化学
甲醇
催化作用
格式化
甲酸甲酯
氢
反应机理
二聚体
转移加氢
组合化学
有机化学
钌
作者
Ya‐Fan Zhao,Yong Yang,Charles A. Mims,Charles H. F. Peden,Jun Li,Donghai Mei
标识
DOI:10.1016/j.jcat.2011.04.012
摘要
Methanol synthesis from CO2 hydrogenation on supported Cu catalysts is of considerable importance in the chemical and energy industries. Although extensive experimental and theoretical efforts have been carried out in the past decades, the most fundamental questions such as the reaction mechanisms and the key reaction intermediates are still in debate. In the present work, a comprehensive reaction network for CO2 hydrogenation to methanol on Cu(1 1 1) is studied using periodic density functional theory calculations. All of the elementary reaction steps in the reaction network are identified in an unbiased way with the dimer method. Our calculation results show that methanol synthesis from direct hydrogenation of formate on Cu(1 1 1) is not feasible due to the high activation barriers for some of the elementary steps. Instead, we find that CO2 hydrogenation to hydrocarboxyl (trans-COOH) is kinetically more favorable than formate in the presence of H2O via a unique hydrogen transfer mechanism. The trans-COOH is then converted into hydroxymethylidyne (COH) via dihydroxycarbene (COHOH) intermediates, followed by three consecutive hydrogenation steps to form hydroxymethylene (HCOH), hydroxymethyl (H2COH), and methanol. This is consistent with recent experimental observations [1], which indicate that direct hydrogenation of formate will not produce methanol under dry hydrogen conditions. Thus, both experiment and computational modeling clearly demonstrate the important role of trace amounts of water in methanol synthesis from CO2 hydrogenation on Cu catalysts. The proposed methanol synthesis route on Cu(1 1 1) not only provides new insights into methanol synthesis chemistry, but also demonstrates again that spectroscopically observed surface species are often not critical reaction intermediates but rather spectator species.
科研通智能强力驱动
Strongly Powered by AbleSci AI