Grantee Research Project Results
2009 Progress Report: Improved Prediction of In-Cloud Biogenic SOA: Experiments and CMAQ Model Refinements
EPA Grant Number: R833751Title: Improved Prediction of In-Cloud Biogenic SOA: Experiments and CMAQ Model Refinements
Investigators: Turpin, Barbara , Seitzinger, Sybil
Institution: Rutgers
EPA Project Officer: Chung, Serena
Project Period: November 1, 2007 through August 31, 2010 (Extended to October 31, 2011)
Project Period Covered by this Report: September 1, 2008 through August 31,2009
Project Amount: $598,544
RFA: Sources and Atmospheric Formation of Organic Particulate Matter (2007) RFA Text | Recipients Lists
Research Category: Air Quality and Air Toxics , Particulate Matter , Air
Objective:
1) Develop mechanistic/ kinetic data needed to simulate in-cloud formation of secondary organic aerosol (SOA) in the presence of HNO3, 2) Identify conditions for which predicted in-cloud SOA formed from isoprene decreases with reductions in interstitial concentrations of ·OH and HNO3 (atmospheric oxidants with anthropogenic precursors), and 3) Incorporate an in-cloud SOA formation pathway into the Community Multiscale Air Quality (CMAQ) model and explore the magnitude of in-cloud SOA formation through a limited set of model simulations.
Overview: It is now reasonably well accepted that aqueous oxidation of glyoxal and methylglyoxal by OH radical in clouds and wet aerosols results in the formation of secondary organic aerosol (SOA). Laboratory studies conducted through our previous STAR grant, together with ambient measurements have provided important evidence for this. Our laboratory experiments with glyoxal/methylglyoxal + ·OH verify that organic acids commonly found in atmospheric particles are formed (e.g., oxalic acid) and suggest that oligomers form as well (Altieri, et al., Atmos. Environ. 2008). SOA formation through aqueous photooxidation (e.g., cloud processing) was only recently recognized (Blando and Turpin, 2000), is the only identified process that can explain the atmospheric abundance/temporal dynamics of oxalic acid, has different precursors than “traditional” SOA, and could be a substantial contributor to total SOA globally and regionally. We hypothesized that similar chemistry also occurs in aerosol water, and the work of others support this hypothesis (a smog chamber experiment, Volkamer et al., ACP, 2009; and ambient measurements in Atlanta, Hennigan et al., GRL, 2008).
One limitation of our previous work is that we were only able to perform experiments at concentrations about 100 times greater than those found in clouds (and 1000 times lower than concentrations in aerosol water). Our current STAR grant included support (50%) for the purchase of instrumentation that has allowed us to conduct experiments at cloud-relevant (and higher) concentrations in order to better understand the differences between the chemistry in clouds and in wet aerosols. The postdoctoral fellow recruited to work on this grant has made substantial progress understanding the detailed radical chemistry that is extending the range of concentrations for which we can make accurate predictions. Also, in a logical extension of our past work, we are conducting experiments in the presence and absence of nitric acid (HNO3). Finally, we are collaborating with EPA scientists (and others) to help them incorporate in-cloud SOA formation into CMAQ (and other models).
Progress Summary:
Objective 1: With support from this grant and Rutgers University, we purchased a Dionex Ion Chromatography System (ICS-3000), which has allowed us to conduct aqueous photooxidation experiments at lower concentrations than previously possible because of improved sensitivity and resolution (Fig 1). In year 1 we conducted experiments with glyoxal + ·OH at 30 μM, 300 μM, and 3000 μM with and without sulfuric acid (Tan et al, Environ. Sci. Technol., 2009). In year 2 we conducted the same suite of experiments with methylglyoxal (Tan, et al., in preparation). Previous experiments were conducted only at 1000-3000 μM. The 30 μM experiments are representative of cloud concentrations. Concentrations in aerosol water could be as high as 1-10 M. Samples from these experiments were analyzed for organic acids by Ion Chromatography (IC). Their mass spectra were obtained by Electrospray Ionization Mass Spectroscopy (ESI-MS) and ultra-high resolution Fourier Transform Ion Cyclotron Resonance Electrospray Ionization Mass Spectroscopy (FTICR-MS), and total carbon in each sample was measured using a Total Organic Carbon Analyzer. Dissolved oxygen and pH were measured at the beginning and end of each experiment. The chemistry in the reaction vessel was also modeled using our aqueous phase chemistry model (Lim model). In this way we could use the experiments to validate and refine our chemistry model. The rate constants in the model are reasonably well known, but in many cases the products had not previously been verified.
Figure 1. Separation of organic acids in a mixed standard and example analysis of a glyoxal experiment sample. Separation is much better than previously obtained by HPLC-UV.
This summer we coupled the IC directly to the ESI-MS so that products separated by polarity in the IC column would pass directly into the ESI-MS. This has helped us to identify more products, test mechanistic hypotheses, and verify that these products are formed in the reaction vessel and not in the ESI. IC preseparation also makes the ESI-MS ion abundances more quantitative. This technique was used to examine a subset of samples from methylglyoxal experiments.
Methylglyoxal + ·OH experiments and control experiments have been conducted with and without HNO3. We hypothesize that organic nitrates form in the presence of HNO3. Further, we expect to find organic nitrates in cloud water. We had the opportunity to analyze a few New Jersey rainwater samples to see if similar products are found in rainwater. Two publications have resulted from the rainwater analyses (Altieri, ES&T 2009; Altieri, Atmos Environ 2009). We have exceeded our year 2 goals for objective 1.
Objective 2: Postdoc Yong Bin Lim has made substantial progress refining our chemical model to incorporate radical chemistry and the formation of higher molecular weight products (manuscript in preparation). He has substantially improved our understanding of how aqueous organic chemistry in clouds differs from that in wet aerosols. He is just beginning to use this model to explore how concentrations of OH radical, HNO3 and NOx (with anthropogenic sources) affect formation of SOA from isoprene (a biogenic compound) through aqueous photochemistry. We have also collaborated with Dr. Barbara Ervens, who has incorporated some of our experimental results into a cloud parcel/cloud chemistry model to investigate the cloud processing of isoprene under low NOx and high NOx conditions (Ervens et al., GRL 2008). We still have substantial modeling to do in order to characterize the dependence of in-cloud isoprene SOA on ·OH and HNO3.
Objective 3: In-cloud production of SOA from glyoxal and methylglyoxal has been added to the CMAQ model using a yield based approach similar to that used for other SOA formation processes. The model was used to predict organic carbon (OC) concentrations measured on an airplane during the ICARTT experiment. This work has been published (Carlton et al., EST 2008). We are also cooperating with other modelers who are working toward adding SOA formation through aqueous chemistry into their models.
To further improve the treatment of cloud chemistry in CMAQ, a Rosenbrock solver (ROS3) has been incorporated into the full scale version of CMAQv4.7 and is now being tested. The current ROS3 aqueous chemical mechanism is the same as in the base model version and will be made publicly available via CMAS during the next (FY2011) public release. The ROS3 version implements the same aqueous chemical mechanism as the base model, with the sole exception that gas-to-water partitioning is described kinetically, rather than assuming instantaneous thermodynamic equilibrium. Box model testing continues for the development and expansion of the aqueous chemical mechanism. We have achieved our year 2 goals for objective 3.
Results to date
Objective 1: The Lim model did a good job of capturing the magnitude and concentration dynamics of oxalic acid and total carbon for both glyoxal and methylglyoxal when the experiments were conducted at cloud relevant concentrations (30 μM of organic). Oxalic acid time profiles for 30 μM experiments are shown in Fig 2a and 3a. This suggests that the Lim model is adequate for use in predicting the aqueous chemistry leading to SOA formation through cloud processing. However, as the initial concentrations of the organic precursor increased, the predictive capability of the model degraded, and the measured organic acids accounted for a smaller and smaller portion of the total carbon in the reaction vessel. We found that increasing precursor concentrations resulted in samples with increasingly complex mass spectra and resulted in the formation of organic acids and oligomers with larger carbon numbers (Fig 4, 5). Because higher carbon number products form only in the presence of OH radical and only at high organic concentrations, we hypothesize that these products form through radical-radical reactions.
In FT-ICR-MS analyses of methylglyoxal (+·OH) experiments conducted with HNO3, we have identified many nitrogen-containing organic compounds that are not present in experiments without HNO3. In these experiments, the decrease in nitrate concentration with time (if occurring at all) is very small. Thus we know that we are producing at most small concentrations of nitrogen radicals (e.g., NO2 radicals) by nitrate photolysis, and any organic nitrogen products, if formed, would be formed in very small quantities. The FT-ICR-MS is capable of detecting very small quantities. While the organic nitrogen products in question are not present in the control experiments (without HNO3), we have not yet addressed the possibility that these are organic contaminants in the HNO3. We are also currently considering what organic nitrogen products we would expect to form, given what we know about the chemistry of methyglyoxal + OH radical and about organosulfur products formed from glycolaldehyde + OH radical + H2SO4.
The available literature on nitrate photolysis suggests that the wavelength we are using to photolyze nitrate is appropriate, but that the molar yields of OH and NO2 radical are small.
Thus, it is quite possible that an insignificant amount of organic nitrogen is formed from the aqueous reaction of organics and nitrate photolysis products at cloud-relevant concentrations. This chemistry may very well become important when nitrate is present in much higher concentrations, for example in wet aerosols. That has not been tested.
In analyses of New Jersey rainwater samples we found series of compounds whose elemental compositions were identical to those formed through oligomerization reactions in methylglyoxal+·OH experiments performed earlier in our laboratory. We also found a wide variety of organic nitrogen products in these rainwater samples. Interestingly, only a modest proportion of the organic nitrogen was comprised of organic nitrates. Most of the organic nitrogen products were in the positive mode in the ESI-MS. We are considering whether/when to shift the focus of the nitrogen experiments to investigate a nitrogen additive more likely to explain the rainwater data (e.g., ammonium, amines, or amino acids).
Figure 4. Negative mode ESI-mass spectra 20 min into each glyoxal + ·OH experiment. Complexity increases with increasing precursor concentration.
Figure 5a. IC-ESI-MS spectra of mixed standard. (A) acetic acid (not detected in ESI-MS) and glycolic acid (m/z- 75), (B) formic acid (not detected in ESI-MS), (C) pyruvic acid (m/z- 87), (D) glyoxylic acid (m/z- 73), (E) succinic acid (m/z- 117), (F) malonic acid (m/z- 103), (G) unknown carboxylic acid with m/z- 175, (H) unknown carboxylic acid with m/z- 161, (I) oxalic acid (m/z- 89), (J) unknown carboxylic acids with m/z- 133 and m/z- 177.
Figure 5b. IC-ESI-MS spectra of a sample taken from 300 M methylglyoxal + OH radical batch reactions (180 minutes reaction time). (A) acetic acid and glycolic acid, (B) pyruvic acid (m/z- 177), (C) contaminant peak (m/z- 92), (D) sulfuric acid contaminant and an unknown carboxylic acid (m/z- 97, and 119), (E) oxalic acid (m/z- 89), (F) mesoxalic acid (m/z- 117).
Figure 5c. IC-ESI-MS spectra of a sample taken from 3000 M methylglyoxal + OH radical batch reactions (180 minutes reaction time). (A) acetic acid and glycolic acid, (B) unkown carboxylic acid (m/z- 177), (C) peak with the retention time of succinic acid (m/z- 117 and 177), (D) peak with the retention time of malonic acid (m/z- 103, 133, and 177), (E) unknown carboxylic acids (m/z- 119 and 177), (F) oxalic acid (m/z- 89), (G) unknown tricarboxylic acids with m/z- > 200.
Objective 2: At the American Association for Aerosol Research Annual Conference this year (2009), our postdoc, Dr. Yong Bin Lim, presented a glyoxal model that expands the original Lim model by explicitly accounting for radical - radical reactions (Fig 6). Model performance for 30 μM experiments changed little; model performance was much improved for 3000 μM glyoxal experiments (Lim et al., publication in preparation). This represents our first efforts to understand how the aqueous OH radical chemistry might be different in wet aerosols, where concentrations are still orders of magnitude greater than the concentrations in our experiments.
Figure 6. Expanded glyoxal model. At low concentrations, radicals in red react with oxygen to produce organic acids. At high concentrations, radicals can react with each other to produce higher molecular weight organic acids and oligomers.
We are beginning to use this model to explore how OH radical, HNO3 and NOx (with anthropogenic sources) affect formation of SOA from isoprene (a biogenic compound) through aqueous photochemistry. Dr. Barbara Ervens, in collaboration with our group, showed that cloud processing of isoprene at high NOx resulted in higher SOA yields than it did at low NOx because the yield of carbonyl compounds (e.g., glyoxal and methylglyoxal) through gas phase reactions is higher at high NOx. This suggests that more SOA will form from a biogenic hydrocarbon like isoprene in polluted conditions than in clean conditions when the SOA is formed through cloud chemistry. The opposite is true for “traditional” SOA, like that investigated in smog chambers at low RH. The degree to which the presence of nitric acid in the aqueous phase alters product formation, resulting in aqueous phase organic nitrogen products is not yet clear. Our work to date suggests this is not important in clouds. Its importance in wet aerosols has not been addressed.
Objective 3: After adding in-cloud production of SOA from glyoxal and methylglyoxal to the CMAQ model using a yield based approach, the model was used to predict organic carbon (OC) concentrations measured on an airplane during the ICARTT experiment. The addition of cloud processing to CMAQ improved agreement between modeled OC and measured water-soluble organic carbon (WSOC) for all flights (e.g., Fig 7). SOA formation through cloud processing was negligible in certain areas and substantial in others. This work, and similar modeling for the Northeastern US and the globe conducted by others using GEOSChem (Fu et al., Atmos. Environ. 2009; Fu et al., JGR, 2008) suggests that SOA formation through aqueous chemistry is comparable in magnitude to “traditional” SOA, although the uncertainties are still large. We are cooperating with a number of modelers who are working toward adding SOA formation through aqueous chemistry into their models.
Figure 7. Results of CMAQ model with (red, uncertainties in grey) and without (blue) SOA through cloud processing for August 14th ICARTT flight. Also shown is water-soluble organic carbon (open black circles) and plane altitude (black line).
To further improve the treatment of cloud chemistry in CMAQ, a Rosenbrock solver (ROS3) has been incorporated into the full scale version of CMAQv4.7 and is now being tested. The current ROS3 aqueous chemical mechanism is the same as in the base model version and will be made publicly available via CMAS during the next (FY2011) public release. The ROS3 version implements the same aqueous chemical mechanism as the base model, with the sole exception that gas-to-water partitioning is described kinetically, rather than assuming instantaneous thermodynamic equilibrium. Box model testing continues for the development and expansion of the aqueous chemical mechanism.
Future Activities:
In the subsequent reporting period we will focus on completing Objective 2. We also will use the model to examine how it can be simplified further for inclusion in CMAQ and other models. We will continue to work with our EPA collaborator, Annmarie Carlton, who is improving the treatment of aqueous chemistry in the CMAQ model. In addition, we will proceed with experiments with and without a nitrogen additive (e.g., nitrate, ammonium, amines).Journal Articles on this Report : 6 Displayed | Download in RIS Format
| Other project views: | All 46 publications | 16 publications in selected types | All 16 journal articles |
|---|
| Type | Citation | ||
|---|---|---|---|
|
|
Altieri KE, Seitzinger SP, Carlton AG, Turpin BJ, Klein GC, Marshall AG. Oligomers formed through in-cloud methylglyoxal reactions: chemical composition, properties, and mechanisms investigated by ultra-high resolution FT-ICR mass spectrometry. Atmospheric Environment 2008;42(7):1476-1490. |
R833751 (2008) R833751 (2009) R833751 (2010) R833751 (Final) R831073 (Final) |
Exit Exit Exit |
|
|
Altieri KE, Turpin BJ, Seitzinger SP. Oligomers, organosulfates, and nitrooxy organosulfates in rainwater identified by ultra-high resolution electrospray ionization FT-ICR mass spectrometry. Atmospheric Chemistry and Physics 2009;9(7):2533-2542. |
R833751 (2009) R833751 (Final) |
Exit Exit |
|
|
Altieri KE, Turpin BJ, Seitzinger SP. Composition of dissolved organic nitrogen in continental precipitation investigated by ultra-high resolution FT-ICR mass spectrometry. Environmental Science & Technology 2009;43(18):6950-6955. |
R833751 (2009) R833751 (2010) R833751 (Final) |
Exit Exit Exit |
|
|
Carlton AG, Turpin BJ, Altieri KE, Seitzinger SP, Mathur R, Roselle SJ, Weber RJ. CMAQ model performance enhanced when in-cloud secondary organic aerosol is included:comparisons of organic carbon predictions with measurements. Environmental Science & Technology 2008;42(23):8798-8802. |
R833751 (2008) R833751 (2009) R833751 (2010) R833751 (Final) R831073 (Final) |
Exit Exit Exit |
|
|
Ervens B, Carlton AG, Turpin BJ, Altieri KE, Kreidenweis SM, Feingold G. Secondary organic aerosol yields from cloud-processing of isoprene oxidation products. Geophysical Research Letters 2008;35(2):L02816 (5 pp.). |
R833751 (2008) R833751 (2009) R833751 (2010) R833751 (Final) R831073 (Final) |
Exit Exit |
|
|
Tan Y, Perri MJ, Seitzinger SP, Turpin BJ. Effects of precursor concentration and acidic sulfate in aqueous glyoxal–OH radical oxidation and implications for secondary organic aerosol. Environmental Science & Technology 2009;43(21):8105-8112. |
R833751 (2009) R833751 (2010) R833751 (Final) |
Exit Exit Exit |
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.