Mass-Conserving Reaction-Diffusion Systems

Updated 3 July 2026

McRD systems are spatial models where chemical species interconvert and diffuse while total mass remains constant, enabling precise pattern formation analysis.
They rely on concepts like reactive nullclines, flux-balance subspaces, and mass-redistribution potentials to trigger phase-space instabilities and coarsening phenomena.
Applications include cell polarity, biomolecular phase separation, and synthetic design, supported by rigorous analytical and numerical approaches that preserve mass conservation.

A mass-conserving reaction–diffusion (McRD) system is a class of spatially extended dynamical models in which chemical or molecular species interconvert and diffuse in space while the total mass (i.e., the spatial integral of all relevant species densities) is strictly conserved. Such systems underlie the formation, evolution, and selection of spatial patterns in a wide array of nonequilibrium physical, chemical, and biological contexts—most notably protein-based intracellular organization and cellular polarity, biomolecular condensation, and non-equilibrium phase separation. The mathematical and physical structure of McRD dynamics has been the focus of extensive analytical, numerical, and experimental research, enabling precise understanding of mesoscale pattern phenomena far beyond the regimes accessible to classical Turing instabilities.

1. Mathematical Structure and Mass Conservation

A general McRD system for $n$ interacting species $u_i(\mathbf{x},t)$ on a domain $\Omega$ is formulated as: $\partial_t u_i = D_i \nabla^2 u_i + f_i(\mathbf{u}), \quad i=1,\dots,n,$ subject to the mass conservation constraint: $\frac{d}{dt} \int_\Omega \sum_{i=1}^n u_i(\mathbf{x},t)\,d\mathbf{x} = 0.$ Mass conservation is ensured provided the reaction terms satisfy $\sum_i f_i(\mathbf{u}) = 0$ for all $\mathbf{u}$ . In prototypical two-component McRD models, such as those describing the interconversion between membrane-bound and cytosolic protein states, the species exchange mass locally while each diffuses at distinct rates, creating a global invariant for $\rho(\mathbf{x},t) = m(\mathbf{x},t) + c(\mathbf{x},t)$ , where $m$ is membrane-bound and $c$ is cytosolic (Frey et al., 14 Dec 2025, Brauns et al., 2018, Latos et al., 2020, Latos et al., 2015).

The concept of the mass-redistribution potential $u_i(\mathbf{x},t)$ 0 (also called the quasi-chemical or flux-balance potential) encodes the effect of differences in diffusivities and is defined, for instance, by $u_i(\mathbf{x},t)$ 1. The evolution of the total density obeys a continuity equation of the form $u_i(\mathbf{x},t)$ 2, reflecting the fact that spatial redistribution of mass is mediated by $u_i(\mathbf{x},t)$ 3 (Frey et al., 14 Dec 2025, Brauns et al., 2018).

2. Phase-Space Geometry and Pattern-Forming Instabilities

McRD systems exhibit rich pattern-forming behavior, governed by the geometry of the reactive nullcline (the set of points where net interconversion vanishes) and the location of the system in phase space relative to the mass conservation constraint.

Key constructs include:

Reactive nullcline $u_i(\mathbf{x},t)$ 4: Characterizes local reaction equilibria at fixed total mass.
Flux-balance subspace (FBS): The linear constraint $u_i(\mathbf{x},t)$ 5, which all stationary interfaces must respect (Brauns et al., 2018, Wigbers et al., 2019).
Local equilibria theory: The intersection of the nullcline with lines of constant total density determines the possible plateau states (high and low concentration regions) and interface structures.

The mass-redistribution instability (also called the McRD Turing instability) occurs when the slope of the nullcline is steeper than $u_i(\mathbf{x},t)$ 6, or equivalently when $u_i(\mathbf{x},t)$ 7 at the homogeneous state. This drives the amplification of small fluctuations via mass transport and leads to symmetry-breaking pattern formation (Frey et al., 14 Dec 2025, Brauns et al., 2020, Brauns et al., 2018).

Spatial heterogeneities or imposed templates can further localize patterns via regional nullcline geometry, as in edge-sensing mechanisms (Wigbers et al., 2019).

3. Nonlinear Dynamics: Sharp Interfaces and Coarsening

McRD systems generically manifest sharp internal layers (“transition layers” or “mesas”) separating domains of distinct local equilibria. Matched asymptotic analysis reveals that these interfaces have width $u_i(\mathbf{x},t)$ 8 (or a small singular parameter), with their location and stability fixed by global mass conservation and interface-turnover balance laws (Kuwamura et al., 2023, Ikeda et al., 2024, He et al., 6 Feb 2026). The Maxwell integral condition or its generalizations (involving the net area under the reaction term) selects the globally permitted configuration for a prescribed total mass.

A central feature is coarsening—the tendency for high-density domains to merge at the expense of smaller domains—driven by self-amplifying competition for conserved mass. In strict mass conservation, coarsening is generically uninterrupted in two-component systems, leading eventually to single-domain, phase-separated states (macrophase separation) (Brauns et al., 2020, Tateno et al., 2020, Frey et al., 14 Dec 2025). In higher dimensions, interface curvature induces effective “surface tension” controlling coarsening rates, with late-stage dynamics following the Lifshitz–Slyozov–Wagner (LSW) scaling law $u_i(\mathbf{x},t)$ 9 for droplet radius (Tateno et al., 2020).

In the presence of weak source-sink terms (weakly broken conservation), coarsening can be arrested, leading to finite-wavelength patterns or microphase separation, as captured by the mathematical structure of “Active Model B $\Omega$ 0” derived from a minimal three-component McRD model (Toffenetti et al., 15 May 2026).

Dynamical Regime	Instability Type	Pattern Selection Mechanism
Strict mass conservation	Mass-competition	Uninterrupted coarsening (wavelength unbounded)
Weakly broken conservation	Lateral instability	Splitting, arrest, finite wavelength selection
Density-dependent stiffness	Finite-wavelength	Microphase separation (AMB $\Omega$ 1, stripes, foams)

4. Asymptotic Analysis and Interface Motion

Singular perturbation theory and matched asymptotic expansions provide a rigorous framework for the existence, uniqueness, and stability of transition-layer solutions in McRD systems with bistable nonlinearities (Kuwamura et al., 2023, Ikeda et al., 2024, He et al., 6 Feb 2026). These analyses show that:

For small diffusivity ratios, one obtains sharp transitions between stable outer states.
Layer stability reduces to the sign of the derivative of the Maxwell integral (or an Evans function criterion), with stable interfaces when $\Omega$ 2 (Kuwamura et al., 2023, Ikeda et al., 2024).
In multidimensional or curved geometries, interface dynamics at long times reduce to area-preserving curvature-driven flow, subject to global mass constraints (Miller et al., 2022).

Microphase-separation regimes feature stable, periodic patterns due to the density dependence of interfacial energetics, as in AMB $\Omega$ 3, where a sign change in the effective interfacial stiffness $\Omega$ 4 stabilizes patterns at finite wavelength (Toffenetti et al., 15 May 2026).

5. Theoretical Frameworks, Model Learning, and Dualities

McRD theory has advanced as an interface between analytic theory, phase-field (“chemical potential”) models, and energetic variational approaches:

Energetic variational principles: Mass-conserving models derived from free energy and dissipation functionals yield thermodynamically consistent PDEs with built-in conservation laws (Wang et al., 2020).
Duality to phase-field models: Every Cahn–Hilliard–type chemical-potential model with a conserved order parameter can be embedded as the slow manifold of an McRD system in the fast-interconversion (reaction) limit. This duality clarifies the mapping between Maxwell construction in phase-field models and reactive turnover balance in McRD (Zhou et al., 14 May 2026).
Physically consistent model learning: Parameterized reaction–diffusion models can be constrained to enforce mass conservation and nonnegativity constraints during data-driven identification, ensuring all learned dynamics respect key conservation laws and positivity (Morina et al., 16 Dec 2025).

6. Biological and Physical Applications

McRD models have provided deep physical insight and predictive power in diverse contexts:

Cell polarity and protein patterning: Robust formation and localization of domains in cell polarity are accurately captured by mass-conserving models; such models recover observed coarsening, pattern selection, and edge localization phenomena in Rho GTPase and Min protein systems (Frey et al., 14 Dec 2025, Wigbers et al., 2019).
Biomolecular phase separation: The formation of droplet-like condensates and the associated dynamics of coarsening, nucleation/growth, and microphase separation are quantitatively accounted for by McRD dynamics, even in the absence of detailed balance (Tateno et al., 2020, Toffenetti et al., 15 May 2026).
Synthetic design and control: Experimental reconstitution and engineering of synthetic pattern-forming systems have exploited geometric criteria originating from McRD phase-space structures (Wigbers et al., 2019).

7. Computation, Numerical Methods, and Limitations

Technically, robust and efficient numerical schemes have been developed to integrate McRD models while exactly preserving mass, even on evolving domains or complex geometries (Mackenzie et al., 2019). Error-controlled, conservative finite-element methods enable simulation of McRD-driven processes such as chemotactic cell migration, providing accurate discretization of bulk–surface or bulk–bulk mass-conserving systems.

Current analytical limitations include the necessity of small parameters for singular perturbation arguments, the challenge of extending results to more complex multicomponent networks, and the handling of noise or stochastic effects, though deterministic theory forms the foundation for these extensions (Toffenetti et al., 15 May 2026, Frey et al., 14 Dec 2025). The assumption of strictly local reaction kinetics, and neglect of advection or mechanical coupling, are active areas of research.

References:

“Active Model B $\Omega$ 5 from Mass-Conserving Reaction–Diffusion Systems” (Toffenetti et al., 15 May 2026)
“Pattern Formation Beyond Turing: Physical Principles of Mass-Conserving Reaction--Diffusion Systems” (Frey et al., 14 Dec 2025)
“Wavelength selection by interrupted coarsening in reaction-diffusion systems” (Brauns et al., 2020)
“Phase-space geometry of mass-conserving reaction-diffusion dynamics” (Brauns et al., 2018)
“Single Transition Layer in Mass-Conserving Reaction-Diffusion Systems with Bistable Nonlinearity” (Kuwamura et al., 2023)
“Radially symmetric transition-layer solutions in mass-conserving reaction-diffusion systems with bistable nonlinearity” (He et al., 6 Feb 2026)
“Surface-tension-driven coarsening in mass-conserved reaction-diffusion systems” (Tateno et al., 2020)
“Pattern localization to a domain edge” (Wigbers et al., 2019)
“Stability of Single Transition Layer in Mass-Conserving Reaction-Diffusion Systems with Bistable Nonlinearity” (Ikeda et al., 2024)
“Duality Between Chemical Potential Dynamics and Reaction-Diffusion Systems” (Zhou et al., 14 May 2026)
“Physically consistent model learning for reaction-diffusion systems” (Morina et al., 16 Dec 2025)
“Mass conservative reaction diffusion systems describing cell polarity” (Latos et al., 2020)
“Global dynamics and spectrum comparison of a reaction-diffusion system with mass conservation” (Latos et al., 2015)
“Field Theory of Reaction-Diffusion: Mass Action with an Energetic Variational Approach” (Wang et al., 2020)
“A Conservative Finite Element ALE Scheme for Mass-Conserving Reaction-Diffusion Equations on Evolving Two-Dimensional Domains” (Mackenzie et al., 2019)
“Generation and motion of interfaces in a mass-conserving reaction-diffusion system” (Miller et al., 2022)