🔢 mathematics

The cut-off resolvent can grow arbitrarily fast in obstacle scattering

This paper demonstrates that for time-harmonic acoustic scattering by a compact sound-soft obstacle with a non-smooth boundary, the norm of the cut-off resolvent can grow arbitrarily fast with the frequency $k$ , contrasting with the known exponential growth bound that holds only for smooth obstacles.

Original authors: Simon N. Chandler-Wilde, Siavash Sadeghi

Published 2026-05-14

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Simon N. Chandler-Wilde, Siavash Sadeghi

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 "Echo" Problem

Imagine you are in a large, empty room (the "universe") and you shout a single note. The sound travels out, hits an object in the room (an "obstacle"), and bounces back. In physics, this is called scattering.

The paper asks a very specific question about how loud that echo can get as you change the pitch of your shout (the frequency, or $k$ ).

The Standard Rule: If the object in the room is smooth and round (like a ball or a smooth rock), the echo gets louder as you shout higher pitches, but there is a predictable limit. It grows exponentially, like a snowball rolling down a hill, but it doesn't explode instantly.
The Paper's Discovery: The authors ask: "What if the object isn't smooth? What if it's jagged, weird, or has a very strange shape?" They prove that if you are allowed to build an obstacle with any shape (no matter how jagged or weird), you can design it so that the echo becomes infinitely loud at specific pitches. In fact, you can make it grow as fast as you want—faster than exponential, faster than a rocket, as fast as you can imagine.

The Key Characters

To understand the paper, we need to meet three characters:

The Sound Wave ( $u$ ): This is the acoustic wave traveling through the air.
The Obstacle ( $\Gamma$ ): This is the thing blocking the sound. In this paper, it's a "sound-soft" obstacle, meaning the sound wave must be zero right at its surface (like a perfect sound absorber).
The Resolvent ( $C_{k,R}$ ): Think of this as the "Volume Knob" or the "Amplifier Factor."
- If you shout a sound ( $f$ ) into the room, the Resolvent tells you how loud the resulting echo ( $u$ ) will be inside a specific area.
- The paper investigates: How big can this Volume Knob get as we turn up the pitch ( $k$ )?

The "Smooth" vs. "Jagged" Debate

For decades, scientists knew that if your obstacle is smooth (like a polished marble sphere), the Volume Knob grows at a manageable rate. Even if the shape traps sound in a weird way (like a billiard ball bouncing between two curved walls), the growth is usually just logarithmic or polynomial (slow and steady).

However, the authors wondered: Does the shape have to be smooth?

They say: "No. If we drop the rule that the obstacle must be smooth, we can break the rules of growth entirely."

The Magic Construction: The "Swiss Cheese" Tower

How did they prove this? They didn't just guess; they built a mathematical "monster" obstacle.

Imagine you are building a tower out of cubes (like dice).

The Cubes: You take a sequence of cubes, getting smaller and smaller.
The Holes: On the surface of each cube, you cut out a tiny square hole.
The Stacking: You stack these cubes in a very specific, layered pattern, like a spiral staircase going down into a deep well.
The Tuning: You choose the size of each cube and the size of each hole very carefully. You tune them so that each cube acts like a perfect "resonator" for a specific pitch.

The Analogy:
Think of each cube as a tiny, perfect organ pipe.

If you blow into a pipe at the exact right frequency, it screams loudly.
The authors arranged thousands of these pipes so that they don't interfere with each other.
They tuned the "pipes" so that for a specific sequence of pitches ( $k_1, k_2, k_3...$ ), the sound inside each pipe gets louder and louder.
By making the holes on the pipes smaller and smaller, they made the "leakage" of sound slower, allowing the internal volume to build up to massive levels.

The Result: "Arbitrarily Fast" Growth

The paper proves that for any speed of growth you can name (even something like "double the speed of light" or "factorial growth"), you can build a jagged obstacle that makes the Volume Knob grow that fast at a specific sequence of pitches.

If you want the echo to grow by 10x every step: They can build an obstacle that does it.
If you want it to grow by a billion times every step: They can build an obstacle that does it.
If you want it to grow faster than any formula you can write down: They can still do it.

The only requirement is that the obstacle must be a finite, compact shape (it fits in a box) and the space around it must be connected (you can walk from any point to any other point without jumping over a gap).

Why Does This Matter? (According to the Paper)

The paper doesn't talk about building better speakers or medical imaging. Instead, it focuses on mathematical limits.

Breaking the "Smooth" Assumption: It shows that the "smoothness" of an object is the only thing keeping the echo from exploding uncontrollably. If you allow jagged, fractal-like, or "Swiss cheese" shapes, the standard mathematical safety nets disappear.
The "Quasimode" Trick: The authors use a mathematical tool called a "quasimode." Imagine a note that is almost a perfect resonance. In a smooth room, these notes die out quickly. In their jagged, stacked-cube room, these "almost notes" get trapped and amplified to infinity.
The Counter-Intuitive Finding: Usually, we think "weird shapes" just make things messy. This paper shows that "weird shapes" can be engineered to create extreme order in the chaos of sound, leading to massive amplification.

Summary in One Sentence

If you build an obstacle with a sufficiently jagged and complex shape (specifically, a stack of cubes with tiny holes), you can force the sound echo to grow as fast as you want, proving that without smoothness, there is no limit to how loud the resonance can get.

Technical Summary: The Cut-Off Resolvent Can Grow Arbitrarily Fast in Obstacle Scattering

Problem Statement
The paper investigates the asymptotic behavior of the cut-off resolvent norm for the time-harmonic acoustic scattering problem by a compact sound-soft obstacle $\Gamma \subset \mathbb{R}^n$ ( $n \ge 2$ ). The problem is modeled by the inhomogeneous Helmholtz equation $\Delta u + k^2 u = -f$ in the exterior domain $\Omega = \mathbb{R}^n \setminus \Gamma$ , subject to the Dirichlet boundary condition $u=0$ on $\partial \Gamma$ and the Sommerfeld radiation condition.

The central quantity of interest is the norm of the cut-off resolvent, $C_{k,R}$ , defined as the operator norm mapping a compactly supported source term $f$ (supported in a ball $\Omega_R$ ) to the solution $u$ restricted to that same ball. Specifically, $C_{k,R} = \|\chi_R (-\Delta - k^2)^{-1} \chi_R\|_{L^2(\Omega) \to L^2(\Omega)}$ . The authors address the question: How fast can $C_{k,R}$ grow as the wavenumber $k \to \infty$ ?

Previous literature has established that for smooth obstacles, the growth is at most exponential (specifically $O(e^{\alpha k})$ ), and for non-trapping or specific trapping geometries (star-shaped, hyperbolic, parabolic), the growth is bounded by algebraic or logarithmic rates (e.g., $O(k^{-1})$ , $O(k^{-1}\log k)$ , or $O(k)$ ). The open question addressed here is whether imposing no smoothness or geometric constraints on $\Gamma$ (other than compactness and connected complement) allows for growth rates exceeding the exponential bounds known for smooth scatterers.

Methodology
The authors employ a constructive approach to demonstrate that the growth of $C_{k,R}$ can be arbitrarily fast along a specific sequence of wavenumbers. The methodology involves three main components:

Variational Formulation and Inf-Sup Constants: The scattering problem is reformulated as a variational problem in a bounded domain $\Omega_R$ . The authors relate the norm of the cut-off resolvent to the inf-sup constant ( $\beta_{k,R}$ ) of the associated sesquilinear form. Specifically, they show that a small inf-sup constant implies a large resolvent norm. The relationship is bounded by inequalities involving $C_{k,R}$ and the energy norm of the solution.
Quasimode Construction: To establish a lower bound on the resolvent norm, the authors construct "quasimodes" (approximate eigenfunctions) for the scattering operator. They design an obstacle $\Gamma$ composed of a countable union of disjoint cubes $O_j$ with specific side lengths $\ell_j$ . The side lengths are chosen such that $k_j^2$ corresponds to the first Dirichlet eigenvalue of the Laplacian in each cube $O_j$ .
- The obstacle $\Gamma$ is defined as the boundary of the union of these cubes, with a small "hole" (an $(n-1)$ -dimensional cube of side $\ell_j \epsilon_j$ ) removed from one face of each $O_j$ .
- The eigenfunctions of the individual cubes are extended by zero to the exterior domain. Due to the small holes, these functions satisfy the boundary condition $u=0$ on $\Gamma$ only approximately.
- The authors utilize a multiplicative trace inequality (Lemma 2.2) to bound the error introduced by the holes. By carefully choosing the hole size $\epsilon_j$ relative to the growth rate $a_j$ and wavenumber $k_j$ , they ensure the residual of the approximate solution is sufficiently small to force the inf-sup constant to be small, thereby forcing the resolvent norm to be large.
Geometric Arrangement (The "Layered" Construction): A critical technical challenge is arranging the infinite sequence of cubes $\{O_j\}$ such that:
- They are mutually disjoint.
- Their union is bounded (ensuring $\Gamma$ is compact).
- The complement $\Omega = \mathbb{R}^n \setminus \Gamma$ remains connected.
  The authors provide a constructive proof (Theorem 1.9) showing that if the side lengths $\ell_j$ decay sufficiently fast (specifically satisfying a condition involving logarithmic factors), the cubes can be arranged in "layers" along the $x_n$ -axis. This arrangement ensures the union is bounded and the complement is connected, despite the infinite number of obstacles.

Key Contributions and Results
The paper's primary contribution is the proof that without smoothness assumptions on the obstacle, the cut-off resolvent can grow arbitrarily fast.

Theorem 1.4 (Main Result): Given any positive, increasing, unbounded sequence of wavenumbers $(k_j)$ satisfying a mild growth condition (roughly $k_j \gtrsim (j \log j)^{1/n}$ ) and any arbitrarily rapidly increasing sequence of target bounds $(a_j)$ , there exists a compact obstacle $\Gamma$ with a connected complement such that the cut-off resolvent norm satisfies $C_{k_j, R} \ge a_j$ for all $j$ .
- This result implies that there is no universal upper bound on the growth rate of the resolvent for general compact obstacles; the growth can exceed exponential rates.
Theorem 1.5: Provides the explicit construction of the obstacle $\Gamma$ (as a union of modified cubes) and proves the lower bound on the resolvent norm for a given sequence of wavenumbers, assuming the cubes can be arranged to satisfy disjointness, boundedness, and connectivity.
Theorem 1.9: Proves that the geometric conditions required in Theorem 1.5 can be satisfied by a specific arrangement of the cubes, provided the side lengths $\ell_j$ decay at a rate compatible with the growth of $k_j$ . This establishes the existence of the required obstacle configuration.
Extension of Trace Inequalities: The paper extends known multiplicative trace inequalities and properties of radiating solutions (Lemma A.1) to dimensions $n > 3$ , which are necessary for the variational arguments used in the main proofs.

Significance and Claims
The authors claim that their results fundamentally alter the understanding of resolvent growth in scattering theory when smoothness is not assumed.

Contrast with Smooth Cases: While smooth obstacles are known to have at most exponential growth (Burq, 2002), and specific geometric classes have polynomial or logarithmic bounds, this paper demonstrates that the lack of smoothness allows for "arbitrarily fast" growth.
Sharpness of Smoothness Assumptions: The results highlight that the exponential bound for smooth scatterers is a consequence of the smoothness itself. By relaxing the smoothness to merely a compact set with a connected complement, the constraints on resolvent growth vanish.
Implications for Numerical Analysis: The authors note (in Section 1.4) that these results have been applied to establish lower bounds on the condition number of first-kind integral equation formulations for general compact scatterers, suggesting that numerical methods for such problems may face severe ill-conditioning that cannot be bounded by standard exponential rates.

The paper does not propose new experimental setups or applications but strictly provides a theoretical counterexample to the idea that resolvent growth is universally bounded by exponential rates, emphasizing that the geometry of the obstacle (specifically the lack of smoothness) is the determining factor.