Discovering Intermetallics through Synthesis, Computation, and Data-Driven Analysis



Journal Title

Journal ISSN

Volume Title



Intermetallics adopt an array of crystal structures, boast diverse chemical compositions, and possess exotic physical properties that have led to a wide range of applications from the biomedical to aerospace industries. Despite a long history of intermetallic synthesis and crystal structure analysis, identifying new intermetallic phases has remained challenging due to the prolonged nature of experimental phase space searching or the need for fortuitous discovery. New approaches with a specific focus on realizing novel intermetallics have been proposed that expand on traditional methods for materials synthesis and characterization. One of the most notable methods is to merge traditional intermetallic synthesis and characterization with computation and materials informatics, when combined, provide a new set of tools capable of advancing the discovery of metal-rich solids. Each chapter of this dissertation employs solid-state synthesis, first-principle calculations, and machine learning to modernize how intermetallics are discovered and to better understand their complex structures. For example, we combined exploratory solidstate synthesis with ab initio calculations to investigate gold’s polyanionic bonding in intermetallic phases. The application of density functional theory goes beyond merely studying the electronic structure and chemical bonding of intermetallics. We also utilized an ab initio approach coupled with a structure-search algorithm (CALYPSO) to predict the crystal structure of intermetallics under pressure. Our research revealed the existence of two binaries in the A-Ir (A = Rb, Cs) system over ~10 GPa. Further, a new approach merging experimental work, computation, and data-driven analysis to discover intermetallics was created. We developed a machine-learning model to predict the formation energy of metal-rich solid based only on the constituent elements followed by experimental validation through the subsequent synthesis of a novel compound, YAg0.65In1.35. Finally, exploratory synthesis was carried out on the ternary RE-Au-Ge (RE = La, Ce, Pr, Nd) phase system leading to the discovery of six novel compounds: La3Au3Ge, La2Au2Ge, and RE2AuGe3 (RE = La, Ce, Pr, Nd). The research results presented in this dissertation discuss the opportunities and challenges in the discovery of new intermetallic phases while new approaches that merge synthesis, computation, and data science to accelerate the realization of metal-rich materials are created.



Intermetallics, Solid-state synthesis, DFT computation, Machine learning


Portions of this document appear in: Lotfi, Sogol, and Jakoah Brgoch. "Discovering intermetallics through synthesis, computation, and data‐driven analysis." Chemistry–A European Journal (2020). And in: Lotfi, Sogol, Anton O. Oliynyk, and Jakoah Brgoch. "Polyanionic gold–tin bonding and crystal structure preference in REAu1. 5Sn0. 5 (RE= La, Ce, Pr, Nd)." Inorganic chemistry 57, no. 17 (2018): 10736-10743. And in: Lotfi, Sogol, and Jakoah Brgoch. "Predicting pressure-stabilized alkali metal iridides: A− Ir (A= Rb, Cs)." Computational Materials Science 158 (2019): 124-129. And in: Lotfi, Sogol, Ziyan Zhang, Gayatri Viswanathan, Kaitlyn Fortenberry, Aria Mansouri Tehrani, Jakoah Brgoch, Covalent radius Polarizability, and Crystal radius Density. "Targeting productive composition space through machine-learning directed inorganic synthesis." (2020).