Suppression of magnetism in Co$_3$Sn$_2$S$_2$ under external pressure

TL;DR

The paper investigates how external pressure up to 10.8 GPa affects the magnetization, band topology, and optical conductivity of Co$_3$Sn$_2$S$_2$, a ferromagnetic Weyl semimetal. It identifies that standard DFT overestimates magnetization under pressure and proposes two remedies: a symmetry-preserving adjustment of the sulfur position and a Bayesian-calibrated, mixed exchange-correlation functional. The mixed functional approach yields better agreement with both magnetization trends and optical-conductivity data, though band-structure changes can differ depending on the chosen remedy. Experimental measurements show Drude enhancement, phonon hardening, and interband shifts under pressure, with the interband conductivity at 10 GPa aligning best with the semi-empirical XC approach, underscoring the need for careful theory–experiment tuning and, potentially, more advanced many-body methods for a complete description.

Abstract

The ability to control the magnetic state provides a powerful means to tune the underlying band topology, enabling transitions between distinct electronic phases and the emergence of novel quantum phenomena. In this work, we address the evolution of ferromagnetic state upon applying external pressures up to 10.8~GPa using a combined experimental and theoretical study. The standard \emph{ab initio} Density Functional Theory computation including ionic relaxations grossly overestimates the unit cell magnetization as a function of pressure. In our theoretical analysis we identify two possible mechanisms to remedy this shortcoming. Matching the experimental observations is achieved by a symmetry-preserving adjustment of the sulfur atoms position within the unit cell. Alternatively, we explore various combinations of the exchange and correlation parts of the effective potential which reproduce the experimental magnetization, the structural parameters and the measured optical conductivity spectra. Thus, the pressure-dependent behavior of magnetization demands a careful theoretical treatment and analysis of theoretical and experimental data.

Figures (6)

• Figure 1: DFT self-consistent calculation of the relaxed lattice parameters together with the experimental data of Ref. ch.wa.19.
• Figure 2: (a) Unit cell of Co$_3$Sn$_2$S$_2$ at 10 GPa. The local environment of the Co-Co-Co triplet in the (b) fully relaxed DFT and (c) with S atom position fitted to match the observed magnetization. Relocating S atom closer to the Kagomé plane increases the Co-S-Co angle and decreases the Co-S distance, leading to a significant reconstruction of the band structure (see Fig. \ref{['fig:p10_bands']})
• Figure 3: Electronic band structure at $10$ GPa without SOC obtained with (a) unconstrained DFT, (b) fitted Wyckoff position of the sulfur atoms and (c) with different weights of the exchange and correlation parts of the XC potential.
• Figure 4: The real parts of the computed optical conductivities at ambient and at 10 GPa:(a) the in-plane Kagomé $\sigma_{xx}$, (b) the out-of-plane $\sigma_{zz}$, (c) the total optical conductivity. The dashed line corresponds to the resonant frequency $(\omega_{res})$ of the nodal line sc.ji.22.
• Figure 5: Experimental optical conductivity of Co$_3$Sn$_2$S$_2$ single crystals in the $ab$ plane as a function of pressure. The small, pressure-independent features at around 0.45 eV are due to multiple phonon excitations in the diamond anvil, which cannot be fully corrected by the reference measurement. Due to the same reason, the spectra were interpolated in the range 0.22 - 0.33 eV based on the Drude-Lorentz fit. Inset: Phonon mode as a function of pressure (spectra are offset for clarity).