Ab initio description of non-adiabatic effects in dissociative adsorption at surfaces

基于微观理解的预测材料科学建模需要对原子尺度上的所有基本过程有透彻的了解。单个气相分子在金属表面的（解离）吸附是这样一个基本的过程，对于任何涉及暴露于实际气体环境的表面的应用都是至关重要的，多相催化只是一个突出的例子。不幸的是，我们目前对这一基本过程的理解，即使是对最简单、但无处不在的双原子分子，比如02，也非常肤浅。由于缺乏关于撞击分子与表面电子结构（以所谓的势能面（PES）的形式总结）相互作用的可靠和确定的信息，经验理论援引有争议的情景来解释实验上可获得的积分动力学量，如表面粘着概率。在这些情况中，经常讨论的是非绝热效应在离解中的作用，即可能违反玻恩-奥本海默近似和电子分布不能瞬间跟随核运动。对这一问题的澄清只能来自第一性原理理论，如密度泛函理论（DFT），它涉及对系统电子结构的可靠而近似的描述。这些理论完全基于绝热的DFT-PES，过去主要是针对H2在金属表面的解离而发展起来的，其中非绝热效应显然不起作用。最近的方法发展现在也使在受限DFT方法中计算绝热PESs成为可能。相应地，我们提出了第一个系统的调查，解决在02离解的非绝热效应的真正重要性。为了明确地确定这些效应所起的作用，这需要同时解决许多其他尚未完全理解的问题，如描述电子相关性或表面迁移率影响所需的准确性。关注于具有代表性的一组表面上的02解离动力学，期期表上的新趋势图有望有助于首次全面了解这一最重要分子的（解离）吸附，这也可能为其他广泛存在的双原子如N2， NO或CO的解离提供一些指导。

A predictive materials science modeling based on microscopic understanding requires a thorough knowledge of all underlying elementary processes at the atomic scale. The (dissociative) adsorption of individual gas phase molecules at metal surfaces is such an elementary process that is of crucial relevance for any application involving surfaces exposed to realistic gas environments, with heterogeneous catalysis forming just one prominent example. Unfortunately, our present understanding of this fundamental process is even for a most simple, but ubiquitous diatomic molecule like 02 very shallow. Lacking reliable and resolved information about the interaction of an impinging molecule with the surface electronic structure (summarized in form of a socalled potential-energy surface (PES)), empirical theories have invoked controversial scenarios to account for experimentally accessible, integral kinetic quantities like the sticking probability at the surface. Among these scenarios is the frequently discussed role of non-adiabatic effects in the dissociation, i.e. a possible violation of the Born-Oppenheimer approximation and electron distributions that are not able to follow the nuclear motion instantaneously. Clarifying insight into this matter can only come from firstprinciples theories like density-functional theory (DFT), which involve a reliable, though approximate description of the electronic structure of the system. Based exclusively on the adiabatic DFT-PES, such theories have in the past been predominantly developed for H2 dissociation at metal surfaces, where non-adiabatic effects apparently play no role. Recent methodological developments have now also enabled the calculation of diabatic PESs within constrained DFT approaches. Correspondingly, we propose a first systematic investigation addressing the real importance of non-adiabatic effects in 02 dissociation. To unambiguously pin down the role played by these effects, this requires to concomitantly tackle a number of other, not yet fully understood issues like the required accuracy in the description of electron correlation or the influence of surface mobility. Focusing at the dynamics of 02 dissociation at a representative set of surfaces, the emerging trend picture over the periodic table is expected to contribute to a first comprehensive understanding of the (dissociative) adsorption of this most important molecule, that may also give some guidance on the dissociation of other widespread diatomics like N2, NO or CO.