Fortunelli Alessandro, Goddard William A, Sementa Luca, Barcaro Giovanni, Negreiros Fabio R, Jaramillo-Botero Andrés
CNR-ICCOM and IPCF , Consiglio Nazionale delle Ricerche , via Giuseppe Moruzzi 1 , 56124 , Pisa , Italy . Email:
Materials and Process Simulation Center (MC 139-74) , California Institute of Technology , Pasadena , California 91125 , USA . Email:
Chem Sci. 2015 Jul 1;6(7):3915-3925. doi: 10.1039/c5sc00840a. Epub 2015 Apr 29.
Recently Debe reported that PtNi leads to extraordinary Oxygen Reduction Reaction (ORR) activity. However, several reports show that hardly any Ni remains in the layers of the catalysts close to the surface ("Pt-skin effect"). This paradox that Ni is essential to the high catalytic activity with the peak ORR activity at PtNi while little or no Ni remains close to the surface is explained here using large-scale first-principles-based simulations. We make the radical assumption that processing Pt-Ni catalysts under ORR conditions would leach out all Ni accessible to the solvent. To simulate this process we use the ReaxFF reactive force field, starting with random alloy particles ranging from 50% Ni to 90% Ni and containing up to ∼300 000 atoms, deleting the Ni atoms, and equilibrating the resulting structures. We find that the PtNi case and a final particle radius around 7.5 nm lead to internal voids in communication with the exterior, doubling the external surface footprint, in fair agreement with experiment. Then we examine the surface character of these nanoporous systems and find that a prominent feature in the surface of the de-alloyed particles is a rhombic structure involving 4 surface atoms which is crystalline-like but under-coordinated. Using density-functional theory, we calculate the energy barriers of ORR steps on Pt nanoporous catalysts, focusing on the O-hydration reaction (O + HO → OH + OH) but including the barriers of O dissociation (O → O + O) and water formation (OH + H → HO). We find that the reaction barrier for the O-hydration rate-determining-step is reduced significantly on the de-alloyed surface sites compared to Pt(111). Moreover we find that these active sites are prevalent on the surface of particles de-alloyed from a Pt-Ni 30 : 70 initial composition. These simulations explain the peak in surface reactivity at PtNi, and provide a rational guide to use for further optimization of improved catalytic and nanoporous materials.
最近，德贝报道称铂镍合金具有非凡的氧还原反应（ORR）活性。然而，几份报告显示，在靠近表面的催化剂层中几乎没有镍残留（“铂皮效应”）。本文利用基于第一性原理的大规模模拟解释了这一矛盾现象：镍对于铂镍合金具有高催化活性至关重要，其ORR活性峰值出现在铂镍合金中，但靠近表面的镍却很少或几乎没有残留。我们做出了一个大胆的假设，即在ORR条件下处理铂镍催化剂会使溶剂可接触到的所有镍都浸出。为了模拟这个过程，我们使用ReaxFF反应力场，从含镍量为50%至90%、原子数多达约300000个的随机合金颗粒开始，删除镍原子，并使所得结构达到平衡。我们发现，铂镍合金的情况以及最终约7.5纳米的颗粒半径会导致内部空隙与外部连通，使外部表面积增加一倍，这与实验结果相当吻合。然后我们研究了这些纳米多孔体系的表面特性，发现脱合金颗粒表面的一个显著特征是一种由4个表面原子构成的菱形结构，它类似晶体但配位不足。利用密度泛函理论，我们计算了铂纳米多孔催化剂上ORR步骤的能垒，重点关注氧水合反应（O + HO → OH + OH），但也包括氧解离（O → O + O）和水形成（OH + H → HO）的能垒。我们发现，与铂（111）相比，脱合金表面位点上氧水合速率决定步骤的反应能垒显著降低。此外，我们发现这些活性位点在初始组成为铂 - 镍30 : 70的脱合金颗粒表面普遍存在。这些模拟解释了铂镍合金表面反应活性的峰值，并为进一步优化改进催化和纳米多孔材料提供了合理的指导。
