Cahn–Hilliard Two-Phase Model

• The Cahn–Hilliard two-phase model is an energy-based framework that defines phase separation through a conserved order parameter and free-energy minimization.
• It integrates fourth-order nonlinear dynamics with transport and hydrodynamic coupling to accurately simulate interfacial phenomena and multiphase flows.
• Advanced numerical methods and coupled CH–NS formulations enable stable, efficient predictions of microstructure evolution in complex materials.

The Cahn–Hilliard two-phase model is a widely adopted energy-based framework for describing the dynamics of multiphase systems with diffuse interfaces, particularly in fluid mechanics, materials science, and interfacial phenomena. Its mathematical formulation hinges on the evolution of a conserved order parameter (typically denoted φ), coupled to transport, hydrodynamic, and, in advanced forms, surface or viscoelastic effects. The central mechanism is phase separation driven by the minimization of a free-energy functional, subject to mass conservation and nonlinear, fourth-order dynamics.

1. Free-Energy Functional and Chemical Potential

The Cahn–Hilliard model posits a Lyapunov functional for the system's free energy, typically of the form

$$F[\phi] = \int_\Omega \left[ f(\phi) + \frac{\varepsilon^2}{2}|\nabla \phi|^2 \right] dV$$

where $f(\phi) = \frac{1}{4}(\phi^2 - 1)^2$ is a double-well bulk energy, and $\varepsilon > 0$ is the capillary (interface thickness) parameter (Manna et al., 27 Aug 2025). The chemical potential is the variational derivative

$$\mu(\phi) = \frac{\delta F}{\delta \phi} = \phi(\phi^2 - 1) - \varepsilon^2 \nabla^2 \phi$$

which governs the dissipative evolution towards energy minimization. For degenerate two-phase models, entropy density $H(c) = c \log c - c + 1$ provides further regularity control in the presence of nonlinear mobilities or positivity constraints (Cancès et al., 2020, Cancès et al., 2017).

2. Cahn–Hilliard Evolution and Two-Phase Coupling

The evolution equation for the two-phase field φ is given by the conservative H$^{-1}$-gradient flow

$$\frac{\partial \phi}{\partial t} + \mathbf{u} \cdot \nabla \phi = \nabla \cdot [ M(\phi) \nabla \mu ]$$

where $M(\phi) \geq 0$ is the mobility—either constant for bulk diffusion, or degenerate to localize transport near interfaces (e.g., $M(\phi) \sim (1 - \phi^2)^2$) (Manna et al., 27 Aug 2025). In degenerate models with explicit phase concentrations $c_i$, the system can be extended to

$$\partial_t c_i - \mathrm{div}\, J_i = \theta_i \Delta c_i, \quad J_i = -\frac{c_i}{\eta_i} \nabla (\mu_i + \Psi_i)$$

with the constraint $c_1 + c_2 = 1$ and capillarity coupling $$\mu_1 - \mu_2 = -\alpha \Delta c_1 + \kappa (1 - 2c_1)$$ (Cancès et al., 2020). Mass conservation is strictly enforced: $\int_\Omega \phi\, dV = \text{const}$.

3. Hydrodynamic, Surface, and Coupled Extensions

The Cahn–Hilliard equation is often coupled to incompressible Navier–Stokes equations to create the CH–NS system: $$\begin{aligned} &\rho(\phi)[\partial_t \mathbf{u} + (\mathbf{u} \cdot \nabla)\mathbf{u}] = -\nabla p + \nabla \cdot [\eta(\phi) D(\mathbf{u})] + \mathbf{F}_\mathrm{cap} + \mathbf{F}_\mathrm{body} \ &\nabla \cdot \mathbf{u} = 0 \end{aligned}$$ Here, density and viscosity are phase-field interpolated: $$\rho(\phi) = \frac{1}{2}\left[(1+\phi)\rho_1 + (1-\phi)\rho_2 \right], \quad \eta(\phi) = \frac{1}{2}\left[(1+\phi)\eta_1 + (1-\phi)\eta_2 \right]$$ The capillary force is given by the Korteweg form: $$\mathbf{F}_{\text{cap}} = \tilde{\sigma} \varepsilon^{-1} \mu \nabla \phi$$ enabling accurate modeling of interface-driven flows such as bubble rise, droplet breakup, and hydrodynamic instability (Manna et al., 27 Aug 2025). Extensions to two-phase flow on surfaces and with bending elasticity use extrinsic (surface Laplacian) operators and additional Helfrich-type elasticity terms (Bachini et al., 2023, Abels et al., 2024).

4. Thermodynamic Structure and Energy Dissipation

Central to the model is the energy dissipation law, ensuring that the system's free energy is non-increasing: $$\frac{d}{dt}F[\phi] = -\int_\Omega M(\phi)|\nabla \mu|^2\, dV \leq 0$$ In complex two-phase, multi-component, or two-flux models, entropy-based functionals and constrained Wasserstein gradient flows generalize this principle. For example, in two-flux models, dissipation is enhanced due to independent phase velocities subject only to global incompressibility: $$D_{\rm nonlocal} = \int_\Omega \frac{|J_1 + J_2|^2}{m_1 c_1 + m_2 c_2} \, dx + \int_\Omega \eta(c_1) |\nabla(\mu_1 - \mu_2)|^2\, dx$$ yielding faster free energy decay compared to classical local Cahn–Hilliard models (Cancès et al., 2017).

5. Numerical Methods and Discretization Strategies

Robust numerical integration of Cahn–Hilliard two-phase models necessitates special treatment for stability, conservation, and efficiency. Key techniques include:

• Decoupled pressure-projection methods with staggered grids and explicit Euler time stepping, ensuring accurate hydrodynamic–phase-field coupling (Manna et al., 27 Aug 2025).
• Finite volume and finite element schemes with convex–concave splitting for unconditional energy stability; degenerate mobilities are handled via implicit time stepping and entropy-based estimates (Cancès et al., 2020, Wu et al., 2016).
• Divergence-free hybridizable discontinuous Galerkin (HDG) schemes eliminate the need for pressure solves and are compatible with mass conservation, achieving high-order convergence in convection-dominated flows (Fu, 2020).
• Gradient-stable time stepping (e.g., Eyre splitting) and spectral filtering allow efficient tracking of metastable patterns and microstructure selection (Choksi et al., 2011).
• Dynamic boundary conditions for the Cahn–Hilliard equation enable realistic modeling of contact angle hysteresis and mass exchange between bulk and surface (Gu et al., 2024).
• Projection methods and semi-implicit schemes for steady-state seeking reduce computational cost while preserving mass conservation and energy dissipation properties (Gu et al., 2024, Riethmüller et al., 14 Jan 2026).

6. Applications in Multiphase Flows, Pattern Formation, and Interfacial Phenomena

The Cahn–Hilliard two-phase framework is foundational in modeling:

• Hydrodynamic instabilities (Rayleigh–Taylor, Plateau–Rayleigh), droplet and bubble dynamics in variable property flows (Manna et al., 27 Aug 2025).
• Microphase separation and 2D morphology selection (lamellae, hex-spots) in systems with long-range interactions; the phase diagram is sharply partitioned, and coarsening is arrested by nonlocal terms (Choksi et al., 2011).
• Coupled phase–hydrodynamics (coarsening acceleration via advective transport), where fluid advection dominates late-stage domain growth, drastically accelerating phase-separation compared to pure diffusion (Howard et al., 2024, Feng et al., 2019).
• Surface and interface phenomena, including biomembrane modeling, bending moments, and two-phase flows on evolving or deformable surfaces (Bachini et al., 2023, Abels et al., 2024).
• Viscoelastoplastic two-phase geodynamic flows, with stress-diffusion regularization and relative-energy inequalities yielding well-posedness for complex rheologies (Cheng et al., 29 Sep 2025).
• Phase field crystal models coupling lattice symmetry and composition for binary alloys, capturing microstructural evolution across phase boundaries (Balakrishna et al., 2017).

7. Analytical Well-Posedness, Multiphysics Extensions, and Open Directions

Rigorous existence and uniqueness results are established for both weak and strong solution notions under energy, mass, and entropy dissipation constraints. These include:

• Existence and uniqueness of global weak and strong solutions for matched and variable densities, surfactant-coupled models, and models with dynamic boundary or surface exchange (Primio et al., 2022, Giorgini et al., 2022, Abels et al., 2024).
• Thermodynamically consistent formulations for fluid-solid systems with precipitation/dissolution, recovering sharp-interface transmission conditions in vanishing thickness limits (Riethmüller et al., 14 Jan 2026, Rohde et al., 2019).
• Stochastic models with multiplicative noise components, proven to possess martingale solutions and energy control, extending classical Cahn–Hilliard–Brinkman models to random environments (Brzeźniak et al., 10 Jan 2026).
• Multiphysics extensions: coupling with thermal, electrohydrodynamic, phase-change, and chemical reaction phenomena (e.g., precipitation/dissolution) via augmented energy functionals and splitting methods (Riethmüller et al., 14 Jan 2026).

A plausible implication is that the model's geometric generalizations (to surfaces, biomembranes, or evolving manifolds) and generalized mobility structures are increasingly significant in biophysical and advanced technological applications.

---

The Cahn–Hilliard two-phase model, in its modern high-dimensional, hydrodynamic, and surface-coupled forms, continues to serve as a mathematically rigorous and computationally tractable organizing principle for multiphase systems, interfacial pattern formation, and dynamic evolution across scientific disciplines (Manna et al., 27 Aug 2025, Cancès et al., 2020, Choksi et al., 2011, Bachini et al., 2023, Abels et al., 2024, Cancès et al., 2017, Wu et al., 2016, Balakrishna et al., 2017, Riethmüller et al., 14 Jan 2026, Giorgini et al., 2022, Primio et al., 2022, Brzeźniak et al., 10 Jan 2026).