Jiale Zhang, Aocheng Fan, Ziqi Wang, Lei Zheng, Yu Liu*, Zhijian Wang* and Peng Kang*,
{"title":"基于层次高斯过程回归的耐高温聚酰亚胺数据驱动设计。","authors":"Jiale Zhang, Aocheng Fan, Ziqi Wang, Lei Zheng, Yu Liu*, Zhijian Wang* and Peng Kang*, ","doi":"10.1021/acs.jpcb.5c03717","DOIUrl":null,"url":null,"abstract":"<p >Accurate prediction of Tg for polyimides (PIs) is essential for assessing material performance in high-temperature applications in aerospace, electronics, microelectronics, and flexible display technology. However, experimental measurements remain critically challenging due to the labor-intensive synthesis, conventional instrument limits, and time-consuming characterization processes. Meanwhile, force field limitations, timescale discrepancy, and validation difficulties exist in the prediction of Tg for PIs using molecular dynamic simulation. In this study, we introduce a hierarchical Gaussian process regression machine learning method that integrates prior knowledge to predict Tg for PIs with small-sample data sets. We employ RDKit for molecular descriptor calculation and feature selection. Twenty-one key descriptors are identified, and exceptional model performance with a coefficient of determination <i>R</i><sup>2</sup> of 0.98/0.74 on the training/test set is achieved, surpassing conventional machine learning approaches. We further use Shapley additive explanations analysis to study the actionable insights for designing thermally stable PIs. The number of rotatable bonds and minimum partial charge act as dominant factors influencing Tg. Validations through experimental synthesis and molecular dynamics simulations confirm that the prediction errors are below 15%, while a Bayesian update strategy employing a radial basis function kernel corrected systematic underestimation in the high-Tg regime (>270 °C). This work provides a robust, validated Tg prediction tool, elucidates critical structure–property relationships, and establishes a transferable framework for data-driven materials design, advancing the development of high-performance polymers.</p>","PeriodicalId":60,"journal":{"name":"The Journal of Physical Chemistry B","volume":"129 34","pages":"8821–8831"},"PeriodicalIF":2.9000,"publicationDate":"2025-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Data-Driven Design of High-Temperature-Resistant Polyimides Using Hierarchical Gaussian Process Regression\",\"authors\":\"Jiale Zhang, Aocheng Fan, Ziqi Wang, Lei Zheng, Yu Liu*, Zhijian Wang* and Peng Kang*, \",\"doi\":\"10.1021/acs.jpcb.5c03717\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p >Accurate prediction of Tg for polyimides (PIs) is essential for assessing material performance in high-temperature applications in aerospace, electronics, microelectronics, and flexible display technology. However, experimental measurements remain critically challenging due to the labor-intensive synthesis, conventional instrument limits, and time-consuming characterization processes. Meanwhile, force field limitations, timescale discrepancy, and validation difficulties exist in the prediction of Tg for PIs using molecular dynamic simulation. In this study, we introduce a hierarchical Gaussian process regression machine learning method that integrates prior knowledge to predict Tg for PIs with small-sample data sets. We employ RDKit for molecular descriptor calculation and feature selection. Twenty-one key descriptors are identified, and exceptional model performance with a coefficient of determination <i>R</i><sup>2</sup> of 0.98/0.74 on the training/test set is achieved, surpassing conventional machine learning approaches. We further use Shapley additive explanations analysis to study the actionable insights for designing thermally stable PIs. The number of rotatable bonds and minimum partial charge act as dominant factors influencing Tg. Validations through experimental synthesis and molecular dynamics simulations confirm that the prediction errors are below 15%, while a Bayesian update strategy employing a radial basis function kernel corrected systematic underestimation in the high-Tg regime (>270 °C). This work provides a robust, validated Tg prediction tool, elucidates critical structure–property relationships, and establishes a transferable framework for data-driven materials design, advancing the development of high-performance polymers.</p>\",\"PeriodicalId\":60,\"journal\":{\"name\":\"The Journal of Physical Chemistry B\",\"volume\":\"129 34\",\"pages\":\"8821–8831\"},\"PeriodicalIF\":2.9000,\"publicationDate\":\"2025-08-14\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"The Journal of Physical Chemistry B\",\"FirstCategoryId\":\"1\",\"ListUrlMain\":\"https://pubs.acs.org/doi/10.1021/acs.jpcb.5c03717\",\"RegionNum\":2,\"RegionCategory\":\"化学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"CHEMISTRY, PHYSICAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"The Journal of Physical Chemistry B","FirstCategoryId":"1","ListUrlMain":"https://pubs.acs.org/doi/10.1021/acs.jpcb.5c03717","RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"CHEMISTRY, PHYSICAL","Score":null,"Total":0}
Data-Driven Design of High-Temperature-Resistant Polyimides Using Hierarchical Gaussian Process Regression
Accurate prediction of Tg for polyimides (PIs) is essential for assessing material performance in high-temperature applications in aerospace, electronics, microelectronics, and flexible display technology. However, experimental measurements remain critically challenging due to the labor-intensive synthesis, conventional instrument limits, and time-consuming characterization processes. Meanwhile, force field limitations, timescale discrepancy, and validation difficulties exist in the prediction of Tg for PIs using molecular dynamic simulation. In this study, we introduce a hierarchical Gaussian process regression machine learning method that integrates prior knowledge to predict Tg for PIs with small-sample data sets. We employ RDKit for molecular descriptor calculation and feature selection. Twenty-one key descriptors are identified, and exceptional model performance with a coefficient of determination R2 of 0.98/0.74 on the training/test set is achieved, surpassing conventional machine learning approaches. We further use Shapley additive explanations analysis to study the actionable insights for designing thermally stable PIs. The number of rotatable bonds and minimum partial charge act as dominant factors influencing Tg. Validations through experimental synthesis and molecular dynamics simulations confirm that the prediction errors are below 15%, while a Bayesian update strategy employing a radial basis function kernel corrected systematic underestimation in the high-Tg regime (>270 °C). This work provides a robust, validated Tg prediction tool, elucidates critical structure–property relationships, and establishes a transferable framework for data-driven materials design, advancing the development of high-performance polymers.
期刊介绍:
An essential criterion for acceptance of research articles in the journal is that they provide new physical insight. Please refer to the New Physical Insights virtual issue on what constitutes new physical insight. Manuscripts that are essentially reporting data or applications of data are, in general, not suitable for publication in JPC B.