Generalized Connes-Chern characters in KK-theory with an application to weak invariants of topological insulators

TL;DR

This paper develops a generalized Connes-Chern character within KK-theory to analyze weak topological invariants of topological insulators, providing new index theorems and stability results under disorder.

Contribution

It introduces a new framework for pairing K-theory with cyclic cohomology to study topological invariants in disordered systems, extending index theory to weak invariants.

Findings

Derived local formulas for the Connes-Chern character in crossed product algebras.

Connected numerical invariants to weak topological invariants of insulators.

Established stability regimes for invariants under strong disorder.

Abstract

We use constructive bounded Kasparov K-theory to investigate the numerical invariants stemming from the internal Kasparov products $K_i(\mathcal A) \times KK^i(\mathcal A, \mathcal B) \rightarrow K_0(\mathcal B) \rightarrow \mathbb R$, $i=0,1$, where the last morphism is provided by a tracial state. For the class of properly defined finitely-summable Kasparov $(\mathcal A,\mathcal B)$-cycles, the invariants are given by the pairing of K-theory of $\mathcal B$ with an element of the periodic cyclic cohomology of $\mathcal B$, which we call the generalized Connes-Chern character. When $\mathcal A$ is a twisted crossed product of $\mathcal B$ by $\mathbb Z^k$, $\mathcal A = \mathcal B \rtimes_\xi^\theta \mathbb Z^k$, we derive a local formula for the character corresponding to the fundamental class of a properly defined Dirac cycle. Furthermore, when $\mathcal B = C(\Omega) \rtimes_{\xi'}^{\phi} \mathbb Z^j$, with $C(\Omega)$ the algebra of continuous functions over a disorder configuration space, we show that the numerical invariants are connected to the weak topological invariants of the complex classes of topological insulators, defined in the physics literature. The end products are generalized index theorems for these weak invariants, which enable us to predict the range of the invariants and to identify regimes of strong disorder in which the invariants remain stable. The latter will be reported in a subsequent publication.