Domain Wall Bound Interfaces

• Domain wall bound refers to interfaces that separate distinct phases and host localized modes or constraints, unifying phenomena in varied physical systems.
• It encompasses both trapped eigenmodes in potential wells and dynamically bound pairs of walls, with theory and experiments confirming field-specific behaviors.
• The concept underpins applications ranging from magnonic waveguides and topological channels to cosmological bounds and EFT constraints in modern research.

“Domain wall bound” is a multivalent technical expression used for several non-equivalent constructions centered on a codimension-one interface. In condensed-matter, photonic, and field-theoretic settings, it often denotes a state localized transverse to the wall and propagating or residing along it; in coupled-interface dynamics it can instead denote a pair of walls that move together as a composite object; in cosmology and string theory it can denote an inequality derived from domain wall physics rather than a localized state. Across these uses, the common structure is an interface separating distinct asymptotic phases, vacua, or order-parameter sectors, together with either a localized mode or a quantitative constraint tied to that interface (Metaxas et al., 2010, Lee-Thorp et al., 2016, Lazanu et al., 2015, Cribiori et al., 9 Mar 2026).

1. Terminological scope and basic definitions

A domain wall is an interface interpolating between distinct configurations. In ultrathin ferromagnets it is a one-dimensional elastic interface moving through a weakly disordered two-dimensional medium; in periodic dielectrics it is a heterojunction between two asymptotic periodic phases; in relativistic scalar theories it is a kink interpolating between degenerate vacua; in flux compactifications it is a flux-changing wrapped brane between anti-de Sitter vacua (Metaxas et al., 2010, Lee-Thorp et al., 2016, Morris, 2016, Cribiori et al., 9 Mar 2026).

The adjective “bound” then acquires three principal meanings. First, it can denote a localized eigenmode. Representative examples include scalar bosons trapped by a $-\operatorname{sech}^2$ potential in a wall core, guided TM Maxwell modes created by a sign-changing Dirac mass, magnons and polar waves localized across a wall but dispersing along it, and chiral or helical electronic channels confined to magnetic or pseudo-spin domain walls (Morris, 2016, Lee-Thorp et al., 2016, Wang et al., 2015, Chen et al., 2021, Tiwari et al., 2017, Chakraborty et al., 2024). Second, it can denote a dynamically locked pair of walls, as in coupled ferromagnetic layers where two interfaces in different media move at a common velocity over finite field windows (Metaxas et al., 2010). Third, it can denote a constraint derived from domain wall physics, such as the CMB upper bound on the surface tension of cosmological domain walls or the anti-de Sitter “domain wall bound” relating $L_{\rm AdS}$ to the EFT cutoff (Lazanu et al., 2015, Cribiori et al., 9 Mar 2026).

This semantic multiplicity is itself significant. It implies that “domain wall bound” is not a single universal phenomenon but a class of interface-localized or interface-constrained structures whose mathematical realization depends on the underlying spectral, dynamical, or EFT problem.

2. Dynamically bound walls and composite interface motion

In coupled ultrathin ferromagnetic layers, domain walls can bind to each other dynamically rather than merely statically. The experimentally studied Pt/Co/Pt-based multilayer contains a hard $0.8\,\mathrm{nm}$ Co layer and a soft $0.5\,\mathrm{nm}$ Co layer, with ferromagnetic interlayer coupling $J$. The isolated wall velocities $v_h(H)$ and $v_s(H)$ differ strongly, yet the coupled system exhibits two field ranges in which the two walls move at a common velocity: a low-field bound regime near $H=0$, and a high-field bound regime around the second crossing $H^*\approx 870~\mathrm{Oe}$. The low-field regime extends approximately over $0<H<254~\mathrm{Oe}$, while the one-dimensional theory predicts a high-field bound window $L_{\rm AdS}$0 to $L_{\rm AdS}$1 (Metaxas et al., 2010).

The minimal theory treats only the mean wall positions $L_{\rm AdS}$2 and the separation $L_{\rm AdS}$3. The coupling acts as opposite effective-field shifts on the two walls,

$L_{\rm AdS}$4

with experimentally determined coupling fields

$L_{\rm AdS}$5

A moving bound state exists when the instantaneous velocities match at some separation,

$L_{\rm AdS}$6

so that $L_{\rm AdS}$7 remains constant and the pair propagates with a common velocity $L_{\rm AdS}$8 (Politi et al., 2011).

At low fields both isolated walls are in the thermally activated creep regime,

$L_{\rm AdS}$9

and the bound state itself also obeys a creep law,

$0.8\,\mathrm{nm}$0

For the measured parameters of the hard and soft layers, the theory yields $0.8\,\mathrm{nm}$1 and $0.8\,\mathrm{nm}$2, in good agreement with the observed low-field bound motion (Politi et al., 2011).

A distinct wall-wall binding problem appears in cylindrical nanowires. In parallel Ni nanowires, a domain wall pinned at a radial constriction creates an attractive magnetostatic potential well for a free transverse wall in a neighboring wire, and trapped bound states appear above the depinning threshold; surface roughness facilitates these trapped bound states (Dolocan, 2015). In a single cylindrical nanowire, pairs of transverse walls can form metastable oscillatory bound states with reported lifetimes of about $0.8\,\mathrm{nm}$3 and $0.8\,\mathrm{nm}$4, stabilized dynamically by wall precession and, during transport, by spin-polarized current (Dolocan, 2013). These examples show that “bound” may refer not to a single-wall eigenmode but to a composite many-wall state sustained by coupling, precession, or dynamical recapture.

3. Wall-guided excitations in bosonic, magnonic, and ferroic media

A large class of domain-wall-bound problems reduces to a one-dimensional Schrödinger-type equation with an attractive potential generated by the wall profile. In a two-field relativistic scalar model, a topological wall in $0.8\,\mathrm{nm}$5 creates for a complex scalar $0.8\,\mathrm{nm}$6 the potential

$0.8\,\mathrm{nm}$7

leading to a Pöschl–Teller problem with discrete transverse levels

$0.8\,\mathrm{nm}$8

Because $0.8\,\mathrm{nm}$9 remains continuous along the wall, the paper characterizes the spectrum as a “quasi-discretuum”: discrete in the transverse direction, continuous on the wall worldvolume (Morris, 2016).

In insulating ferromagnets, the linearized spin-wave equation around a Bloch wall yields a bound branch

$0.5\,\mathrm{nm}$0

which is gapless and propagates along the wall, while the bulk branch is gapped by

$0.5\,\mathrm{nm}$1

Below $0.5\,\mathrm{nm}$2, only the wall-bound mode exists, so the wall acts as a magnonic waveguide. Micromagnetic simulations further show transmission through Bloch lines and $0.5\,\mathrm{nm}$3 corners without visible reflection in the reported frequency range $0.5\,\mathrm{nm}$4–$0.5\,\mathrm{nm}$5 (Wang et al., 2015). A related but more restrictive result appears in the discrete-lattice heat-transport problem: continuum micromagnetism yields one familiar bound spin-wave mode and no reduction of heat conductance, whereas atomically narrow walls support an additional bound state, produce finite reflection, and reduce magnon heat conductance (Yan et al., 2012).

In ferroelectrics, linearization of the Landau-Khalatnikov-Tani equation around the static Ising wall profile

$0.5\,\mathrm{nm}$6

gives the fluctuation equation with an attractive index-2 Pöschl–Teller potential

$0.5\,\mathrm{nm}$7

Two bound polar-wave modes appear below the bulk continuum: a symmetric vibration mode,

$0.5\,\mathrm{nm}$8

and an antisymmetric breathing mode,

$0.5\,\mathrm{nm}$9

In a two-dimensional film these become guided branches,

$J$0

so the ferroelectric wall functions as a narrow waveguide (Chen et al., 2021).

For magnons on a Skyrmion-textured antiferromagnetic wall, the effective transverse potential is Rosen–Morse rather than Pöschl–Teller. Supersymmetric factorization gives an exact wall-guided mode,

$J$1

with $J$2. The mode is localized in $J$3, disperses along $J$4, and is chirality- and polarization-dependent through the emergent gauge field of the textured wall (Lee et al., 2022).

4. Topological interface states, Majorana channels, and wall-bound textures

In photonics, a domain wall can bind a state through a genuine Dirac mechanism rather than through a conventional band-edge defect. For the domain-wall-modulated Hamiltonian

$J$5

a Dirac point of the periodic operator $J$6 yields the effective one-dimensional Dirac operator

$J$7

Because the mass term $J$8 changes sign, $J$9 has a zero mode, and this lifts to an exponentially localized bound state of the full Maxwell-guided-wave problem. The paper identifies the resulting state as a topologically protected, transversely localized, guided TM mode and contrasts it with ordinary defect modes bifurcating from band edges, which are not protected against localized perturbations (Lee-Thorp et al., 2016).

On magnetic topological-insulator surfaces, a sign-changing Dirac mass at a magnetic domain wall binds a one-dimensional chiral channel. In the model

$v_h(H)$0

with a magnetization-induced mass $v_h(H)$1, local magnetization reversal changes the sign of the Dirac mass and generates a gapless chiral domain-wall bound state. When neighboring walls are separated by more than $v_h(H)$2, each channel contributes one conductance quantum,

$v_h(H)$3

The paper connects the nucleation and growth of such walls during a magnetic-field sweep to butterfly-shaped hysteresis in magnetoconductance (Tiwari et al., 2017).

A more structured version appears in a BHZ platform that undergoes a QSH-to-QAH transition. A composite domain wall in spin and pseudo-spin degrees of freedom binds a helical interface channel with left- and right-movers orthogonal in both spin and pseudo-spin space,

$v_h(H)$4

With superconducting proximity, the zero-energy transport signature is

$v_h(H)$5

and the superconducting-to-insulating transition occurs at $v_h(H)$6 (Chakraborty et al., 2024). This suggests a domain wall can serve not only as a spectral defect but as an engineered one-dimensional topological superconducting channel.

In FeSe, a diagonal nematic domain wall lowers symmetry enough that, in the presence of $v_h(H)$7-preserving SOC, singlet-triplet mixing becomes allowed locally. For a $v_h(H)$8 wall, the Ginzburg–Landau coupling

$v_h(H)$9

induces a $v_s(H)$0-wave component parallel to the wall, so the wall acts as an emergent one-dimensional $v_s(H)$1-wave superconducting wire. In simplified BdG calculations, sufficiently strong wall-localized $v_s(H)$2-wave pairing yields zero-energy states localized at the ends of a finite wall segment and satisfying the Majorana condition (Lee et al., 2017).

In magnetic topological semimetals, domain walls can bind broad electronic interface bands rather than narrow arc-like states. For EuB$v_s(H)$3, the bulk topological semimetal structure depends strongly on magnetization direction; first-principles calculations for experimentally motivated $v_s(H)$4 and $v_s(H)$5 walls show robust domain-wall bound states distributed over a large portion of the wall Brillouin zone, with dispersion of about $v_s(H)$6 eV and localized charge on the order of one electron per primitive cell (Li et al., 2022).

A further topological texture problem concerns skyrmions trapped on walls. In a chiral magnet with easy-axis anisotropy and no Zeeman term, a domain-wall skyrmion is a $v_s(H)$7 skyrmion bound to a domain wall. In the ferromagnetic phase it is described by a kink in the wall phase $v_s(H)$8 together with a geometric cusp in the wall position $v_s(H)$9. The cusp amplitude diverges as the FM–CSL boundary is approached,

$H=0$0

and in the CSL an isolated wall-bound skyrmion decays by reconnection into a pair of merons, while an alternating chain of skyrmions and anti-skyrmions on alternating walls and anti-walls is stable or metastable (Amari et al., 2023).

5. Cosmological and anti-de Sitter bounds derived from domain walls

In cosmology, “domain wall bound” usually denotes an upper bound on the wall surface tension or the symmetry-breaking scale of a long-lived scaling network. Using $H=0$1 simulations, unequal-time correlators, and COSMOMC parameter estimation, one analysis obtained

$H=0$2

corresponding to a wall energy scale of $H=0$3; a radiation+matter-only treatment gave

$H=0$4

corresponding to $H=0$5. The CMB amplitude scales as

$H=0$6

so the observable constraint is fundamentally on $H=0$7, later expressed as a bound on the wall formation scale. The result is described as close but below the Zel’dovich bound of about $H=0$8 (Lazanu et al., 2015).

A complementary CMBACT-based phenomenological study modeled domain wall networks with a velocity-dependent one-scale model and a wall version of the Unconnected Segment Model. From a conservative low-$H=0$9 TT normalization argument it inferred

$H^*\approx 870~\mathrm{Oe}$0

Here again the interpretation is a CMB realization of the Zel’dovich statement that stable walls must form at or below the MeV scale (Sousa et al., 2015).

In string compactifications, the phrase has a different meaning. For anti-de Sitter flux vacua connected by flux-changing domain walls, demanding that the wall be fundamental from the EFT viewpoint leads to

$H^*\approx 870~\mathrm{Oe}$1

Combining this with the curvature–tension inequality derived from ten-dimensional flux quantization yields the anti-de Sitter domain wall bound

$H^*\approx 870~\mathrm{Oe}$2

For supersymmetric AdS vacua this implies

$H^*\approx 870~\mathrm{Oe}$3

The paper reports that classical flux vacua and LVS are compatible with this bound, whereas racetrack and KKLT-like AdS vacua face a non-trivial constraint when attempting very large scale hierarchies (Cribiori et al., 9 Mar 2026).

6. Limits, caveats, and non-universality

A recurring misconception is that any domain wall automatically hosts a physically relevant bound state. The NJL domain-wall analysis gives an explicit counterexample. There the mean-field kink reduces the fermion problem to the Jackiw–Rebbi equation with discrete spectrum

$H^*\approx 870~\mathrm{Oe}$4

But for the specific NJL wall generated in the approximation used, $H^*\approx 870~\mathrm{Oe}$5, exactly the threshold value, so no genuine nonzero bound states occur. Only the zero-energy localized mode remains, and it is interpreted as part of the fermionic vacuum rather than as a physical quark bound state. Higher-order corrections could change the wall profile and allow true bound states, but under the stated assumptions there are none (Kutnii, 2011).

A second caveat is that bound-state phenomenology is highly regime-dependent. In insulating ferromagnetic wires, continuum micromagnetism predicts one familiar wall-bound spin-wave mode and no effect on heat conductance; only when the wall becomes atomically narrow on a discrete lattice does an additional bound state emerge and finite reflection appear (Yan et al., 2012). Likewise, in coupled ultrathin ferromagnets the low-field bound creep state is captured quantitatively by the one-dimensional model, whereas the high-field bound velocity is underestimated; the proposed missing ingredients are full two-dimensional interface elasticity and dipolar fields (Metaxas et al., 2010).

Claims of topological protection are also model-specific. In photonics, the protected branch is tied to the sign change of the effective Dirac mass and persists under arbitrary spatially localized perturbations of the domain wall function $H^*\approx 870~\mathrm{Oe}$6; ordinary defect modes created from band edges do not share this property (Lee-Thorp et al., 2016). In FeSe, by contrast, the Majorana interpretation is suggestive rather than definitive: the zero modes are demonstrated numerically in a simplified one-band model with a large induced wall $H^*\approx 870~\mathrm{Oe}$7-wave component, while realistic multiorbital structure, disorder, wall roughness, and even the actual superconducting symmetry of FeSe remain open issues (Lee et al., 2017).

The literature therefore supports a precise but plural conclusion. “Domain wall bound” may denote a guided mode, a bound pair of interfaces, a topological channel, a meronized decay product, or an EFT inequality. What unifies these uses is not a single universal spectrum or mechanism, but the role of the domain wall as a spatial locus where asymptotic phases meet and new localized dynamics or new consistency conditions become unavoidable.