The hydrogen economy stands as one of the most promising alternatives to generate green energy and slow down carbonization. This technology is based on two main processes: the generation of hydrogen (limited by the HER and the OER) in electrolyzer cells, and the production of energy (limited by the ORR and the HOR) in fuel cells. These reactions need
the use of a heterogeneous catalyst, usually a precious metal, such as platinum. The high cost and the low availability of this material hinder the widespread use of this technology, boosting the search for new catalysts and techniques to improve the catalytic activity of these reactions. Four techniques are commonly used to tune the active surface of the socalled heterogeneous catalysts: the introduction of different facets, alloying, the addition of defects, and the application of elastic strains.

Within this framework, this thesis analyzes, from a computational point of view, the catalytic activity of pure metals and multimetallic materials in two of the main reactions in the Hydrogen Economy: the Hydrogen Evolution Reaction (HER) and the Oxygen Reduction Reaction (ORR). Additionally, the effect of elastic strains is also considered. To this end, Density Functional Theory calculations were performed for different pure metals, binary, and ternary intermetallic compounds in order to calculate the adsorption energy of the three main adsorbates present in the reactions: H, O, and OH, with and without
the application of elastic strains. The effect of different types of strains -biaxial, uniaxial, and shear- at different adsorption positions was analyzed in depth. These calculated adsorption energies were used as estimators of the catalytic activity, following the BellEvans-Polanyi principle [53] and the d-band theory proposed by J. K. Norskov [92, 94, 182].

Complimentary, all the DFT calculations with and without the application of strains for pure metals and intermetallic compounds were used to generate a Machine Learning database and train a Random Forest Regressor Model that was able to predict the hydrogen, oxygen, and hydroxyl adsorption energies of many multimetallic materials. The obtained predictions for binary intermetallic compounds had very low errors in the case of H (MAE=0.07eV) and O/OH (MAE=0.18 eV) models. The model was able to predict the effect of elastic strains in the adsorption energy and was used to propose new intermetallic compounds that can replace noble metals for the HER and the ORR.

This work represents a high-throughput computational strategy to discover new catalysts for the Hydrogen Economy including Elastic Strain Engineering.