Validation of a semi-continuous instrument for elemental carbon and organic carbon using a thermal-optical method

A commercial semi-continuous elemental and organic carbon (ECOC) analyzer has been developed by Sunset Laboratory Inc. for the analysis of atmospheric particulate matter. The field deployable instrument is an alternative to off-line laboratory-based analysis of samples collected in the field. Although the Sunset Laboratory analyzer provides time-resolved particulate ECOC measurements using a methodology comparable to the laboratory-based NIOSH 5040 ECOC measurement, the performance of the semi-continuous ECOC analyzer has not been fully validated in the past. As part of the St. Louis-Midwest Supersite, two semi-continuous Sunset Labs ECOC field analyzers were operated to measure fine particle organic and elemental carbon for the entire year of 2002. The dual instrument operation provided a means to have a true continuous ECOC measurement, which would not be possible with only one instrument since the instrument cannot continue sample collection during the sample analysis period. In parallel with the semi-continuous ECOC analyzers, daily 24-h integrated PM2.5 samples were collected for ECOC analysis in the laboratory. Both the 24-h integrated filter samplers and the semi-continuous ECOC instruments were operated downstream of an organics denuder to remove positive adsorption artifacts associated with semi-volatile organic compounds. Comparison of 24-h averaged semi-continuous fine particle total carbon (TC), organic carbon (OC) and elemental carbon (EC) showed good agreement with the daily integrated off-line measurements. The linear regression of the averaged semi-continuous TC on the integrated off-line TC had a coefficient of regression (R2) of 0.89 and a regression slope of 0.97±0.02 when at least 20 h of the day were measured. The OC measurements had a R2 equal to 0.90 and a regression slope of 0.93±0.02. The EC measurements had a poorer R2 due to the very low levels of EC at the sampling site but a regression slope of 0.95±0.05. The accuracy of using a single semi-continuous ECOC instrument, configured to measure ECOC every other hour to represent the daily 24-h average, was also evaluated.