Modeling lifetime and degradability of organic compounds in air, soil, and water systems (IUPAC Technical Report)

Abstract

Republication or reproduction of this report or its storage and/or dissemination by electronic means is permitted without the need for formal IUPAC permission on condition that an acknowledgmentAbstract: Degradability of organic compounds in air, soil, and water is the most important factor for evaluating their environment fate as well as possible adverse effects to humans and the environment. The primary degradation process in the troposphere is the reaction with the hydroxyl radical. For water and soil compartments, the primary degradation process is biodegradation. The objectives of this report are: (i) to review published models and their evaluation studies, (ii) to perform an in-house evaluation of general models for estimating tropospheric degradation and biodegradation of organic compounds, and (iii) to recommend reliable procedures for estimating degradability of organic compounds in the environment. The extensive evaluation procedure has shown that the most accurate method for estimating tropospheric degradation is Atkinson's group contribution method. Although this method has some limitations, it seems to be a method of choice. A viable alternative to Atkinson's method is a direct calculation, performed today almost routinely, of the reaction rate constants with hydroxyl radicals. Recently, a methodology based on reliable semiempirical potential energy surfaces was developed that enables the calculation of reaction rate constants within a factor of 2 of their measured values. A partial least squares (PLS) model and a set of seven biodegradation rules have been found to be the most reliable in estimating complete biodegradation of organic compounds. However, it is recommended to use all four evaluated methods to estimate biodegradation in the environment. If their results agree, such estimates are very reliable.

Evaluation of Artificial Intelligence Based Models for Chemical Biodegradability Prediction

“…A number of detailed reviews of modeling efforts are available [4,5,6,23]. This work does not intend to repeat that work, but rather present a brief discussion of general modeling efforts to-date with a more detailed discussion and evaluation of an inductive machine learning method utilized by Gamberger et al [7].…”

“…The general concept of utilizing multiple models and concluding that reliable results cannot be obtained given conflicting results has been suggested in previous studies [6,15]. However this concept has more recently been rigorously evaluated.…”

“…These Rules use the structural descriptors noted in Table 3, but have different outcomes regarding the significance of those descriptors in biodegradability based on the nature of the data on which they were built. The MITI-I data has been reported to have a tendency to under-predict biodegradability and therefore classifies some compounds as non-degradable that are classified as degradable under other test conditions, such as those conditions that the chemicals in the BIODEG database were tested under [6]. This under-prediction has been reported in part to be potentially caused by the relatively high chemical concentration used in the MITI-I test which is higher than what is likely to be experienced in the environment, and may produce toxic effects on the test inoculum [4].…”

“…Each of these techniques has individual advantages and disadvantages and tradeoffs are managed such that all models have various limitations in their utility and predictive ability [4][5][6]. For example, individual models tend to have some level of chemical class specificity, either by design, or as an artifact of the breadth of the model training data set.…”

Tropospheric Degradation of Perfluorinated Aromatics: A Case of Hexafluorobenzene

“…EACTION with the hydroxyl (OH) radical is the primary chemical loss channel for organic compounds emitted into the troposphere [1][2][3][4][5] and a large number of experimental [1,5] and theoretical studies [4,6] has been published in this research area. The large fraction of those studies have been carried out for almost all classes of aliphatic organic compounds, i.e.…”