Environ. Sci. Technol. 2003, 37, 3145-3151
Sampling Atmospheric Carbonaceous Aerosols Using an Integrated Organic Gas and Particle Sampler XINGHUA FAN,† J E F F R E Y R . B R O O K , * ,‡,§ A N D SCOTT A. MABURY† Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada, MSC of Environment Canada, 4905 Dufferin Street, Toronto, Ontario M3H 5T4, Canada, and Gage Occupational and Environmental Health Unit (GOEHU), University of Toronto, 223 College Street, Toronto, Ontario M5T 1R4, Canada
Measurement of particle-bound organic carbon (OC) may be complicated by sampling artifacts such as adsorption of gas-phase species onto particles or filters or evaporation of semivolatile compounds off the particles. A denuderbased integrated organic gas and particle sampler (IOGAPS), specifically designed to minimize sampling artifacts, has been developed to sample atmospheric carbonaceous aerosols. IOGAPS is designed to first remove gas-phase chemicals via sorption to the XAD-coated denuder, and subsequently particles are trapped on a quartz filter. A backup sorbent system consisting of sorbent- (XAD-4 resin) impregnated filters (SIFs) was used to capture the semivolatile OC that evaporates from the particles accumulated on the upstream quartz filter. A traditional filter pack (FP) air sampler, which uses a single quartz filter to collect the particles, was employed for comparison in this study. Elemental and organic carbon were determined from filter punches by a thermal optical transmittance aerosol carbon analyzer. Field measurements show that there was no significant difference between the elemental carbon concentrations determined by the FP and IOGAPS, indicating that particle loss during the transit through the denuder tube was negligible. Compared with the OC determined by FP (3.9-12.6 µg of C/m3), the lower OC observed on the quartz filter in the IOGAPS (2.2-6.0 µg of C/m3) was expected because of the removal of gas-phase organics by the denuder. Higher semivolatile organic carbon (SVOC) on the backup SIFs during the night (1.24-8.43 µg of C/m3) suggests that more SVOC, emitted from primary sources or formed as secondary organic compounds, partitions onto the particles during the night because of the decreased ambient temperature. These data illustrate the utility of an IOGAPS system to more accurately determine the particle-bound OC in comparison to FPbased systems. * Address correspondence to this author at MSC of Environment Canada; phone: (416)739-4916; fax: (416)739-5708; e-mail:
[email protected]. † Department of Chemistry, University of Toronto. ‡ MSC of Environment Canada. § GOEHU, University of Toronto. 10.1021/es026471y CCC: $25.00 Published on Web 06/07/2003
2003 American Chemical Society
Introduction Despite the ubiquity of organic carbon (OC) as a component of fine particles, very little is known regarding its composition, impacts in the environment, or impacts on human heath. In populated areas where human activities contribute significantly to OC, less than half of the organic species by mass have been identified (1, 2). Consequently, the relative contribution from primary emissions versus secondary formation and from anthropogenic versus natural sources is difficult to determine. Understanding of OC is also greatly complicated by the fact that many compounds are semivolatile and change phase readily depending upon atmospheric conditions (e.g., ambient temperature). Furthermore, measurement techniques tend to alter the true ratio of particle-bound to gas-phase OC, potentially leading to substantial sampling artifacts (3, 4). To minimize such sampling artifacts, several approaches to scrub gas-phase semivolatile organic carbon (SVOC) using a diffusion-based denuder prior to the collection of particles have been reported (5-9). Such techniques have been applied for sampling of atmospheric aerosols by several previous investigators. For example, Eatough and co-workers developed a multichannel, parallel plate diffusion denuder that is made up from charcoal-impregnated filter (CIF) strips (7, 10). Any SVOC off-gassing from the quartz filter placed downstream of the denuder is captured by a sequential CIF filter. This approach has been applied to the design of several samplers, including the low-volume Brigham Young University Organic Sampling System (BOSS) (10) and the high-volume BIG BOSS (7). More recently, the same group developed a Particle Concentrator-Brigham Young University Organic Sampling System (PC-BOSS) (11) by coupling the BIG BOSS with a Harvard particle concentrator (12). They operated PC-BOSS in Riverside and Bakersfield, CA (10, 11), and found that OC collected on an un-denuded quartz filter had a large positive artifact (∼25%) and that the amount of SVOC, which is often not collected by other samplers (i.e., the U.S. EPA FRM) was on the order of 50-60%. Therefore, the total PM2.5 mass determined by the PC-BOSS averaged 34% higher than the mass determined by the U.S. EPA FRM (10). In this paper, we evaluate a denuder-based “integrated organic gas and particle sampler”(IOGAPS) for differentiating between particle-bound and gas-phase OC. The IOGAPS samples particle-bound OC on a quartz filter downstream of a denuder that functions to remove gas-phase organic compounds (10, 13, 14). This negates the possibility for artifacts arising from interaction of gas-phase chemicals with the particle-laden filter. Evaporation of SVOC from the filter, which is accentuated by the upstream denuder, is quantified through the use of backup XAD-4-impregnated filters designed to retain the off-gassing organic compounds. The BOSS-type samplers utilize a particle concentrator to increase denuder efficiency. This necessitates a rather large scale-up to the OC and SVOC measurements on the order of 37%. The IOGAPS does not require an upstream particle concentrator, but the true efficiency of the XAD-coated denuder in capturing all organic gases (VOC and SVOC) and the effectiveness of XAD-impregnated filters (SIFs) are not known. Lewtas et al. attempted to compare the two techniques and found that the collection efficiency of the XAD denuder (eight-channel, 27-cm long) and charcoal denuder were 76% and close to 100%, respectively (15). The charcoal denuder was composed of 17 (4.5 cm × 58 cm) strips of charcoal impregnated filter (CIF) papers. They also found that the average loss of particulate SVOC evaporated off the particles collected on the quartz filter was about 30% of the VOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3145
FIGURE 1. Configuration of collocated samplers. Detailed description can be found in the text. PM2.5. While their results were consistent with expectations, they were based upon conditions in Seattle area, which has its own unique climate and amounts of ambient OC and SVOC. Furthermore, their results were inconclusive regarding the overall utility of the IOGAPS and regarding the optimum approaches for operating the IOGAPS and analyzing the samples. There are several concerns regarding the use of the IOGAPS for the measurement of particle-bound OC and elemental carbon (EC): (i) Do the particles suffer a significant loss during their transit in the denuder tube? (ii) Can the SIFs be analyzed for the total amount of SVOC that they capture? (iii) Does the denuder capture all of the gas-phase VOC and SVOC? (iv) Do the SIFs capture all of the SVOC lost from the quartz filter? (v) Does the SVOC associated with the particles desorb from the particles while in transit through the denuder tube? In this study we systematically examined the first four issues in the field (downtown Toronto in spring and summer 2001) and/or laboratory. We hypothesize that compared with a traditional filter pack (FP) sampler, which uses a single quartz filter to collect atmospheric particles, the IOGAPS approach can more accurately determine particle-bound OC. In addition to sampler evaluation results, we provide initial observations of the relative concentrations of nonvolatile and SVOC and their diurnal variations.
Experimental Section Integrated Organic Gas and Particle Sampler (IOGAPS). The IOGAPS is a diffusion-based air sampler, which consists of an XAD-4 resin-coated denuder (16, 17). The components of the IOGAPS are illustrated in Figure 1a. The eight-channel glass annular denuder with annuli spacing of 1.0, 1.0, 1.2, 1.0, 1.0, 1.0, 1.4, and 1.2 mm (from center outward) was available in two lengths, 28 and 60 cm (URG, Inc., Chapel Hill, NC). Ground XAD-4 resin (ca. 0.75 µm) for denuder coating was cleaned in advance with HPLC-grade organic solvents. Coating procedures were based upon the patent of Gundel et al. (16). The denuder was extracted after each sample collection period using a mixture of hexane, dichloromethane, and methanol (1:1:1, v/v). Each denuder was recoated after 10 samples even though the XAD-4 resin could remain on the denuder for at least 20 sampling and extraction cycles (18, 19). The normal configuration of the IOGAPS consisted of a cyclone inlet with 2.5-µm size cut to remove coarse particles followed by a denuder to retain gas-phase organic compounds. The IOGAPS was operated at a rate of 16.7 L/min. Particles (i.e., PM2.5) were collected on a prefired quartz filter 3146
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003
placed downstream of the denuder (Figure 1a). Three stages of XAD-4 resin-impregnated quartz filters (SIFs) were placed after the quartz filter to capture the SVOC that evaporates off the particles collected on the quartz filter. Blank prefired quartz filters and SIFs, which were handled identically to the sampled filters, including placing them in the sampling box in their Petri slides throughout the sample collection period, were collected with every sample. This is because gas-phase OC is known to be collected passively by prefired quartz filters (20) and probably in even larger amounts by the SIFs. All measured concentrations were corrected by their corresponding blank concentrations. Here we refer to these blanks as “travel blanks”. The measurement of OC and SVOC, determined from the quartz filter and SIFs less the corresponding travel blank values, could be improved by instead subtracting the OC and SVOC found in a breakthrough configuration operated in parallel (“dynamic blank”). However, this requires that two configurations of the IOGAPS (normal and breakthrough configurations) run at the same time. We opted for a less intensive compromise, that being the travel blanks because the data we collected during the breakthrough testing showed that the dynamic blank was very close to the travel blank if there was no gas-phase SVOC breakthrough from the denuder (Table 3). All filters used were 47 mm in diameter. Quartz filters (Type QM-A ultrahigh-purity quartz filters, from Whatman) were prefired at 700 °C for 10 h. SIFs were prepared by coating fine XAD-4 particles onto prefired quartz filters based on the procedures described by Gundel et al. (16). Briefly, each filter was dipped in a slurry of XAD in hexane, air-dried, dipped again, air-dried, rinsed twice with hexane, and air-dried. The prepared SIFs were kept in an airtight jar and stored in a freezer (
