Database, Features, and Machine Learning Model to Identify Thermally Driven Metal–Insulator Transition Compounds

Abstract

Metal–insulator transition (MIT) compounds are materials that may exhibit metallic or insulating behavior, depending on the physical conditions, and are of immense fundamental interest owing to their potential applications in emerging microelectronics. An important subset of MIT materials are those with a transition driven by temperature. The number of thermally driven MIT materials, however, is scarce, which makes delineating these compounds from those that are exclusively insulating or metallic challenging. Most research that addresses thermal MITs is limited by the domain knowledge of the scientists to a subset of MIT materials and is often focused on a limited subset of possible features. Here, using a combination of domain knowledge and natural language processing (NLP) searches, we have built a material database comprising thermally driven MITs as well as metals and insulators with similar chemical composition and stoichiometries to the MIT compounds. We featurized this data set using a wide variety of compositional, structural, and energetic descriptors, including two MIT relevant energy scales, the estimated Hubbard interaction and the charge transfer energy, as well as the structure-bond-stress metric referred to as the global-instability index (GII). We then performed supervised classification on this data set, constructing three electronic-state classifiers: metal vs nonmetal (M), insulator vs noninsulator (I), and MIT vs non-MIT (T). This classification allows us to identify new features separating MIT materials from non-MIT materials. These include the 2D feature space consisting of the average deviation of the covalent radius, the range of the Mendeleev number and Ewald energy. We discuss the relationship of these atomic features to the physical interactions underlying MITs in the rare-earth nickelate family. We then elaborate on other features (GII and Ewald energy) and examine how they affect the classification of binary vanadium and titanium oxides. Last, we implement an online and publicly accessible version of the classifiers, enabling quick probabilistic class predictions by uploading a crystallographic structure file. The broad accessibility of our database, newly identified features, and user-friendly classifier models will aid in accelerating the discovery of MIT materials.

Machines for Materials and Materials for Machines: Metal-Insulator Transitions and Artificial Intelligence

“…Most MIT materials are transition metal oxides [7] and, as a model system in which to study MITs, the rare earth nickelates are a shining example for multiple reasons. These include -when synthesised as epitaxial thin films -their high quality and adaptability in terms of multilayer and superlattice structures as well as their intrinsic physics [8] .…”

“…Machine learningas a model-agnostic toolmay be used to remove some of these biases and provide new insight into yet unexplored directions. This may also help overcome some of the limitations inherent in current theoretical studies of MIT materials and may not need the same amount of user input as, for instance, a first principles calculation [7]. ML approaches, for example, using classifiers trained on existing literature [7,76] can be useful in quickly identifying which materials should be studied in more detail from large databases such as Materials Project [63].…”

“…In meeting these challenges some studies have made progress in identifying certain key features of MIT materials such as the rare-earth nickelates [7]. The most important is the transition metal-transition metal distance, which is essential in setting the electronic bandwidth, as well as in determining electronic interaction parameters such as the Hubbard U.…”

Trigonal Symmetry Breaking and its Electronic Effects in Two-Dimensional Dihalides and Trihalides

“…Such configurations may be susceptible to metalinsulator transitions (MIT), which we assess using a recently devised machine-learning classification model [57]. We found that the binary MIT-non MIT classifier tends to predict most of the 2D halides are candidate MIT compounds, giving a positive MIT classification for CrCl 3 , FeCl 3 , IrBr 3 , MnCl 2 , RuCl 3 , TiCl 2 , VCl 3 and ZrI 2 , and a negative classification for CrI 3 , FeI 2 and NiI 2 .…”

Evidence of a cluster spin-glass phase in the skyrmion-hosting GaMo4S8 compound

“…Certain members of this family feature superconductivity, metal-to-insulator transitions and electric-field-induced resistive switching. [6][7][8] Most notably, the magnetic phase diagram of these compounds may host magnetic skyrmion phases. 4 Skyrmions are swirling spin structures, mostly found in chiral magnets.…”