Lattice Viscosity-Conductivity in Bloch Bands

• The lattice viscosity-conductivity relation is a crystalline analogue of the continuum formula, linking Hall viscosity and finite-wavevector Hall conductivity via the band-projected electric quadrupole.
• Quantum geometry—embodied by Berry curvature and quantum metric—governs the noncommutative algebra of projected coordinates and the quadrupolar spread of Bloch wave packets.
• In ideal lattice bands with uniform Berry curvature, the relation simplifies and serves as a diagnostic tool to quantify Berry-curvature fluctuations and geometric idealness.

Searching arXiv for the cited papers to ground the article with current records. The lattice viscosity-conductivity relation denotes, in the modern Chern-band sense, a relation in an isolated filled Bloch band between the Hall-viscous response to strain and the quadratic-in-wavevector part of the finite-wavevector Hall conductivity. In its most developed form, the relation is formulated as a statement about Bloch-band quantum geometry: the same band-projected electric quadrupole governs both observables, with Berry curvature fixing the projected-coordinate algebra and quantum metric fixing the quadrupolar spread of a wave packet (Shu et al., 26 May 2026). The resulting lattice relation is therefore a crystalline analogue of the continuum Hall-viscosity-conductivity formula, but not a direct transplant of Galilean-invariant hydrodynamics to a lattice.

1. Continuum antecedents and the crystalline problem

In the continuum, the relevant antecedent is the exact stress-response/conductivity relation derived for Galilean-invariant systems. Bradlyn, Goldstein, and Read showed that, for continuum systems with momentum density proportional to current, the stress response tensor is related to the $q^2$ part of the conductivity tensor at all frequencies, both with and without magnetic field; in two dimensions and at low frequency with $B\neq 0$, this yields a relation between Hall viscosity, the $q^2$ part of the Hall conductivity, the inverse compressibility, and a possible divergent shear-viscosity contribution (Bradlyn et al., 2012). That framework is the canonical continuum reference point.

The lattice problem arises because the assumptions behind the continuum identity fail in a crystal. The finite-wavevector Hall response in lattice quantum Hall systems was analyzed in detail in later work, which showed that the neat continuum relation breaks down and develops corrections due to broken rotational symmetry; at weak applied magnetic fields generic lattice wavefunctions connect smoothly to Landau levels, whereas at moderate field strengths lattice corrections perturb wavefunctions, energy levels, and transport coefficients from continuum values (Harper et al., 2018). Independent lattice studies of Hall viscosity in strong magnetic fields likewise found agreement with continuum integer-quantum-Hall results when the magnetic length is much larger than the lattice constant, with deviations increasing as field strength grows and becoming more pronounced when $C_4$ symmetry is broken to $C_2$ (Tuegel et al., 2015).

Against that background, the central modern development is the explicit formulation of a lattice viscosity-conductivity relation in an isolated filled Bloch band. The key claim is not merely that a known continuum formula can be generalized, but that the precise quantum-geometric object surviving the loss of Galilean invariance can be identified: the band-projected electric quadrupole (Shu et al., 26 May 2026).

2. Quantum geometry and the band-projected quadrupole

The lattice formulation begins from Bloch states

$$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$

for which the position operator has matrix elements

$$\langle \psi_n(\mathbf k)|\hat{\mathbf x}|\psi_m(\mathbf k')\rangle = \big[\delta_{nm} i\nabla_{\mathbf k}+\mathbf A_{nm}(\mathbf k)\big]\delta(\mathbf k-\mathbf k'),$$

with non-Abelian Berry connection

$$\mathbf A_{nm}(\mathbf k)=i\langle u_n(\mathbf k)|\nabla_{\mathbf k}u_m(\mathbf k)\rangle.$$

Projection onto an isolated single band gives the covariant coordinate

$$\hat{\mathbf X}=i\nabla_{\mathbf k}+\mathbf A(\mathbf k),$$

which is the lattice analogue of a guiding-center coordinate (Shu et al., 26 May 2026).

Its algebra is fixed by Berry curvature: $$[\hat X_\mu,\hat X_\nu]=i\Omega_{\mu\nu}(\mathbf k),\qquad \Omega_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu, \qquad \Omega=\frac12\epsilon_{\mu\nu}\Omega_{\mu\nu}.$$ This is the crystalline replacement for the continuum guiding-center noncommutativity set by magnetic field. The second geometric datum is the quantum metric,

$B\neq 0$0

the real part of the quantum geometric tensor.

The physical meaning of $B\neq 0$1 is fixed semiclassically. For a wave packet sharply peaked near $B\neq 0$2, the center $B\neq 0$3 is determined by $B\neq 0$4, while the spatial variance obeys

$B\neq 0$5

Thus the metric is the intrinsic second moment, or quadrupolar spread, of the packet. This motivates the primitive electric quadrupole

$B\neq 0$6

whose expectation value is

$B\neq 0$7

That identification is the geometric pivot of the lattice relation: Berry curvature sets the projected-coordinate algebra, and quantum metric gives the electric quadrupole carried by a localized state (Shu et al., 26 May 2026).

3. Hall viscosity in an isolated lattice band

Hall viscosity is introduced through the viscosity tensor

$B\neq 0$8

with the dissipationless Hall part antisymmetric under exchange of index pairs $B\neq 0$9. Using the strain generator $q^2$0, the static response is written as the equal-time commutator

$q^2$1

In the projected-band construction, the strain generator is expressed as

$q^2$2

where $q^2$3 is a dilation operator and the $q^2$4-dependent pieces drop out of the dissipationless commutator algebra relevant for Hall viscosity (Shu et al., 26 May 2026).

For the isotropic Hall-viscosity component, the main lattice-band result is

$q^2$5

This is a genuinely new statement for lattice bands: Hall viscosity is expressed not by Berry-curvature moments alone, but by the full quantum geometry through the ratio $q^2$6. With the paper’s normalization for a filled single band,

$q^2$7

The isotropic scalar $q^2$8 is only a special projection. More generally, Hall viscosity on a lattice remains a rank-4 antisymmetric tensor. In two dimensions the antisymmetric Hall-viscosity tensor has six independent components in general, and only with continuous rotational symmetry does it reduce to a single scalar coefficient multiplying the rotationally invariant antisymmetric tensor structure (Shu et al., 26 May 2026). This point is essential in reduced-symmetry lattices, where the $q^2$9, $C_4$0, and $C_4$1 projections of $C_4$2 generate distinct Hall-viscosity components.

4. Nonlocal Hall conductivity and the common geometric content

On the transport side, the relevant observable is the small-wavevector Hall conductivity

$C_4$3

with $C_4$4. In the adiabatic single-band approximation for a filled insulating band, the quadratic term is

$C_4$5

Its derivation uses semiclassical wave-packet dynamics in a spatially nonuniform electric field,

$C_4$6

where the second term in $C_4$7 is the geometric correction due to the wave packet’s electric quadrupole (Shu et al., 26 May 2026).

The comparison with Hall viscosity is most transparent after reorganizing the conductivity. One form is

$C_4$8

which makes the first lattice-specific correction explicit through the covariance of metric and curvature. In the isotropic $C_4$9 channel, the same quantity is rewritten as

$C_2$0

This is the most direct generic formulation of the lattice viscosity-conductivity relation: the same Brillouin-zone average $C_2$1 that enters Hall viscosity also enters the nonlocal Hall conductivity, now multiplied by $C_2$2 and corrected by a covariance term (Shu et al., 26 May 2026).

The conceptual consequence is that the “same geometric object” appearing in both responses is the band-projected electric quadrupole, whose expectation value is $C_2$3, dressed by the noncommutative projected-coordinate algebra set by $C_2$4. The lattice relation is therefore not an exact Galilean-invariant identity; in a crystal, velocity is not proportional to momentum. Instead, it isolates the common geometric content that survives in Bloch bands.

5. Ideal bands, trace condition, and transport diagnostics

The relation simplifies sharply in the ideal-band regime. The relevant ideal-band criterion is the trace condition

$C_2$5

With the additional assumption that Berry curvature has definite sign over the Brillouin zone, $C_2$6 becomes constant up to sign, the covariance term in the reorganized conductivity vanishes, and the lattice relation becomes

$C_2$7

with

$C_2$8

The magnetic-length factor of the continuum Landau-level formula is replaced by the band-geometric average

$C_2$9

Accordingly, the deviation from the Landau-level form is quantified entirely by the dimensionless Berry-curvature fluctuation $$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$0 (Shu et al., 26 May 2026).

The geometric idealness criterion is tied to the local quantum-geometric inequality

$$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$1

Integrating this inequality yields a lower bound $$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$2 for the isotropic Hall-viscosity density. In ideal sign-definite bands, the trace condition implies a holomorphic or antiholomorphic structure in $$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$3, and if $$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$4 is also uniform then $$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$5, so the continuum result is recovered exactly (Shu et al., 26 May 2026).

The same structure yields an experimentally motivated nonlocal Hall ratio,

$$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$6

For a filled isolated Chern band, $$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$7. In an ideal sign-definite band,

$$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$8

while in the Landau-level limit $$|\psi_n(\mathbf k)\rangle = e^{i\mathbf k\cdot \hat{\mathbf x}}|u_n(\mathbf k)\rangle,$$9. Nonlocal Hall transport is therefore proposed as an electrical diagnostic of geometric idealness, because the finite-$$\langle \psi_n(\mathbf k)|\hat{\mathbf x}|\psi_m(\mathbf k')\rangle = \big[\delta_{nm} i\nabla_{\mathbf k}+\mathbf A_{nm}(\mathbf k)\big]\delta(\mathbf k-\mathbf k'),$$0 Hall response measures not only Hall-viscous content but also Berry-curvature fluctuations (Shu et al., 26 May 2026).

6. Tensorial scope, assumptions, and related usages of the term

The lattice-band relation rests on a specific regime. The analysis assumes an isolated single Bloch band in the adiabatic regime, with external perturbations weak compared to interband gaps so that interband mixing can be neglected. The conductivity formula applies to a filled insulating band with chemical potential in the bulk gap, in the clean limit, and to second order in small wavevector or electric-field gradients. Rotational symmetry is not assumed in the construction of the quadrupole or viscosity tensor, but the compact formulas for $$\langle \psi_n(\mathbf k)|\hat{\mathbf x}|\psi_m(\mathbf k')\rangle = \big[\delta_{nm} i\nabla_{\mathbf k}+\mathbf A_{nm}(\mathbf k)\big]\delta(\mathbf k-\mathbf k'),$$1 and the $$\langle \psi_n(\mathbf k)|\hat{\mathbf x}|\psi_m(\mathbf k')\rangle = \big[\delta_{nm} i\nabla_{\mathbf k}+\mathbf A_{nm}(\mathbf k)\big]\delta(\mathbf k-\mathbf k'),$$2 Hall conductivity relation emphasize the isotropic sector. In reduced-symmetry lattices, one must retain the full tensorial structure, replacing $$\langle \psi_n(\mathbf k)|\hat{\mathbf x}|\psi_m(\mathbf k')\rangle = \big[\delta_{nm} i\nabla_{\mathbf k}+\mathbf A_{nm}(\mathbf k)\big]\delta(\mathbf k-\mathbf k'),$$3 by $$\langle \psi_n(\mathbf k)|\hat{\mathbf x}|\psi_m(\mathbf k')\rangle = \big[\delta_{nm} i\nabla_{\mathbf k}+\mathbf A_{nm}(\mathbf k)\big]\delta(\mathbf k-\mathbf k'),$$4 and replacing the scalar Hall viscosity by the appropriate tensor projections (Shu et al., 26 May 2026).

This tensorial scope also clarifies a common misconception. The modern lattice viscosity-conductivity relation does not restore the continuum Galilean identity in exact form. Rather, it reconstructs the relation from Bloch-band quantum geometry. Earlier lattice finite-$$\langle \psi_n(\mathbf k)|\hat{\mathbf x}|\psi_m(\mathbf k')\rangle = \big[\delta_{nm} i\nabla_{\mathbf k}+\mathbf A_{nm}(\mathbf k)\big]\delta(\mathbf k-\mathbf k'),$$5 studies had already shown that the continuum Hall-viscosity formula acquires nonuniversal lattice corrections once rotational symmetry and continuum kinematics are broken (Harper et al., 2018). The geometric formulation sharpens that observation by isolating the common quadrupolar content and by identifying the ideal-band limit in which the remaining deviation is quantified solely by Berry-curvature fluctuations (Shu et al., 26 May 2026).

A second source of confusion is terminological. In the rigorous theory of non-interacting lattice fermions at equilibrium, a distinct “viscosity-conductivity relation” appears in which the conductivity measure of a finite lattice region is reconstructed as the boundary value of the Laplace-Fourier transform of a quantum current viscosity. That viscosity is explicitly not a mechanical shear viscosity; it is a current-response kernel describing how diamagnetic current induces paramagnetic current through equilibrium current commutators (Bru et al., 2016). The Hall-viscosity relation in lattice Chern bands and the current-viscosity reconstruction in equilibrium lattice fermions therefore concern different observables, different generators, and different response problems, despite the shared phrase.

In the Chern-band setting, the defining content of the lattice viscosity-conductivity relation is therefore precise. Berry curvature provides the noncommutative projected-coordinate algebra, quantum metric provides the electric quadrupole or wave-packet spread, and that same projected quadrupole controls both the Hall-viscous response to strain and the nonlocal Hall response to electric-field gradients. The relation is not universal in the continuum sense, because it contains explicit metric-curvature covariance corrections and Berry-curvature-fluctuation corrections. Precisely for that reason, finite-wavevector Hall transport becomes a diagnostic of quantum geometry and of geometric idealness in lattice quantum Hall bands.