🔢 mathematics

Kernel estimates and weak (1,1)-boundedness of pseudo-differential operators on compact Lie groups

This paper establishes the weak (1,1)-boundedness and $H^1$ - $L^1$ -continuity of pseudo-differential operators in global Hörmander classes on compact Lie groups by deriving kernel estimates, thereby providing new end-point $L^1$ -estimates for subelliptic problems like the sub-Laplacian and heat-type operators on $SU(2)$ that are inaccessible via standard Euclidean calculus.

Original authors: Duván Cardona, Rafik Yeghoyan, Michael Ruzhansky

Published 2026-02-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Duván Cardona, Rafik Yeghoyan, Michael Ruzhansky

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 are trying to navigate a complex, multi-layered city. This city isn't built on a flat map like New York or London; it's a Compact Lie Group. Think of it as a perfectly round, finite sphere (like the surface of a ball, but in higher dimensions) where every point has a special symmetry. You can rotate, flip, and twist the city, and it always looks the same.

In this city, there are Pseudo-Differential Operators. Let's call them "The Magic Filters."

The Problem: The Magic Filters are Fickle

Imagine you have a machine (the Magic Filter) that takes a messy, noisy signal (like a chaotic crowd of people shouting) and tries to smooth it out or extract a specific pattern.

In the flat, easy world of Euclidean space (our standard flat maps), mathematicians have known for decades how to predict exactly how these filters behave. They know that if the filter is "sharp" enough, it won't turn a manageable crowd into an uncontrollable riot.

However, on our special, curved city (the Compact Lie Group), things get tricky. The rules change depending on how "curved" the city is and how the filter is tuned.

The Goal: The authors want to prove that even in this tricky, curved city, if you tune your Magic Filter correctly, it won't cause a disaster. Specifically, they want to prove it has "Weak (1,1) Boundedness."
What does that mean? In plain English: If you feed the machine a "messy" input (an $L^1$ function, which might have spikes or be very loud in one spot), the output might get a little wild, but it won't explode into infinity. It will stay "mostly under control," even if it's not perfectly smooth.

The Solution: The Kernel as a "Distance Map"

To prove this, the authors didn't just look at the machine's settings (the symbols). They looked at the Kernel.

Think of the Kernel as a Distance Map or a Rulebook that tells the machine how to react to two points in the city.

If Point A and Point B are close together, the rulebook says, "Hey, your reaction to A and B should be almost the same."
If they are far apart, the rulebook says, "Your reaction can be different, but it must fade away quickly."

The authors spent most of the paper doing the hard math to prove that this Rulebook is well-behaved. They showed that no matter how you zoom in or out (whether the distance $R$ is huge or tiny), the difference in the machine's reaction to two nearby points is always bounded by a constant. It never goes crazy.

The Analogy: Imagine you are standing in a crowded room.

The Kernel is your ability to hear two people whispering next to each other.
The Estimate proves that if two people whisper almost the same thing, your ears won't suddenly go deaf or start hearing a symphony. The change in what you hear is proportional to how close they are.

The "Full Range" Breakthrough

Previously, mathematicians could only prove this safety for filters that were "very strict" (where the parameters $\rho$ and $\delta$ had to be far apart). It was like saying, "This machine is safe only if you turn the volume knob to the very top."

This paper is a breakthrough because they proved it works for the Full Range.

They showed the machine is safe even when the volume knob is in the middle, or slightly off-center.
They covered every possible setting where the machine doesn't break the laws of physics (specifically, where $\rho \neq 0$ and $\delta \neq 1$ ).

The Real-World Application: The "Sub-Laplacian"

Why do we care? The authors apply this to a specific, difficult problem on the group SU(2) (which is mathematically the same as a 3D sphere, $S^3$ ).

Imagine a Heat Engine or a Sub-Laplacian (a machine that tries to smooth out heat or vibrations).

In the standard flat world, we know exactly how to solve the equation "Heat = Source."
On this curved sphere, the "Source" might be a weird, jagged distribution of heat (like a sudden, intense spark).

Using their new "Safe Filter" proof, the authors show that even if the source of the heat is very messy (in a specific mathematical sense called $W^{1, -1/4}$ ), the resulting temperature distribution ( $u$ ) will still be manageable. It won't blow up. It will belong to a class of functions that, while maybe not perfectly smooth, are at least "weakly bounded" (they won't cause a singularity).

Summary in a Nutshell

The City: A curved, symmetric mathematical world (Compact Lie Group).
The Machine: A complex filter (Pseudo-differential operator) that processes data.
The Fear: That the machine might turn a small, messy input into an infinite, uncontrollable output.
The Proof: The authors checked the machine's "Rulebook" (the Kernel) and proved that the rules are always sensible, no matter how you tune the machine.
The Result: They proved the machine is safe for all reasonable settings, not just the easy ones.
The Payoff: This allows us to solve difficult physics equations (like heat diffusion on a sphere) even when the starting data is very messy, guaranteeing that the solution won't explode.

They essentially built a universal safety net for these mathematical machines, ensuring that even in the most complex, curved geometries, the math stays under control.

1. Problem Statement

The paper addresses the boundedness of pseudo-differential operators (ΨDOs) on compact Lie groups, specifically focusing on the weak (1,1) continuity and $H^1(G) \to L^1(G)$ boundedness.

Context: While the $L^p$ boundedness ( $1 < p < \infty$ ) of ΨDOs in Hörmander classes $\Psi^m_{\rho,\delta}$ is well-established (following Fefferman and Hörmander), the endpoint cases ( $p=1$ and $p=\infty$ ) are more delicate.
The Gap: Previous results on compact Lie groups (e.g., by Delgado and Ruzhansky) often required the strict condition $\delta < \rho$ . The authors aim to extend these results to the full range $0 \le \delta \le \rho \le 1$ (with $\rho \neq 0, \delta \neq 1$ ), including the subelliptic case where $\delta = \rho$ .
Specific Challenge: Standard pseudo-differential calculus based on local coordinates (Euclidean charts) fails for certain subelliptic operators on groups like $SU(2)$ because the principal symbol is not invariant under coordinate changes. The paper utilizes a global quantization approach based on the group's representation theory to overcome this.

2. Methodology

The authors employ a global quantization framework using the group Fourier transform and non-commutative phase space $G \times \widehat{G}$ . Their methodology relies on three main pillars:

Global Symbol Calculus:
- Symbols $\sigma(x, \zeta)$ are defined on $G \times \widehat{G}$ , where $\widehat{G}$ is the unitary dual.
- Derivatives in the spatial variable $x$ are handled via left-invariant differential operators.
- Derivatives in the frequency variable $\zeta$ are handled via difference operators $\Delta_\zeta$ acting on the representation matrices.
- The classes $S^m_{\rho,\delta}(G)$ are defined by the decay of these derivatives: $\|\Delta^\alpha_\zeta \partial^\beta_x \sigma(x, \zeta)\|_{op} \le C \langle \zeta \rangle^{m - \rho|\alpha| + \delta|\beta|}$ .
Kernel Estimates (The Core Innovation):
- The authors derive precise estimates for the distributional kernel $K(x, y)$ of operators in $\Psi^m_{\rho,\delta}(G)$ .
- They prove that for an operator $T$ with order $m \le -\frac{n}{2}(1-\rho)$ , the kernel satisfies a Hörmander-type condition adapted to the group geometry:
  $\int_{B(z, 2R)^\rho} |K(x, y) - K(x, z)| dx \le C$
  (where the radius of integration depends on whether $R \ge 1$ or $R < 1$ , and the exponent involves $\rho$ ).
- This is achieved by decomposing the kernel using a dyadic decomposition in the frequency domain (using a cutoff function $\phi(\langle \zeta \rangle/t)$ ) and applying Plancherel's theorem and Weyl's law for the eigenvalues of the Laplacian.
Calderón-Zygmund Decomposition on Groups:
- To prove weak (1,1) boundedness, they adapt the classical Calderón-Zygmund decomposition to the compact Lie group setting.
- They decompose a function $f \in L^1(G)$ into a "good" part $g \in L^2(G)$ and a "bad" part $b = \sum b_j$ (atoms supported on disjoint balls with mean zero).
- The "good" part is handled via $L^2$ boundedness.
- The "bad" part is handled by splitting it further ( $F'$ and $F''$ ) and utilizing the derived kernel estimates to control the singularity near the support of the atoms.

3. Key Contributions

Extension of Parameter Range: The paper establishes weak (1,1) boundedness for the entire range $0 \le \delta \le \rho \le 1$ (excluding trivial cases $\rho=0, \delta=1$ ). This includes the critical subelliptic case $\delta = \rho$ , which was previously difficult to handle with standard methods.
Alternative Proofs: The authors provide new, simpler proofs for the boundedness of these operators from:
- $L^1(G) \to L^{1,\infty}(G)$ (Weak (1,1)).
- $H^1(G) \to L^1(G)$ .
- $L^\infty(G) \to BMO(G)$ .
- These proofs rely directly on the kernel estimates rather than complex interpolation or duality arguments used in previous works.
Sharp Order Conditions: They identify the sharp order $m = -\frac{n}{2}(1-\rho)$ required for these boundedness results.
Application to Sub-Laplacians: The theory is applied to specific subelliptic operators on $SU(2) \cong S^3$ , specifically the sub-Laplacian $L_{sub} = X^2 + Y^2$ and the heat-type operator $T = Z - X^2 - Y^2$ .

4. Main Results

Theorem 1.1 (Kernel Estimates):
For $T \in \Psi^m_{\rho,\delta}(G)$ with $m \le -\frac{n}{2}(1-\rho)$ , the kernel $K(x,y)$ satisfies:

If $R \ge 1$ : $\int_{B(z, 2R)^c} |K(x, y) - K(x, z)| dx \le C$ .
If $R < 1$ : $\int_{B(z, 2R^\rho)^c} |K(x, y) - K(x, z)| dx \le C$ .
Here, $y \in B(z, R)$ . The constant $C$ is independent of $R$ .

Theorem 1.2 (Weak (1,1) Boundedness):
Under the same conditions, $T$ extends to a bounded operator from $L^1(G)$ to $L^{1,\infty}(G)$ .

Theorem 3.4 & 3.5 (Endpoint Boundedness):

$T$ maps the Hardy space $H^1(G)$ continuously to $L^1(G)$ .
$T$ maps $L^\infty(G)$ continuously to $BMO(G)$.
These hold for the full range $0 \le \delta \le \rho \le 1$ .

Theorem 4.1 (Subelliptic Estimates on $SU(2)$):
For the sub-Laplacian $L_{sub}$ and heat operator $T = Z - L_{sub}$ on $SU(2)$, the following endpoint estimate holds:
$\|u\|_{L^{1,\infty}} \le C \|Tu\|_{W^{1, -1/4}} + C_N \|u\|_{W^{1, -N}}$
This implies that if $f \in W^{1, -1/4}(SU(2))$ , then the solution $u$ to $Tu=f$ belongs to $L^{1,\infty}(SU(2))$ .

5. Significance

Resolution of Subelliptic Endpoint Problems: The results are crucial for understanding the regularity of solutions to subelliptic partial differential equations (PDEs) on compact manifolds, particularly where standard elliptic theory fails (e.g., Hörmander's sum of squares operators).
Global vs. Local: By utilizing the global quantization on Lie groups, the authors bypass the limitations of local coordinate charts, which is essential for operators whose symbols are not invariant under coordinate transformations (a common issue in sub-Riemannian geometry).
Simplification of Theory: The kernel-based approach unifies the treatment of $L^1$ , $H^1$ , and $BMO$ boundedness, offering a more direct and transparent proof strategy compared to previous duality-heavy approaches.
Sharpness: The paper confirms that the order $m = -\frac{n}{2}(1-\rho)$ is the sharp threshold for these estimates, aligning with known Euclidean results but extending them to the non-commutative setting with greater generality in the parameters $\rho$ and $\delta$ .

In summary, this paper significantly advances the harmonic analysis of pseudo-differential operators on compact Lie groups by establishing the full range of weak-type estimates through novel kernel estimates, with direct applications to the regularity theory of subelliptic PDEs.