The antiferromagnetic XYZ spin chain with quenched bond randomness has a critical line separating localized x- and y-polarized Ising antiferromagnetic phases. Roberts and Motrunich (2021) showed with unbiased tensor-network ("rigorous renormalization group") numerics that this line is governed by infinite-randomness fixed points with continuously varying exponents for disorder-averaged correlations, proved analogous varying exponents in a solvable "locally correlated" random XY model, and conjectured that the microscopic JzJ^z couplings — four-fermion interactions in the Majorana representation — are renormalization-group irrelevant for all coupling bandwidths J~z<1\tilde J^z<1 beyond inducing local correlations between the renormalized transverse couplings. The conjecture and its corollaries remain open: no subsequent work has treated the interacting random XYZ chain nonperturbatively. The closest result (Chepiga and Laflorencie, 2024) established stability of the noninteracting infinite-randomness fixed point against interactions in the related single-chain Ising–Majorana model, which does not settle the two-chain, (Z2)2(\mathbb Z_2)^2-symmetric XYZ problem. The surviving open core is to determine nonperturbatively the fate of interactions along the entire critical line, the dynamical exponent ψ\psi and the quantitative map to the effective local-correlation parameter δeff\delta_{\rm eff}, and the nature of the S3S_3-symmetric tricritical (XYZC) fixed point governing the transition into the z-antiferromagnetic phase.

Background

Strong-disorder renormalization group. The strong-disorder renormalization group (SDRG), introduced by Ma, Dasgupta, and Hu for random Heisenberg chains and systematized by Fisher, treats one-dimensional quantum spin chains with quenched randomness by repeatedly eliminating the largest local energy scale: the spins connected by the strongest coupling lock into a cluster, and the weaker couplings to their neighbors are renormalized perturbatively. For random antiferromagnetic XX and Heisenberg chains the flow reaches an infinite-randomness fixed point (IRFP), at which the distribution of coupling log-weights broadens without limit: the ground state is a random-singlet phase (spins paired into nearly independent singlets at all length scales), disorder-averaged spin correlations decay as power laws (η=2\eta=2 for the XX chain) while typical correlations decay much faster, and the energy EE of a singlet of length LL obeys activated scaling ln⁡(1/E)∼Lψ\ln(1/E)\sim L^{\psi} with tunneling exponent ψ=1/2\psi=1/2 (formally an infinite dynamical exponent).

Random XXZ chains and multicriticality. Fisher extended this analysis to the random XXZ chain. The critical point of the random XY chain (between x- and y-polarized Ising-ordered phases) maps to two decoupled copies of the random transverse-field Ising chain, giving ηx=ηy=4−2ϕ\eta_x=\eta_y=4-2\phi with ϕ=(1+5)/2\phi=(1+\sqrt 5)/2 the golden ratio; weak in-plane anisotropy is a marginal perturbation about the XX point, and Fisher conjectured a line of fixed points connecting the XX and XY IRFPs. In the XXZ chain with dominant JzJ^z, a transition into a z-polarized antiferromagnet occurs at a fixed point Fisher called XXZC, which is random-singlet-like. Damle and Huse analyzed a family of permutation-symmetric multicritical points at which nn types of locally ordered domains meet; their IRFPs have ψ=1/n\psi=1/n and ν=(4n+1+1)/2\nu=(\sqrt{4n+1}+1)/2 (equivalently 2n/(4n+1−1)2n/(\sqrt{4n+1}-1); for n=2n=2 this reproduces the random-singlet value ν=2\nu=2).

The random XYZ chain. Consider the antiferromagnetic XYZ chain with independent random positive bond couplings

H=∑j=1N−1(Jjx σjxσj+1x+Jjy σjyσj+1y+Jjz σjzσj+1z),H=\sum_{j=1}^{N-1}\left(J^x_j\,\sigma^x_j\sigma^x_{j+1}+J^y_j\,\sigma^y_j\sigma^y_{j+1}+J^z_j\,\sigma^z_j\sigma^z_{j+1}\right),

with all couplings drawn from distributions of the same shape and bandwidths J~x,y,z\tilde J^{x,y,z}. The model has a Z2×Z2\mathbb Z_2\times\mathbb Z_2 symmetry (global spin flips about two axes) and, when all three bandwidths are equal, a statistical S3S_3 permutation symmetry; this statistically isotropic point (denoted XYZC) is a tricritical point where the x–y critical line meets the transitions into the z-antiferromagnetic (z-AFM) phase, which sets in for J~z\tilde J^z larger than the other two bandwidths. A Jordan–Wigner transformation (the standard exact mapping of spin-1/21/2 operators to fermions) maps the model onto two coupled hopping chains of Majorana fermions — real fermion operators that are their own Hermitian conjugates — (the X and Y chains): the JxJ^x and JyJ^y couplings become the alternating hopping amplitudes on the two chains, while the JzJ^z couplings become four-fermion interactions coupling them. At Jz=0J^z=0 the problem is free-fermion; for Jz≠0J^z\neq0 it is interacting, and the SDRG generates an uncontrolled hierarchy of four-, six-, and higher-fermion terms.

Status from the source paper. Slagle, You, and Xu studied this chain with the spectrum bifurcation renormalization group (SBRG), an excited-state SDRG applicable to Hamiltonians written as sums of products of Pauli operators, and reported continuously varying exponents along the critical line, interpreting the transition as "marginal many-body localization" — eigenstates that fail to thermalize while showing a logarithmic violation of the entanglement area law. This interpretation relies on many-body localization (MBL) stability at finite energy density, which resonance-based arguments have since called into question. Roberts and Motrunich re-examined the ground state with the "rigorous renormalization group" (RRG), an unbiased tensor-network method targeting the low-energy subspace, up to N=80N=80 spins. They confirmed infinite-randomness phenomenology with continuously varying exponents for averaged correlations and an unchanged effective central charge (ceff=cln⁡2c_{\rm eff}=c\ln 2, entering S(n)∼(ceff/6)ln⁡nS(n)\sim(c_{\rm eff}/6)\ln n). A self-consistent Hartree–Fock (mean-field) treatment of the JzJ^z interactions captures much of this physics, and an exactly solved "locally correlated" random XY model — in which the JxJ^x and JyJ^y bond strengths on each link have correlation δ∈[0,1]\delta\in[0,1] — proves continuously varying exponents via a mapping to two-dimensional random walks: ηze=π/arccos⁡(−δ)\eta_z^e=\pi/\arccos(-\delta) and ηz=2π/arccos⁡(−δ)\eta_z=2\pi/\arccos(-\delta) for end-to-end and bulk zz-correlations, interpolating from the XY values (δ=0\delta=0) to the XX values (δ=1\delta=1), with ψ=1/2\psi=1/2 along the whole free-fermion line. Their central conjecture: along the interacting critical line (J~x=J~y≥J~z\tilde J^x=\tilde J^y\ge\tilde J^z), the generated interactions flow to zero while inducing a nonzero local correlation δeff\delta_{\rm eff} between the renormalized transverse couplings, so that the long-distance physics is controlled by the free line of IRFPs with exponents fixed by δeff\delta_{\rm eff}, for all J~z<1\tilde J^z<1 up to the transition into the z-AFM phase. RRG gap scaling gives ψ\psi close to but possibly drifting from 1/21/2, and at J~z→1\tilde J^z\to1 appears inconsistent with n=3n=3 Damle–Huse universality, though the authors regard this evidence as finite-size limited and state that their methods cannot access the XYZC fixed point.

Problem Statement

Determine, nonperturbatively and at strong disorder, the critical behavior of the random antiferromagnetic XYZ chain

H=∑j=1N−1(Jjx σjxσj+1x+Jjy σjyσj+1y+Jjz σjzσj+1z),H=\sum_{j=1}^{N-1}\left(J^x_j\,\sigma^x_j\sigma^x_{j+1}+J^y_j\,\sigma^y_j\sigma^y_{j+1}+J^z_j\,\sigma^z_j\sigma^z_{j+1}\right),

with independent positive random couplings of identical distribution shape — following the source, p(Jiα)=1ΓJ~α(Jiα)1/Γ−1p(J^\alpha_i)=\frac{1}{\Gamma\tilde J^\alpha}(J^\alpha_i)^{1/\Gamma-1} for Jiα∈[0,(J~α)Γ]J^\alpha_i\in[0,(\tilde J^\alpha)^\Gamma], where Γ>0\Gamma>0 is the shape parameter of the common distribution family (the source's numerics use Γ=2\Gamma=2; the specific value should have little effect on the universal low-energy physics for large enough systems) — and bandwidths J~x,y,z\tilde J^{x,y,z}, along the critical line J~x=J~y≥J~z\tilde J^x=\tilde J^y\ge\tilde J^z separating the localized x-AFM and y-AFM phases, in the normalization J~x=J~y=1\tilde J^x=\tilde J^y=1 used by the source, so that statistical isotropy sits at J~z=1\tilde J^z=1, and at its endpoint. The components of the common objective are:

(i) Fate of interactions. Establish or refute the conjecture that the four-fermion and higher interaction terms generated under renormalization are irrelevant for all J~z<1\tilde J^z<1, so that the line is controlled at long distances by the free-Majorana line of infinite-randomness fixed points with locally correlated hoppings. In particular, determine — e.g. by a systematic numerical SDRG that tracks all generated multi-fermion terms and decimates interaction terms when they are the strongest, by ground-state-projected SBRG, or by controlled unbiased numerics at larger sizes — whether the interaction strengths progressively decrease relative to the Majorana hoppings under the flow, or whether a genuine crossover to a new interacting infinite-randomness fixed point occurs at some finite J~z<1\tilde J^z<1.

(ii) Exponents along the line. Determine whether the tunneling exponent ψ\psi (the exponent in the activated dynamical scaling ln⁡(1/E)∼Lψ\ln(1/E)\sim L^{\psi} relating energy and length scales) equals 1/21/2 along the entire interacting critical line or varies with J~z\tilde J^z, separating true drift from slow crossover; and, should the irrelevance scenario of component (i) be confirmed, compute the quantitative map δeff(J~z,Γ)\delta_{\rm eff}(\tilde J^z,\Gamma) from the microscopic bandwidths and disorder distribution (of shape parameter Γ\Gamma) to the renormalized local-correlation parameter, so that the disorder-averaged correlation exponents follow from the locally correlated model (bulk and end-to-end zz-correlation exponents ηz=2π/arccos⁡(−δeff)\eta_z=2\pi/\arccos(-\delta_{\rm eff}) and ηze=π/arccos⁡(−δeff)\eta_z^e=\pi/\arccos(-\delta_{\rm eff}), with the transverse exponents obtained by the same characterization).

(iii) Tricritical point. Characterize the S3S_3-symmetric tricritical (XYZC) fixed point at statistical isotropy that controls the transition into the z-AFM phase, and the universality class of that transition — free-fermion-like IRFP, Damle–Huse-type multicritical IRFP (for n=3n=3: ψ=1/3\psi=1/3, ν=(13+1)/2\nu=(\sqrt{13}+1)/2), or a distinct non-free-fermion IRFP.

An answer must resolve components (i)–(iii) for the stated Hamiltonian; partial resolutions of individual components do not complete the problem.

References

  1. Brenden Roberts and Olexei I. Motrunich, "Infinite randomness with continuously varying critical exponents in the random XYZ spin chain", Physical Review B 104, 214208 (2021). DOI: 10.1103/PhysRevB.104.214208; arXiv:2107.12937; https://arxiv.org/abs/2107.12937 (source of the problem statement).
  2. Natalia Chepiga and Nicolas Laflorencie, "Resilient Infinite Randomness Criticality for a Disordered Chain of Interacting Majorana Fermions", Physical Review Letters 132, 056502 (2024). DOI: 10.1103/PhysRevLett.132.056502; arXiv:2305.10363; https://arxiv.org/abs/2305.10363.
  3. Daniel S. Fisher, "Random antiferromagnetic quantum spin chains", Physical Review B 50, 3799 (1994). DOI: 10.1103/PhysRevB.50.3799; https://journals.aps.org/prb/abstract/10.1103/PhysRevB.50.3799.
  4. Daniel S. Fisher, "Critical behavior of random transverse-field Ising spin chains", Physical Review B 51, 6411 (1995). DOI: 10.1103/PhysRevB.51.6411; https://journals.aps.org/prb/abstract/10.1103/PhysRevB.51.6411.
  5. Kedar Damle and David A. Huse, "Permutation-Symmetric Multicritical Points in Random Antiferromagnetic Spin Chains", Physical Review Letters 89, 277203 (2002). DOI: 10.1103/PhysRevLett.89.277203; arXiv:cond-mat/0207244; https://arxiv.org/abs/cond-mat/0207244.
  6. Kevin Slagle, Yi-Zhuang You, and Cenke Xu, "Disordered XYZ spin chain simulations using the spectrum bifurcation renormalization group", Physical Review B 94, 014205 (2016). DOI: 10.1103/PhysRevB.94.014205; arXiv:1604.04283; https://arxiv.org/abs/1604.04283.
  7. Cécile Monthus, "Strong disorder real-space renormalization for the many-body-localized phase of random Majorana models", Journal of Physics A: Mathematical and Theoretical 51, 115304 (2018). DOI: 10.1088/1751-8121/aaad14; arXiv:1712.02092; https://arxiv.org/abs/1712.02092.
  8. Sanjay Moudgalya, David A. Huse, and Vedika Khemani, "Perturbative instability towards delocalization at phase transitions between MBL phases", arXiv:2008.09113 (2020); https://arxiv.org/abs/2008.09113.
  9. Brayden Ware, Dmitry Abanin, and Romain Vasseur, "Perturbative instability of nonergodic phases in non-Abelian quantum chains", Physical Review B 103, 094203 (2021). DOI: 10.1103/PhysRevB.103.094203; arXiv:2010.10550; https://arxiv.org/abs/2010.10550.
  10. Rahul Sahay, Francisco Machado, Bingtian Ye, Chris R. Laumann, and Norman Y. Yao, "Emergent Ergodicity at the Transition between Many-Body Localized Phases", Physical Review Letters 126, 100604 (2021). DOI: 10.1103/PhysRevLett.126.100604; arXiv:2008.08585; https://arxiv.org/abs/2008.08585.
  11. Nicolas Laflorencie, Gabriel Lemarié, and Nicolas Macé, "Topological order in random interacting Ising-Majorana chains stabilized by many-body localization", Physical Review Research 4, L032016 (2022). DOI: 10.1103/PhysRevResearch.4.L032016; arXiv:2201.00556; https://arxiv.org/abs/2201.00556.
  12. Jeanne Colbois, Fabien Alet, and Nicolas Laflorencie, "Statistics of systemwide correlations in the random-field XXZ chain: Importance of rare events in the many-body localized phase", Physical Review B 110, 214210 (2024). DOI: 10.1103/PhysRevB.110.214210; arXiv:2410.10325; https://arxiv.org/abs/2410.10325.
  13. Kevin Vervoort, Wei Tang, and Nick Bultinck, "Extracting average properties of disordered spin chains with translationally invariant tensor networks", SciPost Physics Core 9, 040 (2026). DOI: 10.21468/SciPostPhysCore.9.2.040; arXiv:2504.21089; https://arxiv.org/abs/2504.21089.
  14. Yi J. Zhao, Samuel J. Garratt, and Joel E. Moore, "Superspin renormalization and slow relaxation in random spin systems", Physical Review B 112, 054436 (2025). DOI: 10.1103/ppkd-dff1; arXiv:2502.09612; https://arxiv.org/abs/2502.09612.
  15. Saurav Kantha and Nicolas Laflorencie, "Strong zero modes in random Ising-Majorana chains", arXiv:2603.05313 (2026); https://arxiv.org/abs/2603.05313.