with a Sunset Laboratory instrument because, depending on the sample type, differences in instrument design can affect the results to varying degrees. Adjustment of the temperature program, according to another method’s specifications, does not necessarily produce the same results as that method. For example, lower EC results [35] were obtained with a Sunset Laboratory instrument when a sucrose standard was analyzed according to the temperature program specified for a method used in Europe. No pyrolysis correction was made, and the total analysis time was shorter, yet the EC result was much lower (about 3 µg with the Sunset instrument and 11 µg with a different instrument). When the NMAM 5040 thermal program was used on a DRI instrument, the filter transmittance reportedly exceeded its initial value before the addition of oxygen [73]. In a recent audit by CARB (California Air Resources Board), the same problem was seen with a DRI instrument, but not a Sunset Laboratory instrument (report [2003 Performance Evaluation Evaluation Sample Audit of the CARB Lab] available at http://www.epa.gov/ttnamti1/pmspec.html). Char loss was extensive when the NMAM 5040 program was used on a DRI instrument. Again, the filter transmittance exceeded its initial value in helium, but the split point was not assigned until after oxygen was introduced. These results contrast with those obtained by our laboratory and others, and such behavior was not seen in another comparison [74]; however, that comparison (and the CARB audit) was a direct comparison of the two methods (samples were analyzed with Sunset and DRI instruments). Only 2 of 52 samples analyzed by Sunset Laboratory showed an increase in transmittance before the addition of oxygen, and the OC-EC split was near the point where oxygen was added [74]. Both samples had low carbon loadings (OC = 4.0 µg/cm2 , EC = 0.6 µg/cm2 , OC = 3.3 µg/cm2 , EC = 0.5 µg/cm2 ) and were thought to be wood-smoke dominated. As pyrolysis was evident, the increase in filter transmittance was attributed to char removal. Comparable EC results would have been obtained if the char had not been removed until after oxygen was added, because char is assigned to the OC fraction. In the NMAM 5040 analysis of organic compounds (e.g., sucrose), partial char loss in helium sometimes occurs. The varying degree of loss may relate to differences in filter purity. 4.
OCCUPATIONAL EXPOSURE CRITERIA (U.S.) In 1995, the American Conference of Governmental Industrial Hygienists (ACGIH) proposed a Threshold Limit Value (TLV® ) for diesel exhaust (see Notice of Intended Changes for 1995–1996). A TLV of 150 µg of submicrometer particulate matter per cubic meter of air was proposed. Four years later, a value of 50 :g/m3 was proposed [80]. Because EC is a demonstrated exposure marker for diesel particulate exhaust and can be accurately quantified at low levels [81], the standard was recast in terms of an equivalent elemental carbon measurement in 2001 [81]. A TLV-TWA (time-weighted average) of 20 :g EC per cubic meter of air was recommended [82]. By comparison, the proposed TLV is high relative to an environmental standard supported by the California Office of Environmental Health Hazard Assessment (OEHHA). The California OEHHA has classified particulate emissions from diesel-fueled engines as a toxic air contaminant (TAC) and supports 5 :g/m3 as the chronic inhalation reference exposure level (REL) [5]. Because EC is a fraction of the diesel particulate emissions, the difference between the two air standards is greater than four fold, and the magnitude of the difference is dependent on the EC mass fraction of the diesel particulate. For example, if EC constituted 40% of the diesel particulate mass, the proposed
3/15/03
247
NIOSH Manuual of Analytical Methods
