Non-semisimple 3-manifold invariants derived from the Kauffman bracket (Q2125359)

A family of quantum invariants \(Z_r\) of closed oriented \(3\)-manifolds has been constructed (for each fixed integer \(r \ge 3\)) by using the non-semisimple representation theory of some quantum groups. The present paper proposes a description of these non-semisimple invariants in purely algebraic and combinatorial terms, bypassing the need to use the machinery of Hopf algebras and their representations. This construction will follow and adapt one used by \textit{C. Blanchet} [Comment. Math. Helv. 67, No. 3, 406--427 (1992; Zbl 0771.57005)] for the semisimple WRT invariants \(\tau_r\), and relies on the definition of the Temperly-Lieb category, the Kauffman bracket polynomials, and the modified trace. The Temperly-Lieb category \(TL\), which carries a ribbon structure, is presented in Section 2.1, and is roughly constructed from isotopy classes of framed tangles modulo two relations. These tangles have to be unoriented images of proper embeddings into the unit cube \(I^3\) of a disjoint unit of finitely many copies of \(S^1\) and \(I\), and are indexed by their (discrete) boundaries on the \textit{top} and \textit{bottom} lines of \(I^3\). There are two versions of special idempotent elements in \(TL\), called the Jones-Wenzl idempotents, that are defined recursively as \(f_m\) (for \(0 \le m \le r-1\)) (the \textit{simple} ones) and as \(g_m\) (for \(r \le m \le 2r-2\)) (the \textit{non-semisimple} ones). These elements are important in the technical proofs of the latter parts of the paper; for example, the \textit{ideal of projective objects} of \(TL\) is generated by the element \(f_{r-1}\) (Proposition 2.1). Crucially, a specific linear combination of the \(f_m\) (suppressing the last one) is defined as the \textit{semisimple Kirby color} and plays an important role in the definition of the WRT invariants. One of the fundamental ingredients in the new description of the \(Z_r\) here is the definition of a \textit{non-semisimple Kirby color}, which is a specific linear combination of the \(g_m\) (Definition 3.1). Blanchet's construction of a \(SO(3)\) version of the WRT invariant \(\tau_r\) relies on the concept of \textit{bichrome tangle}, which consists of the union of a \textit{blue} framed tangle, which is oriented and labeled by idempotent morphisms of \(TL\), and a \textit{red} framed link. If a bichrome tangle is embedded inside a \(3\)-manifold, the blue part should be viewed as an element of the corresponding Kauffman bracket skein module, and the red part as a surgery description. The topological invariant \(\tau_r\) is defined for each pair \((M,T)\), where \(M\) is a closed oriented \(3\)-manifold and \(T\) is a bichrome link inside \(M\), and depends on the construction of a topological invariant \(F_\omega\) for bichrome links, which depends on the Kirby color and the Kauffman bracket polynomial. The non-semisimple \(Z_r\) are defined in a similar way (once the relevant concepts have been adapted to the new situation), but, for the technical reasons described in Section 3.2 and at the end of Section 5.2, only some bichrome links can be used (the so-called \textit{ admissible} ones.) The main result here is then Theorem 3.5, stating that the \(Z_r (M,T)\) are indeed topological invariants. Its proof appears in Section 6.3, and needs another important concept, namely that of finite-dimensional representations of the small quantum group \({\bar U} _q {\mathfrak sl}_2\), where \(q\) is the \(r\)-th root of unity. It is the study of a functor between the category of bichrome tangles and that of finite-dimensional representations of \(\bar U _q \mathfrak sl_2\) that allows for an interpretation of the \(Z_r\) as renormalized Hennings invariants (Section 4), and Section 5 then recovers the HRK algorithm in the context of the \textit{bead category} of \(\bar U _q \mathfrak sl_2\), allowing for the desired proof. The authors also note that the \(Z_r\) can be viewed as first-order extensions of the WRT invariants \(\tau_r\), at least for rational homology spheres, and that these invariants, in the form that is presented here, can serve as a first step towards a purely combinatorial construction of non-semisimple TQFTs.