Stability of the Shrinking Semi-Circle Under the Free Boundary Curve Shortening Flow

This paper establishes a sharp rate of convergence for a free-boundary curve shortening flow within a convex domain in $\mathbb{R}^2$, demonstrating that the flow evolves to a round half-point in finite time.

The Gist

Here is an explanation of the paper "Stability of the Shrinking Semi-Circle Under the Free Boundary Curve Shortening Flow," translated into simple language with everyday analogies.

The Big Picture: A Soap Bubble on a Wall

Imagine you have a soap film stretched across a wire frame. If the wire frame is a perfect circle, the soap film is a circle. But what happens if the wire frame is a half-circle resting on a flat table (a wall), and the film is trying to shrink itself to the smallest possible area?

This is the scenario the authors are studying. They are looking at a curve (like a soap film edge) that is shrinking because of its own "tension" (curvature). The catch is that the ends of the curve are stuck to a wall (the boundary of a convex domain), and they can slide along the wall freely.

The Question: As this curve shrinks and eventually disappears (pops), what shape does it look like right before it vanishes?

The Answer: It looks like a perfect half-circle (a semi-circle).

The Problem: "How Fast" and "How Steady"?

Previous mathematicians (like Gage, Hamilton, and Grayson) had already figured out that:

1. If you have a closed loop (like a rubber band), it shrinks into a perfect circle and vanishes.
2. If you have a curve with ends on a wall, it shrinks into a perfect half-circle and vanishes.

However, there was a missing piece of the puzzle: How exactly does it get there?

Imagine a wobbly, lopsided rubber band shrinking. Does it just wiggle around and then suddenly snap into a perfect circle? Or does it smoothly and steadily straighten out? And how fast does it straighten out?

The authors of this paper wanted to prove that the curve doesn't just eventually look like a half-circle; it converges to that shape at a specific, sharp speed. They wanted to measure the "wobble" and prove it disappears at a predictable rate.

The Challenge: The "Unstable" Modes

Think of the shrinking half-circle as a spinning top that is perfectly balanced.

• Time Translation: If you start the clock one second later, the shape is the same, just at a different stage of shrinking.
• Horizontal Translation: If you slide the half-circle left or right along the wall, it's still the same shape.

In math terms, these are "unstable modes." If you try to analyze the curve, it might look like it's wobbling, but actually, it's just sliding left or right, or the clock is ticking at a slightly different speed. This makes it hard to prove that the curve is actually stabilizing into a perfect shape.

The Solution: The "Normalization" Trick
To fix this, the authors invented a clever way to "normalize" (adjust) the view.

• Imagine you are filming the shrinking curve.
• Every time the curve slides left or right, you move the camera to keep it centered.
• Every time the clock seems to tick too fast or too slow, you adjust the frame rate.

By constantly adjusting the camera and the clock to keep the curve's area and center of mass perfectly fixed, they removed the "wobble" caused by sliding and timing. Once they did this, they could see that the curve was indeed settling down into a perfect half-circle very smoothly.

The Metaphor: The Melting Ice Cream Cone

Imagine a melting ice cream cone that is stuck to a flat plate.

1. The Flow: The ice cream is melting, trying to minimize its surface area.
2. The Shape: As it melts, it gets smaller and smaller.
3. The Instability: Sometimes the ice cream might slide a bit to the left, or melt a tiny bit faster on one side.
4. The Math: The authors proved that if you ignore the sliding and the speed differences, the shape of the melting ice cream becomes a perfect semi-circle at a very specific, predictable speed.

They calculated that the "imperfections" (the parts that aren't a perfect semi-circle) shrink away at a rate of roughly $(T-t)^{1-\delta}$. In plain English: As the time $T$ (when it vanishes) approaches, the errors get smaller and smaller very quickly.

Why Does This Matter?

1. Precision: It's not enough to know that something happens; in physics and engineering, we need to know how fast and how precisely it happens. This paper gives a "speed limit" for how fast the curve becomes perfect.
2. Uniqueness: It helps mathematicians prove that there is only one way for these shapes to behave. If you see a curve shrinking this way, you know exactly what it is doing.
3. Higher Dimensions: This work on 2D curves (lines) helps mathematicians understand more complex shapes in 3D or 4D space. If we understand the simple case perfectly, we can build better models for complex ones.

Summary

The paper is like a high-speed camera analysis of a shrinking soap bubble on a wall. The authors proved that:

1. The bubble always turns into a perfect half-circle before it pops.
2. Even if it starts out wobbly or lopsided, it smooths out perfectly.
3. They calculated the exact speed at which the "wobbles" disappear, proving that the process is stable and predictable.

They achieved this by creating a special mathematical "camera" that automatically corrects for sliding and timing errors, allowing them to see the true, stable behavior of the shrinking curve.

Technical Summary

Here is a detailed technical summary of the paper "Stability of the Shrinking Semi-Circle Under the Free Boundary Curve Shortening Flow" by Theodora Bourni, Nathan Burns, and Mat Langford.

1. Problem Statement

The paper addresses the asymptotic behavior of the free-boundary curve shortening flow (CSF) in a convex domain $\Omega \subset \mathbb{R}^2$.

• Context: In the classical closed-curve CSF, Gage and Hamilton proved that convex curves shrink to a round point with exponential convergence. For free-boundary curves (where the curve endpoints move along the boundary $\partial \Omega$), previous work by Stahl (in higher dimensions) and recently by Bourni et al. (in 2D) established that such flows either converge to a critical chord (in infinite time) or shrink to a "round half-point" (a semicircle orthogonal to the boundary) in finite time $T$.
• The Gap: While the existence of the limit (the shrinking semicircle) was known, the rate of convergence to this limit after rescaling about the extinction point was not established. Specifically, it was unknown whether the convergence was sharp (exponential) or merely polynomial, and how to handle the instability modes inherent in the linearization.
• Goal: The authors aim to establish a sharp rate of convergence for the rescaled flow to the unit semicircle in all $C^k$ norms, quantifying the stability of the shrinking semicircle.

2. Methodology

The proof relies on a dynamic stability analysis of the shrinking semicircle as a self-similar solution. The methodology involves several sophisticated steps:

A. Rescaling and Linearization

The authors introduce a rescaling of the flow around the extinction point $p \in \partial \Omega$ at time $T$.

• Rescaled Variables: They define a rescaled curve $\tilde{\Gamma}_t$ and a new time variable $\tilde{t}$.
• Support Function: The flow is parametrized using the Gauss map (angle $\theta$), representing the curve via its support function $\sigma(\theta, t)$.
• Linearized Operator: Linearizing the flow equation around the unit semicircle ($\sigma = 1$) yields a linear operator with eigenvalues corresponding to modes $j$.
  • $j=0$: Corresponds to time translation (unstable).
  • $j=1$: Corresponds to horizontal translation (unstable).
  • $j \ge 2$: Corresponds to stable modes (decay).
• Obstacle: The presence of unstable modes ($j=0, 1$) prevents direct application of standard stability estimates, as perturbations in these directions would grow or persist, obscuring the decay of the shape itself.

B. Dynamic Normalization (Modulation)

To overcome the instability, the authors introduce a dynamic normalization of the flow. They do not fix the coordinate system rigidly; instead, they dynamically adjust the scaling factor $\lambda(t)$ and the translation $p(t)$ to "mod out" the unstable modes.

• Constraints: They define two geometric functionals:
  1. Enclosed Area Functional ($U(t)$): Related to the integral of $(\sigma - 1)$.
  2. Center of Mass Functional ($q(t)$): Related to the integral of $(\sigma - 1)\cos\theta$.
• Strategy: They choose $\lambda(t)$ and $p(t)$ such that $U(t) \equiv 0$ and $q(t) \equiv 0$ for all time. This effectively forces the projection of the solution onto the unstable modes to be zero (or decaying rapidly), allowing the stability analysis to focus on the stable subspace.

C. Evolution Equations and Estimates

The authors derive precise evolution equations for the normalization parameters $\lambda$ and $p$ by differentiating the constraints $U(t)=0$ and $q(t)=0$.

• They establish that the deviation of $\lambda$ from the standard parabolic scaling $\Lambda(t) = (2(T-t))^{-1/2}$ and the translation $p$ are small ($O(\Lambda^{-1})$).
• They prove that under this normalization, the projections onto the unstable modes decay exponentially in the rescaled time $\tilde{t}$.

D. $L^2$ Decay and Interpolation

• Energy Estimate: They construct an $L^2$ energy functional for the perturbation $\tilde{\sigma} - 1$. By analyzing the linearized operator and using the fact that the unstable modes are suppressed by the normalization, they prove that the $L^2$ norm decays as $e^{-2\tilde{t}}$.
• Higher Order Norms: Using interpolation inequalities (Gagliardo-Nirenberg), they extend the $L^2$ decay to all $C^k$ norms, albeit with a slightly slower rate ($e^{-\alpha \tilde{t}}$ for any $\alpha < 1$).

3. Key Contributions

1. Sharp Convergence Rate: The paper establishes that the rescaled free-boundary CSF converges to the unit semicircle at a rate of $(T-t)^{1-\delta}$ (in original time) or $e^{-\alpha \tilde{t}}$ (in rescaled time) for any $\alpha < 1$. This is the first quantitative rate for this specific free-boundary problem.
2. Handling Boundary Instabilities: The authors successfully adapt the "modulation method" (previously used for closed hypersurfaces) to the free-boundary setting. This is non-trivial because the domain of the Gauss map changes with time, and the boundary conditions introduce additional geometric complexities compared to the closed case.
3. Dynamic Modulation of Symmetries: The explicit construction of the time-dependent scaling and translation to eliminate unstable modes provides a robust framework for analyzing singularities in geometric flows with boundaries.
4. Refined Asymptotics: The paper provides refined estimates for the support function $\sigma$ and curvature $\kappa$ of the original (unscaled) flow near the singularity.

4. Main Results

Theorem 1.1 (Main Result):
Let $\Omega \subset \mathbb{R}^2$ be a convex domain with $C^2$ boundary. Let $\{\Gamma_t\}$ be a maximal free-boundary CSF shrinking to a point $p \in \partial \Omega$ in finite time $T$.

• The rescaled curves $\tilde{\Gamma}_t$ (rotated and scaled by $(T-t)^{-1/2}$) converge to the unit semicircle in the upper half-plane.
• The convergence is uniform in the $C^k$-topology for every $k \ge 0$.
• The rate of convergence is no slower than $(T-t)^{1-\delta}$ for any $\delta > 0$.

Theorem 5.2 (Original Flow Asymptotics):
The support function $\sigma$ and curvature $\kappa$ of the original flow satisfy:
$$\sigma = \sqrt{2(T-t)} + O((T-t)^{\frac{1+\alpha}{2}})$$
$$\kappa = \frac{1}{\sqrt{2(T-t)}} + O((T-t)^{\frac{\alpha-1}{2}})$$
for any $\alpha < 1$.

5. Significance

• Completing the Picture: This work completes the qualitative picture of free-boundary CSF in convex domains, moving from "it converges to a semicircle" to "it converges to a semicircle exponentially fast (up to a logarithmic factor)."
• Analogy to Closed Curves: The result brings the free-boundary case to the same level of quantitative precision as the classical closed-curve case (Gage-Hamilton), confirming that the boundary does not fundamentally alter the stability of the singularity profile, provided the domain is convex.
• Foundation for Classification: Sharp asymptotic rates are crucial for uniqueness and classification problems. For instance, knowing the precise forward asymptotics allows one to classify ancient solutions or prove uniqueness of the singularity profile, similar to how Gage-Hamilton's rates were used to classify shrinking circles.
• Technical Innovation: The successful application of dynamic normalization to a problem with a time-varying domain (due to the moving boundary points in the Gauss map parametrization) offers a new toolkit for analyzing other geometric flows with free boundaries.

In summary, the paper resolves a long-standing open question regarding the quantitative stability of the shrinking semicircle in free-boundary curve shortening flow, proving that the flow approaches the singularity profile with a sharp, exponential-like rate after appropriate rescaling and normalization.