A Remark on the Non-Compactness of $W^{2,d}$ Immersions of $d$-Dimensional Hypersurfaces

Siran Li

Abstract

We consider the continuous $W^{2,d}$ immersions of $d$-dimensional hypersurfaces in $\mathbb{R}^{d+1}$ with second fundamental forms uniformly bounded in $L^d$. Two results are obtained: first, a family of such immersions is constructed, whose limit fails to be an immersion of a manifold. This addresses the endpoint cases in J. Langer and P. Breuning. Second, under the additional assumption that the Gauss map is slowly oscillating, we prove that any family of such immersions subsequentially converges to a set locally parametrised by Hölder functions.

Introduction

In differential geometry and geometric analysis, a central question is whether a sequence of surfaces with controlled "energy" or curvature will converge to a well-behaved limit. Landmark results by J. Langer and P. Breuning established that for $d$-dimensional hypersurfaces, if the second fundamental form (a measure of the surface's extrinsic curvature) is bounded in $L^p$ where $p > d$, the sequence converges to a smooth immersed manifold.

This convergence relies heavily on the Sobolev–Morrey embedding theorem, which provides the necessary regularity to prevent the surface from degenerating. However, this embedding fails exactly at the "endpoint" case where $p = d$. This paper investigates whether compactness still holds at this critical boundary and explores what additional geometric conditions might be required to salvage regularity in the limit.

The Failure of Compactness: The Spiraling Counterexample

The first major contribution of this work is the construction of a "wildly spiraling" family of immersions that demonstrates the failure of compactness in $L^d$.

The author notes that the $L^d$ norm of the second fundamental form is scaling-invariant in $d$ dimensions. This property allows for a specific type of geometric degeneration that is impossible when $p > d$. By using symmetric mollifiers and carefully defined angle functions in spherical coordinates, the author constructs a family of smooth immersions that wrap around an origin point an increasing number of times as a parameter $\epsilon \to 0$.

Crucially, the author uses a topological degree argument to prove non-compactness. Even as the surfaces concentrate near a point, the topological degree of these maps remains constant (e.g., $10^m$). Because any potential limit manifold would necessarily have a degree of zero in this construction, a topological contradiction arises. This proves that the sequence cannot converge to an immersed manifold, effectively settling the endpoint case left open by previous literature.

Restoring Regularity via BMO Gauss Maps

If $L^d$ bounds alone are insufficient for compactness, what additional constraints might suffice? The paper shifts to the framework of "chord-arc surfaces" and harmonic analysis to provide an answer.

The author introduces a condition on the Gauss map—the field of unit normal vectors to the surface. By assuming the Gauss map is "slowly oscillating," specifically that its Bounded Mean Oscillation (BMO) norm is sufficiently small, the author proves a partial regularity result.

Under this BMO constraint, any sequence of $W^{2,d}$ immersions contains a subsequence that converges to a Hölder graph system. This means that while the limit might not be a smooth manifold, it is locally representable by $C^{0,\gamma}$ functions. This result demonstrates that controlling the oscillation of the normal vector prevents the wild spiraling seen in the counterexample and ensures the limit remains somewhat regular, even if it is potentially fractal-like.

Key Insights and Implications

  • Scaling Invariance: The $L^d$ norm's scaling invariance is the mathematical driver behind the non-compactness, allowing curvature to concentrate without blowing up the total integral.
  • Topological Obstructions: The use of topological degree provides a robust method for proving that certain geometric limits are unattainable, regardless of how "close" the surfaces appear to be in a metric sense.
  • BMO as a Regularizer: The paper highlights BMO as a natural geometric bridge. It is a weaker condition than the $L^p$ ($p>d$) bounds used in classical theorems but stronger than the $L^d$ bound alone, providing a middle ground for studying surface limits.

Conclusion and Future Work

While this paper settles the question of non-compactness at the $p=d$ boundary, it opens several new avenues for research. The limiting objects under the BMO condition can still be highly irregular, such as nowhere differentiable Weierstrass-type functions. A significant open question remains: what natural geometric conditions beyond BMO would guarantee more regular limits, such as Lipschitz or $BV$ (Bounded Variation) surfaces? Additionally, the author suggests that extending these results to isometric immersions in higher dimensions ($d \geq 3$) remains a challenging and important problem for the field.

Summary updated

Abstract

We consider the continuous $W^{2,d}$ immersions of $d$-dimensional hypersurfaces in $\mathbb{R}^{d+1}$ with second fundamental forms uniformly bounded in $L^d$. Two results are obtained: first, a family of such immersions is constructed, whose limit fails to be an immersion of a manifold. This addresses the endpoint cases in J. Langer and P. Breuning. Second, under the additional assumption that the Gauss map is slowly oscillating, we prove that any family of such immersions subsequentially converges to a set locally parametrised by Hölder functions.
BibTeX
@inproceedings{li2019remark,
  author = {Siran Li},
  title = {{A Remark on the Non-Compactness of $W^\{2,d\}$ Immersions of $d$-Dimensional Hypersurfaces}},
  booktitle = {Proceedings of the American Mathematical Society},
  year = {2019},
  doi = {10.1090/proc/14710},
}

Loading related content...

Figure