Page:NIOSH Manual of Analytical Methods - Chapter Q.pdf/12

 The mean carbon loading (Table 1) on the sample filters (QQ1) was about 4 :g/cm2 on four different days (sample sets 1-4). A diesel truck was operating briefly on one of these days (set 4). On two other days (sets 5 and 6), the truck was running for longer periods, as evidenced by the higher EC loadings (Figure 3). The corresponding TC loadings on these two days were about 8 :g/cm2 and 23 :g/cm2. Over the loading range, the amount of adsorbed OC had no dependence on the TC loading, and the OC results for the bottom quartz filters (1.35 to 3.44 :g/cm2 ) were higher than those for the passive field blanks (0.17 to 0.74 :g/cm2 ). Two of the results (sets 4 and 5) were slightly lower than the others. In one case (set 4), the sampling period (23 minutes) was shorter, but the adsorbed OC results for the bottom filters in both sampling configurations (i.e., QQ2 and TQQ2) were equivalent and comparable to the upper quartz filters (TQQ1), below the Teflon. This implied that partitioning [65] (between gas and adsorbed states) on the sample filters (QQ1) had reached equilibrium; thus, the lower result was not due to lack of equilibration. In large part, the results for passive blanks provided a likely explanation for the lower value. The mean blank (passive) for this set was 0.17 (± 0.10) :g/cm2, which was the lowest value obtained. Passive blank results for the other sets (excluding set 5) ranged from 0.25 (± 0.07) to 0.74 (± 0.14) :g/cm2. It is not known whether the same explanation applied to set 5 because passive blanks for this set were accidentally contaminated (results voided). As in sets 1–3 and 6, set 5 was collected for an 8-hour period, but it was collected during the spring, while the other sets were collected during mid-summer. Changed weather conditions may explain the lower value. On three days, the adsorbed OC fraction (calculated as the TQQ1 TC mean divided by QQ1 TC mean) constituted about 80% of the TC on the sample filters (QQ1). On three other days (sets 4, 5, and 6), the adsorbed OC fraction was 39%, 22%, and 16% of the total. A backup quartz filter provided a better correction for adsorbed carbon than a traditional blank would have. Traditional blanks (media and field) underestimated the adsorbed OC, causing overestimation of the true particulate OC (TC) concentration. The need for this correction, and the number of backup filters analyzed, depends on the sampling strategy and environment. The lower the OC loading, the greater the influence of adsorbed vapor on the particulate OC measurement. If the EC loading also is low, the same holds true for the TC result. Thus, when carbon loadings are low, correction for adsorbed OC is important for accurate measurement of particulate OC and TC concentrations. However, the correction addresses adsorbed vapor only—not interference of less volatile materials (e.g., components in cigarette and wood smokes, oils) collected primarily on the top filter. Depending on vapor pressure, sampling conditions (temperature, flow rate, sampling period), and filter loading, these materials also collect on the bottom filter to some extent. The bottom filter cannot correct for this interference. If high OC loadings are found on the bottom filter, less volatile OC interferences should be suspected.

g. EC Oxidation in Helium In the NMAM 5040 analysis, oxidation of original EC (as opposed to char) is sometimes seen during the last temperature step in helium, but this is generally not common. If early removal occurs, it is important to ensure that oxygen contamination is not responsible. This check can be performed by analyzing a sucrose standard solution applied to a clean, unused filter punch. Although char loss is sometimes observed with sucrose, the filter transmittance normally does not reach its initial value until after oxygen is introduced. If oxygen 3/15/03

240

NIOSH Manuual of Analytical Methods