Recommender System of Successful Processing Conditions for New Compounds Based on a Parallel Experimental Data Set
Recommender System of Successful Processing Conditions for New Compounds Based on a Parallel Experimental Data Set
复制标题
DOI：
10.1021/acs.chemmater.9b01799
复制
发表时间：
2019-12-24
影响因子：
8.6
通讯作者：
Tanaka, Isao
中科院分区：
文献类型：
--
作者：
Hayashi, Hiroyuki;Hayashi, Katsuyuki;Tanaka, Isao
我们提出了一种机器学习方法，在平行实验的基础上为新化合物推荐成功的加工条件。首先，采用23种起始材料和23种阳离子混合比，对注册在无机晶体结构数据库（ICSD）中的67种伪二元氧化物进行平行实验，建立了实验数据库。采用固相反应、聚合配合物、环醚溶胶-凝胶和喷雾共沉淀法四种合成方法，在五种不同温度下烧制前驱体粉体。这产生了1648个独特的化学合成条件和数据库条目。使用带有自动样品交换器的粉末x射线衍射设备对反应物进行了顺序表征。综合结果以一个分数的形式进行评分，将其置于一个包含243 340个元素的五阶张量中。塔克分解法用于预测未实验处理条件下的评分。交叉验证表明该模型具有良好的预测性能。通过检查另一个数据库ICDD-PDF（国际衍射数据-粉末衍射文件中心）中高评价成分的存在，进一步对其进行了评估。对未实验成分的成功加工条件进行了很好的推荐。
We propose a machine-learning method to recommend successful processing conditions for new compounds on the basis of parallel experiments. Initially, an experimental database was constructed for 67 pseudobinary oxides registered in the Inorganic Crystal Structure Database (ICSD) by parallel experiments using 23 starting materials and 23 cation mixing ratios. Precursor powders were obtained by four synthesis methods (solid-state reaction, polymerized complex, cyclic ether sol-gel, and spray coprecipitation), which were fired at five different temperatures. This resulted in 1648 unique chemical synthesis conditions and database entries. The reactants were characterized sequentially using powder X-ray diffraction equipment with an automatic sample exchanger. The synthesis results were rated as a score, which was placed into a fifth-order tensor with 243 340 elements. The Tucker decomposition method was used to predict yet-to-be-rated scores for unexperimented processing conditions. Good predictive performance of the present model was demonstrated by cross validation. It was further evaluated by examining the presence of highly rated compositions in another database, ICDD-PDF (International Center for Diffraction Data-Powder Diffraction File). Successful processing conditions for unexperimented compositions were found to be well recommended.