In Silico Simulation And Prediction For Adsorption Of Organic Pollutants Onto Carbon Nanomaterials

Investigating adsorption behavior of organic pollutants onto carbon nanomaterials is essential for understanding their environmental fate and evaluating their potential ecological risks.Considering that structures of carbon nanomaterials and organic pollutants in the environment are diverse,and the effects for environmental factors on the adsorption are intricate,experimental determination for exploring the adsorption onto carbon nanomaterials cannot meet the need for risk assessment and management.Therefore,it is urgent to develop in silico simulation and prediction methods for probing the adsorption behavior of organic pollutants onto carbon nanomaterials.In this study,we explored effects for the diameter and chirality of single-walled carbon nanotubes(SWNTs),the functional groups of organic pollutants and graphene oxides on the adsorption via density functional theory(DFT)and molecular dynamics(MD)methods.Differences between the adsorption mechanisms in gaseous phase and that in aqueous phase were revealed.Moreover,we developed quantitative structure activity relationship(QSAR)models for predicting the adsorption of organic pollutants on SWNTs,graphene and graphene oxides with the experimental and computed adsorption data,and further estimated the relative contribution of different adsorption mechanisms to the overall adsorption.Main contents and conclusions for the current research are as follows:(1)Based on experimental adsorption equilibrium coefficient(K)values on SWNTs and theoretical molecular structural descriptors values for 61 organic pollutants,we developed a linear model and a non-linear model for predicting logK values with multiple linear regression(MLR)and support vector machine(SVM)algorithms.The applicability domains for these two models cover organic pollutants including various functional groups,i.e.,>C=C<,-C?C-,-C6H5,>C=O,-COOH,-C(O)O-,-OH,-O-,-F,-Cl,-Br,-NH2,-NH-,>N-,>N-N<,-NO2,>N-C(O)-NH2,>N-C(O)-NH-,-S-and-S(O)(O)-.The SVM model performs slightly better than the MLR one.The MLR model indicated that the adsorption of organic pollutants onto SWNTs is mainly determined by van der Waals forces and hydrophobic interactions.(2)By utilizing DFT method,we simulated adsorption for 38 organic compounds(aliphatic hydrocarbons,benzene and its derivatives,and polycyclic aromatic hydrocarbons)onto graphene in both gaseous and aqueous phases,and developed two poly-parameter linear free energy relationship(pp-LFER)models for estimating the adsorption energies of organic compounds on graphene.The relative contributions of different interactions to the overall adsorption were estimated by using the values for different terms in pp-LFER models.Dispersion and electrostatic interactions are main driving forces during the gaseous adsorption process,while the aqueous adsorption is mainly governed by dispersion and hydrophobic interactions.It was also found that the adsorption energies for organic compounds onto SWNTs become stronger with the increase of the nanotube diameter,and graphene has stronger adsorption energies than SWNTs.(3)By means of MD method,we simulated and predicted logK values for 43 aromatic compounds on graphene and different models of graphene oxides with various functional groups(hydroxyl,epoxy and carbonyl).The results showed that graphene has a stronger adsorption capacity than graphene oxides.The hydroxyl and carbonyl groups for graphene oxides were found to form hydrogen bonds with the aromatic adsorbates,while epoxy groups did not.Furthermore,two theoretical linear solvation energy relationship models were established for predicting the logK values of organic compounds,namely,benzene,phenyl ethanol,phenol,aniline,nitrobenzene,nitrile,halogenated benzene,ketone,ester,biphenyl and their derivatives,polycyclic aromatic hydrocarbons and polybrominated diphenyl ethers onto graphene and graphene oxides.These two TLSERs models indicated that the logK values onto graphene were significantly influenced by H-donating ability(e?),dispersion and hydrophobic interactions(V),while for the logK values onto graphene oxide,e? was the most influencial factor.