Loading...

Prediction of Density, Surface Tension and Viscosity of Ionic Liquids using Artificial Intelligence

Torrik, Abdolhalim | 2017

446 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 49484 (03)
  4. University: Sharif University of Technology
  5. Department: Chemistry
  6. Advisor(s): Rahman Setayesh, Shahrbanoo
  7. Abstract:
  8. Ionic liquids due to their unique physical and chemical properties are widely used in various fields of chemistry and chemical engineering. Determining properties of ionic liquids can be difficult, time consuming or even impossible. In first step of this study the IL thermo database was chosen as source of experimental data of ionic liquids. Viscosity, density, surface tension and melting temperature of ionic liquids based on bis(trifluoromethylsulfonyl)imide anion with cations imidazolium, pyridinium, pyrrolidinium, piperidinium and ammonium were modeled using QSPR approach. Linear and nonlinear models were built by means of multiple linear regression (MLR), multilayer perceptron neural networks (MLP) and using genetic algorithm. Viscosity of ionic liquids was modeled using experimental data for 87 ionic liquids. Both models showed good accuracy in terms of average absolute relative deviation and the AARD value was 4.29% for linear model and 4.28% for non-linear model. Melting temperature of ionic liquids was modeled using experimental data for 74 ionic liquids. AARD values were 4.29% for linear model and 4.28% for non-linear model. For density, experimental data of 48 ionic liquids was used for modeling and two models were developed. AARD values were 0.90% for linear model and 0.98% for non-linear model. Surface tension was modeled for 24 ionic liquids. AARD values were 2.72% for linear model and 1.75% for non-linear model. The applicability domains of the proposed models were also determined. Also using obtained descriptors, effect of molecular structure on the physical properties was addressed. Structural information presented in this work can be used for rational design of new series of ionic liquids
  9. Keywords:
  10. Ionic Liquids ; Genetic Algorithm ; Artificial Neural Network ; Chemometrics Method ; Viscosity ; Surface Tension

 Digital Object List

 Bookmark

No TOC