Machine-learning-assisted materials discovery using failed experiments
Machine-learning-assisted materials discovery using failed experiments
复制标题
DOI：
10.1038/nature17439
复制
发表时间：
2016-05-05
期刊：
影响因子：
64.8
通讯作者：
Norquist, Alexander J.
中科院分区：
文献类型：
--
作者：
Raccuglia, Paul;Elbert, Katherine C.;Norquist, Alexander J.
无机-有机杂化材料（1-3）如有机模板化的金属氧化物（1）、金属-有机骨架（MOF）（2）和有机卤化物钙钛矿（4）已经研究了几十年，并且水热和（非水）溶剂热合成已经产生了数千种新材料，其共同包含元素周期表中的几乎所有金属（5-9）。然而，这些化合物的形成并不完全清楚，新化合物的开发主要依赖于探索性合成。模拟和数据驱动的方法（由材料基因组计划（10）等努力推动）提供了实验试错的替代方案。三大战略是：基于模拟的物理性质预测（例如，电荷迁移率（11）、光伏特性（12）、气体吸附能力（13）或锂离子嵌入（14）），以鉴定用于合成工作的有希望的目标候选物（11，15）;从大量实验数据中确定结构-性能关系（16，17），通过与高通量合成和测量工具（18）的集成而实现;以及基于相似的晶体结构（例如，沸石结构分类（19，20）或气体吸附性质（21））的聚类。在这里，我们展示了一种替代方法，该方法使用在反应数据上训练的机器学习算法来预测模板化钒亚硒酸盐结晶的反应结果。我们使用的信息“黑暗”的反应失败或不成功的水热合成，从我们的实验室存档的实验室笔记本收集，并添加物理化学性质的描述，以原始笔记本信息，使用化学信息学技术。我们使用由此产生的数据来训练机器学习模型，以预测反应的成功。当使用以前未经测试的商业可用有机构建块进行水热合成实验时，我们的机器学习模型优于传统的人类策略，并成功预测了新的有机模板无机产物形成的条件，成功率为89%。反转机器学习模型揭示了关于成功产物形成条件的新假设。
Inorganic-organic hybrid materials(1-3) such as organically templated metal oxides(1), metal-organic frameworks (MOFs)(2) and organohalide perovskites(4) have been studied for decades, and hydrothermal and (non-aqueous) solvothermal syntheses have produced thousands of new materials that collectively contain nearly all the metals in the periodic table(5-9). Nevertheless, the formation of these compounds is not fully understood, and development of new compounds relies primarily on exploratory syntheses. Simulation-and data-driven approaches (promoted by efforts such as the Materials Genome Initiative(10)) provide an alternative to experimental trial-and-error. Three major strategies are: simulation-based predictions of physical properties (for example, charge mobility(11), photovoltaic properties(12), gas adsorption capacity(13) or lithium-ion intercalation(14)) to identify promising target candidates for synthetic efforts(11,15); determination of the structure-property relationship from large bodies of experimental data(16,17), enabled by integration with high-throughput synthesis and measurement tools(18); and clustering on the basis of similar crystallographic structure (for example, zeolite structure classification(19,20) or gas adsorption properties(21)). Here we demonstrate an alternative approach that uses machine-learning algorithms trained on reaction data to predict reaction outcomes for the crystallization of templated vanadium selenites. We used information on 'dark' reactions-failed or unsuccessful hydrothermal syntheses-collected from archived laboratory notebooks from our laboratory, and added physicochemical property descriptions to the raw notebook information using cheminformatics techniques. We used the resulting data to train a machine-learning model to predict reaction success. When carrying out hydrothermal synthesis experiments using previously untested, commercially available organic building blocks, our machine-learning model outperformed traditional human strategies, and successfully predicted conditions for new organically templated inorganic product formation with a success rate of 89 per cent. Inverting the machine-learning model reveals new hypotheses regarding the conditions for successful product formation.