🔢 mathematics

Quadratic flatness and Regularity for Codimension-One Varifolds with Bounded Anisotropic First Variation Part II

This paper extends Brakke's perpendicularity theorem to the anisotropic setting and establishes the locality of the anisotropic mean curvature vector as fundamental consequences of previously proven quadratic flatness and $\mathscr{C}^{2}$ -rectifiability results for codimension-one varifolds with bounded anisotropic mean curvature.

Original authors: Sławomir Kolasiński, Mario Santilli

Published 2026-03-24

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Sławomir Kolasiński, Mario Santilli

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: The Shape of a Wobbly Sheet

Imagine you are looking at a giant, invisible sheet of fabric floating in space. This sheet isn't necessarily smooth like a table; it might be crumpled, folded, or have weird kinks. In mathematics, we call this a varifold. It's a way of describing a shape that might be messy, but still has a "surface" to it.

The authors are studying how this sheet behaves when it tries to minimize its energy. Think of it like a soap bubble trying to shrink to the smallest size possible. Usually, soap bubbles want to be perfect spheres (isotropic). But in this paper, the "soap" is special: it has anisotropy.

The Anisotropy Analogy:
Imagine the fabric is made of a material that stretches easily in one direction but is very stiff in another (like a piece of wood grain or a woven basket). Because of this, the "perfect shape" for this material isn't a sphere; it's a weird, stretched-out blob. The math behind this is the anisotropic mean curvature.

The paper asks a simple question: If this weird, stiff fabric is trying to find its perfect shape, does it actually look like a smooth, curved surface at the microscopic level?

The Two Main Discoveries

The authors prove two major things about this fabric:

1. The "Perpendicularity" Rule (The Compass Needle)

In the world of standard soap bubbles (Euclidean geometry), the force that pulls the bubble tight (the mean curvature) always points straight out, perpendicular to the surface. It's like a compass needle pointing directly away from the ground.

The Problem: When the fabric is "stiff" (anisotropic), the usual math tools break. The "Pythagorean theorem" (the rule that $a^2 + b^2 = c^2$ ) doesn't work the same way, so the compass needle might point in the wrong direction.
The Solution: The authors prove that even for this stiff, weird fabric, the force still points straight out (perpendicular) to the surface, but only at the points where the fabric is "single-layered" (where the density is 1).
The Trick: To prove this, they used a clever mental gymnastics move. They imagined taking a tiny, crumpled piece of the fabric and running it through a "straightening machine" (a linear map). This machine temporarily turns the stiff fabric into a normal, stretchy one just for a moment. They proved the rule holds for the normal fabric, and then reversed the machine to show it holds for the stiff one too.

2. The "Locality" Rule (The Neighborhood Watch)

This is about how the shape at one specific point depends on its neighbors.

The Concept: If you look at a tiny speck on the fabric, does the "curvature" (how much it bends) at that speck depend on what's happening miles away? Or is it determined only by the immediate neighborhood?
The Result: They proved that the curvature is local. If you have two different pieces of fabric that look exactly the same in a small neighborhood around a point, they will have the exact same curvature at that point. It's like saying the temperature at your front door depends only on the weather right outside your house, not on the weather in a different country.

Why Was This Hard? (The "Bad Set" Problem)

In the past, mathematicians could prove these things for normal soap bubbles because they had a "magic formula" (monotonicity) that guaranteed the fabric would smooth out nicely.

However, for this "stiff" fabric, that magic formula doesn't exist. The fabric can be messy in a way that standard math tools can't handle. The authors had to rely on a previous result (from their Part I paper) which showed that if the fabric is "single-layered" (density 1), it is actually very smooth (a $C^{1,\alpha}$ manifold).

They used this smoothness as a foundation. Because they knew the fabric was smooth in these specific areas, they could finally prove the perpendicularity and locality rules.

The "Straightening" Metaphor

The most creative part of the paper is how they handled the anisotropy (the stiffness).

Imagine you are trying to walk on a floor that is tilted and slippery in a weird pattern. It's hard to tell which way is "up."

The Problem: You can't use your standard compass because the floor is warped.
The Solution: The authors imagine a magical elevator. You step into the elevator at a specific point on the floor. The elevator instantly rotates and stretches the floor so that, just for that one spot, the floor becomes perfectly flat and level.
The Proof: Now that the floor is flat, you can easily prove your compass points North.
The Return: You step out of the elevator. The floor is warped again, but because you proved the rule works in the "flat" version, and the elevator transformation is reversible, the rule must also work in the "warped" version.

Summary for the General Audience

This paper is a victory for understanding shapes that aren't perfectly round or smooth.

Before: We knew these shapes were messy, and we couldn't be sure if the forces acting on them behaved predictably.
Now: We know that if the shape is "single-layered" (not double or triple stacked), it behaves beautifully. The forces point straight out, and the shape at any point is determined only by its immediate neighbors.
The Method: They solved the problem by temporarily "flattening" the weird geometry to use standard math, then proving the results hold true even when the geometry gets weird again.

This is a significant step forward in Geometric Measure Theory, helping us understand how complex, irregular surfaces (like cell membranes, crystal structures, or even the boundaries of black holes in certain models) behave when they are under tension.

1. Problem Statement and Context

This paper serves as a sequel to the authors' previous work ([KS25]), which established quadratic flatness and $C^2$ -rectifiability for the support of codimension-one varifolds ( $V$ ) in $\mathbb{R}^{n+1}$ with bounded anisotropic mean curvature, under the hypothesis that the support has locally finite $\mathcal{H}^n$ measure.

The central problem addressed in this paper is to explore the fundamental geometric consequences of that regularity result, specifically focusing on two critical properties of the anisotropic mean curvature vector ( $h_F(V, \cdot)$ ):

Perpendicularity: Does the anisotropic mean curvature vector remain orthogonal to the approximate tangent plane of the varifold?
Locality: Is the anisotropic mean curvature vector compatible with the approximate second-order structure of the varifold (i.e., does it coincide with the mean curvature of a smooth hypersurface where the varifold is locally represented by one)?

In the Euclidean setting (isotropic case), Brakke proved perpendicularity for integral varifolds with bounded first variation. However, extending this to the anisotropic setting is non-trivial because standard Euclidean tools, such as the Pythagorean theorem and monotonicity formulas, fail or are unavailable.

2. Methodology and Key Technical Innovations

The authors employ a sophisticated combination of geometric measure theory, variational calculus, and linear algebraic transformations to overcome the lack of Euclidean structure.

A. Linear Straightening of the Anisotropic Structure

The core innovation is a technique to "locally straighten" the anisotropic geometry.

The Challenge: In the anisotropic setting, the projection operators associated with the integrand ( $P_F$ and $Q_F$ ) do not behave orthogonally in the Euclidean sense, making standard estimates (like those used by Brakke) invalid.
The Solution: For a point $b$ with approximate tangent plane $T$ , the authors introduce a linear map $\varphi_T$ that depends on $T$ . They push forward the varifold $V$ via $\psi = \varphi_T^{-1}$ and pull back the integrand $F$ to a new integrand $G = \varphi_T^\# F$ .
Effect: This transformation aligns the anisotropic structure with the Euclidean structure at the specific point $b$ . Specifically, for the transformed integrand $G$ , the projection $P_G(T)$ becomes the standard orthogonal projection $T^\natural$ . This allows the authors to apply Euclidean-style estimates locally while retaining the global anisotropic context.

B. Analysis of Push-Forwards and First Variation

The paper rigorously analyzes how the first variation behaves under diffeomorphisms (Section 4).

They establish that if $V$ has bounded anisotropic first variation, the push-forward $W = \psi_\# V$ also has bounded first variation with respect to the pulled-back integrand $G$ .
Crucially, they prove that the Lebesgue points of the mean curvature vector are preserved under this transformation (Lemma 4.3), ensuring that the regularity properties of the original varifold translate to the "straightened" varifold.

C. Perpendicularity via Tilt-Excess Decay

To prove perpendicularity (Theorem 5.2), the authors adapt the strategy of Brakke but replace the Pythagorean theorem with the quadratic flatness result from [KS25].

They construct a specific test vector field $\chi_{\lambda, r} u$ (where $u$ is a tangent vector) and analyze the first variation $\delta_G W$ .
By decomposing the variation into terms involving tilt-excess ( $\|S^\natural - T^\natural\|^2$ ) and height ( $|T^\perp(x)|^2$ ), they show that these terms vanish at the correct rate ( $r^{-k-1}$ ) due to the quadratic flatness hypothesis.
This forces the tangential component of the mean curvature vector to vanish, proving $h_F(V, b) \perp T$ .

D. Locality via Ambrosio-Masnou Adaptation

For the locality result (Theorem 5.6), the authors adapt ideas from Ambrosio and Masnou. They show that if the varifold $V$ has bounded anisotropic mean curvature and satisfies a quadratic tilt-excess decay condition, then on the set where $V$ coincides with a $C^2$ hypersurface $M$ , the anisotropic mean curvature of $V$ coincides with the classical anisotropic mean curvature of $M$ .

3. Key Results

Theorem 1.1 (Main Theorem)

Let $\phi$ be a uniformly convex $C^3$ -norm on $\mathbb{R}^{n+1}$ and $F$ be the associated elliptic integrand. Let $V$ be an $n$ -dimensional integral varifold in an open set $\Omega$ with bounded anisotropic first variation ( $\|\delta_F V\| \le H \|V\|$ ) and assume the support has locally finite $\mathcal{H}^n$ measure absolutely continuous with respect to $\|V\|$ .
Define $Q$ as the set of points in the support where the density is 1.
Then, for $\mathcal{H}^n$ -almost all $a \in Q \cap M$ (where $M$ is any embedded $C^2$ hypersurface):

Perpendicularity: The anisotropic mean curvature vector $h_F(V, a)$ is orthogonal to the tangent plane of $M$ at $a$ .
$h_F(V, a) \in \text{Nor}(M, a)$
Locality: The anisotropic mean curvature vector of the varifold coincides with the anisotropic mean curvature vector of the hypersurface $M$ .
$h_F(V, a) = h_F(M, a)$

Theorem 5.2 (Perpendicularity on Unit Density Layer)

Under the hypothesis of quadratic flatness (specifically the decay of tilt-excess and height), the mean curvature vector is perpendicular to the approximate tangent plane for almost all points in the unit-density layer. This theorem holds for general $C^1$ integrands of arbitrary codimension.

Theorem 5.6 (Locality)

If $V$ has bounded anisotropic first variation and satisfies a quadratic tilt-excess decay condition, then on the intersection of the support of $V$ and a $C^2$ hypersurface $M$ , the mean curvature vectors coincide.

4. Significance and Implications

Extension of Brakke's Theorem: This work successfully generalizes Brakke's fundamental perpendicularity theorem from the Euclidean (isotropic) setting to the anisotropic setting, a major open problem in geometric measure theory.
Overcoming Monotonicity Limitations: In the anisotropic case, monotonicity formulas (which are crucial for controlling "bad sets" in the isotropic case) are generally unavailable. The authors circumvent this by relying on the quadratic flatness established in their previous work, which provides the necessary decay estimates without monotonicity.
Regularity Theory: The results bridge the gap between weak solutions (varifolds) and classical smooth geometry. By proving that the mean curvature vector behaves locally like that of a smooth surface on the unit-density layer, the paper strengthens the understanding of the regularity of anisotropic minimal surfaces and mean curvature flow.
Methodological Advance: The technique of "linear straightening" via a point-dependent diffeomorphism to align anisotropic structures with Euclidean ones is a powerful new tool that may be applicable to other problems involving anisotropic energies and non-Euclidean geometries.

5. Limitations

The authors note that the perpendicularity result is currently restricted to the unit-density layer ( $Q$ ). Extending this to higher density points in the anisotropic setting remains difficult because:

Monotonicity formulas are not available.
The available Lipschitz approximation theorems for anisotropic integrands do not provide sufficiently strong decay estimates for the "bad set" (the part of the varifold not covered by the graph of the Lipschitz map) at higher densities.

Conclusion

This paper is a significant advancement in the regularity theory of varifolds. By combining quadratic flatness with a novel linearization technique, the authors establish that anisotropic mean curvature vectors are geometrically well-behaved (orthogonal and local) on the unit-density layer, effectively extending classical Euclidean results to the anisotropic realm.