Prediction of gas collection efficiency and particle collection artifact for atmospheric semivolatile organic compounds in multicapillary denuders

A modeling approach is presented to predict the sorptive sampling collection efficiency of gaseous semivolatile organic compounds (SOCs) and the artifact caused by collection of particle-associated SOCs in multicapillary diffusion denuders containing polydimethylsiloxane (PDMS) stationary phase. Approaches are presented to estimate the equilibrium PDMS–gas partition coefficient (Kpdms) from a solvation parameter model for any compound, and, for nonpolar compounds, from the octanol–air partition coefficient (Koa) if measured Kpdms values are not available. These estimated Kpdms values are compared with Kpdms measured by gas chromatography. Breakthrough fraction was measured for SOCs collected from ambient air using high-flow (300 L min−1) and low-flow (13 L min−1) denuders under a range of sampling conditions (−10 to 25 °C; 11–100% relative humidity). Measured breakthrough fraction agreed with predictions based on frontal chromatography theory using Kpdms and equations of Golay, Lövkvist and Jönsson within measurement precision. Analytes included hexachlorobenzene, 144 polychlorinated biphenyl congeners, and polybrominated diphenyl ethers 47 and 99. Atmospheric particle transmission efficiency was measured for the high-flow denuder (0.037–6.3 μm diameter), and low-flow denuder (0.015–3.1 μm diameter). Particle transmission predicted using equations of Gormley and Kennedy, Pich, and a modified filter model, agreed within measurement precision (high-flow denuder) or were slightly greater than (low-flow denuder) measured particle transmission. As an example application of the model, breakthrough volume and particle collection artifact for the two denuder designs were predicted as a function of Koa for nonpolar SOCs. The modeling approach is a necessary tool for the design and use of denuders for sorptive sampling with PDMS stationary phase.