Spin and charge transport in single crystal magnetic tunnel junctions – SPINCHAT

The understanding of the complex transport mechanisms in single crystal MTJs represents the key parameter towards the improvement and the better control of the tunnel magneto-resistive (TMR) effects in MTJ devices such as sensors, read-heads or magnetic non-volatile memories (MRAM). Within this project, we will address fundamental aspects related to the physics of the charge and spin transport. Combining experimental and theoretical approaches, we will explore how the magneto-resistive effects can be optimized and controlled in MTJ grown either by MBE or by sputtering. One of the project objectives is to investigate the connection between magneto-transport properties and the magnetic coupling/interactions between the two ferromagnetic electrodes in single crystal MgO based tunnel junctions since the theory predicts a direct relation between these three properties: TMR, exchange coupling at zero bias and spin transfer torque at non-zero bias. This represents an extremely important milestone towards the spin-torque/ magnetic switching implementation in non-volatile MRAM devices. Furthermore, we want to correlate the effect of the torque (or spin wave excitations) on the voltage variation of the TMR. This control is extremely important for potential applications of MTJ which are always operated at a finite voltage. Within this project two different transport regimes will be addressed in detail: the equilibrium regime (the tunnel junction is not biased) and the out of equilibrium regime (biased tunnel junction). In the equilibrium regime one of the objectives is to elucidate the controversial point concerning the mechanisms (origin) of the magnetic interactions mediated by tunnelling transport in epitaxial MTJs. Taking into account the complexity of the transport in single crystal systems, we will address independently the influence on the coupling of the insulator (thickness, resonant impurity/vacancy levels), of the ferromagnetic electrodes and of the interfacial chemistry and electronic structure. The morphology of the Fe/MgO interfaces and of the MgO barrier will be modulated by light ion irradiation techniques. Alternatively, interfacial impurities or metallic adlayers will be intercalated in the MTJ stack in order to engineer the interfacial electronic structure and the spin/symmetry filtering efficiency. In all these samples Electron Holography, Transmission Electron Microscopy (TEM) and magnetometry experiments will be involved to study the coupling at macroscopic and microscopic level. The understanding of the equilibrium coupling represents the first step toward the out of equilibrium spin transfer phenomena. In the out-of-equilibrium regime one of the objectives of our project is to control/enhance the amplitude and the voltage variation of the TMR in two different classes of systems: samples elaborated by MBE (LPM) and samples elaborated by sputtering (SPINTEC, LPM). The large TMR may be obtained by spin and symmetry filtering effects across epitaxial barriers and enhanced by a careful control of the electronic structure of the ferromagnetic electrodes and of the interfaces. An important point is the study of the TMR variation with the MgO barrier thickness. Theoretically, when the MgO thickness is reduced, a reduction of the filtering efficiency within the MgO is expected with a negative effect on the TMR amplitude. However, low resistance-area product in epitaxial magnetic tunnel junctions is required for the integration of MTJ in read-heads, high-density MRAMs or MTJ devices where the magnetization is switched by a critical current by spin-torque mechanisms. One of the objectives of our project is to engineer systems with large TMR and low RA product, by playing with the chemical structure of the interfaces. The correlation between the TMR effects and the magnetic interactions (spin-torque effects) will be investigated by two different techniques: noise measurements in patterned micrometric size MTJ devices and by magnetization switching experiments on sub micrometric pillars. By noise experiments, voltage fluctuations due to magneto-resistance will provide the whole spectrum of the ferromagnetic resonance (FMR) excitations. The noise and transport measurements performed on the same sample will allow the direct correlation of the magneto-transport properties and the spin transfer torque effects. In pillar-shape sub micrometric patterned systems the spin-torque effects will be investigated directly in a regime where larger current density will be involved. Within this regime, the magnetization switching phenomena and high frequency oscillation phenomena will be addressed directly. All the experimental activity within this project will be assisted by theoretical modelling for: the electronic properties of the MTJ systems, the charge transport (modelling of TMR effects) and the spin transport (modelling of spin-torque/coupling effects).