Novel Measurements of Volatility- and Polarity-Separated Organic Aerosol Composition and Associated Hygroscopicity to Investigate the Influence of Mixed Anthropogenic-Biogenic Emissions on Atmospheric Aging Processes
April 1, 2013 through March 31,2014
Anthropogenic Influences on Organic Aerosol Formation and Regional Climate Implications (2012)
Coupling of anthropogenic and biogenic emissions and subsequent atmospheric aging processes are hypothesized to be the leading single contribution to global organic aerosol (OA) mass concentrations, and is also perhaps the least understood single contribution to OA. Improved understanding of this coupling is a high priority topic in the field of atmospheric chemistry in order to determine mitigation strategies for OA control, a pollutant that alters climate and causes detrimental health effects. The primary objective of this proposed work is to better characterize the controlling factors in enhanced secondary organic aerosol (SOA) formation from combined anthropogenic and biogenic emission sources through innovative laboratory and field studies using novel instrumentation.
In year one of this project, we completed development of a volatility and polarity separator (VAPS) for improved chemical characterization of organic aerosol and deployed it on two different field studies. We also built a custom multi-channel tandem differential mobility analyzer (MC-TDMA) for particle sizing, volatility, and hygroscopicity, and deployed it on a first field study. We completed an initial round of laboratory-based measurements of the oxidation of isoprene (the dominant biogenic emission globally and in the southeastern United States) under varying concentrations, oxidant exposures, and mixtures with NOx. A major lesson in the first round of lab studies was to improve the design and operation of the PAM reaction chamber to ensure stable mixing and oxidation of precursor gases, and development of improved methods of calibrating oxidant exposures for each individual experiment. Major observations from VAPS deployment to Centreville, AL, as part of the Southern Oxidant and Aerosol Study (SOAS) include observation of highly oxygenated aerosol and identification of several major types of oxygenated biogenic aerosol, the largest contribution coming from low-NOx isoprene oxidation. Very little anthropogenic (from human activity) aerosol was observed. VAPS did observe minor components of organonitrates and organosulfates, which will be investigated further and compared to observations from offline filter-based analyses.
A wide array of gas and particle measurements were also made during this reporting period in East St. Louis, IL, a site with significantly higher anthropogenic pollutants and periodic impact from air masses with high concentration of biogenic isoprene and its oxidation products. Initial analysis reveals evidence of night-time particle growth events when biogenic (isoprene) emissions (transported from the Ozark Mountains to the southwest of St. Louis) were impacting the region along with locally elevated nitrate radical concentrations. These night-time growth events were always smaller in scale to daytime growth events a decade ago (Qian Shi, Masters Thesis with P. McMurry, U.Minnesota) but overwhelmed any evidence of strong daytime growth events during our recent measurement period (Aug.-Oct., 2013). SO2 has dropped by approximately a factor of 5, and NOx has decreased by approximately a factor of 2 since that time period; however, O3 has actually slightly increased in the same time period. It is estimated that nighttime O3 is not completely titrated away as often as it once had been, leaving the potential for increased night-time NO3 formation and associated increased potential for nocturnal chemistry and particle formation. Chemical characterization of this aerosol is currently underway utilizing several online aerosol mass spectrometry instruments (including VAPS) along with offline filter analyses. This increase of nocturnal particle growth events demonstrates the non-linear atmospheric response to decreased anthropogenic emissions.
Activity 1: Development of new instrumentation to obtain new insights on the inherent complex chemistry and composition of SOA formation and evolution.
The volatility and polarity separator (VAPS) instrument for improved organic aerosol (OA) chemical and physical characterization was completed and tested in the laboratory using known chemical standard mixtures prior to deployment to the SOAS field study. Additional development of the instrument occurred in the field to adjust for highly oxygenated OA and elevated humidity present at the Centreville, AL, ground station. Alterations included focusing trap redesign to continue to use compressed air trapping without having to go to liquid nitrogen trapping, as well as a change of GC column materials to achieve sufficient molecular separation of highly oxygenated compounds.
A multi-channel tandem differential mobility analyzer (MC-TDMA) was constructed during the summer of 2013 in preparation for a second field deployment to East St. Louis, IL. The instrument is currently capable of high time resolution measurements of particle volatility, hygroscopicity, size distributions, and has an auxiliary channel for other particle treatments of interest. Development resumed after initial field measurements in East St. Louis, and were recently completed in preparation for a summer 2014 laboratory intensive.
Activity 2: Deployment to a large field study (Southern Oxidant & Aerosol Study).
The VAPS instrument was deployed to the Centreville, AL, ground site between May-July, 2013 during the Southern Oxidant & Aerosol Study (SOAS) as part of the larger Southeast Atmosphere Study (SAS). A graduate student (Martinez) and undergraduate student (Hagan) were present for the full study, and PI Williams was present for ¾ of the study. The instrument was operated in two different sample modes. The first mode was typical VAPS analysis where volatility and polarity are separated in multiple dimensions (See Fig.1). The second mode was a faster GC mode with volatility and oxidation separation through high resolution mass spectrometry (See Fig.2). Both modes detected the same organic aerosol components, but separated the molecules in a slightly different order. Individual compounds can be extracted from both techniques with preserved concentration variability, meaning neither mode was at the sacrifice of lost data, but both were utilized to inform on best practice for future deployment.
Organic aerosol observed at SOAS was highly oxygenated and dominantly produced from gas-to-particle conversion processes. The VAPS observed some highly volatile oxygenated components that are not thought to participate in gas-particle phase partitioning (e.g., acetic acid and 3-methylfuran), and it may be that these species were present due to aqueous uptake. Fine aerosol during SOAS had significant aerosol water content (Carlton et al.) that was most abundant in the morning hours, when many of the oxygenated VAPS positive matrix factorization (PMF) components were at their highest abundance. In other locations, it is often observed that secondary organic aerosol (SOA) is highest during the afternoon when photochemistry is most active.
Preliminary PMF characterization of VAPS data has provided individual components that include: carboxylic acids, ketones, several types of oxygenated biogenics, alkanes, organonitrates, and organosulfates. The dominant oxygenated biogenic source is termed IEPOX SOA. This is the same factor that is observed by the AMS instrument in the southeastern United States, including during the SOAS campaign (e.g., observed by Ng et al. and Jimenez et al.). This component has a characteristic mass spectral pattern with distinct ion fragments at m/z 53 and 82. The VAPS has additional volatility resolution on this factor and further investigation may provide new insight on its composition. It is thought to be derived through the low-NOx isoprene epoxydiol pathway, but it is uncertain what final aerosol component is actually being detected by these thermal desorption/chemical ionization/mass spectrometry instruments. Further investigation of this factor and the organonitrates factor are of high priority for SOAS science objectives.
Activity 3: Controlled laboratory studies using a custom emissions/combustion chamber, a flow-tube Potential Aerosol Mass (PAM) reaction chamber, and a large suite of chemical and physical measurement systems.
Initial work on laboratory studies occurred between October-December, 2013. We performed mostly isoprene oxidation experiments under various concentrations, oxidant exposures, and NOx concentrations. Measurements were made on a PTRMS for isoprene degradation and OVOC product generation and eventual decay at highest oxidation. Particle phase measurements were made on the AMS and VAPS systems for chemical composition and the SMPS and TDMA for sizing and volatility/hygroscopicity. Supporting measurements were also recorded (O3, SO2, NO/NO2/NOx, Temperatures, Relative Humidity, Flows, etc.).
While interesting observations were made regarding variation in oxidation products under all experiments, our greatest lessons were in regard to correct PAM reaction chamber operation in a laboratory setting. OH concentrations are typically calculated using SO2 decay with a known first order reaction rate constant. Being among the first PAM users to calibrate using a range of VOCs, we observed potential for OH suppression at high concentrations of VOC precursor. We performed these calibrations using a constant SO2 injection with varying amounts of precursor VOC (e.g., isoprene, toluene, xylene). SO2 decay does slow down with elevated VOC concentrations. This has informed us that OH exposures will need to be measured explicitly for each VOC and for each variation of VOC concentration. Finally, we have determined that the PAM chamber needs improved mixing of precursor VOCs and oxidants (introduced on a separate inlet) to promote uniform oxidation timescales. However, turbulent mixing will increase undesirable wall interactions. We have thus added an annulus ring at the inlet of the PAM chamber for introduction of oxidant precursors and at the outlet to be used for gas-phase sample. The flow field is thus altered to improve uniform mixing and develops a better plug-type flow, which is desired for this particular type of chamber.
During the Fall 2013 lab experiments, we also combusted oak wood and leaves to test the combustion chamber residence times and resulting concentrations. We are currently beginning a new round of laboratory studies (Summer 2014) where we will utilize the combustion chamber as well as a passivated gas canister for introduction of pure VOC precursors and the PAM chamber in its modified flow configuration for repeat tests of isoprene oxidation under various conditions. We are also planning to test a full range of biogenic and anthropogenic precursor VOCs (e.g., alpha- and beta-pinene, limonene, beta-caryophyllene, xylene, toluene), as well as combustion and natural emissions from various leaves/needles, wood types, and other urban emissions (e.g. tobacco smoke, coffee roasting, food cooking).
Activity 4) A second field study (East St. Louis) which incorporates key lessons from the initial field study and extensive laboratory studies to perform in-field controlled oxidation experiments (utilizing the PAM reaction chamber) to test the sensitivity of ambient SOA formation and transformation to perturbations in oxidation intensity and precursor composition.
We performed an initial East St. Louis field study between August-October, 2013. We deployed a wide range of gas and particle instruments for chemical and physical characterization. Include instruments were: Williams Lab (WashU): VAPS, AMS, TAG, MC-TDMA, O3, SO2, NO/NO2/NOx, PAM chamber; Turner Lab (WashU): 7-channel Aethalometer, particle filters, full suite of meteorological equipment; Millet Lab (U.Minnesota): PTR-MS, CO; Weber Lab (GeorgiaTech): ACSM, MAAP, TEOM, NO2, particle filters, OC/EC; McMeeking (DMT): SP2.
Data processing is currently underway, but there are several highlights to report. East St. Louis provided an interesting contrast to the Centreville, AL, site in that there was persistent daily anthropogenic influence (e.g., elevated NOx, hydrocarbons, O3), with episodic large impact from biogenics, whereas the Centreville, AL, site had persistent biogenics with episodic anthropogenic influence. We observed occasional isoprene events in excess of 8 ppb, and due to a combination of strong photochemistry and transport time from the Ozark Mountains, isoprene and its first generation oxidation products (MVK, MACR) were highest at night. With NOx levels dropping over the last decade in the St. Louis region, it was observed on some nights with high isoprene, that O3 was not completely titrated away, likely allowing the formation of nitrate. On these nights, we observed particle growth events. We are currently processing VAPS data to determine if any organic nitrates were also observed on these nights, which could be contributing to the enhanced particle growth. PAM experiments were performed in the field for altered oxidation; however, limited days were covered due to an interest in obtaining sufficient unperturbed ambient samples. Future PAM experiments in the field will need to be automated to allow for fast time switching between ambient and altered oxidation.
Although it was only proposed to deploy to East St. Louis once during this project, we have high interest to perform another large-scale measurement campaign upwind and downwind of St. Louis, and have received good feedback from other SOAS investigators regarding their interest in studying this region.
In year one of this project, we completed development of a volatility and polarity separator (VAPS) for chemical characterization of organic aerosol and deployed it on two different field studies. We also built a custom multi-channel tandem differential mobility analyzer (MC-TDMA) for particle sizing, volatility, and hygroscopicity, and deployed it on a first field study. We completed an initial round of laboratory-based measurements of the oxidation of isoprene under varying concentrations, oxidant exposures, and mixtures with NOx. A major lesson in the first round of lab studies was to improve the design and operation of the PAM reaction chamber to ensure stable mixing and oxidation of precursor gases, and development of improved methods of calibrating oxidant exposures for each individual experiment.
Major observations from VAPS deployment to Centreville, AL, include observation of highly oxygenated aerosol and identification of several major oxygenated biogenic aerosol types, the largest being IEPOX SOA from low-NOx isoprene oxidation. VAPS also observed components of organonitrates and organosulfates, which will be investigated further and compared to observations from filter-based analyses (e.g., from Surratt et al. and Stone et al.). A wide array of measurements in East St. Louis were also made during this reporting period. Initial analysis reveals evidence of nighttime particle growth events when biogenic (isoprene) emissions (from the Ozark Mnts. to the southwest of St. Louis) were impacting the region along with elevated nitrate radical concentration. These nighttime growth events were always smaller in scale to daytime growth events a decade ago (Qian Shi, Masters Thesis with P. McMurry, U.Minnesota) but overwhelmed any evidence of daytime growth events during our measurement period. SO2 has dropped by approximately a factor of 5 and NOx has decreased by approximately a factor of 2 since that time period; however, O3 has actually slightly increased in the same time period. It is estimated that nighttime O3 is not completely titrated away as often as it once had been, leaving the potential for increased nighttime NO3 formation and thus increased nighttime chemistry, leading to particle production. Chemical characterization of this aerosol is currently underway utilizing the TAG, VAPS, and AMS along with offline filter analyses. This increase of nocturnal particle growth events demonstrates the non-linear atmospheric response to decreased anthropogenic emissions.
In the next reporting period, we will continue to improve VAPS and MC-TDMA operation, including implementation of a miniature-GC on the VAPS and faster mode-switching on the MC-TDMA (especially as we switch to hygroscopicity mode that currently takes additional time to stabilize relative humidity settings). We are currently performing a second intensive lab study using our emissions/combustion chamber, PAM reaction chamber, and a wide variety of gas and particle characterization instrumentation (including aerosol chemical characterization by VAPS and aerosol volatility/hygroscopicity by MC-TDMA). Finally, we will begin planning of a second field deployment to the St. Louis region and will work to scale up the project with hopes of having at least two field sites, one that would be upwind of the urban area relative to the Ozark Mountains and another downwind of the urban area to compare differences with the addition of anthropogenic pollutants.