Get access

Methods for estimating the bioconcentration factor of ionizable organic chemicals

Authors

  • Wenjing Fu,

    1. Department of Environmental Engineering, Technical University of Denmark, Miljoevej, Building 113, DK-2800 Kgs. Lyngby, Denmark
    2. Department of Chemical and Biochemical Engineering, Technical University of Denmark, Building 229, DK-2800 Kgs. Lyngby, Denmark
    Search for more papers by this author
  • Antonio Franco,

    1. Department of Environmental Engineering, Technical University of Denmark, Miljoevej, Building 113, DK-2800 Kgs. Lyngby, Denmark
    Search for more papers by this author
  • Stefan Trapp

    Corresponding author
    1. Department of Environmental Engineering, Technical University of Denmark, Miljoevej, Building 113, DK-2800 Kgs. Lyngby, Denmark
    • Department of Environmental Engineering, Technical University of Denmark, Miljoevej, Building 113, DK-2800 Kgs. Lyngby, Denmark
    Search for more papers by this author

  • Published on the Web 2/26/2009

Abstract

The bioaccumulation potential is an important criterion in risk assessment of chemicals. Several regressions between bioconcentration factor (BCF) in fish and octanol-water partition coefficient (KOW) have been developed for neutral organic compounds, but very few approaches address the BCF of ionizable compounds. A database with BCFs of 73 acids and 65 bases was collected from the literature. The BCF estimation method recommended by the Technical Guidance Document (TGD) for chemical risk assessment in the European Union was tested for ionizing substances using log KOW (corrected for the neutral species, log[fn·KOW]) and log D (sum of log KOW of neutral and ionic molecule, apparent log KOW) as predictors. In addition, the method of Meylan et al. (Environ Toxicol Chem 1999; 18:664–672) for ionizable compounds and a dynamic cell model based on the Fick-Nernst-Planck equation were tested. Moreover, our own regressions for the BCF were established from log KOW and pKa. The bioaccumulation of lipophilic compounds depends mainly on their lipophilicity, and the best predictor is log D. Dissociation, the pH-dependent ion trap, and electrical attraction of cations impact the BCF. Several methods showed acceptable results. The TGD regressions gave good predictions when log(fn·KOW) or log D were used as a predictor instead of log KOW. The new regressions to log KOW and pKa performed similarly, with mean errors of approximately 0.4. The method of Meylan et al. did not perform as well. The cell model showed weak results for acids but was among the best methods for bases.

Get access to the full text of this article

Ancillary