; Worsnop, D. R.; Kroll, J. H.; Knighton, W. B.; Seila, R.; Zavala, M.; Molina, L. T.; DeCarlo, P. F.; Jimenez, J. L.; Weinheimer, A. J.; Knapp, D. J.; Jobson, B. T.; Stutz, J.; Kuster, W. C.; Williams, E. J. Investigation of the correlation between odd oxygen and secondary organic aerosol in Mexico City and Houston. Atmos. Chem. Phys. 2010, 10 (18), 8947−8968. (52) Canagaratna, M. R.; Jayne, J. T.; Jimenez, J. L.; Allan, J. D.; Alfarra, M. R.; Zhang, Q.; Onasch, T. B.; Drewnick, F.; Coe, H.; Middlebrook, A.; Delia, A.; Williams, L. R.; Trimborn, A. M.; Northway, M. J.; DeCarlo, P. F.; Kolb, C. E.; Davidovits, P.; Worsnop, D. R. Chemical and microphysical characterization of ambient aerosols with the aerodyne aerosol mass spectrometer. Mass Spectrom. Rev. 2007, 26 (2), 185−222. (53) Fu, T. M.; Jacob, D. J.; Wittrock, F.; Burrows, J. P.; Vrekoussis, M.; Henze, D. K., Global budgets of atmospheric glyoxal and methylglyoxal, and implications for formation of secondary organic aerosols. J. Geophys. Res.: Atmos. 2008, 113, (D15). (54) Warneke, C.; de Gouw, J. A.; Del Negro, L.; Brioude, J.; McKeen, S.; Stark, H.; Kuster, W. C.; Goldan, P. D.; Trainer, M.; Fehsenfeld, F. C.; Wiedinmyer, C.; Guenther, A. B.; Hansel, A.; Wisthaler, A.; Atlas, E.; Holloway, J. S.; Ryerson, T. B.; Peischl, J.; Huey, L. G.; Hanks, A. T. C. Biogenic emission measurement and inventories determination of biogenic emissions in the eastern United 7841

DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842

Article

Environmental Science & Technology States and Texas and comparison with biogenic emission inventories. J. Geophys. Res.: Atmos. 2010, 115 (D7), D00F18. (55) Carlton, A. G.; Baker, K. R. Photochemical modeling of the Ozark isoprene volcano: MEGAN, BEIS, and their impacts on air quality predictions. Environ. Sci. Technol. 2011, 45 (10), 4438−4445. (56) Hogrefe, C.; Isukapalli, S. S.; Tang, X.; Georgopoulos, P. G.; He, S.; Zalewsky, E. E.; Hao, W.; Ku, J.-Y.; Key, T.; Sistla, G. Impact of biogenic emission uncertainties on the simulated response of ozone and fine particulate matter to anthropogenic emission reductions. J. Air Waste Manage. Assoc. (1995) 2011, 61 (1), 92−108. (57) Luecken, D. J.; Hutzell, W. T.; Strum, M. L.; Pouliot, G. A. Regional sources of atmospheric formaldehyde and acetaldehyde, and implications for atmospheric modeling. Atmos. Environ. 2012, 47, 477−490.

7842

DOI: 10.1021/acs.est.5b02514 Environ. Sci. Technol. 2015, 49, 7834−7842