Advancing the prediction of crystalline phases in glass-ceramics via machine learning

IF 3.2 3区 材料科学 Q1 MATERIALS SCIENCE, CERAMICS Journal of Non-crystalline Solids Pub Date : 2024-10-05 DOI:10.1016/j.jnoncrysol.2024.123251
Jiaqian Zhu, Guohao Sun, Linfeng Ding, Lianjun Wang
{"title":"Advancing the prediction of crystalline phases in glass-ceramics via machine learning","authors":"Jiaqian Zhu,&nbsp;Guohao Sun,&nbsp;Linfeng Ding,&nbsp;Lianjun Wang","doi":"10.1016/j.jnoncrysol.2024.123251","DOIUrl":null,"url":null,"abstract":"<div><div>Lithium aluminosilicate (LAS) glass-ceramics are widely utilized in diverse application, owing to their outstanding properties such as high transparency, high fracture toughness and ultra-low thermal expansion. The development of LAS glass-ceramics has traditionally relied on the phase diagram to identify primary crystalline phases. However, this approach is limited by constrained composition ranges and unknown heating treatment parameters, hindering the efficient development of high-performance glass-ceramics. In this study, we establish a comprehensive small-scale database of LAS glass-ceramics, comprising 751 samples characterized by 27 compositions, nucleation temperature, nucleation time, crystallization temperature, crystallization time and 13 crystalline phases. We employ five algorithms, i.e. Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Classification and Regression Trees (CART), K-Nearest Neighbors (K-NN) and Multi-Layer Perceptron (MLP) Classifier to predict the potential crystalline phases. Our results demonstrate that RF achieves the best overall performance, with the highest accuracy of 0.8609, the lowest hamming loss of 0.0142, and the highest micro F1 score of 0.9234. This work advances the understanding and prediction of crystalline phases in LAS glass-ceramics, providing valuable insights for the development and optimization of these materials.</div></div>","PeriodicalId":16461,"journal":{"name":"Journal of Non-crystalline Solids","volume":"646 ","pages":"Article 123251"},"PeriodicalIF":3.2000,"publicationDate":"2024-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Non-crystalline Solids","FirstCategoryId":"88","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0022309324004289","RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"MATERIALS SCIENCE, CERAMICS","Score":null,"Total":0}
引用次数: 0

Abstract

Lithium aluminosilicate (LAS) glass-ceramics are widely utilized in diverse application, owing to their outstanding properties such as high transparency, high fracture toughness and ultra-low thermal expansion. The development of LAS glass-ceramics has traditionally relied on the phase diagram to identify primary crystalline phases. However, this approach is limited by constrained composition ranges and unknown heating treatment parameters, hindering the efficient development of high-performance glass-ceramics. In this study, we establish a comprehensive small-scale database of LAS glass-ceramics, comprising 751 samples characterized by 27 compositions, nucleation temperature, nucleation time, crystallization temperature, crystallization time and 13 crystalline phases. We employ five algorithms, i.e. Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Classification and Regression Trees (CART), K-Nearest Neighbors (K-NN) and Multi-Layer Perceptron (MLP) Classifier to predict the potential crystalline phases. Our results demonstrate that RF achieves the best overall performance, with the highest accuracy of 0.8609, the lowest hamming loss of 0.0142, and the highest micro F1 score of 0.9234. This work advances the understanding and prediction of crystalline phases in LAS glass-ceramics, providing valuable insights for the development and optimization of these materials.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
通过机器学习推进玻璃陶瓷结晶相的预测
锂铝硅酸盐(LAS)玻璃陶瓷具有高透明度、高断裂韧性和超低热膨胀等优异性能,因此被广泛应用于各种领域。传统上,LAS 玻璃陶瓷的开发主要依靠相图来确定主要晶相。然而,这种方法受限于有限的成分范围和未知的加热处理参数,阻碍了高性能玻璃陶瓷的高效开发。在本研究中,我们建立了一个全面的 LAS 玻璃陶瓷小规模数据库,其中包括 751 个样品,这些样品的特征包括 27 种成分、成核温度、成核时间、结晶温度、结晶时间和 13 种晶相。我们采用了五种算法,即随机森林 (RF)、极梯度提升 (XGBoost)、分类和回归树 (CART)、K-近邻 (K-NN) 和多层感知器 (MLP) 分类器来预测潜在的结晶相。结果表明,RF 的整体性能最佳,准确率最高,为 0.8609;hamming loss 最低,为 0.0142;micro F1 分数最高,为 0.9234。这项工作推进了对 LAS 玻璃陶瓷中晶相的理解和预测,为这些材料的开发和优化提供了宝贵的见解。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Journal of Non-crystalline Solids
Journal of Non-crystalline Solids 工程技术-材料科学:硅酸盐
CiteScore
6.50
自引率
11.40%
发文量
576
审稿时长
35 days
期刊介绍: The Journal of Non-Crystalline Solids publishes review articles, research papers, and Letters to the Editor on amorphous and glassy materials, including inorganic, organic, polymeric, hybrid and metallic systems. Papers on partially glassy materials, such as glass-ceramics and glass-matrix composites, and papers involving the liquid state are also included in so far as the properties of the liquid are relevant for the formation of the solid. In all cases the papers must demonstrate both novelty and importance to the field, by way of significant advances in understanding or application of non-crystalline solids; in the case of Letters, a compelling case must also be made for expedited handling.
期刊最新文献
FTIR-ATR spectroscopic study and statistical modeling of composition-structure-property of MgO-CaO-Al2O3-SiO2 Glasses with and without Boron Improving the amorphous forming ability of FeSiBPCu nanocrystalline alloys by substituting Cu with C An integrate study of the effects of CaF2 on the viscous behavior and structure of CaO-SiO2-MgO-Al2O3-CaF2 blast-furnace slag Influence of ZnO content on liquid-liquid phase separation in photothermal refractive glass A comprehensive study on tunable structural, optical and mechanical properties of recycled windscreen glasses
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1