环境水体中化学品光化学持久性的预测模型

Prediction models for photochemical persistence of chemicals in aquatic environments

化学品是水环境中新污染物的重要来源,光化学转化是其在表层水体中的重要去除途径,光化学持久性是评价其环境暴露行为特性的重要指标.化学品的环境光化学转化受其分子结构、水体特性等多重因素的影响,仅通过实验测定,难以确定其在自然水体中光转化动力学参数值,需要构建预测模型.本文总结了影响化学品环境光转化动力学的因素和机理,介绍了化学品光吸收、光化学转化、水中光生活性中间体的生成以及太阳光在大气/水体中的衰减等过程中相关参数的预测模型,评述了化学品光化学持久性参数预测模型的前沿问题和研究需求,展望了环境水体中化学品光化学持久性研究的重点发展方向.

Chemicals can be released into aquatic environments during their lifecycles,becoming emerging pollutants that cause adverse effects on health of humans and ecosystems.Photochemical transformation is important in determining environmental persistence and exposure concentrations of chemicals.It is of importance to assess photochemical persistence of chemicals in aquatic environments for sound management of chemicals and emerging pollutants.Quantum yields of direct photolysis,phototransformation rate constants,and phototransformation half-lives are key parameters characterizing environmental photochemical persistence of chemicals.Although the parameters on photochemical persistence can be determined experimentally,the experimental determination is of low throughput,expensive,time-consuming,and restricted by availability of authentic chemical standards,due to diversity of aquatic environmental factors that have impacts on the photochemical processes.It is necessary to develop prediction models on aquatic photochemical persistence of chemicals,including quantitative structure-activity relationship(QSAR)models,for high-throughput prediction of the aquatic photochemical persistence parameters of the vast number of chemicals.In this review,environmental phototransformation pathways and influencing factors on the phototransformation of chemicals were summarized.Research frontiers on prediction models of photochemical persistence were discussed.For direct phototransformation,light absorbance parameters of chemicals can be predicted by machine learning models,and the method is more efficient than experimental determination or quantum chemical calculation.Application of quantum chemical descriptors has ensured QSAR models to predict the photochemical behavior parameters of chemicals.Although several QSAR models have been developed on direct phototransformation quantum yields,rate constants or halflives,insufficiency in the experimental data is still a limitation of the modeling.Direct phototransformation rates of chemicals depend on both intrinsic factors that are governed by molecular structures,and environmental factors such as solar spectrum and optical path length.Therefore,incorporating environmental factors with molecular structures to construct multimodal models may improve prediction performance on direct phototransformation kinetics of chemicals.For indirect phototransformation,machine learning and satellite remote sensing can be potentially employed for prediction of quantum yields and steady-state concentrations of photochemically produced reactive intermediates(PPRIs).Molecular composition,electron donor capacity,and optical properties of dissolved organic matter(DOM)could be employed as features in constructing the prediction models.Second-order reaction rate constants of chemicals with PPRIs can be predicted with QSAR models.More efforts are needed to develop prediction models on reactivity of chemicals with excited triplet states of DOM.By incorporating environmental factors such as underwater irradiance intensities,integrated models that consider both direct and indirect phototransformation of chemicals have been developed.Further research efforts are needed to construct integrated models that suit specifically the main waterbodies in China,and to incorporate more photochemical processes and environmental factors into the modeling.Overall,to improve prediction on aquatic photochemical persistence of chemicals or emerging pollutants;phototransformation pathways of chemicals in natural water bodies should be further clarified;molecular modeling such as density functional theory calculation methods should be employed for predicting some parameters that are relevant with photochemical persistence of chemicals;data sets related with phototransformation should be collected and curated,and employed to construct machine learning models for the prediction;and more efforts should be paid to prediction models on photochemical persistence of chemicals in estuarine waters,coastal seawaters and oceanic waters.