Beyond Ordered Materials: Understanding Catalytic Sites on Amorphous Solids

Amorphous materials are widely used as components of solid catalysts and have been the subject of much applied research. In some instances, their catalytic performance is demonstrably superior to that of their crystalline counterparts, due in part to their greater flexibility. Amorphous or disordered phases can also be generated from crystalline phases under reaction conditions, and thus, ex situ observations of long-range order may provide an incomplete or misleading picture. Until recently, theorists and experimentalists have mostly neglected these important materials in fundamental studies, preferring instead to study "well-defined" (often crystalline) catalysts that are potentially more tractable and amenable to computational modeling of their structure-activity relationships. The amorphous materials were assumed to be simply nonuniform versions of compositionally similar materials with long-range order, having the same key features at short and medium length scales. In this Perspective, shortcomings of this assumption are discussed, as well as challenges inherent in tackling amorphous catalysts more directly, namely, identifying and describing the active sites (especially under reaction conditions), discerning how subtle structural variations modulate site activity, and building atomically detailed models of amorphous catalysts. Three important classes of amorphous catalytic materials are highlighted to illustrate key issues: amorphous oxides, metal ions atomically dispersed on amorphous supports, and supported metal clusters. Amorphous and disordered silicas, aluminas, and silica-aluminas are discussed in terms of challenges and progress toward identifying how their local structural disorder and surface heterogeneity may impact the behavior of active sites. Promising models of amorphous materials with atomistic detail and increased fidelity to experiment are becoming available. However, for reactions in which small fractions of sites dominate the total activity, computational estimates of the observed kinetics will require statistical sampling methods, even for the most detailed catalyst models. Further developments in in situ and operando characterization techniques and computational modeling will advance our understanding of amorphous catalytic materials and the impact of structural disorder.