Boundary epsilon regularity for incompressible Navier--Stokes equations via weak-strong uniqueness

This paper establishes the boundary $\epsilon$-regularity for finite-energy weak solutions to the incompressible Navier–Stokes equations on a three-dimensional bounded smooth domain by proving that solutions are regular up to the boundary whenever their $L^4_tL^4_x$-norm is sufficiently small, thereby resolving a problem posed by Albritton, Barker, and Prange through a novel slicing construction near the boundary.

The Gist

Imagine a pot of thick, swirling soup (representing a fluid like water or air) moving inside a smooth, round bowl. Mathematicians have long been trying to predict exactly how this soup will move. The equations governing this motion are called the Navier-Stokes equations.

For decades, mathematicians knew that if you look at the soup deep inside the pot (away from the walls), you can usually predict its smooth flow, provided the soup isn't swirling too wildly. This is called "interior regularity." However, a big mystery remained: What happens right at the edge, where the soup touches the bowl? Could the soup suddenly develop a chaotic, infinite-speed whirlpool right against the wall?

This paper, by Siran Li, solves that mystery. It proves that if the soup isn't swirling too wildly overall, it will remain smooth and predictable all the way up to the very edge of the bowl.

Here is how the author cracked the code, using some creative mental tricks:

1. The Old Problem: The "Slicing" Trap

To prove the soup is smooth, the author uses a method called "slicing." Imagine taking a loaf of bread and slicing it into thin pieces to check the texture inside.

• The Interior Trick: In the middle of the pot, you can slice the soup using perfect spheres (like cutting an orange). If the soup is calm inside a small sphere, you know it's calm everywhere inside that sphere.
• The Wall Problem: When you get to the wall of the bowl, you can't just use spheres. If you slice a sphere against a flat wall, you get a hemisphere. The problem is that the "crust" of the soup (the part touching the wall) might look messy even if the inside is calm. The old slicing method failed here because the math couldn't guarantee the "crust" was calm enough to prove the inside was safe.

2. The New Trick: The "Clam" Shell

The author's breakthrough was inventing a new shape for slicing, which the paper calls a "clam."

Instead of slicing with spheres, imagine a smooth, convex shell that looks like a clam or a seashell.

• The Shape: This shell is shaped like a bowl inside a bowl. The bottom of the shell is a curved parabola (like a satellite dish), and the top is a rounded cap.
• The Magic Touch: The author designs these shells so that they touch the wall of the main bowl at exactly one single point, and they touch it very gently (mathematically, they are "tangent").
• Why it works: Because the shell touches the wall so gently at just one point, the "messy crust" of the soup on the wall is minimized. By shrinking these clam-shells down toward the wall, the author creates a series of layers.

3. The "Pigeonhole" Principle

Now, imagine you have a huge amount of data about how the soup is moving. You can't check every single point.

• The author uses a logic trick called the Pigeonhole Principle. Think of it like this: If you have a lot of pigeons (energy in the soup) and a limited number of holes (the layers of your clam-shells), at least one hole must be relatively empty.
• The author proves that among all these "clam" layers, there must be at least one specific layer where the soup is very calm and quiet.

4. The "Weak-Strong" Handshake

Once the author finds that one calm "clam" layer, they use a technique called Weak-Strong Uniqueness.

• Think of this as a handshake between two versions of the soup:
  1. The Real Soup: The actual, messy fluid we are studying.
  2. The Ideal Soup: A perfectly smooth, mathematical version of the fluid that we know how to calculate.
• The author shows that because the "Real Soup" is calm enough on that specific clam-layer, it is forced to behave exactly like the "Ideal Soup."
• Since the "Ideal Soup" is smooth and has no explosions or infinite speeds, the "Real Soup" must be smooth too.

The Conclusion

By using these "clam" slices to get right up to the wall, and then proving that the fluid must behave like a smooth, ideal fluid in that region, the author proves that the soup cannot suddenly go crazy at the edge.

If the overall energy of the fluid is kept below a certain limit, the fluid will remain smooth and predictable everywhere, from the very center of the pot to the very rim of the bowl. This answers a question that had been open for years, confirming that the "edge" of the fluid is just as safe as the "middle."

Technical Summary

1. Problem Statement

The paper addresses a fundamental open problem in the regularity theory of the incompressible Navier–Stokes equations (NSE) in three dimensions: establishing an $\epsilon$-regularity theorem up to the boundary.

• Context: While interior regularity is well-understood (e.g., Caffarelli–Kohn–Nirenberg, Escauriaza–Seregin–Šverák), boundary regularity remains challenging.
• Specific Gap: Recent work by Albritton, Barker, and Prange (2023) established a new proof of interior $\epsilon$-regularity using weak-strong uniqueness arguments and the theory of very weak solutions (pioneered by Foias, Amann, Farwig, Galdi, and Sohr). However, they noted that their "spatial-temporal slicing" technique could not be directly applied near the boundary.
• The Obstacle:
  1. Trace Incompatibility: A finite-energy weak solution $U$ has a trace in $L^2_t L^4_x$ on the boundary, but the linear theory for very weak solutions (Farwig et al.) requires boundary data in $L^4_t L^4_x$ (or similar) to guarantee uniqueness and regularity. Smallness in the bulk $L^4_t L^4_x$ norm does not automatically imply smallness of the trace.
  2. Slicing Failure: Standard slicing over concentric spheres fails near the boundary because one cannot guarantee small boundary data on hemispheres centered at a boundary point without violating the trace regularity constraints.

Goal: Prove that if a finite-energy weak solution has a sufficiently small $L^4_t L^4_x$ norm in a space-time cylinder near the boundary, the solution is smooth (regular) up to that boundary.

2. Methodology

The author employs a perturbative approach based on weak-strong uniqueness, but introduces a novel geometric construction to overcome the boundary difficulties.

A. The Core Strategy: Weak-Strong Uniqueness

The proof treats the Navier–Stokes solution $U$ as a perturbation of the Stokes problem.

1. Linearization: View the NSE as a Stokes equation with a forcing term $F = U \otimes U$.
2. Very Weak Solutions: Utilize the linear theory of very weak solutions for the Stokes problem with $L^4_t L^4_x$ boundary data.
3. Fixed Point: Use the smallness of the $L^4_t L^4_x$ norm to set up a contraction mapping (fixed point argument) in the space $L^4_t L^6_x$.
4. Uniqueness: Show that the constructed strong solution coincides with the original weak solution $U$ via a weak-strong uniqueness argument (energy estimates).
5. Bootstrap: Once $U$ is shown to be in $L^4_t L^6_x$, apply the Ladyženskaja–Prodi–Serrin (LPS) criterion to bootstrap regularity to $L^\infty_t L^\infty_x$ (and thus $C^\infty$).

B. The Novel Contribution: "Slicing by Clams"

The critical innovation is Proposition 8, a new spatial slicing construction designed to handle the boundary trace issue.

• The Construction: Instead of concentric spheres, the author constructs a smooth, convex body $V_0$ (resembling a "clam") in the upper half-space.
  • $V_0$ is foliated by surfaces $\Sigma_s$ ($s \in [0, 1)$).
  • Each $\Sigma_s$ is convex, rotationally symmetric, and intersects the boundary $\partial \mathbb{R}^3_+$ at a single point $x_\star$ with $C^1$-tangency.
  • The lower part of $\Sigma_s$ is a paraboloid tangent to the boundary; the upper part is a spherical cap.
• The Logic:
  1. The author straightens the boundary of the domain $\Omega$ near a point $x_\star$ using a diffeomorphism.
  2. They apply the Pigeonhole Principle over the foliation $\{\Sigma_s\}$. Since the total $L^4_t L^4_x$ norm is small, there must exist a specific slice $s$ where the trace of $U$ on $\Sigma_s$ is small in $L^4_t L^4_x$.
  3. Crucially, because $\Sigma_s$ is tangent to the boundary at a single point and convex, the region enclosed by $\Sigma_s$ (denoted $R$) has a boundary $\partial R$ that intersects $\partial \Omega$ only at $x_\star$.
  4. This allows the definition of boundary data for the Stokes problem on $\partial R$ that is small in the required $L^4_t L^4_x$ norm, satisfying the hypotheses of the linear theory (Proposition 5).

3. Key Contributions

1. Resolution of an Open Problem: The paper answers the specific question raised by Albritton, Barker, and Prange regarding the feasibility of boundary $\epsilon$-regularity via weak-strong uniqueness.
2. Geometric Innovation: The "clam" slicing technique (Proposition 8) provides a robust method to generate sub-domains with small boundary traces, bypassing the limitations of standard spherical slicing near boundaries.
3. Pressure-Free Regularity: Unlike classical $\epsilon$-regularity theorems (e.g., Caffarelli–Kohn–Nirenberg) which often require smallness conditions on the pressure or specific "suitable weak solution" criteria, this result relies only on the smallness of the velocity field in $L^4_t L^4_x$. The pressure control is implicitly handled via the Stokes semigroup estimates in the very weak solution framework.
4. Regularity Class: The result establishes that finite-energy weak solutions are smooth up to the boundary provided the $L^4_t L^4_x$ norm is below a universal threshold $\epsilon_0$.

4. Main Results

Theorem 3 (Boundary Epsilon Regularity):
Let $\Omega \subset \mathbb{R}^3$ be a smooth bounded domain. There exists a constant $\epsilon_0 > 0$ (depending only on the $C^2$-geometry of $\Omega$) such that:
If $U$ is a finite-energy weak solution to the NSE in $]-1, 0[ \times \Omega$ with $\|U\|_{L^4_t L^4_x} < \epsilon_0$, then $U$ is smooth ($C^\infty$) up to the boundary $]-1, 0] \times \Omega$.
Furthermore, the $C^\infty$ norm of $U$ is bounded by a polynomial of $M$ (the energy bound) and $\epsilon_0$.

Technical Steps in the Proof:

1. Boundary Straightening: Map a neighborhood of a boundary point to the upper half-space.
2. Spatial Slicing: Use the "clam" foliation to find a sub-domain $R$ where the boundary trace of $U$ is small.
3. Temporal Slicing: Select a time $t_0$ where the initial data is small.
4. Data Correction: Use Bogovskiĭ operators and heat kernels to construct divergence-free extensions of the boundary and initial data.
5. Linear Estimates: Apply Farwig–Galdi–Sohr theory to solve the linear Stokes problem with these small data, obtaining a solution in $L^4_t L^6_x$.
6. Fixed Point & Uniqueness: Construct the NSE solution via a fixed point argument and prove it coincides with the original weak solution using energy estimates.
7. Bootstrap: Use the LPS criterion to upgrade regularity from $L^4_t L^6_x$ to $L^\infty_t L^\infty_x$.

5. Significance

• Theoretical Advancement: It bridges the gap between the modern theory of very weak solutions and classical boundary regularity, demonstrating that weak-strong uniqueness is a powerful tool for boundary problems, not just interior ones.
• Methodological Impact: The "clam" slicing technique offers a new geometric tool for analyzing PDEs on domains with boundaries, potentially applicable to other problems where trace regularity is a bottleneck.
• Robustness: The result holds for general smooth bounded domains and does not require the solution to be a "suitable weak solution" (which imposes extra entropy inequalities), making it applicable to the standard class of Leray-Hopf weak solutions.
• Future Directions: The author notes that the regularity requirement on the domain ($C^{2,1}$) could potentially be lowered, and the technique might be adaptable to unbounded domains with compact boundaries.

In summary, Siran Li successfully extends the weak-strong uniqueness framework to the boundary of the Navier–Stokes equations by ingeniously constructing a geometric foliation that circumvents the trace regularity limitations, thereby proving that small $L^4$ energy implies smoothness up to the boundary.