KWC-Type Grain Boundary Motion

• KWC-type grain boundary motion is modeled by pseudo-parabolic gradient systems that introduce higher-order regularizations to secure uniqueness and well-posedness.
• The system employs a variational gradient-flow structure with time-discretization and convex minimization techniques, yielding rigorous energy dissipation and optimal control.
• This approach bridges classical and pseudo-parabolic models, enhancing spatial and temporal regularity for accurate simulation of grain boundary dynamics in polycrystals.

A pseudo-parabolic gradient system is a class of evolution equation characterized by gradient-flow dynamics in a generalized, Hilbert-product metric, incorporating additional time-derivative-in-Laplacian (“pseudo-parabolic”) regularization. These systems generalize classical parabolic gradient flows by introducing “inertia-like” terms involving higher spatial regularity of the time-derivative. The pseudo-parabolic Kobayashi–Warren–Carter (KWC)-type equations, a prime instance, resolve central difficulties in modeling grain boundary motion, notably ensuring mathematical well-posedness—including uniqueness—without sacrificing the variational gradient-flow structure. These models have immediate significance for materials science, particularly polycrystal evolution and grain-boundary dynamics.

1. The Pseudo-Parabolic KWC System: Definition and Structure

Given a bounded domain $\Omega \subset \mathbb{R}^N$ ($N=1,2,3$), pseudo-parabolic KWC-type systems describe the evolution of two order parameters:

• $\eta=\eta(t,x)$: orientation order parameter,
• $\theta=\theta(t,x)$: orientation angle (crystal orientation).

The canonical pseudo-parabolic KWC system is

$$\begin{cases} \partial_{t}\eta - \Delta\left(\eta + \mu^2 \partial_{t}\eta\right) + g(\eta) + \alpha'(\eta)|\nabla\theta| = u(t,x), & (t,x) \in Q,\ \nabla\left(\eta + \mu^2 \partial_{t}\eta\right)\cdot n = 0, & (t,x) \in \Sigma,\ \eta(0,x) = \eta_0(x), & x\in\Omega, \end{cases}$$

$$\begin{cases} \alpha_0(\eta)\partial_{t}\theta - \mathrm{div}\left( \alpha(\eta) \frac{\nabla\theta}{|\nabla\theta|} + \nu^2 \nabla\partial_{t}\theta \right) = v(t,x), & (t,x)\in Q,\ \left(\alpha(\eta)\frac{\nabla\theta}{|\nabla\theta|} + \nu^2 \nabla\partial_{t}\theta\right)\cdot n = 0, & (t,x)\in\Sigma,\ \theta(0,x) = \theta_0(x), & x\in\Omega. \end{cases}$$

where $\mu,\nu > 0$ are pseudo-parabolic parameters; $g(\eta)=G'(\eta)$, $\alpha_0>0$ and $\alpha\geq 0$ are prescribed mobilities (Antil et al., 2024).

The terms $\mu^2\Delta(\partial_t\eta)$ and $\nu^2\Delta(\partial_t\theta)$ are the pseudo-parabolic regularizations, which act as higher-order spatial “inertia” terms and are crucial for increasing temporal and spatial regularity.

The associated nonsmooth free energy is

$\E[\eta,\theta]=\frac12 \int_\Omega |\nabla \eta|^2 \,dx + \int_\Omega G(\eta)\,dx + \int_\Omega \alpha(\eta)|D\theta|$

with $|D\theta|$ the total variation measure, capturing the grain-boundary energy as $\alpha(\eta)|\nabla\theta|$-type contributions (Antil et al., 2024).

2. Variational Gradient-Flow Structure

Pseudo-parabolic KWC systems retain a full energy-dissipation (Lyapunov) structure: $\frac{d}{dt}\E[\eta(t),\theta(t)] + \int_\Omega \left|\partial_t\eta + \mu^2 \partial_t(-\Delta \eta)\right|^2\,dx + \int_\Omega \alpha_0(\eta)|\partial_t\theta|^2\,dx + \nu^2\int_\Omega |\nabla\partial_t\theta|^2\,dx = \int_\Omega u\,\partial_t\eta + v\,\partial_t\theta\,dx,$ implying non-increasing energy in the absence of external forcing (Antil et al., 2024, Mizuno, 2024).

The system is the $L^2$-gradient flow of the energy in a product Hilbert metric determined by a variable-coefficient (state-dependent) operator

$\A_0(\eta):\begin{bmatrix}\xi\\psi\end{bmatrix} \mapsto \begin{bmatrix}\xi-\mu^2\Delta\xi\ \alpha_0(\eta)\psi - \nu^2\Delta\psi\end{bmatrix}$

which persists for forced or generalized anisotropic variants (Antil et al., 17 Dec 2025).

3. Analytical Results: Well-Posedness, Uniqueness, Regularity

The pseudo-parabolic regularization fundamentally improves mathematical properties:

• Existence and uniqueness: For convex $\alpha$, locally Lipschitz $g$, and regular initial data, there exists a unique global-in-time solution in strong Sobolev spaces; and continuous dependence on initial data and forcing holds (Antil et al., 2024, Mizuno, 2024, Antil et al., 17 Dec 2025, Mizuno et al., 7 Dec 2025).
• Energy dissipation and regularity: Both $\partial_t\eta$ and $\partial_t\theta$ enjoy $L^2$- and $H^1$-type estimates, with $\eta, \theta \in W^{1,2}(0,T; H^2(\Omega))$ under suitable conditions (Antil et al., 2024, Mizuno et al., 7 Dec 2025).
• Abstract extension: This structure generalizes to nonlinear systems with state-dependent metrics and coefficients, enabling analysis of anisotropic grain-boundary models and related applications (Antil et al., 17 Dec 2025).

The pseudo-parabolic terms enforce spatial coercivity in energy estimates, systematically ruling out multiple solutions—a core advance over the original KWC system, for which uniqueness typically fails except under severe simplifications (e.g., constant mobility, presence of additional elliptic regularization) (Antil et al., 2024, Mizuno et al., 7 Dec 2025).

4. Physical and Modeling Implications

Pseudo-parabolic gradient systems closely align with physical grain boundary motion:

• Grain boundary localization: The $\alpha(\eta)|D\theta|$ term in the energy imposes total variation penalization on orientation gradients, localizing misorientation energy in regions where $\eta$ is small (the grain boundary) (Antil et al., 2024, Mizuno, 2024, Watanabe et al., 2020).
• State-dependent mobility: Functions $\alpha_0(\eta)$ and $\alpha(\eta)$ model physical grain boundary mobility and viscosity. Taking them as functions of the order parameter allows spatial variation, representing, e.g., high mobility near boundaries and “pinned” behavior in the bulk (Antil et al., 17 Dec 2025).
• Dynamic regularization: The pseudo-parabolic regularization ($\mu,\nu$) has a physical interpretation as finite relaxation (or inertia) time. In the sharp interface limit, the macroscopic boundary velocity is governed by classical curvature- and misorientation-driven kinetics (Antil et al., 2024, Mizuno, 2024, Antil et al., 17 Dec 2025).
• Anisotropic extensions: With appropriate choice of energy density, pseudo-parabolic KWC systems accommodate orientation-dependent anisotropy, via e.g. $\gamma_0(R(\theta)\nabla\theta)$, and more general state-dependence (Antil et al., 17 Dec 2025).

5. Comparison with Classical KWC and Parabolic Systems

Feature	Classical KWC	Pseudo-parabolic KWC
Regularization	None (μ,ν=0)	μ > 0, ν > 0
Uniqueness	Fails except in special cases	Always holds under convexity assumptions
Energy structure	Formal only	Rigorous, fully variational
Solution regularity	BV, low spatial/temporal regularity	$H^2$-spatial, $L^2$/$H^1$-temporal
Applicability	Limited to over-simplified mobilities	Handles state-dependent, anisotropic mobilities

In the classical (parabolic) KWC system, lack of pseudo-parabolic regularization leads to weak regularity and violation of uniqueness except when mobilities are constant or additional elliptic penalty is imposed (Antil et al., 2024, Mizuno et al., 7 Dec 2025). The pseudo-parabolic approximation is thus indispensable for analytic tractability, especially in the presence of variable (order-parameter or orientation-dependent) coefficients (Mizuno et al., 7 Dec 2025).

6. Methodology: Analytical and Numerical Approaches

Key methodological features established in recent pseudo-parabolic KWC work include:

• Time-discretization and convex minimization: Backward Euler-type time-discretization reduces each temporal increment to an elliptic variational problem, ensuring existence of minimizers and facilitating rigorous a priori estimates (Antil et al., 2024, Mizuno, 2024, Antil et al., 17 Dec 2025, Mizuno et al., 7 Dec 2025).
• Energy-compactness arguments: Uniform bounds on energy and time-derivatives, together with Aubin–Simon compactness and Mosco/Γ-convergence, guarantee passage to the continuous-time limit (Antil et al., 2024, Mizuno, 2024).
• Numerical and optimization theory: The well-posed (regularized) pseudo-parabolic system uniquely permits optimal control theory, including derivation of first-order optimality systems and efficient adjoint-based numerical optimization in grain boundary motion (Antil et al., 11 Jun 2025, Kubota et al., 2020).
• Extension to anisotropy: The abstract theory for pseudo-parabolic gradient systems with state-dependent coefficients, including anisotropic/homogenization operators, fits the KWC framework and a variety of other physically relevant systems (Antil et al., 17 Dec 2025).

7. Relevance and Generalization

Pseudo-parabolic gradient systems now provide a mathematically robust, physically faithful framework for

• polycrystal grain boundary motion,
• coupled order-parameter and orientation evolution,
• anisotropic and state-dependent mobility scenarios,
• free-boundary problems in which higher regularity and uniqueness are essential,
• effective optimal control and inverse problems for interface-evolution PDEs (Antil et al., 2024, Mizuno, 2024, Antil et al., 17 Dec 2025, Mizuno et al., 7 Dec 2025, Antil et al., 11 Jun 2025).

Limitations remain: analysis inherently relies on convexity of mobilities, sufficient smoothness of the initial data, and the singular flux structure of the orientation equation. Nonetheless, the pseudo-parabolic KWC system stands as the first fully variational, uniqueness-guaranteed PDE realization of grain-boundary motion with the physically correct energy, offering a rigorous bridge between geometric interface models and regularized phase field descriptions (Antil et al., 2024, Antil et al., 17 Dec 2025).