Non-equilibrium coupling to a diffusing density breaks Ising universality

Technical Summary: Non-equilibrium coupling to a diffusing density breaks Ising universality

Problem Statement
The Ising universality class is widely regarded as remarkably robust against non-equilibrium perturbations. In standard renormalization group (RG) analyses, terms violating detailed balance typically flow to zero at the fixed point, preserving equilibrium critical exponents. However, the authors investigate whether this robustness holds when an order parameter ($\psi$) is coupled nonreciprocally to a conserved, diffusing density ($\rho$). Such a setup arises in systems like bistable motile particles (e.g., active Ising models) or reaction networks with mobile catalysts, where the density evolves independently of the order parameter, explicitly breaking detailed balance. The central question is whether this specific non-equilibrium coupling alters the large-scale critical behavior or if the system remains in the Ising universality class.

Methodology
The authors introduce the Brownian Ising Model (BIM), a minimal field theory describing the hydrodynamic coupling between a $Z_2$ order parameter and a conserved scalar density. The dynamics are governed by:

1. A Model A-like equation for $\psi$ with a mass term dependent on the local density $\rho$.
2. A purely diffusive equation for the density fluctuation $\delta\rho$.
3. Nonreciprocal coupling: The density affects the order parameter, but the order parameter does not feed back into the density current (unlike Model C).

The analysis employs the dynamical renormalization group using the Martin-Siggia-Rose path-integral formalism. The study focuses on the behavior near the upper critical dimension $d_c = 4$ using an $\epsilon = 4-d$ expansion. Key steps include:

• Deriving the beta functions for the dimensionless couplings: the nonlinearity $u_R$, the density-magnetization coupling $f_R$, and the timescale ratio $w_R$.
• Analyzing the flow in the $(u_R, \tilde{f}_R)$ plane, where $\tilde{f}_R$ is a rescaled coupling that remains finite in both fast-diffusion ($w_R \to 0$) and quenched ($w_R \to \infty$) limits.
• Performing one-loop and two-loop calculations to determine the stability of fixed points and critical exponents.
• Applying an all-orders modified Harris criterion derived from the shift symmetry of the field theory to assess stability beyond perturbation theory.

Key Contributions and Results

1. Discovery of the BIM Fixed Point:
   The RG flow reveals that the standard Ising (Model A) fixed point is unstable to the density coupling. Instead, the system flows to a novel, IR-stable fixed point (the BIM fixed point) characterized by:

   • $X_R^* = 0$ (corresponding to the fast-diffusion limit where density relaxes much faster than the order parameter).
   • Non-zero fixed-point values for the couplings: $u_R^* = \epsilon/6$ and $\tilde{f}_R^* = \epsilon/8$.

2. Novel Universality Class Signatures:
   The BIM fixed point defines a new universality class distinct from Ising in two fundamental ways:

   • Violation of Fluctuation-Dissipation Theorem (FDT): The anomalous dimensions for the correlation function ($\eta$) and the response function ($\eta'$) split.
     • $\eta = -\epsilon/8 + O(\epsilon^2)$ (negative, indicating long-range correlations enhanced by the noise).
     • $\eta' = O(\epsilon^2)$ (positive).
     • The inequality $\eta \neq \eta'$ is a definitive signature of non-equilibrium criticality.
   • Modified Critical Exponents: The correlation length exponent $\nu$ deviates from the Ising value even at one-loop order:
     • $\nu = 1/2 + \epsilon/8 + O(\epsilon^2)$.
     • In $d=3$, the two-loop calculation yields $\nu \approx 0.65$.

3. Mechanism of Instability:
   The instability of the Ising fixed point arises because the fast-diffusion limit ($w_R \to 0$) transforms the density coupling into a long-range multiplicative noise with spatial correlations decaying as $r^{-(d-2)}$. Unlike short-range noise, this long-range noise is RG-relevant near $d_c=4$ and drives the system away from the Ising fixed point.

4. Stability Analysis:

   • Two-loop confirmation: While the one-loop analysis leaves the stability of the BIM fixed point ambiguous (due to a vanishing linear eigenvalue in the $X_R$ direction), a two-loop calculation confirms stability with a positive correction-to-scaling exponent $\omega$.
   • All-orders criterion: Using a generalized Harris criterion, $\nu > 2/(d + z - 2)$, the authors prove that neither the standard Ising nor the diluted-Ising fixed points are stable in $d=3$. The BIM fixed point is the only viable attractor for $d \leq 4$.
   • Dimensional dependence: The analysis predicts that Ising universality is recovered in $d=2$, consistent with existing numerical results.

5. Finite-Size Effects:
   The correction-to-scaling exponent $\omega$ is found to be anomalously small in $d=3$ ($\omega \approx 0.020$). This implies that finite-size corrections decay very slowly ($L^{-\omega}$), meaning numerical simulations will require exceptionally large system sizes or explicit correction analyses to reliably extract asymptotic exponents.

Significance
The paper establishes that the Ising universality class is not robust to nonreciprocal coupling with a conserved diffusive density. This coupling creates a new universality class (BIM) characterized by:

• A genuine non-equilibrium nature manifested by the splitting of correlation and response exponents ($\eta \neq \eta'$).
• A negative anomalous dimension $\eta$, reflecting enhanced spatial correlations due to long-range noise.
• A distinct set of critical exponents ($\nu \approx 0.65$ in $d=3$).

The authors argue that this mechanism is likely relevant for a broad class of living and driven systems, including cell populations, neural assemblies, and active matter, where bistable order parameters interact with fluctuating environments without feedback. The splitting of $\eta$ and $\eta'$ is identified as the primary measurable signature for detecting such non-equilibrium criticality in simulations and experiments.