催化作用
四面体
材料科学
拉伤
电化学
电化学能量转换
电极
反应性(心理学)
氢
化学物理
燃料电池
化学工程
纳米技术
还原(数学)
功率密度
化学
能量转换
应变工程
氧气
粒子(生态学)
氢燃料
压缩(物理)
可逆氢电极
化学能
电催化剂
多相催化
膜
功率(物理)
密度泛函理论
质子交换膜燃料电池
氧化还原
储能
析氧
分解水
微生物燃料电池
作者
Dafu Zhao,Dafu Zhao,Zisheng Tang,Jinfeng Liu,Zhiyi Hu,Zhiwen Yin,Jieheng Lv,Xiaobin Liao,Xiaoqian Wang,Yingfei Liu,Yingfei Liu,Damin Liu,Lihua Chen,Bao-Lian Su,Dongyuan Zhao,Dongyuan Zhao,Yong Liu,Yong Liu
标识
DOI:10.1038/s41467-026-75485-9
摘要
Practical electrochemical energy conversion requires electrocatalysts that coordinate multiple elementary steps at spatially distinct active sites, yet atomic-level control of such site-specific reactivity within a single heterogeneous particle remains challenging. Here we propose and realize atomic-scale gradient strain as a design concept for heterogeneous electrocatalysis. Using Pd@Pt core-shell tetrahedra as a model system, we construct a continuous lattice-strain gradient across individual 3–4 atomic-layer Pt {111} epitaxial shells, where lattice-mismatch-driven compression relaxes from edges to center regions. This single-particle strain gradient, spanning approximately −8% to −2%, spatially links *O2 activation at highly compressed sites with *OH weakening at moderately compressed sites through kinetically accessible intermediate redistribution. The catalysts exhibit competitive oxygen reduction reaction performance, with mass and specific activities of 2.19 A mgPt⁻1 and 3.01 mA cm⁻2 at 0.9 V vs reversible hydrogen electrode, while retaining 91% activity after 20 k cycles. In membrane electrode assemblies, they achieve 0.57 A mgPt⁻1, and peak power densities of 2.10 W cm⁻2 in H2 − O2 and 1.16 W cm⁻2 in H2−air, with over 90% performance retention after 20 k cycles. Strain engineering tunes catalyst reactivity, but single strain cannot optimize all steps in oxygen reduction concurrently. Here, the authors report strain gradients within individual Pd@Pt core-shell tetrahedra that link distinct local reaction sites, improving fuel-cell activity and durability.
科研通智能强力驱动
Strongly Powered by AbleSci AI