While previous manipulation of magnetization has been based on magnetic fields and spin transfer torque due to the spin-spin interaction, more efficient approaches can be based on spin-orbit effects.

Transfer of orbital angular momentum can overcome the fundamental limit of adiabatic spin transfer torque that transfers 1ħ for every e- transferred across for instance a domain wall [1].

We have investigated using photoemission electron microscopy in detail the dynamics of domain walls and we find that in addition to conventional spin transfer torque, also spin orbit torques play a key role [2]. By comparing the wall motion with current-induced magnetization switching (CIMS) in our systems, we can deduce the spin-orbit torques independently of the DMI [3] and we find that the motion observed Ta(5nm)/Co20Fe60B20(1nm)/MgO(2nm) can be attributed to a DMI that is opposite to such stacks with a magnetic CoFe layer pointing to the B at the interface that governs the sign of the DMI [2]. Furthermore switching at zero fields can be achieved due to spin orbit torques that result from the spin Hall effect [3].

In addition to moving domain walls, we have recently studied more complex skyrmion spin structures [4]. Using spin orbit torques, we have displaced trains of skyrmions in a skyrmion “racetrack” [5] and imaged the dynamics using scanning transmission x-ray microscopy.

Even more efficient is to use the interplay between mechanical degrees of freedom and the spin degree of freedom as mechanical changes can influence the spin degree of freedom due to spin – orbit coupling. This is exploited in (artificial-)multiferroics, where one uses electric fields to vary the lattice and via magnetoelectric coupling thus the magnetic properties. The appeal of this approach is the possibility to manipulate magnetization with very low power, as electric fields without continuous electric currents suffice to change the magnetization.
We have exploited different systems, where the interplay between mechanical and spin degrees of freedom is studied to understand the underlying physical phenomena.

The first system is Ni on top of a piezoelectric crystal (PMN-PT). The application of an electric field across the PMN-PT substrate generates a uniaxial in-plane stress that strains the magnetic layer on top as well. The uniaxial stress then generates a uniaxial magnetocrystalline anisotropy in the magnetic material due to magnetoelastic coupling allowing us to reversibly displace domain walls in nanostructures including non-volatile switching paving a way for low power devices [6]. Using an all oxide stack with Fe3O4 on top of PMN-PT allows us to induce even phase transitions using electrical fields to apply strain [7].

[1] O. Boulle et al., Mater. Sci. Eng. R 72, 159 (2011)
[2] R. Lo Conte et al.; Phys. Rev. B 91, 014433 (2015)
[3] R. Lo Conte et al., Appl. Phys. Lett. 105, 122404 (2014)
[4] F. Büttner et al., Nature Phys. 11, 225 (2015)
[5] S. Woo et al., arxiv:1502.07376 (2015)
[6] S. Finizio et al., Phys. Rev. Appl. 1, 021001 (2014); A. Tkach et al., Appl. Phys. Lett. 106, 62404 (2015)
[7] A. Tkach et al., Phys. Rev. B 91, 24405 (2015)