Planar Free Elastic Flow

• Planar free elastic flow is defined as the L²-gradient flow of Euler’s elastic energy for curves, governed by a fourth-order degenerate parabolic PDE.
• It extends to diverse configurations such as closed and open curves and networks, with specific boundary conditions ensuring existence and regularity.
• Applications span material science, interface dynamics, and image processing, highlighting both analytic challenges and practical implications.

Planar free elastic flow describes the evolution of planar curves or interfaces driven by the $L^2$-gradient of the elastic (bending) energy, commonly termed Euler’s elastic energy. The flow is represented by a fourth-order geometric evolution equation and arises in diverse settings, including closed and open curves, networks, surfaces with translation symmetry, and in the context of surface diffusion equations perturbed by nonlocal elasticity. In its canonical form for a curve $\gamma \colon S^1 \to \mathbb{R}^2$, the elastic energy is defined by

$E[\gamma] = \frac{1}{2} \int_\gamma k^2\, ds,$

where $k$ is the curvature and $ds$ is the arclength element. The main analytic, geometric, and applied interest in planar free elastic flow is due to its high-order parabolic structure, lack of compactness or constraint (unlike length- or area-constrained flows), and its relevance to material science phenomena such as epitaxial growth, interface dynamics, and stress-driven rearrangement instabilities.

1. Mathematical Formulation of Planar Free Elastic Flow

The planar free elastic flow corresponds to the $L^2$-gradient flow of the elastic energy functional for curves, or more generally, boundaries in the plane. For a smooth immersion $\gamma \colon S^1 \to \mathbb{R}^2$, the evolution equation is

$$\partial_t \gamma = -\big( k_{ss} + \frac{1}{2} k^3 \big) \nu$$

where $k_{ss}$ denotes the second derivative of curvature with respect to arc length, and $\nu$ is the inward-pointing unit normal. This equation is a fourth-order degenerate parabolic PDE, highly sensitive to boundary and initial data.

Much of the literature extends this framework:

• Networks: For networks of curves, such as $\Theta$-networks and triods, the flow generalizes to multiple curves joined at junctions with appropriate coupling and geometric boundary conditions at the junctions (Garcke et al., 2017).
• Surface Diffusion with Elasticity: When considering the evolution of a planar interface as the boundary of an elastic domain, the flow can be written as

$$V_t = \Delta_\tau \big( g(v_t) k_t - Q(E(u_t)) \big)$$

(Equation 1), where $V_t$ is the normal velocity, $\Delta_\tau$ is the tangential Laplacian, $k_t$ is curvature, $g$ encodes anisotropy, and $Q(E(u_t))$ is an elastic term defined via the solution to an elliptic system in the bulk (Fusco et al., 2017).

• Open Curves and Boundary Conditions: The flow can also be considered for open curves with fixed, free, or partially free endpoints (using e.g. Navier or Neumann-type conditions) (Diana, 2023, Kemmochi et al., 2023).
• Generalizations: Elastic flows for $p$-elastic energies (with $p\neq2$), length-penalized/constrained variants, or equipped with area constraints via an $H^{-1}$-gradient paradigm (yielding a sixth-order PDE) are also established (Okabe et al., 2018, Okabe et al., 2021, Langer, 7 Aug 2024, Gazwani et al., 22 Nov 2024).

Formulation Table for Various Cases:

Setting	Evolution Equation	Constraints
Closed curve	$\partial_t\gamma = - (k_{ss} + \frac{1}{2}k^3)\nu$	None (free)
Length-penalized	$$\partial_t\gamma = - (k_{ss} + \frac{1}{2}k^3 - \lambda k)\nu$$	Fixed $\lambda>0$
Area-preserving	$$\partial_t^\perp\gamma = \nabla_s^2( k_{ss} + \frac{1}{2} k^3 - \lambda k )$$	Area constraint (Langer, 7 Aug 2024)
Networks	As above, one equation per curve component + coupled boundary/junction conditions	Concurrency, angle, various, see (Garcke et al., 2017)

2. Existence, Uniqueness, and Regularity

• Short- and Long-Time Existence:
  • For closed curves, the free elastic flow yields smooth solutions for all time when the initial data are sufficiently regular, even in the absence of length or area constraints (Miura et al., 19 Apr 2024, Andrews et al., 14 Sep 2025). The same holds for open curves with generalized Neumann or partial free boundary conditions, provided endpoints do not degenerate (Gazwani et al., 22 Nov 2024, Diana, 2023).
  • For elastic networks (e.g., planar $\Theta$-networks), short-time existence and uniqueness follow from reformulation as a quasilinear parabolic system and subsequent maximal regularity arguments (Garcke et al., 2017).
• Well-posedness Strategies:
  • The primary technical challenge is the high-order, quasilinear nature of the equations and the degeneracy of the parabolic operator.
  • Approaches include fixed-point arguments based on energy identities and contraction mappings (Fusco et al., 2017, Garcke et al., 2017), minimizing-movements variational discretization (Okabe et al., 2018, Okabe et al., 2021), and regularization via parabolic Hölder spaces and the Solonnikov theory (Diana, 2023).
• Special Cases:
  • In the context of surface diffusion with elasticity, short-time existence is proven by treating the nonlocal elastic term as a forcing and employing a fixed-point scheme using a key energy identity to ensure contractiveness (Fusco et al., 2017).

3. Asymptotic Behavior and Stability

• Closed Planar Curves: For curves initially close to a round (possibly multiply-covered) circle (i.e., having small $L^2$-norm of $k_s$ or, in improved results, small $L^2$-norm of $k-1$), the rescaled planar free elastic flow converges smoothly to a circle, with explicit decay rates available as

$$e(t) = \int (k-1)^2 ds \leq 2 e(0) \exp(-\tfrac{7}{4} t)$$

in rescaled time (Miura et al., 19 Apr 2024, Andrews et al., 14 Sep 2025).

• Open Curves in Cones: For open curves inside cones with suitable generalized Neumann conditions, the free elastic flow causes profiles to converge smoothly to an expanding self-similar arc rather than a stationary elastica (Gazwani et al., 22 Nov 2024). This contrasts with length-penalized/constrained flows, which stabilize to stationary arcs.
• Networks: Willmore-type free elastic flows on planar networks, with suitable boundary conditions at multiple junctions, exhibit well-posed dynamics, but convergence and singularity formation can depend delicately on geometric and topological structure (Garcke et al., 2017).
• Surface Diffusion with Elastic Term: Strictly stable stationary sets (i.e., those for which the second variation of the free energy is strictly positive) are exponentially attracting: if the initial set $F_0$ is close (in an appropriate norm) to a strictly stable stationary set $G$, the solution $F_t$ converges exponentially fast toward $G$ (Fusco et al., 2017).

4. Role and Influence of Elasticity

• Nonlocal Effects: When bulk elasticity is present (e.g., when the planar boundary is the edge of a two-phase elastic domain), the evolution law acquires a highly nonlocal term: the chemical potential driving the flow incorporates a contribution $Q(E(u))$ arising from an elliptic system for the elastic equilibrium displacement $u$ in the bulk (Fusco et al., 2017). This alters both the analytic properties and physical behavior, leading to long-range stress effects, altered stability landscapes, and new equilibrium phenomena. The nonlocality is mathematically manifested in the requirement to solve an elliptic system at each time step, with the boundary coupling to the elastic field in the domain.
• Strict Stability and Energy Gaps: The notion of strictly stable stationary configurations is articulated via the positivity of the second variation (i.e., coercivity of the quadratic form associated to the energy functional). This “energy gap” ensures penalization of small perturbations and is crucial for exponential convergence in the flow (Fusco et al., 2017).

5. Boundary and Initial Conditions

• Closed vs. Open Curves: The analytic form of the evolution heavily depends on whether curves are closed, open, or form networks.
  • Closed Curves: Initial data are often parametrized via curvature, tangential angle, or as normal graphs over reference configurations (Andrews et al., 14 Sep 2025, Miura et al., 19 Apr 2024).
  • Open Curves: Properly posed with Dirichlet (fixed), Neumann (free), Navier (partial free), or generalized Neumann boundary conditions; well-posedness requires precise compatibility between the bulk evolution and these boundary prescriptions (Diana, 2023, Gazwani et al., 22 Nov 2024).
  • Networks: At junctions, concurrency (C⁰), angle (C¹), curvature, and higher-order conditions are imposed, and the satisfaction of the Lopatinskii–Shapiro condition is essential for the well-posedness of the parabolic system (Garcke et al., 2017).
• Thresholds for Embeddedness: For planar closed curves under the elastic flow, sharp energy thresholds (expressed via a normalized quantity $$\overline B[\gamma] = L[\gamma] \int |\kappa|^2\, ds$$) guarantee preservation of embeddedness; for the planar case, the threshold is achieved uniquely by an “elastic two-teardrop” configuration (Miura et al., 2021).

6. Applications and Broader Implications

• Materials Science and Continuum Mechanics: Free elastic flows (and their generalizations) rigorously model the evolution of interfaces in stressed solids, the dynamics of epitaxially strained films, phase separation processes in elastic media, and void formation (Fusco et al., 2017).
• Microstructure and Nonlinear Elasticity: In the context of martensitic crystals and nematic elastomers, explicit “Conti-type” constructions provide exactly stress-free deformations in planar nonlinear elasticity, connecting the well-posedness theory of flow to the theory of microstructure and multiwell problems (Cesana et al., 2019).
• Image Processing and Computer Graphics: The analytic and energetic properties of free elastic flow form the backbone of geometric curve fairing, shape optimization, edge detection, and the evolution of filamentary structures in image analysis.
• Mathematical Analysis: The paper of regularity, threshold behavior, convergence, and stability under free elastic flow impacts the understanding of higher-order nonlinear parabolic PDEs, variational calculus, and geometric analysis. The development of energy methods, embedding criteria, and the characterization of rigidity phenomena (such as Li–Yau–type inequalities) for both compact and non-compact curves demonstrates the breadth of influence (Miura et al., 26 Aug 2025).
• Instabilities in Complex Fluids: Extensions to surface diffusion with elastic terms or the analysis of instability mechanisms in elastic channel flows (such as the elastically-driven Kelvin–Helmholtz-like instability) underscore the physical richness of the subject and its intersection with modern hydrodynamics (Jha et al., 2020).

7. Recent Advances and Open Problems

• Sharp Asymptotic Rates and Stability: The latest results improve earlier “closeness” conditions required for stability by establishing convergence directly under small $L^2$-norm of the curvature scalar, with explicit exponential rates (Andrews et al., 14 Sep 2025).
• Preservation of Positivity on Complete Curves: For non-compact (complete) curves with infinite length, the development of adapted elastic energies permits the establishment of sharp optimal thresholds for the preservation of embeddedness and graphicality, along with rigidity results and Li–Yau-type inequalities (Miura et al., 26 Aug 2025).
• Self-similar and Expanding Solutions: In configurations without length penalty or constraint, the natural attractors of the flow are explicitly classified as expanding self-similar arcs or circles, in clear contrast to the static equilibria of length-penalized flows (Gazwani et al., 22 Nov 2024, Miura et al., 19 Apr 2024).
• Area and Volume Preservation through $H^{-1}$-Gradient Flow: Introduction of sixth-order area-preserving flows allows for the paper of geometric evolution under strict area constraint with locally defined operators, avoiding nonlocal Lagrange multipliers (Langer, 7 Aug 2024).
• Higher-Order Flows in Heterogeneous and Network Domains: Analysis now extends to flows in elastic wires with heterogeneous material properties (density/spontaneous curvature dependence) (Dall'Acqua et al., 2022), and to the global existence and regularity of elastic flows in network and free boundary scenarios (Diana, 2023, Novaga et al., 2019).
• Migration and Topology Change: The construction of “migrating” elastic flows—curves that switch from one half-plane to the other—demonstrates nontrivial topology change in fourth-order flows, in stark contrast with maximum principle-governed second-order flows (Kemmochi et al., 2023).
• Open Problems: Key directions involve classification of all self-similar or translating solutions, long-time behavior without symmetry or smallness assumptions, singularity formation and topology change, as well as extension to higher codimension and fully nonlocal elastic effects.

---

Planar free elastic flow stands as a paradigmatic class of high-order geometric evolution equations, unifying analytic, geometric, and physical insights, and continues to stimulate developments across analysis, geometry, and applied mathematics.