WorksPhysics and Astronomy
Topological and spin-orbit effects on orbital moments in ultra-thin magnetic films
Abstract
Topological orbital moments (TOMs) are a direct hallmark of a magnetic texture with a non-trivial spin topology. In addition to giving insight into the topology of the magnetic texture, TOMs could also be used to manipulate magnetic structures with a compensated total spin moment. Experimental evidence of TOMs has been provided via transport measurements of the Hall effect in intercalated van-der-Waals materials. However, a direct observation of TOMs is still missing. Another complication arises for the unambiguous proof of the topological origin since orbital moments can also occur due to spin-orbit coupling. Here, we use first-principles electronic structure theory to investigate the origin of orbital moments in different compensated spin structures with a non-trivial topology. We focus on ultrathin magnetic films at surfaces such as Pd/Mn bilayers on Re(0001) which represent ideal model systems for the detection of TOMs since it is possible to apply experimental techniques with local resolution of magnetic properties such as spin-polarized scanning tunneling microscopy. Due to its trivial topology we use the row-wise antiferromagnetic (RW-AFM) state in a hexagonal monolayer to analyze the spin-orbit induced contributions. The triple-Q (3Q) state is a superposition state of three RW-AFM (1Q) states and thus electronically similar, however, due to its non-trivial spin topology it exhibits TOMs. The comparison between these two spin states allows us to disentangle the topological and spin-orbit contributions to the orbital moments. We find that TOMs have an important contribution in spin-compensated systems, since the spin-orbit coupling induced orbital moments are nearly compensated. First-principles calculations for atomic-scale skyrmion lattices in Fe monolayers on different surfaces exhibit the same general trend found for the 3Q state.
Introduction
Complex non-collinear magnetic structures have attracted considerable attention due to their rich physical properties and potential applications [1, 2, 3, 4]. In recent years, there has been a particular focus on non-collinear spin structures with non-trivial topology on the atomic scale [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17]. In many of these systems, the magnetic unit cell comprises only a few lattice sites, and the spin magnetization is fully compensated, meaning that the vector sum of all spin moments within the unit cell vanishes. Here, the spin moment refers to the magnetic moment arising from the spin momentum of the electrons.
In addition to spin moments, magnetic systems may also exhibit orbital magnetic moments originating from the orbital motion of electrons. These orbital contributions are often neglected because they are typically much smaller than the corresponding spin moments. However, in systems with compensated spin magnetization, orbital moments can become a decisive factor and may dominate the magnetic response of the material [18].
In isolated atoms orbital moments arise due to the orbital quantum number of the occupied orbitals. This is a consequence of Hund’s second rule to reduce the Coulomb interaction between the electrons. In solids, consisting of a very large number of atoms, this orbital moment is quenched due to the crystal field, as long as no global preferred direction exists. Spin moments define such a preferred direction. Spin-orbit coupling (SOC) couples the orbital moments of all electrons to the magnetic moments, leading to a net orbital moment at each lattice site. Energy contributions arising from SOC can be decisive for the magnetic ground state, e.g. due to the Dzyaloshinskii-Moriya interaction (DMI) or magnetocrystalline anisotropy energy (MAE). SOC-induced orbital moments occur in magnetic systems, independent of the topology of the magnetic configuration. As an example, we consider the spin structure of the row-wise antiferromagnetic (RW-AFM) state in a hexagonal magnetic monolayer and its tentative orbital moments as sketched in Figure 1(a).
Figure 1. Schematics of the contributions to the total orbital moment in an ultrathin film. (a) Spin structure of the RW-AFM state. The orbital moments due to SOC, , can approximately be expected to be aligned with the spin moments . Since the RW-AFM state is collinear the topological orbital moments, , vanish. (b) Spin structure of the 3Q state with tetrahedron angles between adjacent spin moments in the orientation with respect to the atomic lattice. Besides , which are nearly aligned with the spin moments, this state exhibits due to its non-trivial spin topology , which are uniform within the unit cell and aligned perpendicular to the film. (c+d) Spin moments, SOC induced orbital moments and TOM for two lattice sites of the state. The SOC induced orbital moments and the TOMs add up to the total orbital moment. The TOM is homogeneous over the spin structure, hence the same for both lattice sites (cf. panel b). The magnitude of the SOC induced orbital moments and its angle to the spin moment depend on the absolute orientation of the spin moment. This leads to different contributions in (c) and (d). Orbital moments are shown schematically (not to scale) for visualization purposes.
Topological orbital moments (TOMs) represent a different type of orbital moments which have recently come into focus [10, 11, 13, 14, 19, 20, 21, 22]. TOMs depend on the topology of the magnetic texture and occur only for a non-vanishing Berry-curvature. Therefore, the RW-AFM state exhibits no topological orbital moments (Figure 1(a)). A non-trivial topology arises for example for the non-coplanar triple-Q (3Q) state [5, 15, 23, 24, 25], which exhibits tetrahedron angles between nearest-neighbor spins as shown in Figure 1(b). Here the TOM form an additional contribution, , besides the orbital moments due to SOC, . The total orbital moment is formed by the sum of the two terms
Note that further contributions exist such as chiral orbital moments [20]. However, it has been shown that these have only a small effect for a skyrmion lattice [20]. Therefore, further contributions are neglected in this work. In an atomistic spin model the TOMs are given by [14]
where is the topological orbital susceptibility, a material dependent parameter. The non-coplanar alignment of the magnetic structure is described by the scalar spin chirality
defined for spin moments , , and at lattice sites , , and . For a collinear or coplanar spin texture the scalar spin chirality is zero. A non-zero scalar spin chirality results in a preferred direction for electrons to move around the three lattice sites [6, 26]. This net chiral motion of electrons induces a topological orbital moment, which is perpendicular to the plane of the chiral current, i.e. the plane of the three lattice sites. The direction of the TOM for each triplet is given by
with lattice vectors , , and . For a single magnetic layer the TOMs are consequently perpendicular to the surface, i.e. either pointing up or down with respect to the surface. In contrast, is usually nearly aligned with the magnetic moment. For the non-coplanar 3Q state, the spin moments of the four atoms in the unit cell compensate leading to a vanishing net spin moment. However, the TOMs are uniform over the lattice resulting in a finite total orbital magnetization (Figure 1(b)).
When the spin structure of the 3Q state is inverted, the scalar spin chirality, Eq. (3), and consequently the TOMs, Eq. (2), change sign and point into the opposite direction. Two domains, which are inverted with respect to each other have antiparallel aligned TOMs. These small orbital moments could be used to manipulate the domain by an external magnetic field even though the net spin moment vanishes. For example a magnetic field perpendicular to the surface should shift the domain wall between two inverted domains in order to enlarge the domain, in which the TOMs are aligned with the field direction.
In the intercalated van-der-Waals material such a manipulation of the 3Q state has recently been reported by transport measurements of a spontaneous topological Hall effect [27, 28]. Similar signatures in transport experiments have been found in the structurally similar material [29], which were also related to the 3Q state [30]. However, this provides only indirect evidence of TOM. A net topological orbital magnetization is also predicted for single skyrmions in a ferromagnetic background and it has been suggested that these are observable via XMCD [13]. However, a direct measurement of TOMs has not been reported so far.
Ultrathin magnetic films on surfaces are good candidate systems to study TOMs locally, since they allow real-space observations, e.g. via spin-polarized scanning tunneling microscopy (SP-STM) [15, 25]. The magnetic state in such films can even be revealed via SP-STM, when a monatomic adlayer is grown on the magnetic layer [25, 31, 32]. Such an adlayer gives additional control over the electronic and magnetic material properties of the system. However, as discussed above, TOMs do not occur in an isolated fashion but are accompanied with orbital moments originating from SOC.
Here we investigate the interplay between topological and spin-orbit coupling induced orbital moments in ultrathin magnetic films on surfaces based on density functional theory (DFT). We consider the 3Q state in Pd/Mn/Re(0001) as an example for a non-coplanar spin structure with a non-trivial spin topology. The 3Q state in Pd/Mn/Re(0001) has been predicted based on DFT and experimentally observed via SP-STM [25]. Since this spin state can couple in different orientations to the atomic lattice it allows us to study different contributions from spin-orbit coupling to the orbital moments. We use the RW-AFM state as a reference state to disentangle orbital moments due to topology and spin-orbit coupling. Our analysis shows that the simple picture of additive contributions to the total orbital moments holds very well for the 3Q state. This conclusion is also true for complex atomic-scale skyrmion lattice states in Fe monolayers on the Re(0001) and Ir(111) surface covered by a few layers of Ir and Rh, respectively.
This paper is structured as follows. In section II we present the computational details of our DFT calculations. In section III we discuss the results of our study. We begin with a comparison of the electronic structure of the collinear RW-AFM state and the non-coplanar 3Q state. Next we analyze the SOC-induced orbital moments of the RW-AFM state and their dependence on the global orientation of the spin moments. The orbital moments of the electronically similar but non-coplanar 3Q spin state are discussed afterwards. Based on these insights of the previous sections we present orbital moments for different nanoskyrmion lattice, where some are spin-compensated, while others are not. In section IV we provide the conclusions of our work.
Computational details
We have performed DFT calculations using the Fleur code [33, 34], which is based on the full-potential linearized augmented plane wave (FLAPW) method [35, 36], for a Pd/Mn/Re(0001) film system and a freestanding Pd/Mn/Re trilayer. The in-plane lattice constant is Å, which is the experimental value for Re. For both systems, DFT calculations were carried out in the RW-AFM and in the 3Q state. The interlayer distances have been optimized for the Pd/Mn/Re(0001) film system in the RW-AFM state, in which six layers of the Re substrate were considered. The three top layers are stacked in an A-B-C (fcc) stacking sequence, while the Re substrate has a C-A (hcp) stacking sequence. Structural optimizations of the film system have been performed in the generalized gradient approximation (GGA) using the PBE exchange-correlation functional [37]. The relaxed interlayer distances obtained for the Pd/Mn/Re(0001) film system are given in Ref. [25]. The interlayer distances of the freestanding Pd/Mn/Re trilayer were chosen according to the relaxed values of the film system. All DFT calculations apart from the structural relaxations were performed in the local density approximation (LDA) using the exchange-correlation potential in the parameterization of Vosko, Wilk, and Nusair [38]. The muffin-tin radii were chosen as a.u. for Mn and Pd and a.u. for Re. The cut-off parameter for the basis functions was . The RW-AFM state has been calculated in a supercell, while the 3Q state was calculated in a supercell. For both states a grid of k-points has been used. All calculations, including those taking SOC into account, have been performed self-consistently.
All nanoskyrmion lattices were investigated with the VASP code [39, 40], which is based on the projector augmented wave (PAW) method [41, 42]. The skyrmion lattice (SkX), the double-SkX (d-SkX), and the multi-Q state were calculated in a 16 atom two-dimensional supercell for the film system of an Fe monolayer (ML) on three atomic Ir layers on the Re(0001) surface, denoted as Fe/Ir-3/Re(0001) in the following. A grid of (15 15 1) k-points was used with an energy cut-off for the wave functions of 268 eV. Further computational details can be found in Ref. [17]. The SkX-12, SkX-19 and SkX-27 lattices were studied for an Fe ML on the Ir(111) surface, an Fe ML on two atomic Rh layers on Ir(111), and for an Fe ML on one atomic Rh layer on Ir(111), respectively. For these DFT calculations a cut-off of 300 eV has been chosen and k-point meshes of (10 20 1), (11 11 1) and (5 15 1), respectively, were used. Further computational details can be found in Ref. [16, 43].
Results
First, we illustrate how topological and SOC contributions can combine to the total orbital moment by a schematic example. We consider a hexagonal magnetic monolayer in the RW-AFM state (Figure 1(a)) and in the 3Q state (Figure 1(b)). Figure 1(c,Figure 1d) show two spin moments and from the state (cf. Figure 1(b)). is aligned with the surface normal (-axis) and has a tetrahedral angle with respect to it. Exemplary SOC induced orbital moments, , and topological orbital moments, , are sketched for both orientations. Since all neighbors in the 3Q state have the same angles
between them and each triplet of neighbors has the same chirality, the TOM is homogeneous over the unit cell. This means all lattice sites have the same , in this case aligned along the -axis (see lower right sketch in Figure 1(b)). Note, that we assumed the sign for according to the Pd/Mn/Re(0001) film system (cf. Tab. I). For the opposite sign would be aligned along the -axis. The SOC induced orbital moments, on the other hand, are typically roughly aligned with the spin moment. The magnitude of the SOC induced orbital moment and the angle with respect to the spin moment depend on the absolute direction of the spin moment. In Figure 1(c), the SOC induced orbital moment is chosen smaller than in Figure 1(d), since the magnitude of the orbital moment can in general depend on the orientation of the spin moment. is nearly, but not perfectly aligned with as is often the case. Note, that in the sketch the size of the orbital moments is exaggerated for better visibility. and are partially compensating, leading to a small total orbital moment. For the SOC induced orbital moment, , is assumed to be larger (Figure 1(d)). Further, and both have negative components, resulting in a relatively large total orbital moment. The two spin moments of the 3Q state demonstrate, that the SOC induced orbital moments are not homogeneous over the lattice and that the addition of and can further enhance or decrease the total orbital moment at a given lattice site.
Since is nearly aligned with the spin moment at each atom site, one might expect the SOC induced orbital moments to cancel upon summing over the unit cell whenever the spin moments compensate. However, the magnitude of varies between different sites as shown in Figure 1 and therefore, the net compensation is not guaranteed. Further, the combination of and raises the question, whether the TOM can be measured at all without explicit knowledge of the SOC induced orbital moments. In the following, we address these issues based on electronic structure theory via DFT calculations for the example of the 3Q state of the freestanding Pd/Mn/Re trilayer and the Pd/Mn/Re(0001) film system and then generalize our conclusions to more complex nanoskyromion spin structures in Fe-based ultrathin films.
Electronic structure of the RW-AFM and 3Q state
To investigate the contributions due to spin topology and spin-orbit coupling to the orbital moments, we use the RW-AFM and 3Q state on a hexagonal lattice. These spin states have a 2D magnetic unit cell of two and four atoms, respectively, making them well suited for a detailed analysis (cf. Figure 1(a,Figure 1b)). The RW-AFM state can be expressed as a spin spiral characterized by a q-vector that is located at the M point of the 2D Brillouin zone. Therefore, this state can also be denoted as a single-Q (1Q) state. The 3Q state can be constructed as a superposition of three RW-AFM (1Q) states propagating along the three equivalent crystallographic directions of the surface. Both magnetic states have the same energy contribution due to exchange while the DMI vanishes [24]. The 3Q state was found via DFT as the ground state in Pd/Mn/Re(0001) and experimentally confirmed using SP-STM [25].
Here, as a model system for the DFT calculations, we use in addition a freestanding Pd/Mn/Re trilayer, which allows for a detailed analysis of all three layers. Note, that we have also performed DFT calculations for the film system Pd/Fe/Re(0001) in which the Re surface has been modeled by six Re layers (see Supplemental Material for details).
Figure 2 shows the local density of states (LDOS) for the Pd, Mn and Re atoms of the freestanding Pd/Mn/Re trilayer. The solid lines represent the LDOS of the 3Q state, while the filled areas represent the RW-AFM state. The shift between the LDOS of the spin-up and spin-down channels of Mn is the exchange splitting leading to a spin moment of 2.99 . For the Pd and Re atoms, there is also a small difference visible between the spin-up and spin-down channels, which corresponds to small spin moments of 0.07 and 0.03 per atom, respectively, which are induced by the hybridization with the adjacent Mn atoms. For all three elements, the LDOS of the RW-AFM (1Q) state is quite similar to that of the 3Q state consistent with their relation by the superposition. Due to the close connection of their electronic structure, we expect that general trends of the orbital moments in the RW-AFM state are transferable to the 3Q state. However, while the spin moments in the RW-AFM state are collinear, the 3Q state has a non-coplanar spin structure. Therefore, the RW-AFM state has only a spin-orbit coupling contribution to the orbital moments, while the 3Q state exhibits both a spin-orbit and a topological contribution to its orbital moments.
Figure 2. LDOS of a Pd/Mn/Re trilayer in the RW-AFM vs. 3Q state. The solid lines represent the LDOS of the freestanding Pd/Mn/Re trilayer obtained via DFT for the 3Q state and the filled areas the LDOS of the RW-AFM state (cf. Figure 1(a,Figure 1b)). The LDOS is shown for (a) Pd, (b) Mn and (c) Re atoms. Note that these DFT calculations include SOC and the 3Q orientation has been chosen for the 3Q state and the spin moments are oriented along the z-axis for the RW-AFM state. While both atoms in the magnetic unit cell of the RW-AFM state have the same LDOS, the four lattice sites of the 3Q state differ slightly due to SOC. Here the lattice site with the spin moment pointing along the positive z-axis is shown, which is the same orientation shown for the RW-AFM state.
Orbital moments of the RW-AFM state
To investigate the orbital contribution independent of the topological contribution , we use the RW-AFM state as a reference state. This collinear spin state has a trivial topology and therefore does not exhibit TOMs. In the RW-AFM state the spin moments form ferromagnetic rows, where neighboring rows are aligned antiparallel to each other, as shown in Figure 1(a). While keeping the angles between all spin moments fixed, the global spin orientation of the RW-AFM state can be varied. The optimal orientation of this state with respect to the atomic lattice is defined by effects of SOC, such as magnetocrystalline anisotropy energy (MAE) and anisotropic symmetric exchange (ASE) [15, 24]. Here we consider different orientations of the RW-AFM state, to observe how the total energy and the orbital moments depend on the orientation.
In Figure 3 the RW-AFM state is rotated from an out-of-plane () to an in-plane () orientation, i.e. from an orientation of the spin moments perpendicular to the surface to an orientation parallel to the surface. The total energy obtained via DFT (Figure 3(a)) decreases when the spin moments are rotated into the plane of the surface. Therefore, the MAE prefers an orientation in the film plane, a so called easy-plane anisotropy. As the energy decreases, the magnitude of the orbital moment increases (Figure 3(b)). This trend of a larger orbital moment for an energetically favored orientation due to the magnetocrystalline anisotropy is consistent with the well-known Bruno formula [44].
Figure 3. Total energy and orbital moments for the RW-AFM state. The spin moments in the RW-AFM state of a Pd/Mn/Re trilayer in fcc stacking are rotated from an out-of-plane orientation to an in-plane orientation. (a) Total energy calculated via DFT including SOC. (b) Magnitude of the orbital moments (blue) and angle between the orbital and spin moments (red). The dashed gray line indicates the angle, which corresponds to the 3Q2 state (cf. Figure 5).
The magnitude of the spin moment is nearly constant for all orientations. While the magnitude of the orbital moments is increasing, the orbital moments are not always perfectly aligned with the spin moments. The angle between the spin moments and the orbital moments is displayed in Figure 3(b). The angle between and can be divided into a polar angle and an azimuthal angle . As shown in Supplementary Fig. S1 and Supplementary Note 1, both and are non-zero. For , i.e. an in-plane orientation of the spin moments, and are perfectly aligned.
In addition to the polar angle , the azimuthal angle of the spin moments can be varied. Upon changing , the MAE contribution can also vary with a three fold symmetry. For an unsupported magnetic monolayer the MAE would even have a six fold symmetry. However, the Pd and Re atoms in the adjacent layers lower this symmetry to a three-fold rotational symmetry. Nevertheless, this contribution is typically very small. A larger contribution can arise from the ASE, which also occurs due to SOC. In contrast to the MAE, the ASE is a two-site interaction. In the atomistic spin model the ASE is given by
where the normalized connection vector between lattice sites and . Depending on the sign of , which depends on the electronic structure of the system, the ASE can favor an alignment of the RW-AFM state along or perpendicular to the ferromagnetic rows [15]. For a positive sign, the ASE favors the same orientation as the magnetic dipole-dipole interaction.
In Figure 4 the spin moments of the RW-AFM state are oriented with a polar angle of with respect to the surface normal, which corresponds to the out-of-plane component of the 3Q2 state (introduced below, cf. Figure 5(c)). For comparison, this angle is also indicated by a dashed gray line in Figure 3. For this fixed polar angle the azimuthal angle is varied, as shown in Figure 4(a). The total energy (Figure 4(b)), including the effects of SOC, has two equivalent minima for and . The two minima indicate that the energy difference can be attributed to the ASE and not to the MAE, which would lead to at least three minima due to symmetry (see Supplementary Note 2 for the MAE contribution). In comparison to Figure 3 the orbital moments (Figure 4(c)) do not scale inverse proportional to the energy any more. This relation apparently only holds for the MAE and not for other SOC contributions, such as the two-site interaction ASE. The angles of and correspond to an orientation of the spins parallel to the ferromagnetic rows of the RW-AFM state (cf. Figure 4(a)), which is the same ordering favored by the dipole-dipole interaction. While the energy and the absolute value of the orbital moment do not scale inverse proportionally, the absolute value of the orbital moment still varies with the direction of the spin moment. [45].
Figure 4. Orbital moments of the RW-AFM state for a varying azimuthal angle. The spin moments of the RW-AFM state in a Pd/Mn/Re trilayer have a polar angle of , which corresponds to the orientation of the spin moments for the state. This angle is indicated by a gray dashed line in Figure 3. The azimuthal angle is varied. (a) Sketch of the orientation of the spin moments with respect to the lattice for different angles . The gray dashed lines mark the angles represented in (e+f). (b) Total energy of system with respect to . (c) Magnitude of the orbital moment (blue) and the spin moment (red). (d) -component of the orbital moment (blue) and total angle between the orbital and spin moment (green). (e+f) orientation of the spin moments for (e) and (f) for the three atoms, which are located on the dashed gray line in (a).
Figure 5. -components of the total orbital moments for the 3Q state in different configurations. The -component of the orbital moment for a Pd/Mn/Re trilayer, , is shown for three orientations of the 3Q state with respect to the atomic lattice: (a+b) the , (c) the , and (d) the orientation. The DFT calculations were performed for (a) in the scalar-relativistic approximation, i.e. neglecting SOC, and for (b-d) including SOC. Note, that without SOC the orbital moments originate only from the spin topology of the 3Q state and are invariant under a global spin rotation, i.e. identical for the , and state. All values are given in units of .
For a rotation of by the orbital moment can differ. This can be seen for example for the angles of and . For both angles the energy exhibits a local maximum, however, with a different value (Figure 4(b)). The difference in energy comes along with a difference in the absolute value of the orbital moment as well as the spin moment (Figure 4(c)). The origin of this energy difference for a rotation of by lies in the orientation of the spin moments with respect to the position of the Re and Pd atoms. Figure 4(e,Figure 4f) show a side view of the atoms along the dashed gray line in Figure 4a. For the spin moments are roughly aligned with the connection vectors between the Mn atom and the Pd or the Re atom (Figure 4(e)). As seen in Figure 4(c), this leads to a larger orbital moment compared to the value for . Here, the spin moments are nearly perpendicular to the connection vectors (Figure 4(f)). This shows that both substrate and adlayer have an influence on the orbital moment of the Mn layer.
While the absolute value of the orbital moments varies for a full rotation of by about (Figure 4(c)), the variation with is significantly larger with a value of about (Figure 3(b)). The magnitude of the orbital moments varies up to 15% from the mean orbital moment. This demonstrates that the substrate and adlayer atoms do not have a large, but still a measurable effect on the orbital moments of Mn. Hence, the system can only be treated approximately as an effective monolayer system. Overall, the magnitude of the orbital moments depends on the alignment of the spin moments with each other as well as the alignment with respect to the induced spin moments of the Re and Pd atoms. The magnitude of the spin moments varies inverse proportional to the orbital moments.
The -component of the orbital moment varies on the order of for a rotation of the azimuthal angle (Figure 4(d)). There is also a difference between the values for angles of and . Therefore, the -component does not only depend on the polar angle , but also on the azimuthal angle . This is especially interesting with respect to the interplay of orbital moments due to SOC and due to the spin topology studied for the state in the next section.
Orbital moments of the 3Q state
The state is a superposition of three RW-AFM (1Q) states. As seen in Figure 2, both have a similar electronic structure. Therefore, we expect that the general trends of the orbital moments due to spin-orbit coupling, observed for the RW-AFM state, can be transferred to the state. While the RW-AFM state is collinear, the state is non-coplanar and has therefore a non-trivial topology. In addition to the orbital moments due to SOC, this leads to the formation of TOMs, which are uniform over the whole unit cell (cf. Figure 1(b)). It has been shown that the spin configuration of the RW-AFM (1Q) state can be continuously transformed into the state [17, 24] and that the TOMs are continuously increasing along the path. As seen for the RW-AFM state (Figure 3), the size of the orbital moment due to SOC depends on the orientation of the spin moment with respect to the atomic lattice. Therefore, one expects the orbital moments to differ for individual lattice sites in the unit cell of a non-collinear spin state. For the state, both effects – orbital moments due to SOC and due to the spin topology – occur. To differentiate between them, one can calculate the TOM in DFT using the scalar-relativistic approximation, in which SOC effects are not included. An additional DFT calculation including SOC yields the total contribution of and . When the TOM is already known, the SOC contribution to the orbital moments can be distinguished. However, the TOM depends on the electronic structure via the topological orbital susceptibility (cf. Eq. (2)). In principle, a change in the electronic states, which occurs upon including SOC, can also modify the TOM. Typically, the influence of SOC on the electronic structure is small. This implies that the TOMs are relatively robust against the influence of small SOC effects. In [20] the TOMs were explicitly calculated with gradually increasing SOC strength and found only small changes in the TOM for small SOC contributions. Here, we show that this conclusion also holds for several small-scale spin textures such as the state and nanoskyrmion lattices. In Figure 5(a) the orbital moments of the state are displayed for the orientation with respect to the lattice (indicated by the tetrahedron shown at the top of the panel) at the different sites of the unit cell by colored hexagons. In this calculation, SOC has been neglected which results in a purely topological orbital moment contribution. At each lattice site the TOM has a magnitude of as obtained via our DFT calculations. A summation over the unit cell results in a total contribution of the TOM of . These TOM only have a -component and are invariant under global rotations. Therefore, all spin orientations of the 3Q state with respect to the lattice have the same TOM. While the TOM is invariant under a global spin rotation since it depends on the spin topology, SOC-induced effects such as the ASE or depend on the global orientation of the state.
We consider three distinct spin orientations: the , the , and the configuration (see Figure 5(b-Figure 5d)). For the 3Q state, all neighboring spins span a tetrahedral angle. Therefore, this state can be visualized by a tetrahedron which can be used to indicate the orientation of the state relative to the atomic lattice (Figure 5). The orientation of the TOM along the negative -axis rather than the positive -axis depends on the scalar spin chirality as well as the topological orbital susceptibility (cf. Eq. (2)). In a different material system the TOMs for the same state could be aligned with the -axis due to a change in the topological orbital susceptibility. If, on the other hand, all spins are inverted, i.e. going from an all-out to an all-in 3Q state, the scalar spin chirality changes sign (cf. Eq. (3)) and all TOM point along the -axis.
Figure 5(b) shows the component of the orbital moment for the orientation with SOC effects included. The total component summed over the unit cell is and very similar to the TOM without SOC (Figure 5(a)). Intuitively one can understand this, since the spin moments are compensated over the unit cell, i.e. the total spin moment vanishes, and the orbital moments are roughly aligned with the spin moments, therefore also the are compensated. Interestingly this still holds true, if the magnitude of the orbital moments due to SOC varies over the unit cell, as our calculations show. While the total contribution is very similar, the distribution within the unit cell is not any more uniform. When SOC is included the three lattice sites II to IV have a negative contribution to the -component of the orbital moment, while the lattice site I has a positive contribution.
The absolute value of the orbital moment’s -component for all four lattice sites is larger in the orientation upon including SOC (Figure 5(b)) than obtained neglecting it (Figure 5(a)). However, the contributions are partially compensating over the unit cell. The variations at the different lattice sites originate from the orientation of the orbital moments due to SOC, . In general, downward pointing spin moments have a larger component, since the TOM and orbital moment due to SOC point in a similar direction. This can be seen e.g. for the atoms I and III of the state (Figure 5(c)). Spin moments, which have a contribution along the positive -direction, also have with a positive component, which are partially canceling the TOM with a negative -component, e.g. atoms II and IV of the state (Figure 5(c)). However, also for the an orientations we obtain a very similar total component upon summing over all atoms in the unit cell (cf. Figure 5(c,Figure 5d)). Therefore, we conclude from these calculations that the total orbital moment of the unit cell of the 3Q state is nearly unaffected by including the effect of SOC and remains basically at the value obtained from the TOM.
To understand how spin-orbit coupling and spin topology contribute to the total orbital moment, , we analyze the orbital moments for the , and configuration in more detail (Figure 6). Figure 6(a) shows the absolute value of the total orbital moment of the Mn atoms in the Pd/Mn/Re trilayer for the state. For Mn atom I the absolute value coincides with the -component (cf. Figure 5(b)). Therefore, the orbital moment is completely aligned with the -axis. As the total orbital moment for this lattice site is pointing along the positive -axis, while the TOM points along the negative -axis (Figure 5(a)), the contribution due to SOC needs to point along the positive -axis and is stronger than the TOM contribution. This alignment of and is sketched in Figure 1(c). The alignment of with the spin moment coincides with the results for the RW-AFM state (Figure 4). This observation can be confirmed, when the TOM (Figure 5(a)) is subtracted from the total orbital moment, i.e. by plotting (Figure 6(b)), which leaves approximately the SOC contributions. Note, that small differences could occur when the TOM obtained with and without SOC differ. Based on our DFT results we conclude that this effect is small. As seen by comparing Figure 5(a) and Figure 6(b) the SOC contribution to the orbital moment of Mn atom I is three times as large as the TOM contribution.
Figure 6. Orbital moments including SOC for three different orientations of the 3Q state. The orbital moments are shown for the Mn atoms of the freestanding Pd/Mn/Re trilayer. (a, d, g) Absolute value of the total orbital moment, obtained via DFT, in which effects due to SOC and spin topology are included. (b, e, h) Absolute value of the orbital moment upon substracting the TOM obtained from the scalar-relativistic calculation (cf. Figure 5a). This value is approximately the absolute value of . (c, f, i) is subtracted from the total orbital moment, leaving approximately . The resulting vector is projected onto the direction of the spin moment. All three quantities are displayed for the three considered orientations of the 3Q state with respect to the atomic lattice. (a-c) 3Q¹ state (d-f) 3Q² state (g-i) 3Q³ state The contributions were obtained by a calculation without SOC (cf. Figure 5). All values are given in units of .
In Figure 6(c), the difference between the total orbital moment and the TOM, , is projected onto the local direction of the spin moment . As discussed above, this is approximately the contribution of SOC to the orbital moment projected onto the direction of the spin moment, i.e. . For Mn atom I the projection and the absolute value of are equal (cf. Figure 6(b) vs. Figure 6(c)). Therefore, is aligned with the spin moment and opposite to . Using the same quantities, we can also investigate the orbital moments of Mn atoms II to IV. The absolute value of the total orbital moment (Figure 6(a)) is around four times larger than the component (Figure 5(b)). Thus in contrast to Mn atom I, the orbital moment is not aligned with the -axis. This case is sketched in Figure 1(d). When the TOM is subtracted (Figure 6(b)), is only slightly smaller than . Therefore, both contributions are not directly competing, as for Mn atom I. When is compared to its projection onto the spin moment (Figure 6(c)) it becomes apparent that is nearly aligned with the spin moment. This is the expected orientation of the orbital moment due to SOC.
In conclusion, the SOC contributions for Mn atom I and atoms II to IV differ, because they have a different out-of-plane component. This is consistent with the observations for the RW-AFM state (cf. Figure 3). Due to the opposite orientation of and for Mn atom I, both contributions are directly competing. As the spin moments of Mn atoms II to IV are not directly aligned with the -axis, the change of the magnitude of the orbital moments due to TOM is relatively small. As , calculated from the difference between the total orbital moment and the TOM, is consistent with the observations for the RW-AFM state, the approximation seems to hold that the TOM is nearly constant under the effect of SOC.
For the configuration all Mn atoms have the same out-of-plane component (Figure 6(d)). Therefore, all atoms are expected to have similar contribution to the orbital moments due to SOC. As seen from Figure 4 for the RW-AFM state, the contribution can still differ slightly, depending on the azimuthal angle . From Figure 6(d) we find that indeed all atoms have a similar total orbital moment. The moments of Mn atoms II and IV are exactly the same. The total orbital moments for Mn atoms I and III are slightly larger and differ from each other. When the TOM contribution, obtained from the scalar-relativistic DFT calculation (Figure 5(a)), is subtracted (Figure 6(e)), the orbital moment becomes slightly larger for Mn atoms II and IV, while it is slightly decreased for atoms I and III. Since Mn atoms II and IV have a positive -component, the TOM reduces the total orbital moment. For Mn atoms I and III, with a negative -component, the TOM increases the total orbital moment, as sketched in Figure 1(d). This can also be seen from Figure 5(c), where Mn atoms I and III have a component of . Here and both point along the negative -axis. For atoms II and IV, is oriented along the negative -axis and exhibits a positive -contribution such that the two contributions are competing, which results in . The comparison between and the projection onto the direction of the spin moment (Figure 6(f)) demonstrates that all orbital moments are roughly, but not exactly aligned with the spin moments. This is expected, since for the RW-AFM state with the exact same out-of-plane component (Figure 4(d)), the angle between orbital and spin moment varies between and .
For the configuration (Figure 6(g)) Mn atoms II and IV have a very similar magnitude of the orbital moment, while differs for atoms I and III. If the contribution of is subtracted (Figure 6(h)) the contributions of Mn atom I and III become nearly equivalent. This shows again that in one case the TOM is increasing the total orbital moment, while in the other case it is decreasing it. The contribution for Mn atoms I and III is smaller than for Mn atoms IV and II, as their spin moments have an out-of-plane contribution, which leads to a smaller orbital moment. Due to their different spin alignment with the spin moments of the Pd and Re atoms, Mn atom II has a slightly smaller than Mn atom IV.
As shown in Figure 6(i), the orbital moments of Mn atoms II and IV align perfectly with their spin moments, while for Mn atoms I and III there is an angle between the orbital and the spin moment. This can be seen from the difference in Figure 6(h) and Figure 6(i). It is in line with the observations for the RW-AFM state since Figure 3 shows that the orbital moments are nearly aligned with the spin moments, when the spin moment has no out-of-plane component.
Since the difference between and , calculated without SOC, is nearly aligned with the direction of the spin moments, one can conclude that . However, there are small deviations, where the combination of TOM and SOC has a different contribution for two lattice sites, which would be symmetry equivalent under just one of the two effects. Examples for this are the lattice sites I and III of the state.
Overall in the case of the 3Q state the component, summed over the unit cell, can be viewed as nearly identical without and with SOC considered. In both cases the unit cell has a net -component of the orbital moment, which flips its sign when all spin moments are inverted. Potentially this allows to manipulate domains of the 3Q state by an external magnetic field which interacts with the orbital moments. Such an effect has been suggested by transport experiments performed for intercalated van der Waals materials [27, 28, 30]. Our DFT calculations indicate that the manipulation of the domains via a field is independent of SOC contributions to the orbital moments. On the other hand, the component of the orbital moments varies for different atoms within the magnetic unit cell of the 3Q state, when SOC effects are considered. This could allow for a unique detection of the 3Q state and its orientation via STM. Harnessing the tunneling anisotropic magnetoresistance (TAMR) effect [46, 47], the 3Q state could be detected even with a non-magnetic STM tip.
In addition to the freestanding Pd/Mn/Re trilayer, which has been discussed so far, we have calculated the orbital moments for a Pd/Mn/Re(0001) film system. The orbital moments of the 3Q state with and without SOC are shown in Supplementary Figs. S4 and S5. These figures are equivalent to Figure 5 and Figure 6, respectively, but consider the full film system, i.e. take the effect of the Re(0001) surface into account. Qualitatively, the trilayer and the film system show the same behavior. The components of the total orbital moments, summed over the unit cell, are nearly identical with and without SOC (Supplementary Fig. S4).
Compared to the trilayer system, however, the TOMs are approximately twice as large. For the film system, the ASE has the opposite sign compared to the freestanding trilayer, favoring an antiferromagnetic alignment along the connection vectors. Consequently, the angular dependence of the SOC-induced orbital moment differs between the two systems. Nevertheless, as in the trilayer, the total orbital moment in Pd/Mn/Re(0001) is approximately given by the sum of the SOC-induced orbital moment and the TOM calculated without SOC, further supporting our conclusions.
Orbital moments of nanoskyrmion lattices
The 3Q state is sometimes referred to as the smallest possible skyrmion lattice (SkX). We have shown previously that TOMs can also play a role in other atomic-scale skyrmion lattices [17]. As an example, an atomic scale SkX with zero net spin magnetization is displayed in Figure 7. This non-coplanar spin state has been predicted as a possible ground state for an Fe monolayer on a Re(0001) surface covered by three atomic layers of Ir – denoted below as Fe/Ir-3/Re(0001) – a system which has also been studied experimentally via SP-STM [17].
Figure 7. Topological and total orbital moments of the SkX state in Fe/Ir-3/Re(0001). (a) Topological orbitals moments as calculated via Eq. (2) and Eq. (3), using the topological orbital susceptibility given in Tab. I. (b) -component of the orbital moment obtained by a DFT calculation for Fe/Ir-3/Re(0001) neglecting SOC. The - and - components are zero for this calculation. (c) -component of the orbital moment obtained by a DFT calculation including SOC. For this calculation the - and -components are locally non-zero.
Figure 7(a) shows the TOM calculated in the atomistic model for this spin texture, i.e. based on Eq. (2). In contrast to the 3Q state, the TOMs are not anymore uniform within the unit cell. For some lattice sites the TOM is pointing along the negative -axis, for others the TOM points along the positive -axis and for some the TOM is zero. Overall the lattice has a net TOM aligned with the positive -axis.
In general, according to Eq. (2), the topological orbital susceptibility can vary over the lattice. Here, we assume for simplicity that the electronic structure is nearly constant over the lattice and approximate the topological orbital susceptibility to be given by a single constant independent of the atom site. We calculate the topological orbital susceptibility by comparing the TOM from DFT with the TOM from Eq. (2). Note, that the scalar-spin chirality can be calculated based on the spin structure alone (cf. Eq. (3)). Therefore, we use the lattice site with the largest contribution to determine , in order to reduce numerical errors. The constant obtained in this way for Fe/Ir-3/Re(0001) is shown in Table I and has been used to obtain the TOM in the model (Figure 7(a)). Note, that the value of depends slightly on the considered spin state of the Fe monolayer.
Figure 7(b) shows the topological orbital moments for the atomic-scale SkX of Fe/Ir-3/Re(0001) as calculated via DFT by neglecting SOC. Since SOC has been excluded the orbital moments can only have a topological origin. The orbital moments are aligned along the -direction and nearly coincide with the TOM, calculated by the atomistic model (Figure 7(a)). This demonstrates that the atomistic model gives a good description of the TOMs and that the assumption of the spatially constant is justified. The -component of the total orbital moment obtained via DFT upon including SOC is shown in Figure 7(c). Fe/Ir-3/Re(0001) has an easy-axis magnetocrystalline anisotropy, meaning that an alignment parallel to the surface normal is favored [17]. According to Ref. [44] and Figure 3 this should lead to larger orbital moments due to SOC for the magnetic moments, which are aligned with the surface normal. Indeed, we find that the magnetic moments pointing along have the largest orbital moments (dark red and blue hexagons).
The -component of the total orbital moments (Figure 7(c)) are similar to the -components of the topological orbital moments (Figure 7(b)) but exhibit some slight differences, e.g. for the magnetic moments pointing along the negative -direction. The net -component averaged over the whole unit cell (marked area) only changes from per atom without SOC to per atom with SOC. In this case the net orbital moments are nearly unchanged, while locally the orbital moments differ when SOC is included or neglected. Tab. I displays the net orbital moments with and without SOC, i.e. total and topological orbital moments, for different film systems and magnetic states. Here also the topological orbital susceptibilities are given. As seen from the three different states in the Fe/Ir-3/Re(0001) system, the topological orbital susceptibility can slightly differ between different magnetic states in the same system. In this case the variation is on the order of 10%. The reason for this lies in the electronic struc-
| system | state | |||
| Pd/Mn/Re(0001) | 3Q | 0.005 | −0.021 | −0.022 |
| Pd/Mn/Re(0001) | 3Q | 0.005 | −0.021 | −0.020 |
| Pd/Mn/Re(0001) | 3Q | 0.005 | −0.021 | −0.020 |
| Fe/Ir-3/Re(0001) | SkX | −0.021 | 0.022 | 0.021 |
| Fe/Ir-3/Re(0001) | multi-Q | −0.017 | 0.000 | 0.000 |
| Fe/Ir-3/Re(0001) | d-SkX | −0.019 | 0.000 | 0.000 |
| Fe/Ir(111) | SkX-12 | −0.021 | 0.042 | 0.019 |
| Fe/Rh-2/Ir(111) | SkX-19 | −0.022 | 0.041 | 0.023 |
| Fe/Rh/Ir(111) | SkX-27 | −0.023 | 0.037 | 0.021 |
Table 1. Topological orbital susceptibility and -component of total orbital moments. The TOM have been calculated via DFT in the scalar-relativistic approximation for the Mn and Fe atoms, respectively, in the ultrathin film systems denoted in the 1st column for the magnetic state given in the 2nd column. The topological orbital susceptibility is assumed to be constant over the lattice and calculated by comparing the DFT calculated TOM with Eq. (2) for the lattice site with the largest contribution. The TOM averaged over the atoms in the magnetic unit cell is given in the 4th column. The averaged -component of the total orbital moments given in the last column have been calculated with the inclusion of SOC. The spatial resolved and contributions are displayed in Figure 5, Figure 7, Supplementary Fig. S6 and Supplementary Fig. S7. All values are given in units of .
ture, which changes slightly with the magnetic structure.
Other non-coplanar spin states which have been studied for Fe/Ir-3/Re(0001) include a multi-Q and a double SkX state [17]. In these spin states the TOM exhibit an antiferromagnetic order and the net topological orbital magnetization vanishes (see Supplementary Fig. 7). Since the spin moments are compensated and the are nearly aligned with the spin moments, the total orbital moments are also compensated. The vanishing contribution of the total orbital moments shows that also is compensated, when the spin moment is compensated.
For Fe monolayers on the clean and Rh covered Ir(111) surface hexagonal skyrmion lattices have been previously studied [16, 43]. These SkX contain 12, 19, and 27 atoms in the 2D unit cell and are therefore denoted as SkX-12, SkX-19, and SkX-27 states (see Supplementary Fig. 8 for sketches of the spin structures). They are constructed in a similar way as the 3Q state by a superposition of three 1Q states. However, in contrast to the 3Q state, here the period of the superimposed spin spirals is significantly larger ranging from 8.1 nm to 12.2 nm. Most importantly – in contrast to the spin structures discussed so far – these states are not spin compensated and therefore possess a non-vanishing net spin moment. Consequently, also the orbital moment is not compensated if we sum over all atoms in the unit cell. Therefore, in these skyrmion lattices, large differences occur in the net contribution of the topological vs. the total orbital moment (see table I). In all spin compensated lattices, on the other hand, is also nearly compensated and there is basically no difference between the averaged topological vs. total orbital moment (see first six lines in table I). This makes the TOM in spin compensated systems still measurable, despite the effect SOC.
In general, Fe monolayer based material systems seem to have a roughly four times larger topological orbital susceptibility of than the Mn based system Pd/Mn/Re(0001) with . Note also the opposite sign between the two elements.
Conclusion
In this work, we have investigated via electronic structure theory based on DFT three different magnetic structures with a net zero spin moment: the RW-AFM state, the 3Q state, and the SkX state. Due to its collinearity the RW-AFM state exhibits no topological orbital moments, while the 3Q and the SkX state are non-coplanar spin structures and possess homogeneous and non-homogeneous topological orbital moments, respectively. We have calculated via DFT the topological and spin-orbit coupling (SOC) induced contributions to the orbital moments in these spin structures.
From the calculations for the RW-AFM state, we conclude that follows roughly the direction of the spin moments. The angle between both moments depends mainly on the direction of the spin moment and can be as large as . The norm of scales with the polar angle , but also with the azimuthal angle . The variation of with respect to is thereby smaller than with respect to . A rotation in does not have to conserve the absolute value of . The difference originates here from the alignment of the spin moment with the connection vectors to the substrate (Re) or adlayer (Pd).
For the 3Q state, we observe that the topological orbital moment calculated in the scalar-relativistic approximation, i.e. neglecting SOC, is a very good approximation for the topological orbital moment obtained including SOC. Apparently, the SOC-induced changes of the electronic structure affect the topological orbital moment only in a negligible way. The contribution differs for the atoms in the magnetic unit cell due to the different orientations of the spin moment. The dependence of the orbital moment on the spin direction is very similar to that observed for the RW-AFM state. Although the components for the individual lattice sites differ, the total component averaged over the unit cell is very similar, when only or when and are considered.
Comparison between the atomistic spin model and DFT calculations for SkX states shows that is nearly constant over the lattice. This behavior is expected, since the electronic structure at the individual lattice sites varies only slightly. The comparison between different materials shows that Fe-monolayer based systems seem to have a larger , than Mn-monolayer based systems.
Key findings of our work are that the magnitude of the SOC-induced orbital moment depends on the direction of the spin moment and that for spin compensated structures, the SOC induced orbital moments are also compensated over the unit cell. The magnitude of the TOMs is nearly not influenced by SOC effects. For spin compensated magnetic structures, the topological orbital moment is nearly identical to the total orbital moment, when averaged over the unit cell. In this respect, spin-compensated states are much better suited for experimental studies than spin-uncompensated states which exhibit in addition to the topological part also a net orbital moment from SOC. On the atomic scale, however, the total orbital moments differ from the topological orbital moments by the SOC-induced orbital moments. This variation of the orbital moments will also be reflected in the local electronic structure and may therefore be detected in STM experiments.
It is our pleasure to thank Yuriy Mokrousov for insightful discussions. We acknowledge financial support by the Deutsche Forschungsgemeinschaft (DFG) via project no. 555842692 and computing time made available to them on the high-performance computer ”Lise” at the NHR center NHR@ZIB. This center is jointly supported by the Federal Ministry of Education and Research and the state governments participating in the NHR (www.nhr-verein.de).
References
- [1]A. Fert, V. Cros, and J. Sampaio, Skyrmions on the track, Nat. Nano. 8, 152 (2013).DOI
- [2]N. Nagaosa and Y. Tokura, Topological properties and dynamics of magnetic skyrmions, Nat. Nanotechnol. 8, 899 (2013).DOI
- [3]S. Koraltan, C. Abert, M. Albrecht, M. Azhar, C. Back, H. Béa, M. T. Birch, S. Blügel, O. Boulle, F. Büttner, P. Che, V. Cros, E. Darwin, L. Desplat, C. Donnelly, H. Du, K. Everschor-Sitte, A. Fernández-Pacheco, S. Finizio, G. Finocchio, M. Garst, R. Gruber, D. Grundler, S. Hayami, T. Hesjedal, A. Hoffmann, A. Hrabec, H. J. Hug, H. Jani, J. Jena, W. Jiang, J. Junquera, K. Karube, L.-M. Kern, J.-V. Kim, M. Kläui, H. Kurebayashi, K. Litzius, Y. Liu, M. Lonsky, C. H. Marrows, J. Masell, S. Mathias, Y. Mokrousov, S. S. P. Parkin, B. Pfau, P. G. Radaelli, F. Radu, R. Ramesh, N. Reyren, S. Rohart, S. Seki, I. I. Smalyukh, S. Sorn, D. Steil, D. Suess, M. Tasinkevych, Y. Tokura, R. Tomasello, V. Ukleev, H. Yang, F. S. Yasin, X. Yu, C. Zhang, S. Zhang, L. Zhao, and S. Wintz, The 2026 skyrmionics roadmap (2026), arXiv:2601.16575 [cond-mat.mes-hall].arxiv.org/abs/2601.16575
- [4]B. H. Rimmler, B. Pal, and S. S. P. Parkin, Non-collinear antiferromagnetic spintronics, Nat. Rev. Mater. 10, 109 (2025).DOI
- [5]P. Kurz, G. Bihlmayer, K. Hirai, and S. Blügel, Three-dimensional spin structure on a two-dimensional lattice: Mn/Cu(111), Phys. Rev. Lett. 86, 1106 (2001).pubmed.ncbi.nlm.nih.gov/11178021
- [6]Y. Taguchi, Y. Oohara, H. Yoshizawa, N. Nagaosa, and Y. Tokura, Spin chirality, Berry phase, and anomalous Hall effect in a frustrated ferromagnet, Science 291, 2573 (2001).pubmed.ncbi.nlm.nih.gov/11283363
- [7]I. Martin and C. D. Batista, Itinerant electron-driven chiral magnetic ordering and spontaneous quantum Hall effect in triangular lattice models, Phys. Rev. Lett. 101, 156402 (2008).
- [8]S. Heinze, K. von Bergmann, M. Menzel, J. Brede, A. Kubetzka, R. Wiesendanger, G. Bihlmayer, and S. Blügel, Spontaneous atomic-scale magnetic skyrmion lattice in two dimensions, Nat. Phys. 7, 713 (2011).
- [9]S. Nakosai, Y. Tanaka, and N. Nagaosa, Two-dimensional p-wave superconducting states with magnetic moments on a conventional s-wave superconductor, Phys. Rev. B 88, 180503 (2013).DOI
- [10]M. Hoffmann, J. Weischenberg, B. Dupé, F. Freimuth, P. Ferriani, Y. Mokrousov, and S. Heinze, Topological orbital magnetization and emergent hall effect of an atomic-scale spin lattice at a surface, Phys. Rev. B 92, 020401(R) (2015).
- [11]J.-P. Hanke, F. Freimuth, A. K. Nandy, H. Zhang, S. Blügel, and Y. Mokrousov, Role of Berry phase theory for describing orbital magnetism: From magnetic heterostructures to topological orbital ferromagnets, Phys. Rev. B 94, 121114(R) (2016).
- [12]A. K. Nandy, N. S. Kiselev, and S. Blügel, Interlayer exchange coupling: A general scheme turning chiral magnets into magnetic multilayers carrying atomic-scale skyrmions, Phys. Rev. Lett. 116, 177202 (2016).
- [13]M. dos Santos Dias, J. Bouaziz, M. Bouhassoune, S. Blügel, and S. Lounis, Chirality-driven orbital magnetic moments as a new probe for topological magnetic structures, Nat. Commun. 7, 13613 (2016).DOI
- [14]G. Grytsiuk, J.-P. Hanke, M. Hoffmann, J. Bouaziz, O. Gomonay, G. Bihlmayer, S. Lounis, Y. Mokrousov, and S. Blügel, Topological–chiral magnetic interactions driven by emergent orbital magnetism, Nat. Commun. 11, 51 (2020).DOI
- [15]J. Spethmann, S. Meyer, K. von Bergmann, R. Wiesendanger, S. Heinze, and A. Kubetzka, Discovery of magnetic single- and triple-Q states in Mn/Re(0001), Phys. Rev. Lett. 124, 227203 (2020).DOI
- [16]M. Gutzeit, T. Drevelow, M. A. Goerzen, S. Haldar, and S. Heinze, Spontaneous square versus hexagonal nanoscale skyrmion lattices in Fe/Ir(111), Phys. Rev. B 108, L060405 (2023).DOI
- [17]F. Nickel, A. Kubetzka, M. Gutzeit, R. Wiesendanger, K. von Bergmann, and S. Heinze, Antiferromagnetic order of topological orbital moments in atomic-scale skyrmion lattices, npj Spintronics 3, 7 (2025).DOI
- [18]D. Go, D. Jo, H.-W. Lee, M. Kläui, and Y. Mokrousov, Orbitronics: Orbital currents in solids, Europhys. Lett. 135, 37001 (2021).
- [19]J.-P. Hanke, F. Freimuth, S. Blügel, and Y. Mokrousov, Prototypical topological orbital ferromagnet γ-FeMn, Sci. Rep. 7, 41078 (2017).DOI
- [20]F. R. Lux, F. Freimuth, S. Blügel, and Y. Mokrousov, Engineering chiral and topological orbital magnetism of domain walls and skyrmions, Commun. Phys. 1, 60 (2018).DOI
- [21]S. Mankovsky, S. Polesya, and H. Ebert, Extension of the standard Heisenberg hamiltonian to multipspin exchange interactions, Phys. Rev. B 101, 174401 (2020).DOI
- [22]S. Mankovsky, S. Polesya, and H. Ebert, Topologically driven three-spin chiral exchange interactions treated from first principles, Phys. Rev. B 104, 054418 (2021).DOI
- [23]T. Momoi, K. Kubo, and K. Niki, Possible chiral phase transition in two-dimensional solid 3He, Phys. Rev. Lett. 79, 2081 (1997).DOI
- [24]S. Haldar, S. Meyer, A. Kubetzka, and S. Heinze, Distorted 3q state driven by topological-chiral magnetic interactions, Phys. Rev. B 104, L180404 (2021).DOI
- [25]F. Nickel, A. Kubetzka, S. Haldar, R. Wiesendanger, S. Heinze, and K. von Bergmann, Coupling of the triple-q state to the atomic lattice by anisotropic symmetric exchange, Phys. Rev. B 108, L180411 (2023).DOI
- [26]G. Tatara and H. Kohno, Permanent current from noncommutative spin algebra, Phys. Rev. B 67, 113316 (2003).DOI
- [27]H. Takagi, R. Takagi, S. Minami, T. Nomoto, K. Ohishi, M.-T. Suzuki, Y. Yanagi, M. Hirayama, N. D. Khanh, K. Karube, H. Saito, D. Hashizume, R. Kiyanagi, Y. Tokura, R. Arita, T. Nakajima, and S. Seki, Spontaneous topological Hall effect induced by non-coplanar antiferromagnetic order in intercalated van der Waals materials, Nat. Phys. 19, 961 (2023).DOI
- [28]P. Park, W. Cho, C. Kim, Y. An, Y.-G. Kang, M. Avdeev, R. Sibille, K. Iida, R. Kajimoto, K. H. Lee, W. Ju, E.-J. Cho, H.-J. Noh, M. J. Han, S.-S. Zhang, C. D. Batista, and J.-G. Park, Tetrahedral triple-q magnetic ordering and large spontaneous Hall conductivity in the metallic triangular antiferromagnet Co1/3TaS2, Nat. Commun. 14, 8346 (2023).DOI
- [29]N. J. Ghimire, A. S. Botana, J. S. Jiang, J. Zhang, Y.-S. Chen, and J. F. Mitchell, Large anomalous hall effect in the chiral-lattice antiferromagnet CoNb3S6, Nat. Commun. 9, 3280 (2018).DOI
- [30]N. D. Khanh, S. Minami, M. M. Hirschmann, T. Nomoto, M.-C. Jiang, R. Yamada, N. Heinsdorf, D. Yamaguchi, Y. Hayashi, Y. Okamura, H. Watanabe, G.-Y. Guo, Y. Takahashi, S. Seki, Y. Taguchi, Y. Tokura, R. Arita, and M. Hirschberger, Gapped nodal planes and large topological nernst effect in the chiral lattice antiferromagnet CoNb3S6, Nat. Commun. 16, 2654 (2025).DOI
- [31]N. Romming, A. Kubetzka, C. Hanneken, K. von Bergmann, and R. Wiesendanger, Field-Dependent Size and Shape of Single Magnetic Skyrmions, Phys. Rev. Lett. 114, 177203 (2015).DOI
- [32]S. Meyer, M. Perini, S. von Malottki, A. Kubetzka, R. Wiesendanger, K. von Bergmann, and S. Heinze, Isolated zero field sub−10 nm skyrmions in ultrathin Co films, Nat. Commun. 10, 3823 (2019).
- [33]D. Wortmann, G. Michalicek, N. Baadji, M. Betzinger, G. Bihlmayer, J. Bröder, T. Burnus, J. Enkovaara, F. Freimuth, C. Friedrich, C.-R. Gerhorst, S. Granberg Cauchi, U. Grytsiuk, A. Hanke, J.-P. Hanke, M. Heide, S. Heinze, R. Hilgers, H. Janssen, D. A. Klüppelberg, R. Kovacik, P. Kurz, M. Lezaic, G. K. H. Madsen, Y. Mokrousov, A. Neukirchen, M. Redies, S. Rost, M. Schilf, A. Schindlmayr, M. Winkelmann, and S. Blügel, FLEUR, Zenodo (2023).
- [34]The FLEUR project, https://www.flapw.de/.
- [35]E. Wimmer, H. Krakauer, M. Weinert, and A. J. Freeman, Full-potential self-consistent linearized-augmented-plane-wave method for calculating the electronic structure of molecules and surfaces: O₂ molecule, Phys. Rev. B 24, 864 (1981).DOI
- [36]D. J. Singh and L. Nordstrom, Planewaves, Pseudopotentials and the LAPW Method, Second Edition (, Berlin, Germany, 2005).
- [37]J. P. Perdew, K. Burke, and M. Ernzerhof, Generalized gradient approximation made simple, Phys. Rev. Lett. 77, 3865 (1996).pubmed.ncbi.nlm.nih.gov/10062328
- [38]S. H. Vosko, L. Wilk, and M. Nusair, Accurate spin-dependent electron liquid correlation energies for local spin density calculations: a critical analysis, Can. J. Phys. 58, 1200 (1980).DOI
- [39]VASP, https://www.vasp.at.
- [40]G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B 54, 11169 (1996).pubmed.ncbi.nlm.nih.gov/9984901
- [41]P. E. Blöchl, Projector augmented-wave method, Phys. Rev. B 50, 17953 (1994).pubmed.ncbi.nlm.nih.gov/9976227
- [42]G. Kresse and D. Joubert, From ultrasoft pseudopotentials to the projector augmented-wave method, Phys. Rev. B 59, 1758 (1999).DOI
- [43]M. Gutzeit, A. Kubetzka, S. Haldar, H. Pralow, M. Goerzen, R. Wiesendanger, S. Heinze, and K. von Bergmann, Nano-scale collinear multi-q states driven by higher-order interactions, Nat. Commun. 13, 5764 (2022).DOI
- [44]P. Bruno, Tight-binding approach to the orbital magnetic moment and magnetocrystalline anisotropy of transition-metal monolayers, Phys. Rev. B 39, 865 (1989).pubmed.ncbi.nlm.nih.gov/9947253
- [45]Note that when the spin moments of the FM state, instead of the RW-AFM state, are rotated with φ, the two-site ASE has no contribution and also the variation of the orbital moment is negligible (see Supplementary Figure 2 and Supplementary Note 2).
- [46]M. Bode, S. Heinze, A. Kubetzka, O. Pietzsch, X. Nie, G. Bihlmayer, S. Blügel, and R. Wiesendanger, Magnetization-direction-dependent local electronic structure probed by scanning tunneling spectroscopy, Phys. Rev. Lett. 89, 237205 (2002).pubmed.ncbi.nlm.nih.gov/12485038
- [47]K. von Bergmann, M. Menzel, D. Serrate, Y. Yoshida, S. Schröder, P. Ferrari, A. Kubetzka, R. Wiesendanger, and S. Heinze, Tunneling anisotropic magnetoresistance on the atomic scale, Phys. Rev. B 86, 134422 (2012).DOI