$L^p$-Sobolev inequalities on minimal submanifolds

This paper establishes Allard-Michael-Simon-type $L^p$-Sobolev inequalities with explicit, asymptotically sharp, and codimension-free constants for Euclidean minimal submanifolds of arbitrary codimension by utilizing optimal mass transport theory, while also providing a unified proof of recent isoperimetric inequalities by Brendle and Eichmair.

The Gist

Imagine you are a cartographer trying to draw a map of a very strange, crumpled piece of paper floating in a giant, empty room. This paper isn't flat; it's a minimal submanifold. In plain English, think of it as a soap film stretched between two wire frames. It's a surface that naturally tries to minimize its area, just like a soap bubble tries to be as small as possible.

The authors of this paper, Zoltán Balogh, Alexandru Kristály, and Ágnes Mester, are mathematicians who study how functions (like temperature or pressure) behave on these crumpled, floating surfaces.

The Big Problem: The "Roughness" Rule

In the flat, empty room (mathematicians call this "Euclidean space"), there is a famous rule called the Sobolev Inequality. You can think of this rule as a "Roughness Law."

It says: If you want to know how much a function varies (how "rough" or "spiky" it is) across the whole surface, you can predict it by looking at how fast it changes at every single point.

In a flat room, this rule is perfect. We know the exact "price tag" (a constant number) for this relationship. But when you move from a flat floor to a crumpled, floating soap film, things get messy.

1. The Shape Matters: The film might be twisted, folded, or exist in a higher dimension than we can see.
2. The "Codimension" Issue: The paper is floating in a room that is much bigger than the film itself. If the film is 2D but the room is 10D, that's a huge gap. Previous rules for these films had a "price tag" that got worse and worse as the room got bigger. It was like paying a higher tax just because the room was larger, even if the film itself was tiny.

The Goal: A Fair Price Tag

The authors wanted to find a new, better rule for these floating films. They wanted a "price tag" (a constant) that:

• Works for any size of the room (codimension-free).
• Is as close to perfect as possible (asymptotically sharp).
• Doesn't depend on how twisted the film is, as long as the film is "minimal" (like a soap film with no extra weight pulling it down).

The Magic Tool: Optimal Mass Transport (OMT)

To solve this, they used a tool called Optimal Mass Transport.

The Analogy: Imagine you have a pile of sand (your function) sitting on your crumpled soap film. You want to move this sand to a perfectly smooth, round target shape in the big room.

• The Old Way: You might just guess how to move the sand, but you'd waste energy and end up with a messy result.
• The OMT Way: You calculate the most efficient path for every grain of sand to get to the target. You want to move the sand with the least amount of "effort" (energy).

The authors used a sophisticated version of this "sand-moving" logic. They treated the crumpled film and the smooth target as two different worlds and found the perfect "bridge" (a mathematical map) to connect them.

The Two Different Strategies

The paper splits the problem into two scenarios, like having two different keys for two different locks:

1. The "Steep" Case ($p \ge 2$):
When the function changes very sharply (like a steep mountain), the authors found a way to move the sand that ignores the size of the room entirely.

• The Result: They found a "price tag" that is codimension-free. It doesn't matter if the room is 3D or 1000D; the rule stays the same.
• The Catch: It's not perfectly perfect, but as the dimensions get huge, it becomes almost perfect. It's like a discount coupon that gets better the more you buy.

2. The "Gentle" Case ($1 < p < 2$):
When the function changes gently (like a rolling hill), the math gets trickier. The "room size" still matters a little bit here.

• The Result: They found a new rule that is still better than the old ones, even if it's not completely independent of the room size. It's a "better deal" than what we had before.

The "Soap Film" Surprise

The authors also used this same "sand-moving" bridge to prove an old, famous rule about the Isoperimetric Inequality (which relates the area of a shape to its boundary).

• The Old Proof: Required the soap film to be a closed, finite loop (compact).
• The New Proof: Their method works even if the film is infinite or has edges! They removed the "compactness" requirement, making the rule much more powerful and universal.

Why Does This Matter?

Think of this paper as upgrading the GPS for navigating complex shapes.

• Before: The GPS said, "To get from A to B, you need to pay a fee that depends on how big the universe is."
• Now: The GPS says, "No matter how big the universe is, here is the most efficient, fair route, and here is the exact cost."

This helps physicists and engineers who model things like:

• Black Holes: Which are essentially minimal surfaces in 4D space-time.
• Materials Science: Understanding how thin films and membranes behave.
• Computer Graphics: Creating realistic simulations of cloth or water.

In short, the authors took a messy, complicated problem involving crumpled surfaces in high-dimensional spaces and used the elegant logic of "moving sand efficiently" to find a cleaner, fairer, and more universal set of rules.

Technical Summary

Here is a detailed technical summary of the paper "Lp-Sobolev Inequalities on Minimal Submanifolds" by Zoltán M. Balogh, Alexandru Kristály, and Ágnes Mester.

1. Problem Statement and Context

The paper addresses the problem of establishing $L^p$-Sobolev inequalities for functions defined on minimal submanifolds of Euclidean space $\mathbb{R}^{n+m}$ with arbitrary codimension $m$.

• Background: Classical Sobolev inequalities in $\mathbb{R}^n$ relate the $L^{p^*}$ norm of a function to the $L^p$ norm of its gradient. On submanifolds, the presence of curvature (specifically the mean curvature vector $H$) typically introduces additional terms in the inequality.
• Existing Results:
  • Michael-Simon (1973): Established a general inequality involving the mean curvature, but with a constant ($4^{n+1}\omega_n^{-1/n}$) that is far from optimal and depends heavily on dimension.
  • Brendle (2021) & Brendle-Eichmair (2024): Proved sharp isoperimetric/Sobolev inequalities for compact submanifolds. Their constants are optimal for codimensions $m=1, 2$ but depend on $m$ for $m \geq 3$, and they formally blow up as $m \to \infty$.
• The Gap: There was no known codimension-free $L^p$-Sobolev inequality with explicit, asymptotically sharp constants for minimal submanifolds ($H \equiv 0$) when $p > 1$ and $m$ is large. The authors aim to close this gap, specifically distinguishing between the cases $p \geq 2$ and $1 < p < 2$.

2. Methodology: Optimal Mass Transport (OMT)

The core of the proof relies on Optimal Mass Transport (OMT) theory adapted to Euclidean submanifolds.

• Key Tool: The authors utilize a generalized version of Brenier's Theorem and the Monge-Ampère equation developed by Balogh and Kristály [4]. This version handles measures that are not necessarily absolutely continuous with respect to the ambient Lebesgue measure (crucial since the source measure is supported on the submanifold $\Sigma$).
• The Map: They construct a transport map $\Phi: A \to \mathbb{R}^{n+m}$ defined on a subset of the normal bundle $T^\perp \Sigma$:
  $$\Phi(x, v) = \nabla_\Sigma u(x) + v$$
  where $u$ is a semiconvex potential function on $\Sigma$.
• Key Properties Used:
  1. Pythagorean Rule: $|\Phi(x, v)|^2 = |\nabla_\Sigma u(x)|^2 + |v|^2$.
  2. Determinant-Trace Inequality: For minimal submanifolds ($H=0$), the Jacobian determinant of the transport map satisfies:
     $$\det D\Phi(x, v) \leq \left( \frac{\Delta_\Sigma^{ac} u(x)}{n} \right)^n$$
     where $\Delta_\Sigma^{ac}$ is the absolutely continuous part of the distributional Laplacian.
  3. Integral Monge-Ampère Equation: Relates the source density (function $f$ on $\Sigma$) to the target density (a specific "Talenti-type bubble" in $\mathbb{R}^{n+m}$).

3. Key Contributions and Results

The paper provides two main theorems depending on the exponent $p$, along with a unified proof for the isoperimetric case ($p=1$).

A. The Case $p \geq 2$ (Theorem 1.1)

• Result: For a complete minimal submanifold $\Sigma \subset \mathbb{R}^{n+m}$ ($n \geq 3$), and $f \in C_0^\infty(\Sigma)$:
  $$\left( \int_\Sigma |f|^{p^*} d\text{vol}_\Sigma \right)^{1/p^*} \leq S(n, p) \left( \int_\Sigma |\nabla_\Sigma f|^p d\text{vol}_\Sigma \right)^{1/p}$$
• The Constant $S(n, p)$:
  $$S(n, p) = \frac{p^*}{n} \left( 1 - \frac{1}{n} \right)^{\frac{p-1}{p}} (2\pi)^{-1/2} \left( \frac{e}{n} \right)^{\frac{1}{p'} - \frac{1}{2}} \left( \frac{\Gamma(n)}{\Gamma(n/p)} \right)^{1/n}$$
• Significance:
  • Codimension-Free: The constant $S(n, p)$ does not depend on the codimension $m$.
  • Asymptotically Sharp: As $n \to \infty$, the ratio $S(n, p) / A_T(n, p) \to 1$, where $A_T(n, p)$ is the optimal Euclidean Sobolev constant (Aubin-Talenti).
  • Comparison: This constant is significantly better than the Brendle constant for $m \geq 3$, which grows with $m$.

B. The Case $1 < p < 2$ (Theorem 1.2)

• Result: A similar inequality holds, but the constant $\tilde{S}(n, m, p)$ depends on the codimension $m$.
  $$\left( \int_\Sigma |f|^{p^*} d\text{vol}_\Sigma \right)^{1/p^*} \leq \tilde{S}(n, m, p) \left( \int_\Sigma |\nabla_\Sigma f|^p d\text{vol}_\Sigma \right)^{1/p}$$
• Improvement: While not codimension-free, this constant is strictly better than previous results (Brendle's and Michael-Simon's) for certain ranges of $m$ and $p$.
• Consistency: When $p=2$, the constants in Theorem 1.1 and 1.2 coincide, recovering the result in Corollary 1.1.

C. Unified Proof of Isoperimetric Inequality (Theorem 3.1)

• The authors provide an alternative proof of the Brendle-Eichmair isoperimetric inequality (the $p=1$ limit case) using OMT.
• Novelty: Unlike previous OMT proofs by Brendle/Eichmair which required the submanifold to be compact, this proof works for complete (potentially non-compact) submanifolds by using cut-off functions and careful limit arguments.
• Sharpness: The constant is proven to be sharp for $m=1$ and $m=2$.

4. Technical Obstacles and Solutions

• The "Blow-up" Problem: In standard OMT approaches for submanifolds, the target density is defined on $\mathbb{R}^{n+m}$. Integrating over the normal fibers introduces terms that depend on $m$.
• Solution for $p \geq 2$: The authors exploit the concavity of the function $|\cdot|^{p'/2}$ (since $p' \leq 2$) combined with the Pythagorean rule. This allows them to separate the tangential and normal components in the integral estimate in a way that permits taking the limit $m \to \infty$, effectively eliminating the codimension dependence in the final constant.
• Limitation for $p < 2$: When $p < 2$ (so $p' > 2$), the convexity/concavity arguments change, and the separation of variables does not yield a codimension-free constant. The authors derive the best possible bound given the current OMT framework, which retains $m$-dependence.
• Sharpness Limitation: The authors note that they cannot recover the exact Aubin-Talenti constant $A_T(n, p)$ because the integral term $J$ (arising from the transport map) can only be estimated rather than computed exactly in the submanifold setting (unlike the pure Euclidean case where $m=0$). This results in the constant being "asymptotically sharp" rather than "sharp" for finite $n$.

5. Significance

1. Resolution of a Long-standing Question: The paper answers whether codimension-free Sobolev inequalities exist for minimal submanifolds with $p > 1$. The answer is yes for $p \geq 2$.
2. Unification: It unifies the treatment of Sobolev inequalities across different codimensions, showing that for high codimensions, the geometry of the minimal submanifold behaves asymptotically like the Euclidean space regarding Sobolev constants.
3. Methodological Advancement: It demonstrates the power of Optimal Mass Transport on submanifolds, extending techniques previously limited to compact or Euclidean settings to complete non-compact minimal submanifolds.
4. Applications: These inequalities are fundamental for the analysis of partial differential equations (PDEs) on manifolds, providing crucial a priori estimates for solutions to geometric evolution equations and variational problems.

In summary, this paper establishes that for minimal submanifolds, the "cost" of the Sobolev inequality (the constant) does not necessarily increase with the codimension of the embedding, provided $p \geq 2$, and provides the explicit, asymptotically optimal constants for this regime.