Using the POD sampler for quantitative diffusive (passive) monitoring of volatile and very volatile organics in ambient air: Sampling rates and analytical performance

POD diffusive samplers loaded with Carbopack X and Carbograph 5TD were exposed to certified calibration mixtures containing a total of 110 different ozone precursor and air toxic compounds. Constant sampling rates were identified for 39 ozone precursors and 33 air toxics. As 9 of these compounds were included in both mixtures, this meant a total of 63 different volatile and very volatile compounds were sampled using the POD with overall expanded uncertainties below 30 % for the sampling rate associated with the whole range of sampling times from 2 to 24 h. Carbograph 5TD exhibited superior performance for diffusive sampling of oxygenated and halogenated compounds in the air toxics mixture, while Carbopack X showed higher sampling efficiencies for aliphatic and aromatic hydrocarbons, as well as halogenated compounds derived from benzene and C₂ carbon number hydrocarbons. A model has been developed and applied to estimate sampling rates, primarily for the more volatile and weakly adsorbed compounds, as a function of the collected amount of analyte and the exposure time. For an additional 9 ozone precursors on Carbopack X, and 11 air toxics on Carbograph 5TD, the expanded uncertainties of modelled sampling rates were reduced to below 30 % and have a significantly reduced uncertainty compared to those associated with an averaged sampling rate. The paper provides Freundlich's isotherm parameters for the estimated (modelled) sampling rates and defines a pragmatic approach to their application. It does so by identifying the best sampling time to use for the expected exposure concentrations and associated analyte masses. This allows for expansion of the sampling concentration range from hundreds ng m^-3 to mg m^-3, while avoiding saturation of the adsorbent. Finally, field measurement comparisons of POD samplers, pumped tube samplers and online gas chromatography (GC), for sampling periods of 3 and 7 days in a semi-rural background area, showed no significant differences between reported concentrations.