Chromatic homology, Khovanov homology, and torsion (Q524328): Difference between revisions

Calculations of Khovanov homology have shown that torsion of order 2 is very common among links, while torsion of higher order is much more scarce. In this paper the authors give a partial explanation of this observation. Loosely speaking, they show that in the first few (and the last few) homological degrees all torsion is of order two. To make this more precise, consider a link diagram \(D\) and resolve every crossing so that the resulting Kauffman state sits over the 0-vertex of the hypercube. Associate a graph \(G\) to this resolution using the components as vertices and the traces of each crossing as edges. Let \(g\) be the length of the shortest cycle in \(G\). The theorem now states that all torsion in \(Kh^{i,j}(D)\) is of order 2 for all \(i \leq -c_-+g-1\) where \(c_-\) is the number of negative crossings in \(D\). The proof uses the result of \textit{L. Helme-Guizon} et al. [Fundam. Math. 190, 139--177 (2006; Zbl 1105.57012)] that Khovanov homology agrees with chromatic homology of \(G\) in this range. In fact the main work of this article consists in showing that all torsion in the chromatic homology of a graph is of order 2.