Heterogeneous
catalysts are the key components in industrial chemical
transformations. Metal oxides are particularly appealing as catalysts
owing to their inherent Lewis acid–base properties that facilitate
the activation of chemically inert paraffinic C–H bonds. Computational
chemistry provides a rich mechanistic understanding of catalyst functionality
through the calculation of accurate thermodynamic and kinetic data
that cannot be experimentally accessible. Using these data, one can
relate the energy needed for elementary reaction steps with properties
of the catalyst, paving the way for computational catalyst discovery.
At the heart of this process is the development of structure–activity
relationships (SARs) that facilitate the rapid prediction of promising
catalytic materials for energy intense industrial transformations,
guiding experimentation. In this review article, we highlight SARs
on oxides for chemical reactions of high industrial relevance including
(i) methane activation and conversion, (ii) alkane dehydrogenation,
and (iii) alcohol dehydration. We also discuss current limitations
and challenges on SARs and propose future steps to advance catalyst
discovery.