Direct observation of imprinted antiferromagnetic vortex states in CoO/Fe/Ag(001) discs

Abstract

In magnetic thin films, a magnetic vortex is a state in which the magnetization vector curls around the centre of a confined structure 1. In a thin-film disc, vortex states are characterized by the vortex polarity and the winding number 2,3. In ferromagnetic (FM) discs, these two parameters have been shown to govern many fundamental properties of the vortex, such as its gyroscopic rotation 4 , polarity reversal 5-7 , core motion 8 and vortex-pair excitation 9. In antiferromagnetic (AFM) discs 10 , in contrast, there has been only indirect evidence for a vortex state, obtained through the observation of induced FM-ordered spins in the AFM disc 11-14. Here we report the direct observation of an AFM vortex state in the AFM layer of an AFM/FM bilayer system. We have fabricated single-crystalline NiO/Fe/Ag(001) and CoO/Fe/Ag(001) discs, and using X-ray magnetic linear dichroism techniques we observe two types of AFM vortex, one of which has no analogue in FM structures. We also show that a frozen AFM vortex can bias an FM vortex at low temperature. Single-crystalline NiO/Fe(12 nm)/Ag(001) and CoO/Fe(12 nm)/ Ag(001) films were grown by molecular beam epitaxy and patterned into discs using a focused ion beam. The FM Fe and AFM NiO and CoO were measured at the Advanced Light Source of Lawrence Berkeley National Laboratory by X-ray magnetic circular dichroism (XMCD) and X-ray magnetic linear dichroism (XMLD). Although the XMCD measurement is a standard method, the XMLD measurement on NiO and CoO in our experiment was made at the Ni L2 edge and Co L3 edge by changing the X-ray linear polarization angle (φ) relative to the Fe [001] magnetization axis, which is parallel to the NiO or CoO [110] crystalline axis (Fig. 1a; ref. 15). Figure 1 represents a typical CoO XMLD result from CoO (3 nm)/Fe (12 nm)/Ag(001) with the XMLD signal defined by the so-called L3 ratio (R L3),which is the X-ray absorption intensity at the photon energy E = 778.1 eV divided by the absorption intensity at E = 778.9 eV (ref. 15). The L3 ratio follows the expected cos 2 φ dependence for all CoO thicknesses. As the L3 ratio under this condition should reach its maximum value for X-ray polarization parallel to the CoO spin axis 16-18 , the R L3 result in Fig. 1b shows that the CoO spins are coupled collinearly to the Fe spins at smaller CoO thickness (d CoO = 0.6 nm) and perpendicularly to the Fe spins at larger CoO thickness (d CoO = 3.0 nm). This collinear to 90 • coupling transition was also reported in the NiO/Fe(001) system as a function of NiO thickness 19. The underlying mechanism of this coupling transition remains unclear so far and has been a focus of research in this field. Element-specific magnetic domains were imaged for both the FM Fe and the AFM NiO (CoO) of the bilayer discs at

Antiferromagnetic spintronics

“…For example, divergent vortices cannot form whenever a ferromagnetic component is present. In contrast, compensated antiferromagnetic materials can form divergent vortices in disks because they do not produce magnetic charges at the disk boundary (Wu et al, 2011). Another example can be found in domainwall dynamics.…”

“…III.B). Finally, imprinting multidomain states and magnetic textures in antiferromagnets can be achieved by preparing the ferromagnet in specific magnetic configurations, e.g., multidomain (Brück et al, 2005;Roshchin et al, 2005) and vortex (Salazar-Alvarez et al, 2009;Wu et al, 2011) states. Section I.C.4 is devoted to antiferromagnetic textures.…”

“…Direct observation usually requires large scale facilities with element-sensitive techniques like x-ray absorption spectroscopy (Weber et al, 2003;Salazar-Alvarez et al, 2009;Wu et al, 2011) or specific techniques with local probes like spin-polarized scanning tunneling microscopy (Bode et al, 2006;Loth et al, 2012) and quantum sensing with single spins (nitrogen vacancies) in diamond (Gross et al, 2017;Kosub et al, 2017). Alternatively, some information about antiferromagnetic domains can be inferred using indirect transport measurements.…”

Self-organized nano-structuring of CoO islands on Fe(001)

“…CoO is a compound of very broad interest in particular in the fields of catalysis [12,13], energy storage [14][15][16] and magnetism [17][18][19][20][21][22][23][24], being an antiferromagnetic insulator. Furthermore, CoO films coupled to ferromagnetic substrates (like Fe), show the well-known and technologically important exchange bias effect [25,26], which makes this kind of systems even more appealing.…”

Magnetic anisotropy at the buried CoO/Fe interface

“…17 In particular, the CoO/Fe(001) interface has been a workbench for many investigations related to AF/FM systems with reactive interfaces, ranging from the exchange bias mechanism, 14,18 to the influence of structure and stoichiometry on the magnetic properties, 19 to the realization of peculiar magnetic configurations, like vortices. 20 Very recently, we have demonstrated that it is possible to prepare a CoO/Fe(001) interface characterized by the absence of any Fe oxide, 21,22 which is a quite unique feature with respect to the common experimental situations. 10,19,23 This result has been accomplished by exploiting an ultra-thin Co buffer layer with a bct cubic structure.…”