Characterization of sorption properties of high-density polyethylene using the poly-parameter linearfree-energy relationships

High-density polyethylene (HDPE) is a known sorbent for non-ionic organic compounds in technical applications. Nevertheless, there is little information available describing sorption to industrial HDPE for a broad range of compounds. With a better understanding of the sorption properties of synthetic polymers, environmental risk assessment would achieve a higher degree of accuracy, especially for microplastic interactions with organic substances. Therefore, a robust methodology for the determination of sorbent properties for non-ionic organic compounds by HDPE is relevant for the understanding of molecular interactions for both technical use and environmental risk assessment. In this work, sorption properties of HDPE material used for water pipes were characterized using a poly-parameter linear free-energy relationship (ppLFER) approach. Sorption batch experiments with selected probe sorbates were carried out in a three-phase system (air/HDPE/water) covering an aqueous concentration range of at least three orders of magnitude. Sorption in the concentration range below 10-2 of the aqueous solubility was found to be non-linear and the Freundlich model was used to account for this non-linearity. Multiple regression analysis (MRA) using the determined distribution coefficients and literature-tabulated sorbate descriptors was performed to obtain the ppLFER phase descriptors for HDPE. Sorption properties of HDPE were then derived from the ppLFER model and statistical analysis of its robustness was conducted. The derived ppLFER model described sorption more accurately than commonly used single-parameter predictions, based i.e., on log K-o/w . The ppLFER predicted distribution data with an error 0.5 log units smaller than the spLFERs. The ppLFER was used for a priori prediction of sorption by the characterized sorbent material. The prediction was then compared to experimental data from literature and this work and demonstrated the strength of the ppLFER, based on the training set over several orders of magnitude. (C) 2019 Elsevier Ltd. All rights reserved.