Quantitative Structure-property Relationship Of Environmental Media Partition Coefficients And Bioconcentration Factor For Perfluorinated And Polyfluoroalkyl Substances

Perfluorinated and polyfluoroalkyl substances(PFASs)are a typical class of emerging persistent organic pollutants(POPs),which are widely used in industrial fields due to their hydrophobicity,oil-repellency,thermal stability and chemical stability.PFASs have been used in large-scale production and have been found in a variety of different environmental media.PFASs can enter and accumulate in organisms through drinking water,food chains and air inhalation.When their concentration in the organism reaches a certain threshold,it will have a corresponding toxic effect on organisms,posing a significant threat to human health.The environmental stock of PFASs is high and they are biologically toxic.The research on their environmental behavior and bioconcentration characteristics has become a hot topic in the field of environmental science and biotoxicology.Organic carbon/water partition coefficient(Koc)and air/particulate partition coefficient(Kp)are important parameters that characterize the distribution behavior of organic compounds among different environmental media,and are crucial for their environmental fate and risk assessment.Bioconcentration factor(BCF)is an important parameter for evaluating the accumulation ability of organic compounds in organisms,and can be used to describe the accumulation effect of organic compounds in organisms.The accurate determination and rapid acquisition of Koc,Kp and BCF of PFASs are of great significance for the study of their environmental behavior and bioconcentration characteristics.Experiment is the basic method for determining Koc,Kp and BCF of PFASs.However,there are many limitations in relying solely on experimental methods to obtain their property parameters due to the wide variety of PFASs and the endless emergence of derivatives.Developing reliable prediction models to quickly and accurately obtain the Koc,Kp and BCF of PFASs is considered to be one of the important ways to enrich the basic property data of PFASs.However,most existing prediction models mainly have the following problems.First,the type and quantity of PFASs in the modeling data set are relatively small.The second is that the model modeling process is not complete enough,and there is a lack of validation of the external prediction ability of the model and a definition of the scope of application.The prediction ability for Koc,Kp and BCF of different types of PFASs is limited.Quantitative structure-property relationship(QSPR)is a class of chemometrics methods that study the quantitative mathematical relationship between the structure and properties of substances and thereby predict their properties.This method is based on computer algorithms,which can effectively predict the molecular properties or activities of compounds while reducing experimental costs and errors.It has certain advantages in filling the gaps in compound property data with fast and scientific methods.At the same time,the QSPR model can reveal the interaction mechanism between substances,providing a new idea for studying the mechanism of substance related reactions.Therefore,in view of the limitations of the existing Koc,Kp and BCF prediction models for PFASs,this paper intends to conduct QSPR studies on the Koc,Kp and BCF of PFASs,respectively.Based on previous research literature,this study comprehensively collected and collated experimental data related to PFASs(logKoc,logKp and logBCF).On the basis of following the QSPR model construction criteria,the QSPR model between the molecular structure parameters of PFASs and their logKoc,logKp and logBCF was established using multiple linear regression(MLR)algorithm and support vector regression(SVR)algorithm,respectively.Conduct comprehensive validation and evaluation of the model to ensure the integrity and standardization of the model modeling process,with a view to achieving reasonable prediction of logKoc,logKp and logBCF for different types of PFASs,and providing basic data for environmental risk assessment of PFASs.At the same time,the mechanism of the QSPR model established is explained,and the key structural variables that affect PFASs,such as logKoc,logKp and logBCF,are identified.The distribution mechanism and bioconcentration mechanism of PFASs between different environmental interfaces are analyzed at the molecular level,providing necessary theoretical guidance for the study of environmental behavior and bioconcentration characteristics of PFASs.The main research contents and conclusions of the paper are as follows:(1)Using the Koc of PFASs in the context of deionized water experiments and field experiments as the modeling dataset,using MLR and SVR algorithms,construct corresponding PFASs-logKoc linear and nonlinear QSPR models to widely predict the Koc value of PFASs.Based on the model construction criteria,a comprehensive evaluation,verification and application domain characterization of the built models were conducted one by one.Under different experimental backgrounds,the goodness of fit,robustness,and external prediction ability of the four models were slightly different,but they all met the model evaluation criteria(R2>0.6,Q2LOO>0.5 and Q2F1,Q2F2,Q2F3>0.5).Within the application range defined by the model,they were able to better predict the logKoc of different types of PFASs.The optimal MLR-QSPR model structure of the model indicates that in the context of deionized water environment,the molar volume(Vm),Mulliken electronegativity(XM),dipole moment(μ)and the molecular van der Waals potential(Vvdw)are key structural variables that affect the logKoc of PFASs.Under field experimental background conditions,molecular weight(MW),dipole moment(μ),the lowest unoccupied molecular orbital energy(ELUMO)and the maximum average local ionization energy of the molecular surface(ALIEmax)have significant effects on the logKoc of PFASs.Studies have shown that the distribution of PFASs between soil/sediment and water phases is generally affected by hydrophobic,electrostatic,and dipole dipole interactions.In the context of deionized water environments,van der Waals interactions also have a certain impact on the distribution of PFASs between the two phases.(2)The QSPR model of PFASs-logKp was established using MLR and SVR algorithms to study the distribution mechanism of PFASs between the air and particulate phases.After system evaluation,validation and application domain analysis,the MLR-QSPR model has good goodness of fit and robustness on the basis of satisfying the application conditions of the model,while the SVR-QSPR model has superior external prediction ability.The mechanism analysis shows that the molecular descriptor that has an important impact on the logKp change of PFASs is electronic hardness(η)and molecular volume(V),where electronic hardness(η)has a leading role.The distribution of PFASs between the atmospheric gas phase and the atmospheric particle phase is dominated by electrostatic and dispersion interactions.(3)Based on the logBCF data in fish liver,QSPR research on the bioconcentration mechanism of PFASs was conducted.The PFASs-logBCF model constructed using MLR and SVR algorithms has good goodness of fit,robustness,and external prediction capabilities.In the application domain,it can achieve extensive prediction of logBCF of various PFASs in fish liver.On the other hand,model mechanism analysis revealed that the bioconcentration of PFASs in fish liver is related to the molecular volume(V)of the substance and the minimum value of molecular surface potential(Vs,min),and the effect of molecular volume(V)is more significant.The bioconcentration effect of PFASs in fish liver is the result of the interaction mechanism of hydrophobic and hydrogen bonding.In summary,this paper establishes QSPR model by quantitatively analyzing between the molecular structure of PFASs and their environmental media partition coefficients(logKoc and logKp)and their biological enrichment factors(logBCF),respectively,to predict their corresponding parameter values,enriching the theoretical research methods for the physical and chemical properties of PFASs.Through model mechanism research,obtain key molecular structures and property descriptors that affect the environmental behavior and ecological risks of PFASs,explore the relationship and mechanism between the structure of PFASs and their soil-water/gas-solid partition coefficients,and provide new ideas for quickly understanding the migration and transformation behavior of PFASs in different environmental media and ecotoxicological research using efficient scientific methods.