---
title: Néel-Order-Locked Polarization
url: https://www.emergentmind.com/topics/neel-order-locked-polarization
type: topic
---

# Néel-Order-Locked Polarization

Néel-order-locked polarization denotes a class of phenomena in which a polarization channel is constrained by the Néel vector of an antiferromagnet or altermagnet, so that reorienting the magnetic order reorients, switches, or suppresses the polarization. In recent literature, the locked quantity is not limited to ferroelectric polarization: it includes spontaneous electric polarization in type-II multiferroics, spin-layer and valley polarization in symmetry-locked bilayer altermagnets, Kerr rotation and ellipticity arising from antisymmetric optical conductivity, and interfacial or supercurrent-induced spin polarization. Across these settings, the defining feature is a symmetry-allowed coupling between a polar observable and the antiferromagnetic order parameter, linear in some realizations and quadratic in others [2505.01964] [2512.07149] [2603.00739] [2508.21404].

## 1. Conceptual scope and formal structure

The term is used across several neighboring subfields to describe different observables that are all tied to the orientation of antiferromagnetic order. In some systems the locked quantity is a true electric polarization $\mathbf P$; in others it is an optical, valley, layer, or spin polarization. What unifies these cases is not the microscopic origin but the one-to-one dependence on the Néel order parameter imposed by symmetry or by an exchange-mediated locking mechanism.

| Realization | Locked observable | Representative relation |
|---|---|---|
| Rare-earth orthochromites [1212.2740] | field-induced switchable electric polarization | $P \approx -(\alpha\,\mathbf L\times\mathbf M+\beta\,\mathbf H_{\rm ex})/a_P$ |
| 2D altermagnetic multiferroics [2505.01964] | spontaneous electric polarization | $P^\alpha=\Pi^\alpha_{\beta\gamma}L^\beta L^\gamma$ |
| Symmetry-locked bilayer altermagnet [2603.00739] | spin-layer or valley polarization | $P_i=\alpha_i(\epsilon)\,n_i$ |
| Hematite MOKE [2512.07149] | Kerr response via antisymmetric optical conductivity | $\boldsymbol{\sigma}^A=\alpha(\omega)\,(\hat{\mathbf c}\times\mathbf N)$ |
| AFI/FM bilayer [2508.21404] | out-of-plane spin polarization | $\sigma_z$ is largest when $\boldsymbol{\mu}_s\parallel\mathbf n$ |
| Superconductor/altermagnet hybrid [2603.22243] | supercurrent-induced spin polarization | $\mathbf P_s=-\,t_m\gamma\,(q_y\,\hat x+q_x\,\hat y)$ |

A recurrent source of confusion is terminological. The same label is applied to electric, optical, momentum-space, and spin polarizations. This suggests that the term functions as a unifying symmetry descriptor rather than as a restriction to ferroelectric order alone.

## 2. Precursor phenomenology in rare-earth orthochromites

An important precursor appears in weakly ferromagnetic rare-earth orthochromites $R$CrO$_3$, where field-induced switchable polarization emerges below the chromium Néel temperature only when the rare-earth ion is magnetic [1212.2740]. In $R=$ Sm, Gd, Tb, Tm, and Er, a spontaneous polarization $P\sim 0.2$–$0.8~\mu{\rm C/cm}^2$ appears on cooling below $T_N^{\rm Cr}$ in the range $120$–$300~{\rm K}$. Pyroelectric-current measurements under $\pm E_p$ poling of approximately $1$–$2~{\rm kV/cm}$ show true sign reversal of $P$, and PUND pulses confirm an intrinsic remanent polarization of approximately $0.01~\mu{\rm C/cm}^2$ under the small PUND effective field.

The temperature dependence is locked to chromium ordering. The polarization onset coincides with $T_N^{\rm Cr}$ and grows on further cooling roughly as
$$
P(T)\approx P_0(1-T/T_N^{\rm Cr})^\beta,
$$
with experimental $\beta\sim 0.3$–$0.5$. In ErCrO$_3$, which undergoes a spin-reorientation or Morin transition at $T_{SR}\approx 22~{\rm K}$, the polarization collapses to zero below $T_{SR}$ exactly when the weak ferromagnetism of the Cr sublattice disappears. Application of magnetic field $H\gtrsim 1$–$2~{\rm T}$ suppresses the spin reorientation, revives the weak-ferromagnetic state, and restores $P$ down to the lowest temperature. Isothermal scans give a roughly linear magnetoelectric coefficient $dP/dH\sim 0.05$–$0.1~\mu{\rm C/cm}^2\,{\rm T}^{-1}$, with weak higher-order curvature.

The symmetry argument is local-to-global. The rare-earth ion sits on a mirror plane with point group $C_s$, so a local electric dipole is allowed; however, in ideal Pbnm the eight $R$ sites in the unit cell are related by inversion and mirror operations, yielding an antiferroelectric cancellation. The proposed mechanism is that the external poling field $E_p$ breaks macroscopic inversion symmetry and seeds a small polar distortion, while the exchange field from the canted Cr sublattice,
$$
H_{\rm ex}\sim J_{CR}M_{Cr},
$$
acts uniformly on the rare-earth ions and stabilizes one polarity. In the schematic Landau expansion,
$$
F = F_0 + \tfrac12 a_P P^2 + \tfrac12 a_L L^2 + \tfrac12 a_M M^2
+ \alpha\,\mathbf P\cdot(\mathbf L\times\mathbf M)
+ \beta\,\mathbf P\cdot\mathbf H_{\rm ex} + \ldots,
$$
minimization gives
$$
\mathbf P \approx -(\alpha\,\mathbf L\times\mathbf M+\beta\,\mathbf H_{\rm ex})/a_P.
$$
Here the one-to-one correspondence between polarization and weak ferromagnetism anticipates later NOLP formulations, although the later altermagnetic terminology was not yet in use.

## 3. Electric polarization in altermagnetic type-II multiferroics

Guo et al. developed the most explicit symmetry-based formulation of Néel-order-locked *electric* polarization in altermagnetic type-II multiferroics [2505.01964]. The starting point is the contrast between a conventional collinear antiferromagnet with combined $\mathcal P\mathcal T$ symmetry and an altermagnet. In the conventional case, the two magnetic sublattices $A$ and $B$ are related by inversion, implying equal and opposite local dipoles and hence $\mathbf P=0$. In the altermagnetic case, the antiparallel spins on $A$ and $B$ are not connected by inversion. Writing the local dipole on sublattice $M$ as
$$
p_M^\alpha=[K_M]^\alpha_{\beta\gamma}S_M^\beta S_M^\gamma,
$$
the total dipole per cell becomes
$$
p_{\rm tot}^\alpha
=
\bigl[K_A+K_B\bigr]^\alpha_{\beta\gamma}L^\beta L^\gamma,
\qquad
\mathbf L=\tfrac12(\mathbf S_A-\mathbf S_B).
$$
Because inversion is absent in $\mathcal G_s\cup\mathcal G_e$, nothing enforces $K_A=-K_B$, and generically $K_A+K_B\neq 0$. The resulting locked polarization is therefore
$$
P^\alpha
=
N\,[K_A+K_B]^\alpha_{\beta\gamma}L^\beta L^\gamma
\equiv
\Pi^\alpha_{\beta\gamma}L^\beta L^\gamma.
$$

For two-dimensional layer groups with two magnetic sites per cell, the analysis yields exactly eight distinct locking behaviors $\mathbf P(\theta,\phi)$. The magnetoelectric coupling is
$$
F_{\rm ME}=-\mathbf E\cdot\mathbf P(\mathbf L)
=
-\,E_\alpha\,[K_A+K_B]^\alpha_{\beta\gamma}L^\beta L^\gamma.
$$

| Category | Layer groups / site symmetry | $\mathbf P(\phi)$ |
|---|---|---|
| 1 | LG 19, 20, 21 / `..2` | $(A\sin\phi,\;C\cos\phi,\;B\sin2\phi)$ |
| 2 | LG 50 / `2..` | $(A\cos\phi+C\sin\phi,\;C\cos\phi-A\sin\phi,\;B\cos2\phi+D\sin2\phi)$ |
| 3 | LG 53, 54, 76 / `4..` or `222.` | $(A\sin\phi,\;-A\cos\phi,\;0)$ |
| 4 | LG 57, 58 / `2.mm` etc. | $(A\sin\phi,\;A\cos\phi,\;B\sin2\phi)$ |
| 5 | LG 59, 60 / `2mm.` | $(A\cos\phi,\;-A\sin\phi,\;B\cos2\phi)$ |
| 6 | LG 67 / `3..` | $(A\sin\phi+B\sin2\phi,\;-A\cos\phi+B\cos2\phi,\;0)$ |
| 7 | LG 68 / `3..` | $(A\sin\phi+B\cos2\phi,\;-A\cos\phi-B\sin2\phi,\;0)$ |
| 8 | LG 79 / `-6..` | $(B\cos2\phi,\;-B\sin2\phi,\;0)$ |

In every category, the functional form is fixed by symmetry once the layer group is specified; the constants $A,B,C,D$ are material dependent. The components of $\mathbf P$ are either $\pi$- or $2\pi$-periodic in $\phi$.

Monolayer MgFe$_2$N$_2$ is presented as a prototypical realization. It crystallizes in LG 59 with point group $\bar 4 2m$ and two Fe sublattices. DFT using VASP + PAW + PBE finds that the collinear altermagnetic Néel order is lower in energy than the ferromagnetic state by approximately $68~{\rm meV/f.u.}$. The electronic bands show momentum-dependent spin splitting up to approximately $1~{\rm eV}$ at certain $k$, both without and with SOC. Using the modern theory of polarization, the Berry-phase calculation yields an out-of-plane polarization
$$
P_z(\phi)=P_0\cos2\phi,
\qquad
P_0\approx 15.2~\mu{\rm C/m}^2,
$$
which vanishes at $\phi=(2n+1)\pi/4$ and changes sign between $\phi=n\pi/2$. The energy barrier for rotating $\mathbf L$ and switching $P_z$ is below $30~\mu{\rm eV}$.

The same work proposes magneto-optical microscopy as a domain probe. At fixed $\hbar\omega=1~{\rm eV}$, the antisymmetric optical conductivities obey $\sigma_{yz}\propto \sin2\phi$ and $\sigma_{zx}\propto \cos2\phi$, while the Faraday-angle maximum occurs at
$$
\varphi_{\rm in}^{\max}=c_1-\phi+c_2\sin2\phi.
$$
This gives a one-to-one map from the optical response to the Néel angle and therefore to the sign and direction of the locked polarization.

## 4. Linear locking in symmetry-locked bilayer altermagnets

A distinct realization appears in the symmetry-locked bilayer altermagnet proposed to combine momentum-space topology with coupled antiskyrmions [2603.00739]. Here the staggered antiferromagnetic order is described by the Néel vector
$$
\mathbf n=\mathbf m_{BL}-\mathbf m_{TL},
$$
and the polarization channels are defined directly in terms of topological or Berry-curvature observables. The principal examples are the spin-layer polarization
$$
P_{sl}\equiv C_{TL}-C_{BL},
$$
the valley polarizations
$$
P_{vx}\equiv \Omega_X-\Omega_{X'},\qquad
P_{vy}\equiv \Omega_Y-\Omega_{Y'},
$$
and charge or orbital polarization, which vanishes at charge neutrality although layer-resolved versions can survive.

The locking is enforced by residual symmetry. In the unstrained bilayer, the overall point group is $D_{2d}$. With the A-AFM configuration and stacking operator $\mathcal P=M_2C_{4z}$, global inversion and one twofold rotation are broken, while the orthogonal mirror class is preserved. Under SOC, the allowed polar vector depends on the Néel-vector direction. For $\mathbf n\parallel [001]$, the invariance group is $D_{2d}$ itself and only $P_z\neq 0$ is allowed. For $\mathbf n\parallel [100]$, the remaining subgroup forces $P_y=P_z=0$ and permits only $P_x$. For $\mathbf n\parallel [010]$, the allowed component is only $P_y$. The general form is therefore
$$
\mathbf P=(P_x,P_y,P_z)\propto (n_x,n_y,n_z),
$$
or equivalently
$$
P_i=\alpha_{ij}n_j,
$$
with only the diagonal component along the residual symmetry axis remaining nonzero.

The phenomenology is summarized by the strain-dependent coupling
$$
F_{\rm coup}=-\sum_{i=x,y,z}\alpha_i(\epsilon)\,n_iP_i,
\qquad
P_i=\alpha_i(\epsilon)\,n_i.
$$
In the unstrained or compressive regime $(\epsilon\lesssim 0)$, first-principles calculations give $\alpha_z(\epsilon)<0$, so $n_z\neq 0$ opens a spin-layer-polarized quantum spin Hall gap. The reported bandgap is $\Delta\sim 45~{\rm meV}$ with spin Chern numbers $(C_{TL},C_{BL})=(-1,+1)$, yielding
$$
P_{sl}=C_{TL}-C_{BL}=-2.
$$
Under tensile strain $(\epsilon\gtrsim +2\%)$, $\alpha_x(\epsilon)$ and $\alpha_y(\epsilon)$ become nonzero while $\alpha_z$ flips sign and may vanish. For $\mathbf n\parallel[100]$, one layer enters a Chern-insulator phase while the other remains Weyl semimetal, generating a finite $P_{vx}\propto n_x$; the $[010]$ case is analogous for $P_{vy}$.

Because $\mathbf P\parallel \mathbf n$, rotating $\mathbf n$ by $90^\circ$ toggles the polarization channel. In the QSH regime with $\epsilon\sim -2\%$ and $\mathbf n\parallel [001]$, the system exhibits spin-layer polarization and two helical edge modes. Rotating $\mathbf n$ into the film plane enforces $P_z=0$ and allows an in-plane channel, producing a spin-layer-polarized QAH + Weyl state together with a valley Hall response. In the same structure, the interlayer co-directional and locked in-plane Dzyaloshinskii-Moriya interactions stabilize coupled antiskyrmion pairs with compensated topological charges. Since $Q_{TL}=-Q_{BL}$, the total gyrovector vanishes and the Hall-like deflection is fully canceled, while the layer-resolved polarizations remain nonzero.

## 5. Magneto-optical Néel-order-locked polarization in hematite

In hematite, NOLP appears not as a static electric dipole but as a Kerr rotation and ellipticity locked to the Néel vector in a collinear antiferromagnetic insulator with broken $PT$ symmetry [2512.07149]. The key observable is the antisymmetric optical conductivity,
$$
\sigma^A_i(\omega)=\tfrac12\,\epsilon_{ijk}\sigma_{jk}(\omega)
=
\alpha_{ij}(\omega)N_j+\beta_{ij}(\omega)M_j+\gamma_{ij}(\omega)H_j,
$$
which determines the Kerr angle through the usual Fresnel formula,
$$
\varphi_K+i\theta_K
\propto
-\frac{\epsilon_{xy}(\omega)}{\sqrt{\epsilon_{xx}(\omega)}\,(\epsilon_{xx}(\omega)-1)}
\approx
-\frac{\sigma^A_x(\omega)}{i\omega[\epsilon_{xx}(\omega)-1]}.
$$

The central result is that the Néel-order term arises at first order in SOC, whereas the magnetization contribution is at least third order. With $H_{\rm SOC}=\lambda_{\rm SO}\,\mathbf L\cdot\mathbf S$, one obtains to leading order
$$
\epsilon_{xy}(\omega;\mathbf N)\simeq \alpha(\omega)\,N_z,
$$
so that
$$
\varphi_K\propto {\rm Re}[\alpha(\omega)]\,N_z,
\qquad
\theta_K\propto {\rm Im}[\alpha(\omega)]\,N_z.
$$
By contrast, the weak ferromagnetic moment satisfies $\mathbf M_S\propto D(\hat{\mathbf c}\times\mathbf N)$ with Moriya parameter $D\propto \lambda_{\rm SO}^2$, giving
$$
\epsilon_{xy}^{(M)}(\omega)\propto \beta(\omega)M_S\sim \lambda_{\rm SO}^3.
$$
The paper therefore concludes that the magnetization term is two orders of magnitude smaller than the Néel term in hematite.

The symmetry analysis is performed for space group $R\bar 3c$ and collinear AFM order $(+ - - +)$. For $\mathbf N\parallel x$, a glide plane combined with time reversal constrains the allowed coupling to $(\hat c\times \mathbf N)$; for $\mathbf N\parallel y$, the response is correspondingly rotated; for $\mathbf N\parallel z$, the threefold and twofold rotations enforce $\boldsymbol{\sigma}^A=0$. Collecting the cases gives
$$
\boldsymbol{\sigma}^A=\alpha(\omega)\,(\hat{\mathbf c}\times\mathbf N),
$$
at first order in $\lambda_{\rm SO}$.

The measured polar-Kerr signal is of order milliradian and is dominated by the Néel contribution rather than by weak ferromagnetism or the applied field. In the visible range $1.6$–$3.1~{\rm eV}$, the calculated coefficient is
$$
\alpha(\omega)=(-4\text{ to }-2)\times 10^{-4}+i(4\text{ to }6)\times 10^{-4}\quad[{\rm S/A}],
$$
which yields
$$
\sigma^A_{x,0}\simeq(-7.3+9.2\,i)\ {\rm S/cm}
$$
at $\lambda=550~{\rm nm}$ and a Kerr rotation of approximately $1.7~{\rm mrad}$. The ferromagnetic contribution is of order $10^{-2}~{\rm S/cm}$, and the field-induced term evaluated at $H=1~{\rm T}$ is another order smaller. This case directly refutes the common assumption that a sizeable MOKE signal necessarily tracks net magnetization.

## 6. Switching protocols and related spin-polarization extensions

Several recent works extend the locking idea from passive readout to active control of the Néel state and to spin-polarization channels that are not strictly electric. In parity-time-symmetric A-type antiferromagnetic multilayers, the two orientations $\pm \mathcal N_0$ are exactly degenerate until symmetry-breaking light is applied [2503.13825]. Under circularly polarized illumination, time reversal is broken by the handedness and inversion by nonreciprocal attenuation through the layered stack. The free-energy contrast
$$
\Delta G^{\sigma\pm}(\omega)
=
G_{-\mathcal N_0}^{\sigma\pm}(\omega)-G_{+\mathcal N_0}^{\sigma\pm}(\omega)
$$
is nonzero only above the band gap and changes sign under $\sigma^\pm$ reversal or by tuning $\omega$. In the six-layer $k\cdot p$ model, $\Delta G=0$ for $\omega<E_g$, while for $\omega>E_g$ one finds $\sigma^+$ favoring $-\mathcal N_0$ and $\sigma^-$ favoring $+\mathcal N_0$. For bilayer MnBi$_2$Te$_4$, $E_g=0.454~{\rm eV}$ and intensities $I\gtrsim 5\times 10^{11}~{\rm W/cm^2}$ yield $\Delta G$ up to $0.06~{\rm meV/nm^2}$, about twice the intrinsic anisotropy of approximately $0.03~{\rm meV/nm^2}$. For bilayer CrI$_3$, $E_g=1.39~{\rm eV}$ and the threshold is $I\sim 8\times 10^{12}~{\rm W/cm^2}$ for $\Delta G\sim 0.08~{\rm meV/nm^2}$. A plausible implication is that any polarization channel locked one-to-one to the Néel vector should be optically switchable by the same handedness- and frequency-selective protocol.

A more direct spintronic realization appears in AFI/FM bilayers, where the adjacent antiferromagnetic insulator engineers a self-generated out-of-plane spin polarization in a single ferromagnet [2508.21404]. A charge current $\mathbf J_e\parallel \hat x$ in the FM produces a self-generated spin accumulation
$$
\boldsymbol{\mu}_s(z)=\theta_{\rm AH}\rho_{\rm FM}\mathbf J_e\times \hat z\approx \mu_0\,\hat y,
$$
subject to the spin-diffusion equation and spin-mixing boundary conditions at the AFI/FM interface. Reflection is maximal when the incident spin is collinear with the Néel vector, and interfacial spin-orbit coupling or spin swapping rotates the reflected spin to generate $\sigma_z$. The formal result is that $\sigma_z$ is largest when $n_i\mu_s^i$ is maximal. Experimentally, large out-of-plane torques are observed for $I_{\rm RF}\perp \mathbf n$, which is precisely the geometry in which $\boldsymbol{\mu}_s\parallel \mathbf n$: the normalized ST-FMR ratios are
$$
\frac{S_{\rm DL}}{A_{\rm FL}}\approx 0.44,\qquad
\frac{S_{\rm FL}}{A_{\rm FL}}\approx 0.30,
$$
whereas for $I_{\rm RF}\parallel \mathbf n$ they fall to approximately $0.049$ and $0.05$. Field-free perpendicular switching occurs only for $I\perp \mathbf n$, with critical current density
$$
J_c\approx 32.3~{\rm MA/cm^2}.
$$

An analogous extension to dissipationless transport is predicted for superconductor/$d$-wave altermagnet hybrids [2603.22243]. Starting from a finite-$\mathbf q$ BdG Hamiltonian with Rashba SOC and momentum-dependent altermagnetic spin splitting, the first-order free-energy correction is
$$
\mathcal F^{(1)}(\mathbf q,\mathbf n)
=
t_m\gamma(q_x n_y+q_y n_x)+\mathcal O(q^2),
$$
which produces the effective spin polarization
$$
\mathbf P_s\equiv \mathbf h
=
-\,\nabla_{\mathbf n}\mathcal F^{(1)}
=
-\,t_m\gamma(q_y,q_x,0).
$$
The corresponding field-like Néel torque is
$$
\boldsymbol{\tau}_N=\mathbf n\times \mathbf P_s.
$$
The symmetry analysis shows that only the component of $\mathbf n$ perpendicular to $\mathbf q$ contributes, so rotating the Néel vector switches the effect on or off. For realistic parameters the estimated torque is approximately $0.1~{\rm mT}$, sufficient in the calculation to drive domain-wall motion or reverse the Néel orientation within a domain wall.

Taken together, the literature distinguishes at least two mechanistic routes. In rare-earth orthochromites, polarization is stabilized by poling-field symmetry breaking plus exchange striction from a weakly ferromagnetic canted sublattice. In altermagnets and related topological or optical systems, the locked response instead follows from the symmetry of the Néel order itself and can persist with vanishing net magnetization. NOLP is therefore neither synonymous with weak ferromagnetism nor guaranteed by antiferromagnetism alone: it appears only when the relevant inversion, mirror, or rotation constraints permit a polar channel to couple to the Néel vector in the required form.

Source: https://www.emergentmind.com/topics/neel-order-locked-polarization