Functional Berry Flux: Operational Topology

• Functional Berry flux is a concept where Berry curvature becomes a tunable, geometry-dependent quantity whose value varies with the integration region or measurement protocol.
• The approach applies methods from exciton fine structure analysis and momentum-space Fermi surface integrations to reveal both quantized and dynamic flux behaviors.
• Experimental and computational frameworks, including DFT post-processing and ARPES, validate its role in determining transport anomalies and topological responses.

Functional Berry flux denotes a class of geometric constructions in which Berry flux is treated not only as a quantized topological invariant on a closed manifold, but as a quantity whose value or physical action depends on the integration region, the real- or momentum-space protocol used to define it, or a tunable or self-consistently generated gauge structure. In the literature, this viewpoint appears in several distinct but related forms: the Berry-curvature flux through the finite $k$-space patch sampled by an exciton, the flux through a Fermi surface carrying monopole charge, the radius-dependent and non-quantized flux inside a $\mathcal{PT}$-symmetric exceptional surface, the dependence of polarization Berry phases on how flux is inserted in real space, and engineered orbital or moiré fluxes that activate Berry curvature and topological response (Zhou et al., 2015, Son et al., 2012, Wang et al., 2024, Watanabe et al., 2018, Tagani et al., 3 Jun 2026, Yu et al., 2019).

1. Meanings and scope of the concept

Across the cited works, the expression does not have a single canonical meaning. This suggests that “functional Berry flux” is best understood as a family of closely related uses in which Berry flux is promoted from a fixed topological number to a geometry-dependent, protocol-dependent, or tunable object. In the exceptional-surface problem, the flux is explicitly a function of the radius and shape of the chosen integration region, and it is therefore not quantized inside the exceptional torus. In the polarization problem, the Berry phase depends on the functional form of the vector potential $A_x(x)$, so different flux profiles represent different real-space current measurements. In kagome altermagnetism, the relevant “flux” is an orbital chiral flux $\chi$ that acts as a control knob for Berry-curvature activation. In the $\lambda_N$-jellium model, both the Berry-curvature distribution and the total integrated flux are tunable model parameters. In Berryogenesis, the central quantity is a self-induced dc Berry flux generated by the system’s own internal plasmonic field rather than imposed externally (Wang et al., 2024, Watanabe et al., 2018, Tagani et al., 3 Jun 2026, Desrochers et al., 18 Sep 2025, Rudner et al., 2018).

A common thread is that Berry flux is treated operationally: it is defined through what region is enclosed, what gauge evolution is performed, what subspace is parallel transported, or what internal field configuration is realized. In some cases the functional dependence is literal, as in the polarization problem where one may distribute the same total flux using $A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$ with $\sum_i p_i=1$, thereby obtaining different Berry phases associated with different seam-weighted current measurements. In other cases it is dynamical or materials-engineering driven, as when an orbital flux term $H_\chi$ breaks $\mathcal{T}C_{2z}$ in a coplanar kagome altermagnet, or when a moiré pseudospin texture produces a real-space gauge field whose total flux per supercell is quantized even though the local field profile is continuously tunable by interlayer bias (Watanabe et al., 2018, Tagani et al., 3 Jun 2026, Yu et al., 2019).

2. Geometric and algebraic formulations

The basic geometric object remains the Berry curvature, whose integral over a surface defines a Berry flux. In the ordinary continuum language, this is the familiar $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$. Several of the cited works reinterpret that integral in system-specific ways. For excitons in multiband semiconductors, the relevant curvature is the sum of electron and hole curvatures, $\mathcal{PT}$0, and the flux is the curvature flux through the $\mathcal{PT}$1-space area explored by the relative motion of the electron-hole pair. Because Berry curvature makes the position operators noncommutative, $\mathcal{PT}$2, one introduces commuting canonical coordinates through $\mathcal{PT}$3, which yields the effective Hamiltonian

$\mathcal{PT}$4

The last term is odd under angular-momentum reversal and is the direct Berry-curvature correction that splits $\mathcal{PT}$5 and $\mathcal{PT}$6 states (Zhou et al., 2015).

In Fermi liquids, Berry curvature modifies the symplectic structure of semiclassical quasiparticle dynamics. With a momentum-space Berry connection $\mathcal{PT}$7 and electromagnetic vector potential $\mathcal{PT}$8, the single-particle action

$\mathcal{PT}$9

leads to deformed Poisson brackets and an invariant phase-space measure

$A_x(x)$0

The physically decisive invariant is the total Berry flux through the Fermi surface,

$A_x(x)$1

which acts as the monopole charge controlling the anomaly and the chiral magnetic effect (Son et al., 2012).

For degenerate subspaces, Berry flux becomes non-Abelian. In the $A_x(x)$2 formulation for Kramers-degenerate bands, the connection is matrix-valued,

$A_x(x)$3

and the gauge-invariant content is extracted from Wilson loops,

$A_x(x)$4

Because $A_x(x)$5, its eigenvalues are $A_x(x)$6, and quantized flux is read from the winding of the gauge-invariant angle $A_x(x)$7. The non-Abelian Stokes theorem supplies the corresponding surface formulation in terms of the parallel-transported curvature $A_x(x)$8 (Tyner et al., 2021).

Strongly correlated metals motivate an additional generalization in which the relevant Berry curvature is defined not from Bloch eigenstates but from eigenvectors of the Hermitian Green’s-function combination $A_x(x)$9. The Green’s-function Berry curvature

$\chi$0

and the spin Chern number

$\chi$1

characterize topological zeros of the Green’s function rather than quasiparticle poles (Chen et al., 2024).

3. Quantized, non-quantized, and branch-dependent flux

A large part of the literature treats Berry flux as quantized. Weyl nodes are the canonical example: a closed surface surrounding a Weyl node carries nonzero and quantized Berry flux, and the node is stabilized by a Chern number $\chi$2. In TaAs, the nontrivial topology of the Berry-flux monopoles is encoded experimentally in the momentum dependence of the orbital angular momentum texture, while the spin angular momentum texture is topologically trivial with $\chi$3 in the reported case. In the Fermi-liquid anomaly problem, the flux through a Fermi surface is the integer monopole charge $\chi$4. In the non-Abelian $\chi$5 setting, the quantized magnitude of the flux is obtained from Wilson-loop winding by $\chi$6. In the orbital-selective Mott phase, the integrated Green’s-function flux is quantized, and the spin Chern number jumps by $\chi$7 when the plane crosses the nodal region, with each crossing contributing $\chi$8. In $\chi$9 topologically ordered states, adiabatic vison transport around a plaquette yields a Berry phase quantized to $\lambda_N$0 or $\lambda_N$1, depending on the symmetry-enriched topological data of the spinon sector (Ünzelmann et al., 2020, Son et al., 2012, Tyner et al., 2021, Chen et al., 2024, Chen et al., 2022).

Other settings make the opposite point: Berry flux need not be quantized once the enclosing geometry or the definition itself is altered. Inside a $\lambda_N$2-symmetric exceptional torus, the operational Berry flux

$\lambda_N$3

depends continuously on the loop radius. In the meridional slice, the explicit result

$\lambda_N$4

shows that the flux is not quantized and diverges as $\lambda_N$5, where the exceptional surface is reached. The same work emphasizes that the dynamical phase then oscillates violently. In the many-body theory of polarization, the Berry phase is likewise not unique: different ways of inserting the same total flux correspond to different Berry phases and different real-space current measurements. Yet the total charge pumped in a full Thouless cycle remains invariant. The moiré problem occupies an intermediate position: the local real-space Berry magnetic field is continuously tunable by moiré period and interlayer bias, but the total flux per supercell is quantized and changes only through a topological transition, for example from $\lambda_N$6 to $\lambda_N$7 at the critical bias $\lambda_N$8 (Wang et al., 2024, Watanabe et al., 2018, Yu et al., 2019).

This contrast is central to the subject. Functional Berry flux is often precisely the regime in which one studies how a topological flux concept survives, deforms, or becomes region- and protocol-dependent once one moves away from the simplest closed, gapped, Abelian setting.

4. Physical manifestations in bound states, transport, and driven matter

In exciton physics, Berry flux produces a geometric fine structure that is absent from the hydrogenic effective-mass model. The splitting of the lowest $\lambda_N$9-states in the two-dimensional hydrogen model is

$A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$0

so the splitting is proportional to the Berry-curvature flux through the characteristic $A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$1-space area $A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$2 sampled by the exciton envelope. In the gapped two-dimensional Dirac model this same structure appears as an effective spin-orbit coupling term, while the Darwin term gives an additional radial but rotationally symmetric shift. The paper emphasizes that in materials such as transition-metal dichalcogenides the resulting exciton fine structure can be on the order of tens of meV (Zhou et al., 2015).

In metallic transport, flux through momentum-space manifolds directly controls anomalous response. For a Fermi surface with Berry-flux charge $A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$3, the density obeys

$A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$4

which is the Fermi-liquid form of the triangle anomaly, and the equilibrium current becomes

$A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$5

In ferromagnetic CoS$A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$6, a distinct transport consequence is proposed for large linear positive magnetoresistance, with the scaling relation

$A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$7

where $A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$8 is the average Berry curvature near the Fermi surface. The reported system exhibits over $A_x(x)=\theta\sum_i p_i\,\delta(x-x_i)$9 magnetoresistance at $\sum_i p_i=1$0 Kelvin and $\sum_i p_i=1$1 Tesla, and the details section reports more than $\sum_i p_i=1$2 at $\sum_i p_i=1$3 K and $\sum_i p_i=1$4 T, remaining non-saturating up to $\sum_i p_i=1$5 T (Son et al., 2012, Zhang et al., 2022).

Weyl semimetals provide the paradigmatic monopole realization. In TaAs, bulk-sensitive soft-X-ray ARPES, circular dichroism ARPES, and spin-resolved photoemission were used to show that the orbital angular momentum texture changes sign across Weyl nodes and winds nontrivially around nodes of opposite chirality. The reported interpretation is that the non-trivial topology of the Berry flux monopoles is encoded in the momentum dependence of the orbital angular momentum, while the measured spin angular momentum reflects spin splitting but not the same topological winding (Ünzelmann et al., 2020).

Driven and self-organized systems introduce a genuinely dynamical notion of functional flux. In Berryogenesis, the time-averaged Floquet Berry flux through the Fermi sea is

$\sum_i p_i=1$6

for off-resonant graphene, so a purely linearly polarized drive does not directly generate dc Berry flux. Near plasmon resonance, however, the internal field can dominate the external drive and generate a self-consistent flux through the chirality of the collective motion, leading above threshold to spontaneous circulating motion and spontaneous time-reversal-symmetry breaking. The transition may be continuous or discontinuous depending on detuning and damping (Rudner et al., 2018).

Engineered lattice settings extend the same logic. In homobilayer transition-metal-dichalcogenide moirés, the real-space Berry phase of the layer pseudospin texture produces an effective magnetic field $\sum_i p_i=1$7 and a quantized flux per moiré supercell, with local field strengths reaching hundreds of Tesla for moiré periods around $\sum_i p_i=1$8. Twisting and biaxial strain yield the same flux sign, whereas area-conserving uniaxial strain yields the opposite sign despite an identical scalar potential landscape. In kagome altermagnets, by contrast, Berry curvature is identically zero in the strictly coplanar $\sum_i p_i=1$9 phase because of the hidden antiunitary symmetry $H_\chi$0; finite Berry curvature appears only when the orbital chiral flux term $H_\chi$1 is introduced, producing local Berry-curvature hot spots even without spin-orbit coupling and scalar spin chirality (Yu et al., 2019, Tagani et al., 3 Jun 2026).

5. Computational and experimental realizations

Berry flux becomes operational only when wavefunctions can be organized into a smooth gauge. The berry post-processing suite addresses this for two-dimensional DFT calculations by reconstructing analytic bands from Quantum ESPRESSO wavefunctions. It first removes random phases by synchronizing states at a reference real-space point, then matches neighboring states using overlaps $H_\chi$2, and finally handles degeneracies with an iterative combination of graph theory and unsupervised machine learning together with local unitary rotations. After that preprocessing, numerical $H_\chi$3-space derivatives yield Berry connections, Berry curvatures, and optical conductivities. The implementation is restricted to two-dimensional materials, and the paper emphasizes that reliable geometry requires sufficiently dense $H_\chi$4-meshes and successful handling of degeneracies (Reascos et al., 2020).

For electric polarization, Berry flux diagonalization computes the difference in formal polarization directly from plaquettes that connect two structures in combined real/reciprocal space. The plaquette holonomy

$H_\chi$5

is built from unitary approximants of overlap matrices obtained by SVD, and its eigenphases $H_\chi$6 are summed to obtain the branch-consistent electronic polarization difference. The reported high-throughput workflow uses a real-space displacement heuristic of about $H_\chi$7 Å, a singular-value criterion $H_\chi$8, and the requirement that plaquette eigenphases remain well below $H_\chi$9. On $\mathcal{T}C_{2z}$0 candidate ferroelectrics, $\mathcal{T}C_{2z}$1 results differed by less than $\mathcal{T}C_{2z}$2 from standard interpolation-based values, $\mathcal{T}C_{2z}$3 differed by less than $\mathcal{T}C_{2z}$4, and the workflow required at most $\mathcal{T}C_{2z}$5 interpolations and often $\mathcal{T}C_{2z}$6 or fewer, whereas the standard workflow typically used $\mathcal{T}C_{2z}$7 or $\mathcal{T}C_{2z}$8 interpolations (Poteshman et al., 23 Nov 2025).

The following summary organizes several representative operational frameworks:

Context	Operational object	Reported output
DFT post-processing	Phase-fixed, analytically ordered Bloch states	Berry connection, curvature, SHG conductivity
Polarization	Plaquette overlap holonomy $\mathcal{T}C_{2z}$9	Branch-consistent $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$0
$$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$1 topology	Wilson-loop eigenvalues $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$2	Quantized non-Abelian flux magnitude
Vison transport	Ground-state overlaps with parity strings	Plaquette Berry phase $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$3 or $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$4

Experimental access is equally diverse. In TaAs, soft-X-ray ARPES, circular dichroism, and spin-resolved photoemission map momentum-resolved signatures of Berry-flux monopoles. In CoS$$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$5, anomalous Hall conductivity and magnetoresistance are used to extract Berry-curvature-dominated scaling. In the non-Hermitian exceptional-surface problem, an electrical-circuit realization is proposed through a circuit Laplacian $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$6, with gain/loss implemented by inductive or negative-impedance elements, to probe the internal Berry-curvature and Berry-flux structure of exceptional surfaces (Ünzelmann et al., 2020, Zhang et al., 2022, Wang et al., 2024).

6. Gauge dependence, symmetry constraints, and conceptual subtleties

Several recurring subtleties define what functional Berry flux is not. It is not, in general, a basis-free scalar obtained by a naive local curvature integral. In the exciton Bethe–Salpeter formulation, the angular-momentum label $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$7 is gauge dependent, so one must choose a basis in which $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$8 reduces to the standard hydrogenic quantum number as $$\Phi_B=\int_S \boldsymbol{\Omega}(\mathbf{k})\cdot d\mathbf{S}$$9. In the non-Abelian setting, the matrix-valued connection itself is gauge dependent, but the Wilson-loop eigenvalues and the extracted angle $\mathcal{PT}$00 are gauge invariant and basis independent. In the many-body polarization problem, different Berry phases are all legitimate, but they correspond to different physical current measurements rather than contradictory values of a single observable (Zhou et al., 2015, Tyner et al., 2021, Watanabe et al., 2018).

Symmetry may also enforce the absence of flux effects until an additional functional ingredient is supplied. In the coplanar kagome altermagnet, $\mathcal{PT}$01 is forced to vanish everywhere by $\mathcal{PT}$02 when $\mathcal{PT}$03, even in the presence of intrinsic spin-orbit coupling. In Berryogenesis, a linearly polarized drive satisfies $\mathcal{PT}$04 and therefore does not directly generate a net dc Berry flux; the flux appears only through self-consistent instability and spontaneous chirality selection. In the exceptional-surface problem, the divergence of curvature and flux near the EP boundary is not a topological quantization effect but the singular consequence of $\mathcal{PT}$05 in the biorthogonal Berry fields. The paper there further argues that curvature streamlines diagnose whether the net flux in a given slice is zero or nonzero (Tagani et al., 3 Jun 2026, Rudner et al., 2018, Wang et al., 2024).

Taken together, these results show that functional Berry flux is a useful organizing idea precisely because it keeps geometry, gauge, symmetry, and dynamics in view simultaneously. In some systems the key question is how a quantized flux is protected; in others it is how the flux depends on radius, seam placement, plaquette construction, Floquet self-consistency, or an orbital control parameter. This suggests a broad unifying perspective: Berry flux is most informative when treated not only as a topological integer, but also as an operational field integral whose physical content depends on the manifold, subspace, and protocol by which it is realized.