Affiliation:
1. Department of Product Safety BASF SE, GBP/RA 67056 Ludwigshafen am Rhein Germany
2. Department of Ecotoxicology Crop Protection BASF SE 67117 Limburgerhof Germany
3. Department of Experimental Toxicology and Ecotoxicology BASF SE, RG/T 67056 Ludwigshafen am Rhein Germany
4. Department of Regulatory Toxicology & Ecotoxicology BASF Services Europe GmbH 10245 Berlin Germany
5. Department of Global Regulatory Solutions BASF SE, EV/SR 67056 Ludwigshafen am Rhein Germany
Abstract
AbstractPhenolic benzotriazoles (BTZs) are used globally as light stabilizers in various plastic products to protect them from photooxidative degradation. The same physical–chemical properties that confer their functionality, like a sufficient photostability and a high octanol–water partition coefficient, also raise concerns on their potential for environmental persistence and bioaccumulation based on in silico predictive tools. To evaluate their bioaccumulation potential in aquatic organisms, standardized fish bioaccumulation studies according to OECD TG 305 were conducted with four of the most commonly used BTZs: UV 234, UV 329, UV P, and UV 326. The resulting growth‐ and lipid‐corrected BCF values revealed that UV 234, UV 329, and UV P were below the bioaccumulation threshold (BCF ≤ 2000), but UV 326 is considered very bioaccumulative (BCF ≥ 5000) with respect to the bioaccumulation criteria under REACH. Comparing these experimentally derived data with quantitative structure activity related or other calculated values using a logarithmic partitioning coefficient octanol–water (log Pow) driven mathematical formula revealed significant discrepancies demonstrating the weakness of current in silico approaches for this group of substances. Furthermore, available environmental monitoring data demonstrate that these rudimentary in silico approaches can lead to unreliable bioaccumulation estimates for this chemical class due to considerable uncertainties in underlying assumptions (e.g., concentration and route of exposure). However, using more sophisticated in silico methods (i.e., CATALOGIC base‐line model), the derived BCF values were better aligned with the experimentally derived ones.