%0 Journal Article
%T Determining the Aggregation Number of Anionic Surfactants based on Conductivity Method: Employing QSAR-ANN Modelling Techniques for Predicting the aggregation number of surfactants
%J Applied Chemistry Today
%I
%Z 2981-2437
%A abdous, behnaz
%A Sajjadi, S. Maryam
%A Bagheri, Ahmad
%D 2022
%\ 06/22/2022
%V 17
%N 63
%P 87-108
%! Determining the Aggregation Number of Anionic Surfactants based on Conductivity Method: Employing QSAR-ANN Modelling Techniques for Predicting the aggregation number of surfactants
%K Molecular Descriptors
%K QSAR
%K Artificial Neural Network
%K Anionic Surfactant
%K Aggregation Number
%R 10.22075/chem.2022.25907.2042
%X In this study, a new modeling method using QSAR model and artificial neural network is used to predict the aggregation number of some anionic surfactants in aqueous solution at 25 °C. The micelle aggregation number was determined using electrical conductivity measurements and the Evans method for anionic surfactants in aqueous solutions. However, the obtained results based on conductibvity strategy were not in good agreement with those of fluorescence method. Since the fluorescence method is a more accurate method for calculating the aggregation number of micelles, the results of the fluorescence method have been used in this study. In order to correlate the molecular structure of these surfactants with their aggregation number, a quantitative structure-property relationship (QSPR) study was performed. An artificial neural network (ANN) model was developed to predict the aggregation number of anionic surfactants by using four out of more than 3200 molecular descriptors, calculated by Dragon software, as input variables. The importance of selected descriptors were computed based on ANN method and listed as follows in descending order: nC> X5V> MWC05> MWC04. The complete set of 24 anionic surfactants was randomly divided into a training set of 16, a test set of 4, and a validation set of 4 compounds. Also, multiple linear regression (MLR) analysis was utilized to build a linear model by using the same descriptors. Correlation coefficient (R2) and root mean square error (RMSE) of the ANN and MLR models (for the whole data set) were 0.94, 4.99 and 0.82, 8.38, respectively. The higher R2 of the ANN method showed that the relationship between the descriptors and the aggregation number of the compounds is nonlinear.
%U https://chemistry.semnan.ac.ir/article_6543_85af38a48f5e4a38dd0197801ae2e98b.pdf