Functoriality for the \(\mathfrak{su}_{3}\) Khovanov homology (Q1017845)

This paper proves the full functoriality under tangle cobordisms of the Morrison--Nieh categorification of the \(\mathfrak{su}_3\) Kuperberg spider web invariant. In [J. Knot Theory Ramifications 17, No.~9, 1121--1173 (2008; Zbl 1300.57013)], \textit{S. Morrison} and \textit{A. Nieh} defined the category \({\mathbf {Kob}}(\mathfrak{su}_3)\) -- roughly speaking the category of chain complexes over the category of, up to some relations, seamed cobordisms between trivalent graphs -- which is endowed with a canopolis structure, i.e., some planar algebra stuctures coherently defined on both objects and morphisms of the categories, cf. [\textit{D. Bar-Natan}, Geom. Topol. 9, 1443--1499 (2005; Zbl 1084.57011)]. Then, they defined a map from the set of tangles to the objects of \({\mathbf {Kob}}(\mathfrak{su}_3)\) up to homotopy. It results in a tangle invariant which contains the \(\mathfrak{su}_3\) Khovanov homology. The present paper extends this map to a canopolis functor \({\mathcal F}\) from the tangle cobordisms category, which is naturally endowed with a canopolis structure, and \({\mathbf {Kob}}(\mathfrak{su}_3)\) (up to homotopy). It means extending the map to tangle cobordisms such that it is invariant under Carter and Saito movie moves and such that it respects the canopolis structures. The paper begins with a simple and readable introduction to canopolis and to \({\mathbf {Kob}}(\mathfrak{su}_3)\). Then the author gives the construction of the Morrison-Nieh map, but with some more technical details in order to have an explicit chain map for every Reidemeister move. He also gives some technical facts on these maps which will be used to prove the main statement. The functor \({\mathcal F}\) is straighforwardly defined using these Reidemeister maps and some obvious Morse ones. The main part of the paper consists in proving the consistency of \({\mathcal F}\) with regard to movie moves. The proof is largely inspired from [\textit{D. Clark, S. Morrison} and \textit{K. Walker}, Geom. Topol. 13, No.~3, 1499--1582 (2009; Zbl 1169.57012)] which deals with the \(\mathfrak{su}_2\) case. However, the \(\mathfrak{su}_3\) case needs some more care concerning the non-existence of non-trivial positive morphisms between some given webs. Moreover, the paper provides a computation-free proof of invariance under move MM10. Finally, the paper ends with some hints concerning a possible extension of the theory to knotted webs.