Deep learning based approach for prediction of glass transition temperature in polymers
| Autor: | Arohan Neog, Bitopan Das, Subhasish Goswami, Rajdeep Ghosh |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
010302 applied physics
chemistry.chemical_classification Materials science business.industry Deep learning Term memory 02 engineering and technology Polymer 021001 nanoscience & nanotechnology 01 natural sciences Condensed Matter::Soft Condensed Matter chemistry 0103 physical sciences Molecule Artificial intelligence Statistical physics Experimental methods 0210 nano-technology business Glass transition |
| Zdroj: | Materials Today: Proceedings. 46:5838-5843 |
| ISSN: | 2214-7853 |
| DOI: | 10.1016/j.matpr.2021.02.730 |
| Popis: | Glass Transition Temperature is one of the most studied fields in material science and measurement of Glass Transition Temperatures for the ever-expanding list of polymers in an accurate and efficient manner has been the scope of many recent works. Glass Transition Temperatures of polymers depend on various physical as well as chemical properties and thus calculation of Glass Transition Temperature using empirical methods require a detailed study of these chemicals. Empirical methods depend on the heating rate and the history of the material and for polymers with a wide range of glass transition temperatures, these experimental methods can pose challenges on the first run. This paper proposes a Long Short- Term Memory (LSTM) model based on Simplified Molecular-Input Line-Entry System (SMILES) structure of polymer molecules to predict the Glass Transition Temperature. The paper goes on to study the results obtained from the model and discuss application of the proposed model in real world. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect