The Kauffman skein module of a connected sum of 3-manifolds (Q1877604)

The author investigates in this paper the behavior of the Kauffman bracket skein module (KBSM for short) of 3-manifolds over a ``certain special ring \(k\)'' under a connected sum. Basically, the KBSM \(K(M)\) for a compact oriented 3-manifold is defined over \(\mathbb{C}[t,t^{-1}]\), see [\textit{J. Przytycki}, J. Math. 16, No. 1, 45--66 (1999; Zbl 0947.57017)], and is closely related to the \(SL(2,{\mathbb C})\)-character ring [\textit{D. Bullock}, Comment. Math. Helv. 72, 521--542 (1997; Zbl 0907.57010)]. In fact, Przytycki has already investigated the behavior of the KBSM under connected sum in a special case [Manuscr. Math. 101, No. 2, 199--207 (2000; Zbl 0947.57018)]. Using another special ring which makes the calculation of the KBSM easier, the author in this paper continues this study of the behavior under connected sum. To state the main theorem, we define a ring \({\mathcal R}\) as follows. Let \({\mathcal I}\) be the submonoid of the multiplicative monoid of \({\mathbb Z}[\alpha, s]\) generated by \(\alpha\), \(s\), \(s^{2n}-1\) and \(1\pm \alpha s^n\) for all positive integers \(n\), and \(c_{\lambda}\), which is associated with the Young diagram \(\lambda\), for all Young diagrams \(\lambda\). Then \({\mathcal R}\) is defined as \({\mathbb Z}[\alpha,s]\) localized at \({\mathcal I}\). In fact, regarding the case of \(k={\mathcal R}\), \({\mathbb Q}(\alpha,s)\), the field of rational functions in \(\alpha\) and \(s\), the KBSM of a connected sum of 3-manifolds \(M_1\) and \(M_2\) is isomorphic to the tensor product of each KBSM. Namely, \(K(M_1 \sharp M_2)\cong K(M_1)\otimes K(M_2)\). The ideas to show the main theorem are as follows. For a handlebody \({\mathcal H}_n\) of genus \(n \in {\mathbb Z}_{\geq 0}\), the set of generators of the KBSM \(K({\mathcal H}_n)\) can be described by the basis of the Birman-Murakami-Wenzl algebra. (Therefore in the definition of the ring \({\mathcal R}\) above we need the Young diagrams). Consider a connected sum of two handlebodies \({\mathcal H}_m\) and \({\mathcal H}_n\). The resulting manifold is a handlebody of genus \(m+n\) attached a 2-handle along the boundary of ``the meridian disk'' separating \({\mathcal H}_m\) and \({\mathcal H}_n\), and the KBSM of the resulting manifold is the same as that of \({\mathcal H}_{m+n}\). Then in the case of \(k={\mathcal R}\), \({\mathbb Q}(\alpha, s)\), the generators of \(K({\mathcal H}_m \sharp {\mathcal H}_n\)) which are in neither \(K({\mathcal H}_m)\) nor \(K({\mathcal H}_n\)) vanish. By using this, we can show that \(K({\mathcal H}_m \sharp {\mathcal H}_n)\) is exactly the tensor product of \(K({\mathcal H}_m)\) and \(K({\mathcal H}_n)\). The observation of the KBSM of a handlebody above immediately leads us to the compact oriented 3-manifolds case via the Heegaard splitting. Note that this approach is quite different from Przytycki's proof. At the end of this paper, the author gives a calculation of \(K(S^1 \times S^2 \sharp S^1 \times S^2)\). \(K(\sharp^m S^1 \times S^2)\) is in fact a free \(k\)-module generated by the empty link in the case of \(k={\mathcal R}\), \({\mathbb Q}(\alpha,s)\).