Comparison of predictive methods for Henrys Law Coefficients of organic chemicals

Predictive methods proposed in the literature for estimating Henrys Law Coefficient (HLC) of organic chemicals are reviewed and compared. Five different models developed from the following methods are discussed: solubility-vapor pressure method; group/bond contribution method; linear solvation energy relationship method; and, molecular connectivity method. The models reviewed cover over 450 chemicals belonging to several different chemical families. Experimental variability and interlaboratory variances of measured HLC data are summarized to establish acceptance limits for the predictive models. The five predictive models are compared and evaluated on the basis of quality of predictions and predictive errors for a common set of 150 chemicals. (C) 1998 Elsevier Science Ltd. All rights reserved.