Skyrmion Lattice Model Overview

• Skyrmion lattice model is a theoretical framework describing periodic arrays of noncoplanar magnetic textures with quantized topological charge in systems like chiral magnets.
• It incorporates diverse formulations ranging from continuum Ginzburg–Landau and CP¹ field theories to discrete lattice-spin Hamiltonians and effective particle models.
• The model connects microscopic interactions—such as exchange, Dzyaloshinskii–Moriya coupling, and anisotropy—to observable phase transitions, melting behaviors, and collective excitations.

A skyrmion lattice model is a theoretical description of a periodic array of topological textures, most commonly a triangular skyrmion crystal in a chiral magnet, but in the current literature also including antiskyrmion, antiferromagnetic, multipolar, and orbital analogues. The model space is not singular: published formulations range from continuum Ginzburg–Landau and CP$^1$ field theories to discrete lattice-spin Hamiltonians, effective point-particle and soft-disk reductions, Brownian and Thiele dynamics, and itinerant-electron models in which RKKY, Kondo, Rashba, or scalar-chirality terms generate the noncollinear order. Across these settings, the recurring problem is to determine when exchange, Dzyaloshinskii–Moriya interaction, scalar chirality, anisotropy, and Zeeman coupling stabilize a periodic topological state and how that state melts, deforms, or acquires collective modes (Han et al., 2010, Buhrandt et al., 2013, Garanin et al., 2024, Bocquet et al., 20 May 2026).

1. Foundational Hamiltonians

The canonical chiral-magnet starting point is a continuum energy for a unit-vector field. In two dimensions, one representative form is

$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$

with ferromagnetic exchange $J$, Dzyaloshinskii–Moriya coupling $D$, and Zeeman field ${\bf B}=B\hat z$. In the CP$^1$ rewrite, the spin field is represented as ${\bf n}=z^\dagger\boldsymbol\sigma z$ with $z^\dagger z=1$, and the resulting mean-field equation can be reduced to an Abrikosov-type problem whose lowest mode generates a triangular skyrmion crystal (Han et al., 2010).

Continuum helimagnet models use the same exchange–DM–Zeeman structure but emphasize symmetry and perturbative departures from circular symmetry. A standard form is

$$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$

for a normalized magnetization field $\mathbf m(\mathbf r)$, with weak-$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$0 perturbation theory predicting systematic asymmetric deformations for both single skyrmions and hexagonal Skyrme crystals (Kundu, 2015).

Discrete lattice-spin Hamiltonians make the microscopic ingredients explicit. In a two-dimensional square lattice of classical spins $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$1, one model is

$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$2

with nearest-neighbor ferromagnetic exchange $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$3, Bloch-type DMI $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$4, perpendicular field $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$5, and uniaxial anisotropy $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$6. For $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$7, $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$8, and $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$9, energy minimization gives a triangular skyrmion lattice of period $J$0 at $J$1 (Garanin et al., 2024).

Three-dimensional chiral magnets are modeled on the cubic lattice by nearest- and next-nearest-neighbor exchange and DM terms together with a Zeeman field,

$J$2

where $J$3 and $J$4 are introduced to cancel leading lattice-discretization anisotropies. For $J$5, the model reproduces a helical pitch of ten sites and yields the finite-temperature A-phase as a triangular lattice of skyrmion tubes (Buhrandt et al., 2013).

2. Reduced and coarse-grained descriptions

A major theme in skyrmion-lattice modeling is reduction from a spin field to effective particles. In a dilute two-dimensional skyrmion crystal, well-separated skyrmions can be treated as point particles interacting through a short-range repulsive potential

$J$6

with skyrmion centers extracted from connected islands of spins satisfying $J$7. In the square-lattice spin model and in the derived point-particle model, the two descriptions produce similar melting behavior (Garanin et al., 2024).

Thin-film multilayers admit an even more aggressive coarse-graining. There each skyrmion is replaced by a “soft disk” of characteristic size $J$8 interacting by a purely repulsive power law,

$J$9

where $D$0 is fixed by the first peak of the experimental pair-correlation function $D$1, and $D$2 is obtained by minimizing a mean-squared deviation $D$3. For the high-density lattice at $D$4, $D$5, the extracted parameters are $D$6, $D$7, and $D$8, while at $D$9, ${\bf B}=B\hat z$0, one finds ${\bf B}=B\hat z$1 (Zázvorka et al., 2020).

Brownian-dynamics and Thiele-type descriptions focus on collective motion rather than equilibrium energetics. In confinement studies, the steady-state equation of motion is

${\bf B}=B\hat z$2

with damping, Magnus force, thermal noise, pair repulsion, and boundary forces. For static ordering, ${\bf B}=B\hat z$3 is set to zero, the pair interaction is approximated by ${\bf B}=B\hat z$4 with cutoff ${\bf B}=B\hat z$5, and the confinement is modeled by a truncated Lennard–Jones repulsion from polygonal or circular boundaries (Gruber et al., 21 Aug 2025).

Rigid-lattice Thiele models also describe collective phasons. For a synthetic antiferromagnet, coupled Thiele equations show that when the interlayer coupling becomes comparable to or larger than the intralayer coupling, the long-wavelength phason branch changes from quadratic to linear, matching the continuum result that the skyrmion-lattice phason becomes a linear gapless mode after cancellation of the Berry term in the Néel-field dynamics (Wang et al., 2022).

3. Order parameters and topological diagnostics

The standard structural diagnostic is sixfold bond-orientational order. In particle or center-coordinate descriptions one defines

${\bf B}=B\hat z$6

For a perfect horizontal hexagon, ${\bf B}=B\hat z$7, while in the high-temperature liquid ${\bf B}=B\hat z$8 fluctuates around ${\bf B}=B\hat z$9, giving $^1$0 and $^1$1. Solid and liquid phases are then distinguished by $^1$2, the radial distribution $^1$3, and Bragg peaks in the structure factor $^1$4 (Garanin et al., 2024).

In Voronoi-based analyses, the local order parameter is written as

$^1$5

with global metrics $^1$6 and $^1$7. A perfect monodomain hexagonal lattice sends both quantities toward unity, whereas a polycrystal can retain a large $^1$8 while suppressing $^1$9 by phase cancellation between domains (Gruber et al., 21 Aug 2025).

Positional and topological information comes from correlation functions and charge formulas. In two-dimensional melting studies, the pair-correlation function ${\bf n}=z^\dagger\boldsymbol\sigma z$0, the hexatic correlator

${\bf n}=z^\dagger\boldsymbol\sigma z$1

and the translational correlator ${\bf n}=z^\dagger\boldsymbol\sigma z$2 separate liquid, hexatic, and solid regimes by exponential versus algebraic decay. In calibrated soft-disk simulations, ${\bf n}=z^\dagger\boldsymbol\sigma z$3 marks the liquid–hexatic crossover, the interval ${\bf n}=z^\dagger\boldsymbol\sigma z$4 is associated with a hexatic phase, and the crystal is identified by algebraic decay in both ${\bf n}=z^\dagger\boldsymbol\sigma z$5 and ${\bf n}=z^\dagger\boldsymbol\sigma z$6 (Zázvorka et al., 2020).

The skyrmion number is usually defined by the continuum integral

${\bf n}=z^\dagger\boldsymbol\sigma z$7

or by a lattice discretization based on signed spherical-triangle areas. In three-dimensional chiral magnets, the projected plaquette charge ${\bf n}=z^\dagger\boldsymbol\sigma z$8 identifies skyrmion tubes in each ${\bf n}=z^\dagger\boldsymbol\sigma z$9-plane; in triangular-lattice RKKY systems, the total scalar chirality $z^\dagger z=1$0 acts as the order parameter of the triple-$z^\dagger z=1$1 SkX, while the staggered chirality $z^\dagger z=1$2 diagnoses a single-$z^\dagger z=1$3 spiral (Buhrandt et al., 2013, Mitsumoto et al., 2021).

4. Thermal phase behavior and melting

The thermal behavior of skyrmion lattices is model dependent rather than universal. In a two-dimensional square-lattice spin model and in its point-particle reduction, the skyrmion lattice exhibits a sharp one-step transition between solid and liquid phases on temperature and magnetic field. At fixed $z^\dagger z=1$4, the transition occurs at $z^\dagger z=1$5, where $z^\dagger z=1$6 and $z^\dagger z=1$7 drop sharply, the Bragg peaks in $z^\dagger z=1$8 vanish, translational order is lost, and the magnetization $z^\dagger z=1$9 shows a clear kink (Garanin et al., 2024).

By contrast, thin-film multilayers mapped to repulsive soft disks exhibit the two-step melting structure associated with Kosterlitz–Thouless–Halperin–Nelson–Young theory. In the $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$0 plane, the liquid–hexatic and hexatic–solid boundaries are approximately

$$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$1

with a small region of two-phase liquid–hexatic coexistence when $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$2. This directly connects skyrmion lattices in multilayers to the soft-disk melting problem (Zázvorka et al., 2020).

Three-dimensional chiral magnets add a further layer of fluctuation physics. Classical Monte Carlo on the cubic lattice shows that the skyrmion-lattice A-phase occupies a small pocket $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$3 and $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$4, with $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$5. The A-phase vanishes as $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$6 and as $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$7, and its stabilization is attributed to thermal fluctuations beyond Gaussian order, in qualitative agreement with the Brazovskii scenario for the helical–paramagnetic transition (Buhrandt et al., 2013).

A third pattern appears in the $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$8-symmetric triangular-lattice SkX-2 model with scalar chirality. There the finite-temperature behavior depends on charge density: the lattice-locked SkX-2 shows a first-order transition, while the free SkX-2 exhibits a sharp, size-independent peak in $$\mathcal{H}[\mathbf{m}] = J\!\int\!d^2r\,(\nabla\mathbf{m})^2 + D\!\int\!d^2r\,\mathbf{m}\!\cdot\!(\nabla\times\mathbf{m}) -\mu_0 M_s H\!\int\!d^2r\,m^3,$$9, a crossover of spin–spin correlations from exponential to algebraic, and proliferation or binding of $\mathbf m(\mathbf r)$0–$\mathbf m(\mathbf r)$1 dislocation pairs consistent with a single Kosterlitz–Thouless–Halperin–Nelson–Young-type melting to a floating solid (Bocquet et al., 20 May 2026).

Field gradients generate coexistence phenomena. When a static gradient $\mathbf m(\mathbf r)$2 is applied to a two-dimensional skyrmion system, the local melting temperature becomes $\mathbf m(\mathbf r)$3, and for $\mathbf m(\mathbf r)$4 a stationary solid–liquid interface appears near $\mathbf m(\mathbf r)$5 such that $\mathbf m(\mathbf r)$6. In the point-particle model, $\mathbf m(\mathbf r)$7 and $\mathbf m(\mathbf r)$8 fall from $\mathbf m(\mathbf r)$9 to $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$00 over a width of order a few $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$01 (Garanin et al., 2024).

5. Itinerant, frustrated, and zero-field stabilization mechanisms

Not all skyrmion-lattice models rely on explicit Dzyaloshinskii–Moriya interactions. In the isotropic RKKY Heisenberg model on the triangular lattice,

$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$02

the oscillating long-range coupling $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$03 produces frustration, while thermal fluctuations generate effective many-body terms that favor multi-$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$04 superpositions. For the canonical choice $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$05, Monte Carlo finds a field-induced chiral-degenerate triple-$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$06 SkX, and replica-symmetry breaking—reported in a related three-dimensional RKKY model—is absent in the strictly two-dimensional case (Mitsumoto et al., 2021).

In the triangular Kondo-lattice model, the skyrmion-lattice parameter is tied to the electronic structure. In the RKKY limit,

$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$07

and varying the conduction-electron filling generates a continuous sequence of SkX states with different magnetic-unit-cell sizes. The four-sublattice chiral antiferromagnetic order appears as the dense $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$08 limit of this sequence, and intermediate coupling stabilizes both $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$09 and field-induced $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$10 skyrmion crystals (Wang et al., 2021).

A related large-$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$11 Kondo-lattice treatment on the triangular lattice shows a different mechanism: conduction electrons moving in the skyrmion background acquire a spin-texture-dependent Peierls phase in the effective hopping $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$12. At electron density $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$13, with $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$14, $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$15, and $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$16, the system exhibits an abrupt jump from $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$17 to $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$18 per sublattice at $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$19. The high-skyrmion-number phase coincides with a downward shift of the band bottom and a deep depletion of the density of states at the Fermi level (Reja, 2020).

Zero-field or weak-DMI skyrmion lattices can also arise from competing exchange terms. On a square lattice with nearest-neighbor bilinear ferromagnetic exchange $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$20, biquadratic exchange $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$21, DMI $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$22, uniaxial anisotropy $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$23, and Zeeman field $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$24,

$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$25

the spin-wave mode acquires a finite-$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$26 minimum once $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$27. For $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$28, simulated annealing yields nanoscale skyrmions at vanishingly small DMI and zero magnetic field, together with field-stabilized skyrmion-lattice phases (Bera et al., 2019).

Scalar-chirality terms provide another zero-field route. On the triangular lattice, the $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$29-symmetric Hamiltonian

$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$30

stabilizes SkX-2, a skyrmion crystal carrying two topological charges per magnetic unit cell and no magnetization at the ferromagnetic point in reciprocal space. Near the onset, a continuum expansion gives a Landau functional in the skyrmion density $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$31 with a continuous onset $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$32 (Bocquet et al., 20 May 2026).

6. Generalized skyrmion lattices and collective modes

Antiferromagnetic skyrmion-lattice models alter both topology and dynamics. In an intrinsic triangular-lattice antiferromagnet with $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$33, $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$34, third-neighbor DMI $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$35, anisotropy $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$36, and field $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$37, isolated and interlinked AFM skyrmions are stabilized on top of a row-wise AFM background. Each ferromagnetic sublattice carries an integer $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$38 charge, while the total charge over the four sublattices vanishes, implying zero net Magnus force (Aldarawsheh et al., 2023).

A distinct antiferromagnetic route begins from itinerant electrons with Rashba spin–orbit coupling and strong Hund’s coupling. After a second-order expansion in the insulating limit, the resulting classical spin Hamiltonian contains antiferromagnetic superexchange, a Dzyaloshinskii–Moriya term proportional to $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$39, and a bond-anisotropy term proportional to $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$40. Monte Carlo on the triangular lattice finds two field-stabilized AF-skyrmion crystals, AF-SkX1 and AF-SkX2, which are distinguished by peaks at distinct momenta in the spin structure factor (Mukherjee et al., 2021).

Synthetic antiferromagnets emphasize collective excitations rather than static topology alone. In the three-coupled-helix continuum theory, a single ferromagnetic skyrmion lattice has a Berry-term-induced quadratic gapless phason, but in the bilayer SAF the Berry term is cancelled and the effective Lagrangian becomes

$${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$41

which yields a linear gapless mode. The same work shows nonzero spin pumping for arbitrary microwave frequency in the lattice case because the excitation spectrum is gapless (Wang et al., 2022).

The notion of a skyrmion lattice also extends beyond ordinary magnetic dipoles. In non-Kramers $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$42 doublet systems on the triangular lattice, bond-dependent hoppings generate a compass-like anisotropy in the effective RKKY interaction, stabilizing zero-field multipolar skyrmion crystals. Two phases are reported: MSkX-I with $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$43 and MSkX-II with $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$44, and the total charge fractionalizes locally into meron and antimeron composites subject to charge conservation within the unit cell (Zhang et al., 2024).

In optical-lattice boson systems, the relevant degree of freedom is orbital pseudospin rather than magnetization. Spinless bosons in degenerate $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$45-orbital bands of a triangular lattice are described by a multi-orbital Bose–Hubbard model whose strong-coupling expansion yields an anisotropic orbital-exchange Hamiltonian. Bosonic dynamical mean-field theory and exact diagonalization show a periodic orbital texture with skyrmion number $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$46 per magnetic unit cell, arising from the interplay of $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$47-orbital symmetry and geometric frustration (Cao et al., 2022).

Symmetry also permits non-skyrmionic topological lattices. For $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$48-type Dzyaloshinskii–Moriya interaction, a three-dimensional lattice-spin model supports an antiskyrmion lattice pocket over $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$49 and $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$50 at $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$51. The same study treats Bloch and Néel skyrmion lattices within the same formalism, while perturbative helimagnet analysis predicts intrinsic sixfold anisotropies and asymmetric lattice distortions rather than perfectly isotropic unit cells (Criado et al., 2021, Kundu, 2015).

7. Interpretation and scope of the model concept

The published usage of “skyrmion lattice model” therefore denotes a modeling framework rather than a unique Hamiltonian. At one end are microscopic spin models with exchange, DM interaction, anisotropy, and field; at the other are reduced particle or continuum descriptions that preserve the lattice symmetry, topological charge, and relevant elastic or thermodynamic observables. Between these limits lie itinerant and multipolar formulations where the periodic topological order is selected indirectly by electronic, orbital, or chirality-mediated interactions (Garanin et al., 2024, Zázvorka et al., 2020, Wang et al., 2021, Zhang et al., 2024).

Two points recur across otherwise different constructions. First, the experimentally visible lattice is usually triangular, but the mechanism of stabilization is not unique: it may be field-induced in a chiral magnet, fluctuation-stabilized near $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$52, generated by frustration and emergent many-body terms in RKKY systems, tied to $${\cal F}[{\bf n}] \;=\;\dfrac{J}{2}\sum_{\mu=x,y}\partial_\mu{\bf n}\!\cdot\!\partial_\mu{\bf n} \;+\;D\,{\bf n}\cdot(\nabla\times{\bf n}) \;-\;{\bf B}\cdot{\bf n},$$53 in Kondo-lattice models, or realized at zero field through scalar chirality, compass anisotropy, or orbital exchange (Buhrandt et al., 2013, Mitsumoto et al., 2021, Bocquet et al., 20 May 2026, Cao et al., 2022). Second, the thermal fate of the lattice is likewise model specific: one-step solid–liquid melting, two-step KTHNY-type melting through a hexatic regime, first-order locking transitions, and floating-solid behavior all occur in the current literature (Garanin et al., 2024, Zázvorka et al., 2020, Gruber et al., 21 Aug 2025, Bocquet et al., 20 May 2026).

A plausible implication is that “skyrmion lattice model” is best understood as a topological-crystal modeling class whose defining ingredients are a periodic noncoplanar texture, a quantized charge diagnostic, and a symmetry-respecting effective theory for ordering, melting, and collective motion. The supplied studies show that this class now encompasses chiral magnets, centrosymmetric itinerant metals, synthetic antiferromagnets, non-Kramers multipolar systems, and orbital optical lattices, with corresponding observables ranging from Bragg peaks and magnetization kinks to Hall responses, defect proliferation, and phason spin pumping (Han et al., 2010, Wang et al., 2022, Zhang et al., 2024, Gruber et al., 21 Aug 2025).