On the Ozsváth-Szabó $d$-invariants for almost simple linear graphs
Tatsumasa Suzuki
TL;DR
This work refines the Karakurt-Şavk computation of the Ozsváth-Szabó $d$-invariant for Brieskorn spheres $\Sigma(p,q,r)$ with almost simple linear graphs, providing an explicit max-formula for odd $p$ via a quadratic form $F_{p,q}$ and a refined index $D(p,q,r)$. It establishes parity- and growth-dependent criteria for when $d(\Sigma(p,q,r))$ equals $D(p,q,r)$ and constructs infinite families where equality holds or fails, yielding new infinite summands in the 3-dimensional homology cobordism group $\Theta_{\mathbb{Z}}^3$ and informing knot concordance through $\mathcal{C}_{TS}$. The paper also derives a parity-independent inequality guiding $D(p,q,r)$ across families, analyzes the Fibonacci and other infinite classes, and demonstrates concrete implications for braids, pretzel knots, and Dehn surgery realizations. Overall, it deepens the link between 3-manifold invariants and knot concordance, providing tools to generate and distinguish infinite families of non-cobordant Brieskorn spheres and their associated concordance classes.
Abstract
Karakurt and Şavk calculate the Ozsváth-Szabó $d$-invariant for Brieskorn homology $3$-spheres where surgery diagrams can be expressed as almost simple linear graphs. In this paper, we introduce a refiniment of Karakurt and Şavk's formula for these $d$-invariants. We also introduce infinite examples of inequalities that appear in this refinement in which the equality holds, and infinite examples in which it does not hold. We also discuss an application to knot concordance group.
