Aromatic pollutant n-caprylic alcohol-air partition coefficient prediction method based on solvation free energy and dimer

Abstract

The invention discloses an aromatic pollutant n-caprylic alcohol-air partition coefficient prediction method based on solvation free energy and dimer, and the method comprises the steps: collecting a data set, respectively calculating the single-point energy of a target aromatic pollutant single molecule and dimer in a gas phase and an n-caprylic alcohol phase, and calculating the n-caprylic alcohol-air partition coefficient of the target aromatic pollutant. Corresponding solvation free energy is obtained according to a single-point energy difference value of the single molecule and the dimer between a gas phase and an n-caprylic alcohol phase, a single molecule KOA predicted value and a dimer KOA predicted value are obtained according to a thermodynamic relationship and the solvation free energy, and the proportion of the target aromatic pollutant existing in the form of the dimer is predicted according to a QSAR (Quantitative Synthetic Aperture Radar) model; calculating the existing proportion of the single molecules according to the existing proportion of the dimers, and obtaining the corrected KOA value of the target aromatic pollutant according to the dimer KOA predicted value, the single molecule KOA predicted value, the existing proportion of the dimers and the existing proportion of the single molecules. According to the method, the KOA value of any aromatic pollutant can be quickly and accurately predicted only according to the molecular structure.

Description

Technical Field

[0001] The invention relates to the field of ecological risk assessment of chemicals, and in particular to a method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers. Background Art

[0002] Octanol-air partition coefficient (K OA ) refers to the ratio of the concentration of an organic compound in the n-octanol phase to the concentration of an organic compound in the air phase when the distribution is in equilibrium (dimensionless). Due to its large value, it is often expressed in logarithmic form (log K OA ) is used to characterize the distribution behavior of pollutants between the gas phase and the organic phases of the environment such as soil, plants, and aerosol particles. It is of great significance for evaluating the distribution, migration, transformation and ecological effects of pollutants. In addition, K OA It has a certain correlation with other physical and chemical properties and environmental behavior parameters. For example, the soil-air partition coefficient (K) of polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs) and chlorobenzenes (CBs) at 21°C SA ) and logK OA There is a good linear relationship. The supercooled liquid vapor pressure (P L ) and K OA There is also a good linear relationship between them. In addition, the amplification effect of organic pollutants between different trophic levels in the food chain is related to K OA There is also a significant relationship, which shows that K OA It can be used to evaluate the bioaccumulation of pollutants. OA The value is also crucial for understanding physical and chemical parameters such as soil-air partition coefficient, supercooled liquid vapor pressure and bioaccumulation. OA The value is of great significance for the environmental regulation and risk assessment of chemicals.

[0003] Common K OA Experimental determination methods include generation column method, fugacity measurement method, solid phase microextraction method and gas chromatography retention time method. However, aromatic pollutants are of various types and huge in number, and it is difficult to determine their K using experimental methods. OA The experimental determination method is time-consuming and laborious, requires specific instruments and equipment such as chromatography, and the standard is difficult to obtain, resulting in high cost, high technical difficulty and delayed risk assessment. OA There are only thousands of them, which is far from meeting the needs of chemical risk assessment. Summary of the invention

[0004] The present invention provides a method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers, so as to overcome the above technical problems.

[0005] The prediction method of n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimer includes:

[0006] S1, n-octanol-air partition coefficient K for collecting aromatic pollutants OA Experimental values ​​and form a data set;

[0007] S2. Using the SMD model at the HF / MiDi! 6D calculation level, the single point energy of the target aromatic pollutant single molecule in the gas phase and the n-octanol phase was calculated respectively. The solvation free energy of the single molecule was obtained based on the difference between the single point energy of the gas phase and the n-octanol phase. The single molecule K was calculated based on the thermodynamic relationship and the solvation free energy of the single molecule. OA Predicted value;

[0008] S3. Using the SMD model at the HF / MiDi! 6D calculation level, the single point energy of the target aromatic pollutant dimer in the gas phase and the n-octanol phase was calculated respectively. The solvation free energy of the dimer was obtained based on the difference in the single point energy between the gas phase and the n-octanol phase. The dimer K was calculated based on the thermodynamic relationship and the solvation free energy of the dimer. OA Predicted value;

[0009] S4. Constructing a QSAR model, wherein the QSAR model is used to OA Predicted value, dimer K OA The predicted value predicts the proportion of aromatic pollutants existing as dimers, randomly selects data from the data set to form a training set, and trains the QSAR model based on the training set;

[0010] S5. Calculate the proportion of dimers in the target aromatic pollutants based on the trained QSAR model, calculate the proportion of single molecules based on the proportion of dimers in the target aromatic pollutants, and calculate the proportion of dimers based on the dimer K OA Predicted value, single molecule K OA The predicted value, the proportion of dimers and the proportion of single molecules were used to obtain the corrected K value of aromatic pollutants. OA value.

[0011] Preferably, S2 includes:

[0012] S21. Use the chemical drawing software CS ChemDraw Ultra to draw the molecular structure of the target aromatic pollutant;

[0013] S22. Using the Gaussian09 platform, the SMD model was used to perform geometry optimization, frequency calculation, and single-point energy calculation on the single molecule of the target aromatic pollutant in the gas phase and the n-octanol phase at the HF / MiDi! 6D calculation level, and the solvation free energy ΔG between the gas phase and the n-octanol phase was obtained. OA ;

[0014] S23. Calculate single molecule K according to formula (1) OA Predicted value,

[0015]

[0016] In the formula, ΔG OA (single molecule) is the solvation free energy of a single molecule between the gas phase and the n-octanol phase, R is the ideal gas constant, T is the absolute temperature, K OA (single molecule) is single molecule K OA Predicted value.

[0017] Preferably, S3 includes:

[0018] S31. Use the chemical drawing software CS ChemDraw Ultra to draw the molecular structure of the target aromatic pollutant;

[0019] S32. Using the Gaussian09 platform, the SMD model was used to perform geometry optimization, frequency calculation, and single-point energy calculation of the target aromatic pollutant dimer in the gas phase and the n-octanol phase at the HF / MiDi! 6D calculation level, and the solvation free energy ΔG between the dimer in the gas phase and the n-octanol phase was obtained. OA ;

[0020] S33, calculate the dimer K according to formula (2) OA Predicted value,

[0021]

[0022] In the formula, ΔG OA (dimer) is the solvation free energy of the dimer between the gas phase and the n-octanol phase, R is the ideal gas constant, T is the absolute temperature, K OA (Dimer) is dimer K OA Predicted value.

[0023] Preferably, S4 comprises randomly selecting data from the data set to form a training set, and selecting five molecular structure descriptors, namely dihedral angle (DA), relative molecular mass (Mw), nonpolar surface area (NSA), polar surface area (PSA) and molecular polarity index (MPI), and establishing a QSAR model using partial least squares regression in Simca-S software, and the expression is:

[0024] X = -1.927 + 7.809 × 10 -3 NSA+1.510×10 -1 MPI - 2.218 × 10 -3 PSA+

[0025] 1.721×10 -3 DA-1.118×10 -3 Mw(3)

[0026] X+Y=100% (4)

[0027] Wherein, X=dimer%, dimer% indicates the proportion of aromatic pollutants existing in the form of dimers, the training set is used to train the QSAR model and obtain the trained QSAR model, Y=single molecule%, single molecule% indicates the proportion of target aromatic pollutants existing in the form of single molecules, and the trained QSAR model and dimer% can be used to obtain the single molecule%.

[0028] Preferably, the S5 includes:

[0029] K OA (single molecule), K OA (dimer), the values ​​of X and Y in the target aromatic pollutant calculated according to the trained QSAR model are substituted into formula (5) to obtain the predicted value of the target aromatic pollutant after correction,

[0030] logK OA (Correction value) = Y*logK OA (single molecule)+X*logK OA (Dimer) (5)

[0031] The positive value indicates the K value of the target aromatic pollutant after correction. OA Predicted value, X=dimer%, dimer% indicates the proportion of dimers in the target aromatic pollutant, Y=monomolecule%, monomolecule% indicates the proportion of the target aromatic pollutant in the form of monomolecule.

[0032] The present invention provides a method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers, and uses a solvation model density (SMD) model to calculate the solvation free energy of aromatic pollutants, and converts the solvation free energy into K through thermodynamic relations. OA The QSAR model was used to establish a dimer prediction model for K OA The K values ​​of aromatic pollutants can be accurately and quickly predicted. OAThe prediction method constructed by the present invention can quickly and accurately predict the K value of any aromatic pollutant based only on the molecular structure. OA The value effectively reduces the experimental measurement method and improves the prediction efficiency, providing a basis for a comprehensive understanding of the migration, transformation and bioaccumulation processes of pollutants in different environmental media, as well as their potential risks to ecosystems and human health, and providing a useful reference for the control of aromatic pollutants. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0034] Figure 1 is a flow chart of the method of the present invention;

[0035] Figure 2(a) is the log K of aromatic pollutants in the embodiment of the present invention. OA Experimental value and log K OA Regression model for predicting values;

[0036] Figure 2(b) is the log K of aromatic pollutants in the embodiment of the present invention. OA Experimental value and logK OA Regression model for corrected values;

[0037] Figure 3 are the experimental and predicted values ​​of dimer ratios of different aromatic pollutants in the embodiments of the present invention;

[0038] Figure 4 is the logK of the four types of aromatic pollutants in the embodiment of the present invention OA Original predicted values ​​and corrected values. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Figure 1 It is a flow chart of the method of the present invention, such as Figure 1 As shown, the method of this embodiment includes:

[0041] S1, n-octanol-air partition coefficient K for collecting aromatic pollutants OA Experimental values ​​and form a data set;

[0042] S2. Using the SMD model at the HF / MiDi! 6D calculation level, the single point energy of the target aromatic pollutant single molecule in the gas phase and the n-octanol phase was calculated respectively. The solvation free energy of the single molecule was obtained based on the difference between the single point energy of the gas phase and the n-octanol phase. The single molecule K was calculated based on the thermodynamic relationship and the solvation free energy. OA Predicted value;

[0043] S3. Using the SMD model at the HF / MiDi! 6D calculation level, the single point energy of the target aromatic pollutant dimer in the gas phase and the n-octanol phase was calculated respectively. The solvation free energy of the dimer was obtained based on the difference in the single point energy between the gas phase and the n-octanol phase. The dimer K was calculated based on the thermodynamic relationship and the solvation free energy. OA Predicted value;

[0044] S4. Constructing a QSAR model, wherein the QSAR model is used to OA Predicted value, dimer K OA The predicted value predicts the proportion of aromatic pollutants existing as dimers, randomly selects data from the data set to form a training set, and trains the QSAR model based on the training set;

[0045] S5. Calculate the proportion of dimers in the target aromatic pollutants based on the trained QSAR model, calculate the proportion of single molecules based on the proportion of dimers in the target aromatic pollutants, and calculate the proportion of dimers based on the dimer K OA Predicted value, single molecule K OA The predicted value, the proportion of dimers and the proportion of single molecules were used to obtain the corrected K value of aromatic pollutants. OA value.

[0046] Based on the above content, the present invention provides a method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers, and uses the solvation model density (SMD) model to calculate the solvation free energy of aromatic pollutants, and converts the solvation free energy into K through thermodynamic relations. OA The QSAR model was used to establish a dimer prediction model for K OA The K values ​​of aromatic pollutants can be accurately and quickly predicted. OA The prediction method constructed by the present invention can quickly and accurately predict the K value of any aromatic pollutant based only on the molecular structure. OAThe value effectively reduces the measurement cost of the experimental method and improves the prediction efficiency, providing a basis for a comprehensive understanding of the migration, transformation and bioaccumulation processes of pollutants in different environmental media, as well as their potential risks to ecosystems and human health, and providing a useful reference for the control of aromatic pollutants.

[0047] Specifically, this embodiment provides K OA The details of the value forecast include:

[0048] First, this embodiment gives the current status of related research. At present, a variety of continuous solvation models have been developed and have been widely used in the calculation of solvation free energy, such as polarized continuum model, conductor-like shielding model, solvation model density (SMD) model and some SMx series models. Among these solvation models, the SMD model is generally considered to be a reliable and effective solvation model, and its performance varies with the different calculation levels formed by the combination of the quantitative theoretical calculation method and the basis set. The calculation level formed by the combination of the HF theoretical method and the MiDi! 6D basis set is more accurate in describing the molecular geometry, so this calculation level is more accurate in calculating the solvation free energy of aromatic pollutants. Based on this, this embodiment selects the HF / MiDi! 6D calculation level combined with the SMD model to calculate the solvation free energy of aromatic pollutants. The distribution coefficient K of the compound between the gas phase and the n-octanol phase OA The thermodynamic relationship to the solvation free energy is shown below:

[0049]

[0050] In the formula, ΔG OA is the free energy change during the transfer of the compound from the gas phase to the n-octanol phase, that is, the solvation free energy, R is the ideal gas constant, and T is the absolute temperature.

[0051] From formula (1), we can know that the equilibrium distribution coefficient of aromatic pollutants between any two phases can be obtained based on the solvation free energy of aromatic pollutants transferred between different phases. The generalized form of formula (1) is:

[0052]

[0053] That is, according to the thermodynamic relationship, as long as the corresponding solvation free energy is obtained, the equilibrium distribution coefficient of any organic compound between two phases can be predicted. Therefore, in theory, the K of the compound can be calculated by the thermodynamic relationship OA Moreover, this method is not limited by the application domain. The corresponding distribution coefficient can be further estimated by calculating the solvation free energy only through the structure of the compound.

[0054] In addition, some studies have shown that aromatic pollutants such as benzene and PAHs are prone to form dimer structures driven by the interaction between aromatic rings. Other studies have shown that the driving force for the formation of dimers between benzene and PAHs is the π-π interaction between aromatic rings. OA In the experimental determination process, the concentration of aromatic pollutants is high and the π-π interaction between molecules is large. Therefore, aromatic pollutants are prone to form dimers driven by π-π interactions. It is speculated that the K of aromatic pollutants OA The experimental value may be a mixture of monomers and dimers rather than a single molecule. Since dimers are more hydrophobic and lipophilic than single molecules, the K value of aromatic pollutants is OA The experimental value will be higher than that of the case where only a single molecule exists. OA The predicted values ​​are based on single molecules, which will lead to K OA The experimental values ​​are all higher than the predicted values. Based on this, the influence of dimer form on the predicted value should be considered in the prediction process of the n-octanol-air partition coefficient of aromatic pollutants. π-π interaction is a non-covalent interaction and a type of weak interaction. Studies have shown that the size and direction of π-π interaction depend on the molecular structure, molecular weight and the size of the relative aromatic ring surface. Substituents on the aromatic ring will also affect the π-π interaction because the substituents will affect the electron density of the π electron cloud, thereby changing the electrostatic interaction. Therefore, the dimer ratio of different types of pollutants is related to the strength of the π-π interaction. As a quantitative structure-activity relationship model, the QSAR model can characterize its environmental behavior through molecular structure, and the aggregation of pollutants through intermolecular interactions to form dimers is a typical environmental behavior. Therefore, this method selects the QSAR model to predict the dimer ratio of aromatic pollutants.

[0055] Based on this, the present invention constructs the K of typical aromatic pollutants based on the open source SMD model and thermodynamic relationship, combined with quantum chemical methods. OA The QSAR model was used to predict the dimer ratio of aromatic pollutants and the K OA The present invention proposes to use dimers to correct the K values ​​of aromatic pollutants. OA Predicted values, for aromatic pollutants K OA Therefore, the invention can be used to make up for the pollutant K OA It can reduce the loss of values, reduce the cost of obtaining physical and chemical properties, reduce the error of theoretical prediction methods, and help chemical companies deal with chemical management regulations such as the EU REACH regulation. It is of great significance for evaluating the long-distance environmental migration ability and bioaccumulation of pollutants.

[0056] The detailed process of this embodiment includes:

[0057] S1. Collect the experimental values ​​of n-octanol-air partition coefficients of aromatic pollutants and form a data set, and divide the data set into a training set and a validation set;

[0058] Specifically, this example collected four types of aromatic pollutants, including 16 PAHs, 30 PBDEs, 24 PCNs and 10 PCDDs in the literature. OA Experimental values, for K in different literatures OA The average of the experimental values ​​was used as the data set for this model. Regression models were established using the experimental values ​​and the predicted values ​​based on single molecules, as well as the experimental values ​​and the corrected predicted values. The goodness of fit (R 2 ), root mean square error (RMSE) and leave-one-out cross-validation coefficient (Q 2 CV )Evaluation K OA Goodness of fit, prediction accuracy, and robustness of the linear regression model between experimental and predicted values.

[0059] S2. Using the SMD model at the HF / MiDi! 6D calculation level, the single point energy of the target aromatic pollutant single molecule in the gas phase and the n-octanol phase was calculated respectively. The solvation free energy of the single molecule was obtained based on the difference between the single point energy of the gas phase and the n-octanol phase. The single molecule K was calculated based on the thermodynamic relationship and the solvation free energy. OA Predicted value, that is, calculated according to formula (3) for single molecule K OA Predicted value,

[0060]

[0061] In the formula, ΔG OA (single molecule) is the solvation free energy of a single molecule between the gas phase and the n-octanol phase, R is the ideal gas constant, T is the absolute temperature, K OA (single molecule) is single molecule K OA Predicted value.

[0062] Specifically, first, the molecular structure of aromatic pollutants was drawn using CS ChemDraw Ultra (Version 14.0, Cambridge Scientific Computing, Inc.). Next, the SMD model was used on the Gaussian 09 platform to perform geometry optimization, frequency calculation, and single-point energy calculation for the target aromatic pollutants in the gas phase and n-octanol phase at the HF / MiDi! 6D calculation level. The solvation free energy (ΔG) of the target aromatic pollutants between the two phases was obtained from the difference in single-point energy between the two phases. OA ). Using the log K of aromatic pollutants OA Experimental value and logK OAThe regression model is shown in Figure 2(a). The statistical parameters show that the goodness of fit of the model (R 2 ) is 0.636, the root mean square error (RMSE) is 2.731, and the cross validation coefficient (Q 2 CV ) is 0.622. The statistical parameter results show that the goodness of fit of the prediction model needs to be improved.

[0063] S3. Using the SMD model at the HF / MiDi! 6D calculation level, the single point energy of the target aromatic pollutant dimer in the gas phase and the n-octanol phase was calculated respectively. The solvation free energy of the dimer was obtained based on the difference in the single point energy between the gas phase and the n-octanol phase. The dimer K was calculated based on the thermodynamic relationship and the solvation free energy. OA Predicted value;

[0064] Specifically, the molecular structure of the aromatic pollutant dimer is drawn in the same manner as step (2). Similarly, on the Gaussian09 platform, the SMD model is used to optimize the geometry, calculate the frequency and calculate the single point energy of the aromatic pollutant dimer in the gas phase and the n-octanol phase at the HF / MiDi! 6D calculation level. The solvation free energy (ΔG) of the aromatic pollutant dimer between the two phases is obtained from the difference in the single point energy between the two phases. OA ). According to the thermodynamic relationship, the dimer K of the target aromatic pollutant is obtained. OA Predicted value, that is, calculated according to formula (4) OA Predicted value,

[0065]

[0066] In the formula, ΔG OA (dimer) is the solvation free energy of the dimer between the gas phase and the n-octanol phase, R is the ideal gas constant, T is the absolute temperature, K OA (Dimer) is dimer K OA Predicted value.

[0067] S4. Constructing a QSAR model, wherein the QSAR model is used to OA Predicted value, dimer K OA The predicted value predicts the proportion of aromatic pollutants existing as dimers, randomly selects data from the data set to form a training set, and trains the QSAR model based on the training set;

[0068] S5. Calculate the proportion of dimers in the target aromatic pollutants based on the trained QSAR model, calculate the proportion of single molecules based on the proportion of dimers in the target aromatic pollutants, and calculate the proportion of dimers based on the dimer K OA Predicted value, single molecule KOA The predicted value, the proportion of dimers and the proportion of single molecules were used to obtain the corrected K value of aromatic pollutants. OA value.

[0069] Specifically, S4 and S5 include:

[0070] Part I: Obtaining the measured value of the proportion of aromatic pollutants existing in dimer form:

[0071] Due to the lack of corresponding experimental values, the experimental values ​​of the proportion of aromatic pollutants in dimer form are estimated by formula (5) and formula (6):

[0072] logK OA (Experimental value) = single molecule % * log K OA (single molecule prediction) + dimer %*logK OA (Dimer

[0073] Predicted value) (5) Single molecule % + dimer % = 100% (6)

[0074] Among them, the experimental value in formula (5) comes from the data set, and the single molecule prediction value and dimer prediction value are calculated by S2 and S3 respectively. In addition, since aromatic pollutants rarely exist in the form of polymers, the present invention only considers the effect of dimers on the distribution coefficient and ignores the effect of polymers, thereby deriving formula (6), indicating that aromatic pollutants exist only in the form of single molecules or dimers.

[0075] Part II: QSAR model predicts the proportion of aromatic pollutants existing in dimer form:

[0076] The analysis results of the components and main influencing factors of π-π interaction show that π-π interaction can be decomposed into electrostatic term, London dispersion term, induction (polarization) term and exchange-repulsion term. Among them, electrostatic term and induction term are related to electrostatic potential, while London dispersion term and exchange-repulsion term are related to van der Waals potential. Therefore, five molecular structure descriptors related to electrostatic potential (ESP) and van der Waals (vdW) potential were selected to establish QSAR model to predict the proportion of aromatic pollutants in dimer form. The selected molecular structure descriptors include dihedral angle (DA), relative molecular mass (Mw), nonpolar surface area (NSA), polar surface area (PSA) and molecular polarity index (MPI).

[0077] Among them, DA is a parameter to determine whether the compound is planar or non-planar, NSA and PSA represent van der Waals potential and electrostatic potential, respectively, and Mw and MPI characterize the polarity of the compound. The calculation method of molecular structure descriptors is as follows: DA is calculated by GaussView 6.0, Mw is calculated by PM7 method in MOPAC 2016, and the other three molecular structure descriptors are calculated by quantitative analysis of molecular surface module in Multiwfn software.

[0078] 75% of the data set was randomly selected as the training set, and the remaining 25% was used as the validation set. The above five molecular structure descriptors were used to establish the QSAR model using partial least squares (PLS) regression in Simca-S software (Version 13.0, Umetri AB & Erisoft AB), and the expression is:

[0079] X = -1.927 + 7.809 × 10 -3 NSA+1.510×10 -1 MPI - 2.218 × 10 -3 PSA

[0080] +1.721×10 -3 DA-1.118×10 -3 Mw(7)

[0081] N TRA =60,p=5,A=4,R 2 X(cum) =0.994,R 2 Y(cum) =0.815,Q 2 (cum) =0.793,

[0082] RMSE=0.071,N VAL =20,RMSE VAL =0.062,Q 2 VAL =0.800

[0083] Where N and p are the number of samples and the number of predictor variables in the model, respectively, and X = dimer%, where dimer% represents the proportion of the target aromatic pollutant in the form of dimers. Statistical parameters show that the model has a high Q 2 (cum) The smaller the RMSE value, the more robust the model is and the better the goodness of fit is. Figure 3It can be seen that the dimer ratio predicted by the QSAR model has a good correlation with the experimental value, indicating that the model has good predictive ability. Therefore, the QSAR model established in the present invention can accurately predict the ratio of aromatic pollutants in the form of dimers.

[0084] Log K OA (single-molecule prediction) and log K OA (dimer predicted value) and the dimer ratio and single molecule ratio obtained according to formula (6) and formula (7) are substituted into formula (8) to use the dimer effect to calculate the K of aromatic pollutants. OA The value is corrected to obtain the corrected predicted value of the aromatic pollutant, Y = single molecule %, single molecule % represents the proportion of single molecules in the target aromatic pollutant.

[0085] logK OA (Correction value) = Y*log K OA (single molecule prediction value)+X*log log K OA (Dimer predicted value) (8)

[0086] Using the logK of aromatic pollutants OA Experimental value and logK OA The regression model is established based on the correction value. The regression model is shown in Figure 2(b). The logK OA The values ​​are listed in Table 1.

[0087] Statistical parameters showed that the model was goodness of fit (R 2 ) is 0.941, the root mean square error (RMSE) is 0.441, and the cross validation coefficient (Q 2 CV ) is 0.932. This shows that the prediction model has high goodness of fit, robustness and prediction accuracy. Figure 4 As shown, compared with logK OA Experimental values ​​and regression model based on single-molecule prediction values ​​and logK OA Experimental value and logK OA The regression model established by the correction value. The results show that the goodness of fit, robustness and prediction accuracy of the model are greatly improved. In addition, the application domain of the present invention is wider and less dependent on experimental data. Therefore, the K OA The prediction model is a reliable way to estimate the logK of aromatic pollutants. OA Value method.

[0088] Table 1 logK of aromatic pollutants OA Correction value

[0089]

[0090]

[0091] Taking the aromatic pollutant Phenanthrene as an example, the logK OA The experimental value is 7.65. First, the SMD model was used in the Gaussian09 platform to optimize the geometry, calculate the frequency and calculate the single point energy of phenanthrene in the gas phase and n-octanol phase at the HF / MiDi! 6D calculation level. The solvation free energy (ΔG OA ). According to the thermodynamic relationship of formula (1), the logK of Phenanthrene based on a single molecule is obtained: OA The predicted value is 6.45.

[0092] Next, the same calculation level is used to obtain the K of the phenanthrene dimer OA The dimer ratio QSAR prediction model based on formula (7) estimates the proportion of phenanthrene in the form of single molecule and dimer respectively. The log K of phenanthrene is calculated using the dimer effect in formula (8). OA The corrected value was 7.55. OA Predicted and corrected logK OA Prediction value. It can be seen that after correction, K OA The deviation between the experimental value and the predicted value is significantly reduced. Therefore, it is suggested that the dimer ratio has a significant effect on K OA The predicted value has a greater impact on K OA The dimer effect needs to be considered when making theoretical predictions of the values.

[0093] Taking the aromatic pollutant PCN-5 as an example, the log K OA The experimental value is 6.93. First, the geometry optimization, frequency calculation and single point energy of PCN-5 in gas phase and n-octanol phase were performed using the SMD model at the HF / MiDi! 6D calculation level on the Gaussian09 platform. The solvation free energy (ΔG OA ). According to the thermodynamic relationship of formula (1), the logK of PCN-5 based on a single molecule is obtained. OA The predicted value is 4.96.

[0094] Next, the same calculation level was used to derive the K of the PCN-5 dimer OAThe dimer ratio QSAR prediction model based on formula (7) estimates the proportion of PCN-5 in the form of single molecule and dimer respectively. The log K of PCN-5 is calculated using the dimer effect in formula (8). OA The corrected value was 6.66. OA Predicted and corrected log K OA Prediction value. It can be seen that after correction, K OA The deviation between the experimental value and the predicted value is significantly reduced. Therefore, it is suggested that the dimer ratio has a significant effect on K OA The predicted value has a greater impact on K OA The dimer effect needs to be considered when making theoretical predictions of the values.

[0095] Taking the aromatic pollutant PBDE-28 as an example, the logK OA The experimental value is 9.5. First, the SMD model was used in the Gaussian09 platform to optimize the geometry, calculate the frequency and calculate the single point energy of PBDE-28 in the gas phase and n-octanol phase at the HF / MiDi! 6D calculation level. The solvation free energy (ΔG OA ). According to the thermodynamic relationship of formula (1), the logK of PBDE-28 based on a single molecule is obtained: OA The predicted value is 7.65.

[0096] Next, the same calculation level was used to derive the K of the PBDE-28 dimer. OA The dimer ratio QSAR prediction model based on formula (7) estimates the proportion of PBDE-28 in the form of single molecules and dimers. The logK of PBDE-28 is calculated using the dimer effect in formula (8). OA The corrected value was 9.82. OA Predicted and corrected log K OA Prediction value. It can be seen that after correction, K OA The deviation between the experimental value and the predicted value is significantly reduced. Therefore, it is suggested that the dimer ratio has a significant effect on K OA The predicted value has a greater impact on K OA The dimer effect needs to be considered when making theoretical predictions of the values.

[0097] Taking the aromatic pollutant PCDD-1 as an example, the logK OAThe experimental value is 7.86. First, the SMD model was used in the Gaussian09 platform to optimize the geometry, calculate the frequency and calculate the single point energy of PCDD-1 in the gas phase and n-octanol phase at the HF / MiDi! 6D calculation level. The solvation free energy (ΔG OA ). According to the thermodynamic relationship of formula (1), the logK of PCDD-1 based on a single molecule is obtained: OA The predicted value is 4.62.

[0098] Next, the same calculation level was used to derive the K of the PCDD-1 dimer OA The dimer ratio QSAR prediction model based on formula (7) estimates the proportion of PCDD-1 in the form of single molecule and dimer respectively. The logK of PCDD-1 is calculated using the dimer effect in formula (8). OA The corrected value was 7.72. OA Predicted and corrected logK OA Prediction value. It can be seen that after correction, K OA The deviation between the experimental value and the predicted value is significantly reduced. Therefore, it is suggested that the dimer ratio has a significant effect on K OA The predicted value has a greater impact on K OA The dimer effect needs to be considered when making theoretical predictions of the values.

[0099] Overall beneficial effects:

[0100] The present invention provides a method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers. The method is a fast and simple method for predicting the n-octanol-air partition coefficient of aromatic pollutants. OA Using the SMD model and thermodynamic relationships, the K of the substance can be obtained only through the molecular structure coordinates. OA In addition, the dimer ratio was used to determine the K of aromatic pollutants. OA After the predicted value is corrected, K OA The accuracy of the predicted values ​​is significantly improved, and this correction method is more in line with the real environment under measured conditions. Statistical parameters show that the model has a high goodness of fit, good robustness and stability, low dependence on measured values ​​in the modeling stage and is not limited by the application domain.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers, characterized in that: include: S1, n-octanol-air partition coefficient K for collecting aromatic pollutants OA Experimental values ​​and form a data set; S2. Using the SMD model at the HF / MiDi! 6D calculation level, the single point energy of the target aromatic pollutant single molecule in the gas phase and the n-octanol phase was calculated respectively. The solvation free energy of the single molecule was obtained based on the difference between the single point energy of the gas phase and the n-octanol phase. The single molecule K was calculated based on the thermodynamic relationship and the solvation free energy of the single molecule. OA Predicted value; S3. Using the SMD model at the HF / MiDi! 6D calculation level, the single point energy of the target aromatic pollutant dimer in the gas phase and the n-octanol phase was calculated respectively. The solvation free energy of the dimer was obtained based on the difference in the single point energy between the gas phase and the n-octanol phase. The dimer K was calculated based on the thermodynamic relationship and the solvation free energy of the dimer. OA Predicted value; S4. Constructing a QSAR model, wherein the QSAR model is used to OA Predicted value, dimer K OA The predicted value predicts the proportion of aromatic pollutants existing as dimers, randomly selects data from the data set to form a training set, and trains the QSAR model based on the training set; S5. Calculate the proportion of dimers in the target aromatic pollutants based on the trained QSAR model, calculate the proportion of single molecules based on the proportion of dimers in the target aromatic pollutants, and calculate the proportion of dimers based on the dimer K OA Predicted value, single molecule K OA The predicted value, the proportion of dimers and the proportion of single molecules were used to obtain the corrected K value of aromatic pollutants. OA value.

2. The method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers according to claim 1, characterized in that: The S2 includes: S21. Use the chemical drawing software CS ChemDraw Ultra to draw the molecular structure of the target aromatic pollutant; S22. Using the Gaussian09 platform, the SMD model was used to perform geometry optimization, frequency calculation, and single-point energy calculation on the single molecule of the target aromatic pollutant in the gas phase and the n-octanol phase at the HF / MiDi! 6D calculation level, and the solvation free energy ΔG between the gas phase and the n-octanol phase was obtained. OA ; S23. Calculate single molecule K according to formula (1) OA Predicted value, In the formula, ΔG OA (single molecule) is the solvation free energy of a single molecule between the gas phase and the n-octanol phase, R is the ideal gas constant, T is the absolute temperature, K OA (single molecule) is single molecule K OA Predicted value.

3. The method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers according to claim 2, characterized in that: The S3 includes: S31. Use the chemical drawing software CS ChemDraw Ultra to draw the molecular structure of the target aromatic pollutant; S32. Using the Gaussian09 platform, the SMD model was used to perform geometry optimization, frequency calculation, and single-point energy calculation of the target aromatic pollutant dimer in the gas phase and the n-octanol phase at the HF / MiDi! 6D calculation level, and the solvation free energy ΔG between the dimer in the gas phase and the n-octanol phase was obtained. OA ; S33. Calculate the dimer K according to formula (2): OA Predicted value, In the formula, ΔG OA (dimer) is the solvation free energy of the dimer between the gas phase and the n-octanol phase, R is the ideal gas constant, T is the absolute temperature, K OA (Dimer) is dimer K OA Predicted value.

4. The method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers according to claim 3, characterized in that: The S4 includes randomly selecting data from the data set to form a training set, and selecting five molecular structure descriptors, namely, dihedral angle DA, relative molecular mass Mw, non-polar surface area NSA, polar surface area PSA and molecular polarity index MPI, and using partial least squares regression in Simca-S software to establish a QSAR model, the expression is: X=-1.927+7.809×10 -3 NSA+1.510×10 -1 MPI-2.218×10 -3 PSA+ 1.721×10 -3 DA-1.118×10 -3 Mw(3) X+Y=100% (4) Wherein, X=dimer%, dimer% indicates the proportion of aromatic pollutants existing in the form of dimers, the training set is used to train the QSAR model and obtain the trained QSAR model, Y=single molecule%, single molecule% indicates the proportion of target aromatic pollutants existing in the form of single molecules, and the trained QSAR model and dimer% can be used to obtain the single molecule%.

5. The method for predicting the n-octanol-air partition coefficient of aromatic pollutants based on solvation free energy and dimers according to claim 4, characterized in that: The S5 includes: K OA (single molecule), K OA (dimer), the values ​​of X and Y in the target aromatic pollutant calculated according to the trained QSAR model are substituted into formula (5) to obtain the predicted value of the target aromatic pollutant after correction, logK OA (Correction value) = Y*logK OA (single molecule)+X*logK OA (dimer) (5) The positive value indicates the K of the target aromatic pollutant after correction. OA Predicted value, X=dimer%, dimer% indicates the proportion of dimers in the target aromatic pollutant, Y=monomolecule%, monomolecule% indicates the proportion of the target aromatic pollutant in the form of monomolecule.

Images