Automated Electrochemical Research based on Deep Learning

在化学系化学催化（CAT）和化学结构、动力学和机理- b （CSDM-B）项目的支持下，加州大学洛杉矶分校的刘冲和顾泉泉以及加州大学欧文分校的Jenny Y. Yang组成的合作团队正在努力建立一个需要最少人为干预的电化学研究自动化平台。该项目的成功完成将展示自主电化学研究的可行性、注意事项和力量，并有望在如何进行电化学和电催化的科学研究调查方面实现范式转变。该项目还引入了在跨学科研究环境中培训具有不同技能的下一代研究人员和劳动力的机会。所开发的软件和方法将向公众免费提供，并纳入以电化学为重点的教育新兵训练营。将建立一个电化学、自动化和人工智能的新兵训练营，面向本科生、研究生和博士后，特别是那些来自社会经济上代表性不足的群体的参与者。这个训练营将吸引高级本科生和研究生以及博士后学者，将促进跨学科，并将帮助建立一支精通人工智能的化学劳动力队伍。根据该奖项，加州大学洛杉矶分校的刘充和谷泉泉以及加州大学欧文分校的Jenny Y. Yang的三方合作正在建立一个概念验证平台，以自主进行高通量的电化学研究，至少在一定程度上补充（如果不是取代）人工过程。该团队将开发基于深度学习的算法，自动分析电化学数据，并构建电化学中质子耦合电子转移机理研究的实验平台。具体而言，本提案的目的是：(1)开发基于深度学习的自动算法，自动分析循环伏安图作为电化学数据的经典示例；(2)基于小组对深度学习算法和贝叶斯优化的理解，构建自动化电化学测试和迭代设计新实验的自主实验平台；(3)利用该平台开展电化学中质子耦合电子转移（PCET）的机理研究，发现CO2固定均相电催化的新反应。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

With support from the Chemical Catalysis (CAT) and Chemical Structure, Dynamics, and Mechanisms-B (CSDM-B) programs in the Division of Chemistry, the collaborative team of Chong Liu and Quanquan Gu of the University of California, Los Angeles and Jenny Y. Yang of the University of California, Irvine is working to establish an electrochemical research automation platform that requires minimal human intervention. Successful completion of this project will showcase the feasibility, caveats, and power of autonomous electrochemical research and promise a paradigm shift in how scientific research investigation in electrochemistry and electrocatalysis will be conducted. The project also introduces the opportunity of training the next-generation researchers and workforce with diverse skill sets in an interdisciplinary research environment. The software and methodology developed will be made publicly accessible free of charge and incorporated into an educational boot camp focused on electrochemistry. A boot camp on electrochemistry, automation, and artificial intelligence for undergraduate, graduate, and postdoctoral participants, particularly those from socio-economically underrepresented groups, will be established. This boot camp will engage senior-level undergraduate and graduate students, as well as postdoctoral scholars, will foster interdisciplinarity and will help to build an AI-savvy chemistry workforce.Under this award, the tripartite collaborative of Chong Liu and Quanquan Gu of the University of California, Los Angeles and Jenny Y. Yang of the University of California, Irvine are establishing a proof-of-concept platform to autonomously conduct electrochemistry research with high throughput and at least partly supplement, if not replace, the manual process. The team will develop algorithms based on deep learning to automatically analyze electrochemical data and construct an experimentation platform for mechanistic studies of proton-coupled electron transfer in electrochemistry. Specifically, the aims of this proposal are: (1) to develop automatic algorithms based on deep learning that automatically analyze cyclic voltammograms as a classic example of electrochemical data; (2) to construct an autonomous experimentation platform that automates electrochemical testing and iteratively designs new experiments based on the group's understanding of the deep-learning algorithm and Bayesian optimization and (3) to employ the established platform to conduct mechanistic studies of proton-coupled electron transfer (PCET) in electrochemistry and discover new reactivities in homogenous electrocatalysis of CO2 fixation.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.