On the Pin(2)-equivariant monopole Floer homology of plumbed 3-manifolds (Q1654433)

\(\mathrm{Pin}(2)\)-Equivariant monopole Floer homology is a variant of the usual Seiberg-Witten Floer homology designed to make use of the additional data of a spin structure. It was originally defined by \textit{C. Manolescu} [J. Am. Math. Soc. 29, No. 1, 147--176 (2016; Zbl 1343.57015)] for a pair \((Y,\mathfrak{s})\) of a rational homology \(3\)-sphere \(Y\) equipped with a \(\mathrm{spin}\)-structure \(\mathfrak{s}\). For such \(Y\) there is a one-to-one correspondence between \(\mathrm{spin}\)-structures and self-conjugate \(\mathrm{spin}^c\)-structures, and the presence of the latter gives rise to an extra \(\mathbb{Z}_2\)-symmetry in the Chern-Simons-Dirac functional which Manolescu's construction is designed to exploit. \textit{C. Manolescu}'s actual definition of the \(\mathrm{Pin}(2)\)-homology follows his earlier work [Geom. Topol. 7, 889--932 (2003; Zbl 1127.57303)], and is defined as the Borel homology of a certain \(\mathrm{Pin}(2)\)-equivariant spectrum produced in [\textit{C. Manolescu}, Zbl 1343.57015]. On the other hand, another version of \(\mathrm{Pin}(2)\)-equivariant monopole Floer homology has been constructed by \textit{F. Lin} [A Morse-Bott approach to monopole Floer homology and the triangulation conjecture. Providence, RI: American Mathematical Society (AMS) (2018; Zbl 1446.57013)] which is more in the spirit of \textit{P. Kronheimer} and \textit{T. Mrowka}'s original construction [Monopoles and three-manifolds. Cambridge: Cambridge University Press (2007; Zbl 1158.57002)]. It extends Manolescu's theory to all closed \(3\)-manifolds equipped with a self-conjugate \(\mathrm{spin}^c\)-structure, and is functorial with respect to \(\mathrm{spin}^c\)-cobordisms. However, while the nonequivariant theories are understood to be isomorphic [\textit{T. Lidman} and \textit{C. Manolescu}, The equivalence of two Seiberg-Witten Floer homologies. Paris: Société Mathématique de France (SMF) (2018; Zbl 1415.57002)], at present Lin's and Manolescu's refined constructions are only conjecturally equivalent. In the present paper Dai computes the \(\mathrm{Pin}(2)\)-homology of a certain class of 3-manifolds first considered by \textit{P. Ozsváth} and \textit{Z. Szabó} [Geom. Topol. 7, 185--224 (2003; Zbl 1130.57302)]. More precisely, these are the \(3\)-manifolds obtained by plumbing according to a connected, negative-definite graph with at most one bad vertex in the sense of [loc. cit.]. Dai's computations run in Lin's framework, making use of the identification of monopole and Heegaard Floer homology [\textit{Ç. Kutluhan} et al., ``HF=HM I: Heegaard Floer homology and Seiberg-Witten Floer homology'', Preprint, \url{arXiv:1007.1979}] to allow him to introduce techniques from \textit{A. Némethi}'s lattice cohomology [Publ. Res. Inst. Math. Sci. 44, No. 2, 507--543 (2008; Zbl 1149.14029)]. Dai's insight is to notice that a certain composite of maps in Lin's Gysin sequence [\textit{F. Lin}, A Morse-Bott approach to monopole Floer homology and the triangulation conjecture. Providence, RI: American Mathematical Society (AMS) (2018; Zbl 1446.57013)] may be identified with a \(\mathbb{Z}_2\)-symmetry in the lattice cohomology. This observation provides enough rigidity to allow for effective computation of the \(\mathrm{Pin}(2)\)-homology. Following his main result Dai presents several sample computations. Although the results here are not new, they serve to illustrate the power of his theoretical framework. Another upside is that their inclusion helps strengthen the still conjectural equivalence between Manolescu's and Lin's viewpoints of \(\mathrm{Pin}(2)\)-equivariant monopole Floer homology. Indeed, Dai acknowledges an overlap with \textit{M. Stoffregen}'s work [Compos. Math. 156, No. 2, 199--250 (2020; Zbl 1471.57017)], where the \(\mathrm{Pin}(2)\)-homology of Seifert fibred spaces was studied from within Manolescu's formulation. The next section of the paper is devoted to showing that for \((Y,\mathfrak{s})\) as above the relation \(\beta(-Y,\mathfrak{s})=\overline\mu(Y,\mathfrak{s})\) holds, where \(\beta(-Y,\mathfrak{s})\) is \textit{C. Manolescu}'s correction term [Zbl 1343.57015] of the orientation reversal \(-Y\), and \(\overline\mu(Y,\mathfrak{s})\) is the Neumann-Siebenmann invariant [\textit{W. D. Neumann}, Lect. Notes Math. 788, 125--144 (1980; Zbl 0436.57002); \textit{L. Siebenmann}, ibid. 788, 172--222 (1980; Zbl 0444.57003)]. Such a relation was conjectured by Manolescu [loc. cit.], and here Dai confirms it in greater generality, albeit from within Lin's formulation of \(\mathrm{Pin}(2)\)-homology. The calculation itself is accomplished using algebraic considerations derived from the lattice cohomology. Dai finishes the paper by providing a view towards future applications of his theory to manifolds obtained by more complicated plumbings. In particular Dai shows by example that his methods are capable of understanding the \(\mathrm{Pin}(2)\)-homology of some manifolds obtained by plumbing connected, negative-definite graphs with two bad vertices [\textit{P. Ozsváth} and \textit{Z. Szabó}, Geom. Topol. 7, 185--224 (2003; Zbl 1130.57302)].