Zeeman Berry Curvature Dipole

• Zeeman Berry curvature dipole is defined as either the conventional dipole modified by Zeeman coupling or its Zeeman-weighted variant incorporating magnetic moment textures.
• It reveals how magnetic fields break symmetry to induce nonlinear Hall, photocurrent, and magnetoelectric responses in both centrosymmetric and noncentrosymmetric systems.
• Applications include Weyl semimetals, moiré systems, and low-symmetry materials, where Zeeman interactions reshape band geometry and topological transitions.

Searching arXiv for the cited papers to ground the article in the published record. arxiv_search(query="(Sinha et al., 2022)") Zeeman Berry curvature dipole is not yet a fully standardized term in the arXiv literature. The current literature suggests two closely related usages. One treats it as the ordinary Berry curvature dipole evaluated in a Zeeman-modified band structure, $D_{ab}(\mathbf{B})$, so that magnetic or exchange fields tune the first momentum-space moment of Berry curvature. The other identifies a Zeeman-weighted Fermi-surface tensor, $$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$, as the natural magnetic-moment analogue of the conventional dipole and the key object in magnetoelectric electro-optic response (Zhang et al., 2017, Sousa et al., 2024). In both usages, Zeeman coupling mediates between band geometry, symmetry lowering, and nonlinear transverse response.

1. Definition and formal scope

The conventional Berry curvature dipole is the first moment of Berry curvature over occupied states. In the semiclassical form used for noncentrosymmetric metals it is written as

$$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$

with $\int_k\equiv \int d^3k/(2\pi)^3$ in three dimensions. In two dimensions, where only the out-of-plane component $\Omega_z$ is nonzero, the same object is often written as a pseudovector $\Lambda_\alpha$ or $D_{\alpha z}$ (Zhang et al., 2017, Sinha et al., 2022). For twisted double bilayer graphene, for example,

$$\Lambda_{\alpha}=\sum_n \int_{\rm mBZ}\frac{d\mathbf{k}}{(2\pi)^2}\,\Omega_z^n(\mathbf{k})\,\frac{\partial \epsilon^n_{\mathbf{k}}}{\hbar\partial k_\alpha}\,\frac{\partial f(\epsilon^n_{\mathbf{k}})}{\partial \epsilon^n_{\mathbf{k}}}, \qquad D_{xz}\equiv \Lambda_x,\quad D_{yz}\equiv \Lambda_y.$$

A second, more specific construction appears when Berry curvature is combined with the magnetic moment texture. In that case the tensor

$$G^{\alpha\beta}=\sum_{n\mathbf{k}}\left(-\frac{\partial f^0_{n\mathbf{k}}}{\partial\epsilon_{n\mathbf{k}}}\right)\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$

is the natural Zeeman-weighted Berry-curvature object controlling a magnetoelectric electro-optic effect (Sousa et al., 2024).

Usage	Object	Response channel
Zeeman-tuned conventional BCD	$D_{ab}(\mathbf{B})$	nonlinear Hall, photocurrent
Zeeman-weighted BCD	$$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$0	magnetoelectric electro-optic response

This dual usage is not a contradiction. It reflects two levels of generalization. The first keeps the usual BCD definition and asks how Zeeman coupling changes it. The second treats the magnetic moment itself as part of the geometric weight and thereby accesses response functions that are absent in the purely electric sector (Zhang et al., 2017, Sousa et al., 2024).

2. Symmetry requirements and what Zeeman coupling changes

For the conventional BCD in the Sodemann–Fu setting, the canonical symmetry conditions are stringent. Time-reversal symmetry may be preserved, but inversion symmetry must be broken, and additional point-group symmetries must be sufficiently reduced so that the first moment of $$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$1 does not cancel. In two dimensions, the largest symmetry compatible with a nonzero in-plane dipole is a single mirror line; $$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$2 or higher rotational symmetry forces the dipole to vanish. This structure is emphasized across TMDCs, strained graphene, and field-induced WTe$$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$3 (You et al., 2018, Qin et al., 2020, Battilomo et al., 2019, Ye et al., 2023).

A common misconception is that adding a Zeeman term automatically generates a Berry curvature dipole. The isotropic Rashba-plus-Zeeman problem already shows why this is false. The lower band acquires a finite Berry curvature,

$$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$4

but because the resulting $$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$5 is radial in the isotropic model, any first moment vanishes by symmetry. This comparison indicates that Zeeman coupling is a source of Berry curvature, but not by itself a guarantee of a nonzero dipole; anisotropy, warping, or reduced point-group symmetry remain essential (Price et al., 2014).

Once Zeeman coupling breaks time-reversal symmetry, the symmetry landscape changes in two distinct ways. First, a linear anomalous Hall conductivity becomes allowed because the net Berry curvature need no longer cancel. Second, the conventional BCD $$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$6 remains well defined, but its allowed tensor structure is no longer restricted to the time-reversal-invariant case. By contrast, the Zeeman-weighted tensor $$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$7 is even under both $$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$8 and $$G^{\alpha\beta}=\sum_{n\mathbf{k}}(-\partial f^0/\partial\epsilon)\,\Omega^\alpha_{n\mathbf{k}}\,m^\beta_{n\mathbf{k}}$$9, because $$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$0 and $$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$1 transform in the same way under those operations. As a result, $$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$2 can exist either in noncentrosymmetric systems that preserve time reversal or in centrosymmetric systems that break time reversal (Sousa et al., 2024).

3. Microscopic mechanisms

At the band-structure level, Zeeman coupling enters through a term such as $$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$3, or more generally through a $$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$4-dependent $$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$5-tensor or exchange field. In ab initio and continuum formulations alike, the effect is conceptually direct: one adds the Zeeman term to the Hamiltonian, recomputes eigenstates and eigenvalues, evaluates

$$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$6

and then forms the corresponding dipole tensor (Zhang et al., 2017). Because the Berry-curvature denominator is controlled by interband splittings, Zeeman-induced lifting of degeneracies and avoided crossings can strongly enhance local curvature.

In Weyl semimetals the dominant mechanism is Fermi-surface asymmetry around concentrated Berry-curvature monopoles. Type-II Weyl points are especially effective because strong tilt makes the occupied region around the node highly asymmetric, which yields large $$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$7; the same paper also shows that near-mirror-plane spin-orbit gaps can produce large BCD even when the dominant contribution is not directly Weyl-node-derived (Zhang et al., 2017). This implies that a Zeeman field can tune the dipole not only by moving Weyl nodes relative to the Fermi level, but also by reshaping other Berry-curvature hot spots.

In moiré systems the same logic appears in a different guise. In twisted double bilayer graphene, perpendicular electric displacement field tunes the valley Chern numbers and the BCD simultaneously, and the dipole changes sign sharply across a topological transition. The paper explicitly notes that its formalism carries over essentially unchanged if the tuning parameter is Zeeman field rather than displacement field. This suggests that a Zeeman-driven topological transition should generically be accompanied by enhanced BCD and often by a sign reversal of $$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$8 at or near the critical field (Sinha et al., 2022).

A further microscopic route is purely orbital. In a low-symmetry SU(3) orbital model, crystal fields and orbital Rashba-type couplings generate Berry-curvature hot spots and singular pinch points even without hole-like excitations, and the resulting BCD can be giant. This is not itself a Zeeman mechanism, but it identifies a favorable background on which Zeeman coupling could act as a secondary knob: once sharp hot spots and pinch points exist, small magnetic perturbations can strongly skew their Fermi-surface weighting (Mercaldo et al., 2023).

4. Response theory

For the conventional BCD, the low-frequency nonlinear response is governed by the standard semiclassical susceptibility

$$D_{bd}=\int_k f_0(\mathbf{k})\,\frac{\partial \Omega_d(\mathbf{k})}{\partial k_b},$$9

so that in the dc limit

$\int_k\equiv \int d^3k/(2\pi)^3$0

This is the basic relation connecting $\int_k\equiv \int d^3k/(2\pi)^3$1 to nonlinear Hall current, dc photocurrent, and second-harmonic generation in noncentrosymmetric metals (Zhang et al., 2017). In a Zeeman-controlled setting, nothing in this structure changes except that $\int_k\equiv \int d^3k/(2\pi)^3$2 becomes field dependent.

The magnetic-moment generalization leads to a different response sector. With a dc electric bias $\int_k\equiv \int d^3k/(2\pi)^3$3 and weak ac fields, the current takes the generalized form

$\int_k\equiv \int d^3k/(2\pi)^3$4

The new Zeeman-weighted tensor $\int_k\equiv \int d^3k/(2\pi)^3$5 enters the magnetoelectric electro-optic conductivity through

$\int_k\equiv \int d^3k/(2\pi)^3$6

This term vanishes in the dc limit and is therefore inherently ac, in contrast with the ordinary BCD contribution, which remains finite as $\int_k\equiv \int d^3k/(2\pi)^3$7 (Sousa et al., 2024).

A further extension appears in superconductors, where the BCD is reformulated as a many-body geometric quantity in vector-potential or Cooper-pair-momentum space. The superconducting BCD is defined from derivatives of the many-body projector, and in Bogoliubov–de Gennes form it decomposes into a parent-band BCD contribution and an order-parameter contribution that is explicitly sensitive to the phase structure of the gap. The cited superconducting work does not include Zeeman coupling, but its formalism can accommodate Zeeman-modified normal bands or Zeeman-dependent pairing and therefore provides a direct route to a Zeeman-dependent BdG dipole (Matsyshyn et al., 2024).

5. Material platforms and empirical paradigms

The most mature Zeeman-relevant platforms are noncentrosymmetric semimetals, low-symmetry two-dimensional materials, and tunable moiré systems. The existing literature does not yet provide a single canonical “Zeeman Berry curvature dipole material,” but several classes already supply the required ingredients.

In Weyl semimetals, ab initio calculations show substantial zero-field BCD in TaAs-family compounds and in the MoTe$\int_k\equiv \int d^3k/(2\pi)^3$8 family, with type-II Weyl points generally outperforming type-I nodes because of stronger tilt. The same work identifies NbP and NbAs as especially large-BCD materials and therefore natural candidates for field tuning (Zhang et al., 2017). This makes Weyl systems the most direct setting for $\int_k\equiv \int d^3k/(2\pi)^3$9.

In moiré graphene, strained twisted double bilayer graphene establishes a different paradigm: a tunable flat-band system in which BCD sign reversal tracks a topological transition and where hysteresis appears in both longitudinal and nonlinear Hall response. Although no Zeeman term is included, the paper states that its mechanism for probing topological transitions by NLH/BCD is not specific to electric displacement field and can apply when the tuning parameter is a Zeeman field (Sinha et al., 2022).

In monolayer WSe$\Omega_z$0, uniaxial strain breaks $\Omega_z$1 symmetry and creates a finite BCD, which in turn generates out-of-plane current-induced orbital magnetization and a nonlinear Hall effect. The paper explicitly frames this as a template for Zeeman-induced BCD, because a Zeeman field or exchange coupling can modify Berry curvature distribution and valley occupation in an analogous way (Qin et al., 2020).

In thick WTe$\Omega_z$2, a dc electric field generates and reorients a BCD that is symmetry-forbidden in the bulk at zero field. The observed BCD is linear in the control field, and its polarization follows the field direction subject to mirror constraints. This provides a direct analogy for Zeeman control: any external vector field that lowers response symmetry to at most a single mirror can induce a BCD (Ye et al., 2023).

In chiral materials such as Te, Se, CoSi, and RhSi, the conventional BCD tensor $\Omega_z$3 and the Zeeman-weighted tensor $\Omega_z$4 are both symmetry-allowed and diagonal in the classes analyzed. These systems are therefore especially relevant for optical or THz probes of the magnetoelectric electro-optic effect governed by $\Omega_z$5 (Sousa et al., 2024).

6. Experimental signatures and diagnostics

The most robust transport signature of a conventional Zeeman-tuned BCD is a second-order Hall response whose magnitude or sign tracks the magnetic control parameter. In the standard low-frequency regime, one expects a transverse signal quadratic in electric field and linear in the relevant dipole component through the susceptibility above (Zhang et al., 2017). In practice, the experimentally decisive signature is often a sign reversal or lobe rearrangement rather than merely a large magnitude.

The clearest precedent for such a sign reversal comes from twisted double bilayer graphene. There, the BCD changes sign at the same critical displacement field at which the band topology changes, producing an order-parameter-like transport signature of a topological transition. Because the authors explicitly state that the same framework carries over when the tuning parameter is Zeeman field, this establishes a concrete diagnostic expectation for field-driven magnetic tuning: a Zeeman-induced topological transition should be detectable through a sign-changing nonlinear Hall coefficient (Sinha et al., 2022).

Experimental extraction generally requires separating intrinsic geometric response from extrinsic skew-scattering or side-jump mechanisms. In TDBG the scaling form

$\Omega_z$6

was used, with the intercept $\Omega_z$7 identifying the intrinsic BCD-dominated contribution (Sinha et al., 2022). In strained WSe$\Omega_z$8, the second-harmonic Hall signal scales quadratically with drive current, and the extracted BCD reaches a few nanometers; in WTe$\Omega_z$9, the induced BCD is linear in the control field and can be reoriented by rotating that field relative to the crystal axes (Qin et al., 2020, Ye et al., 2023). These protocols are directly transferable to Zeeman tuning, with the additional complication that broken time reversal introduces a linear anomalous Hall background that must be separated from the quadratic signal.

For the Zeeman-weighted tensor $\Lambda_\alpha$0, the preferred probes are ac or optical rather than dc. The magnetoelectric electro-optic effect is linear in both $\Lambda_\alpha$1 and $\Lambda_\alpha$2, proportional to $\Lambda_\alpha$3, and suppressed as $\Lambda_\alpha$4 by the factor $\Lambda_\alpha$5. THz and mid-infrared spectroscopy in biased chiral materials are therefore the natural experimental setting (Sousa et al., 2024).

7. Related concepts and terminological boundaries

A useful distinction must be maintained between Zeeman Berry curvature dipole and other “dipole” notions in band topology. “Berry-dipole semimetals,” for instance, are three-dimensional gapless phases whose nodes carry quantized half-space Berry fluxes and whose low-energy Berry curvature has a dipolar structure in momentum space. That concept concerns quantized point dipoles associated with band nodes, not the Fermi-surface dipole tensor $\Lambda_\alpha$6 or the magnetic-moment-weighted tensor $\Lambda_\alpha$7 that control nonlinear response (Zhuang et al., 2024).

The present literature therefore supports two complementary but distinct encyclopedia-level meanings of Zeeman Berry curvature dipole. One is operational and transport-oriented: the conventional Berry curvature dipole made field dependent by Zeeman coupling, $\Lambda_\alpha$8. The other is constitutive and magnetoelectric: the tensor $\Lambda_\alpha$9, which weights Berry curvature by magnetic moment and governs a bias-controlled response to ac magnetic fields (Zhang et al., 2017, Sousa et al., 2024). This terminological split is not yet resolved by universal convention. What is already clear is the underlying principle shared by both usages: Zeeman coupling reorganizes Berry-curvature-carrying states near the Fermi surface, and the resulting asymmetry is accessible through nonlinear Hall, optical, and magnetoelectric observables.