Bottom-up Design of MAX-phase Borides

The discovery of new materials together with advancing the performance, biocompatibility and sustainability of materials in use pose great challenges for modern technologies. The chemical makeup, structure, and properties give materials their specific flavour, which, in turn, de- termines their applicability. Most typically, engineering materials belong to the class of metals, polymers, elastomers, ceramics, or glasses. My proposal aims on designing novel materials that combine desirable properties of metals and ceramics. Metals show excellent thermal and electrical conductivity, effectively resist to crack propagation and are generally very damage tolerant, but cannot withstand high temperatures and are very dense, thus, of high weight. Ceramics, in contrast, are characterised by remarkable high-temperature strength, ultra-high melting points, and light weight, but suffer from brittleness and damage intolerance. Opening doors to unprecedented applications, nanolayerd materials called MAX-phase borides com- bine desirable properties of the two worlds. Their unrivalled diversity in crystal chemistry and bonding motifs offers an excellent basis for tuning mechanical properties and oxidation resis- tance. In the last few years, theoretical as well as experimental research of MAX-phase borides started, suggesting a great potential of these novel materials. As the field is very young, funda- mental questions remain and many outstanding atomically-laminated boride systems are yet to be discovered. In my project, the vast and rather uncharted territory of MAX-phase borides will be explored employing non-empirical as well as data-driven methods based on quantum- mechanical calculations, in combination with targeted experiments. 1

My project focused on the design and atomic-scale understanding of nanolayered boron-based materials applicable in the field of protective and wear-resistant coatings, magnetic cooling, electrocatalysis or electrochemical sensing, or radiation shielding. Particular attention was paid to systematic and efficient screening of phase stability trends, followed by in-depth case studies of elastic, deformation, and fracture properties for the most promising material systems. As a computational materials scientist, I decided to tackle this topic by combining quantum-mechanical ab initio calculations, finite temperature molecular dynamics, and machine learning, where concerning the latter two methodologies I benefited from the collaboration with colleagues at the Theoretical Physics group, Linköping University, Sweden. Experimental support has been provided via colleagues from Prof. P. H. Mayrhofer's Thin Film Group at TU Wien as well as by collaborators in Leoben and Aachen. Among main project's outcomes is high-throughput computational screening of atomically-laminated borides called MAB phases, where the searched chemical and phase space contained all combinations of the group 4-7 transition metals (M); Al, Si, Ga, Ge or In (A); and boron (B), with 10 phases prototypes for each elemental combination. Out of these, the proposed candidates for promising MABs (considering stability and mechanical performance) included Ta-Al-B, W-Al-B, Cr-Si-B, or Mn-Si-B. Furthermore, the project focused on the development of methods and workflows to study these materials, and ceramics in general, at finite temperatures, from the atomic to nano scales, and including both the thermodynamic equilibrium as well as far-from-equilibrium conditions relevant for many applications. In particular, (quantum-mechanical) ab initio molecular dynamics simulations of various mechanical tests have been performed under room and other temperatures and the underlying data has been used to train and validate machine-learning interatomic potentials for classical molecular dynamics. These molecular dynamics simulations allowed predicting mechanical response and crack resistance of MAB phases (and other reference material systems) at length scales accessible to high-resolution transmission electron miscroscopy as well as understanding of strain-activated growth of crystallographic defects and their effect (positive or negative) on the performance of the material in question. In summary, the project's results should accelerate rational design of laminated borides with optimised structure-property relationships. Additionally, large sets or accurate quantum-mechanical data has been produced, which can serve-and has already partly served-to train machine-learning models etc.