Study On Quantitative Structure-property Relationship Between Molecular Structure And Solubility Of Nitrogen Heterogeneous Compounds In Petroleum

Nitrogen heterocyclic compounds (NHCs) are a kind of toxic and refractory organics. It’s essential for the development of denitrification technology and wastewater treatment to get information about the solubility of NHCs in water, especially in aqueous solutions of different pH. In this work,25 representative NHCs from petroleum fraction were selected as sample compounds, and the solubility of them was measured in pure water, buffer solutions of different pH and salt solutions of different concentration. Based on the measured solubility, the molecular structure-solubility relationship of NHCs was investigated and models for prediction of the solubility of NHCs in pure water and solutions of different pH were built.UV-VIS spectrophotometry or potentiometric titration was selected to analyze the solubility based on the structure properties of NHCs and a systematic analytical method of solubility measurement of NHCs was thus established. With the solubility measurement method established,160 solubility values of NHCs were obtained in pure water, different buffer solutions of pH 4.0,7.0,9.2 and NaCl solutions (CNaCl= 0.2、0.6、 1.0 mol/L) at 298.15 ± 0.02 K. The results showed that the effect of pH on the solubility of NHCs was different for NHCs with different ring types, which could be explained as the changes of the electron density on heterocyclic Nitrogen atom of NHCs. The study on the solubility of NHCs in NaCl solutions showed that the salting-out effect could be neglected in the concentration range of the buffer solutions used in the study since the values of Ks was small. The structural parameters of NHCs and Ks were also correlated and the correlation between the partial charge of heterocyclic nitrogen atom (QN) and Ks was found to a good linearity, which proved that Qk could describe the molecular structures of NHCs well and was suitable as one of ideal structural parameters of NHCs.Five best molecular descriptors (BIC4, RDFOlOv, Mor29e, Mor14v, R6m) were selected from 1664 molecular descriptors calculated by Dragon software by Genetic Algorithm-Multiple Linear Regression (GA-MLR) and the prediction model of the solubility of NHCs in pure water was built by MLR. Meanwhile, an ANN model for the prediction of the solubility of NHCs in pure water was built with the same five molecular descriptors and the comparison between the two models showed that ANN was better than MLR for the prediction of the solubility of NHCs in pure water. Since MLR was built based on linear structure-property relationship while ANN was based on nonlinear relationship, the comparison result indicated that the relationship between the molecular descriptors and the solubility of NHCs was nonlinear to some extent. But the prediction result was unsatisfying only with the five descriptors selected by GA-MLR for the solubility of NHCs in solutions of different pH. QN was then introduced and the ANN model of the solubility of NHCs in solutions of different pH could predict well with the five descriptors, pH and QN as model inputs. A comparison was made between the model and a conventional modeling approach (HH equation) and the comparison result indicated that the model built with QN as an extra descriptor was better than HH approach. Since QN quantified the electron density of heterocyclic nitrogen atom and reflected the degree of forming salts of NHCs in solutions of different pH,QN should be introduced to describe the tendency of structure changes of NHCs in solutions of different pH and was the key of building quantitative structure-solubility relationship of NHCs in solutions of different pH.