Two-Valley Quantum Hall Model

• The two-valley quantum Hall model is a framework describing systems with two inequivalent valleys acting as a pseudospin-½ degree of freedom in Landau-quantized or Chern-band settings.
• It combines valley-selective single-particle terms—arising from strain, displacement fields, or moiré effects—with Coulomb interactions that drive polarization, coherence, and domain formation.
• The model unifies semiconductor quantum Hall ferromagnets, graphene Landau levels, and moiré heterobilayers, providing insights into ferromagnetic, nematic, and fractional quantum Hall phases.

Searching arXiv for recent and foundational papers on two-valley quantum Hall models. The two-valley quantum Hall model is a family of effective descriptions in which two inequivalent valleys, usually denoted $\tau=\pm1$ or $\xi=\pm1$, furnish a pseudospin-$\tfrac12$ degree of freedom in a Landau-quantized or Chern-band system. Its basic structure is the competition between a valley-selective single-particle term—generated, depending on platform, by strain, displacement field, Zeeman coupling, sublattice mass, or moiré pseudogauge fields—and Coulomb interactions that can favor valley polarization, valley coherence, domain formation, or paired states. In this sense the model unifies semiconductor quantum Hall ferromagnets, graphene zero-energy Landau levels, strained honeycomb lattices, and moiré heterobilayers, and it underlies both conventional quantum Hall ferromagnetism and valley-resolved topological responses such as the quantum valley Hall and valley-polarized quantum anomalous Hall effects (Padmanabhan et al., 2010, Xie et al., 2021, Khanna et al., 2023).

1. Microscopic Hamiltonians and valley pseudospin structure

In its most direct electronic form, the two-valley model appears in AlAs as a two-dimensional electron system with two in-plane valleys treated as an isospin-$\tfrac12$ degree of freedom. The effective Hamiltonian is written as

$H=H_{\rm kin}+H_v+H_{\rm int},$

with anisotropic kinetic terms

$$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$

a strain-induced valley splitting

$$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$

and Coulomb interactions projected into the Landau-level basis. In this realization the deformation potential is $E_2\simeq 5.8$ eV, so in-plane uniaxial strain acts as a valley pseudospin Zeeman field (Padmanabhan et al., 2010).

Other realizations preserve the same two-valley logic while modifying the one-body sector. In strained honeycomb lattices the continuum Hamiltonian near the two Dirac points is

$$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$

with a strain-induced gauge potential $$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$; the associated effective magnetic field is opposite in the two valleys (Yuan et al., 2021). In AA′-stacked bilayers with space-time inversion symmetry, a minimal continuum model uses

$\xi=\pm1$0

where $\xi=\pm1$1 act on sublattice and $\xi=\pm1$2 on layer (Ghadimi et al., 2024). In bilayer graphene projected into the zeroth Landau level, the single-particle Hamiltonian is instead

$\xi=\pm1$3

so the valley degree of freedom is intertwined with spin and orbital indices from the outset (Khanna et al., 2023).

Across these platforms, the common formal step is to represent the valley label as a two-component internal space acted on by Pauli matrices. This provides a uniform language for valley polarization ($\xi=\pm1$4), intervalley coherence ($\xi=\pm1$5), and valley-selective external fields.

2. Integer quantum Hall ferromagnetism and valley nematic order

At integer filling, exchange can spontaneously polarize the valley pseudospin even when the bare valley splitting vanishes. In the AlAs model, the two valleys become degenerate as $\xi=\pm1$6, and at $\xi=\pm1$7 exchange favors a fully pseudospin-polarized ferromagnetic ground state. For an ideal 2DES the gap is

$\xi=\pm1$8

and mean-field/skyrmion theory predicts an initially infinite slope of $\xi=\pm1$9 versus $\tfrac12$0, with crossover to slope $\tfrac12$1 at large valley splitting (Padmanabhan et al., 2010).

In anisotropic valleys this ferromagnet is simultaneously a nematic. Projecting to the lowest Landau level, one introduces valley-pseudospin densities $\tfrac12$2 obeying the Girvin–MacDonald–Platzman algebra, and the local Ising order parameter is

$\tfrac12$3

The corresponding coarse-grained energy functional can be written as

$\tfrac12$4

where $\tfrac12$5 is the exchange stiffness, $\tfrac12$6 is an Ising anisotropy, and $\tfrac12$7 are uniform and random valley-Zeeman fields (Parameswaran et al., 2018).

This anisotropic structure has two immediate consequences. First, in a clean system the valley-pseudospin ordering is also spatial nematic ordering, and the clean problem supports a finite-temperature Ising transition; second, in weakly disordered systems random fields generate domains of opposite valley polarization, destroying macroscopic nematic order while leaving the quantum Hall effect asymptotically intact (Abanin et al., 2010). Domain walls between $\tfrac12$8 regions have width

$\tfrac12$9

and tension

$\tfrac12$0

and in a monodomain sample the longitudinal transport anisotropy is

$\tfrac12$1

with a sign change as the system is tuned through $\tfrac12$2 (Parameswaran et al., 2018).

3. Fractional quantum Hall states and multicomponent many-body phases

The two-valley model extends directly into the fractional regime. In AlAs, persistent quantum Hall states were observed at $\tfrac12$3 and $\tfrac12$4 even at zero strain when the two valleys are degenerate, indicating a fractional analogue of the $\tfrac12$5 quantum Hall ferromagnet. In the composite-fermion picture, $\tfrac12$6 maps to $\tfrac12$7 integer quantum Hall physics of composite fermions, which suggests

$\tfrac12$8

Experimentally, however, sample A yielded

$\tfrac12$9

with initial slopes $H=H_{\rm kin}+H_v+H_{\rm int},$0 and $H=H_{\rm kin}+H_v+H_{\rm int},$1, respectively. These values are only $H=H_{\rm kin}+H_v+H_{\rm int},$2 of the ideal single-valley spin-$H=H_{\rm kin}+H_v+H_{\rm int},$3 prediction for the zero-strain gap, and the measured gaps and slopes are lower by factors of $H=H_{\rm kin}+H_v+H_{\rm int},$4–$H=H_{\rm kin}+H_v+H_{\rm int},$5. The proposed causes are finite quantum-well thickness, large Landau-level mixing from the heavy AlAs mass, anisotropic Coulomb interactions from anisotropic Fermi contours, disorder, and possible low-energy domain-wall excitations at zero strain (Padmanabhan et al., 2010).

A more general two-component composite-fermion formulation writes

$H=H_{\rm kin}+H_v+H_{\rm int},$6

and labels candidate states by valley-resolved composite-fermion fillings $H=H_{\rm kin}+H_v+H_{\rm int},$7. A transition between two polarization sectors occurs at

$H=H_{\rm kin}+H_v+H_{\rm int},$8

while a free-composite-fermion estimate gives a fan of straight boundaries

$H=H_{\rm kin}+H_v+H_{\rm int},$9

This phase-diagram construction is intended for situations in which the active internal degree of freedom is either spin or valley (Archer et al., 2013).

Recent work has also formulated an idealized two-valley quantum Hall model with one Landau level per valley and opposite valley Chern numbers $$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$0. On a torus at total filling $$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$1, exact diagonalization of

$$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$2

finds phases denoted CFL$$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$3CFL, MR$$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$4MR, phase-separated, partially particle-hole Halperin $$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$5, partially particle-hole Halperin $$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$6, and intervalley-paired. The abstract formulation states that, in the physically relevant regime, the system initially exhibits phase-separated or valley-polarized states and can transition into paired states when onsite Coulomb repulsion is reduced (Das et al., 19 Sep 2025). This suggests that the two-valley quantum Hall model is not limited to ferromagnetic ordering, but also furnishes a controlled setting for time-reversal-symmetric fractional topological order.

4. Valley Hall, quantum anomalous Hall, and opposite-chirality formulations

A distinct branch of the subject concerns two-valley systems in which the two valleys carry opposite topological charge. In strained honeycomb lattices, the engineered gauge field produces exactly quantized Landau levels

$$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$7

and the $$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$8 sector forms a multiplet that can be relabeled by a Schwinger pseudo-spin with $$H_{\rm kin}=\sum_{\tau=\pm}\int d^2r\,\psi_\tau^\dagger(\mathbf r)\Bigl[-\frac{\hbar^2}{2m_x^\tau}\partial_x^2-\frac{\hbar^2}{2m_y^\tau}\partial_y^2\Bigr]\psi_\tau(\mathbf r),$$9. When a small inversion-breaking mass $$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$0 is present, the Berry curvature yields half-integer valley Chern numbers

$$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$1

and the valley-Hall conductivity becomes

$$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$2

Within the zeroth Landau level, the valley Hall current and the Haldane chiral edge current are both represented as pseudo-spin precession, but around different axes (Yuan et al., 2021).

In moiré MoTe$$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$3/WSe$$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$4 heterobilayers, the same opposite-valley structure emerges in a continuum setting with lattice-relaxation-induced pseudo-magnetic fields. One calculation finds that the topmost moiré bands acquire

$$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$5

At filling $$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$6, both bands are occupied, so

$$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$7

which is a time-reversal-invariant quantum valley Hall insulator. At $$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$8, Hartree–Fock valley polarization selects a single $$H_v=\frac{\Delta_v}{2}\sum_{\tau=\pm}\tau\int d^2r\,\psi_\tau^\dagger(\mathbf r)\psi_\tau(\mathbf r)=\frac{\Delta_v}{2}\sigma_z,\qquad \Delta_v=\epsilon E_2,$$9 valley band and yields a valley-polarized quantum anomalous Hall insulator with $E_2\simeq 5.8$0 (Xie et al., 2021).

A related Hartree–Fock analysis of moiré MoTe$E_2\simeq 5.8$1/WSe$E_2\simeq 5.8$2 at $E_2\simeq 5.8$3 uses a valley-resolved $E_2\simeq 5.8$4 Bloch Hamiltonian in layer-pseudospin space. There, interaction-generated interlayer tunneling alone can produce a quantum valley Hall insulating phase with

$E_2\simeq 5.8$5

even when the corresponding single-particle hopping is absent. The same framework predicts that a small Zeeman field can remove band inversion in one valley; the critical condition is

$E_2\simeq 5.8$6

beyond which the system becomes a quantum anomalous Hall insulator with one topological valley and one trivial valley. When both interaction-induced and single-particle band-mixing terms are retained, a competition between $E_2\simeq 5.8$7-wave and $E_2\simeq 5.8$8-wave topological states can appear (Saha et al., 2 Dec 2025).

5. Edge strips, domain walls, and the limits of valley topology

Transport in two-valley quantum Hall systems is often controlled by spatial structure rather than by a uniform bulk gap alone. In low-disorder Si quantum wells, activation-gap measurements at odd integer filling are well described by

$E_2\simeq 5.8$9

with

$$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$0

The proposed explanation is an edge-strip transport picture: in the quantum Hall regime, electrons cross alternating compressible and incompressible strips, and the relevant activation energy is the valley-splitting increment across one compressible strip,

$$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$1

This makes the measured mobility gap proportional to $$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$2 and strikingly independent of Hall density, and implies a density slope

$$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$3

(Wuetz et al., 2020).

Domain walls play an equally central role in interaction-driven valley order. In valley nematics, a domain wall is the interface between regions of opposite valley polarization and can be viewed microscopically as an edge between two $$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$4 liquids in different valleys. In the sharp-anisotropy limit it supports a pair of counterpropagating modes of opposite $$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$5, while in a multidomain sample the percolating domain-wall network becomes the dominant conduction path. Uniform strain biases the valleys, suppresses percolation, and restores the monodomain transport regime (Parameswaran et al., 2018).

A persistent interpretive issue is that the conventional quantum valley Hall effect is not an exact topological phenomenon. One formulation states explicitly that the total Berry curvature vanishes, and that a valley Chern number is meaningful only when inter-valley scattering is negligible. A complementary approach relates the QVHE to an exact quantum spin-Hall insulator by a unitary mapping, which permits precise statements about domain-wall spectra and robustness under $$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$6-preserving disorder (Qian et al., 2018). A stronger revision is provided by the zero-Berry-curvature QVHE, where space-time inversion symmetry forces the Berry curvature to vanish identically at every momentum but leaves a nontrivial valley Euler number

$$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$7

Across a mass domain wall, the effective one-dimensional Hamiltonian has

$$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$8

so a helical pair of metallic states can be protected without any local Berry-curvature hotspot (Ghadimi et al., 2024).

6. Phase diagrams, direct probes, and broadened realizations

In bilayer graphene, the two-valley quantum Hall model acquires an especially rich phase diagram because valley competes with spin and orbital polarization. At $$H^\xi(\mathbf p)=\xi\,v_F\bigl(\mathbf p-e\mathbf A^\xi(\mathbf r)\bigr)\!\cdot\!\boldsymbol\sigma+\Delta\,\sigma_z,$$9, a mean-field description with short-range couplings $$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$0, $$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$1, and $$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$2 predicts competition between a valley-unpolarized state, a valley-polarized state, and, when spin–valley coupling is included, a canted-antiferromagnet. The critical displacement field is

$$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$3

and the same model predicts that valley coherence may emerge for $$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$4 in the high-$$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$5 regime, corresponding to Kekulé bond ordering (Khanna et al., 2023).

The direct experimental identification of valley order in bilayer graphene uses the mapping of valley quantum number onto layer polarization in the zero-energy Landau level. In the basis $$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$6, the single-particle Hamiltonian is

$$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$7

and the measured layer polarization satisfies

$$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$8

The reported data show 32 electric-field-tuned phase transitions across states of different valley, spin, and orbital polarization, and include orbitally polarized phases stabilized by skew interlayer hopping (Hunt et al., 2016).

The same two-valley logic extends beyond conventional Landau levels. In large-angle twisted bilayer graphene, each valley carries

$$\mathbf A^\xi(\mathbf r)=\xi(\hbar/e)[A_x(\mathbf r),A_y(\mathbf r)]$$9

and a small deviation from the commensurate angle generates a two-dimensional network of one-dimensional domain-wall channels. The tunable masses are

$\xi=\pm1$00

so external electric field, lateral layer shift, and twist-angle deviation independently control the valley Hall response, the emergent sub-valley Haldane mass, and the network density (Mondal et al., 2023). In photonic analogs of staggered honeycomb lattices, the corresponding valley Chern difference across an interface produces valley-polarized propagating modes, while an interface between a QVH domain and a Zeeman-driven QAHE domain yields a perfect valley filter (Bleu et al., 2017).

Taken together, these formulations show that the two-valley quantum Hall model is not a single Hamiltonian but a structural paradigm. Its invariant content is the promotion of valley to an active internal quantum number, the use of valley-selective one-body fields as pseudospin Zeeman terms, and the competition between those fields and exchange, coherence, and topology. What varies from platform to platform is the symmetry class—Ising nematic, $\xi=\pm1$01-like pseudospin ferromagnet, opposite-Chern QVHI, zero-Berry-curvature Euler phase, or multicomponent graphene zero-energy manifold—and therefore the character of the lowest excitations: skyrmions, domain walls, edge-strip activation processes, helical domain-wall modes, or intervalley-paired states.