🔢 mathematics

Some universal inequalities for Dirichlet eigenvalues of the Laplacian on a Euclidean convex domain

This paper establishes two universal inequalities governing the Dirichlet eigenvalues of the Laplacian on Euclidean convex domains.

Original authors: Kei Funano

Published 2026-04-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Kei Funano

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

Imagine you have a musical instrument, but instead of strings or air columns, it's a shape—a room, a field, or a blob of clay. If you were to "pluck" this shape, it would vibrate at specific frequencies. In mathematics, these frequencies are called eigenvalues. The lowest frequency is the deep, fundamental hum (the first eigenvalue), and the higher ones are the overtones (the second, third, and so on).

The shape of the room determines the pitch. A long, skinny hallway sounds different from a square room. But here's the big question mathematicians have asked for decades: Can we predict the relationship between the low notes and the high notes just by knowing the shape is "convex" (bulging outward, like a ball or a cube, rather than a star or a crescent)?

This paper by Kei Funano says, "Yes, we can!"

Here is the breakdown of the paper's discovery using simple analogies.

1. The Problem: The "Universal" Rule

Mathematicians love "universal" rules—laws that work for everything in a certain category without needing to measure every single detail.

The Old Way: Previously, we had rules that worked for any shape, but they were very vague. They could tell us that the 100th note is higher than the 1st, but they couldn't tell us how much higher with any precision.
The New Goal: The author wanted to find a rule that says: "If you know the pitch of the 10th note, you can calculate a very specific range for the 100th note, provided the shape is convex."

2. The Main Discovery: The "Stretchy Rubber Band"

The paper establishes two main rules (Theorems) that act like a stretchy rubber band connecting the low notes to the high notes.

The Upper Bound (The "Ceiling")

Theorem 1.1 says: If you have a convex shape, the higher notes can't get too loud compared to the lower notes.

The Analogy: Imagine the notes are rungs on a ladder. If you are on rung 10, you know you can't jump to rung 10,000 in a single bound. There is a "ceiling" on how fast the pitch can rise as you go up the ladder.
The Math: The paper proves that the ratio between the $k$ -th note and the $l$ -th note is bounded by a simple formula involving the numbers $k$ and $l$ . It's like saying, "No matter how weird your convex room is, the 100th note will never be more than $X$ times the 10th note."

The Lower Bound (The "Floor")

Theorem 1.2 says: If the lower notes are already high enough (a specific condition), the higher notes can't get too quiet.

The Analogy: This is the "floor." If you are already climbing high up the ladder, you can't suddenly drop down to the basement. The pitch must keep rising at a certain minimum speed.
The Catch: This rule only works if the "room" isn't too weirdly shaped in a way that traps the sound (mathematically, if the lower eigenvalue is large enough compared to the very first one). If the room is a long, thin hallway, the first note is very low, and this specific rule doesn't apply until you get to higher notes.

3. How Did They Do It? (The "Box" Trick)

Proving this for a weird, curvy blob is hard. So, the author used a clever trick involving boxes.

The Problem: Convex shapes can be curvy and complex.
The Solution: The author used a mathematical tool (Hatcher's Lemma) that says: "Any convex shape can be squeezed inside a box, and a smaller box can be squeezed inside it."
The Magic: Instead of calculating the music of a curvy blob, the author calculated the music of a perfect box (an orthotope). Boxes are easy to solve; their notes are just simple sine waves.
The Result: Because the blob is "sandwiched" between two boxes, its notes must fall somewhere between the notes of the big box and the small box. By doing the math on the boxes, the author proved the rule holds for the blob too.

4. Why Does This Matter?

Before this paper, we had a patchwork of rules. Some worked for the first two notes, some for the first ten, some for huge numbers. They were all different formulas.

This paper unifies them. It says: "For any convex shape, the relationship between any two notes follows this simple pattern."

It's like discovering that while every car engine sounds different, they all follow the same basic rule of how the RPMs relate to the speed, as long as the car isn't broken (convex).

Summary in One Sentence

Kei Funano proved that for any "bulging" shape (convex domain), there is a strict, predictable mathematical relationship between the low and high vibration frequencies, acting like a universal ruler that measures how fast the pitch can rise, regardless of the shape's specific details.

1. Problem Statement

The paper addresses the problem of establishing universal inequalities for the Dirichlet eigenvalues of the Laplacian operator on bounded domains in $\mathbb{R}^n$ . Specifically, it seeks to bound the ratio of arbitrary eigenvalues $\lambda_k(\Omega)/\lambda_l(\Omega)$ (where $k \ge l$ ) in terms of the ratio of their indices $(k/l)^{2/n}$ .

While universal inequalities are known to hold for arbitrary bounded domains (e.g., Payne-Pólya-Weinberger, Yang's inequalities), these typically focus on consecutive eigenvalues ( $\lambda_{k+1}/\lambda_k$ ) or sums of eigenvalues. Establishing bounds for the ratio of arbitrary eigenvalues that hold uniformly for a specific class of domains is more challenging. This paper focuses specifically on the class of bounded convex domains, aiming to derive both upper and lower bounds for $\lambda_k/\lambda_l$ that depend only on the dimension $n$ and the indices $k, l$ .

2. Methodology

The author employs a combination of geometric approximation, domain monotonicity, and spectral estimates. The core methodological steps include:

Geometric Approximation (Hatcher's Lemma): The proof relies on a lemma by Hatcher, which states that any bounded convex domain $\Omega$ can be approximated by an orthotope (a rectangular box) $R$ . Specifically, there exists an orthotope $R$ such that $\frac{1}{\sqrt{n}}R \subseteq \Omega \subseteq nR$ (after appropriate coordinate scaling).
Domain Monotonicity: The authors utilize the property that if $\Omega_1 \subseteq \Omega_2$ , then $\lambda_k(\Omega_2) \le \lambda_k(\Omega_1)$ . This allows the reduction of the problem from general convex domains to orthotopes.
Spectral Estimates for Orthotopes: The core technical work involves proving the inequalities for orthotopes. This is achieved by:
- Constructing maximal $r$ -separated point sets within the orthotope.
- Using volume arguments to relate the number of points to the eigenvalues.
- Applying the Berezin-Li-Yau inequality (a lower bound for eigenvalues based on volume).
- Applying the Hersch-Protter inequality (a lower bound for the first eigenvalue $\lambda_1$ in terms of the inradius for convex domains).
Contradiction Arguments: For the lower bound proof, the author assumes the opposite of the desired inequality and derives a contradiction using the Hersch-Protter inequality and the properties of the inradius.

3. Key Contributions and Results

The paper establishes two main theorems providing upper and lower bounds for the ratio of eigenvalues on convex domains.

Theorem 1.1: Universal Upper Bound

There exists a constant $c_n$ depending only on the dimension $n$ such that for any bounded convex domain $\Omega$ and any $k \ge l$ :
$\lambda_k(\Omega) \le c_n \left(\frac{k}{l}\right)^{2/n} \lambda_l(\Omega)$

Constant: $c_n = 12n^3 (j_{\frac{n}{2}-1, 1})^2$ , where $j_{\nu, 1}$ is the first positive zero of the Bessel function $J_\nu$ .
Significance: This provides a uniform upper bound for the growth of eigenvalues relative to a lower index eigenvalue, valid for all convex domains.

Theorem 1.2: Universal Lower Bound

There exist constants $\alpha_n$ and $\beta_n$ depending only on $n$ such that for any bounded convex domain $\Omega$ and $k \ge l$ :
If $\lambda_l(\Omega) > \alpha_n \lambda_1(\Omega)$ , then:
$\beta_n \left(\frac{k}{l}\right)^{2/n} \lambda_l(\Omega) \le \lambda_k(\Omega)$

Constants:
- $\alpha_n = \frac{4^2}{\pi^2} (j_{\frac{n}{2}-1, 1})^2 n^3$
- $\beta_n = \frac{1}{12n^3 (j_{\frac{n}{2}-1, 1})^2}$
Condition: The lower bound requires that the eigenvalue $\lambda_l$ is sufficiently large compared to the first eigenvalue $\lambda_1$ . The paper demonstrates via a counterexample (a thin rectangle $\Omega_k = [0, k] \times [0, 1]$ ) that this condition is necessary; without it, the ratio $\lambda_k/\lambda_l$ can be arbitrarily small for fixed $k, l$ as the domain geometry changes.

Corollary 1.3: Multiplicity Bound

As a consequence of Theorem 1.1, the paper derives an upper bound for the multiplicity $m_k(\Omega)$ of the $k$ -th eigenvalue. If $\lambda_k(\Omega) > \alpha_n \lambda_1(\Omega)$ , then:
$m_k(\Omega) \le C_n \cdot k$
where $C_n$ is a constant depending only on $n$ . This implies that for high-energy eigenvalues on convex domains, the multiplicity grows at most linearly with $k$ .

4. Comparison with Previous Work

The author contextualizes these results against classical inequalities:

Payne-Pólya-Weinberger (PPW) & Yang: These established inequalities for consecutive eigenvalues ( $\lambda_{k+1}/\lambda_k$ ) for arbitrary domains.
Levitin-Parnovski: Provided bounds for sums of eigenvalues.
Cheng-Yang: Provided bounds involving $\lambda_1$ .

Comparison:

The inequalities in Theorem 1.1 are technically "weaker" in terms of the specific constants compared to the best-known bounds for consecutive eigenvalues in general domains (e.g., Yang's inequality).
The Advantage: The novelty lies in the unification of the ratio $\lambda_k/\lambda_l$ for arbitrary $k, l$ (not just consecutive) and the derivation of a lower bound (Theorem 1.2), which is the first result of its kind for Dirichlet eigenvalues in terms of lower eigenvalues.
Trade-off: These results require the convexity assumption, whereas classical universal inequalities hold for non-convex domains. The author notes that it is currently unknown whether convexity is strictly necessary for these specific forms of the inequalities.

5. Significance

First Lower Bound: Theorem 1.2 represents a significant breakthrough as the first result providing a lower bound for Dirichlet eigenvalues in terms of lower-order eigenvalues for a specific class of domains.
Convexity as a Tool: The paper demonstrates that assuming convexity allows for the control of eigenvalue ratios in a way that is difficult or impossible for general domains, specifically enabling the derivation of lower bounds.
Geometric Insight: The proof technique linking the geometry of the domain (via Hatcher's lemma and inradius) directly to spectral properties offers a robust framework for future spectral geometry research.
Multiplicity Control: The corollary provides new constraints on the degeneracy of eigenvalues for convex domains, a topic of interest in spectral theory and mathematical physics.

In summary, Kei Funano's paper extends the theory of universal inequalities by leveraging the geometric properties of convex domains to establish two-sided bounds for arbitrary eigenvalue ratios, filling a gap in the literature regarding lower bounds and non-consecutive eigenvalue comparisons.