| Abstract |
| Collaborators |
| Project Background |
| Objectives |
| Hypotheses |
| Approach |
| Instrumentation |
| Why Pittsburgh? |
| Maps |
| Links |
| Home |
Approach Ambient Monitoring
Instrument Development and Evaluation
Indoor Exposure Research
Epidemiological Study
Three-Dimensional Modeling Study
The central sampling site will be located next to the Carnegie Mellon University campus near downtown Pittsburgh. The site location is described in Section 8. The basic suite of sampling equipment at this site (FRMs, TEOMs, gas analyzers, etc.) will be provided by DOE/NETL by relocating the equipment at the Lawrenceville site of their UORVP. Satellite sites are planned for Holbrook, Pennsylvania (a rural area); Athens, Ohio; Morgantown, West Virginia; and South Park, Pennsylvania (a suburb of Pittsburgh). These four sites are currently being operated and will continue to be operated by NETL as part of their UORVP. Existing PM sampling sites of the regulatory network in the area will be used as additional satellite sites. Discussions are currently underway with the Air Quality Program of the Allegheny County Health Department to determine which regulatory network sites will be chosen for sampling. A letter from Roger Westman, Manager of the Air Quality program, is included in Appendix 1. The Health Department has agreed to increase the frequency of sampling to accommodate the proposed Supersite program, as explained in the letter. A letter from Jim Salvaggio, Director of the Bureau of Air Quality of the Pennsylvania Department of Environmental Protection, is also attached explaining their support of the proposed Supersite program. Figure 1 shows the location of the central site, the four satellite sites of UORVP, and some additional potential satellite sites.
Figure 1. Map of Pittsburgh region showing the approximate location of the Supersite, the four DOE/NETL UORVP sites that will be used as satellite sites, and five candidate satellite sites that are currently Allegheny County Health Department sampling sites.
The measurement campaign will last for 18 months (May 2001-October 2002) to include two summers and will consist of regular measurement periods and three 14-day intensive periods. The measurements can be categorized as follows:
PM size distributions Continuous PM composition PM Hygroscopicity PM mass characterization Bioaerosols Meteorology PM chemical composition Aerosol acidity Gases Single particle characterization Aerosol light scattering Fogs and clouds Each of these measurement categories is listed in Table 2 with instrumentation, frequency of sampling, and name of investigator. A brief discussion of each measurement category follows. Additional details can be found in Appendix 2.
PM Size Distributions
This category includes number and surface area size distributions. The Pandis group using a variety of real-time instruments will measure at the central site and at the Holbrook rural site number distributions. These instruments will provide data in the full range 3 nm to 10 mm (see Table 1). The separate overlapping distributions from these instruments will be inverted and combined using the MICRON code (Wolfenbarger and Seinfeld, 1989; Pandis et al., 1990; Weber et al., 1998). The Baltensperger group using Epiphaniometers will measure at the central site and at the Holbrook rural site surface area concentrations. This instrument provides a signal proportional to the Fuchs surface (Gäggeler et al., 1989; Baltensperger et al., 1997), namely, the surface of aerosols actually "seen" by a diffusing atom or molecule (Pandis et al., 1991; Rogak et al., 1991). The instrument operates continuously and averaging times of 10-30 min have been used in the past (Lugauer et al., 1998; Pandis et al., 1991). This will provide a unique opportunity to investigate potential health effects of aerosol surface area without assuming that the atmospheric particles are spherical.
PM Mass Characterization
The PM mass concentration will be measured by the Davidson group using the gravimetric approach and a variety of samplers. The PM2.5 and PM10 mass concentrations will be measured using Federal Reference Method (FRM) samplers, Tapered Element Oscillating Microbalance (TEOM) samplers, and denuder/filter systems. Dichotomous samplers with a 2.5 mm cut-off will be used for independent measurements of the coarse fraction (2.5< dp <10 mm). This will enable us to determine the extent to which the difference between PM10 and PM2.5 mass concentrations reflects the concentration in this size range.
An electrical low pressure impactor (ELPI) and a Micro-Orifice Uniform Deposit Impactor (MOUDI) will be used to sample PMx, i.e., mass concentrations associated with particle diameters less than x micrometers (where x will be determined by the cut-offs of the various impactor stages). These measurements will be performed daily during the regular sampling periods and five times per day during the intensive runs. Gravimetric analysis will be conducted in the class 100 Clean Lab at CMU used in several other studies (e.g., Davidson et al., 1993, 1996, Bergin et al., 1995). A humidity-controlled chamber will be used for equilibrating the filters according to standard protocol.
PM Chemical Composition
Inorganic Components. The inorganic speciation samplers (PM2.5 and PM10) for the Supersite will consist of a combination of denuders and filters. The sampler arrangement for PM2.5 is shown in Figure 2; a modified arrangement will be used for PM10. The figure lists several ionic species as well as organics (discussed below) and trace elements; the last category will include more than 20 species analyzed by Inductively Coupled Plasma Mass Spectrometry (ICP-MS). The sampling frequency will be once per day during the regular period and five times per day during the intensives, with 4-hour samples during the day
Figure 2. CMU PM2.5 sampler for chemical analysis. A similar system will be used for PM10 but without the potassium carbonate cellulose filter and without the quartz filter analyzed for organics. Dichotomous samplers will also be used to investigate the hypothesis that coarse particle concentrations can be determined from PM2.5 and PM10. Methods of analysis are listed in Table 1. and 6 hour samples at night. There will also be chemical analyses from the dichotomous sampler; the species will be chosen after quantifying uncertainties for different chemical species in the filter samples. A detailed characterization of the losses of semivolatile particulate matter in the denuders will be conducted during the first year of the project. These experiments will involve monodisperse model (ammonium nitrate, ammonium sulfate, and organics) and ambient particle size and composition measurements before and after the denuder. The results of these characterization experiments will allow better design of the denuder operation to minimize losses as well as provide information about the magnitude of the losses that can be used for correction of the measurements. Potential losses of HNO3 on the sampling system inlets and walls will be directly measured during the sampler development phase and the sampling system will be designed to minimize them.
MOUDI impactor samplers will also be analyzed for inorganic chemical composition during the intensive periods.
Besides these samples, OC and EC will be measured during intensive runs and during regular sampling periods by the Turpin group using their improved in situ carbon analyzer. The in situ carbon analyzer collects PM2.5 samples on a quartz fiber filter mounted inside a thermal-optical carbon analyzer. Thus, the analytical method is very similar to that used for the speciation sampler. Sample collection and analysis are performed continuously and automatically through computer control. The prototype instrument compares favorably with the in situ photoacoustic spectrometer developed at Ford Motor Company (Turpin et al., 1990), and has been used to investigate secondary organic aerosol formation (Turpin and Huntzicker 1991; 1995).
Two types of size-resolved organics sampling will be conducted; both restricted to periods of intensive runs. First, samples from an electrical low-pressure impactor (ELPI) will be analyzed for OC and EC by the Robinson group. Second, samples from a Hering Low Pressure Impactor (LPI) will be analyzed by the Turpin group by Fourier Transform Infrared Spectroscopy (FTIR) to provide information on the functional groups of the organics. Such information is valuable for understanding aerosol processes (Blando et al., 1998; Carlton et al., 1999).
Artifacts during sampling will be investigated by the Eatough group using the PC-BOSS diffusion denuder sampler (Eatough, 1999) during the intensive sampling periods. This sampler includes a PM2.5 inlet with a fine particle concentrator, denuder, and several filters (Quartz, Teflon, Nylon, and Charcoal-impregnated), as illustrated and discussed in Appendix 2. The sampler will provide data for OC and EC as well as semivolatile organics lost from particles during sampling. In addition, analyses will be conducted for sulfate, nitrate, and ammonium.
Special Organics Study: Polycyclic Aromatic Hydrocarbons (PAHs). A sampling system assembled by the Miguel group will be used during intensive runs to collect size-resolved samples for analysis for PAH compounds. Sampled air will first pass through an AIHL-design cyclone separator (John and Reischl, 1980) at 30 lpm allowing particles with diameter less than 1.8 mm to pass though it together with the gas-phase. A XAD-4 coated annular denuder placed downstream of the cyclone will trap semi-volatile organics while allowing transmission of particles smaller than 1.8 mm, which then flow into a MOUDI impactor. Blow off from the impactor's backup filter will be trapped with two polyurethane foam plugs (PUF) placed in series behind it. Samples will be extracted and analyzed using recently developed methods (Miguel et al., 1998; Marr et al., 1999; Schauer et al, 1996; Schauer et al, 1998; Schauer, 1998). All data will be corrected for chemical reactivity (Friedlander et al., 1996, Miguel et al., 1986). The ambient PAH concentrations at each site will be used by the Miguel group to separately apportion black carbon and associated toxic components to gasoline- and diesel-fueled vehicles using the EPA Chemical Mass Balance receptor model. PAH source signature data are available in the literature (Miguel et al., 1998, Marr et al., 1999, Venkataraman and Friedlander, 1994).
Single Particle Characterization
Single Particle Mass Spectroscopy. The groups of Wexler and Johnston will use an on-line single particle analysis technique (RSMS-II) to measure the particle-by-particle size and composition over the size range from 10 nm to 2 microns. As shown in Figure 3, particles of a narrow size range are focused aerodynamically to the source region of a mass spectrometer. The size that is focused can be selected from 10 nm to 2 microns by adjusting the upstream pressure. An excimer laser beam collinear with the particle beam is periodically fired. If a particle is in the beam, it is desorbed and ionized. The ions are analyzed in a time-of-flight mass spectrometer. Spectra from each particle are recorded and stored on a PC. The instrument is described in more detail in Carson et al. (1997) and Ge et al. (1998).
RSMS-II can analyze for a wide range of compounds and compound classes including a) speciation of inorganics such as metals and metal oxides, refractory crustal materials such as silicon dioxide, and electrolytic compounds such as sulfates and nitrates, b) speciation of aromatic organic compounds, and c) distinguishing elemental from organic carbon. A new RSMS-II instrument will be built for the Supersite that will require relatively little operator intervention and will have improved analytical capabilities. This instrument will be deployed for entire sampling period. Further information about RSMS-II and its operation is given in Appendix 2.
The Wexler/Johnston group plan to expand the capabilities of the RSMS-II to examine the polar organic compounds in single particles using an aerosol matrix-assisted laser desorption ionization (MALDI) technique (see, e.g., Mansoori et al., 1996). MALDI is a widely used technique to obtain mass spectra of highly polar compounds. The proposed on-line, single particle experiment mixes ambient aerosol with a small flow of hot vapor matrix material that strongly absorbs the laser radiation. The matrix vapor will condense on aerosol particles creating a matrix/analyte mixture similar to that created in a conventional MALDI experiment. The matrix-aerosol mixture will then be irradiated with a pulsed laser beam to eject intact analyte ions without fragmentation, which are then characterized with the mass spectrometer.
Figure 3. Schematic of RSMS-II, an on-line single particle technique for sizing and analyzing particles from 10 nm to 2 µm.
Laser-Induced Breakdown Spectroscopy (LIBS). The Buckley group will use Laser-Induced Breakdown Spectroscopy (LIBS) to measure the elemental composition of single particles in the atmosphere. In LIBS, a tightly focused, pulsed laser is used to create a microplasma. The high temperature of the plasma breaks all of molecular bonds, and thermally excites the electronic states of the atoms. Using atomic emission, the mass concentration of a number of elements (Be, Cd, Cr, Na, K, V, Ni, Si and Pb) within the spark volume can be quantitatively measured (Buckley et al., 1999). To concurrently measure particle size, a single-shot spectrum containing emission from a single particle is analyzed, and the total mass of each atomic constituent is determined (Hahn, 1998). The LIBS system will be deployed during two of the intensive sampling periods and during 1 month of the baseline-sampling period. Additional information about the system can be found in Appendix 2.
SEM analysis. During the intensive periods, samples will also be collected for analysis by scanning electron microscopy (SEM). Analyses will be conducted at the laboratories of R.J. Lee, Inc. Under the conditions required for SEM analysis, wet aerosol particles lose their water rapidly. During this rapid evaporation their morphology changes significantly and the SEM images are not representative of the actual particle shape. However, during periods of low relative humidity the SEM morphology information is expected to be relatively accurate. This will yield complementary information about the elemental composition and morphology (during periods of low RH) of individual larger particles that will not be provided by the single particle spectrometer.
Continuous and Semi-Continuous PM Composition
In addition to the previously mentioned continuous (or semi-continuous) measurements of the aerosol size distribution, OC and EC, and single particle size and composition a number of additional state-of-the-art techniques will be further developed and used in the proposed Supersite program.
Semi-continuous elements. The Ondov group will measure the concentration of 18 metals species (As, Cu, Mn, Ni, Cr, Cd, Se, Ag, Pb, Al, Fe, Zn, Ca, Bi, V, Ti, Be, and Ba) during the intensive periods at both the central Supersite and the rural satellite site in Holbrook. The semi-continuous system consists of a high-frequency aerosol sampler (HFAS) (dynamic aerosol concentrator) and a true simultaneous multi-element Graphite Furnace Atomic Absorption (GFAA) spectrometer. After as little as 10 minutes of sampling (depending on ambient concentrations), enough slurry is collected to permit 4 suites of 4 or 5 elements to be determined, each in triplicate. In addition to high temporal resolution, tests with NIST Standard Reference Material 1648 ("Urban Particulate Material") confirm that analytical concentration measurements are accurate. The initial target collection rate will be 6 samples per hour, pending ambient concentrations.
Nitrates, sulfates, and aerosol carbon. Aerosol Dynamics Inc. (ADI) will provide automated, near-continuous measurements of aerosol nitrate, sulfate, and carbon in airborne particles below 2.5 µm diameter over the 18-month sampling period following the method of Stolzenburg and Hering (1999). With this method, the aerosol stream is first denuded to remove interfering vapors such as vapor organics. The sample stream is then humidified to prevent bounce-off (Winkler, 1974; Stein et al., 1994) and the particles are collected by impaction on a metal substrate inside an integrated collection and vaporization cell (ICVC). At the end of collection the cell is filled with a carrier gas, the substrate is heated resistively, and the particles thermally decompose to vapors that are measured using a commercial gas phase analyzer. A unique flow system eliminates the need for valves on the sampling line.
Sulfate analysis will be achieved using air as the carrier gas, with quantification of the evolved vapors for SO2, as described by Roberts and Friedlander (1974). Nitrate analysis will be achieved by low-temperature heating in nitrogen carrier gas, with analysis of the evolved nitrogen oxides as described by Yamamota and Kosaka (1994). Carbon will be detected through quantification of the evolved CO2. Because the sample is concentrated, the analyses can be done using proven, robust gas analyzers. Results using this method compare favorably with more conventional methods as illustrated in Appendix 2.
Bioaerosols
The Hernandez group will investigate the combined utility of high volume sampling, direct epi-fluorescent microscopy, and newer molecular biology methods to characterize outdoor bioaerosols in the size range between 0.2-20 µm. Initial laboratory experiments will be performed to determine the effects of sample capture on the structure, survival and activity of indigenous atmospheric microbioaerosols. Once the response to high volume sample capture is determined, field studies will be initiated. Multi-season surveys will be performed with samples collected at the Pittsburgh Supersite. These surveys will: (i) determine the effects of common sampling devices on outdoor microbiological aerosols during extended sampling periods, (ii) obtain accurate measurements of microbiological particulate inventories in outdoor air including, biomass, volume, and size distributions (between 0.2- 20 µm), and (iii) determine the abundance and identity of microbioaerosols (bacteria, fungi, and their spores) using direct microscopy combined with novel molecular biological assays. Details are presented in Appendix 2.
Aerosol Acidity
The aerosol acidity will be estimated using the method proposed by Saxena et al. (1993) combining the measured aerosol composition with the state-of-the-art aerosol thermodynamic models GFEMN, AIM2, and SCAPE2 (Ansari and Pandis, 1999). This method provides a better characterization of the actual ambient aerosol acidity compared to the extractable acidity by the pH method (e.g., using the method of Koutrakis et al., 1988). These estimates will be available for all aerosol-sampling periods, with the intensive periods providing the most useful information. The extractable acidity will be measured for selected samples by the Eatough group using the PC-BOSS and the pH method of Koutrakis et al. (1988).
Aerosol Light Scattering
The aerosol scattering coefficient and backscatter will be measured using a three wavelength (450, 550, and 700 nm) integrating nephelometer (TSI Model 3653) (Bodhaine et al., 1991). Periodically, an automated ball-valve built into the inlet will be activated to divert the air sample through a high-efficiency filter, allowing the measurement of the particle-free air signal. The sample temperature and relative humidity will be measured inside the nephelometer and will be kept practically equal to the ambient conditions (for RH<95%). The visual range will also be measured during the measurement periods and additional observations will be collected from the airports in the area and archived. Pictures (in electronic form) of the area surrounding the Supersite will be taken every hour during the intensive sampling period and every six hours during the rest of the study period for the calculation of the visual range and documentation of the prevailing conditions.
PM Hygroscopicity
The ability of ambient fine particles to absorb water and grow will be quantified using the Tandem Differential Mobility Analyzer (TDMA) technique (Zhang et al., 1993) by the Pandis group. These measurements during the intensive periods will provide the growth factors as a function of particle size and particle group. The measurement of aerosol size as function of RH (from around 10 to 95 percent) will allow the quantification of the aerosol liquid water content at the RH of PM mass measurement and will provide input for the visibility calculations. The Cloud Condensation Nuclei concentration will be measured using the CCN counter of DH Associates. The CCN concentration is directly related to the ability of the particles to become cloud droplets and thus provides valuable information about their atmospheric lifetimes.
The TDMA system will be combined with the single particle instrument during specific periods to establish the links between the hygroscopic particle properties and their chemical composition. Two experiments are planned. In the first monodisperse particles will be selected by the first DMA and then they will be humidified to a RH of around 90%. The second DMA will be used for the quantification of the size change of the particles (how many particles grew how much) and the RSMS-II will be used for the measurement of their chemical composition. In the second experiment the particles will be pre-humidified (RH around 80-90%) and then a monodisperse part will be selected by the first DMA. The particles will then be exposed to a low relative humidity environment (around 10%). The second DMA will quantify once more their size change (how may particles lost how much water) while the RSMS-II will measure the chemical composition of the various particle groups. These experiments will provide the link between chemical composition and hygroscopic properties of individual ambient particles.
Meteorology
Several meteorological parameters will be measured during the sampling period, including temperature, relative humidity, precipitation, wind speed and direction, UV intensity, and solar intensity. In addition, the Kahl group will compute backward 10-day airmass trajectories twice daily over the 18-month sampling period (Harris and Kahl, 1994). This will provide (i) an assessment of the day-to-day variability in transport pathways and source regions for air sampled at the Supersite, (ii) an assessment of the dependence of aerosol chemical composition on the possible source regions, and (iii) an atmospheric transport climatology for the Pittsburgh region. Additional information on the trajectories is provided in Appendix 2.
Gases
Several gases will be measured continuously during the 18-month period and reported as 1-hour average concentrations. These include O3, NO, NOx, CO, and SO2. In addition, air will be collected in canisters and analyzed for VOCs by GC-FID and GC-MS techniques. Methods of Lewis et al. (1999) will be applied here; these techniques enabled quantification of roughly 130 compounds in Atlanta air. Samples will be collected over six-day periods for analysis, except during intensive runs where there will be five samples collected and analyzed per day. In addition, the Collett group will measure hydrogen peroxide and soluble organic peroxides using a monitor based on the method of Lazrus et al. (1986). Details of the VOC analyses and the peroxide measurements are given in Appendix 2.
Fogs and Clouds
The Collett group will measure the cloud and fog composition during the winter period using the compact version of the Caltech Active Strand Cloudwater Collector known as the CASCC2 (Demoz et al., 1996). Fog will be sampled at the central site while cloudwater will be sampled on the roof of the 160 m high Cathedral of Learning at the nearby University of Pittsburgh. The building is often immersed in clouds, which will give us a unique opportunity to characterize the cloud composition in the area.
Collected fog/cloud samples will be analyzed on-site for pH and sample aliquots will be prepared for later analysis of major anion and cation concentrations at CSU. A subset of samples will also be aliquotted and stabilized for later analysis of total organic carbon (TOC), formaldehyde, and trace metal catalysts (Fe and Mn).
Green Design Considerations
The proposed sampling program will have adverse environmental effects, including use of resources and energy as well as production of pollutants. We will take the first steps in an effort to minimize these effects to the extent possible without compromising the quality of the data. For example, we will identify those materials where production is the most resource and energy intensive, and will investigate whether substitution of alternate materials is feasible. We will also consider the energy consumption of the project and determine whether alternative strategies can save energy. This includes energy used by the sampling equipment, air conditioning, clean laboratories, analytical instruments, and transportation of people and equipment. The effort will be assisted by faculty and students in the Green Design Initiative at Carnegie Mellon where there have been several projects on minimizing environmental effects of products and processes (e.g., Hendrickson and McMichael, 1994; Lankey and Davidson, 1997).
Table 1. Measurements in the Pittsburgh Supersite Program
IntensivesAerosol number distribution Ultrafine SMPS 1, SMPS 2, APS3, ELPI 4, Ultrafine CPC 5 Aerosol surface distribution Ultrafine SMPS, SMPS, APS, ELPI, Ultrafine CPC,Epiphaniometer Aerosol volumeDistribution Ultrafine SMPS ,SMPS, APS, ELPI, Ultrafine CPC
PM2.5 mass FRM 6, TEOM 7, CMU Sampler8, LPI 9, MOUDI 10 PM10 mass FRM, TEOM, CMU Sampler, LPI, MOUDI PMx mass LPI, MOUDILPI, MOUDI
PM2.5 ions and elements CMU Sampler/ IC 11 & ICPMS12 PM2.5-10 ions, elements CMU Sampler/ IC & ICPMS HNO3 vapor CMU Sampler/ IC & ICPMS NH3 vapor CMU Sampler/ IC & ICPMS
Size-resolved ions and metals MOUDI/IC and ICPMS
PM 2.5 OC and EC 13 Organic sampler/thermal PC-BOSS system 14 In-situ carbon analyzer PM10 OC and EC Organic sampler/thermal R&P sampler 15 PMx OC and EC ELPI/thermal Organic speciation Organic sampler/GC-MS 16 Organic size-resolved characterization LPI/FTIR 17 Polycyclic Aromatic Hydrocarbons MOUDI-PUF system 18/GC-MS, HPLC 19-fluorescence Polar Organics RSMS-II 20
Single Particle RSMS-II Chemical Composition Filter/SEM 21 LIBS
Semi-continuous metals HFAS/GFAA 22 Continuous nitrate ICVC Continuous sulfate ICVC Continuous carbon ICVC
Bioaerosols Epi-fluorescent microscopy,Molecular biology assays
Acidity Filter-Thermodynamics
Visibility Nephelometer
Photos/Visual Range
6 hr
1 hr
Growth with RH TDMA 23/RSMS-II CCN concentration CCN Counter 24
RH, T, Wind UV and Solar Trajectories
VOCs GC-FID 25, GC-MS Hydrogen Peroxide CSU Monitor Organic Peroxides CSU Monitor O3 Ozone Monitor NO and NOx NOx Monitor SO2 SO2 Monitor, Filter CO CO Monitor
Fog and cloud composition CSU Collector 26 See Appendix 2 for a list of instrumentation acronyms
Instrument Development and Evaluation
The proposed Pittsburgh Supersite program will allow the further development of a number of PM measurement methods, provide comparisons among methods to be used in the next few years, and will serve as a platform for field comparisons of emerging methods that have the potential of addressing current PM measurement needs.
Further development of methods
Single Particle Measurements (RSMS-II). A number of improvements will made to the current single particle instrument of the University of Delaware to make it more suitable for long-term Supersite use. A new RSMS-II instrument will be built for the Supersite that is relatively free of operator intervention and has improved analytical capabilities. A power monitor will be added to the rear of the laser to record pulse energies. The on-site computer will record the pulse power with each spectrum and if the power becomes too low automatically refill the laser gas. The instrument will record inlet flow and pressure and store them with each spectrum. The relationship between flow and pressure will be checked by the computer. Off calibration will indicate that the flow/pressure control orifices are clogged and need to be cleaned. Extra orifices of the smallest sizes will be included with each instrument since these clog more easily. Operator intervention to change orifices is straightforward.
Epiphaniometer. A calibration method for the conversion of the epiphaniometer signal to aerosol surface area will be developed by the Baltensperger and Pandis teams based on the laboratory work of Pandis et al. (1991) and Baltensperger et al. (1997). This will allow the use of the epiphaniometer as a continuous surface area monitor in urban areas.
Organic Aerosol Speciation. The Rogge team will add one more derivatization step, silylation using BSTFA [bis(trimethylsilyl)trifluoroacetamide], to their state-of-the-art organic extraction procedure. A suitable derivatization procedure will be developed that can be used in sequence with the current derivatization technique that uses diazomethane. Furthermore, a library of organic compounds, which are susceptible to BSTFA derivatization, will be generated that contains ion fragmentation patterns for those target derivatized compounds. This approach will among others detect levoglucosan, a tracer for cellulose formed from biomass burning (Simoneit et al., 1999).
Aside applying whole sample derivatization, a method will be developed based on the work of Simoneit et al. (1999) that fractionates sample extracts according to polarity into up to eight fractions that then will be derivatized and analyzed separately. By fractionating a sample extract according to polarity, more single compound resolution is obtained that will allow us to increase the pool of identifiable compounds.
Polar Organics. The University of Delaware team will develop a technique (described in 4.2.4) for the continuous measurement of the polar organic aerosol concentration on single particle basis.
Continuous-measurements. The proposed program will allow the continued development of a number of continuous PM measurement techniques to make them suitable for long-term monitoring. These methods include the continuous elemental (Ondov, Buckley), sulfate, nitrate, and carbon (Hering), ultrafine PM (Pandis), OC and EC (Turpin) measurements.
Bio-aerosols. The Hernandez group will develop rapid, quantitative aerosol assays to characterize the identity, distribution, and activity of microbiological components present in ambient aerosols.
Comparisons and evaluation of methods
The existence of several overlapping techniques will allow the intercomparison of existing measurement approaches and also the evaluation of new and emerging approaches. These intercomparisons are summarized in Table 3.
Table 3. Comparison of methods
Observable Methods Number distribution SMPS - ELPI - APS - RSMS-II Surface area Epiphaniometer - SMPS/APS - ELPI PM2.5 and PM10 mass FRM - TEOM - Speciation Sampler - LPI - ELPI - MOUDI PM Elements Speciation Sampler/ICPMS - HFAS/GFAS - RSMSII - LIBS PM Sulfate, Nitrate Speciation Sampler/ICPMS - ICVC -RSMSII PM Carbon Speciation Sampler - Rutgers Sampler - ICVC - PC BOSS -RMSII - R&P Sampler Polar Organics Detailed Speciation - FTIR - RSMSII/MALDI
Researchers in the Indoor Environment Department (IED) at Lawrence Berkeley National Laboratory (LBNL) will conduct an indoor PM2.5 study in Pittsburgh in conjunction with the Supersite program (not supported by the EPA Supersite funds). A letter of intent from LBNL is attached. The objective of this work is to better understand the relationship between ambient PM2.5 and actual exposures. The proposed work in Pittsburgh will build upon previous and current indoor aerosol studies conducted by LBNL. In particular, it will build upon a just-initiated LBNL project to develop and test a semi-empirical model for predicting indoor PM concentrations based on outdoor PM measurements. This LBNL project will be conducted in collaboration with the on-going San Joaquin Valley study (see, for example, Chow et al., 1993) and the new Supersite study in Fresno, CA (located in the San Joaquin Valley). The LBNL research group has also just initiated a complementary study with EPA sponsorship to investigate factors influencing indoor exposures to particles of outdoor origin.
The proposed indoor study in Pittsburgh is an important extension to the current LBNL research because of differences between Pittsburgh and the San Joaquin Valley. Significant differences include: differences in housing stock (e.g. older, with different construction), differences in climate (e.g. colder wintertime temperatures and higher summertime relative humidities), and differences in PM composition (e.g. sulfate instead of nitrate). As is the case in the LBNL San Joaquin Valley project, the proposed Pittsburgh study would take advantage of the extensive and detailed outdoor data collected by the intensive Supersite study. Knowledge about the composition and temporal variability of the outdoor 'source term' is critical to understanding the exposures indoors.
The indoor study will employ both experiments in an unoccupied research house and monitoring in one or more occupied homes. The research house will permit detailed examination of some of the factors influencing aerosol transport into and behavior within the house under controlled conditions. Examples of these factors include aerosol characteristics, housing characteristics, and weather conditions. Detailed measurements will include indoor aerosol size and concentration profiles and chemical speciation, as well as building and indoor environment characteristics. The research house will be located close to the central Supersite. For the duration of the Supersite monitoring period, a more limited set of instruments will be used in one or more occupied homes or other residential buildings to examine the effect of human activity on relationship between indoor and outdoor PM.
An important result of this work will be the ability to test the semi-empirical indoor-PM2.5 model derived from the LBNL San Joaquin Valley study with data from Pittsburgh. Such a model would enable estimation of indoor exposures from the Supersite data. Detailed planning for the indoor study will be based on results from the LBNL study in the San Joaquin Valley (sampling is expected to begin 11/99) and on the detailed and intensive outdoor measurements to be performed at the Pittsburgh Supersite.
The proposed monitoring study with its wide range of epidemiology-relevant measurements (see Tables 1 and 2), its daily sampling schedule, and 18 month duration in a populated urban area will result in a dataset of high quality and quantity. Availability of not only PM2.5, PM10 mass concentration and composition, but also PMx, number distributions including the ultrafine size range, surface area (measured directly with the epiphaniometer), surface area distribution, organic speciation and polar organics, concentrations of aerosols from specific sources, etc., will provide a comprehensive data set for epidemiological studies. Samet and his group (in a study not funded by the Supersites program) will address the effects of particulate air pollution on readily available health indicators, including daily mortality counts and morbidity measures, such as hospitalization rates in general or among the elderly, rates of emergency room utilization, and rates of outpatient utilization. A number of publicly available data resources will be used for such analyses, supplemented by data from health care organizations or other provider networks. The basic study design will draw on time-series methods, already widely applied in air pollution research. The Pittsburgh Supersite program will provide a unique opportunity to discover the size, composition, and source of the particulate matter that most likely causes observed increases in mortality and morbidity. This first level analysis using traditional timeseries analysis techniques will be funded with available resources of the Johns Hopkins group. Additional resources for a panel epidemiological study in the Pittsburgh region will be requested by other funding agencies as soon as the EPA funds are committed. Groups of potentially susceptible individuals (persons with asthma, both children and adults, chronic obstructive pulmonary disease (COPD), ischemic heart disease, etc.) from the Pittsburgh community will be identified and their health status prospectively monitored and evaluated in relation to air pollution concentrations. The basic design will involve the identification of a feasible number of patients, generally ranging from 20 to 100, who are enrolled into a protocol involving daily assessment of health status, typically using both questionnaires and physiological measurements, and estimated exposure. A letter by Samet is included in the proposal.
The EPA Supersites program would be most informative if parallel data collection efforts were initiated at all sites. Coordinating as much as possible the various Supersites to use a wide range of common instrumentation, will allow the use of all Supersites data in conjunction to produce a much richer data set for epidemiological study. If gradients of heterogeneity in exposures within the selected communities could be defined, then analyses could be further stratified by location.
Three-Dimensional Modeling Study
The Pandis and Davidson groups are members of the EPA-STAR funded research consortium (Research Consortium on Ozone and Fine Particle Formation in California and in the Northeastern United States) investigating the interactions between the ozone and the PM problems in the Eastern United States. The objective of this research is to advance the understanding of emissions/air quality relationships for ozone and fine particles in the two most populous areas of the United States: California and the Northeastern states. The models developed are currently applied in both to study the effects of emissions controls on both ozone and fine particle air quality.
In collaboration with Ted Russell in Georgia Tech., these groups have created a comprehensive three-dimensional model for the study of PMx in the region (see Appendix 2 for a description). The model includes a state-of-the-art description of PM processes and describes the complete aerosol size/composition distribution using user-selected chemical and size resolution. The model is also coupled to a sensitivity analysis module so it can calculate directly the sensitivities of PM concentrations to small changes in source strength.
The Pittsburgh Supersite Program will use the results of this parallel activity. For the EPA-STAR program the region around Pittsburgh will be described with high spatial resolution (5x5 km) and selected simulations will be run for the intensive periods focusing on:
- the lifetime and transport distances of PM reaching Pittsburgh
- the relationships between total nitric acid and sulfate in Pittsburgh and the NOx and SO2 emissions in the modeling domain, and
- the sensitivity of the PM in the area to these NOx, SO2, and VOC emissions
- the contribution of primary organic aerosol sources to the OC in Pittsburgh
- the contribution of primary and secondary biogenic aerosol to the organic PM in Supersite area.