催化作用
材料科学
氢
氧化物
化学工程
甲醇
制氢
铟
烧结
氢溢流
多相催化
金属
无机化学
碳纳米管
纳米颗粒
纳米技术
降级(电信)
可逆氢电极
氧化铈
碳纤维
兴奋剂
催化剂载体
作者
Lei Zhang,Mingyi Wu,Hua An,Keyi Zhou,Bingqing Yao,Qian He,Kyungho Lee,Xi Chen,Ning Yan
标识
DOI:10.1021/acscatal.6c06714
摘要
Abstract Metal promotion is a widely adopted strategy to enhance the activity of indium oxide (In2O3) catalysts for CO2 hydrogenation to methanol. However, intimately contacted metal promoters on In2O3 often result in rapid catalyst deactivation, as excessive local H2 supply induces undesirable structural degradation of the In2O3 domain. Here, we address this limitation by implementing a carbon nanotube (CNT)-mediated hydrogen transport pathway, which spatially separates the H2 activation sites (Pd) from the CO2 hydrogenation sites (In2O3) at the catalyst preparation stage. Intriguingly, upon reductive activation, the initially separated components spontaneously evolve into a distinctive interfacial configuration. In this configuration, InOx migrates onto and partially covers the Pd nanoparticles, forming an InOx@Pd inverse interface. Unlike the directly constructed Pd–In2O3 interface, which suffers from over-reduction to metallic In0 accompanied by sintering and alloying with Pd, this CNT-derived interface preserves the In3+ oxidation state and serves as a non-destructive junction for hydrogen delivery. Combined with long-range hydrogen spillover across the CNT network, this architecture transforms a concentrated and destructive hydrogen flux into a regulated supply that reaches distal In2O3 domains without driving over-reduction. Consequently, the Pd/CNT + In2O3 system exhibits both improved catalytic activity and enhanced long-term stability (>230 h), whereas the conventional Pd/In2O3 catalyst deactivates continuously. These findings demonstrate that a controlled hydrogen delivery system between H2 activation sites and active sites on the oxide domain offers an effective way for developing active and durable catalysts involving reduction-sensitive oxide phases.
科研通智能强力驱动
Strongly Powered by AbleSci AI