Rasmussen's spectral sequences and the sl_N-concordance invariants

Abstract: Combining known spectral sequences with a new spectral sequence relating reduced and unreduced sl(N)-homology yields a relationship between the Homflypt-homology of a knot and its sl(N)-concordance invariants. As an application, some of the sl(N)-concordance invariants are shown to be linearly independent.

Review text

In this paper, the author presents a spectral sequence going from reduced (tensorized by a fixed graded space) to unreduced \(\mathfrak{sl}_N\) Khovanov--Rozansky (KR) homologies. Then he uses it to compute \(\mathfrak{sl}_N\)--Rasmussen invariants of some pretzel knots, and those computations discard linear dependances between these invariants. KR homologies are links invariants categorifying the HOMFLYPT polynomials and their \(\mathfrak{sl}_N\) specifications. There are several versions of them.\newline 1. The unreduced version \([\![\;.\;]\!]_N\) is a bigraded homology whose Euler characteristic recovers the \(\mathfrak{sl}_N\)--polynomial [\textit{M. Khovanov} and \textit{L. Rozansky}, Fundam. Math. 199, No. 1, 1--91 (2008; Zbl 1145.57009)].\newline 2. The construction of \([\![\;.\;]\!]_N\) can be deformed into a filtred one \([\![\;.\;]\!]^f_N\) [\textit{B. Gornik}, ``Note on Khovanov cohomology, (2004); \url{arXiv:math.0402266}]. By looking at the grading of the surviving terms in the spectral sequence associated to the filtration, one defines a knot invariant, the \(\mathfrak{sl}_N\)--Rasmussen invariant \(s_N\), which is known to be a slice-torus invariant, that is a knot invariant giving a lower bound for the slice genus which is sharp for torus knots [\textit{A. Lobb}, Adv. Math. 222, No. 4, 1220--1276 (2009; Zbl 1200.57011)].\newline 3. There is an action by \(A:=\mathbb{C}[X]/X^N=[\![ U ]\!]_N\) on \([\![\;.\;]\!]_N\), induced by the connected sum with the unknot \(U\). By tensoring the chain complex underlying \([\![\;.\;]\!]_N\) by \(A/X\), one obtains \(\overline{[\![\;.\;]\!]}_N\), an invariant of links with a marked component, which categorifies the reduced \(\mathfrak{sl}_N\)--polynomial.\newline 4. Lastly, \(\overline{[\![\;.\;]\!]}_\infty\) is a triply graded homology categorifying the reduced HOMFLYPT polynomial [\textit{M. Khovanov} and \textit{L. Rozansky}, Geom. Topol. 12, No. 3, 1387--1425 (2008; Zbl 1146.57018)] which can be seen as a stabilization of all \(\overline{[\![\;.\;]\!]}_N\) since there are spectral sequences from the former to the latter which degenerate at the first page for \(N\gg0\) [\textit{J. Rasmussen}, ``Some differentials on Khovanov--Rozansky homology, (2006); \url{arXiv:math.0607544}]. Oddly enough, among all KR homologies, \(\overline{[\![\;.\;]\!]}_\infty\) is the most easily computed. The present paper fills the gap between reduced and unreduced worlds by setting a spectral sequence starting at \(\overline{[\![\;.\;]\!]}_N\otimes A\) and converging to \([\![\;.\;]\!]_N\). Then, using decategorified versions of all the spectral sequences mentioned above, the author gives lower and higher bounds for \(s_N\), obtained by evaluating the bi-degree of potential surviving terms in \(\overline{[\![\;.\;]\!]}_\infty\). These bounds, together with some slice-Bennequin inequalities that follow from Livingston's work on slice-torus invariants [\textit{C. Livingston}, Geom. Topol. 8, 735--742 (2004; Zbl 1067.57008)], allow the author to compute \(s_N\) for pretzel knots \(P(\ell,-m,n)\) with \(\ell>m\geq3\) odd integers and \(n\) an even positive integer. These computations are sufficient to prove that neither \(s_2\), \(s_3\) nor \(\tau\), a similar slice-torus invariant coming from Heegaard--Floer homology, is a linear combination of higher Rasmussen invariants, and that \(\tau\), \(s_2\), \(s_3\) and \(s_N\) are linearly independent for any given \(N\). The paper is organized as follows. Section 2 recalls known material about the distinct KR homologies. Definitions are actually not given, but it emphasizes the relevant properties, and particularly the known spectral sequences. Section 3 introduces a filtration on the chain complex underlying \([\![\;.\;]\!]_N\), which comes from the filtration of \(A\) by polynomial degrees and the action of \(A\) on the chain complex. The spectral sequence associated to this filtration relates \(\overline{[\![\;.\;]\!]}_N\otimes A\) to \([\![\;.\;]\!]_N\). In section 4, the author expresses all the mentioned spectral sequences as polynomial equalities relating the Poincaré polynomials of the different homologies. Starting from the surviving element in \([\![\;.\;]\!]^f_N\), this allows to climb up all the spectral sequences and bound \(s_N\) in term of bi-degrees of generators of \(\overline{[\![\;.\;]\!]}_\infty\) in homological degree 0. Section 5 recalls known material about slice-torus invariants and, in particular, it proves a sharper slice-Bennequin inequality defined combinatorially using Seifert graphs. In section 6, the author combines the results from the two previous sections to compute \(s_N\big(P(\ell,-m,n)\big)\) and proves the linear independency statements. The paper ends with a knot, namely \(12n_{340}\), for which \(s_N\) with \(n\geq4\) gives a sharper bound for the slice genus than \(s_2\) or \(s_3\).

Recommendations

Cites work

Cited in

(28)

Describes a project that uses

Uses Software

This page was built for publication: Rasmussen's spectral sequences and the \(\mathfrak{sl}_N\)-concordance invariants

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2452069)