Christopher C. Price, Yansong Li, Guanyu Zhou, Rehan Younas, Spencer S. Zeng, Tim H. Scanlon, Jason M. Munro, Christopher L. Hinkle
{"title":"利用机器学习功能化预测和加速纳米材料合成","authors":"Christopher C. Price, Yansong Li, Guanyu Zhou, Rehan Younas, Spencer S. Zeng, Tim H. Scanlon, Jason M. Munro, Christopher L. Hinkle","doi":"arxiv-2409.08054","DOIUrl":null,"url":null,"abstract":"Solving for the complex conditions of materials synthesis and processing\nrequires analyzing information gathered from multiple modes of\ncharacterization. Currently, quantitative information is extracted serially\nwith manual tools and intuition, constraining the feedback cycle for process\noptimization. We use machine learning to automate and generalize feature\nextraction for in-situ reflection high-energy electron diffraction (RHEED) data\nto establish quantitatively predictive relationships in small sets ($\\sim$10)\nof expert-labeled data, and apply these to save significant time on subsequent\nepitaxially grown samples. The fidelity of these relationships is tested on a\nrepresentative material system ($W_{1-x}V_xSe2$ growth on c-plane sapphire\nsubstrate (0001)) at two stages of synthesis with two aims: 1) predicting the\ngrain alignment of the deposited film from the pre-growth substrate surface\ndata, and 2) estimating the vanadium (V) dopant concentration using in-situ\nRHEED as a proxy for ex-situ methods (e.g. x-ray photoelectron spectroscopy).\nBoth tasks are accomplished using the same set of materials agnostic core\nfeatures, eliminating the need to retrain for specific systems and leading to a\npotential 80\\% time saving over a 100 sample synthesis campaign. These\npredictions provide guidance for recipe adjustments to avoid doomed trials,\nreduce follow-on characterization, and improve control resolution for materials\nsynthesis, ultimately accelerating materials discovery and commercial scale-up.","PeriodicalId":501234,"journal":{"name":"arXiv - PHYS - Materials Science","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Predicting and Accelerating Nanomaterials Synthesis Using Machine Learning Featurization\",\"authors\":\"Christopher C. Price, Yansong Li, Guanyu Zhou, Rehan Younas, Spencer S. Zeng, Tim H. Scanlon, Jason M. Munro, Christopher L. Hinkle\",\"doi\":\"arxiv-2409.08054\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Solving for the complex conditions of materials synthesis and processing\\nrequires analyzing information gathered from multiple modes of\\ncharacterization. Currently, quantitative information is extracted serially\\nwith manual tools and intuition, constraining the feedback cycle for process\\noptimization. We use machine learning to automate and generalize feature\\nextraction for in-situ reflection high-energy electron diffraction (RHEED) data\\nto establish quantitatively predictive relationships in small sets ($\\\\sim$10)\\nof expert-labeled data, and apply these to save significant time on subsequent\\nepitaxially grown samples. The fidelity of these relationships is tested on a\\nrepresentative material system ($W_{1-x}V_xSe2$ growth on c-plane sapphire\\nsubstrate (0001)) at two stages of synthesis with two aims: 1) predicting the\\ngrain alignment of the deposited film from the pre-growth substrate surface\\ndata, and 2) estimating the vanadium (V) dopant concentration using in-situ\\nRHEED as a proxy for ex-situ methods (e.g. x-ray photoelectron spectroscopy).\\nBoth tasks are accomplished using the same set of materials agnostic core\\nfeatures, eliminating the need to retrain for specific systems and leading to a\\npotential 80\\\\% time saving over a 100 sample synthesis campaign. These\\npredictions provide guidance for recipe adjustments to avoid doomed trials,\\nreduce follow-on characterization, and improve control resolution for materials\\nsynthesis, ultimately accelerating materials discovery and commercial scale-up.\",\"PeriodicalId\":501234,\"journal\":{\"name\":\"arXiv - PHYS - Materials Science\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-12\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Materials Science\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.08054\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Materials Science","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.08054","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
要解决材料合成和加工的复杂条件,就必须分析从多种表征模式中收集到的信息。目前,定量信息是通过手动工具和直觉连续提取的,这限制了工艺优化的反馈周期。我们利用机器学习来自动和通用原位反射高能电子衍射(RHEED)数据的特征提取,在专家标记的小型数据集中($\sim$10)建立定量预测关系,并将这些关系应用于后续片晶生长样品,从而节省大量时间。我们在两个合成阶段的代表性材料系统(在 c 平面蓝宝石衬底 (0001) 上生长的 $W_{1-x}V_xSe2$)上测试了这些关系的保真度,目的有两个:1) 根据生长前基底表面数据预测沉积薄膜的晶粒排列,以及 2) 使用原位 RHEED 代替原位方法(例如 x 射线光电子能谱)估算钒(V)掺杂浓度。这两项任务都是使用同一套材料无关核心特征完成的,无需针对特定系统进行重新训练,从而在 100 个样品的合成过程中节省 80% 的时间。这些预测为配方调整提供了指导,以避免注定失败的试验,减少后续表征,提高材料合成的控制分辨率,最终加速材料发现和商业放大。
Predicting and Accelerating Nanomaterials Synthesis Using Machine Learning Featurization
Solving for the complex conditions of materials synthesis and processing
requires analyzing information gathered from multiple modes of
characterization. Currently, quantitative information is extracted serially
with manual tools and intuition, constraining the feedback cycle for process
optimization. We use machine learning to automate and generalize feature
extraction for in-situ reflection high-energy electron diffraction (RHEED) data
to establish quantitatively predictive relationships in small sets ($\sim$10)
of expert-labeled data, and apply these to save significant time on subsequent
epitaxially grown samples. The fidelity of these relationships is tested on a
representative material system ($W_{1-x}V_xSe2$ growth on c-plane sapphire
substrate (0001)) at two stages of synthesis with two aims: 1) predicting the
grain alignment of the deposited film from the pre-growth substrate surface
data, and 2) estimating the vanadium (V) dopant concentration using in-situ
RHEED as a proxy for ex-situ methods (e.g. x-ray photoelectron spectroscopy).
Both tasks are accomplished using the same set of materials agnostic core
features, eliminating the need to retrain for specific systems and leading to a
potential 80\% time saving over a 100 sample synthesis campaign. These
predictions provide guidance for recipe adjustments to avoid doomed trials,
reduce follow-on characterization, and improve control resolution for materials
synthesis, ultimately accelerating materials discovery and commercial scale-up.