Sharp mapping properties of Poisson transforms and the Baum-Connes conjecture

1. Problem Statement and Context

The paper addresses two interconnected problems in the representation theory of semisimple Lie groups and operator K-theory:

1. Quantitative Helgason Conjecture: The classical Helgason conjecture establishes an isomorphism between joint eigenfunctions of invariant differential operators on a symmetric space $K\backslash G$ and their boundary values (hyperfunctions) on the Furstenberg boundary $P\backslash G$. However, this is an algebraic/isomorphism result at the level of smooth functions or hyperfunctions. The authors seek a sharp, quantitative analogue at the level of distributions, specifically determining the optimal Sobolev regularity required for the Poisson transform (specifically the Knapp-Wallach Szegö map) to act as a bounded operator with a closed range between specific function spaces.
2. The Baum-Connes Conjecture: This conjecture relates the topological K-theory of a group's reduced C*-algebra to the equivariant K-homology of its classifying space. While proven for many classes of groups (e.g., a-T-menable, hyperbolic), it remains open for closed subgroups of semisimple Lie groups of real rank one (e.g., $SL(3, \mathbb{Z})$). Julg proposed a program to resolve this for real rank one groups, reducing the problem to two conjectures: one representation-theoretic (proven by Julg-Nishikawa) and one analytical. The analytical conjecture requires proving that the commutator of the Szegö map with smooth functions on the Furstenberg compactification is a compact operator.

2. Methodology

The authors employ a sophisticated blend of microlocal geometric analysis, representation theory, and non-commutative geometry.

• Heisenberg Calculus: The core analytical tool is the Heisenberg calculus (developed by van Erp, Yuncken, and others) adapted to the filtered manifold structure of the boundary $P\backslash G$. Since $G$ has real rank one, $P\backslash G$ is a Carnot manifold of depth 2. The authors utilize the associated scale of Heisenberg-Sobolev spaces ($W^s_H$), defined via complex powers of a sub-Laplacian $\Delta_H$.
• Parabolic Tangent Groupoids: To analyze the operators, the authors construct an extended parabolic tangent groupoid ($\widehat{TH}X$) for $X = P\backslash G$. This geometric object allows them to treat operators as distributions on a groupoid that interpolates between the boundary and the bulk, facilitating the study of their asymptotic behavior and homogeneity.
• Ewert's Averaging Technique: They utilize a technique by Ewert involving the averaging of smooth functions over the dilation action of the group $A$ (the split component of the parabolic subgroup). This allows them to represent operators as meromorphic families of Heisenberg pseudodifferential operators.
• Knapp-Stein Intertwiners: The proof relies heavily on the relationship between the Szegö map and Knapp-Stein intertwiners (intertwining operators between principal series representations). The authors show that the composition of the Szegö map with its adjoint yields a Knapp-Stein intertwiner, which is a known object in the Heisenberg calculus.

3. Key Contributions and Results

A. Sharp Mapping Properties of the Szegö Map (Theorem 1)

The primary result establishes the optimal Sobolev regularity for the Szegö map $S_V$.

• Setup: Let $G$ be a connected, semisimple Lie group of real rank one. Let $(V, \tau)$ be an irreducible $K$-representation with highest weight $\lambda$. The Szegö map $S_V$ maps distributions on the boundary $P\backslash G$ (twisted by a character $\nu_V$) to smooth functions on the symmetric space $K\backslash G$.
• Result: The map extends to a bounded operator with closed range:
  $$S_V: W^{-s_V}_H(P\backslash G, H_{\nu_V}) \to L^2(K\backslash G, V)$$
  where $s_V$ is a specific exponent determined by the character $\nu_V = s_V \alpha$ (with $\alpha$ being the simple restricted root).
• Isomorphism: The map induces a $G$-equivariant isomorphism between a closed subspace of the Heisenberg-Sobolev space on the boundary and the discrete series subrepresentation on $K\backslash G$.
• Source Projection: The source projection $P_V$ (the orthogonal projection onto the image of the adjoint) is shown to be a zeroth-order operator in the Heisenberg calculus.

B. Compactness of Commutators (Theorem 3)

This result resolves the analytical conjecture in Julg's program.

• Result: For any smooth function $b$ on the Furstenberg compactification $Z = K\backslash G \cup P\backslash G$, the commutator $[S_V, b]$ is a compact operator from $W^{-s_V}_H(P\backslash G, H_{\nu_V})$ to $L^2(K\backslash G, V)$.
• Mechanism: The proof demonstrates that the composition $[S_V, b]^\dagger [S_V, b]$ is a Heisenberg pseudodifferential operator of order $-2s_V - 1$ with a vanishing principal symbol. By the properties of the Heisenberg calculus, such operators are compact.

C. Implications for the Baum-Connes Conjecture

• By proving the compactness of the commutator, the authors complete the analytical part of Julg's program.
• Combined with the representation-theoretic result of Julg and Nishikawa, this implies that the Baum-Connes conjecture with coefficients holds for all closed subgroups of semisimple Lie groups of real rank one.

D. K-Homological Interpretation (Section 8)

• The source projection $P_V$ defines a non-trivial class in the equivariant K-homology group $K^G_1(P\backslash G)$.
• The authors show that this class corresponds to the boundary of the K-homology class of the Dirac operator on $K\backslash G$ under the boundary map in the exact sequence associated with the Furstenberg compactification. This links the analytic properties of the Poisson transform directly to the topological invariants of the group.

4. Significance

1. Resolution of a Major Open Problem: The paper provides the final piece of the puzzle for the Baum-Connes conjecture in the real rank one case, a long-standing open problem in geometric group theory and operator algebras.
2. Sharpness of Estimates: Unlike previous works (e.g., by Julg-Kasparov) which established continuity in some Sobolev scale, this paper identifies the exact optimal exponent and proves the closed range property, which is crucial for the invertibility and Fredholm properties required in K-theory.
3. Methodological Advancement: The successful application of the Heisenberg calculus and extended parabolic tangent groupoids to the study of Poisson transforms and intertwiners provides a powerful new toolkit for analyzing singular integral operators on boundaries of symmetric spaces.
4. Bridge Between Analysis and Topology: The work explicitly connects the analytic regularity of the Szegö map (via the Heisenberg calculus) with the topological structure of the Baum-Connes assembly map, demonstrating how microlocal analysis can solve deep problems in non-commutative geometry.

5. Limitations and Future Directions

The authors note that their proof relies heavily on the real rank one assumption. In higher ranks, the boundary $P\backslash G$ is a multifiltered manifold, and the standard Heisenberg calculus is insufficient. They suggest that a "multifiltered" calculus (as studied by Yuncken) would be required to extend these results to higher-rank groups, though they acknowledge that obstructions related to Lafforgue's strong property (T) may make the higher-rank case significantly more complex.