Fractional Sobolev isometric immersions of planar domains

Siran Li, Mohammad Reza Pakzad, Armin Schikorra

TL;DR

This paper proves that "rough" isometric immersions of flat domains into $\mathbb{R}^3$—specifically those belonging to the fractional Sobolev space $W^{1+s, 2/s}$ for $s \ge 2/3$—are surprisingly well-behaved. The authors demonstrate that these mappings are $C^1$ regular and "developable" (meaning they consist of flat or ruled sections), bridging a major gap between classical smooth geometry and the "wild" world of low-regularity embeddings.

What Problem Does This Paper Address?

In differential geometry, an isometric immersion is a way to bend a flat sheet (like a piece of paper) into 3D space without stretching or tearing it.

There is a famous "flexibility vs. rigidity" dichotomy in this field. If the immersion is very smooth ($C^2$), it is rigid and must be developable—you can only bend it into cylinders, cones, or tangent developable surfaces. However, the famous Nash-Kuiper theorem shows that if the immersion is only $C^1$ (once continuously differentiable), it can be incredibly "flexible" and "wild," allowing you to crumple a large sheet into an arbitrarily small ball.

The challenge is finding the exact threshold of "roughness" where flexibility ends and rigidity begins. Previous research established results for standard Sobolev spaces ($s=1$) and Hölder spaces ($C^{1,\alpha}$). This paper tackles the fractional Sobolev regime, specifically the critical threshold $s = 2/3$, which is the borderline case where standard embedding theorems often fail.

Methods

The authors employ several sophisticated tools from mathematical analysis and partial differential equations (PDEs):

  • Mollification and Commutator Estimates: They smooth out the rough mappings using a convolution kernel (mollification) and then carefully analyze the error (the commutator) as the smoothing parameter goes to zero.
  • Weak Codazzi-Mainardi Equations: These are fundamental geometric equations relating the curvature of a surface to its metric. Since the mappings aren't smooth enough for standard derivatives, the authors derive a "weak" version of these equations that holds in a distributional sense.
  • Distributional Jacobian Determinants: A core technical hurdle was handling the Jacobian determinant for maps that are not twice-differentiable. The authors prove a new result (Theorem 3) showing that the distributional Jacobian behaves like a standard determinant under multiplication by scalar functions, which is essential for proving developability.
  • Hodge Decomposition: They use this technique to break down complex vector fields into simpler components (gradient and curl-free parts) to analyze the second fundamental form of the surface.

Datasets

Not discussed in the paper. This is a theoretical research paper in pure mathematics; it relies on rigorous proofs and logical derivation rather than experimental data or training sets.

Key Insights and Ideas

  • The 2/3 Threshold: The paper identifies $s = 2/3$ as a critical regularity threshold. For $s \ge 2/3$, the surface retains its geometric rigidity.
  • Defining the Second Fundamental Form: The authors successfully define a "weak" second fundamental form for mappings that are only in $W^{1+s, p}$. This allows them to talk about the "curvature" of a surface that isn't technically twice-differentiable.
  • Determinant-like Behavior: One of the most significant "aha moments" is Theorem 3, which proves that the distributional Jacobian determinant satisfies a product-rule-like identity ($\text{Jac}(f)[\phi] = \text{Jac}(g)[\lambda^n \phi]$ when $\nabla f = \lambda \nabla g$). This provides the technical bridge needed to show that if one component of the mapping is developable, the others must be as well.
  • Connection to Elasticity: The work is motivated by nonlinear elastic plate theory (the Kirchhoff model), suggesting that the "shapes" paper takes when bent are governed by these specific regularity classes.

Key Results

  • Theorem 1 (Main Result): Any isometric immersion in the local fractional Sobolev space $W^{1+s, 2/s}$ for $2/3 \le s < 1$ is $C^1$ regular and developable.
  • Theorem 2: They prove that planar deformations with symmetric Jacobians and vanishing distributional Jacobians are constant along lines, providing a new analytical tool for studying surface deformations.
  • Corollary 1.2: They show that for $s \ge 2/3$, there is a minimum radius $\rho_0$ such that a disk cannot be isometrically immersed into a ball smaller than that radius—effectively ruling out the "Nash-Kuiper crumpling" in this regularity class.

Limitations and Future Work

  • The $sp < 2$ Limit: The authors acknowledge that Theorem 1 fails if $sp < 2$. In this lower-regularity regime, "conical singularities" can occur (like the point of a paper cone), which breaks developability.
  • Boundary Regularity: While the paper establishes $C^1$ regularity in the interior of the domain, Remark 1.5 notes that proving $C^1$ regularity up to the boundary remains an open expectation, likely requiring the boundary itself to be smooth ($C^{1,\alpha}$).
  • Super-critical vs. Critical: The proof for the super-critical case ($s > 2/3$) is somewhat simpler; the "borderline" case of $s = 2/3$ required the development of the new distributional Jacobian tools presented in Theorem 3.

Summary updated

Abstract

We discuss $C^1$ regularity and developability of isometric immersions of flat domains into $\mathbb R^3$ enjoying a local fractional Sobolev $W^{1+s, \frac2s}$ regularity for $2/3 \le s&lt; 1 $, generalizing the known results on Sobolev and Hölder regimes. Ingredients of the proof include analysis of the weak Codazzi-Mainardi equations of the isometric immersions and study of $W^{2,\frac2s}$ planar deformations with symmetric Jacobian derivative and vanishing distributional Jacobian determinant. On the way, we also show that the distributional Jacobian determinant, conceived as an operator defined on the Jacobian matrix, behaves like determinant of gradient matrices under products by scalar functions.
BibTeX
@inproceedings{li2022fractional,
  author = {Siran Li and Mohammad Reza Pakzad and Armin Schikorra},
  title = {{Fractional Sobolev isometric immersions of planar domains}},
  booktitle = {ANNALI},
  year = {2022},
  doi = {10.2422/2036-2145.202107\_007},
}
Figure