Chemical Analyses of Particle Filter Deposits* - Aerosols: Measurement, Dosimetry, and Health Effects - page 190

U.S. ambient networks (Flanagan et al., 2006; Hansen et al., 2006; Watson, 2002) and most of the
source proiles in the U.S. EPA's SPECIATE library (U.S. EPA, 2008) have applied the IMPROVE_
TOR or IMPROVE_A_TOR (Chow et al., 1993b, 2004a, 2005a, 2007a, 2011b) method (DRI Model
2001 Thermal/Optical Carbon Analyzer, Atmoslytic, Inc., Calabasas, CA). This protocol heats the
sample in steps, irst in an inert 100% helium (He) atmosphere, to obtain OC evolving between
ambient temperature (∼25°C) and 140°C (OC1), 140°C-280°C (OC2), 280°C-480°C (OC3), and
480°C-580°C (OC4). The analysis atmosphere is then changed to a 98% He/2% O 2 composition,
producing three EC fractions at 580°C (EC1), 580°C-740°C (EC2), and 740°C -840°C (EC3). The
evolved carbon peaks, detected by an FID as CH 4 , are deined by their return to a stable baseline;
and analysis time depends on the composition of each peak rather than on preset windows. These
carbon fractions have been found useful as source markers (Cao et al., 2005b; Kim and Hopke,
2004; Lee et al., 2003; Maykut et al., 2003). Relectance from and transmittance through the sample
of a 633 nm laser beam is monitored before, during, and after analysis to detect darkening of the
ilter deposit as some of the OC chars to EC in the inert He atmosphere.
The optical pyrolysis fraction (OP) is deined as carbon evolved between the time O 2 is added and
the relected or transmitted signal returns to its original value. This OP is added to the sum of the
OC fractions and subtracted from the sum of the EC fractions to account for OC that changed to EC
during analysis. Chow et al. (2004a) discovered that transmitted light is dominated by charring of
the adsorbed organic vapors within the ilter, while relected light is dominated by charring of the
surface deposit. EC by TOR is typically higher and less sensitive to the temperature program than
EC by TOT, so the default EC is that by TOR. By summing the different temperature fractions and
selecting OP by TOR or TOT, however, an EC value can be achieved from a single IMPROVE mea-
surement that approximates many of the EC values obtained by other methods (Watson et al., 2005).
During analysis, a 0.5 cm 2 circle is removed from the quartz-iber ilter with a precision punch
and placed in a sample boat that is inserted into the analysis oven and immersed in the appropriate
carrier gas. If carbonate carbon (CO 3 = ) is suspected to be in the sample (Cao et al., 2005a; Chow
and Watson, 2002; Li et al., 2008), it can be removed and quantiied by injecting 20 μL of 0.4 M HCl
through the septum port onto the sample punch. The CO 2 evolved is measured as CO 3 = and after
600 s, the carbon analysis method for OC and EC speciied earlier is initiated. The FID response is
calibrated by analyzing samples of known amounts of CH 4 , CO 2 , sucrose, and potassium hydrogen
phthalate (KHP).
7.4.9 t HerMal a nalysis For w ater -s oluble o rganic c oMPounds
Water-soluble organic compounds (WSOC) are related to aerosol radiative forcing in clouds
(Novakov and Corrigan, 1996; Novakov and Penner, 1993), secondary organic aerosol (SOA), and
hygroscopic aerosol (Saxena and Hildemann, 1996). WSOC can be measured on the water extract
with a total organic carbon (TOC) analyzer (e.g., Shimadzu TOC-VCSH with ASI-V autosampler,
Columbia, MD) or with the OC/EC carbon analyzer by replacing the quartz boat used for the ilter
punch with a platinum boat that can retain 20-100 μL of the water extract. The extract is evaporated
at 50°C, followed by ramping the temperature to 900°C in a 98% He/2% O 2 atmosphere. Peak inte-
gration and calibration standards are the same as they are for the OC/EC analyses.
For the TOC analyzer, ∼5 mL of the quartz-iber ilter sample extract is placed in 9 mL prebaked
glass tubes. Once injected, the sample is mixed with 1.5% of 2 M HCl and the mixture is sparged in
zero air for 1.5 min. Acidiication and agitation of the sample eliminate interference from inorganic
carbon (i.e., carbonates). The sample is then conveyed to the combustion chamber, where the extract
is catalytically oxidized at 680°C to CO 2 . Combustion products in zero-air carrier gas move into a
dehumidiier for moisture removal and cooling and then into a halogen scrubber before progressing
to the sample cell for nondispersive infrared (NDIR) detection. Moisture and chlorine gas cause
positive interference in CO 2 measurements. NDIR peaks are integrated for comparison with KHP
and glucose standards that span the range of expected concentrations.