Predicting the volumetric properties of pure and mixture of amino acid-based ionic liquids

This work deals with the equation of state (EOS) and artificial neural network (ANN) modeling prediction of the volumetric properties of several pure and mixtures of amino acid ionic liquids (AAILs). Concerning EOS modeling, we extended a perturbed hard-sphere (PHS) EOS to predict the densities of pure and their mixtures of AAILs. The PHS EOS used the surface tension and liquid densities both at room temperature as the scaling constants. The accuracy of the PHS EOS has been checked by comparing the results with 344 experimental data points of 28 AAILs in 283-373.15 K temperature range. The overall average absolute deviation (MD) of the calculated densities from the literature data was found to be 0.55%. The PHS EOS has also been employed to calculate the densities and excess volumetric properties of 8 mixtures including AAIL + solvent in 298.15-318.15 K temperature range. From 279 data points examined, the MD of the predicted densities from the measurements was found to be 1.73%. Also, the ANN has been trained to predict the density of pure compounds and mixtures of interest. The optimized network contains one hidden layer for each layer and 29 and 37 neurons for pure and mixture respectively and a sigmoid transfer function. The results obtained reveal that the developed ANN technique is able to estimate accurately the density of AAILs providing an MD of 0.26% and 0.025% for pure and mixtures, respectively. (C) 2019 Elsevier B.V. All rights reserved.