Abstract :
[en] Magnetoelectricity (ME), the coupling between magnetic and electric order parameters in crystals, plays a critical role in advancing multifunctional materials for applications in sensors, actuators, and memory devices. However, comprehending the underlying mechanisms driving this coupling at the atomic scale remains a challenge. Traditional computationally intensive methods, such as ab initio calculations under external fields, give the system's total response rather than revealing the localized effects. In this context, dynamical magnetic effective charges (DMCs), the magnetic analog of Born effective charges is a promising alternative. By directly giving the magnetic response to atomic displacements, DMCs enable a local and global (by projecting those charges onto polar modes) exploration of the ME effects.
To compute DMCs, we performed DFT calculations with spin-orbit coupling (SOC) using both VASP and ABINIT codes. A finite difference approach is employed, where atoms are slightly displaced from their equilibrium positions, and the induced magnetization response is analyzed to capture the local ME effects (from both spin and orbital contributions). The convergence is tedious, as the atomic displacements introduce small perturbations to the magnetism, with energy changes at computer noise levels. However, we found that the two different DFT codes and their different implementation and pseudopotentials give very similar results if one fine tune the convergence residual on the magnetization.
We investigated two materials: the room temperature BiCoO₃ multiferroic material and KCoF₃, a non-functional first order Jahn-Teller distorted material with unquenched orbital moments at cobalt sites. In those materials both spin and orbital contributions have similar amplitudes, such that the orbital contribution cannot be neglected, and that some atoms have surprisingly large DMCs, like the F atom in KCoF3 with a value close to 200×10−2 μB/Å, i.e. one of the largest values ever reported in real bulk material without f-electrons. Another surprising result is that the Co atom have a strictly zero value of its DMCs, while it is the magnetic cation that give the magnetic response of the system. We will show that a symmetry analysis explain why this situation happens and we will demonstrate that some magnetic crystals can even have all of their DMCs strictly zero. We will give some generalities of how DMC can be present in magnetic materials and how this allows to understand better some hidden magnetic properties of materials which opens new possibilities for controlling AFM with E-fields and antiferroelectricity with B-fields, offering a novel way to manipulate these properties.
This method holds potential for broader applications, offering a new way to control and predict magnetoelectric effects. Our work paves the way for future research aimed at designing and optimizing multiferroic and multifunctional materials with enhanced ME properties.
