STRUCTURE-ACTIVITY RELATIONSHIPS OFHETEROGENEOUS CATALYSTS FOR THE LIQUID-PHASE H2PRODUCTION FROM NITROGEN-HYDRIDES

This work aims to study catalytic systems specifically designed to address selected structure-performance relationships for the rational design of materials in the conversion of hydrazine hydrate and ammonia borane to hydrogen. As outlined in Chapter 1, the unique properties of heterogenous catalysts are determined by their complexity, and the surface chemistry results from an ensemble of phenomena and interactions. Establishing structure-property relationships is complex, but correlating the composition of catalysts to their overall activity is the doorstep to rational improvements of the existing technologies. Indeed, understanding this complexity lies at the heart of modern research in catalysis, and requires comprehensive approaches combining experiments, characterizations and theoretical knowledge. The methods used in this work are thoroughly presented in Chapter 2. Knowledge of catalytic materials is critical to achieve the goals needed to address modern challenges, such as the transition to a zero-emission hydrogen-based economy. In this context, metals are inherently expensive and their use might raise concerns about the overall sustainability of the hydrogen produced. Hence, the choice of metal-free analogues is more and more intriguing for a zero-emissions future. Chapter 3 disclosed the role of oxygen groups of CNFs for the hydrazine hydrate decomposition. The functionalized carbons were used as catalysts and the performance compared to that of the pristine material. The surface composition was tuned through different treatments (HNO3 and H2O2) at variable times (1 and 6h), and a thorough characterization with ICP-OES, Raman and XPS unveiled the active sites for the reaction, confirmed through DFT simulations. Although the absence of metals is intriguing from an economical and ecological point of view, the direct use of hydrogen in specific applications, such as fuel cells, requires large quantities of hydrogen to operate, which can be met uniquely by the performance of noble metals. Hence, the attention focused on studying MSIs of Ir, the most active metal for the hydrazine hydrate decomposition, with the heteroatoms of carbons to enhance its performance. In Chapter 4, a low-loading of highly active Ir particles was deposited on commercial graphite and GCN, and compared to an Ir/graphite reference. The results showed enhanced stability of Ir/GCN, which was attributed to the MSIs via HR-TEM, XPS and ICP-OES, and confirmed by DFT studies on the interaction of Ir15 clusters GCN and graphite. The MSIs were proven effective in altering the performance of Ir catalysts for the hydrazine hydrate decomposition. Hence, Chapter 5 focused on a specific subset of MSIs and their synergy, metal/oxide interfaces. An Iridium-Cobalt-Cerium (IrCoCeOx) catalyst was synthesized and compared to CoCeOx, CeO2 and IrCeO2. A comprehensive characterization (XRD, XPS, XAS, DRIFTS, HR-TEM) unveiled the promotion of IrCoCeOx due to the Ir/CoO interface, which was confirmed as the active site from DFT. On these sites, the role of NaOH in enhancing H2 productivity was clarified via operando ATR-IR. In metal/oxide interfaces, the ensemble of atomic species and oxidation states deeply affect their surface and profoundly affect the catalysis, including the hydrazine hydrate conversion. However, other dehydrogenation benefits from their use, such as ammonia borane hydrolysis. PtCo3O4 catalysts represent the state-of-the-art for this reaction and, in Chapter 6, novel insights into the role of metal/oxide interfaces in this process were given. PtCo3O4 was tested and compared to Pt deposited on a non reducible (PtAl2O3) oxide, and the pristine oxide (Co3O4). The catalytic profiles were analyzed with kinetic models and the results were combined with a thorough characterization via in-situ ATR-IR, XPS, DRIFTS, XAS, XRD and HR-TEM. The strong MSIs of PtCo3O4 enhanced the reaction rate and generated the active site in-situ upon interaction with the ammonia borane. Atomic scale insights into the redox properties of the metal/oxide interface were obtained through DFT. Most of the work done in this thesis concerned the study of MSIs from an experimental and computational point of view. The computational modelling of supported catalyst is a non-trivial task, and GO methods screening a large number of configurations were required to obtain reliable results. To this end, Chapter 7 presented the Milan Cardiff Genetic Algorithm (MCGA), a modern and flexible Python framework for the parallel optimization of nanostructures, implemented for the first time and benchmarked in the frame of this thesis. The efficient and flexible mapping of complex PESs laid the foundation for robust investigations of catalytic surfaces by coupling with a well-known atomistic modeling framework, ASE, and the ability to interface with virtually any level of theory. The MCGA was extensively used to obtain all the supported cluster DFT models presented in this work. Lastly, Chapter 8 presents the main conclusions and the possible directions to continue the research of this thesis.