The unbounded Kasparov product by a differentiable module (Q2073788)

One of the main features of the \(K\!K\)-theory of \(C^*\)-algebras developed by \textit{G. G. Kasparov} [Math. USSR, Izv. 16, 513--572 (1981; Zbl 0464.46054)] is the bilinear and associative product \[ K\!K_n(A,B)\times K\!K_m(B,C)\to K\!K_{n+m}(A,C) \] between the \(K\!K\)-groups associated to (separable) \(C^*\)-algebras \(A,B\) and \(C\). This paper focuses on the problem of constructing the above product at the level of \textit{unbounded} cycles: given two unbounded Kasparov modules, the goal is finding an \textit{explicit} unbounded Kasparov module which represents the Kasparov product. This is important for at least two reasons: \begin{itemize} \item it allows for bypassing technical results like the so-called ``absoprtion'' and ''technical'' theorems; \item the unbounded version of the product retains a larger amount of geometric data (relating to the asymptotic behavior of eigenvalues of differential operators). \end{itemize} The specific case considered here is where the class in \(K\!K(A,B)\) is represented by a \(C^*\)-correspondence whose left action (the action of \(A\)) is compact, while the class in \(K\!K(B,C)\) is represented by an unbounded self-adjoint and regular operator (acting on a \(C^*\)-correspondence from \(B\) to \(C\)). One of the main contributions of the paper is that the construction is carried out without imposing any kind of smooth projectivity condition on the \(A\)-\(B\)-correspondenece. This is relevant in view of examples arising from non-complete manifolds. Secondly, the text introduces a notion of unbounded \textit{modular cycle}, somewhat related to the concept of twisted spectral triple considered by \textit{A. Connes} and \textit{H. Moscovici} [in: Traces in number theory, geometry and quantum fields. Wiesbaden: Vieweg. 57--71 (2008; Zbl 1159.46041)] and the half-closed chains of \textit{M. Hilsum} [Math. Scand. 107, No. 1, 73--89 (2010; Zbl 1198.19005)]. Unbounded Kasparov modules are a special case of unbounded modular cycles. Given such a modular cycle, the paper also introduces a bounded transform, similarly to the work of \textit{S. Baaj} and \textit{P. Julg} [C. R. Acad. Sci., Paris, Sér. I 296, 875--878 (1983; Zbl 0551.46041)], turning a modular cycle into a \(K\!K\)-class. The transform considered here entails much technical care, mainly due to the (seemingly innocent) passage from straight commutators to twisted commutators. The identification of the Kasparov product involving the aforementioned objects relies on the concept of \(F_2\)-connection introduced by \textit{A. Connes} and \textit{G. Skandalis} [Publ. Res. Inst. Math. Sci. 20, 1139--1183 (1984; Zbl 0575.58030)]. Another key ingredient in the paper is the local-global principle for unbounded regular operators [\textit{J. Kaad} and \textit{M. Lesch}, J. Funct. Anal. 262, No. 10, 4540--4569 (2012; Zbl 1251.46030); \textit{F. Pierrot}, J. Lie Theory 16, No. 4, 651--689 (2006; Zbl 1152.22006)]. It is worth noting that standard methods (in the unbounded setting), like those of \textit{D. Kucerovsky} [J. Oper. Theory 44, No. 2, 255--275 (2000; Zbl 0996.46030)], do not apply in the general context of modular cycles, contributing to the technical difficulty of the proof.