化学
选择性
电解
波形
新戊酸
产量(工程)
选择(遗传算法)
基础(拓扑)
交流电
组合化学
电化学
贝叶斯概率
价值(数学)
建设性的
电流(流体)
计算化学
直流电
反应机理
管道(软件)
电解质
先验与后验
动作(物理)
失真(音乐)
校准
作者
Eric S. Larsen,Louis P. J.‐L. Gadina,Yaroslav I. Sobolev,Jonathan Sabaté del Río,Rafał Frydrych,Kisung Lee,Juhyeong Lee,Bartosz A. Grzybowski
摘要
Abstract This paper reassesses the selectivity, efficiency, and mechanism of sequential electroreductions of carbonyl compounds under alternating current (AC) versus direct current (DC) conditions. It reflects the constructive dialogue─through multiple rounds of revisions of the present work─between the Baran group (authors of the original report) and the Grzybowski group (who sought to reproduce the reported phenomena and develop theoretical models). Ultimately, consensus was reached that under the originally described conditions, AC offers no systematic yield or selectivity advantage over DC; side reactions are outcompeted by mass action rather than kinetically suppressed by AC waveforms, while DC is faster and more energy efficient (less F mol–1). Higher AC yields and selectivities can be reached when a small amount of base is added to the sacrificial pivalic acid reductant. This improvement, documented in Baran’s Corrigendum, was confirmed in our own reproductions. However, the addition of base fundamentally alters the originally proposed mechanism: the observed behavior can no longer be attributed to frequency or waveform effects imparted by AC forcing, but is instead explained by mass-action considerations and pH conditions at the electrodes. Moreover, with base present, AC yields and selectivities can still be matched or even surpassed under appropriately optimized DC conditions. These findings (i) clarify the mechanistic origins of the reported phenomena, (ii) establish a transparent framework for choosing between AC and DC electrolysis in future synthetic applications, and (iii) showcase the value of using Machine Learning techniques (here reaction selection aided by Bayesian Optimization) in optimizing electrochemical processes.
科研通智能强力驱动
Strongly Powered by AbleSci AI