The 3D-index and normal surfaces (Q2359553)

The 3D-index as introduced by \textit{T. Dimofte}, \textit{D. Gaiotto} and \textit{S. Gukov} [Adv. Theor. Math. Phys. 17, No. 5, 975--1076 (2013; Zbl 1297.81149) and Commun. Math. Phys. 325, No. 2, 367--419 (2014; Zbl 1292.57012)] is obtained by assigning certain integer weights to the edges of an ideal triangulation \(\mathcal{T}\) of \(int\,M\), where \(M\) is a compact orientable \(3\)-manifold with \(\partial M\) consisting of \(r\geq 1\) tori. Summing up the weights leads to a \(q\)-series that depends on \([\gamma ]\in H_1 (\partial M;\mathbb{Z})\) and determines a function \(I_{\mathcal{T}}: H_1 (\partial M;\mathbb{Z})\to \mathbb{Z}((q^{1/2}))\). \textit{S. Garoufalidis} et al. [Geom. Topol. 19, No. 5, 2619--2689 (2015; Zbl 1330.57029)] showed that the index sum for \(I_{\mathcal{T}}( [\gamma ])\) converges for all \([ \gamma ]\) if and only if \(\mathcal{T}\) is \(1\)-efficient, that is it contains no embedded normal spheres or tori other than peripheral tori. The main goal of the present paper is to show a connection of the 3D-index with classical normal surface theory. After reviewing the definitions of the 3D-index and \(Q\)-normal surface theory the authors give a new formulation of the index as a sum over certain singular normal surfaces. They compute the 3D-index for many interesting examples. The authors achieve their goal by exhibiting the 3D-index as a generating series of normal surfaces associated to \(\mathcal{T}\) and explain how this fulfills a dream of constructing topological invariants of \(3\)-manifolds using normal surfaces.