Skip to Main content Skip to Navigation
New interface

Transport de spin dans des matériaux magnétiques en couches minces, par simulations Monte Carlo

Abstract : The study of spin transport properties in magnetic materials started in the 1950s. The intrinsic magnetism of this kind of materials represents an additional diffusion source compared to non-magnetic materials, in which diffusion sources are mainly phonons, impurities and crystalline defects of the lattice. Magnetic scattering is more complex than the two previous sources expressed. On one hand the magnetic diffusion induces different behaviors of resistivity due to the material, as it can be a metal or a semiconductor. On the other hand, the type of magnetism exhibited by a structure is also at the origin of different magnetic behaviors of resistivity. During the 1950s many authors focused on metals and magnetic semiconductors. It follows that each of the various theories works for a particular case, but failed for all other behaviors observed experimentally. For example, the resistivity of magnetic metals is a monotonic function of temperature (it is only the derivative that presents a singularity with respect to the temperature), while magnetic semiconductors show a peak at the critical temperature (Tc) , temperature separating the order to the disorder magnetic phases. Note that the different interpretations concerning the magnetic diffusion around the critical temperature are still a controversial point. Moreover, the theories available to date about transport in magnetic materials should be selected according to the material studied (type of structure, metal or semiconductor, type of magnetism). A general method which could be used for all behaviors of resistivity, whatever the nature of the solid (metal or semiconductor) and the type of magnetism was lacking at the beginning of this work. First, it is to this lack we want to answer with this thesis. Indeed, the spectacular development of spintronics relies on the use of magnetic semiconductors and metals, so a rigorous understanding of the scattering mechanisms near the critical temperature is needed today. Our approach consists in using a Monte Carlo Metropolis simulation to study this problem. I have developed an algorithm that applies to a fairly general Hamiltonian of the system. Thanks to this Hamiltonian and with an appropriate set of parameters, we are able to reproduce experimental trends and/or generic behaviors presented in the literature for ferromagnetic and antiferromagnetic (frustrated and non-frustrated) semiconductors. The method also allows to perform a systematic study of materials experimentally characterized taking into account units to realize direct comparisons between our simulation results and experimental measurements, with good agreement. This as been applied to the case of MnTe, a magnetic semiconductor, for example. Whereas this work was limited to the study of semiconductor materials, our next objective will consist in extending the method to diluted magnetic semiconductors (DMS), as well as scattering at the magnetic interfaces of GMR devices, an issue that currently raises many questions. So far, our approach is unique and seems able to reproduce a large number of experimental behaviors of resistivity, in particular in the case of semiconductors, and providing a new explanation to the controversy about scattering mechanisms around the critical temperature. My contribution to this work was to develop, in collaboration with my PhD advisor Prof. Hung The Diep, the calculation model and the Monte Carlo Metropolis spin transport algorithm. Thanks to this algorithm, I obtained many results which led to the interpretation of different experimental results presented in my thesis. My main contribution to this work is: -The development of a Monte Carlo Metropolis algorithm with numerical tricks for the reduction of statistical fluctuations samples (multi-step averaging) - The study of different ferromagnetic and antiferromagnetic materials (not frustrated). A very significant observation of resistivity behavior around Tc and at low T was seen: the dependence of the peak has Tc depends on well-identified physical variables (interactions: range, nature, relaxation time, lattice, ...) - Study of the frustration effect on the resistivity. Effect of a first order transition on resistivity. - Interpretation of the behavior of the resistivity using numerical data, especially in terms of clusters and relaxation time. - Quantitative study of MnTe and comparison with experiment: excellent agreement. Let us briefly explicit these points in a few words. The first point I would like to emphasize is the new interpretation we gave regarding the scattering mechanisms near the order/disorder temperature transition for non-frustrated lattices. I showed during my PhD that the shape of the resistivity peak around Tc is a direct consequence of two related effects induced by the magnetism of the lattice. Around Tc the lattice percolates and shows large clusters of parallel (antiparallel) spins which represent many areas of low (high) energies for electrons which to go through. In addition lattice spin fluctuations undergo a critical slowing down which produce an extremely slow evolution of the magnetic landscape. The magnetic lattice can be seen as a frozen magnetic state which allows itinerant spins to be trapped in parallel clusters (low energies), giving rise to a peak resistivity at Tc. As a result we emphasize that the study of transport in magnetic systems must take into account the local energy landscape of the magnetic lattice, as well as the dynamics of spin fluctuations as a function of temperature (which were not included in the previous works). These two points allow us to generalize the treatment of resistivity to non-frustrated systems. Frustrated antiferromagnetic systems are characterized by a large number of degenerate states (often infinite in case of Ising spins). However, the thin film structure allows us to make it as a finite degenerate state. We firstly showed that the degeneration of the system induces a first order resistivity transition, and according to the degenerate state of the system, the resistivity can present two kinds of transition: a transition from low to high resistivities with increasing temperature, or vice versa.
Complete list of metadata

Cited literature [121 references]  Display  Hide  Download
Contributor : Yann Magnin Connect in order to contact the contributor
Submitted on : Tuesday, October 2, 2012 - 10:33:29 AM
Last modification on : Friday, August 5, 2022 - 2:45:59 PM
Long-term archiving on: : Thursday, January 3, 2013 - 3:45:15 AM


  • HAL Id : tel-00737547, version 1


Yann Magnin. Transport de spin dans des matériaux magnétiques en couches minces, par simulations Monte Carlo. Physique Numérique [physics.comp-ph]. Université de Cergy Pontoise, 2011. Français. ⟨NNT : ⟩. ⟨tel-00737547⟩



Record views


Files downloads