Counting Invariant Curves: a theory of Gopakumar-Vafa invariants for Calabi-Yau threefolds with an involution (2203.12077v2)

Abstract: We develop a theory of Gopakumar-Vafa (GV) invariants for a Calabi-Yau threefold (CY3) $X$ which is equipped with an involution $\imath$ preserving the holomorphic volume form. We define integers $n_{g,h}(\beta) $ which give a virtual count of the number of genus $g$ curves $C$ on $X$, in the class $\beta \in H_{2}(X)$, which are invariant under $\imath$, and whose quotient $C/\imath$ has genus $h$. We give two definitions of $n_{g,h}(\beta) $ which we conjecture to be equivalent: one in terms of a version of Pandharipande-Thomas theory and one in terms of a version of Maulik-Toda theory. We compute our invariants and give evidence for our conjecture in several cases. In particular, we compute our invariants when $X=S\times \mathbb{C}$ where $S$ is an Abelian surface with $\imath (a)=-a$ or a $K3$ surface with a symplectic involution (a Nikulin $K3$ surface). For these cases, we give formulas for our invariants in terms of Jacobi modular forms. For the Abelian surface case, the specialization of our invariants $n_{g,h}(\beta) $ to $h=0$ recovers the count of hyperelliptic curves on an Abelian surface first computed by Bryan-Oberdieck-Pandharipande-Yin.