A neural network approach to prediction of glass transition temperature of polymers

Abstract

Polymeric materials are finding increasing application in commercial optical communication systems. Taking advantage of techniques from the field of artificial intelligence, the goal of our research is to construct systems that can computationally design polymer formulations, including polymer optical fibers, with specified desirable consumer characteristics. Through the use of an extensive structure-property correlation database, properties of polymers can be predicted by an artificial network and the structure of novel polymers with desired properties can be optimized by a genetic algorithm. In this paper, we are focusing on one of the parameters, glass transition temperature (T g ) that influences a desired outcome in polymer optical fibers. Performance of such fibers can be optimized by engineering a polymer to exhibit a lower refractive index and T g . This paper compares and discusses a neural network model and a linear model that have been developed to correlate T g and repeating units of polymers. A neural network and multiple linear regression analysis were used in the study. A set of descriptors, chosen based on previous studies on the relations between T g and polymer structure, were used to describe the structure of repeating units, individual bond energies, and intermolecular forces, especially hydrogen bonding, which is the strongest intermolecular force and exerts the greatest influence on T g compared with other intermolecular interactions. A comprehensive neural network model with 28 descriptors was developed to predict T g values of 6 randomly selected polymers from a database containing 71 polymers. The network was trained with the remaining 65 polymers and had a typical training root mean square error of 17 K (R 2 = 0.95) and prediction average error of 17 K (R 2 = 0.85). A linear regression model developed for comparison had an average error of 30 K (R 2 = 0.88).

Design, synthesis, characterization and optimization of PTT-b-PEO copolymers: A new membrane material for CO2 separation

“…Other approaches such as theoretical simulation and optimization are also way to design new polymers, but they require data bases, special programs, experts, etc. Actually, different procedures for designing and optimization of new polymers are reported in the literature [1][2][3][4][5][6].…”

A simple three‐descriptor model for the prediction of the glass‐transition temperatures of vinyl polymers

“…This indicates that the correlation between T g and the structural parameters outlined previously was nonlinear rather than linear. In comparison with previous models, [5][6][7][8][9] this ANN model showed better statistical quality. Table II shows that the three descriptors were all significant descriptors from the significance test.…”

Soft Matter Informatics: Current Progress and Challenges

“…As early as 1994, Sumpter and Noid showed a neural network approach improved predictions of glass transition temperatures from polymer structural descriptors over more traditional QSPR methods . Glass transition temperature has proved to be a fertile test case for ML application, with many studies building on the work of Sumpter and Noid by more advanced applications of neural networks to larger and more diverse data sets . Yu et al.…”