🔬 condensed matter

Nonequilibrium nucleation theory for nonconserved fields: from active matter to population dynamics

This paper develops a nonequilibrium nucleation theory for systems with nonconserved order parameters by defining a reaction coordinate that accounts for interfacial profile deviations, successfully applying the framework to models in active matter and population dynamics with strong agreement to numerical results.

Original authors: Michalis Chatzittofi, Noah Ziethen, Cesare Nardini, Michael E. Cates

Published 2026-06-18

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Michalis Chatzittofi, Noah Ziethen, Cesare Nardini, Michael E. Cates

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are watching a calm lake. Suddenly, a tiny ripple forms. In the world of physics, this is like a "metastable" state—a situation that looks stable but is actually waiting to snap into something else. Usually, for a new phase (like ice forming in water, or a new species taking over an ecosystem) to appear, a "critical droplet" must form. If this droplet is too small, it shrinks and disappears. If it's big enough, it grows uncontrollably, taking over the whole system.

This process is called nucleation.

For decades, scientists have had a rulebook for how this happens in calm, balanced systems (equilibrium). It's like a game where the rules are fair and predictable: you calculate the energy cost of the droplet's surface versus the energy gain of its inside, and you can predict exactly how likely it is to happen.

The Problem: The Rules Change When Things Get "Active"
The real world, however, is often messy and "active." Think of a school of fish swimming together, bacteria moving on their own, or a population of animals fighting for resources. In these systems, the usual rules of balance (called "detailed balance") are broken. The old rulebook doesn't work anymore.

Previously, scientists could only write a new rulebook for systems where the total amount of "stuff" (like the number of particles) stays the same. But many important systems—like populations growing or shrinking, or active fluids where particles consume energy—don't keep a constant count. They are "non-conserved." Until now, no one had a reliable way to predict nucleation in these chaotic, non-conserved systems.

The Solution: A New Way to Measure the "Hill"
This paper introduces a new theory (Nonequilibrium Nucleation Theory, or NNT) specifically for these messy, non-conserved systems.

Here is the core idea using an analogy:

Imagine trying to push a heavy boulder over a hill to get to a valley on the other side.

The Old Way (Equilibrium): You assume the boulder rolls down a smooth, symmetrical path. The path up is the exact mirror image of the path down. You just measure the height of the hill, and you know the answer.
The New Reality (Active/Non-conserved): The hill is slippery, and the ground is shifting. The path the boulder takes when it's being pushed up (by random noise) is not the same as the path it takes when it rolls down.
- If you try to use the "down" path to guess the "up" path, you get the wrong answer. You might think the hill is twice as high as it really is, meaning you'd predict the event is impossible when it's actually quite likely.

The Breakthrough: Defining the "Reaction Coordinate"
The authors solved this by inventing a very specific way to measure the size of the droplet (the "reaction coordinate").

Think of the droplet not just as a ball of stuff, but as a ball with a fuzzy, wiggly edge. In active systems, the edge of a growing droplet looks different than the edge of a shrinking one.

The authors realized that if you define the "size" of the droplet in a very clever, mathematical way, you can ignore the messy wiggles of the edge.
By doing this, they could project the complex, chaotic movement of the whole system down to a simple, one-dimensional story: "How big is the droplet?"
This allowed them to calculate the true height of the energy barrier (the hill) without getting confused by the different shapes the droplet takes while growing versus shrinking.

What They Tested It On
They didn't just do the math; they tested it on two real-world scenarios:

Population Dynamics: Imagine a species of animals in a forest. Sometimes, a small group of a new, stronger species tries to take over. The paper showed how to calculate the odds of this new species successfully taking over the forest, even when the population numbers are fluctuating wildly and the "noise" (random births/deaths) is part of the system.
Active Matter (Active Model A): This models things like self-propelled particles (e.g., bacteria or synthetic robots) that move on their own. They found that in these systems, the "hill" the droplet has to climb is significantly different from what old theories predicted.

The Big Takeaway
The paper proves that in active, non-conserved systems, you cannot simply assume that the path of growth is the reverse of the path of decay. If you do, you will be wrong.

By carefully defining how we measure the size of a "nucleus" (the seed of the new phase), they created a new, accurate theory that matches computer simulations perfectly. This means scientists can now better predict rare, dramatic events in everything from how bacteria colonies form to how new species invade ecosystems, without relying on outdated, equilibrium-based assumptions.

Technical Summary: Nonequilibrium Nucleation Theory for Nonconserved Fields

Problem Statement
Classical Nucleation Theory (CNT) successfully describes the formation of a stable phase from a metastable one in equilibrium systems by quantifying the free-energy competition between bulk gain and interfacial cost. In equilibrium, the nucleation barrier is determined by the saddle point of the free energy, and the optimal trajectory (instanton) is the time-reversal of the deterministic relaxation path. However, for systems without detailed balance (nonequilibrium), these simplifications fail: free-energy arguments cannot estimate rare event probabilities, and the instanton trajectory is not known a priori.

While a nonequilibrium nucleation theory (NNT) has been developed for systems with a conserved order parameter (e.g., Active Model B+), a rigorous extension to systems with a nonconserved order parameter has remained elusive. Such nonconserved dynamics govern many reaction-diffusion systems, including population dynamics, biomolecular condensates, and active matter (e.g., Active Model A). The central challenge is that in nonequilibrium systems, the interfacial profile of a noise-driven growing droplet deviates from the profile observed during deterministic relaxation, potentially altering the nucleation barrier in ways not captured by previous phenomenological applications of CNT.

Methodology
The authors construct a rigorous NNT for field theories with a single, scalar, nonconserved order parameter $\phi(x,t)$ governed by the stochastic equation:
$\partial_t \phi = -\mu[\phi] + \sqrt{2T}D(\phi)\xi$
where $\mu$ is the deterministic drift (not necessarily a functional derivative of a free energy) and $\xi$ is Gaussian white noise.

The core methodological innovation is the systematic expansion of the stochastic dynamics of the droplet radius $R(t)$ into a 1D Langevin equation:
$\dot{R} = M(R)[- \partial_R U(R)] + \sqrt{2T M(R)}\zeta$
where $U(R)$ is the quasipotential, $M(R)$ is the emergent mobility, and $\zeta$ is noise.

Key steps in the derivation include:

Definition of the Reaction Coordinate: The authors define the droplet radius $R(t)$ not arbitrarily, but as the unique solution to an integral condition involving an auxiliary function $\psi'(r-R)$ . This specific choice is designed to project out deviations ( $\epsilon$ ) in the interfacial profile from the Langevin equation for $R$ .
Handling Interfacial Deviations: Unlike equilibrium CNT or conserved NNT, the authors demonstrate that the interfacial profile of a growing droplet (instanton trajectory) differs from the relaxation profile. They show that these deviations, denoted $\epsilon_A$ , are small ( $O(v_0, 1/R, \sqrt{T})$ ) but crucially affect the energy barrier if not handled correctly.
Selection of $\psi'$ : By choosing $\psi'$ such that it lies in the kernel of the adjoint linearized operator ( $L^\dagger$ ), the authors eliminate the dependence on the unknown field deviations $\epsilon$ in the drift term, allowing for an analytical derivation of $M(R)$ and $U(R)$ .
Validation: The analytical results are validated by numerically minimizing the Freidlin–Wentzell action of the original field theory using a geometric minimum action method (GMAM) combined with a Ritz method.

Key Contributions and Results

Analytical Formulas for NNT: The paper derives explicit expressions for the critical radius $R_c$ $R_{c}$ , the mobility $M(R)$ $M (R)$ , and the quasipotential barrier $U(R)$ $U (R)$ for nonconserved nonequilibrium systems.
- The critical radius is given by $R_c = \frac{(d-1)\sigma}{\int \psi' \mu_0[\phi_0]}$ .
- The quasipotential barrier $U(R_c)$ depends on the noise structure $D(\phi)$ and the specific choice of $\psi'$ , meaning it cannot be inferred solely from the noiseless dynamics or the interfacial tension $\sigma$ alone.
Failure of Time-Reversed Relaxation (TRR) Ansatz: The authors demonstrate that for generic nonequilibrium systems, assuming the instanton follows a time-reversed relaxation trajectory (TRR) leads to incorrect results. Specifically, the TRR ansatz overestimates the nucleation barrier $U(R_c)$ by an $O(1)$ factor (e.g., by more than a factor of 2 in Active Model A), leading to predictions of nucleation times that are infinitely too long in the low-noise limit.
Application to Specific Models:
- Population Dynamics: Applied to Fisher-type equations with strong Allee effects. The results show that nucleation barriers depend on the specific form of the multiplicative noise $D(\phi)$ , distinguishing them from equilibrium predictions.
- Active Model A (AMA): Applied to a minimal field theory for nonequilibrium phase ordering. The theory correctly predicts the dependence of the barrier on the activity parameter $\lambda$ and external field $h$ . Numerical minimization of the action confirms the analytical predictions for both $R_c$ and $U(R_c)$ , while the TRR ansatz fails significantly.
Interfacial Profile Distortion: The paper quantifies the deviation between the instanton profile and the relaxation profile, showing it scales as $\epsilon_A \sim O(v_0)$ , consistent with the analytical assumptions.

Significance
The paper claims to provide the first well-founded extension of nucleation theory to nonconserved nonequilibrium field theories. Its primary significance lies in correcting the misconception that the nucleation barrier in nonequilibrium systems can be approximated by time-reversing the relaxation path. The authors show that deviations in the interfacial profile are a leading-order effect in the calculation of the barrier for nonconserved dynamics.

By carefully defining the reaction coordinate to project out these deviations, the authors recover a closed Langevin equation that yields the correct quasipotential. This framework allows for the analytical prediction of nucleation rates in systems ranging from active matter to ecological population dynamics, where detailed balance is broken. The work bridges the gap between phenomenological applications of CNT in active matter and rigorous statistical mechanics, providing a benchmark for numerical rare-event algorithms and suggesting that similar approaches may be applicable to biomolecular condensates and multi-order-parameter systems.

Technical Summary: Nonequilibrium Nucleation Theory for Nonconserved Fields

More like this