Magnetic skyrmion lattice domain manipulation enables new memory technology

Magnetic skyrmions – swirls of magnetic moments found in certain magnetic materials - are promising candidates for next generation memory devices. Thanks to their unique properties they can provide an increase in storage density, while at the same time dramatically saving electrical energy compared to their current ferromagnetic counterparts.

However within a given material skyrmions tend to form highly ordered long-range lattice states that need to be broken into smaller domains before they can be used as memory storage devices. To overcome this, researchers at Diamond Light Source turned to BLADE (Beamline for Advanced Dichroism Experiments), also known as I10, both to characterise the skyrmion structure and to create and control skyrmion lattice domains.

By using the beamline’s soft X-ray diffractometer, RASOR, the researchers were able to clearly observe the magnetic signal within their chosen material, Cu2OSeO3. Characterising the skyrmion structure both accurately and quickly was a fundamental experimental challenge. RASOR allowed them to obtain detailed structural information of the skyrmion lattice state and its dynamics as it offered ultrahigh reciprocal space resolution, as well as a unique sample environment - ultrahigh vacuum, high temperature stability, magnetic vector field and multi-axis goniometer.

At Diamond the researchers experimentally identified a new magnetic state in which smaller domains within the material were identified – which they termed the multidomain skyrmion lattice state. In addition the researchers were also able to manipulate the size, shape, and distribution of the domains in situ, work which is crucial to enable and design skyrmion-based devices in the future.

Figure 1
Figure 1: Skyrmion-based racetrack memory and the skyrmion lattice state. a) Illustration of the skyrmion racetrack memory concept, with an individual skyrmion structure highlighted in the centre. Adapted from Sci. Rep. 5, 15773 (2015). b,c) Single-domain and multi-domain skyrmion lattice state, respectively. Adapted from Nano Lett. 16, 3285 (2016).

Modern information technology is based on the storage and manipulation of binary units (bits). Consequently, the performance of a computing system is largely governed by the properties of the bit-carrying material. Magnetic materials are excellent bit-carrying materials and have led to magnetic random access memory technology and future spintronics concepts, owing to their non-volatility, small bit size, and elegant bit manipulation schemes relying on spin currents alone. In the framework of magnetic technologies, bits are represented by magnetic domains in ferromagnetically ordered systems. Each domain contains hundreds of uniformly aligned spins. The binary information is encoded by the direction of these spins, e.g., ‘1’ for spin-up domains and ‘0’ for spin-down domains. One example of an advanced magnetic storage scheme is the so-called racetrack memory1, in which the magnetic domains can be read and written in a shift-register-like fashion (similar to the concept illustrated in Fig. 1a).

Recently, it has been recognised that such memory schemes can be further upgraded, allowing for much higher storage densities and much improved energy consumption (five orders of magnitude lower compared to current-induced domain wall motion)2. The idea behind this improvement is inspired by the mathematical concept of topology, which states that in certain materials, the magnetic domains can form unique and stable textures that are topologically protected against decaying. Such non-trivial, swirl-like magnetisation configurations are called magnetic skyrmions2, as shown in the centre of Fig. 1a. Combining these two concepts led to the proposal of a skyrmion-based racetrack memory structure (Fig. 1a).

Some of the most promising skyrmion-carrying materials are the cubic 3d chiral magnets2. Their crystal structure belongs to the space group P213. The primitive cubic structure only presents a 21 screw-axis and three-fold diagonal symmetry axis, which excludes inversion symmetry. As a consequence, the spin-orbit coupling of the 3d magnetic ions allows for a superexchange-type Hamiltonian to become relevant, the so-called Dzyaloshinskii-Moriya interaction. It promotes the formation of a long-range-ordered periodic vortex structure, the skyrmion lattice. The individual skyrmions assemble themselves into a rigid, hexagonally-ordered magnetic crystal, and the correlation length reaches hundreds of micrometres (Fig. 1b). Such ordered lattice structures are not directly useful for memory applications (Fig. 1a). In order to encode skyrmions as bits, the individual skyrmions have to be uncoupled from the rigid, two-dimensional arrays. In other words, skyrmion lattice domains have to be formed, analogous to the formation of ferromagnetic domains in classical memory concepts (Fig. 1c).

Theoretically, spontaneous symmetry breaking of the skyrmion lattice state can occur just on the surface of a bulk material, due to the O(2)-symmetric surface anisotropy, assisted by a tilted magnetic field3. However, the experimental observation remained a great challenge. First, there is a very limited number of magnetic characterisation techniques suited for the study of 3d chiral magnets. So far, only neutron diffraction, Lorentz transmission electron microscopy, and magnetic force microscopy have been successfully applied to the observation of the skyrmion lattice phase2. However, transmission-type neutron and electron techniques are not surface-sensitive; while force microscopy has a limited probing area. Second, in situ magnetic characterisation has to be performed, which requires a complex sample environment with high temperature stability and magnetic vector field capability.

The RASOR diffractometer on beamline I10 at Diamond Light Source is uniquely suited for such a task. The soft X-ray resonant magnetic diffraction is the ideal experimental technique for the unambiguous characterisation of the skyrmion lattice state on the surface of the sample4. Moreover, RASOR has ultrahigh reciprocal space resolution – a necessary condition for the observation of these unique magnetic structures. RASOR’s large ultra-high vacuum (UHV) chamber allows for the incorporation of a magnetic vector field setup – the key ingredient for creating skyrmion lattice domains. Furthermore, the size of the X-ray beam can be reduced so that a diffraction-based scanning-type imaging can be performed, with which the real-space skyrmion domain pattern can be reconstructed.

The scattering geometry is shown in Fig. 2a, in which the magnetic field can be freely rotated by an angle γ within the scattering plane. The material investigated was single-crystalline Cu2OSeO3 – a typical skyrmion-carrying ferroelectric system. The characteristic six-fold-symmetric diffraction pattern is shown in (Fig. 2b), representing the single-domain skyrmion lattice state3,5. The raster scan confirms that the skyrmion crystal is long-range-ordered. By tilting the magnetic field angle, the diffraction pattern splits into a ‘necklace-like’ pattern (Fig. 2d), suggesting that the skyrmion lattice breaks up into domains. This is further elaborated by the domain image in Fig. 2e3,5.

Figure 2: Experimental setup and summary of the main results. a) Scattering geometry used in resonant soft X-ray diffraction experiments. b) Diffraction pattern plotted in reciprocal space in the (hk1)-plane for the single-domain skyrmion lattice state. c) Real-space domain image of the single-domain skyrmion lattice state. d) Diffraction pattern of the multidomain skyrmion lattice state and e) real-space domain pattern. Adapted from Appl. Phys. Lett. 109, 192406(2016).

By performing a systematic resonant X-ray scattering study, a new skyrmion phase, the multidomain skyrmion lattice state, was discovered. The significance of this work is twofold. First, a control parameter was identified that allows for the creation and manipulation of the domains – an important step towards memory applications. Second, it was demonstrated that resonant X-ray scattering is a powerful technique suited for studying magnetic skyrmions. The advantages of soft X-rays, such as their polarisation, surface-sensitivity, element specific features, as well as the fast measurement time, will allow for a much deeper understanding of magnetic skyrmion systems in the future.

References:

  1. Parkin, S. S. P. et al. Magnetic Domain-Wall Racetrack Memory. Science 320, 190-194, doi:10.1126/science.1145799 (2008).
  2. Nagaosa, N. et al. Topological Properties and Dynamics of Magnetic Skyrmions. Nat. Nanotech. 8, 899-911, doi:10.1038/NNANO.2013.243 (2013).
  3. Zhang, S. L. et al. Multidomain Skyrmion Lattice State in Cu2OSeO3. Nano Lett. 16, 3285-3291, doi:10.1021/acs.nanolett.6b00845 (2016).
  4. Zhang, S. L. et al. Resonant Elastic X-Ray Scattering from the Skyrmion Lattice in Cu2OSeO3. Phys. Rev. B 93, 214420, doi:10.1103/PhysRevB.93.214420 (2016).
  5. Zhang, S. L. et al. Imaging and Manipulation of Skyrmion Lattice Domains in Cu2OSeO3. Appl. Phys. Lett. 109, 192406, doi:10.1063/1.4967499 (2016).

Funding acknowledgement:

We thank Diamond Light Source for beamtime awarded on I10 (RASOR). We acknowledge financial support by the Semiconductor Research Corporation (SRC) and EPSRC (EP/N032128/1).

Corresponding author:

Dr Shilei Zhang, University of Oxford, shilei.zhang@physics.ox.ac.uk

Related publication:

Zhang SL, Bauer A, Burn DM, Milde P, Neuber E, Eng LM, Berger H, Pfleiderer C, van der Laan G, Hesjedal T. Multidomain Skyrmion Lattice State in Cu2OSeO3. Nano Letters 16(5), 3285-3291, doi: 10.1021/acs.nanolett.6b00845 (2016).

Publication keywords:

Magnetic skyrmions; Magnetic memory; Resonant X-ray scattering