Racetrack memory based on in-plane-field controlled domain-wall pinning

Abstract

Magnetic domain wall motion could be the key to the next generation of data storage devices, shift registers without mechanically moving parts. Various concepts of such so-called 'racetrack memories' have been developed, but they are usually plagued by the need for high current densities or complex geometrical requirements. We introduce a new device concept, based on the interfacial Dzyaloshinskii-Moriya interaction (DMI), of which the importance in magnetic thin films was recently discovered. In this device the domain walls are moved solely by magnetic fields. Unidirectionality is created utilizing the recent observation that the strength with which a domain wall is pinned at an anisotropy barrier depends on the direction of the in-plane field due to the chiral nature of DMI. We demonstrate proof-of-principle experiments to verify that unidirectional domain-wall motion is achieved and investigate several material stacks for this novel device including a detailed analysis of device performance for consecutive pinning and depinning processes.

Figure 1

Device concept. (a) Schematic graph of the change in H _depin (both for up-down and down-up domain walls) as a function of H _x. If H _z is below H _depin a DW stays pinned at an anisotropy barrier while above H _depin it will move past it, as indicated by the black inset cartoons. (b) An initial magnetic configuration is shown in the top cartoon (red shading = down, blue shading = up) together with a schematic energy landscape for the domain walls, which are represented by circles. H _x lowers the energy barriers for one type of DW, which is subsequently moved by an H _z pulse, and the system ends up in the configuration shown in the middle cartoon. A following H _x and H _z with opposite sign move the other type of DWs in the same direction, ending up in the configuration shown in the bottom cartoon.

Figure 2

Magnetic configuration (red = down, blue = up) of a strip for every cycle in the propagation sequence. (a) Domain walls successfully being moved to the right, fields strengths H _x and H _z are 140 mT and 10.4 mT respectively. (b) Domain walls successfully being moved to the right by changing the sign of combined fields, field strengths H _x and H _z are 140 mT and 9.2 mT respectively). (c) In-plane fields (120 mT) are applied transverse to instead of along the strips, H _z = 10.8 mT. (d) Unirradiated strip, field strengths H _x = 150 mT and H _z = 5.0 mT are used. (e) Sample with top Pt layer grown under higher argon pressure, field strengths H _x = 80 mT and H _z = 22.8 mT are used. (f) Sample with top Pt layer replaced by Ir, field strengths H _x = 160 mT and H _z = 8.0 mT are used.

Figure 3

Success rate (indicated by colour scale) as a function of H _x and H _z. (a) Chance on a successful pinning step. (b) Chance on a successful depinning step. (c) Chance that a complete procedure (both pinning and moving) is successful for the proof-of-principle sample. (d) Chance that a complete procedure is successful for a sample with the top Pt layer grown at a higher pressure. (e) Chance that a complete procedure is successful for an irradiated Pt/Co/Ir sample. (f) Chance that a complete procedure is successful for an unirradiated Pt/Co/Ir sample.

Figure 4

Boundaries for which a domain wall has 50 percent chance to depin, for both the irradiated and unirradiated Pt/Co/Ir sample and for both domain walls that are suppose to remain pinned and for domain walls that are supposed to move.