Nondegenerate neck pinches along the mean curvature flow

This paper establishes that for generic smooth compact initial surfaces, the mean curvature flow in $\mathbb{R}^3$ develops only isolated spherical or nondegenerate neck pinch singularities at the first singular time.

The Gist

Imagine you have a piece of elastic, soapy skin floating in space. If you let it shrink under its own surface tension (a process mathematicians call Mean Curvature Flow), it will eventually get smaller and smaller until it pops or pinches off.

For a long time, mathematicians wondered: What exactly happens right before it pops? Does it just vanish into a single point like a soap bubble? Or does it get weird and complicated, forming strange shapes or multiple pinches at once?

This paper by Gábor Székelyhidi answers that question with a very specific, reassuring result: If you start with a generic (randomly chosen) smooth surface, the only way it can pinch off is in a very simple, predictable way.

Here is the breakdown using everyday analogies:

1. The Two Ways to Pop

The paper confirms that there are essentially only two "safe" ways for these surfaces to develop a singularity (a point where the math breaks down):

• The Spherical Pinch: Imagine a balloon shrinking. It gets smaller and smaller until it vanishes into a single point. This is simple and well-understood.
• The Neck Pinch: Imagine a dumbbell or a figure-eight shape. The middle part (the "neck") gets thinner and thinner until it snaps.
  • The "Bad" Neck: Sometimes, the neck can get thin in a messy, unstable way. It might pinch off in a way that depends on tiny, random imperfections in the material. This is called a degenerate pinch. It's like a neck that is about to snap but is wobbling unpredictably.
  • The "Good" Neck: The paper proves that for almost any starting shape, the neck will snap in a non-degenerate way. This means the pinch is stable, isolated (happens at one specific spot and time), and looks exactly like a perfect cylinder snapping.

2. The Main Discovery: "Generic" Means "Simple"

The author's main point is that messy, complicated singularities are rare.

Think of it like balancing a pencil on its tip.

• If you try to balance a pencil perfectly, it will eventually fall.
• If you balance it exactly on the very edge of a razor, it might fall in a weird, unpredictable direction (degenerate).
• But if you just pick up a pencil and drop it (generic), it will fall in a very predictable, standard way.

Székelyhidi shows that if you take any smooth surface and make a tiny, almost invisible tweak to its shape (a "perturbation"), you can guarantee that when it eventually pinches, it will do so in that clean, "non-degenerate" way. The messy, complicated pinches are like the pencil balanced on the razor edge—they are so unstable that a tiny nudge makes them disappear.

3. How Did They Prove It? (The "Tuning" Analogy)

The proof is like tuning a radio to find a clear signal.

1. The Problem: The researchers knew that "degenerate" pinches (the messy ones) existed, but they wanted to show they could be "tuned out."
2. The Method: They imagined adding a tiny, specific "vibration" or "nudge" to the starting shape. They didn't just nudge it randomly; they nudged it in a specific direction (mathematically, along the axis of the cylinder).
3. The Result: They showed that if you nudge the shape just right, the "messy" pinch gets pushed away. It's like if you have a wobbly table leg, and you slide a tiny shim under it. The wobble (the degenerate singularity) disappears, and the table becomes stable.
4. The Covering Argument: Since there might be many places where the surface could pinch, they used a mathematical "sweeping" technique. They showed that if you pick a random set of nudges, you will almost certainly hit a combination that fixes all the potential messy spots at once.

4. Why Does This Matter?

In the world of physics and geometry, "singularities" are where our equations break down. If a singularity is messy and unpredictable, we can't trust our models to tell us what happens next.

By proving that non-degenerate (clean, isolated) singularities are the norm, this paper tells us:

• Predictability: We can trust that these flows behave in a standard way.
• Uniqueness: Because the pinches are clean and isolated, mathematicians can now define exactly how the flow continues after the pinch. It's like knowing exactly where a broken bridge fell so you can build a new one right next to it.

Summary

Imagine a complex, shrinking soap film. For decades, mathematicians worried it might get tangled and form a chaotic knot before popping. This paper says: "Don't worry." If you start with a normal, smooth shape, the film will never get tangled. It will either shrink to a dot or snap its neck in a clean, isolated spot. Any "messy" behavior is so fragile that a tiny change in the starting shape eliminates it completely.

The universe of these shrinking surfaces is much more orderly and predictable than we feared.

Technical Summary

Here is a detailed technical summary of the paper "Nondegenerate Neck Pinches Along the Mean Curvature Flow" by Gábor Székelyhidi.

1. Problem Statement and Context

The paper addresses the nature of singularities in the Mean Curvature Flow (MCF) of compact surfaces in $\mathbb{R}^3$. While MCF is well-understood for smooth initial data, it inevitably develops singularities in finite time.

• The Conjecture: A major conjecture by Ilmanen (and related works by Colding-Minicozzi) posits that for generic initial data, the only singularities occurring are spherical (modeled on shrinking spheres) and cylindrical (modeled on shrinking cylinders $C = \mathbb{R} \times S^1$).
• Recent Progress: Chodosh-Choi-Mantoulidis-Schulze recently resolved the genericity of spherical and cylindrical singularities (ruling out other types). However, cylindrical singularities can be degenerate, meaning they are not isolated in spacetime (e.g., a "marriage ring" torus collapsing to a circle).
• The Specific Problem: The paper aims to prove that for generic initial data, even the cylindrical singularities are non-degenerate and isolated in spacetime. Specifically, it seeks to show that one can perturb any smooth compact initial surface so that the resulting flow encounters only spherical and non-degenerate cylindrical singularities at the first singular time.

2. Methodology and Technical Approach

The author employs a perturbation strategy combined with a detailed analysis of the rescaled mean curvature flow near singularities. The methodology relies on several advanced tools from geometric analysis:

A. Rescaled Flow and Linearization

The analysis focuses on the flow near a singularity $(X, T)$ by rescaling:
$$M_\tau = e^{\tau/2} L_{T - e^{-\tau}}$$
where $\tau \to \infty$ corresponds to approaching the singularity.

• Cylindrical Singularity: The rescaled flow $M_\tau$ converges to the cylinder $C = \mathbb{R} \times S^1(\sqrt{2})$.
• Degenerate vs. Non-degenerate:
  • Non-degenerate: The convergence rate is governed by the eigenfunction $y^2 - 2$ (eigenvalue $\lambda = 0$ of the linearized operator), leading to a specific asymptotic expansion.
  • Degenerate: The convergence is faster, dominated by eigenfunctions with eigenvalues $\lambda \geq 1/2$ (e.g., $y$, $z_1$, $z_2$), causing the singularity to be non-isolated.

B. Perturbation via Initial Data

The core strategy involves perturbing the initial surface $S_0$ to eliminate degenerate singularities.

• Perturbation Function: The author constructs perturbations of the form $a \chi_{R_0} y$, where $a$ is a small parameter, $\chi_{R_0}$ is a cutoff function, and $y$ is the coordinate along the axis of the cylinder.
• Density Argument: Using a result by Proposition 13 (based on linearized MCF), the author shows that the image of the linearized flow operator is dense in $W^{1,p}$. This allows the construction of initial perturbations that approximate the desired function $a \chi_{R_0} y$.

C. Growth Estimates and Three-Annulus Lemma

To prove that a perturbation eliminates a degenerate singularity, the author must show that the perturbation grows fast enough to dominate the original flow's behavior before the singularity forms.

• Linearized Growth: The perturbation $v(\tau)$ is expected to grow like $e^{\tau/2}$ (associated with the eigenvalue $\lambda=1/2$).
• Three-Annulus Lemma (Proposition 9): A crucial technical tool is a "three-annulus" type lemma. It states that if a solution to the linearized equation grows at a certain rate on an inner annulus and satisfies specific bounds on an outer annulus, this growth rate persists. This ensures the perturbation does not decay or stagnate.
• Global Barriers: To control the flow globally and prevent the perturbation from causing new singularities elsewhere, the author constructs barrier flows (Proposition 12). These barriers utilize the mean convexity of the flow near cylindrical singularities (a result by Choi-Haslhofer-Hershkovits) to sandwich the perturbed flow between time-translated versions of the original flow.

D. The "Covering" Argument

The proof concludes with a measure-theoretic argument (Proposition 18):

1. If a flow $L^a_t$ has a degenerate singularity at a specific $y$-coordinate, a nearby flow $L^{a'}_t$ cannot have a degenerate singularity at a $y'$-coordinate if $|y - y'| < |a - a'|^{2/3}$.
2. This implies that the set of parameters $a$ leading to degenerate singularities in a fixed neighborhood has measure zero.
3. By an inductive covering argument over the finite set of singular points, one can find a dense set of perturbations that remove all degenerate singularities.

3. Key Contributions

1. Verification of Ilmanen's Conjecture (Partial): The paper proves that for generic initial data, the first singular time consists only of spherical and non-degenerate cylindrical singularities.
2. Isolation of Singularities: It establishes that non-degenerate cylindrical singularities are isolated in spacetime. This is a significant improvement over the general case where cylindrical singularities can form curves.
3. Stability: The result implies that non-degenerate singularities are stable under small perturbations, allowing the flow to be uniquely defined through the singularity (continuing past the first singular time).
4. Technical Innovation: The paper introduces a refined perturbation scheme using the specific eigenfunction $y$ (associated with translations along the cylinder axis) combined with a global barrier argument and a three-annulus lemma to control the growth of perturbations over long time intervals in the rescaled flow.

4. Main Results

Theorem 1: Let $S_0 \subset \mathbb{R}^3$ be a smooth compact surface. There exist arbitrarily small $C^2$ perturbations $\tilde{S}_0$ of $S_0$ such that the mean curvature flow $\tilde{S}_t$ admits only spherical and non-degenerate cylindrical singularities at its first singular time.

Corollary: Consequently, the singularities at the first singular time are isolated in spacetime.

5. Significance and Implications

• Uniqueness of Flow: The isolation of singularities is a prerequisite for defining a unique continuation of the mean curvature flow past singularities. This supports the program of defining "canonical" flows through singularities.
• Genericity: The result strengthens the understanding of the "generic" behavior of MCF, suggesting that the most complex singularities (degenerate neck pinches) are non-generic and can be avoided by slight adjustments to the initial shape.
• Future Directions: The author notes that while the result holds up to the first singular time, extending it to all time requires a deeper understanding of the flow's behavior after passing through a non-degenerate cylindrical singularity. The techniques also hold promise for higher-dimensional mean convex flows, though new difficulties regarding degenerate directions in higher dimensions remain.

In summary, Székelyhidi provides a rigorous proof that "bad" (degenerate) neck pinches are not generic, establishing that the mean curvature flow of a generic surface in $\mathbb{R}^3$ encounters only well-behaved, isolated singularities.