Spontaneous magnetization and long-range order in the isotropic quantum Heisenberg ferromagnet at positive temperature ($d \geq 3$)

In 1978 Dyson, Lieb and Simon introduced the method of reflection positivity, Gaussian domination and infrared bounds for quantum spin systems, and proved the existence of phase transitions for the spin-$1/2$ $x$–$y$ model and the spin-$S \geq 1$ Heisenberg antiferromagnet in three or more dimensions. The isotropic quantum Heisenberg ferromagnet — whose treatment requires complex matrix representations of the spin algebra — was conjectured by the authors to follow by the same route, but the proof works only with real matrices and the case was left open. It has remained open for almost fifty years: Speer (1985) proved that reflection positivity itself fails for the spin-$1/2$ ferromagnet, so the specific route conjectured in the 1978 paper is obstructed and genuinely new methods are required. Modern work has validated the spin-wave approximation (the low-temperature expansion around the fully aligned state) for the free energy of the ferromagnet in dimensions one and two and for all spins $S \geq 1/2$ in three dimensions, but the order parameter itself remains uncontrolled; a 2022 authoritative review (Björnberg–Ueltschi) states that proving spontaneous magnetization or long-range order in the Heisenberg ferromagnet "remains open to this day". The open problem is to prove (or refute) that the nearest-neighbor isotropic quantum Heisenberg ferromagnet on $\mathbb{Z}^d$, $d \geq 3$, exhibits long-range order / spontaneous magnetization at sufficiently low positive temperature.

Background

Consider a finite box $\Lambda \subset \mathbb{Z}^d$ with a spin-$S$ degree of freedom at each site, i.e. the Hilbert space $\bigotimes_{x \in \Lambda} \mathbb{C}^{2S+1}$ with spin matrices $S_x^a$ ($a = 1,2,3$) satisfying $[S^a, S^b] = i\varepsilon_{abc} S^c$ and $\sum_a (S^a)^2 = S(S+1)$. The isotropic (Heisenberg) ferromagnet has Hamiltonian $H_\Lambda = -J \sum_{\langle x,y \rangle} \mathbf{S}_x \cdot \mathbf{S}_y$ with $J > 0$ and the sum over nearest-neighbor pairs; the antiferromagnet has $+J$. The finite-volume Gibbs state at inverse temperature $\beta$ is $\langle A \rangle_\Lambda = \mathrm{Tr}\, A \, e^{-\beta H_\Lambda} / \mathrm{Tr}\, e^{-\beta H_\Lambda}$, and the thermodynamic limit is $\Lambda \uparrow \mathbb{Z}^d$. "Long-range order" (LRO) means that the two-point correlation $\langle \mathbf{S}_0 \cdot \mathbf{S}_x \rangle$ stays bounded away from zero as $|x| \to \infty$ (uniformly in $\Lambda$); "spontaneous magnetization" means that a symmetry-breaking equilibrium state exists in which the SU(2)-invariant finite-volume states are averaged against an external field that is switched off after the limit. The ferromagnet is the standard model of quantum magnetism and its expected low-temperature phase is fully aligned (magnetized), while the Mermin–Wagner theorem (the rigorous result that a continuous symmetry cannot be spontaneously broken in one or two dimensions at positive temperature) forbids ordering in $d \leq 2$ at $\beta < \infty$, so $d \geq 3$ is the critical case.

The Dyson–Lieb–Simon (DLS) 1978 method, which founded the modern rigorous theory of phase transitions with continuous symmetries, works as follows. (i) Reflection positivity: the Gibbs state is positive under reflection of the lattice through a plane between sites, which DLS could establish whenever the interaction can be written with real matrices in a suitable representation of the spin algebra. (ii) Gaussian domination: the partition function with a spatially varying external field coupled linearly to the spins is bounded by the corresponding Gaussian (free-field) expression; expanded to second order this bounds the Duhamel two-point function (the time-integrated correlation $\int_0^\beta \langle S_x^a(-i t) S_y^a \rangle\, dt$), a quantum analogue of the classical Gaussian domination of Fröhlich–Simon–Spencer (1976). (iii) The infrared bound: for $k \neq 0$, $\hat{\eta}(k) \leq [2\beta\,\varepsilon(k)]^{-1}$, where $\hat{\eta}$ is the Fourier transform of the Duhamel two-point function, $\hat{J}(k) = \sum_y J(y) e^{i k \cdot y}$ is the Fourier transform of the coupling, and $\varepsilon(k) = \hat{J}(0) - \hat{J}(k)$ behaves as $|k|^2$ near $k = 0$ (the infrared, long-wavelength region), so the bound on correlations diverges there. (iv) Mode condensation: the Falk–Bruch inequality (a convexity bound, due to Falk and Bruch, relating the Duhamel two-point function to ordinary correlations) transfers the infrared bound to $\langle \mathbf{S}_0 \cdot \mathbf{S}_x \rangle$ and yields LRO at sufficiently low temperature when the infrared-mode sum $\sum_{k \neq 0} 1/\varepsilon(k)$ is finite — which holds precisely in $d \geq 3$ for short-range couplings — with an antiferromagnetic criterion that also involves the ground-state energy density. This covered the spin-$1/2$ $x$–$y$ model and the spin-$S \geq 1$ Heisenberg antiferromagnet in $d \geq 3$ (the $S = 1/2$, $d = 3$ antiferromagnet was left inconclusive by their criterion).

For the antiferromagnet one rotates the spin operators on one sublattice by $\pi$ about a spin axis; the transformed interaction can then be written with real matrices satisfying the DLS coupling condition, and the argument reaches the staggered (Néel) order parameter of anti-aligned neighboring spins. For the isotropic ferromagnet — all coupling constants of the same, ferromagnetic sign — the corresponding decomposition forces complex matrix representations of the spin algebra, the positivity inequality breaks down, and DLS could only conjecture that the Gaussian-domination bound survives (their bound (10Q)). Speer (1985) showed the situation is worse: reflection positivity fails in the spin-$1/2$ Heisenberg ferromagnet — in any dimension with a mild anisotropy above a volume-independent temperature, and in the isotropic one-dimensional model below a volume-dependent temperature — so the specific DLS route is obstructed, not merely unfinished.

The antiferromagnetic side of the program was subsequently completed: Kennedy–Lieb–Shastry (1988) proved ground-state Néel order for the spin-$1/2$ antiferromagnet in $d \geq 3$ and LRO for the $x$–$y$ model for all spins in $d > 1$; Kubo–Kishi (1988) covered the XXZ interpolating family (the anisotropic Heisenberg model with couplings $J_x = J_y \neq J_z$, interpolating between the $x$–$y$ model and the Ising model); Koma–Tasaki (1993) showed that DLS-type LRO implies spontaneous symmetry breaking for these models; and the 2022 review of Björnberg and Ueltschi records that within this reflection-positive family LRO is now proved for all $d \geq 2$ and all spins $S \geq 1/2$ except the $d = 2$, $S = 1/2$ case (where it is restricted to models close to the $x$–$y$ model). On the ferromagnetic side, a separate line of work (Conlon–Solovej 1991; Tóth 1993; Correggi–Giuliani 2012; Correggi–Giuliani–Seiringer 2014/2015; Napiórkowski–Seiringer 2021; Klippel 2026) rigorously established the validity of spin-wave theory — Bloch's low-temperature expansion in long-wavelength fluctuations above the fully aligned ground state — for the free energy of the quantum Heisenberg ferromagnet, but no result controls any correlation or the magnetization at positive temperature in $d \geq 3$.

Problem Statement

Prove, or refute, that the nearest-neighbor isotropic quantum Heisenberg ferromagnet on the hypercubic lattice $\mathbb{Z}^d$ has a phase transition at positive temperature for every dimension $d \geq 3$ and every spin $S \geq \tfrac{1}{2}$: specifically, that there exists $\beta_c(d,S) < \infty$ such that for all $\beta > \beta_c(d,S)$ the infinite-volume Gibbs state exhibits long-range order, i.e. there is $c > 0$ such that $\langle \mathbf{S}_0 \cdot \mathbf{S}_x \rangle \geq c$ for all $x$ uniformly in $\Lambda$ (equivalently, spontaneous magnetization in a symmetry-breaking equilibrium state obtained by an external field switched off after the thermodynamic limit; in the averaged form taken by Björnberg–Ueltschi 2022, the hypothesis of their Theorem 2.2 is $\frac{1}{|\Lambda|^2}\sum_{x,y}\langle S^3_x S^3_y\rangle \geq c > 0$ uniformly in $\Lambda$, which yields the corresponding first-order transition in the external field). A stronger, optional target (not required for the main claim) is any quantitative low-temperature lower bound on the spontaneous magnetization, since such a bound would imply the main claim (the DLS program's magnetization bounds are non-sharp for the models they cover and, for the ferromagnet, entirely absent). The DLS-specific route via reflection positivity is known to be obstructed (Speer 1985 proved its failure for the spin-$1/2$ ferromagnet), so the problem demands a method that does not currently exist in the literature.

References

1. F. Bloch, Zur Theorie des Ferromagnetismus, Zeitschrift für Physik 61, 206–219 (1930). doi:10.1007/BF01339661, https://doi.org/10.1007/BF01339661
2. N. D. Mermin, H. Wagner, Absence of ferromagnetism or antiferromagnetism in one- or two-dimensional isotropic Heisenberg models, Physical Review Letters 17, 1133–1136 (1966). doi:10.1103/PhysRevLett.17.1133, https://doi.org/10.1103/PhysRevLett.17.1133
3. H. Falk, L. W. Bruch, Susceptibility and fluctuation, Physical Review 180, 442–444 (1969). doi:10.1103/PhysRev.180.442, https://doi.org/10.1103/PhysRev.180.442
4. J. Fröhlich, B. Simon, T. Spencer, Infrared bounds, phase transitions and continuous symmetry breaking, Communications in Mathematical Physics 50, 79–95 (1976). doi:10.1007/BF01608557, https://doi.org/10.1007/BF01608557
5. F. J. Dyson, E. H. Lieb, B. Simon, Phase transitions in the quantum Heisenberg model, Physical Review Letters 37, 120–123 (1976). doi:10.1103/PhysRevLett.37.120, https://doi.org/10.1103/PhysRevLett.37.120
6. F. J. Dyson, E. H. Lieb, B. Simon, Phase transitions in quantum spin systems with isotropic and nonisotropic interactions, Journal of Statistical Physics 18, 335–383 (1978). doi:10.1007/BF01106729, https://doi.org/10.1007/BF01106729
7. J. Fröhlich, R. Israel, E. H. Lieb, B. Simon, Phase transitions and reflection positivity. I. General theory and long range lattice models, Communications in Mathematical Physics 62, 1–34 (1978). doi:10.1007/BF01940327, https://doi.org/10.1007/BF01940327
8. E. R. Speer, Failure of reflection positivity in the quantum Heisenberg ferromagnet, Letters in Mathematical Physics 10, 41–47 (1985). doi:10.1007/BF00704585, https://doi.org/10.1007/BF00704585
9. T. Kennedy, E. H. Lieb, B. S. Shastry, Existence of Néel order in some spin-1/2 Heisenberg antiferromagnets, Journal of Statistical Physics 53, 1019–1030 (1988). doi:10.1007/BF01023854, https://doi.org/10.1007/BF01023854
10. T. Kennedy, E. H. Lieb, B. S. Shastry, The XY model has long-range order for all spins and all dimensions greater than one, Physical Review Letters 61, 2582–2584 (1988). doi:10.1103/PhysRevLett.61.2582, https://doi.org/10.1103/PhysRevLett.61.2582
11. K. Kubo, T. Kishi, Existence of long-range order in the XXZ model, Physical Review Letters 61, 2585–2587 (1988). doi:10.1103/PhysRevLett.61.2585, https://doi.org/10.1103/PhysRevLett.61.2585
12. J. G. Conlon, J. P. Solovej, Upper bound on the free energy of the spin-1/2 Heisenberg ferromagnet, Letters in Mathematical Physics 23, 223–231 (1991). doi:10.1007/BF01885500, https://doi.org/10.1007/BF01885500
13. B. Tóth, Improved lower bound on the thermodynamic pressure of the spin-1/2 Heisenberg ferromagnet, Letters in Mathematical Physics 28, 75–84 (1993). doi:10.1007/BF00739568, https://doi.org/10.1007/BF00739568
14. T. Koma, H. Tasaki, Symmetry breaking in Heisenberg antiferromagnets, Communications in Mathematical Physics 158, 191–214 (1993). doi:10.1007/BF02097237, https://doi.org/10.1007/BF02097237
15. M. Correggi, A. Giuliani, The free energy of the quantum Heisenberg ferromagnet at large spin, Journal of Statistical Physics 149, 234–245 (2012). doi:10.1007/s10955-012-0589-4, https://doi.org/10.1007/s10955-012-0589-4
16. M. Correggi, A. Giuliani, R. Seiringer, Validity of spin-wave theory for the quantum Heisenberg model, EPL 108, 20003 (2014). doi:10.1209/0295-5075/108/20003, https://doi.org/10.1209/0295-5075/108/20003; arXiv:1404.4717, https://arxiv.org/abs/1404.4717
17. M. Correggi, A. Giuliani, R. Seiringer, Validity of the spin-wave approximation for the free energy of the Heisenberg ferromagnet, Communications in Mathematical Physics 339, 279–307 (2015). doi:10.1007/s00220-015-2402-0, https://doi.org/10.1007/s00220-015-2402-0; arXiv:1312.7873, https://arxiv.org/abs/1312.7873
18. M. Napiórkowski, R. Seiringer, Free energy asymptotics of the quantum Heisenberg spin chain, Letters in Mathematical Physics 111, 31 (2021). doi:10.1007/s11005-021-01375-4, https://doi.org/10.1007/s11005-021-01375-4; arXiv:1912.01967, https://arxiv.org/abs/1912.01967
19. J. E. Björnberg, D. Ueltschi, Reflection positivity and infrared bounds for quantum spin systems, book chapter, preprint (2022). arXiv:2204.12896, https://arxiv.org/abs/2204.12896
20. Y. Goto, T. Koma, Antiferromagnetic long-range order in a lattice fermion model, preprint (2025). arXiv:2508.12421, https://arxiv.org/abs/2508.12421
21. X. Han, A few observations around Gaussian domination and continuous symmetry breaking for spin O(N) model, preprint (2025). arXiv:2512.13449, https://arxiv.org/abs/2512.13449
22. A. Klippel, Free-energy asymptotics of the two-dimensional quantum Heisenberg ferromagnet, preprint (2026). arXiv:2608.25506, https://arxiv.org/abs/2608.25506