Wilson-Fisher Fixed Point

Updated 29 January 2026

Wilson-Fisher fixed point is a nontrivial infrared stable fixed point in d<4 scalar field theories that governs universal scaling behavior in models like Ising and O(N).
It arises via an epsilon-expansion from the Gaussian fixed point, using perturbative RG techniques to compute critical exponents and anomalous dimensions.
The fixed point is studied through various approaches including analytic bootstrap, functional RG, and lattice simulations, with applications to large-charge effective theories and fermionic extensions.

The Wilson-Fisher fixed point is a paradigmatic nontrivial infrared stable fixed point in the renormalization group (RG) flow of scalar quantum field theories in spatial dimension $d < 4$ , notably controlling the universal scaling behavior of the Ising, XY, and Heisenberg universality classes, as well as their O(N) and more general extensions. It arises by analytic continuation from the Gaussian fixed point in four dimensions via @@@@2@@@@ in $\epsilon = 4-d$ and underlies a wide array of critical phenomena and conformal field theories (CFTs), both in continuum field theory and in lattice models.

1. Definition and General Structure

The Wilson-Fisher fixed point is defined as a nontrivial zero of the beta function in scalar field theory as the space dimension is analytically continued below the upper critical value. In the canonical case, the theory is a real scalar field $\phi$ with quartic self-interaction,

$S = \int d^d x \left[ \frac{1}{2} (\partial_\mu \phi)^2 + \frac{1}{2} m^2 \phi^2 + \frac{\lambda}{4!} \phi^4 \right].$

In minimal subtraction schemes within dimensional regularization, the one-loop beta function for the dimensionless coupling $g$ (rescaled $\lambda$ ) is

$\beta_g = -\epsilon g + 3 g^2 + O(g^3).$

The Wilson-Fisher fixed point occurs at

$g_* = \frac{\epsilon}{3} + O(\epsilon^2),$

for $d = 4 - \epsilon$ , with $\epsilon > 0$ (Herbut, 2023). This nontrivial RG fixed point governs the long-wavelength physics at criticality in three dimensions and below.

2. Critical Exponents and Operator Content

At this fixed point, the theory becomes conformal, and scaling exponents can be systematically computed in the $\epsilon$ -expansion. The critical exponents are encoded in anomalous dimensions of fundamental and composite operators:

The anomalous dimension $\eta_\phi$ of the elementary field appears at two loops, $\eta_\phi = 0 + O(\epsilon^2)$ for pure $\phi^4$ theory (Roumpedakis, 2016).
The inverse correlation length exponent is

$\nu^{-1} = 2 - \beta_g'(g_*) = 2 - \left(-\epsilon + 6g_*\right) = 2 - \epsilon/3 + O(\epsilon^2),$

which gives

$\nu = \frac{1}{2} + \frac{\epsilon}{12} + O(\epsilon^2).$

For O(N) extensions, the one-loop expressions generalize to

$g_* = \frac{6 \epsilon}{N+8}, \quad \nu = \frac{1}{2} + \frac{N+2}{4(N+8)}\epsilon + \cdots,$

with the anomalous dimension at order $\eta_\phi = \frac{N+2}{2(N+8)^2} \epsilon^2$ (Herbut, 2023).

The spectrum of local operators at the fixed point is determined by the interplay of RG scaling and conformal invariance (Roumpedakis, 2016, Liendo, 2017, Khanikar et al., 11 Jan 2026). These dimensions can be extracted by matching OPEs, crossing symmetry, and via algebraic analysis of the dilatation operator, which at order $\epsilon$ is completely fixed by symmetry once normalization is set (Liendo, 2017).

3. Extensions: O(N) Models and Large-Charge Sectors

The Wilson-Fisher fixed point extends to multicomponent fields with O(N) symmetry, yielding families of critical points as a function of N. In three dimensions, the critical O(N) model for integer N exhibits a nontrivial CFT structure. Associated with these models is a large-charge effective field theory (EFT) that enables the computation of operator scaling dimensions in sectors with fixed large global charge $Q$ (Banerjee et al., 2019, Banerjee et al., 2021, Singh, 2022).

In this framework, the ground state energy (thus scaling dimension) for large $Q$ admits a systematic $1/Q$ expansion: $\Delta(Q) = c_{-1} Q^{3/2} + c_0 Q^{1/2} + c_1 + c_2 Q^{-1/2} + \ldots$ where $c_{-1}, c_0, \ldots$ are low-energy constants that encode universal CFT data. Monte Carlo simulations on specially regularized lattice models (e.g., qubit worldline representations) yield high-precision results for these coefficients. For O(4), the leading coefficients are found as $c_{3/2} = 1.068(4), c_{1/2} = 0.083(3)$ , and in the subleading sector, coefficient differences such as $\lambda_{1/2} = 2.08(5), \lambda_1 = 2.2(3)$ appear in fractional inverse powers of the charge (Banerjee et al., 2019, Banerjee et al., 2021). These confirm universal predictions of the large-charge EFT paradigm.

4. Analytical, Numerical, and Functional Techniques

A variety of methods elucidate the Wilson-Fisher fixed point structure:

Perturbative RG and $\epsilon$ -expansion: Systematic expansion in $\epsilon = 4-d$ for couplings, scaling dimensions, and correlation functions (Herbut, 2023, Roumpedakis, 2016).
Conformal bootstrap and analytic functionals: Exploitation of crossing symmetry and analyticity of conformal blocks to extract OPE data and scaling dimensions, even reconstructing all order- $\epsilon$ anomalous dimensions by symmetry (Liendo, 2017, Khanikar et al., 11 Jan 2026, Gliozzi et al., 2017).
Functional and nonperturbative RG: Solution of functional flow equations (e.g., Wetterich equation) in the Local Potential Approximation or beyond. Local and global fixed-point solutions are constructed via expansions about small, large, purely imaginary fields, or matched numerically (Jüttner et al., 2017, Litim et al., 2016, Jüttner et al., 2017).
Lattice simulations: Worldline and worm algorithm Monte Carlo techniques, including qubit regularization for efficient sampling at large charge, yielding quantitative agreement with EFT predictions (Singh, 2022, Banerjee et al., 2021).
Gradient flow: RG flows constructed via the gradient flow of lattice fields allow for direct nonperturbative observation of the fixed point and estimation of its universal features (Morikawa et al., 2024).

5. Universality, Non-universality, and Generalizations

While the critical exponents and certain amplitude ratios are universal within a given universality class, the fixed-point coordinates themselves (e.g., fixed-point couplings) and various nonuniversal amplitudes do depend on regularization details. For example, the position of the Wilson-Fisher fixed point in coupling space shifts depending on the scheme and the form of the cutoff in Wilsonian RG (Goldstein, 2024). At lowest order, critical exponents are universal, but higher-order corrections introduce scheme-dependence.

Generalizations include nonlocal field theories in arbitrary dimension, where one can dial the interaction to remain marginal and access Wilson-Fisher–like criticality in all d, with critical exponents varying smoothly across dimensions but unitarity violations emerging for $d > 4$ (Trinchero, 2019).

In noncommutative field theory, analogous fixed points exist and interpolate between the standard commutative Wilson-Fisher fixed point and novel noncommutative, strongly coupled regimes with dramatically altered universality and criticality structure (Ydri et al., 2012, Ydri et al., 2015).

6. Wilson-Fisher Fixed Point in the Presence of Fermions

In extensions such as Gross-Neveu-Yukawa (GNY) models, the order parameter couples to Dirac fermions, leading to the chiral Ising or chiral Heisenberg universality classes. The RG flow now involves coupled beta functions for the bosonic quartic and Yukawa couplings: $\begin{aligned} \beta_u &= -\epsilon u + \frac{1}{2}(N+8) u^2 - 2N_f uy + 4N_f y^2 + \cdots, \ \beta_y &= -\epsilon y + \frac{N_f + 6}{2} y^2 + \cdots, \end{aligned}$ with fixed points at $(u_*, y_*) = O(\epsilon)$ (Herbut, 2023). The anomalous dimensions are enhanced, e.g., $\eta_\phi \sim O(\epsilon)$ , $\eta_\psi \sim O(\epsilon)$ , and the universality class is modified. These GNY fixed points are central to the theory of the Néel-ordered Mott transition in models like the half-filled honeycomb lattice Hubbard model, where agreement is sought between high-order $\epsilon$ -expansion, large-N, and Monte Carlo studies (Herbut, 2023).

7. Subtleties and Open Issues

The Wilson-Fisher construction is exact only for non-integer d at fixed order in $\epsilon$ ; for physical integral dimensions, subtle features emerge:

In $d = 2$ , the Wilson-Fisher fixed point reduces to containing an Ising subsector but does not coincide with the full 2d Ising CFT owing to the emergence of Virasoro symmetry and the decoupling/cancellation of nonunitary or negative-norm sectors (Zan, 27 Jan 2026).
In noninteger d, the fixed point entails an infinite number of "evanescent" operators (vanishing at integer d) and negative-norm states, leading to genuinely nonunitary behavior, e.g., complex anomalous dimensions already at leading order for high-dimension operators (Hogervorst et al., 2015).
The choice of cutoff or regularization in functional RG flows (e.g., real-space block, momentum-shell, or smooth mass-like cutoff) affects finite-order coordinates of the fixed point, as well as emergent symmetry in multicomponent models (Goldstein, 2024).
In large-charge expansions, the precise matching of Wilson coefficients with analytic EFT calculations remains an ongoing challenge, as does the extension to higher dimensions, other symmetry classes, or inclusion of fermions or disorder (Banerjee et al., 2021).