Exponential decay in $O(n)$-invariant quantum spin systems

Jakob E. Bj\"ornberg; Kieran Ryan

REVIEW 3 major objections 6 minor 21 references

Exponential decay in $O(n)$-invariant quantum spin systems

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read For O(n)-invariant quantum spin systems on the integer lattice, taking the spin dimension n large forces exponential decay of spin-spin correlations, answering Ueltschi's question about the persistence of long-range order.

desk verdict A likely-correct answer to Ueltschi's question, with a real but non-fatal gap in a combinatorial lemma and a small overstatement at the origin. read the letter →

arxiv 2506.22254 v1 pith:R4QEXL7S submitted 2025-06-27 math-ph math.MPmath.PR

Jakob E. Björnberg , Kieran Ryan This is my paper · ORCID

classification math-phmath.MPmath.PR MSC 82B2082B1060K3582B26

keywords O(n)-invariantquantumspinsystemsrandomloopmodelexponentialdecayofcorrelationsreflectionpositivitychessboardestimatesPeierlsargumentspin-spinlong-rangeorder

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper establishes that O($n$)-invariant quantum spin systems on the integer lattice have exponentially decaying spin-spin correlations once the spin dimension $n$ is large, for any dimension $d \ge 1$ and anisotropy $u \in [0, \tfrac12]$. This answers a question of Ueltschi, who had proved long-range order for small $n$ in dimensions $d \ge 3$: large $n$ gives the opposite behavior, in every dimension. The proof works through the random loop representation, where the central object is the probability that two space-time points lie on the same loop, and shows this probability decays exponentially. The result matters because it sharply separates a small-$n$ ordered regime from a large-$n$ disordered regime for a family of models including Heisenberg-type spin systems, and because the proof introduces a Peierls-plus-chessboard mechanism for quantum loop models.

What carries the argument

The load-bearing object is the random loop model on the space-time torus $\Lambda(k)\times[0,\beta]$: a Poisson process of crosses and double bars on edges, reweighted by $n^{\ell(\omega)}$, where $\ell(\omega)$ is the number of loops; large $n$ therefore favors many small loops. The proof runs through three linked mechanisms: reflection positivity of an auxiliary spin-colouring measure, which yields the chessboard estimate (Proposition 3.1) bounding probabilities of events in many cubes by powers of their distributed probabilities; a Peierls-type geometric lemma (Lemma 3.2) showing that a loop connection forces a path through cubes with a positive fraction of 'bad' events (crowded, empty, or transposition); and loop-counting estimates (Lemma 3.3) bounding the distributed probabilities of these bad events. The core of the counting is Lemma 3.4, which asserts that on the distributed crowded or transposition events there are at least $m_0 = n\beta K/(4R)$ links that do not close a loop; this deficit produces the $n^{-1/5}$ suppression that makes the Peierls sum converge.

What would settle it

Check the unverified 'one can check' step of Lemma 3.4 case (a) by enumerating all relative temporal orders of the four links in a reflected boundary 4-cycle within one time slab; if any order yields fewer than two disjoint switch pairs on non-incident edges, the $K/4$-per-slab bound fails and the large-$n$ suppression mechanism collapses.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: for $u \in [0, \tfrac12]$ and $d \ge 1$, given any decay rate $c>0$ there exist $n_d \in \mathbb{N}$ and $\alpha_d>0$ such that for all integers $n>n_d$ and all $\beta>\alpha_d/n$, the truncated spin-spin correlations on the torus $\Lambda(k)$ satisfy $|\langle S_0^{(i)}; S_x^{(i)}(t)\rangle_{\Lambda(k),\beta}| \le e^{-c(\|x\|_1+|t|)}$ for $i=1,2,3$, uniformly in the torus side lengths and in $\beta$. The probabilistic counterpart (Theorem 2.1) states the same exponential bound for loop-connection probabilities $P_n[(0,0)\leftrightarrow(x,t)]$. The mechanism is that large $n$ makes loops small: the weight $n^{\ell(\omega)}$ favors many loops, and a connection between distant points forces a long chain of 'bad' space-time cubes whose probability is shown to be exponentially small. This answers Ueltschi's question in the negative, since his long-range order for small $n$ in $d \ge 3$ cannot persist for large $n$.

Load-bearing premise

The proof hinges on a geometric counting step, checked only by inspection, that in each reflected four-edge boundary cycle two opposite edges always produce two disjoint pairs of links that cannot close loops in every time slab; if this count failed, the large-$n$ suppression of the bad events would collapse.

Editorial extensions

If this is right

For every dimension $d \ge 1$ and every $u \in [0,\tfrac12]$, once $n$ is large enough spin-spin correlations decay exponentially, so Ueltschi's small-$n$ long-range order in $d\ge3$ cannot extend to all $n$.
The exponential bound transfers to any correlation expressible as a loop-connection probability, including the elementary-operator correlations $\langle E^{a,b}_x(s) E^{a,b}_y(t)\rangle$ for $a\neq b$ (Remark 1.2).
The same result applies to Ueltschi's closely related model with the projection $P_{xy}$ for odd $n$, giving exponential decay of correlations of the form $\langle(S_x^{(i)}(s))^2;(S_y^{(i)}(t))^2\rangle$.
In two dimensions the result upgrades the previously known polynomial decay of correlations to exponential decay for large $n$ and $u\in[0,\tfrac12]$, consistent with the absence of continuous symmetry breaking.
The temperature restriction $\beta>\alpha_d/n$ is an artifact of the proof: a separate stochastic-domination argument gives the same exponential decay for all small $\beta$, leaving only an intermediate gap that the authors expect to close.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the switch-counting technology of Lemma 3.4 is as robust as it appears, the same chessboard-plus-Peierls scheme could deliver exponential decay for other loop-model observables, such as probabilities of two disjoint loops connecting specified pairs of points.
The large-$n$ exponential decay suggests viewing the loop model as a 2D loop-$O(n)$ model in which $n$ plays the role of the classical spin dimension; the paper leaves open the minimal $n$ for which decay holds for all $u\in[0,1]$, and $n=3$ (the bilinear-biquadratic Heisenberg model) is the natural test case.
A direct Monte Carlo measurement of $P_n[(0,0)\leftrightarrow(x,t)]$ for $n$ in the range 8–16 could empirically locate the crossover from small-$n$ order to large-$n$ disorder and test whether the rate $c$ grows with $n$ as the proof suggests.
Because the proof relies on reflection positivity, it cannot reach $u>\tfrac12$; a version avoiding reflection positivity would be needed to decide whether exponential decay persists into the ferromagnetic-anisotropy region where long-range order remains open.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

The paper proves that for O(n)-invariant, reflection-positive quantum spin systems with u in [0,1/2], taking n large forces exponential decay of spin-spin correlations in every dimension d. That is new, and it is the expected complement to Ueltschi's small-n long-range order for d ≥ 3. The loop-model theorem (Theorem 2.1) is the real content; Theorem 1.1 follows from Ueltschi's correlation identities.

The structure is sound. The proof uses a chessboard estimate on an auxiliary reflection-positive measure, coupled to the loop measure, plus a Peierls argument. The key quantitative steps—the loop-closing probability bound (1-u)/(2d), the stochastic domination of X_j by Bernoullis, and the dominance of the (n(1-u))^{-1/5} factor over the O(K'β) exponentials—check out. The paper is honest about the beta-window gap and the u ≤ 1/2 restriction.

The soft spots are real but not load-bearing. First, the statement of Theorem 1.1 is false at (x,t) = (0,0): the truncated correlation is (n^2-1)/12, while the bound gives 1. Trivial fix—exclude the origin or handle it separately. Second, Lemma 3.4(a), the switch-counting in the boundary case, really is deferred to 'one can check.' That is the one place I could not fully verify. The stress-test note worries that failure would kill the n-suppression; I think that overstates it. Even if the count drops from K/4 to K/8 per slab, the Peierls factor still gets an n^{-1/8} suppression, which is enough to drive the bound to zero as n → ∞ for any fixed c. So the theorem is unlikely to hinge on the exact constant, but the claim needs a written proof. Third, equation (3.21) contains an arithmetic slip—the generating function appears to be missing a factor d in the exponent. It does not change the qualitative conclusion.

The citation pattern is fine: Ueltschi's loop identities, Biskup's chessboard framework, Georgii-Küneth, and the authors' companion paper are all used appropriately. No circularity.

Who is this for? Anyone working on random loop models or rigorous quantum spin order/disorder. It deserves a serious referee; I would send it out. With the (0,0) statement fixed and Lemma 3.4(a) expanded, I'd accept it.

Referee Report

3 major / 6 minor

Summary. The paper proves exponential decay of spin-spin correlations in O(n)-invariant quantum spin systems on the torus Λ(k) for large n, for u in [0,1/2] and β > α_d/n, complementing Ueltschi's long-range order for small n in d ≥ 3. The proof works through a random loop representation, a Peierls-type argument with chessboard estimates, and a large-n suppression of loop-closing events. The main probabilistic result is exponential decay of loop-connection probabilities, and the spin correlation bound is derived from it via Ueltschi's identities.

Significance. If the proof is completed as indicated, this is a substantial contribution: it answers a question of Ueltschi and provides the first rigorous large-n exponential decay result for this family in all dimensions. The paper clearly identifies the external black boxes (Ueltschi's correlation identities, Biskup's chessboard framework, Georgii–Küneth domination) and organizes the argument carefully. The central mechanism is falsifiable: the bound (3.10) has explicit constants and the decay rate can be made arbitrarily large. The paper also provides a probabilistic theorem (Theorem 2.1) that is stronger than the spin statement and likely transferable to other O(n)-invariant observables.

major comments (3)

[Theorem 1.1, Section 1.1] The statement of Theorem 1.1 is false at (x,t) = (0,0) for n ≥ 4. By Lemma 2.2, the truncated correlation ⟨S^{(i)}_0 ; S^{(i)}_0(0)⟩ equals (n²−1)/12, which exceeds 1, while the claimed bound is e^{-c·0} = 1. The 'immediate' derivation from Lemma 2.2 and Theorem 2.1 is also not immediate for other points because the prefactor (n²−1)/12 must be absorbed into the exponential rate, which requires a rate adjustment that depends on n and, for short distances, on the distance itself. Please rephrase the theorem to exclude (0,0) or to include the prefactor (e.g., bound by (n²−1)/12 e^{-c(‖x‖₁+|t|)}), and give the absorption argument.
[Lemma 3.4, Section 3.3] The switch-counting claim in case (a) is deferred to 'one can check.' This is load-bearing because it produces the m0 = nβK/(4R) bound on links that do not close a loop, which feeds the n-suppression in Lemma 3.3 and hence the decisive Peierls factor in (3.10). I verified the claim: in each 4-edge component, the edge with the largest temporal coordinate has an opposite edge; the two edges opposite the maximum edge each have an outgoing switch, and their lower links are non-incident. The sharing bound (each link used by at most two switches) then gives at least K/4 disjoint switches per slab. The argument is short and should be included explicitly, as the present text leaves a central combinatorial step to the reader.
[Section 2.2, Lemma 2.2 and Theorem 1.1] The factor (n²−1)/12 in Ueltschi's correlation identities is incompatible with the no-prefactor form of Theorem 1.1 without further reasoning. In particular, absorbing this factor into the exponential rate requires the rate to depend on n, so the uniformity in n asserted in Theorem 1.1 needs a careful statement. Specify the intended theorem (for example, with the prefactor included) and provide the absorption argument, or state the theorem only for (x,t) with ‖x‖₁+|t| ≥ 1 and a rate that may depend on n through the prefactor.

minor comments (6)

[Section 3.1, equation (3.8)] In (3.8), 'θtiB' should read 'θqiB'.
[Section 3.1, definition of Δ] The symbol T* is used in the discussion of Δ but never defined, and the phrase 'maximum distance between points in neighbouring cubes' is inaccurate if taken literally for arbitrary continuous points; the value 3d+R/n appears to be a conservative upper bound for the relevant endpoints (lattice vertices and arbitrary times). Please rephrase and define T*.
[Section 3.3, equation (3.27)] The factor n^{3R2^{d−1}/βn} in (3.27) is easy to misread; please rewrite it with explicit parentheses or a displayed fraction, and double-check the exponent.
[Section 4, Proposition 4.3] The proof of reflection positivity is sketched; the treatment of the i.i.d. random variables for edges not bisected by planes is terse but acceptable. A sentence explaining why conditioning on η_p does not affect these variables would improve clarity.
[Section 1, equation (1.8)] The statement that (1.3) is equivalent to the bilinear-biquadratic Hamiltonian (1.8) is a bit abrupt; a brief indication of the parameter identification (e.g., how u and the coefficient of (S_x·S_y)² relate) would help the reader.
[References] The companion paper [6] is cited for Lemma 2.4; please verify that the lemma numbering is correct in the final version and that the citation to [11, Theorem 1.1] for stochastic domination is the intended precise statement.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the exponential-decay bound is derived in-paper from the model via external black boxes and a direct Peierls argument.

full rationale

The paper's central claim is derived in-paper from the model definition. The loop representation and the identity (Lemma 2.2) relating spin-spin correlations to loop-connection probabilities are imported from Ueltschi [21], an external and independent source; the authors' own papers [5] and [6] appear only for context and for an optional extension in Remark 1.2, not as load-bearing support. The chessboard estimate (Proposition 3.1) is proved in Section 4 via a reflection-positive measure and a coupling to the loop measure, with the reflection-positivity argument given in detail rather than imported. The Peierls bound in Theorem 2.1 is a direct computation: the large-n suppression comes from the switch-counting Lemma 3.4 and the stochastic domination used in Lemma 3.3, neither of which assumes the target exponential decay. The only unproven step is the geometric 'one can check' claim in Lemma 3.4(a), and the translation from Theorem 2.1 to Theorem 1.1 involves a prefactor (n^2-1)/12 that is not addressed; both are potential correctness gaps, not instances of circular reasoning. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work, and no ansatz is smuggled in via citation. The derivation is therefore self-contained modulo external black boxes from other authors, and the honest finding is no significant circularity.

Assumptions & free parameters 2 free parameters · 6 assumptions · 1 invented entities

Pure mathematics: the central claim depends on no fitted parameters. The inputs are: (i) the loop representation of the Gibbs state with correlation identities imported from Ueltschi [21] (Lemma 2.2); (ii) the reflection-positivity and chessboard-estimate framework (Biskup [4] plus Section 4 of this paper), which requires u ≤ 1/2; (iii) the stochastic domination result of Georgii-Küneth [11]; (iv) Mecke's formula [15]. The only hand-chosen quantities are proof constants (R, R0, thresholds n_d, α_d), not data fits. The auxiliary reflection-positive measure P_{ι2} is a fully explicit construction whose coupling to the loop measure is proven, so it carries its own verification rather than an external falsifiable handle.

free parameters (2)

Coarse-graining scale R (and R0) = large, depending on d and the target rate c (no explicit value given)
Chosen by hand in Section 3.1 to define the big cubes of height R/n; R must be large enough to make the Peierls bracket in (3.10) small and the geometric series converge. It is a proof device, not a fit to data.
Thresholds n_d and α_d = existentially quantified, no explicit bounds
The theorem asserts their existence; the proof constructs them implicitly by 'R large enough, then n large enough'. They depend on u, d, c but not on k or β, and they are not fitted to any data.

assumptions (6)

domain assumption Loop-representation correlation identities of Ueltschi: ⟨S^(1)S^(1)⟩ = ⟨S^(3)S^(3)⟩ = (n²-1)/12 P_n(connection), with the S^(2) version an inequality (Lemma 2.2).
Imported from [21, Theorem 3.3] as a black box; this is the bridge converting the probabilistic Theorem 2.1 into the spin-statement Theorem 1.1, including the prefactor (n²-1)/12 that creates the (0,0) overclaim.
domain assumption Reflection positivity holds exactly for u ∈ [0,1/2] and fails for u > 1/2 (Ueltschi [21]); the auxiliary measure P_{ι2} in (4.5) is reflection positive across the plane set π'.
The chessboard estimate (Proposition 3.1) is restricted to u ∈ [0,1/2] because the intensity set II_{a,b} requires 1-2u ≥ 0 (Section 4). The whole theorem inherits this restriction.
standard math Standard chessboard-estimate framework for reflection-positive measures, including subadditivity [4, Lemma 5.9].
Invoked in (3.6), (3.7), (3.15), (3.25); the paper proves Proposition 3.1 by reducing to this framework for P_{ι2}.
standard math Stochastic domination of the link process by a Poisson process of intensity n (Georgii-Küneth [11, Theorem 1.1]).
Used in Lemma 3.3 to bound the total number of links on the event W (M = e 2dnβK' bound, leading to P_n[W^c] ≤ e^{-ndK'β}).
standard math Mecke's formula for conditioning on Poisson points (Last-Penrose [15, Theorem 4.4]).
Used in Lemma 4.2 and in the loop-counting argument of Section 3.3 to condition on the locations of exactly m links.
domain assumption Symmetry reduction ⟨S^(i)_0 ; S^(i)_x(t)⟩ = ⟨S^(i)_0 S^(i)_x(t)⟩ via [21, Lemma 3.1].
Used immediately after Theorem 1.1 to drop the expectation product in the truncated correlation.

invented entities (1)

Auxiliary reflection-positive measure P_{ι2} with coupling Q_n independent evidence
purpose: The loop measure P_n itself is not reflection positive; P_{ι2} on (η, σ) pairs is reflection positive across cube boundaries and is coupled to P_n (Propositions 4.3, 4.4) so that the chessboard estimate applies to P_n-events independent of link type on boundaries.
This is a fully explicit mathematical construction (intensities (4.5), coupling (4.9)-(4.11)) whose properties are proven inside the paper; it is a proof device with an internal certificate, not a physical entity with external predictive consequences.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Exponential decay in $O(n)$-invariant quantum spin systems." pith.science (2026). https://pith.science/paper/R4QEXL7S

@misc{pith2026250622254,
  author       = {Pith},
  title        = {Pith review of: Exponential decay in $O(n)$-invariant quantum spin systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R4QEXL7S}},
  note         = {Machine review of arXiv:2506.22254}
}

abstract

We consider $O(n)$-invariant and reflection-positive quantum spin systems on the integer lattice in any dimension, and prove that spin-spin correlations decay exponentially fast provided n is large enough. This answers a question of Ueltschi, who proved that for small n there is instead long-range order (for d at least 3).

Figures

Figures reproduced from arXiv: 2506.22254 by the authors.

**Figure 1.** Example of a configuration ω in dimension d = 1. The links ( and ) create loops, which wrap around in the vertical direction. 2.2. Main result for the loop model. We write (x, s) ↔ (y, t) for the event that (x, s) and (y, t) belong to the same loop. The following is the probabilistic counterpart to Theorem 1.1. Recall the torus Λ(k) from (1.6). Theorem 2.1 (Exponential decay in the loop model). Let d ≥ 1 and u ∈ [0,… view at source ↗

**Figure 2.** The same configuration ω as in [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Illustration of the three bad events C, E, T. The big cube S0,0 is drawn with a solid outline, its four constituent small cubes are indicated with dotted lines. Left: a configuration in C ∩ E. The two edges with common endpoint x2 both have links in S0,0, as required for C, and the top left small cube containing x1 (shaded grey) has no links, as required for E. Right: the transposition event T, where at least one cr… view at source ↗

Figures from the paper (7 more)

**Figure 4.** Figure 4: The two possibilities for the first visit of γ to a type 3 cube sk upon leaving sk−1 for the last time (γ drawn red). (5) If sk is neither crowded nor empty we say that sk is type 3. As γ leaves sk−1 for the last time and enters sk, there are two possibilities (see [P…

**Figure 5.** Figure 5: The possibilities for case 3a. Top left: sk ∪ s˜k+1 is crowded. Top right: s˜k+1 is empty. Bottom left: s˜k+1 ∪ s˜k+2 has a cross. Bottom right: s˜k+2 is crowded. (7) Now consider the case when sk is type 3b. For simplicity, assume that γ enters sk in the positive time…

**Figure 6.** Figure 6: The possibilities for case 3b. Top left: sk∪s˜k+1 has a cross. Top right: s˜k+1∪s˜k+2 is crowded. Bottom left: s˜k+2 is empty. Bottom right: the remaining case is that γ returns to sk−1, which is excluded by construction. This concludes the definition of γ ′ = (s1, s2,…

**Figure 7.** Figure 7: Two links forming a switch, illustrated in the case when both are double bars. The higher link cannot close a loop. To prove the claim, first recall that for q ∈ T we denote by θq the composition of reflections that maps S0,0 to Sq. We write T ′ for the projection of T…

**Figure 8.** Figure 8: The event DCe,e′ in the case (a), meaning e, e′ are on the boundary of S0,0. The figure shows Λ(k) in the case d = 2, with light orange outlining the boxes of T ′ . The components of D(e, e′ ) are highlighted with turquoise colour. For d = 2 as depicted here, the compo…

**Figure 9.** Figure 9: The event DCe,e′, in the cases (b) and (c), i.e. one or both edges inside S0,0. respectively. Thus each slab contains 2 ⋅ K/4 switches whose lower links are all on non-incident edges. The higher links may be shared by such switches (in the same or different components)…

**Figure 10.** Figure 10: Pairings produced by revealing successive links zj , with minimal pairs indicated by underlining. The initial pairing ξ0 is indicated with dotted lines. Later pairings ξj can be determined by, for any vertex x, starting at (x, t+ j ) and moving downwards, following th…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

[1]

Aizenman and B

M. Aizenman and B. Nachtergaele,Geometric aspects of quantum spin states, Comm. Math. Phys. 164 (1994), 17–63

work page 1994
[2]

Benassi, J

C. Benassi, J. Fröhlich and D. Ueltschi,Decay of Correlations in 2D Quantum Systems with Continuous Symmetry, Ann. Henri Poincaré 18 (2017), 2831–2847

work page 2017
[3]

Improved bounds for connection probabilities in random loop models

V. Betz and A. Klippel and J. Nauth,Improved bounds for connection probabilities in random loop models, arXiv:2503.17857

work page Pith review arXiv
[4]

Biskup,Reflection Positivity and Phase Transitions in Lattice Spin Models, in R

M. Biskup,Reflection Positivity and Phase Transitions in Lattice Spin Models, in R. Kotecky (ed): Methods of Contemporary Mathematical Statistical Physics. Lecture Notes in Mathematics, vol 1970. Springer, Berlin, Heidelberg (2009)

work page 2009
[5]

J. E. Björnberg, P. Mühlbacher, B. Nachtergaele and D. Ueltschi,Dimerization in quantum spin chains with O(n) symmetry, Comm. Math. Phys. (2021)

work page 2021
[6]

Dimerization in $O(n)$-invariant quantum spin chains

J. Björnberg, K. Ryan,Dimerization in O(n)-invariant quantum spin chains, arXiv:2506.06103

work page Pith review arXiv
[7]

Björnberg, D

J. Björnberg, D. Ueltschi,Reflection positivity and infrared bounds for quantum spin systems, in R. Frank, A. Laptev, M. Lewin, R. Seiringer (ed): The Physics and Mathematics of Elliott Lieb, Vol 1. European Mathematical Society (2022)

work page 2022
[8]

Chayes, L

L. Chayes, L. P. Pryadko, K. Shtengel,Intersecting loop models onZd: rigorous results, Nuclear Physics B 570, 590–614 (2000)

work page 2000

Show all 21 references

[9]

Dyson, E

F. Dyson, E. Lieb, B. Simon,Phase Transitions in Quantum Spin Systems with Isotropic and Nonisotropic Interactions, Jour. Stat. Phys. 18 (1978), 335-383

1978
[10]

Fröhlich, C

J. Fröhlich, C. Pfister,On the Absence of Spontaneous Symmetry Breaking and of Crystalline Ordering in Two-Dimensional Systems, Comm. Math. Phys. 81 (1981), 277–298. 20 J. E. BJÖRNBERG AND K. RYAN

1981
[11]

Georgii and T

H.-O. Georgii and T. Küneth.Stochastic comparison of point random fields.Journal of Applied Probability, 34(4): 868-881, 1997

1997
[12]

Kennedy, E

T. Kennedy, E. Lieb, B. S. Shastry,The XY Model Has Long-Range Order for All Spins and All Dimensions Greater than One, Phys. Rev. Lett. 61 (1988), 2582-2584

1988
[13]

Kennedy, E

T. Kennedy, E. Lieb, B. S. Shastry,Existence of Néel Order in Some Spin-l/2 Heisenberg Antiferromagnets, Jour. Stat. Phys. 53 (1988), 1019-1030

1988
[14]

T. Koma, H. Tasaki,Decay of superconducting and magnetic correlations in one- and two-dimensional Hub- bard models, Phys. Rev. Lett. 68 (1992)

1992
[15]

Last and M

G. Last and M. Penrose.Lectures on the Poisson process. Cambridge University Press, 2017

2017
[16]

Lees,Existence of Néel Order in the S=1 Bilinear-Biquadratic Heisenberg Model via Random LoopsComm

B. Lees,Existence of Néel Order in the S=1 Bilinear-Biquadratic Heisenberg Model via Random LoopsComm. Math. Phys. 347, (2016) 83–101

2016
[17]

Lees,Long-range order for the spin-1 Heisenberg model with a small antiferromagnetic interactionJ

B. Lees,Long-range order for the spin-1 Heisenberg model with a small antiferromagnetic interactionJ. Math. Phys. 55, (2014) 093303

2014
[18]

McBryan, T

O. McBryan, T. Spencer,On the decay of correlations in SO(n)-symmetric ferromagnets, Comm. Math. Phys. 53 (1977), 299–302

1977
[19]

Mermin, H

N. Mermin, H. Wagner,Absence of Ferromagnetism or Antiferromagnetism in One- or Two-Dimensional Isotropic Heisenberg Models, Phys. Rev. Lett. 17 (1966)

1966
[20]

Tóth,Improved lower bound on the thermodynamic pressure of the spin1/2 Heisenberg ferromagnet, Lett

B. Tóth,Improved lower bound on the thermodynamic pressure of the spin1/2 Heisenberg ferromagnet, Lett. Math. Phys. 28 (1993), 75–84

1993
[21]

Ueltschi,Random loop representations for quantum spin systems, J

D. Ueltschi,Random loop representations for quantum spin systems, J. Math. Phys. 54 (2013), 083301

2013

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Aizenman and B

M. Aizenman and B. Nachtergaele,Geometric aspects of quantum spin states, Comm. Math. Phys. 164 (1994), 17–63

work page 1994

[2] [2]

Benassi, J

C. Benassi, J. Fröhlich and D. Ueltschi,Decay of Correlations in 2D Quantum Systems with Continuous Symmetry, Ann. Henri Poincaré 18 (2017), 2831–2847

work page 2017

[3] [3]

Improved bounds for connection probabilities in random loop models

V. Betz and A. Klippel and J. Nauth,Improved bounds for connection probabilities in random loop models, arXiv:2503.17857

work page Pith review arXiv

[4] [4]

Biskup,Reflection Positivity and Phase Transitions in Lattice Spin Models, in R

M. Biskup,Reflection Positivity and Phase Transitions in Lattice Spin Models, in R. Kotecky (ed): Methods of Contemporary Mathematical Statistical Physics. Lecture Notes in Mathematics, vol 1970. Springer, Berlin, Heidelberg (2009)

work page 2009

[5] [5]

J. E. Björnberg, P. Mühlbacher, B. Nachtergaele and D. Ueltschi,Dimerization in quantum spin chains with O(n) symmetry, Comm. Math. Phys. (2021)

work page 2021

[6] [6]

Dimerization in $O(n)$-invariant quantum spin chains

J. Björnberg, K. Ryan,Dimerization in O(n)-invariant quantum spin chains, arXiv:2506.06103

work page Pith review arXiv

[7] [7]

Björnberg, D

J. Björnberg, D. Ueltschi,Reflection positivity and infrared bounds for quantum spin systems, in R. Frank, A. Laptev, M. Lewin, R. Seiringer (ed): The Physics and Mathematics of Elliott Lieb, Vol 1. European Mathematical Society (2022)

work page 2022

[8] [8]

Chayes, L

L. Chayes, L. P. Pryadko, K. Shtengel,Intersecting loop models onZd: rigorous results, Nuclear Physics B 570, 590–614 (2000)

work page 2000