Insights into critical sources of bias in quantitation of volatile organic compounds based on headspace extraction approach

Abstract

Headspace (HS)-based vaporization of volatile organic compounds (VOCs) in a glass vial is an effective option for standard preparation or sample quantitation. However, adsorptive losses can occur systematically during the preparation of HS VOC samples between a liquid working standard (L-WS) and the inner surface of the vial. Here, a mixture of 22 VOCs (aldehydes, ketones, esters, alcohols, aromatics, chlorinated aromatics, and fatty acids) was used to assess the adsorptive loss characteristics based on in-vial vaporization of an L-WS. Quantified losses between the L-WS and vapor were between 48.2% (propionaldehyde [PA]) and 99.8% (heptanoic acid [HPA]) as the lowest and highest boiling point of all targets, respectively. The logarithmic partition coefficient values exhibited a quadratic relationship with the VOC boiling points. The detection threshold limit (DTL; the inception of mass quantification) ranged from 369 ng (m-xylene) to 4,107 ng (HPA), which were far higher than their corresponding method detection limits. The relative recovery (RR) of each VOC was also estimated as its mass ratio between HS (after vaporization) and liquid sample injected into the glass vial (before vaporization). Accordingly, the RR results of VOCs varied in the range from 0.14% (HPA) to 44.3% (PA). Further, there were little variations in RR patterns between VOCs, when the effect of temperature on their extraction was examined across 25 to 80 °C. HS-based quantification of VOCs should, therefore, be carried out by considering the DTL and associated VOC losses to reduce substantial biases that can otherwise arise from these complicated variables.