---
title: Orbital-Moment Octupole in Crystals
url: https://www.emergentmind.com/topics/orbital-moment-octupole
type: topic
---

# Orbital-Moment Octupole in Crystals

Orbital-moment octupole denotes a rank-3 magnetic multipole associated with orbital degrees of freedom. In current usage, the term covers several closely related but not identical constructions: a bulk orbital magnetic octupole tensor \(M_{ijk}\) in periodic crystals; local time-reversal-odd octupolar operators such as \(T_i^y\) or \(T_{xyz}\) in orbital and spin-orbit-entangled manifolds; and octupolar contributions to anomalous Hall transport extracted from symmetry-resolved angular decompositions. A central recent development is the formulation of a gauge-invariant bulk expression for the orbital magnetic octupole in crystalline solids and its connection to a higher-rank Hall response induced by spatially nonuniform electric fields, including cases in which symmetry forbids the conventional anomalous Hall effect against uniform electric fields [2512.24269].

## 1. Terminology and scope

In the periodic-crystal setting, the orbital magnetic octupole is defined thermodynamically as the response of the free-energy density to the second spatial derivative of a magnetic field. This places it in the same hierarchy as dipole and quadrupole magnetic moments, but at the next rank. The resulting object is a rank-3 axial tensor \(M_{ijk}\) [2512.24269].

In local orbital models, the same octupolar designation is often attached to a time-reversal-odd operator acting within a restricted orbital manifold. In the spin-less doubly degenerate \(e_g\) model, the site-octupole operator is identified as
\[
\hat O_i \equiv T_i^y
=\frac1{2i}\bigl(c_{iu}^\dagger c_{iv}-c_{iv}^\dagger c_{iu}\bigr),
\]
whose eigenstates are the complex orbitals
\[
|\,\pm\>\;=\;\tfrac1{\sqrt2}\bigl(d_u\pm i\,d_v\bigr),
\]
and which transforms with \(A_{2g}\) octupolar symmetry [1002.0178]. In the \(J=2\) non-Kramers-doublet setting, the fully symmetrized octupole is
\[
T_{xyz}\equiv \overline{J_xJ_yJ_z}
=\frac1{6}\sum_{\pi\in S_3}J_{\pi(x)}J_{\pi(y)}J_{\pi(z)},
\]
and projection into the low-energy doublet gives
\[
\overline{J_xJ_yJ_z}=-\sqrt3\,\tau^y
\quad\text{or equivalently}\quad
\tau^y=-\frac1{\sqrt3}\overline{J_xJ_yJ_z}
\]
[1909.03089].

A further symmetry-based usage appears in dipole-octupole doublets on the triangular lattice. There, \(\tau^x\) is invariant under the two-fold rotation \(C_2\), odd under time reversal, and is identified as the component of a rank-3 octupolar moment, whereas \(\{\tau^y,\tau^z\}\) transform as the two components of an ordinary magnetic dipole [1608.07008]. In a surface-state \(e_g\) model on the simple cubic lattice, the single-site octupole operator is written as
\[
\hat T_{xyz}(i)=\sum_{\alpha\beta}d_{i\alpha}^\dagger\,\sigma^y_{\alpha\beta}\,d_{i\beta},
\]
with \(\sigma^y\) mixing the two \(e_g\) orbitals with an imaginary coefficient [2408.12046].

A common source of confusion is that not every octupole-related transport theory starts from the same microscopic object. In the multilayer anomalous-Hall literature, the “octupole term” is a rank-4 tensor \(o_{ijkl}\) entering the magnetization-direction expansion
\[
\sigma_i^H
= p_{ij}\,\hat M_j
+ o_{ijkl}\,\hat M_j\,\hat M_k\,\hat M_l
+\cdots,
\]
where \(o_{ijkl}\) is symmetric and traceless in its last three indices. That work explicitly does not write down an explicit second-quantized or Bloch-basis operator \(\hat O_{abc}(\mathbf k)\), and instead parameterizes the Hall conductivity by the coefficients \(o_{ijkl}\) [2508.19033]. The bulk-crystal tensor \(M_{ijk}\) and the transport-expansion tensor \(o_{ijkl}\) therefore belong to different but related frameworks.

## 2. Gauge-invariant bulk formulation in periodic crystals

For a periodic crystal at finite chemical potential \(\mu\) and temperature \(T\), the orbital magnetic octupole is given by the gauge-invariant bulk expression
\[
M_{\,i j k} =\frac{e}{12}\,\sum_{n}\int\!\frac{d^d k}{(2\pi)^d}\, \Bigl[ A_{n}^{\,i j k}\,f'_n \;+\;B_{n}^{\,i j k}\,f_n \;+\;C_{n}^{\,i j k}\,\mathcal{G}_n \Bigr],
\]
with
\[
f_n=f(\epsilon_{n\bm k}-\mu)=\bigl(e^{\beta(\epsilon_{n\bm k}-\mu)}+1\bigr)^{-1},
\qquad
\mathcal{G}_n=-T\ln\bigl(1+e^{-(\epsilon_{n\bm k}-\mu)/T}\bigr).
\]
The band-resolved tensors \(A_n^{ijk}\), \(B_n^{ijk}\), and \(C_n^{ijk}\) are fully gauge-invariant rank-3 axial tensors built out of the Berry curvature \(\Omega_n^{ab}\), quantum metric \(g_n^{ab}\), orbital moment \(m_n^{ab}\), and higher derivatives of the Hamiltonian [2512.24269].

The geometric building blocks are
\[
T_n^{ab}=\langle\partial_{k_a}u_{n\bm k}\vert \hat Q_{n}\vert\partial_{k_b}u_{n\bm k}\rangle
=g_n^{ab}-\tfrac{i}{2}\,\Omega_n^{ab},
\]
\[
\tilde T_n^{ab}=\langle\partial_{k_a}u_{n\bm k}\vert(\hat H(\bm k)-\epsilon_{n\bm k})\vert\partial_{k_b}u_{n\bm k}\rangle,
\qquad
m_n^{ab}=-\mathrm{Im}\,\tilde T_n^{ab},
\]
together with the covariant derivative
\[
\ket{D_a n}=\hat Q_n\ket{\partial_{k_a}u_{n\bm k}}.
\]
An explicit example is
\[
A_n^{\,i j k} =\epsilon_{i a b}\,\Bigl[\, \tfrac{1}{24}\,m_n^{ab}\,\partial_{k_j}\partial_{k_k}\,\epsilon_{n\bm k} \;+\;\text{(5 perms. of }a,b,j,k)\Bigr].
\]
The expressions for \(B_n^{ijk}\) and \(C_n^{ijk}\) are likewise completely gauge-invariant.

The gauge structure is central. Under a local phase redefinition \(\ket{u_{n\bm k}}\to e^{i\eta_n(\bm k)}\ket{u_{n\bm k}}\), all terms remain unchanged. The position operator never appears explicitly; only derivatives with respect to \(\bm k\) enter. Gauge invariance is guaranteed by using projectors \(\hat Q_n\) and covariant derivatives [2512.24269].

## 3. Thermodynamic definition and response-theory derivation

The thermodynamic starting point is the free-energy density \(F(\bm r)\) in a slowly varying magnetic field \(\bm B(\bm r)\). Expanding \(F\) in gradients of \(\bm B\) introduces the dipole, quadrupole, and octupole magnetic moments. By definition,
\[
M_{ijk}=-\frac{\partial F}{\partial(\partial_j\partial_k B_i)}.
\]
This identifies the orbital magnetic octupole as the coefficient conjugate to the second spatial derivative of the magnetic field [2512.24269].

A Maxwell relation makes the calculation algebraically simpler:
\[
\frac{\partial M_{ijk}}{\partial\mu}
=
\frac{\partial N}{\partial(\partial_j\partial_k B_i)}.
\]
The derivation then proceeds through linear response for the density variation \(\Delta N\). One computes the density-current correlation \(\chi_{N,J_a}(\bm q,\omega)\) up to \(\mathcal O(q^3)\), expands \(\chi_{N,J_a}\sim q_b q_j q_k\,\partial_q^3\chi\), and uses
\[
\Delta N(\bm q)=\chi_{N,J_a}A_a(\bm q)
\]
to isolate the contribution proportional to \(\partial_j\partial_k B_i\). Integrating the \(\mu\)-derivative and reorganizing the result yields the final gauge-invariant bulk formula [2512.24269].

This framework clarifies the status of \(M_{ijk}\). It is not introduced as a phenomenological fitting parameter but as a thermodynamic quantity derivable from Bloch-state response theory. The formulation also makes explicit that the bulk theory relies on Bloch periodicity on the Brillouin zone and on a representation in terms of quantum-geometric objects rather than real-space position operators. A plausible implication is that the formulation places the orbital magnetic octupole on the same conceptual footing as other modern-band-theory observables built from Berry curvature, quantum metric, and orbital moments.

## 4. Symmetry constraints and model realizations in crystalline solids

A minimal realization of a nonzero orbital magnetic octupole is provided by the two-sublattice, collinear antiferromagnet described as a “\(d_{xy}\)-wave altermagnet” on a tetragonal lattice:
\[
\hat H_{\bm k} =\varepsilon_{0,\bm k}\,\hat1 \;+\;t_{x,\bm k}\,\hat\tau_x \;+\;t_{z,\bm k}\,\hat\tau_z \;+\;\lambda_{z,\bm k}\,\hat\tau_{y}\,\hat\sigma_z \;+\;J\,\hat\tau_z\,\hat\sigma_z.
\]
Here \(\hat\tau_{x,y,z}\) act in sublattice space, \(\hat\sigma_z\) in spin space, and
\[
\varepsilon_{0,\bm k}=t_1(\cos k_x+\cos k_y)-\mu+\cdots,\quad
t_{x,\bm k}=t_8\cos\frac{k_x}{2}\cos\frac{k_y}{2}\cos\frac{k_z}{2},
\]
\[
t_{z,\bm k}=t_6\sin k_x\sin k_y+\cdots,\quad
\lambda_{z,\bm k} =\lambda\,\cos\tfrac{k_x}{2}\cos\tfrac{k_y}{2}\cos\tfrac{k_z}{2}\,(\cos k_x-\cos k_y).
\]
The term \(J\,\tau_z\sigma_z\) breaks time-reversal but preserves mirror symmetries \(\mathcal M_x,\mathcal M_y,\mathcal M_z\) and \(C_{4z}\mathcal T\). As a result, the only nonzero octupole components are
\[
M_{zxy},\quad M_{xyz}=M_{yzx}.
\]
Numerical evaluation shows that both vanish for \(\lambda=0\) and grow roughly linearly with spin-orbit coupling \(\lambda\), which highlights the relativistic origin of the orbital contribution [2512.24269].

A distinct realization arises in \(C_{4v}\)-symmetric magnetic multilayers, where the anomalous Hall response is decomposed into dipole and octupole harmonics. For in-plane magnetization \(\mathbf M=(\cos\alpha,\sin\alpha,0)\) and current \(j_z\) along \(z\),
\[
j_H^x(\alpha)=q_1\,\sin\alpha + q_3\,\sin3\alpha + q_5\,\sin5\alpha + \cdots,
\]
\[
j_H^y(\alpha)=-\,q_1\,\cos\alpha + q_3\,\cos3\alpha - q_5\,\cos5\alpha + \cdots.
\]
The first two coefficients are
\[
q_1=\frac14\bigl(4\,p_{Xx}+3\,o_{Xxxx}+3\,o_{Xxyy}\bigr),
\qquad
q_3=\frac14\bigl(-\,o_{Xxxx}+3\,o_{Xxyy}\bigr).
\]
Here \(q_1\) is the conventional dipole-dominated anomalous Hall effect with a small octupole correction, whereas \(q_3\) is exclusively the third-order octupole contribution. In \(C_{4v}\), only combinations such as
\[
o_{Xxxx},\quad o_{Xxyy},\quad o_{Yyyy},\quad o_{Yyxx}
\]
can be nonzero, and they reduce to two independent parameters in the Hall geometry [2508.19033].

The multilayer calculations use a fully-relativistic, spin-orbit-coupled exact muffin-tin orbitals scattering-wave-function method for alternating bcc-Fe and fcc-Ag layers, sandwiched between semi-infinite Ag leads, with the Ag lattice rotated \(45^\circ\) to obtain \(C_{4v}\) in-plane symmetry. Charge currents between atoms \(R,R'\) are evaluated from
\[
J_{RR'}
= \frac1{i\hbar}\Bigl[\,\langle\Psi_R|\mathcal H_{RR'}|\Psi_{R'}\rangle
-\langle\Psi_{R'}|\mathcal H_{R'R}|\Psi_R\rangle\Bigr],
\]
and a dense \(5120\times5120\) \(k_\parallel\) mesh in the 2D Brillouin zone ensures convergence of the transport integrals. In \((\mathrm{Ag}_2\mathrm{Fe}_5)_n\) multilayers, the pure-octupole term \(q_3\) reaches up to \(\approx 6\times |q_1|\) in the \((\mathrm{Ag}_2\mathrm{Fe}_5)_8\) ferromagnet, while in a simple Fe\(|\)Ag interface it is only \(\approx 0.12\times |q_1|\). The relative octupole strength \(q_3/|q_1|\) changes from \(\approx 2.5\) to \(\approx 6\) to \(\approx 1.2\) as \(n\) varies from \(6\to8\to10\), demonstrating tunability by layer number, and the “\((\mathrm{Ag}_2\mathrm{Fe}_5)_8\)-AFM” configuration reverses the sign of both \(q_1\) and \(q_3\) [2508.19033].

## 5. Higher-rank Hall response and experimental access

The principal transport consequence of the bulk orbital magnetic octupole is a higher-rank Hall response driven by a spatially nonuniform electric field. Since a spatially nonuniform electric field can be written as a gradient of a local chemical potential, \(\bm E=-\nabla(\mu/e)\), the octupole induces a transverse current under \(\nabla\nabla \bm E\). The corresponding “octupolar anomalous Hall conductivity” is
\[
\sigma_{abjk}
=
\epsilon_{abi}\,\frac{\partial M_{ijk}}{\partial\mu}
=
e\,\epsilon_{abi}\,\frac{\partial N}{\partial(\partial_j\partial_k B_i)}.
\]
In real space, the leading linear response is
\[
J_a=\sigma_{abjk}\,\partial_{r_b}\partial_{r_j}E_k+\cdots.
\]
This relation is the higher-rank analogue of the ordinary Středa formula \(\sigma_{xy}=\partial M_z/\partial\mu=\partial N/\partial B_z\) [2512.24269].

This framework has a direct symmetry implication for altermagnets. Mirror or rotation symmetries can forbid the conventional anomalous Hall effect \(\sigma_{xy}\) against uniform electric fields while still allowing the octupolar conductivity \(\sigma_{abjk}\). In that situation, the first nonzero Hall response under a spatially varying \(\bm E\) is purely octupolar. The proposed detection route is to engineer a controlled electric-field gradient, for example via split-gate electrodes or near-field optical excitation, and measure the resulting transverse current. A static alternative is the quadrupolar magnetoelectric effect \(\bm P\propto\nabla\bm B\) with
\[
\alpha_{ijk}=e\,\frac{\partial M_{ijk}}{\partial\mu},
\]
so that applying a magnetic-field gradient generates an electric polarization with the symmetry of \(M_{ijk}\). Because \(\alpha_{ijk}\) is linear in \(\mu\) within an insulating gap, one expects a strictly linear-in-\(\mu\) plateau in the octupolar response versus doping [2512.24269].

The multilayer literature provides a complementary experimental strategy based on angle-resolved transport. Because \(q_3\) is a pure octupole contribution, measuring the \(\sin3\alpha\) and \(\cos3\alpha\) components of the Hall current completely isolates the octupole piece. This method explicitly incorporates discrete crystal symmetries and is stated to enable the investigation of octupole contribution in non-periodic systems, particularly at interfaces and surfaces [2508.19033].

A recurrent misconception is that anomalous Hall phenomena in magnetic conductors are exhausted by the dipole moment or net magnetization. The multilayer results explicitly state that even the conventional contribution arises not only from the dipole moment, and the crystalline-octupole theory shows that a Hall response can persist in a higher-rank form when the uniform-field anomalous Hall channel is symmetry-forbidden [2508.19033; 2512.24269].

## 6. Correlated-orbital, surface, and collective manifestations

In Mott-insulating orbital models, the orbital-moment octupole appears as an ordered degree of freedom rather than as a band-structure response coefficient. In the doubly degenerate \(e_g\) model, fourth-order ring exchange generates the plaquette interaction
\[
{\cal H}_R
=K_R\sum_{[ijkl]_a}\frac12
\bigl(\tau_i^{a+}\tau_j^{a-}\tau_k^{a+}\tau_l^{a-}+\mathrm{H.c.}\bigr),
\qquad
\tau_i^{\pm a}=\tau_i^a\pm i\,\frac{\sqrt3}{2}T_i^y,
\]
with \(K_R=40t^4/U^3\). Because \({\cal H}_R\) naturally involves the imaginary combination \(\tau^a\pm i(\sqrt3/2)T^y\), it directly couples to the time-reversal-odd operator \(T_i^y\), identified as the magnetic octupole. The reported phase structure includes quadrupole order up to \(r_R\simeq 0.02\), onset of genuine \(\mathrm{OP}(0,0,\pi)\) order at \(r_R\simeq0.22\), and collapse of both quadrupole and octupole moments by \(r_R\sim0.35\) because of strong fluctuations [1002.0178].

In \(5d^2\) double perovskites, the octupole resides in a low-lying non-Kramers doublet of a \(J=2\) ion. The doublet carries purely quadrupolar and octupolar moments and no matrix elements of \(J^\alpha\) within it. A second-order perturbative mechanism from Heisenberg exchange and orbital repulsion produces the ferro-octupolar coupling
\[
K=\frac{6\gamma_m\gamma_2}{\Delta},
\qquad
H_{\rm oct}=-K\sum_{\langle ij\rangle}\tau_i^y\tau_j^y.
\]
The same work identifies experimental signatures in \(\mu\)SR, Raman scattering, X-ray diffraction, inelastic neutron scattering, and a gapped, dispersive magnetic exciton that can condense into conventional type-I antiferromagnetic order when \(\gamma_m=\Delta/16\) [1909.03089].

In dipole-octupole doublets on the triangular lattice, octupolar order may be hidden from conventional magnetization measurements. The nearest-neighbor pseudospin Hamiltonian is
\[
H_0=\sum_{\langle ij\rangle}\Bigl[
J_x\,\tau_i^x\tau_j^x
+J_y\,\tau_i^y\tau_j^y
+J_z\,\tau_i^z\tau_j^z
+J_{yz}\,(\tau_i^y\tau_j^z+\tau_i^z\tau_j^y)
\Bigr].
\]
After an \(SO(2)\) rotation in pseudospin space, \(T^x\) remains purely octupolar while \(\{T^y,T^z\}\) form the dipole doublet that couples linearly to an external field. The model supports ferro-octupolar order with
\[
m_x=\langle T_i^x\rangle\neq 0,\qquad m_y=m_z=0,
\]
and antiferro-octupolar three-sublattice order. In the ferro-octupolar phase, the octupolar susceptibility \(\chi^{xx}\) diverges at \(\mathbf q=0\), whereas the dipolar susceptibilities remain finite. The predicted experimental consequences include finite Kerr rotation below the ordering temperature and sharp octupolar-wave modes in inelastic neutron scattering [1608.07008].

Surface-localized orbital octupoles also occur in a purely orbital tight-binding setting. For the simple-cubic \(e_g\) model with finite thickness, (110) and (111) surfaces host electronic states localized around the surfaces, and the (111) surface supports a perfectly flat band when the bulk band projected onto the surface Brillouin zone is gapped. The surface-state spinor is proportional to the \(\sigma^y\) direction, \(\phi\propto [1,\pm i]^\top\), and the expectation value
\[
\langle\psi_{\rm surf}(k_\parallel)|\hat T_{xyz}(\ell=1)|\psi_{\rm surf}(k_\parallel)\rangle=\pm1
\]
shows that the surface-state manifold carries a quantized octupole moment density of unit magnitude on the top surface and the opposite sign on the bottom [2408.12046].

A different use of orbital moment appears in the octupole collective Hamiltonian. In the intrinsic-frame description of octupole vibrations, the operators
\[
J_s^{(f)}=\frac{3}{2}\,i\Bigl(f_t\frac{\partial}{\partial f_u}-f_u\frac{\partial}{\partial f_t}\Bigr),
\qquad (s,t,u)\text{ cyclic in }(x,y,z),
\]
satisfy the usual angular-momentum algebra and are identified as an intrinsic orbital moment carried by the octupole vector vibrations \((f_x,f_y,f_z)\). This establishes a collective-mechanical notion of orbital moment that is distinct from the Bloch-band and local-orbital octupoles discussed above [1802.01703].

Taken together, these results show that orbital-moment octupoles are not restricted to a single formalism. They appear as gauge-invariant bulk tensors in periodic crystals, as local order parameters in orbital and spin-orbit-entangled Mott systems, as symmetry-resolved transport coefficients in interfacial anomalous Hall effects, as quantized surface-state moments in \(e_g\) bands, and as intrinsic orbital moments of octupole vibrations in collective Hamiltonians. This suggests that the unifying content of the term is not a unique operator but the organization of time-reversal-odd, higher-rank orbital structure and its electromagnetic or dynamical consequences.

Source: https://www.emergentmind.com/topics/orbital-moment-octupole