Ultra-fast spintronics

The booming demand on data consumption pushes new technologies to be able to process bigger volumes of data at a faster rate. Thus, ultrafast manipulation and processing of information has now become a challenge of tomorrow’s information technologies.

Ultra-fast phenomena in magnetism is a 20-year-old subject [BEAU1996] that still leads to strong controversies among active researchers. Indeed, it is still not entirely clear why a ferromagnetic layer hit by an intense femtosecond laser pulse demagnetizes well below one picosecond. Relevant phenomena involve the almost instantaneous generation of hot electrons, which have to occupy bands with different populations and spin polarizations before relaxing to the Fermi level. Despite the lack of a full understanding, it has been shown that a picosecond spin pulse is emitted in the process. One can therefore envision the possibility of ultrafast spintronic components and the emergence of terahertz spintronic devices. The latter is nowadays achieved by ultra-fast demagnetization of a ferromagnetic layer along with the electrical conversion of the associated spin pulse (see Fig. 3). This is realized in bi-layers composed of a ferromagnet (CoFeB for example) and a heavy metal like Pt where the large spin-orbit coupling generates the ‘Inverse Spin Hall Effect’ converting spin into charge currents. Such structures generate THz bursts with a power comparable to that of the conventional semiconductors (ZnTe) when submitted to intense and ultra-fast optical pulses. The simplicity of such devices makes them competitive and an intense research effort is now undertaken in the international community. It has also been shown that another spin/charge conversion based on Rashba split 2D states can also be used. Despite this burgeoning success, the processes at stake are not yet clear and a model allowing describing and analysing quantitatively the spin-carrier dynamics is presently lacking. We propose here to develop a dynamical model of spin-polarized carrier transport beyond the superdiffusive current approach involving the quantum spin-orbit polarized electronic transmission and relaxation, able to probe the spin properties of the interface. The model will be developed in the frame of a multiband tight-binding formalism involving plane-to-plane calculations of the electronic structure and spin-currents in order to treat the 3d/5d transition metals (e.g Fe/Pt or Co/Pt) as well as NiFe/LAO/STO oxide systems (Ti 3d bands).

Figure 3: Schematics of THz emission from ultra-fast demagnetization of a FM layer with subsequent spin/charge conversion.

On the other hand, other materials called antiferromagnets (AF) are attracting a surge of interest because of their potential dynamical properties. In these materials, the magnetic moments of atoms align in a regular pattern with neighbouring spins pointing in opposite directions. Because of their zero net moment, antiferromagnets are rather insensitive to a magnetic field and difficult to probe. However, a key property of AFs lies in their dynamics. Because at resonance each sublattice has to precess in the exchange field of the other one, the corresponding frequencies are in the THz range. However, it remains experimentally difficult to efficiently trigger AF resonance and detect the THz signals. Recent predictions suggest genuinely interesting new avenues to shortcut these problems using spin current injection. This allows envisioning THz sources controlled by spin currents [CHEN2016, KHYM2017], as well as ultra-fast motion of antiferromagnetic domain walls although nothing of the kind has been reported yet. With the development of ultra-fast light sources, recent works on AF dynamics are in the time domain, using pump-probe techniques [DUON2004, RUBA2010]. Coherent spin waves in insulating NiO have thus been triggered by single cycle-Terahertz pulses [KAMP2010], but a precise control of their emission is currently lacking. In particular, the coupling between the pump (often a femtosecond optical pulse) and the AF order generally remains unclear. Importantly, most of the pump-probe studies are on (bulk) insulators because of their low damping, but rarely in thin films. It is also important to image the AF domains and try to move domain walls at fast speeds (as predicted theoretically). At this point in time, the scientific community has realized that the introduction of functionalities offered by antiferromagnetic materials would represent a new paradigm in spintronics, but nothing conclusive has been published on the dynamical side.


[BEAU1996] E. Beaurepaire et al.
Ultrafast Spin Dynamics in Ferromagnetic Nickel
Phys. Rev. Lett. 76, 4250 (1996)

[CHEN2016] R. Cheng et al.
Terahertz Antiferromagnetic Spin Hall Nano-oscillator
Phys. Rev. Lett. 116, 207603 (2016)

[KHYM2017] R. Khymyn et al.
Antiferromagnetic THz-frequency Josephson-like oscillator driven by spin current
Sci. Rep. 7, 43705 (2017)

[DUON2004] N. P. Duong et al.
Ultrafast Manipulation of antiferromagnetism of NiO
Phys. Rev. Lett. 93, 117402 (2004)

[RUBA2010] A. Rubano et al.
Influence of laser pulse shaping on the ultrafast dynamics in antiferromagnetic NiO
Phys. Rev. B 82, 174431 (2010)

[KAMP2010] T. Kampfrath et al.
Coherent terahertz control of antiferromagnetic spin waves
Nat. Photon. 5, 31 (2010)