BPS cohomology for rank 2 degree 0 Higgs bundles (and more) (Q6054787)

In this work, the author calculates the cohomology of moduli spaces of rank 2, degree 0 semistable Higgs bundles, extending the investigation to a broad class of 2-Calabi-Yau (2CY) settings. The paper presents a detailed formula that compares the E-series of the moduli stacks of these Higgs bundles in genus \(g \geq 2\) with the intersection E-polynomials of their coarse moduli spaces. This comparison is not only restricted to Higgs bundles but is also shown to be valid in diverse 2CY contexts such as sheaves on \(K3\) surfaces and preprojective algebras of \(g\)-loop quivers. A substantial part of this study focuses on validating a conjecture by Ben Davison concerning the BPS cohomology of Higgs bundles. The BPS cohomology sheaf, as introduced in [\textit{B. Davison}, J. Reine Angew. Math. 804, 105--154 (2023; Zbl 07761116)], aims to categorify BPS invariants. This framework was initially developed to explore the categorical PT/DT correspondence and has since been extended to other BPS invariants, such as Gopakumar-Vafa invariants [\textit{Y. Toda}, J. Differ. Geom. 123, No. 1, 141--193 (2023; Zbl 07690280)] and the P=W conjecture [\textit{B. Davison}, Adv. Math. 415, Article ID 108889, 61 p. (2023; Zbl 1516.14026)]. The rationale behind defining BPS sheaves stems from scenarios where the quotient stack is a quotient of the moduli space by \(\mathbb{G}_m\). In such cases, the pushforward of a mixed Hodge module on the moduli stack to the moduli space can be explicitly described as generated by the element defined by the first perverse cohomology. Generally, by comparing the pushforward of intersection cohomology from a stack to a scheme, Davison postulated that one could explicitly define a perverse sheaf as a generator; this conjectural sheaf is termed the BPS sheaf. Furthermore, the compactly supported cohomology of this sheaf is referred to as BPS cohomology. This paper conducts a detailed analysis of the E-polynomial of the compactly supported cohomology of the moduli stack and, through examination of the stratification, establishes the specific compactly supported cohomology for the rank 2 degree 0 moduli stack of Higgs bundles relative to the corresponding moduli space. Thus, it confirms that Davison's conjecture may be true for moduli of semistable objects in the 2CY category, at least topologically. The structure of the paper is organized as follows: Section 1 provides an overview of the BPS cohomology sheaf by Ben Davison, emphasizing the main results on compactly supported cohomology and its implications for the existence of the BPS sheaf. Section 2 delves into the E-polynomial and associated concepts. Section 3.1 examines semistable objects within a 2-CY category, utilizing quivers to stratify the moduli stack. Section 3.2 focuses on strictly semistable objects, categorizing the remaining cases in Sections 3.3 and 3.4, which cover particular non-split semistable objects and polystable objects, respectively, culminating in a complete formula for the E-polynomials. Section 4 applies the findings from Section 3 to Higgs bundles, presenting an explicit calculation of the E-polynomial for the moduli space. The author's rigorous algebraic geometry computations provide profound insights into the topology of moduli stacks and spaces. This paper is set to be an invaluable resource for researchers interested in the topology of moduli spaces for Higgs bundles and sheaves defining 2CY categories, offering significant theoretical and practical contributions.