Variability of PM measurement results due to subsampling of filters and water content analysis

This study investigates the effects of filter cutting on particulate matter (PM) analysis using gravimetric measurements, Karl Fischer titration (KF), and ion chromatography. Six Whatman™ Quartz fiber filters were exposed to PM at an urban background site in Zabrze, Poland, collecting PM1 and PM2.5 fractions. Filters were conditioned before and after exposure, and their masses were measured gravimetrically. Cutting was performed using a 3D-printed plate under controlled humidity and temperature to minimize material loss or addition. KF titration assessed water content in PM sub-samples, while ion chromatography analyzed water-soluble ions before and after titration to examine potential changes in ion composition due to the KF oven's heating process. Results show significant material loss when filters were cut into sectors during sub-sampling. For PM1 filters, the average material loss was 0.060 ± 0.073 mg, while PM2.5 filters experienced a loss of 0.192 ± 0.142 mg. Relative deposit mass losses for PM1 filters were 3.34%, 3.35%, and 3.54%, and for PM2.5 filters, 5.96%, 7.53%, and 9.65%, highlighting variations based on deposit amount and composition. The relative variability in water content measured in filter sectors ranged from 2.65% to 4.79%, whereas variability from cutting sectors ranged from 0.0011% to 0.0304%, indicating that water content contributes more significantly to variability in PM assessment. ANOVA revealed negligible effects of sub-sampling timing on material loss (p = 0.7908). This study highlights uncertainties in filter sub-sampling (cutting) during PM analysis and their implications for PM measurement accuracy. By combining analytical techniques, it underscores the need to refine sub-sampling methodologies. The findings contribute to improving the reliability of PM assessments, particularly regarding water content variability and deposit losses.