Correlation of selected molecular properties and recovery values in volatile organic compounds analysis: comparison of two water matrices

Abstract

This study investigates if certain molecular properties can influence the recovery of 18 volatile organic compounds (VOCs) in water under the applied analytical conditions, a purge and trap gas chromatographic method with mass spectrometric detection (P&T GC MS). Statistical and quantitative structure property relationship (QSPR) analyses were applied to find correlations between molecular parameters and analytical recoveries in two different water matrices (clean water (CW) produced in the laboratory and natural groundwater (GW)) at two different concentration levels. At 1 μg L⁻¹, most compounds had higher recoveries in CW than in GW, whereas at 15 μg L⁻¹, the recoveries were higher in the GW matrix. Polarity number and hydrophilic factor were correlated with the recovery differences at both concentration levels in GW. Polarity was significant in the distinction of recovery differences for CW and GW matrices at low concentration, while air diffusivity had an acceptable correlation with recovery differences for both matrices at the higher concentration. Further correlation of the recoveries themselves with the molecular properties was made by multivariate linear regression (MLR) resulting in a QSPR model. This was only possible for GW at the low 1 μg L⁻¹ concentration. Partial least square analysis indicated that hydrophilic factor, polarity, and molecular weight were the most important properties investigated. No significant correlation was found in CW matrix or in the higher concentration level in GW matrix, which implies that the most significant properties might only be relevant for the VOCs recovery at low concentrations and only if the matrix contains other water constituents (DOC, salt).