Prediction of Glass Transition Temperatures from Monomer and Repeat Unit Structure Using Computational Neural Networks

Abstract

Quantitative structure-property relationships (QSPR) are developed to correlate glass transition temperatures and chemical structure. Both monomer and repeat unit structures are used to build several QSPR models for Parts 1 and 2 of this study, respectively. Models are developed using numerical descriptors, which encode important information about chemical structure (topological, electronic, and geometric). Multiple linear regression analysis (MLRA) and computational neural networks (CNNs) are used to generate the models after descriptor generation. Optimization routines (simulated annealing and genetic algorithm) are utilized to find information-rich subsets of descriptors for prediction. A 10-descriptor CNN model was found to be optimal in predicting T(g) values using the monomer structure (Part 1) for 165 polymers. A committee of 10 CNNs produced a training set rms error of 10.1K (r2 = 0.98) and a prediction set rms error of 21.7 K (r2 = 0.92). An 11-descriptor CNN model was developed for 251 polymers using the repeat unit structure (Part 2). A committee of CNNs produced a training set rms error of 21.1K (r2 = 0.96) and a prediction set rms error of 21.9 K (r2 = 0.96).

Prediction of refractive index of vinyl polymers by using density functional theory

“…But molecular descriptors calculated directly from the structure of the monomers can be used on the study of QSPRs for polymers, 8,9 since all the properties depend on the chemical structure of the polymer molecule, and all this structure is conditioned by the monomer structure. Therefore this method is adopted in this paper to calculate molecular quantum chemical descriptors.…”

Glass transition temperature prediction of polymers through the mass-per-flexible-bond principle

“…These values are remarkably low considering the simplicity of the model, the large data scatter in the T g vs. M/f plots, and the structural variety in the library. In fact, this accuracy is better than most of the results from (semi-) empirical T g prediction methods found in the literature [30][31][32][33][34][35][36][37][38][39][40][41][42][43]50].…”

“…Other commonly used methods for T g prediction are: ab initio quantum mechanical calculations [21,22], Monte Carlo [23,24], and molecular dynamics simulations [25][26][27][28][29] and semi-empirical or empirical methods based on group contribution methods, often using the QSPR approach through neural network computation [30][31][32][33][34][35][36][37][38][39][40][41][42][43][44]. While some of these methods yield good results by predicting glass transition temperatures with errors as good as 3-10 K, most predictive accuracies are on the order of 20-100K.…”

Prediction of Thermal Decomposition Temperature of Polymers Using QSPR Methods

“…[1][2][3][4][5][6][7] Thermal stability of a material is stability against degradation upon exposure to elevated temperatures in an inert environment. Polymers are often exposed to high temperatures during processing and/or use.…”

“…In the forward selection the variable considered for inclusion at any step is the one yielding the largest single degree of freedom F-ratio among the variables that are eligible for inclusion. Consequently, at each step, the j th variable is added to a k-size model if (1) In the above inequality RSS is the residual sum of squares and s is the mean square error. The subscript k + j refers to quantities computed when the j th variable is added to the k variables that are already included in the model.…”