Yanbo Pang

High-Volume Diffusion Denuder Sampler for the Routine Monitoring of Fine Particulate Matter: I. Design and Optimization of the PC-BOSS

Fine particulate mass will frequently be significantly underestimated in urban environments using current EPA federal reference method (FRM) single filter samplers because of loss of semivolatile organic material and nitrate from the particles during sampling. Previously described diffusion denuder samplers, such as the Brigham Young University organic sampling system (BOSS) and BIG BOSS, eliminate both positive and negative sampling artifacts by removing the gas-phase material before fine particulate mass is collected on a quartz filter and a sequential sorbent filter or bed. However, these samplers are not amenable to routine field use because they require multiple systems to accurately account for the efficiency of the denuder for each sample. The denuder efficiency is substantially improved by combining a conventional 2.5 w m cutpoint inlet and a low cutpoint virtual impactor particle concentrator to provide concentrated fine particles (0.1-2.5 w m) to a BOSS diffusion denuder sampling system (PC-BOSS). The performance of both the particle concentrator and the diffusion denuder in the PC-BOSS has been evaluated. Under optimized conditions, the minor to total flow ratio for the concentrator is 25% and the concentrator efficiency is 70-90% for sulfate and 60-85% for soot, depending on the particle size distribution of these species. The particle loss in the concentrator is about 6%. The particle loss in the denuder system is negligible. At the recommended optimum minor flow (concentrated particles, 25% of the total flow), most of the gas-phase material is discarded by the particle concentrator before the concentrated fine particles enter the diffusion denuder. At an optimized inlet flow of 150 L/min with a flow through the denuder of 38 L/min, the overall efficiency of the PC-BOSS for the removal of gas-phase organic material, SO 2 , and HNO 3 is over 99%. Therefore independent determination of the denuder breakthrough can be eliminated in the PC-BOSS.

Semi-volatile species in PM2.5: comparison of integrated and continuous samplers for PM2.5 research or monitoring.

ABSTRACT Fine particles in urban atmospheres contain substantial quantities of semi-volatile material [e.g., NH4NO3 and semi-volatile organic compounds (SVOCs)] that are lost from particles during collection on a filter. Several diffusion denuder samplers have been developed for the determination of both NO3 - and organic semi-volatile fine particulate components. The combination of technology used in the BOSS diffusion denuder sampler and the Harvard particle concentrator has resulted in the Particle Concentrator-Brigham Young University Organic Sampling System (PC-BOSS) for the 24-hr (or less) integrated collection of PM2.5, including NH4NO3 and semi-volatile organic material. Modification of the BOSS sampler allows for the weekly determination of these same species. Combination of BOSS denuder and tapered element oscillating microbalance (TEOM) monitor technology has resulted in the real-time ambient mass sampler (RAMS) for the continuous measurement of PM2.5, including the semi-volatile components. Comparison of the results obtained with the BOSS and with each of the newly developed modifications of the BOSS indicates that the modified versions can be used for the continuous, daily, or weekly monitoring of PM2.5, including semi-volatile species, as appropriate to the design of each sampler.

Effect of Semivolatile Material on PM 2.5 Measurement by the PM 2.5 Federal Reference Method Sampler at Bakersfield, California

The chemical composition of fine particulate material was determined for samples collected in Bakersfield, CA, during February-March, 1998 using several diffusion denuder samplers, including the PC-BOSS, which measures both semivolatile fine particulate nitrate and organic material. An average of 56% of the fine particulate carbonaceous material was lost from the filters of the Particle Concentrator-Brigham Young University Organic Sampling System (PC-BOSS). A comparable amount of fine particulate semivolatile organic material was also lost from collected particles with single filter samplers, such as the PM 2.5 Federal Reference Method. The fraction of nitrate lost from collected particles was a function of temperature and humidity, with the biggest effect being due to temperature. The fraction of nitrate lost was comparable for conventional annular denuder samplers and the PM 2.5 FRM, averaging 33%. The nitrate loss from particles for the PC-BOSS was smaller, averaging 11%, possibly due to the concentration of particulate material prior to collection with this sampler. The loss of nitrate and semivolatile organic material during sample collection resulted in the PM 2.5 FRM sampler giving PM 2.5 mass that was an average of 30% (7.3 w g/m 3 ) lower than the true value and different from the true value from negligible to 20 w g/m 3 .

High-Volume Diffusion Denuder Sampler for the Routine Monitoring of Fine Particulate Matter: II. Field Evaluation of the PC-BOSS

The high-volume Brigham Young University organic sampling system with a particle concentrator (PC-BOSS) has been field evaluated for the determination of airborne fine particulate matter including semivolatile chemical species during 3 intensive sampling programs in 1997: Tennessee Valley Authority (TVA), Lawrence County, TN; Riverside, CA; and Provo, UT. The PC-BOSS precision was tested using 2 collocated PC-BOSS samplers. In addition, the PC-BOSS results were compared with results from a prototype PM 2.5 U.S. EPA federal reference method (FRM sampler), a filter pack sampler (quartz and charcoal sorbent filters), the BIG BOSS, an annular denuder sampler, and the ChemSpec sampler for the determination of major fine particulate species. Fine particulate mass, sulfate, nitrate, and organic carbonaceous material (OC) determined by 2 PC-BOSS samplers agreed within - 10%. Possibly due to absorption of SO 2 by a quartz filter, the sulfate concentrations determined by the filter pack sampler and the BIG BOSS were higher (by 10 - 3%) than concentrations obtained with the other samplers. No absorption of SO 2 (g) by the quartz filters of the PC-BOSS occurred due to the high efficiency (>99%) of its denuder. The PC-BOSS, annular denuder, and ChemSpec samplers agreed with each other (to within - 0.5 w g/m 3 , - 17%, with no bias) for the determination of fine particulate nitrate concentrations, including volatilization losses. The prototype PM 2.5 FRM sampler collected only particle-retained nonvolatile mass. The mass concentrations determined by the PM 2.5 FRM agreed with those collected by the post-denuder Teflon filters of the PC-BOSS (to within - 1.1 w g/m 3 , - 10%, with no bias). The overall loss of material from particles and the resultant underestimation of the particulate mass concentrations by the PM 2.5 FRM depended on the fine particle composition and the ambient temperature.

Evaluation of the RAMS Continuous Monitor for Determination of PM2.5 Mass Including Semi-Volatile Material in Philadelphia, PA

Abstract The real-time ambient mass sampler (RAMS) is a continuous monitor based on particle concentrator, denuder, drier, and tapered element oscillating microbalance (TEOM) monitor technology. It is designed to measure PM2.5 mass, including the semi-volatile species NH4NO3 and semi-volatile organic material, but not to measure PM2.5 water content. The performance of the RAMS in an urban environment with high humidity was evaluated during the July 1999 NARSTO-Northeast Oxidant and Particles Study (NEOPS) intensive study at the Baxter water treatment plant in Philadelphia, PA. The results obtained with the RAMS were compared to mass measurements made with a TEOM monitor and to constructed mass obtained with a Particle Concentrator-Brigham Young University Organic Sampling System (PC-BOSS) sampler designed to determine the chemical composition of fine particles, including the semi-volatile species. An average of 28% of the fine particulate material present during the study was semi-volatile organic material lost from a filter during particle collection, and 1% was NH4NO3 that was also lost from the particles during sampling. The remaining mass was dominantly nonvolatile (NH4)2SO4 (31%) and organic material (37%), with minor amounts of soot, crustal material, and nonvolatile NH4NO3. Comparison of the RAMS and PC-BOSS results indicated that the RAMS correctly monitored for fine particulate mass, including the semi-volatile material. In contrast, the heated filter of the TEOM monitor did not measure the semi-volatile material. The comparison of the RAMS and PC-BOSS data had a precision of ±4.1 μg/m3 (±9.6%). The precision of the RAMS data was limited by the uncertainty in the blank correction for the reversible adsorption of water by the charcoal-impregnated cellulose sorbent filter of the RAMS monitor. The precision of the measurement of fine par-ticulate components by the PC-BOSS was ±6-8%.

Evaluation of the Performance of Annular Denuder Samplers

Two annular denuder samplers were evaluated for the determination of atmospheric nitric acid, SO 2 , and fine particulate sulfate and nitrate. One, the University Research Glassware (URG) sampler, used an elutriator-impactor to remove coarse particles prior to sampling the aerosol through the denuder sections for the removal of acid gases. The second, the Rupprecht and Patashnick (R&P) ChemSpec, removes the acid gases in diffusion denuder sections and then removes coarse particles prior to the collection of fine particles. Possible sampling artifacts examined included the removal of HNO 3 (g) by the inlet of the URG sampler before the collection of this gas-phase species by the diffusion denuder, resulting in the underestimation of HNO 3 (g) and the deposition of some coarse particles in the R&P denuders prior to the removal of these larger particles, and the collection of fine particles on a filter, resulting in the overestimation of HNO 3 (g) and SO 2 (g). Samples were collected in Riverside, CA, Bakersfield, CA, and Provo, UT. Both sampling artifacts were found to be present. However, the effect on the determination of HNO 3 (g) by the URG sampler and on the determination of particulate-phase sulfate and nitrate by the R&P ChemSpec were both small, on the order of a few tenths of a microgram per cubic meter.