2D O(n) Model: Criticality & Integrability

• The two-dimensional O(n) model is a central theoretical framework in statistical mechanics characterized by global O(n) symmetry and a rich spectrum of critical phenomena.
• It exhibits distinct critical and massive phases, with conformal field theories describing its critical behavior and a mass gap emerging in the non-critical regime.
• Multiple formulations—including nonlinear sigma models, lattice loop-gas, and supersymmetric extensions—provide rigorous insights into integrability and nonperturbative effects.

The two-dimensional $O(n)$ model is a central paradigmatic model in statistical mechanics, quantum field theory, and mathematical physics. It encompasses a broad class of systems characterized by global $O(n)$ symmetry in two spatial dimensions, and admits a rich variety of critical phenomena, exact solutions, and deep connections to conformal field theory (CFT), integrable systems, and quantum statistical systems. The model exists in multiple formalizations—including nonlinear sigma models, loop-gas formulations, lattice spin models, and their supersymmetric deformations—and serves as a precise laboratory for the study of universality, mass generation, critical exponents, integrability, and non-invertible symmetries.

1. Definitions, Formulations, and Symmetry

The two-dimensional $O(n)$ model admits several inequivalent but deeply related definitions:

• Nonlinear Sigma Model (NLSM): On a Riemann surface $\Sigma$, the field is $\mathbf{n}(x)\in S^{n-1}\subset \mathbb{R}^n$ with $|\mathbf{n}(x)|^2=1$; the action reads

$$S[\mathbf{n}] = \frac{1}{2g^2} \int_\Sigma d^2x\, \partial_\mu \mathbf{n} \cdot \partial^\mu \mathbf{n}.$$

The target manifold $S^{n-1}$ is a symmetric space $O(n)/O(n-1)$; the model has global $O(n)$ invariance $\mathbf{n}(x)\to R \mathbf{n}(x)$ for $R\in O(n)$ (Krichever et al., 2021).

• Linear Sigma Model and Large-$N$ Formulation: The “Linear Sigma Model” relaxes the constraint $|\Phi(x)|^2=N\beta$ with a quartic potential for $\Phi\in\mathbb{R}^N$; under 't Hooft scaling ($\beta\mapsto N\beta$, $\lambda\mapsto \lambda/N$), the action is

$$S_N(\Phi) = \int_{\mathbb{R}^2}\Bigl\{\frac12\|\nabla\Phi(x)\|^2 + \frac{\lambda}{4N}(\|\Phi(x)\|^2-N\beta)^2\Bigr\}dx.$$

(Delgadino et al., 27 Jan 2026)

• Lattice Loop-Gas and Spin Model: On a trivalent (e.g., honeycomb or square) lattice, $O(n)$-invariant lattice models admit a high-temperature expansion as a loop-gas:

$$Z_{\text{loops}} = \sum_{\mathcal{C}} K^{|\mathrm{edges}(\mathcal{C})|} n^{\# \text{loops}(\mathcal{C})},$$

where $n$ is now analytically continued to $\mathbb{R}$ or $\mathbb{C}$. The $O(n)$ spin model Hamiltonian is $$\beta H = -\sum_{\langle ij\rangle}\log(1 + K\,\mathbf{S}_i\cdot\mathbf{S}_j )$$ (Gorbenko et al., 2020, Jacobsen et al., 2023, Grans-Samuelsson et al., 2021).

• Supersymmetric and Heterotic Extensions: Deformations include supersymmetric $O(n)$ sigma models in $\mathcal{N}=(1,1)$ or heterotic $\mathcal{N}=(0,1)$, built by coupling left- or right-moving superpartners or Majorana-Weyl fermions to the bosonic sector (Koroteev et al., 2010, Peterson et al., 2015).

The continuum limit at criticality exhibits rich conformal structures with different universality classes depending on the range of $n$.

2. Phase Structure, Criticality, and Conformal Field Theory

The model exhibits distinct critical and massive phases determined by the value of $n$ and, for lattice models, the loop tension (fugacity):

• Critical Points and Phases ($|n| \le 2$):
  • There exist two branches of critical points—dilute ($g=\beta^2\in[1,2]$) and dense ($g\in(0,1)$)—each associated with a CFT of central charge $c=1-6(g-1)^2/g$, $n=-2\cos(\pi g)$ (Jacobsen et al., 2023, Grans-Samuelsson et al., 2021, Gorbenko et al., 2020).
  • CFTs describe both lattice and continuum (field-theoretic) versions. Critical exponents and operator dimensions follow Coulomb gas predictions:

  $$\Delta_{(r,s)} = \left(\frac{\beta r - \beta^{-1}s}{2}\right)^2 - \left( \frac{\beta - \beta^{-1}}{2} \right)^2,$$

  and $c = 1-6(g-1)^2/g$ (Grans-Samuelsson et al., 2021). - At $n=2$, the dilute and dense branches meet in a Kosterlitz-Thouless (KT) transition; $O(2)$ is the compact boson with $c=1$ (Gorbenko et al., 2020).

• Massive Phase ($n > 2$):

  • For $n>2$, real fixed points annihilate and move into the complex plane. The theory is asymptotically free for $N>2$ (in sigma-model language): the $\beta$-function is $\beta(g) = -b_1 g^2 - b_2 g^3 + \dots$ with $b_1=(N-2)/(2\pi)$, so no real critical point exists at positive coupling (Yang et al., 5 Jan 2026).
  • The spectrum is massive, with exponential decay of correlations, dynamically generated mass gap, and no spontaneous symmetry breaking by the Mermin–Wagner theorem (Yang et al., 5 Jan 2026, Haldar et al., 2023).
• Complex CFTs and “Walking” RG ($n > 2$):
  • Upon complexification, nontrivial fixed points appear as complex conjugate pairs (“CCFTs”) with complex central charge and scaling dimensions, confirmed both by analytic continuation and numerical transfer-matrix studies up to $n_g\approx12.34$ (Yang et al., 5 Jan 2026, Haldar et al., 2023).

A table summarizing phase structure in the two-dimensional $O(n)$ model:

$n$	Critical CFT?	Central Charge $c$	Phase
$|n|\le2$	Dilute & Dense	$1-6(g-1)^2/g$	Critical or quasicritical
$n=2$	KT critical point	$1$	Massless (compact boson)
$n>2$	No real fixed point	Complex-conjugate $c_\pm$	Massive, "walking" RG

3. Mass Generation, Large-$N$ Analysis, and Nonperturbative Phenomena

The $O(n)$ model provides a classic setting for nonperturbative phenomena in two dimensions:

• Dynamical Mass Generation:
  • In the continuum linear sigma model at large $N$, the self-consistent gap equation determines a mass $m$ dynamically generated by quantum fluctuations:

  $$\frac{m_*^2}{\lambda}+\frac{1}{2\pi}\ln m_* = -\beta,$$

  leading, at low temperatures ($\beta\gg1$), to

  $m_*(\beta) \sim \exp(-2\pi\beta),$

  with exponential decay of correlations at large distances (Delgadino et al., 27 Jan 2026). - The two-point function takes the leading-order form of a massive Gaussian field,

  $$\langle \Phi_i(x)\Phi_j(0)\rangle \approx \delta_{ij} ((-\Delta+m^2)^{-1})(x,0) \sim \delta_{ij}\frac{e^{-m|x|}}{\sqrt{|x|}}$$

  for $|x|\gg1$ (Delgadino et al., 27 Jan 2026).

• Quantified Gaussianity and Marginals:

  • For each individual component, the law converges, in Wasserstein distance, to the massive Gaussian Free Field (GFF), with the deviation bounded as $O(1/\sqrt{N})$ (for the 2-Wasserstein distance in a weighted $H^1$ norm) (Delgadino et al., 27 Jan 2026).
• Lattice Mass Gap and Strong-Coupling Methods:
  • On the lattice, strong-coupling expansions and Padé-Borel resummations extract the non-perturbative mass gap and compare it to exact and RG results. The continuum limit is governed by known asymptotic scaling, and non-perturbative constants agree with Bethe-ansatz predictions (Yamada, 2011).

4. Logarithmic and Nonunitary Conformal Structures

The two-dimensional $O(n)$ model for generic (non-integer) $n\ne 2$ explores territory beyond unitary rational CFT:

• Logarithmic CFTs (LCFT):
  • For $n<2$, the CFT describing both critical and low-temperature phases are nonunitary and logarithmic: correlators exhibit logarithmic terms, negative-norm (“ghost”) states are generic, and conserved $O(n)$ currents themselves reside in staggered logarithmic multiplets (Gorbenko et al., 2020).
  • Negative-norm states decouple in the $n\to2$ limit, restoring unitarity as in the compact boson case (Gorbenko et al., 2020).
  • Physical observables include Jordan-cell dilatation actions and logarithmic conformal blocks in both two- and four-point functions.
• Dangerously Irrelevant/Relevant Operators:
  • The four-leg (watermelon) singlet operator $S$ is irrelevant at the UV critical point but becomes relevant in the IR dense phase; RG flow from dilute to dense is governed by the only relevant singlet $\varepsilon=\mathcal{O}_{e_0-2,0}$, with $S$ never generated in the flow, protected by selection rules and, more fundamentally, by non-invertible topological defect symmetries (Gorbenko et al., 2020, Jacobsen et al., 2023).
• Non-invertible Topological Defects:
  • Recent work has established that RG protection from dilute to dense $O(n)$ is enforced by non-invertible topological defect lines (TDLs), leading to stringent fusion rules and exclusion of dangerous operators from the RG flow. These TDLs manifest as algebraic topological constraints in both lattice and continuum theories (Jacobsen et al., 2023).

5. Symmetries, State Spaces, and Algebraic Structures

The $O(n)$ model in two dimensions realizes symmetry in nontrivial algebraic and representation-theoretic forms:

• Spectrum Decomposition:
  • The primary fields of the $O(n)$ CFT are labeled by $(r,s)$, belong to representations $\Lambda_{(r,s)}$ of $O(n)$, and the torus partition function decomposes accordingly. For generic $n$, explicit formulae for multiplicities and representation content are available based on Chebyshev polynomials and combinatorial diagrammatic algebra (Grans-Samuelsson et al., 2021, Jacobsen et al., 2022).
  • The fusion rules and conformal blocks exhibit full compatibility with $O(n)$ invariants, saturating known bounds for a large set of correlators (Grans-Samuelsson et al., 2021).
• Diagram Algebra and Brauer/Jones–Temperley–Lieb Structures:
  • The space of states and operator content is constructed via the representation theory of the Brauer algebra and its restriction to Jones–Temperley–Lieb (JTL) subalgebras, providing a precise framework for $O(n)$ decomposition in the scaling limit (Jacobsen et al., 2022).
  • Twisted torus partition functions, with $O(n)$ group elements inserted along noncontractible cycles, resolve ambiguities and enable unique expansions in terms of irreducible representations.
• Supersymmetric Generalizations:
  • $\mathcal{N}=(1,1)$ and $\mathcal{N}=(0,1)$ deformations possess vacuum structures controlled by the Witten index and its modifications, with massless Goldstino modes and explicit control of supersymmetry breaking in response to connectivity number and heterotic couplings (Koroteev et al., 2010, Peterson et al., 2015).

6. Integrability, Hierarchies, and Geometric Aspects

The $O(n)$ sigma model is one of the archetypal integrable field theories in two dimensions:

• Novikov–Veselov Hierarchy and Spectral Curves:
  • The equations of motion are part of a larger family of commuting flows (Novikov–Veselov hierarchy), and all periodic (double-periodic) solutions are algebraic-geometric: their spectral data are finite-genus curves (Fermi curves), and the integrable structure connects harmonic maps into spheres with spectral curves of the elliptic Calogero–Moser system (Krichever et al., 2021).
  • The zero-curvature representation, Baker–Akhiezer function construction, and connections to quantum integrable systems enable the explicit parameterization of the solution space in terms of algebraic geometry and spectral theory.

7. Modern Extensions and Experimental Realizations

Recent research has led to novel understandings and prospective applications:

• Complex CFTs and Non-Hermitian Realizations:
  • For $N>2$, continuation into the complex coupling plane reveals generic complex conformal fixed points that describe non-unitary, dissipative, or monitored quantum systems—demonstrating their universality in open quantum systems and in the description of “walking” RG and weakly first-order transitions (Yang et al., 5 Jan 2026, Haldar et al., 2023).
  • Non-Hermitian spin-1 Heisenberg chains provide microscopic realizations of these CCFTs, both at the level of spectrum and entanglement scaling (Yang et al., 5 Jan 2026).
• Dissipative Preparation and Lindbladian Dynamics:
  • It is possible to engineer Lindbladian or monitored dynamics such that the system conditionally relaxes to the ground state of a given CCFT, dynamically producing universal long-range entanglement with complex central charge (Yang et al., 5 Jan 2026).
• Universality, Phase Diagrams, and Experimental Probes:
  • The $O(n)$ model serves as an archetype for universality classes in statistical mechanics, with robust confirmation of universality for special transitions, effects of non-invertible superselection sectors, and the analytic continuity of phase diagrams across a wide range of real and complex $n$ (Fu et al., 2016, Jacobsen et al., 2023).

The two-dimensional $O(n)$ model thus continues to serve as a testing ground for theoretical innovation and rigorous methodology in field theory, statistical mechanics, representation theory, and quantum many-body physics.