Khovanov homology of a unicolored B-adequate link has a tail (Q2018369)

Rozansky's main goal in this paper is to give a categorification of results from [\textit{C. Armond}, Algebr. Geom. Topol. 13, No. 5, 2809--2826 (2013; Zbl 1271.57005)] and [\textit{S. Garoufalidis} and \textit{T. T. Q. Lê}, Res. Math. Sci. 2, Article ID 1, 55 p. (2015; Zbl 1334.57021)] showing that the Jones polynomial of a unicolored B-adequate link has a stable tail at large color. When computing the Jones polynomial using the Kauffman bracket, one starts by ``splicing'' every crossing in two ways, called here A-splicing and B-splicing. Looking at the case where one chooses all B-splicings, a crossing is called B-adequate if it takes place between two different circles, and a knot or link that admits a diagram with only B-adequate crossings is called B-adequate. This class of links includes in particular all alternating links. Interested in lifting Armond's, Garoufalidis and Lê's results, Rozansky uses categorified Jones-Wenzl projectors as defined by Frenkel, Stroppel and Sussan [\textit{I. B. Frenkel} et al., Quantum Topol. 3, No. 2, 181--253 (2012; Zbl 1256.17006)], \textit{B. Cooper} and \textit{V. Krushkal} [Quantum Topol. 3, No. 2, 139--180 (2012; Zbl 1362.57015)], and himself [Fundam. Math. 225, 305--326 (2014; Zbl 1336.57025)], to define the colored Khovanov homologies he wants to compare. The main result, Theorem 2.2, states that for a B-adequate link diagram, there is a sequence of degree preserving maps from the Khovanov homology of the link colored by \(N\) to the one where the link is colored by \(N+1\). Furthermore, these maps are isomorphisms in low homological degree. This yields the notion of tail homology of a B-adequate link, that categorifies the tail of the unicolored Jones polynomial. An important part of the paper is devoted to a careful step-by-step construction of the maps relating successive colors for Khovanov homology, which is based on a succession of local transformations giving isomorphisms at low degree. Rozansky remarks that this construction might as well be applied for any kind of link, but conjectures (Conjecture 2.14) that only B-adequate links have a non-trivial tail.