An Autonomous Electrochemical Test stand for Machine Learning Informed Electrolyte Optimization
An Autonomous Electrochemical Test stand for Machine Learning Informed Electrolyte Optimization
复制标题
用于机器学习通知电解质优化的自主电化学测试台
DOI：
--
复制
发表时间：
2019
影响因子：
3.9
通讯作者：
V. Viswanathan
中科院分区：
文献类型：
--
作者：
J. Whitacre;Jared M. Mitchell;Adarsh Dave;Wei Wu;S. Burke;V. Viswanathan
描述了一种全自动、计算机控制的测试台，能够快速创建和电化学表征任意液体电解质溶液。对数百种不同的电解液进行了研究，并用结果验证了系统的精密度和准确性。为了测试该方法的功能，我们快速创建并自动检查了几种含有硫酸盐和硝酸盐混合物的二维共溶剂电解液。试验台进行这些搜索只花了不到一天的时间，而传统的人工方法需要更长的时间。试验台的电导率标准误差为0.5ms/cm，电压稳定窗口测量标准误差为0.02V，其中几种组合显示出令人惊讶的高电压稳定性和电导值。该试验台在2维搜索空间中展示的成功表明，特别是在与基于机器学习的实时/在环数据评估计算程序包配合使用时，有望对液体电解质进行高速联合优化研究。
A fully automated, computer-controlled test stand capable of rapidly creating and electrochemically characterizing any arbitrary liquid electrolyte solution is described. Hundreds of different electrolytes were studied, and the results were used to verify the precision and accuracy of the system. To test the functionality of the approach, several 2-dimensional co-solvated electrolyte solutions containing blends of aqueous sulfates and nitrates were rapidly created and examined automatically. The test stand took less than a day to conduct these searches, while conventional manual methods would have taken much longer. The demonstrated standard error of the test-stand was 0.5 mS/cm on conductivity and 0.02 V for voltage stability window measurements, and several of the combinations studied revealing surprisingly high voltage stability and conductivity values. The demonstrated success of the test-stand in a 2-dimensional search spaces shows the promise of conducting high speed co-optimization studies of liquid electrolytes in particular when used in concert with a machine learning-based real time/in-loop data assessment computational package.