The exact density of states of a two-dimensional electron gas confined to the lowest Landau level with local impurity disorder was obtained in the 1980s — by Wegner for Gaussian white noise and, in a supersymmetric formulation valid for any local disorder law, by Brézin, Gross and Itzykson. The same papers left transport open: the Kubo conductivity requires the disorder average of a product of one retarded and one advanced Green function, and the field-doubling needed to represent the two opposite i0i0 prescriptions generates difference interaction terms (four-fermion terms with two independent Grassmann pairs) for which the superspace identity underlying the solvable one-particle problem fails, so the boson–fermion symmetry is lost. The subsequent literature contains the static-field Hall conductivity in this formalism (Kunz–Zumbach 1986, not quantized), 45-term Borel–Padé resummations of the divergent transport series (≈0.46 e2/h\approx 0.46\,e^2/h), an exact diagonal conductivity but only for a modified two-subband model (Hikami–Shirai–Wegner 1993), numerical benchmarks (σxx≈0.5 e2/h\sigma_{xx}\approx 0.5\,e^2/h at the band center; Huo–Hetzel–Bhatt 1993), a single-superfield representation for mixed retarded/advanced correlators tested only in random matrix theory (Barci–Oxman 2004), and a 2024 revival of the supersymmetric/Parisi–Sourlas structure that explicitly leaves the conductivity tensor uncomputed — but no exact or controlled conductivity for the original lowest-Landau-level model. The open problem is to extend the exact treatment from spectral to transport observables: derive the disorder-averaged dc conductivity tensor of the disordered lowest Landau level, equivalently its retarded/advanced two-particle Green function, exactly or with an explicitly controlled error.

Background

Setting. A two-dimensional electron gas in a strong perpendicular magnetic field BB has its single-particle spectrum quantized into highly degenerate Landau levels separated by the cyclotron energy ℏωc\hbar\omega_c. When ℏωc\hbar\omega_c is the largest scale, the dynamics can be projected onto the lowest Landau level (LLL), whose states have degeneracy κ2/2π\kappa^2/2\pi per unit area (with κ2=eB/ℏ\kappa^2=eB/\hbar) and whose wavefunctions are holomorphic functions of z=x+iyz=x+iy times a Gaussian (the Bargmann representation). A random impurity potential V(r)V(\mathbf{r}) couples the degenerate states and broadens the level; the quantity exactly computed in the 1980s is the density of states (DOS) per unit area, ρ(E)=−π−1 Im Tr (E+i0−H)−1‾\rho(E)=-\pi^{-1}\,\mathrm{Im}\,\overline{\mathrm{Tr}\,(E+i0-H)^{-1}}, the disorder-averaged spectral density of the retarded Green function.

Exact spectral solution and its supersymmetric mechanism. For Gaussian white-noise disorder, Wegner (1983) derived the exact LLL density of states from topological properties of its Feynman diagrams. Brézin, Gross and Itzykson (BGI, 1984) rederived and generalized this with a supersymmetric representation: the disorder-averaged Green function is written as a functional integral over a chiral boson–fermion superfield Φ(z,θ)=u(z)+κ2θv(z)\Phi(z,\theta)=u(z)+\frac{\kappa}{\sqrt{2}}\theta v(z), where vv is a Grassmann (anticommuting) field and θ\theta a Grassmann coordinate. An identity for powers of the superfield density (BGI's eq. (36)), combined with the holomorphy of LLL fields, reduces all functional integrals for the DOS to finite-dimensional quadratures: for any local single-site impurity statistics, characterized by the logarithm g(α)g(\alpha) of the impurity characteristic function, ρ(E)\rho(E) is given by a one-dimensional integral, with closed forms for Gaussian, Poisson (exhibiting a condensation phenomenon), and Lorentzian laws. The underlying structure is a superspace rotation symmetry mixing zz and θ\theta, implying a Parisi–Sourlas-type dimensional reduction from two dimensions to zero for the one-particle correlator. Rigorous versions of the spectral result were given by Klein and Perez (1985), and Weidenmüller (1987) extended the path-integral treatment to the level density of arbitrary Landau levels.

The transport obstruction. The static (dc) conductivity tensor σab(E)\sigma_{ab}(E), a,b∈{x,y}a,b\in\{x,y\}, is given at zero temperature by the Kubo formula, which requires the disorder average of products of one retarded and one advanced Green function, ⟨GR(E) va GA(E) vb⟩\langle G^R(E)\,v_a\,G^A(E)\,v_b\rangle (with vav_a the velocity operator along direction aa), carrying opposite i0i0 prescriptions (+i0+i0 and −i0-i0). Representing both prescriptions forces two independent field copies, and the resulting interaction terms have the schematic form [∫dθ dθˉ e−12κ2θθˉ(Φˉ1Φ1−Φˉ2Φ2)]n[\int d\theta\,d\bar\theta\,e^{-\frac{1}{2}\kappa^2\theta\bar\theta}(\bar\Phi_1\Phi_1-\bar\Phi_2\Phi_2)]^n — difference combinations to which the identity behind the z↔θz\leftrightarrow\theta superspace symmetry does not apply. Four-fermion terms depending on two distinct Grassmann pairs survive, the boson–fermion symmetry is lost, and the reduction that solved the spectral problem fails. BGI's paper records this obstruction, and the calculation of the conductivity within the formalism was thereby posed as the open problem.

Physical stakes and partial results. In the disordered LLL, all states except at the band center E=0E=0 are expected and numerically observed to be Anderson-localized, so the dc longitudinal conductivity σxx(E)\sigma_{xx}(E) vanishes except at the critical energy E=0E=0, where the localization length diverges; this is the continuum description of the integer quantum Hall plateau transition, in which the Hall conductivity σxy(E)\sigma_{xy}(E) jumps by e2/he^2/h between quantized plateaus. The band-center value σxx(0)\sigma_{xx}(0) has been benchmarked numerically at ≈0.5 e2/h\approx 0.5\,e^2/h (Huo, Hetzel and Bhatt 1993), while resummed high-order perturbation theory (Borel–Padé, a resummation technique for divergent asymptotic series; Singh and Chakravarty 1986) gives ≈0.46 e2/h\approx 0.46\,e^2/h; no exact analytic value exists. Kunz and Zumbach (1986) computed the Hall conductivity within the BGI supersymmetric formalism for a very weak static electric field, finding it not quantized — in contrast to the adiabatically switched-on case — but did not obtain the dissipative component. Hikami, Shirai and Wegner (1993) constructed a modified two-subband impurity model, different from the original single-band BGI model, for which the exact diagonal conductivity at the band center can be evaluated. Barci and Oxman (2004) proposed a single-superfield representation for mixed retarded/advanced correlators to circumvent the doubling obstruction, tested only against random matrix theory (the statistics of the eigenvalues of large random matrices). A 2024 revival of the supersymmetric structure (Kaviraj and van Vliet, identifying BGI's chiral supersymmetry as a limit of a magnetic Parisi–Sourlas theory) extended the symmetry relations to a larger class of observables and to higher Landau levels but explicitly leaves the conductivity tensor uncomputed. The exact transport theory of the disordered LLL — the disorder-averaged retarded/advanced two-particle Green function — therefore remains the open target.

Problem Statement

The problem, as left open by Brézin, Gross and Itzykson (1984) and unresolved by the subsequent literature, is to compute the transport coefficients of the disordered lowest Landau level within an exact analytic framework — extending the supersymmetric (or an equivalent exact) treatment from the density of states to the conductivity. Concretely: for a two-dimensional electron gas projected to the lowest Landau level (degeneracy κ2/2π\kappa^2/2\pi per unit area, κ2=eB/ℏ\kappa^2=eB/\hbar) with specified local impurity statistics (characterized by g(α)g(\alpha), the logarithm of the single-site characteristic function; canonical case Gaussian white noise), derive the disorder-averaged zero-temperature dc conductivity tensor σab(E)\sigma_{ab}(E), a,b∈{x,y}a,b\in\{x,y\} — equivalently the disorder-averaged two-particle Green function ⟨GR(E) va GA(E) vb⟩\langle G^R(E)\,v_a\,G^A(E)\,v_b\rangle with opposite i0i0 prescriptions — exactly, or with an explicitly controlled error.

The derivation must accommodate the two-prescription (retarded/advanced) structure that breaks the chiral superspace symmetry which made the one-particle problem solvable: in the doubled theory, interaction terms of the schematic form [∫dθ dθˉ e−12κ2θθˉ(Φˉ1Φ1−Φˉ2Φ2)]n[\int d\theta\,d\bar\theta\,e^{-\frac{1}{2}\kappa^2\theta\bar\theta}(\bar\Phi_1\Phi_1-\bar\Phi_2\Phi_2)]^n (four-fermion terms depending on two independent Grassmann pairs) survive the identities that previously reduced the functional integrals.

The conductivity tensor σab(E)\sigma_{ab}(E) sought here is specified by a definite Kubo prescription and order of limits (broadening η→0\eta\to0, frequency →0\to0, system size →∞\to\infty); it must reproduce the exact one-particle sector (the known Brézin–Gross–Itzykson density of states), be consistent with the expected and numerically observed localization of all states away from the band center (vanishing dc σxx(E≠0)\sigma_{xx}(E\neq0)), determine the critical value σxx(0)\sigma_{xx}(0), numerically benchmarked at ≈0.5 e2/h\approx 0.5\,e^2/h, including whether it is universal across local impurity laws, and reconcile the Hall component with the static-field result of Kunz–Zumbach and with the adiabatic quantization expectation, with the switching prescription made explicit.

Scope boundaries: exact extensions of the density of states to nonlocal impurity correlations, to higher Landau levels, or to the Lorentzian law above two dimensions are recorded follow-on threads of the source paper, not requirements for completion; likewise an exact result for a modified (e.g. two-subband) impurity model does not complete this problem, which concerns the original single-band lowest-Landau-level model.

References

  1. É. Brézin, D. J. Gross, C. Itzykson, Density of states in the presence of a strong magnetic field and random impurities, Nuclear Physics B 235 [FS11] (1984) 24–44. DOI: 10.1016/0550-3213(84)90146-9, https://doi.org/10.1016/0550-3213(84)90146-9
  2. F. Wegner, Exact density of states for lowest Landau level in white noise potential. Superfield representation for interacting systems, Zeitschrift für Physik B 51 (1983) 279–285. DOI: 10.1007/BF01319209, https://doi.org/10.1007/BF01319209
  3. A. Klein, J. Fernando Perez, On the density of states for random potentials in the presence of a uniform magnetic field, Nuclear Physics B 251 [FS13] (1985) 199–211. DOI: 10.1016/0550-3213(85)90257-3, https://doi.org/10.1016/0550-3213(85)90257-3
  4. H. Kunz, G. Zumbach, The static Hall conductivity in the presence of a strong magnetic field and random impurities, Nuclear Physics B 270 [FS11] (1986) 347–364. DOI: 10.1016/0550-3213(86)90558-4, https://doi.org/10.1016/0550-3213(86)90558-4
  5. R. R. P. Singh, S. Chakravarty, A disordered two-dimensional system in a magnetic field: Borel–Padé analysis, Nuclear Physics B 265 [FS17] (1986) 265–292. DOI: 10.1016/0550-3213(86)90309-3, https://doi.org/10.1016/0550-3213(86)90309-3
  6. H. A. Weidenmüller, Single electron in a random potential and a strong magnetic field, Nuclear Physics B 290 [FS20] (1987) 87–110. DOI: 10.1016/0550-3213(87)90179-9, https://doi.org/10.1016/0550-3213(87)90179-9
  7. Y. Huo, R. E. Hetzel, R. N. Bhatt, Universal conductance in the lowest Landau level, Physical Review Letters 70 (1993) 481–484. DOI: 10.1103/PhysRevLett.70.481, https://doi.org/10.1103/PhysRevLett.70.481
  8. S. Hikami, M. Shirai, F. J. Wegner, Anderson localization in the lowest Landau level for a two-subband model, Nuclear Physics B 408 (1993) 415–426. DOI: 10.1016/0550-3213(93)90373-W, https://doi.org/10.1016/0550-3213(93)90373-W
  9. D. G. Barci, L. E. Oxman, Single superfield representation for mixed retarded and advanced correlators in disordered systems, Journal of Physics A: Mathematical and General 37 (2004) 9607–9620. DOI: 10.1088/0305-4470/37/41/001, arXiv:cond-mat/0404295, https://arxiv.org/abs/cond-mat/0404295
  10. M. Ippoliti, S. D. Geraedts, R. N. Bhatt, Integer quantum Hall transition in a fraction of a Landau level, Physical Review B 97 (2018) 014205. DOI: 10.1103/PhysRevB.97.014205, arXiv:1711.04688, https://arxiv.org/abs/1711.04688
  11. A. Sinner, G. Tkachov, Diffusive transport in the lowest Landau level of disordered 2d semimetals: the mean-square-displacement approach, The European Physical Journal B 95 (2022), article 95. DOI: 10.1140/epjb/s10051-022-00358-1, arXiv:2202.02873, https://arxiv.org/abs/2202.02873
  12. A. Kaviraj, P. van Vliet, Supersymmetry in the Landau Level Problem with a Disorder Potential, arXiv:2410.00958 (2024), DOI: 10.48550/arXiv.2410.00958, https://arxiv.org/abs/2410.00958 (preprint)
  13. B. Le Floch, G. Patashuri, E. Trevisani, Unleash QQ! Cohomology, Localization, and Interpolation in Parisi–Sourlas Supersymmetry, SciPost Physics 21 (2026) 039. DOI: 10.21468/SciPostPhys.21.2.039, arXiv:2512.15899, https://arxiv.org/abs/2512.15899