Antiferromagnetic Domain Walls

• Antiferromagnetic domain walls are localized interfaces separating regions with distinct Néel order, characterized by sharp, symmetry-dictated spin rotations and atomic-scale widths.
• Their energetics and dynamics are modeled using micromagnetic theory and ab initio simulations, revealing ultrafast, spin-torque driven motion and strain-tuned confinement.
• Advanced imaging techniques like XMLD-PEEM, aberration-corrected STEM, and NV magnetometry elucidate defect-induced pinning and interfacial engineering for next-generation spintronic devices.

Antiferromagnetic domain walls (AFM DWs) are spatially localized, topologically stable interfaces between domains of distinct Néel order in antiferromagnetic materials. In contrast to ferromagnetic domain walls, AFM DWs possess no net magnetization in equilibrium and are characterized by sharp, symmetry-dictated rotations of the staggered order parameter over nanometer to atomic length scales. Their microscopic spin textures, energetics, interaction with defects and external fields, and ultrafast dynamics distinguish them as central objects in spintronics and topological condensed matter physics.

1. Structural Varieties and Microscopic Textures

The microscopic structure of an AFM DW is determined by the crystallography, magnetic symmetry, and defect landscape of the host material. In collinear antiferromagnets such as NiO and Cr₂O₃, 180° domain walls separate domains of opposite Néel vector. In NiO, below the Néel temperature, the crystal contracts along one of four equivalent ⟨111⟩ axes, producing four T-domains; the Néel vector lies in the corresponding contracted {111} plane. In thin NiO/Pt films on MgO(001), 180° T-domain walls are circular and exhibit a rotation path wherein the local spin orientation at the wall points along the normalized vector sum of the adjoining domains, i.e., $$\mathbf{n}(s) = (\mathbf{n}_A + \mathbf{n}_B)/|\mathbf{n}_A + \mathbf{n}_B|$$, not along the common axis of the two planes. This “average-direction” rotation minimizes exchange energy costs and leads to a strongly confined, non-chiral wall structure (Schmitt et al., 2022).

In itinerant systems such as chromium and monolayer Mn films, DWs can connect domains differing by 90° in spin-density-wave propagation vector (Cr) or by discrete 60°, 120° rotations in the in-plane AFM “row-wise” texture (Mn). Notably, in Mn/Re(0001), the domain wall interpolates between 1Q ground states via a central 2Q region characterized by transient 90° spin angles (a “2Q” state), with a width of ~2 nm set by higher-order exchange (Spethmann et al., 2020).

In defect-rich systems such as CuMnAs, nanoscale crystalline twins create lines of locally rotated easy axes, pinning and orienting AFM domain walls. The presence of such defects can confine 180° or 90° walls to mesoscale regions, modulating wall width and allowing for robust, geometry-imposed pinning landscapes (Reimers et al., 2021).

AFM domain walls can also organize on atomic scales. In bulk CuMnAs, aberration-corrected STEM has revealed domain walls of width ≤ 1 unit cell (~4 Å), residing outside the continuum-micromagnetic paradigm and stabilized by quantum sublattice-specific energetics (Krizek et al., 2020).

2. Energetics, Anisotropy, and Micromagnetic Models

Classical micromagnetic theory for AFM DWs is based on the energy functional

$$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$

where $A$ is exchange stiffness, $K$ is magnetocrystalline anisotropy, and $w_{\text{me}}$ represents magnetoelastic contributions. The canonical wall profile for uniaxial, collinear AFMs is

$$\theta(x) = 2 \arctan \left[ \exp \left( \frac{x}{\delta} \right) \right], \quad \delta = \pi\sqrt{A/K}$$

yielding a wall energy per unit area $\sigma = 4\sqrt{A K}$ (Hedrich et al., 2020, Krizek et al., 2020, Schmitt et al., 2022). The presence of magnetoelastic coupling or local strain $\varepsilon$ modifies $K$ to an effective anisotropy $K_{\mathrm{eff}} = K + B\varepsilon$, thereby tuning $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$0.

In thin-film AFMs, substrate-induced strain can dominate over crystalline anisotropy, as in NiO/Pt on MgO(001), where biaxial in-plane strain shifts the easy axes and enables circular domain formation with reduced wall widths. Detailed XMLD-PEEM measurements determine domain wall widths as $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$1 nm, much less than the expected hundreds of nm in bulk, demonstrating strong strain-induced confinement (Schmitt et al., 2022).

Atomistic simulations and relativistic ab initio calculations in CuMnAs reveal that energetic minima can occur for atomically abrupt walls—the wall energy becomes non-monotonic in width, in contrast to classical micromagnetism, due to compensations in the sublattice quantum structure (Krizek et al., 2020).

3. Defects, Pinning, and Interfacial Engineering

Crystalline defects and engineered heterostructures play pivotal roles in dictating domain wall existence, orientation, and mobility. In Cr thin films capped by Fe, interfacial steps generate magnetic frustration, constraining the spin-density-wave (SDW) propagation vector $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$2 to be in-plane. Selective removal of the Fe/Au cap thus creates sharp interfaces where $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$3 changes by 90°, forming lithographically defined AFM domain walls. The wall width, $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$4, is empirically determined to be $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$5 in these systems (Logan et al., 2012).

In CuMnAs thin films, high-density microtwin defects set local easy-axis orientations and pin walls along twin boundaries, imposing both geometric and micromagnetic constraints. The critical strain for wall nucleation is linked to $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$6, and typical twin-induced strains far exceed this threshold, ensuring robust wall pinning (Reimers et al., 2021).

In multiferroics, periodic ferroelectric domain structures create a periodic potential for the AFM walls, with energy

$$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$7

where $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$8 is the ferroelectric domain half-period. The resulting pinning landscape creates finite coercivity and periodic equilibrium positions for the wall center (Gareeva et al., 2010).

4. Dynamics: Spin Torques, Spin Waves, and Relativistic Effects

AFM DWs exhibit ultrafast dynamics governed by coupled equations for the Néel order and net magnetization, often reducing in the collective coordinate approach to effective equations of motion for the wall center $$E[\mathbf{n}] = \int d^3 r \left\{ A|\nabla \mathbf{n}|^2 + K f_{\text{an}}[\mathbf{n}] + w_{\text{me}}[\mathbf{n}, \varepsilon] \right\}$$9. In antiferromagnet/heavy metal bilayers, spin–orbit torques (SOTs) can propel AFM DWs at velocities limited only by the AFM magnon group velocity ($A$0–$A$1 km/s), with the wall width Lorentz-contracting as

$A$2

(Shiino et al., 2016). As $A$3, walls emit terahertz spin-waves, establishing a mechanism for high-frequency electrically tunable AFM oscillators.

Under spin-polarized currents, the adiabatic component of spin-transfer torque is symmetry-suppressed in AFMs, making the non-adiabatic torque dominant, with its magnitude inversely related to wall width. For narrow DWs, this torque is strongly enhanced, resulting in velocities much larger than those achievable in ferromagnets (Park et al., 2020).

Spin-wave-driven motion can be achieved via magnonic torque, particularly when the wall is subjected to polarized spin waves with polarization-selectivity enforced by Dzyaloshinskii-Moriya interaction (DMI). The direction and velocity of the wall can be precisely controlled by the polarization of incident spin waves, enabling magnonic gating of walls for memory and logic applications (Yu et al., 2017, Seyler et al., 10 Nov 2025).

For AFMs under a thermal gradient and applied magnetic field, domain walls experience competing entropic and magnonic-reflection forces, with the total force

$A$4

where $A$5 is the net wall moment induced by field canting; the direction of wall motion can be reversed by tuning $A$6 (Yanes et al., 2020).

5. Topological and Symmetry Aspects

AFM domain walls can host topologically protected electronic and magnonic excitations. In uniaxial A-type AFM topological insulators, 180° domain walls produce a $A$7 kink of the axion angle $A$8, leading to 1D chiral bound states predicted by topological field theory (Sass et al., 2019, Naselli et al., 15 May 2025). In dual topological insulators, the character of wall-bound states is dictated by the symmetry (spinful or spinless mirror) of the crystalline phase: bulk-gapped walls can host chiral edge states at terminations on ferromagnetic surfaces or, in certain symmetry classes, embedded 2D Dirac semimetal states. This opens design routes for engineering 0D–2D topological channels controllable by domain structure (Naselli et al., 15 May 2025).

In non-coplanar antiferromagnets such as Nd₂Ir₂O₇, domain walls intrinsically break symmetries (e.g. twofold rotations) present in the bulk, allowing for an unconventional anomalous Hall effect (AHE) with the Hall conductivity $A$9 localized at the wall plane and not present in the bulk domain (Kim et al., 2018).

6. Experimental Probes and Imaging

Direct imaging of AFM DWs employs a range of techniques that exploit either the weak net moment inside the wall or anisotropy in magneto-optical, electron, or x-ray signals. XMLD-PEEM and scanning x-ray microscopy enable nanometer-scale visualization of the spin axis and wall morphology in systems such as NiO, CuMnAs, and Cr (Schmitt et al., 2022, Reimers et al., 2021, Logan et al., 2012). Cryogenic magnetic force microscopy (MFM) achieves susceptibility-dominated imaging of AFM walls in MnBi₂Te₄ and related topological antiferromagnets, revealing wall widths down to ~10 nm and field-tunable domain architectures (Sass et al., 2019). Aberration-corrected STEM with differential phase contrast enables direct, atomically resolved mapping of wall profiles in CuMnAs, resolving abrupt Néel reversals over a single lattice plane (Krizek et al., 2020). Single-spin NV magnetometry provides full vector field mapping and mechanical response in collinear AFM insulators such as Cr₂O₃ (Hedrich et al., 2020).

7. Functional Properties and Spintronic Applications

AFM DWs set ultimate bounds for bit cell scaling (down to atomic planes), offer immunity from stray magnetic fields, and can be manipulated at velocities on the km/s to tens-of-km/s scale. Strain engineering, defect control, and tailored interfacial anisotropy enable deterministic domain and wall patterning and pinning, as demonstrated in Cr and CuMnAs (Logan et al., 2012, Reimers et al., 2021). In synthetic AFMs with strong perpendicular magnetic anisotropy, vertical domain walls can be nucleated and pinned by controlled surface spin flop transitions, allowing device integration and real-time magnon dynamics tuning (Böhm et al., 2019). AFM DWs under spin-orbit torque or spin-wave drive are attractive for ultrafast memory and logic circuits, including potential implementations of collision-based NOT and XOR gates by harnessing relativistic soliton-like wall scattering (Otxoa et al., 2022). The polarization-selective filtering of magnons by DMI-stabilized walls establishes foundational elements for magnonic logic and information transduction (Lan et al., 2017).

A plausible implication is that continued progress in atomic-scale imaging, interface engineering, and ultrafast magnetization control will further expand the functional role of antiferromagnetic domain walls in next-generation spintronic and quantum devices.