The Chang-Marshall Trace Inequality for Sobolev functions in domains in higher dimensional space $\mathbb{R}^n$

Jungang Li, Guozhen LuarXiv 2021 · DOI · arXiv

Abstract

In their celebrated work, Chang and Marshall established a critical trace inequality of Moser-Trudinger type for holomorphic functions with mean value zero on the unit disc in the complex plane. This paper addresses a question proposed by S. Y. Alice Chang regarding whether the Chang-Marshall type inequality for holomorphic functions on the unit disk holds for Sobolev functions on general domains in higher dimensional Euclidean space $\mathbb{R}^n$ for all $n \ge 2$. We partially answer her question affirmatively, establishing sharp exponential-type bounds for the boundary values of these functions.

Introduction

In the study of mathematical analysis, Sobolev inequalities provide a fundamental framework for understanding how the smoothness of a function—quantified by the integrability of its derivatives—restricts its potential for local growth or "blow-up." A cornerstone of this field is the Moser-Trudinger inequality. This result addresses a borderline case: when a function's gradient is integrable to the $n$-th power in an $n$-dimensional space, the function itself is exponentially integrable over the domain.

While the interior behavior of such functions is well-documented, the behavior on the boundary (the "trace") presents a more complex challenge. In 1985, Alice Chang and David Marshall proved a sharp inequality specifically for the boundary of the unit disk in the complex plane ($n=2$). In 2015, Alice Chang challenged the mathematical community to extend this result to general domains in higher dimensions ($\mathbb{R}^n$ for $n \geq 3$). The primary difficulty lies in the dimensional mismatch: the boundary has a lower dimension than the interior, and standard scaling techniques often fail to bridge the gap between interior gradient norms and boundary integrals.

Methodology: The Trace Function Approach

To overcome the geometric hurdles of higher dimensions, the authors depart from classical techniques, such as Adams' method, and introduce a novel strategy centered on a variational "trace function."

1. Defining the Trace Function ($w$)

The authors define a specific function $w$ that minimizes a functional balancing the interior gradient and the boundary values. This function $w$ serves as a mathematical bridge, effectively translating information from the $n$-dimensional interior to the $(n-1)$-dimensional boundary.

2. Conversion to Partial Differential Equations

By analyzing the Euler-Lagrange equations associated with this functional, the authors demonstrate that $w$ satisfies a $p$-Laplacian type equation. This equation is coupled with specific Neumann boundary conditions, allowing the problem to be framed within the context of well-understood elliptic operators.

3. Integration by Parts and Interior Mapping

Using the properties of the trace function, the authors utilize integration by parts to "pull" the boundary integral into the interior of the domain. This transformation is critical because it allows the application of existing, powerful Moser-Trudinger results for the interior (specifically those developed by Cianchi) to resolve the boundary problem.

4. Establishing Sharpness

To ensure the constants identified are the best possible, the authors construct a "counter-sequence" of functions. These functions are designed to concentrate their mass near a single point on the boundary. By showing that the integral diverges if the exponent exceeds the identified threshold, they prove the sharpness of their results.

Key Insights and Results

The Critical Constant

The research identifies a sharp constant $\beta_n$ defined as: $$\beta_n = (n-1)\left(\frac{\omega_{n-1}}{2}\right)^{1/(n-1)}$$ This value represents the "speed limit" for exponential growth on the boundary. If the exponent $\alpha$ remains below $\beta_n$, the function remains well-behaved and integrable; if it exceeds $\beta_n$, the function can become infinite.

Main Theorems

  • Theorem 2.1 (The Inequality): For any $\alpha < \beta_n$, the supremum of the boundary integral $\int_{\partial \Omega} e^{\alpha |u|^{n/(n-1)}} d\mathcal{H}_{n-1}$ is bounded by a constant $C$, provided the function $u$ has a mean value of zero on the domain ($\int_\Omega u = 0$) and its gradient norm satisfies $\|\nabla u\|_n \leq 1$.
  • Theorem 2.2 (Sharpness): The constant $\beta_n$ is sharp. For any $\alpha > \beta_n$, there exists a sequence of functions such that the boundary integral tends toward infinity.

Unlike the 1985 predecessor, which was restricted to holomorphic functions on a 2D disk, this new result is significantly more general, applying to any Sobolev function on a bounded domain with a sufficiently smooth ($C^{1, \alpha}$) boundary.

Limitations and Future Research

While this work provides a substantial answer to Chang's challenge, several avenues for further investigation remain:

  1. Mean Value Constraints: The current proof relies on the function having a mean value of zero over the entire domain. Future work may explore whether the inequality holds if the mean value is zero only on the boundary ($\int_{\partial \Omega} u = 0$).
  2. Boundary Regularity: The assumption of a $C^{1, \alpha}$ boundary is a standard regularity condition, but it remains to be seen if the inequality persists for "rougher" domains, such as those with Lipschitz boundaries.
  3. The Critical Case: This paper focuses on the subcritical case where $\alpha < \beta_n$. Determining the behavior of the integral at the exact critical point $\alpha = \beta_n$ for general domains remains an open and technically demanding question in geometric analysis.

Summary updated

Abstract

In their celebrated work [5], Chang and Marshall established a critical trace inequality of Moser-Trudinger type for holomorphic functions with mean value zero on unit disc in the complex plane. The main purpose is to address a question proposed to us by S. Y. Alice Chang who asked whether the Chang-Marshall type inequality for holomorphic functions on unit disk in the complex plane holds for Sobolev functions on general domains in higher dimensional Euclidean space $\mathbb{R}^n$ for all $n\ge 2$. We partially answer her question affirmatively.
BibTeX
@misc{li2021chang,
  author = {Jungang Li and Guozhen Lu},
  title = {{The Chang-Marshall Trace Inequality for Sobolev functions in domains in higher dimensional space $\textbackslash\{\}mathbb\{R\}^n$}},
  year = {2021},
  eprint = {2108.06792},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2108.06792},
}

Loading related content...

Figure