通过机器学习发现CO2的有机金属接口的电子海绵行为,通过机器学习进行电还原.
Haochen Shen1, Bin Jiang1, Xiaodong Yang1
1School of Chemical Engineering and Technology, Tianjin University, Tianjin, 300072, China.
本研究介绍了一种可解释的机器学习定量结构-活动关系 (ML-QSAR) 框架,以了解二氧化碳电还原中的分子调节. 它揭示了一个"电子海绵"机制,增强了多碳产品的形成.
科学领域:
- 电化学和材料科学 材料科学
- 计算化学和机器学习
背景情况:
- 在有机金属接口的分子调节是二氧化碳电还原中的C-C合的关键,影响多碳 (C2+) 产品的选择性.
- 建立预测性的定量结构-活动关系 (QSAR) 是具有挑战性的,因为复杂的分子描述器相互作用,限制了机械的理解.
研究的目的:
- 开发一种可解释的机器学习 (ML) -QSAR框架,以将分子特征与铜 (Cu) 表面的C─C合自由能量障碍 (ΔG‡) 相关联.
- 阐明控制二氧化碳电还原的占主导地位的界面"电子海绵"机制.
主要方法:
- 使用一种可解释的ML-QSAR框架,将分子描述符与Cu表面的C-C合自由能量屏障 (ΔG‡) 联系起来.
- 沙普利添加式解释 (SHAP) 分析确定了关键的电子描述:最小的局部电子亲和力 (LEAmin),HOMO-LUMO差距和HOMO能量.
- 选择和合成了一种代表性的分子 - - 3,4-米诺拉桑 (DAF).
主要成果:
- 该研究发现了一种"电子海绵"机制,其中修饰分子向Cu捐赠电子,促进中间稳定和C2+产物形成.
- 关键的电子描述符 (低LEAmin,狭窄的HOMO-LUMO差距,高的HOMO能量) 被确定为减少ΔG‡的关键.
- 用DAF进行的实验验证显示,C2+法拉第效率从42%提高到77%显著增加.
结论:
- 开发的ML-QSAR框架有效地预测了由电子海绵机制驱动的二氧化碳电还原中的分子性能.
- 描述器驱动的方法为设计高效的下一代电催化剂提供了可扩展的途径,用于Cu,Au和Ag表面上的CO2转化.
更多相关视频
09:53Quantifying the Brain Metastatic Tumor Micro-Environment using an Organ-On-A Chip 3D Model, Machine Learning, and Confocal Tomography
Published on: August 16, 2020
08:53Author Spotlight: Automated Lifespan Monitoring – Discovering Aging Dynamics with the Lifespan Machine
Published on: January 26, 2024
相关概念视频
09:34A Virtual Machine Platform for Non-Computer Professionals for Using Deep Learning to Classify Biological Sequences of Metagenomic Data
Electron Behavior
Electrons are negatively charged subatomic particles that are attracted to an orbit around the positively-charged nucleus of an atom. They reside in locations that are associated with energy levels called shells and are further organized into sub-shells and orbitals within each shell.
Electrons Orbit the Nucleus
Electrons are found in specific locations outside of the nucleus. The shell in which an electron resides indicates the general energy level of the electron: those closer to the...
Electron Behavior
Electrons Orbit the Nucleus
Electrons are found in specific locations outside of the nucleus. The shell in which an electron resides indicates the general energy level of the electron: those closer to the nucleus have less energy,...
09:53Quantifying the Brain Metastatic Tumor Micro-Environment using an Organ-On-A Chip 3D Model, Machine Learning, and Confocal Tomography
08:53High-Throughput Behavioral Aging and Lifespan Assays Using the Lifespan Machine
16:38Bridging the Bio-Electronic Interface with Biofabrication
