Data mining of transition metal borides, nitrides and carbides

NAISS 2024/6-154


NAISS Medium Storage

Principal Investigator:

Igor Dimarco


Uppsala universitet

Start Date:


End Date:


Primary Classification:

10304: Condensed Matter Physics

Secondary Classification:

10403: Materials Chemistry

Tertiary Classification:

20501: Ceramics




Transition metal ceramics are relatively easy to synthesize and are suitable for a variety of applications, due to their limited costs. With the rise of databases and machine learning, computational research on ceramics has become very important for materials design and optimization. In the last few years, the PI and co-investigators have performed the calculation of electronic, structural and mechanical properties of 240 compounds of chemical formula M2X, using density-functional theory (DFT). These calculations were run on resources provided by SNIC, NAISS, and other foreign facilities in USA, South Korea and China. In a first study [W. Sun. et al., Ceram. Int. 50, 20796 (2024)], we analysed the ground state structures for identifying materials with interesting hardness, ductility and brittleness. We are currently investigating the remaining mestastable states, including mostly transition metal borides, but also promising carbides and nitrides. We will focus on the search of materials with adequate hardness, ductility and melting point, as well as on the understanding of the descriptors of these characteristics, based on a combination of DFT and machine learning. Next, we will focus on identifying what stable compounds are suitable for mechanical exfoliation, in accordance to our previous work on Mo2C [W. Sun, et al., Nanoscale 8, 15753 (2016)]. This will offer a realistic path for the synthesis of novel MXenes without resorting to chemical etching, which involves several disadvantages. Closely associated with 2-dimensional transition metal ceramics are transition metal dichalcogenides, which often exhibit superconductivity and charge-density waves. Analogously, 2-dimensional Mo2C is a superconductor below 8 K, but this superconductivity can be suppressed by doping with a magnetic atom, as eg Cr [S. Li, et al., ACS Nano 15, 14938 (2021)]. In this situation, a charge-density wave emerges, whose origin is still debated. We intend to perform electronic structure calculations to understand the origin of the charge-density wave and its relation to magnetism and superconductivity. DFT calculations will be accompanied by the Topological Data Analysis (TDA) [M. Bykov, et al., Phys. Rev. Lett. 126, 175501 (2021)], which will provide insight on the periodic lattice distortions, including topological data descriptors of the charge-density waves of various periodicity. The resources requested in this proposal are for an adequate storage space where we can collect all our data in one single location. So far, we have been storing partial calculations on various clusters, but this has become a problem over the years, since it has been prone to data loss and an inefficient communication among all the researchers involved in the data analysis. Additional calculations that will be necessary to complete our scientific tasks will be performed on computational resources already managed by the PI and co-investigators.