The objective of this study was to sample particulate matter <2.5 m (PM2.5) with distinctive source influences to optimize thermal-optical analysis for black carbon measurement. High-volume samples of PM2.5 were collected at different PM mass levels and at 3 urban sites to represent distinct source influences. These included samples dominated or heavily influenced by secondary organic aerosol (SOA) (Atlanta, GA in summer), vehicle emissions (Los Angeles, CA near a freeway), and woodburning emissions (Seattle, WA in winter). A randomized blocked design consisting of replicate suites involving two samplers was used to minimize effects from compositional variation during sampling. Statistically distinct levels of TC were acquired, and variance F-tests indicated that any temporal shift in the sample composition did not affect replicate samples. Principal components of the normalized TOT laser time series revealed an OC pyrolysis pattern that correlated highly with laser signals of source samples (lab-generated SOA, diesel exhaust reference material, and wildfire emissions) that represented the intended source influences. In contrast, correlations of the TOT flame ionization detector (FID) time series between the PM2.5 samples and sources were in all cases highest with SOA. Cross-correlations showed that the FID time series from the Los Angeles samples did not reveal a distinctive pattern for diesel soot. In general, the TOT laser series provided a superior indicator of source influence on PM2.5 samples than the FID time series.
Citation: Journal of Geophysical Research
Pub Type: Journals
central composite designs, DOE, particulate matter, PM2.5, response surface modeling, thermal-optical transmission