How do magnetic fields induce phenomena in Kitaev spin liquids?

The field changes the fractionalized response itself

In Kitaev's honeycomb model a spin-1/2 splits into two kinds of fractionalized objects: matter Majorana fermions and Z2 fluxes tied to an emergent Z2 gauge field. In the integrable limit each sector has well-defined dynamics. Real materials, however, carry additional symmetry-allowed interactions that break integrability and hybridize the gauge and matter sectors, and a magnetic field adds further mixing between the responses of different quasiparticles.

This is why the paper frames the problem as one of isolation rather than mere detection. The question is under which field-stabilized regimes, and under what conditions, the response of one specific fractionalized quasiparticle can be separated and positively understood. Fields strong enough to suppress competing order may also push the system into phases that are not perturbatively connected to the integrable limit, so the standard weak-field picture does not carry over automatically.

A spin operator flips fluxes and moves a Majorana

The paper makes the coupling concrete by rewriting Pauli operators in terms of bond fermions. A local spin operator becomes a product of a matter Majorana operator and operators that flip a pair of adjacent fluxes sharing the same link. Acting on a state therefore alters both the gauge and matter sectors at once, which is the microscopic reason a spin probe cannot, in general, address one sector alone.

This rewriting also explains why the pure Kitaev ground state has extremely short-ranged spin correlations despite gapless Dirac nodes: the extensive number of conserved fluxes constrains both static and dynamical spin structure factors. The gauge sector contributes the topological entanglement entropy of −ln 2, while the matter sector behaves as a p-wave spinon superconductor conditioned on the Z2 gauge fields. Any measurement of spin correlations is therefore a joint measurement of both sectors.

Flux disorder turns the Majorana sector metallic

The paper's central mechanism for the intermediate-field regime is flux proliferation. The magnetic field does not enter the flux-Majorana representation directly; its main effect is captured indirectly through a time-reversal-breaking perturbation, notably next-nearest-neighbor Majorana hoppings. As the field increases, the ensemble-averaged flux expectation Wp falls from unity, the flux-free value, toward zero, the half-filled flux state.

To model this, the authors introduce random sign flips on link variables with a predefined probability, mimicking flux proliferation at moderate field. Tracing out the flux sector leaves a density matrix for the Majorana fermions that is an ensemble average over disorder, producing a thermal-Majorana-metal-like state with multifractal wavefunctions and a weak anti-localization divergence in the localization length. The paper is explicit that this extended character is not directly observable in magnetic systems before parton decomposition: it cannot be probed by spin transport because there is no global spin conservation, and the lowest-energy extended states carry too little energy for conductivity measurements. What matters for experiments is instead the finite-energy Majorana density of states, which links to the spectral function.

A gapless density of states without a Fermi surface

The quantum Majorana metal is the paper's leading candidate for the intermediate-field phase of the antiferromagnetic Kitaev honeycomb model under an out-of-plane field along c. Its appeal is that it ties flux proliferation to a finite zero-energy Majorana density of states, akin to a filled-pocket Fermi surface, and to the spin and dimer response. It also offers a natural explanation for why heat transport can remain weak despite a gapless spectrum.

The paper presents this as a concrete alternative to Dirac spin-liquid or U(1) spinon Fermi-surface interpretations of recently observed gapless spin-liquid candidates, and it predicts sub-diffusive behavior when translation-invariant flux disorder is induced by the field. The authors stress that the scenario still needs further tests, especially two-dimensional real-time calculations at larger bond dimension, purification for finite temperatures, and direct comparison with material-specific extended Hamiltonians.

Entangling the sectors is not optional

The paper contrasts its entangled-ground-state ansatz with microscopic mean-field theories that solve self-consistent equations for quadratic partons. Those approaches impose translation invariance on the emergent gauge fields and therefore miss spatially fluctuating flux sectors that may be central to the intermediate-field regime. They also fail to capture entanglement between the two fractionalized sectors and do not treat fluxes as physical degrees of freedom with many-body entanglement among the six link operators inside a hexagon.

The alternative ansatz writes the ground state as a sum over disordered flux configurations, each paired with a Majorana state conditioned on that flux configuration, weighted by complex coefficients that arise from field-induced mixing. Because each individual flux configuration breaks translation symmetry while the full Hamiltonian does not, the true ground state is expected to restore it, possibly with an enlarged unit cell. This is what makes infinite-size tensor-network validation possible.

Weak fields make composite anyons sub-dimensional

In the anisotropic Abelian limit, a weak out-of-plane field produces a direction-dependent response of the Z2 gauge fluxes. The paper reports DMRG calculations on 96-site cylinders, arranged as 12 × 4 unit cells, for the two-spin-flip dynamical structure factor. The two-anyon composite built from an electric charge e and a magnetic flux m shows a one-dimensional dispersion, and the equal-energy slice of the structure factor near its gap reflects those one-dimensional soft modes.

The four-anyon composite, by contrast, shows a two-dimensional dispersion, and its equal-energy slice has large weight near the M points. The paper reads this contrast as a separation of mobility characteristics and suggests that scattering experiments such as Raman and resonant inelastic X-ray scattering could provide direct evidence of sub-dimensional dynamics of Abelian anyons and of the underlying Z2 lattice gauge theory in anisotropic Kitaev materials. The comparison is between composites of different size within the same model, not between different materials.

A critical field can reduce the 2D liquid to chains

Sub-dimensional physics can also arise without gauge-theoretic constraints. Near a critical field, the two-dimensional Kitaev spin liquid can cross over into effectively decoupled one-dimensional spin chains, because spin-orbit coupling intertwines spatial directions with exchange interactions. The paper's schematic phase diagram in anisotropy and field along the 001 crystallographic direction identifies four regions: a toric code, a gapless nodal Kitaev spin liquid, an emergent gapless phase under moderate field along the 001 crystallographic direction showing a dimensional crossover, and a partially polarized phase.

The mechanism differs from the weak-field anyon case: here the inter-chain coupling asymptotically vanishes as the field reaches the quantum critical point, so quasiparticle mobility across z bonds becomes negligible. The paper also points to an experimentally verified analogue of field-driven dimensional reduction in the three-dimensional frustrated insulator BaCuSi2O6, where a crossover to two-dimensional behavior occurs near a critical magnetic field. That analogue is offered as precedent for the mechanism, not as evidence that the Kitaev crossover has been observed.

Match the observable to the quasiparticle

The paper's practical priority is a matching problem. Weak fields in the anisotropic Abelian limit give a direction-dependent flux response and expose gapless Majorana modes in the dynamical spin structure factor without obscuring them behind the vison gap; the authors say these predictions are well captured by a simple effective Majorana tight-binding model in a field and can be tested with inelastic neutron or Raman scattering. Moderate fields, by contrast, are where the quantum Majorana metal scenario applies, with its finite zero-energy density of states and predicted sub-diffusive behavior under field-induced flux disorder.

A third regime uses the field as part of the probe. Fields along special directions decouple Kitaev chains and generate field-tunable anyon motion whose nonlinear response offers a controlled benchmark. The paper proposes pump-probe spectroscopy with two pulses separated by a variable delay: anyons created by the first pulse braid in space-time with those from the second, imprinting a statistical phase on the nonlinear response. Under nontrivial mutual statistics the high-order response shows phase-interference patterns that grow in time rather than decaying exponentially as in purely bosonic media.

The paper notes that prior simulations validated this nonlinear pump-probe technique by capturing mutual-semion braiding in the toric code, but did not address the one-dimensional mutual-fermion braiding that emerges when a weak field is applied along specific directions. Because one-dimensional fermionic anyons interfere differently along space-time trajectories, the early-time scaling of the dynamical susceptibility should produce qualitatively distinct nonlinear susceptibilities from the two-dimensional case. Varying the field angle in pump-probe experiments is therefore the paper's proposed way to tune the effective dimension of anyon dispersion and distinguish one-dimensional mutual-fermion braiding from the two-dimensional semion case.

The authors' own standard for robustness is that an interpretation should be over-constrained by independent diagnostics: ground-state observables, real-space correlations, momentum- and energy-resolved spin and dimer spectra, and the qualitative behavior of thermal and Hall transport should all agree. By that standard the quantum Majorana metal is attractive but not settled, and the field-angle-dependent nonlinear response remains a proposal rather than a demonstrated measurement.

Sources used in this answer