Thesis presented April 08, 2021

Abstract:
Mott insulators are systems that according to band theory should be conductors, but in reality are found to be insulators because of the strong electronic correlations. However, the metallic state can be induced by the application of pressure or an electric field. In some of these systems, a relatively weak electric field pulse has been shown to control the insulator to metal transition: this is called resistive switching. The possibility of switching in Mott insulators is of great interest to realize a new type of resistive memories (ReRAM), called Mott memories. Unlike other types of ReRAM, where the conductive filament created during switching is caused by ion migration, the filament formation mechanism in Mott insulators is of purely electronic origin, which could solve the major problem of ReRAM which is variability. Switching in Mott insulators is probably due to an electronic avalanche breakdown phenomenon, but the mechanism of filament formation is not yet understood. By analogy with the metallic state induced under pressure, one possible hypothesis is a local compressive effect of the material.

In this thesis we study the GaV₄S₈ compound, a narrow gap Mott insulator of the AM₄X₈ lacunar spinel family in which the switching phenomenon has already been demonstrated and which is a good model system to advance the understanding of the phenomenon and try to determine if Mott memories actually have an advantage over other ReRAMs. We present several instrumental developments where we have adapted existing pressure measurement techniques to the study of insulators. We have performed measurements of ac calorimetry, capacitance or dielectric constant, loss, resistivity and ac magnetic susceptibility in a diamond anvil cell with an in-situ pressure tuning system that can be used over a wide temperature range. This enables us to present a complete study of the temperature - pressure phase diagram in GaV4S8, highlighting the transition to the metallic state at around 14 GPa and the evolution of the Mott gap with pressure. This compound also presents a very rich phase diagram with temperature and magnetic field with a ferroelectric Jahn-Teller structural transition at 43 K then a magnetic transition at 12.7 K towards a cycloidal phase then a ferromagnetic phase at lower temperature and towards a skyrmionic phase of the type Néel under magnetic field. Our study under pressure allowed us to follow these different transitions and the evolution of the different phases with pressure.

Evaluating a material's performance for applications involves well-defined test protocols. In general this involves the long and costly process of the production of thin layers, then of test devices. We explored the feasibility of using lithography techniques on single crystals that would make it much easier to perform such tests to explore new materials. We present numerical simulations which show that it is possible in principle to define geometries on single crystals which would have characteristics close to thin film devices. We then present the different steps carried out to obtain such devices and the difficulties encountered.

Finally, we present an original study of the switching phenomenon by combining the two parameters that induce the metallic state: pressure and electric field. We show that the threshold electric field necessary to induce a volatile transition changes little with the pressure while the Mott gap decreases sharply. We discuss the implications of these results to better understand the formation of the metallic state in this system.

Keywords:
metal-to-insulator transition, resistive switching, pressure, mott insulators, strong correlations, memory

On-line thesis.