Ferromagnetic-impurity and gapless-bulk regimes of the exactly solved Kondo impurity at the edge of a superconducting wire

A spin-1/2 Kondo impurity at the edge of a charge-conserving one-dimensional 'superconducting' wire (a spin-gapped, Gross-Neveu-type interacting wire) is exactly solvable by Bethe ansatz, but every published analysis - the original solution of Pasnoori, Rylands and Andrei, its refinement that uncovered a Yu-Shiba-Rusinov regime, and the subsequent thermodynamic Bethe ansatz - is restricted to antiferromagnetic impurity coupling together with an attractive bulk coupling that opens a spin gap. The phases, ground-state structure, excitation spectra, boundary bound states, impurity thermodynamics, and renormalization-group flow of the same model in the ferromagnetic-impurity regime (ferromagnetic exchange $J<0$, impurity parameter $c<0$ at weak coupling, with a gapped bulk) and in the regime of bulk couplings that leave the wire gapless ($b\le 0$) have never been determined. Constructing them would complete the coupling-space phase diagram of this paradigmatic exactly solvable quantum impurity model and establish how strong quantum fluctuations treat ferromagnetic impurities in a fully fluctuating, charge-conserving superconducting wire.

Background

A magnetic impurity coupled by a spin-exchange interaction $J$ to conduction electrons defines the Kondo problem: for antiferromagnetic $J>0$ the impurity spin is screened below a dynamically generated Kondo scale $T_K$, while for ferromagnetic $J<0$ the coupling is marginally irrelevant and the impurity spin remains asymptotically free in a metallic host. When the host is a conventional mean-field BCS superconductor, the excitation gap competes with screening: the impurity creates Yu-Shiba-Rusinov (YSR) bound states localized inside the gap, and the ground state crosses between an even-parity screened singlet and an odd-parity unscreened doublet as the coupling is varied.

In one dimension the setting changes qualitatively. Long-range superconducting order is forbidden, and a faithful treatment must conserve charge: an attractive spin-exchange (backscattering) interaction between right- and left-moving electrons instead opens a spin gap $\Delta$ dynamically, producing a charge-decoupled, Luther-Emery-type 'superconducting' wire (a Luther-Emery liquid is a spin-gapped Luttinger liquid whose charge sector remains gapless). Pasnoori, Rylands and Andrei introduced an exactly solvable model of precisely this setting: a spin-1/2 impurity at the edge of a semi-infinite interacting wire, with bulk coupling $g_1$ and boundary Kondo exchange $J$ (the impurity-free limit is the Gross-Neveu model, the standard exactly solvable interacting-fermion field theory). The model is integrable for arbitrary couplings - its two-particle and particle-impurity scattering matrices satisfy the Yang-Baxter and reflection equations (the algebraic consistency conditions of factorized scattering on the half line) - and its spectrum follows from Bethe ansatz equations, i.e. algebraic quantization conditions for auxiliary rapidities $\lambda_\alpha$ that encode the spin degrees of freedom; these roots are real or grouped into strings, complexes with a common real part and equally spaced imaginary parts. The couplings enter through $b=(1-g_1^2/4)/(2g_1)$ and $c=2J/(1-3J^2/4)$; the impurity parametrization is singular at $J=\pm 2/\sqrt{3}$, where $c$ diverges and changes sign, and the combination $d=\sqrt{b^2-2b/c-1}$ is invariant along the renormalization-group flow, so it labels whole flow lines rather than bare couplings.

All published analyses of this model are restricted to $b>0$ together with antiferromagnetic impurity coupling $J>0$ (the source paper further to $c>0$, stating explicitly 'We consider here only $b,c>0$'; at weak coupling $c>0$ is the antiferromagnetic branch). In that regime the model exhibits a Kondo-screened phase (spin-singlet, odd-parity ground state) and an unscreened local-moment phase (even-parity doublet), separated at $T_K/\Delta\approx 0.32$ by a boundary quantum phase transition that the authors attributed to quantum fluctuations rather than a level crossing (a claim later refined, see below). Later work by the same groups found an additional narrow YSR regime with midgap boundary bound states and a first-order singlet-doublet level crossing between the screened and unscreened phases, and a thermodynamic Bethe ansatz (a finite-temperature integral-equation reformulation of the exact solution) then gave the impurity free energy and entropy across all antiferromagnetic phases with the full ultraviolet-to-infrared crossover. The weak-coupling renormalization-group equations $dJ/dl=-(2/\pi)J(J-g_1)$ and $dg_1/dl=-(2/\pi)g_1^2$ are known. Crucially, the Bethe equations and scattering matrices are analytic in $b$ and $c$, so the exact machinery formally extends beyond the studied regime - but no ground state, spectrum, or thermodynamics has ever been constructed for the model at ferromagnetic impurity coupling $J<0$ (whose weak-coupling image is $c<0$) or at $b\le 0$ (where the bulk interaction no longer opens a spin gap - the region $-2\le g_1<0$ or $g_1\ge 2$ in the parametrization - so the wire is a Luttinger liquid, a critical one-dimensional quantum fluid with no spin gap, rather than a spin-gapped superconductor). In nearby lattice models - Heisenberg and XXZ spin chains with boundary impurity spins - ferromagnetic couplings produce free-impurity and ferromagnetic bound-mode phases, so the corresponding regimes of the continuum wire model are expected to be nontrivial and are not settled by any existing result.

Problem Statement

Consider the exactly solvable impurity model of Pasnoori, Rylands and Andrei: a spin-1/2 magnetic impurity at the edge $x=0$ of a one-dimensional electron gas on $x\in[-L/2,0]$ with Hamiltonian $H=H_0+H_{\rm int}+H_{\rm imp}$, where $H_0=\sum_{\sigma,a}\int\psi^\dagger_{\sigma,a}\{-i\sigma\partial_x\}\psi_{\sigma,a}$ (chirality $\sigma=\pm$, spin $a=\uparrow,\downarrow$), $H_{\rm int}=2g_1\int\psi^\dagger_{+,a}\psi_{+,b}\psi^\dagger_{-,b}\psi_{-,a}$ is a bulk spin-exchange (backscattering) interaction, and $H_{\rm imp}=J\,\vec\sigma_{ab}\cdot\vec S_{\alpha\beta}\,\psi^\dagger_{-,a}(0)\psi_{+,b}(0)$ couples the impurity spin to the boundary spin density: a right mover scattering off the edge at $x=0$ flips its chirality while exchanging spin with the impurity (repeated spin indices are summed). The model is integrable for arbitrary couplings; the interactions enter the Bethe ansatz equations through $b=(1-g_1^2/4)/(2g_1)$ and $c=2J/(1-3J^2/4)$, the weak-coupling renormalization-group equations are $dJ/dl=-(2/\pi)J(J-g_1)$ and $dg_1/dl=-(2/\pi)g_1^2$, and the RG-invariant combination $d=\sqrt{b^2-2b/c-1}$ labels whole flow lines. Published exact analyses cover only $b>0$ (attractive bulk, dynamically generated spin gap $\Delta$) together with antiferromagnetic impurity coupling $J>0$ - the original 2020 solution is further restricted to $c>0$. In that domain the screened and unscreened phases, the intervening YSR window, and the impurity thermodynamics are known.

The problem is to extend the exact solution of this same model to the coupling regimes that every published analysis excludes, and to determine the model's phases and spectral properties there. Specifically:

(i) For $b>0$ (gapped bulk) and ferromagnetic impurity coupling $J<0$: construct the ground state (Bethe-root/string structure, total spin, fermion parity), determine whether the impurity is screened, remains a free local moment, or is screened only through boundary bound modes, and establish whether intragap or boundary bound states exist, including at strong $|J|$. At weak coupling this regime is the region $c<0$; since $c=2J/(1-3J^2/4)$ is singular at $J=\pm 2/\sqrt{3}$ and non-monotonic in $J$, delimiting the corresponding region of the $(b,c)$ parameter plane - including the relation of its strong-coupling parts to the antiferromagnetic analyses that approach the same region - is part of the problem.

(ii) For $b\le 0$ - the region $-2\le g_1<0$ or $g_1\ge 2$ of the parametrization, where the bulk interaction no longer opens a spin gap and the wire is gapless: analyze the same model with the impurity, identifying the resulting phases and boundary critical behavior and their relation to the known $b>0$, $J>0$ phase diagram and to the Kondo-in-Luttinger-liquid setting.

The problem requires both regimes: a solution must address (i) and (ii), and neither alone resolves the task. The overall objective they serve - and the target of a complete account - is a complete phase diagram and characterization of the ground states, excitation spectra, impurity thermodynamics, and renormalization-group flow structure of the model over its full coupling space, together with the fate of boundary bound states and screening in the ferromagnetic and gapless-bulk regimes.

References

1. P. R. Pasnoori, C. Rylands, N. Andrei, 'Kondo impurity at the edge of a superconducting wire', Phys. Rev. Research 2, 013006 (2020), DOI: 10.1103/PhysRevResearch.2.013006, arXiv:1906.06728, https://arxiv.org/abs/1906.06728
2. P. R. Pasnoori, N. Andrei, C. Rylands, P. Azaria, 'Rise and fall of Yu-Shiba-Rusinov bound states in charge-conserving s-wave one-dimensional superconductors', Phys. Rev. B 105, 174517 (2022), DOI: 10.1103/PhysRevB.105.174517, arXiv:2111.05909, https://arxiv.org/abs/2111.05909
3. P. Kattel, A. Zhakenov, N. Andrei, 'Thermodynamics in a split Hilbert space: Quantum impurity at the edge of a one-dimensional superconductor', Phys. Rev. B 113, 195155 (2026), DOI: 10.1103/qw2d-k8wv, arXiv:2508.19330, https://arxiv.org/abs/2508.19330
4. P. Kattel, A. Zhakenov, N. Andrei, 'Kondo overscreening in the presence of superconductivity', Phys. Rev. B 112, 085103 (2025), DOI: 10.1103/rg1x-bztv, arXiv:2412.01924, https://arxiv.org/abs/2412.01924
5. P. Kattel, A. Zhakenov, N. Andrei, 'Multichannel Kondo effect in one-dimensional superconducting leads', Phys. Rev. B 113, 165130 (2026), DOI: 10.1103/fzw2-xk49, arXiv:2512.11965, https://arxiv.org/abs/2512.11965
6. P. Kattel, P. R. Pasnoori, J. H. Pixley, P. Azaria, N. Andrei, 'Kondo effect in the isotropic Heisenberg spin chain', Phys. Rev. B 109, 174416 (2024), DOI: 10.1103/PhysRevB.109.174416, arXiv:2311.10569, https://arxiv.org/abs/2311.10569
7. P. Kattel, P. R. Pasnoori, J. H. Pixley, N. Andrei, 'Edge modes and boundary impurities in the anisotropic Heisenberg spin chain', Phys. Rev. B 111, 174430 (2025), DOI: 10.1103/PhysRevB.111.174430, arXiv:2412.00194, https://arxiv.org/abs/2412.00194
8. Hua-Chen Zhang, Ying-Hai Wu, Hong-Hao Tu, 'Generalized Calogero-Sutherland models for Kondo physics in the Luttinger liquid', Phys. Rev. B 108, L161102 (2023), DOI: 10.1103/PhysRevB.108.L161102, arXiv:2204.04669, https://arxiv.org/abs/2204.04669
9. L. Yu, 'Bound state in superconductors with paramagnetic impurities', Acta Physica Sinica 21, 75 (1965), DOI: 10.7498/aps.21.75, https://doi.org/10.7498/aps.21.75
10. Hiroyuki Shiba, 'Classical spins in superconductors', Progress of Theoretical Physics 40, 435 (1968), DOI: 10.1143/PTP.40.435, https://doi.org/10.1143/PTP.40.435
11. A. I. Rusinov, 'Superconductivity near a paramagnetic impurity', JETP Letters 9, 85 (1969) [Russian original: Pis'ma Zh. Eksp. Teor. Fiz. 9, 146 (1968)], NASA ADS bibcode 1969JETPL...9...85R, https://ui.adsabs.harvard.edu/abs/1969JETPL...9...85R