The rational homotopy groups of virtual spheres for rank 1 compact Lie groups

J.P.C.Greenlees · 2025

Read The rational homotopy groups of virtual spheres for rank 1 compact Lie groups by J.P.C.Greenlees (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-18085.txt.

Introduction

Greenlees calculates the rational homotopy groups of virtual spheres for rank one compact Lie groups in a 2025 mathematics paper. The groups covered include SU(2), SO(3), and several related circle and pin groups. The calculation grades the groups by representations of each group, following an earlier treatment for finite groups.

The work shows the power of algebraic models for these spectra categories. It also shows how a classic splitting result fails when applied to desuspensions in this setting. Greenlees notes the paper grew from a colleague's interest and thanks a Cambridge institute for support. The results give topologists concrete values where abstract machinery can be hard to test. The style is that of a specialist research note.

Catalog record

CREATORJ.P.C.Greenlees
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-18085

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The rational homotopy groups of virtual spheres for rank 1 compact Lie groups Thanks: The author is grateful to J.D.Quigley for his interest, which encouraged the author to write this down. The author would also like to thank the Isaac Newton Institute for Mathematical Sciences, Cambridge, for support and hospitality during the programme Equivariant Homotopy Theory in Context, where this paper was written. This work was supported by EPSRC grant EP/Z000580/1. J.P.C.Greenlees Abstract . We calculate the rational representation-ring-graded stable stems for rank 1 groups, S ​ U ​ ( 2 ) , S ​ O ​ ( 3 ) , P ​ i ​ n ​ ( 2 ) , O ⁡ ( 2 ) , S ​ p ​ i ​ n ​ ( 2 ) SU(2),SO(3),Pin(2),O(2),Spin(2) and S ​ O ​ ( 2 ) SO(2) , in the same spirit as the calculations for finite groups in [ 6 ] . 1. Context There are now algebraic models for rational G -spectra for many compact Lie groups G . The models are designed to make calculation straightforward, and the purpose of this note is to show this in practice. Following [ 6 ] for finite groups, we have chosen the test case of calculating the rational homotopy of all virtual spheres, this time for groups of rank 1: the calculation is again entirely reduced to real representation theory. 1.A. Groups The title is a slight misnomer, since we will concentrate on the 6 most obvious rank 1 compact Lie groups. However, in the light of [ 5 ] this can easily be adapted to any rank 1 group. In any case these groups all come from the semisimple rank 1 Lie algebra. So there are two forms, the simply connected form SU(2) and the adjoint form SO(3) . For each of those we consider also the maximal torus and its normalizer. Altogether we adopt the following abbreviations \textstyle{SU(2)\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{SO(3)} \textstyle{\widetilde{\Gamma}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\Gamma} \textstyle{Pin(2)\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{O(2)\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\widetilde{G}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{G\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{Spin(2)\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{SO(2)\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\widetilde{SG}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{SG\ignorespaces\ignorespaces\ignorespaces\ignorespaces} In some sense G=O(2) is the core case, which helps explain the naming scheme. In all cases the double cover is indicated by a tilde. Although Spin(2) and SO(2) are both copies of the circle group, one needs to remember the quotient map between them is of degree 2 and does not induce a bijection on subgroups. 1.B. Rational Throughout, without further explicit mention, all spectra are rational. 1.C. Homotopy The point of this note is to outline the calculation of [S^{0},S^{U}]^{H}_{*} when H is one of the 6 rank 1 groups we named above, and where U^{H}=0 . Throughout we write \Gamma=SO(3),G=O(2),SG=SO(2) and W=G/SG ; the double cover will be indicated by a tilde. 1.D. Contents The key input is for the circle group, so we begin with that in Section 2 . Building on this, we move to O(2) , with general notation in Section 3 and then the cyclic and dihedral parts in the following two sections. The group SO(3) only involves noting fusion and the change in representation theory: this is dealt with in Section 6 . Finally we turn to the double covers, which are minor variants. We deal with Pin(2) in Section 7 and with SU(2) in Section 8 . 2. The group SO(2) In this section the ambient group is SG=SO(2) . It is reasonable to do this first, since it is the ingredient with the most substance. All the calculations in this section can be made with [ 3 ] , since the Adams spectral sequence collapses. 2.A. Subgroups The set of finite subgroups of SG is \mathcal{C}=\{C_{1},C_{2},\ldots\} , and we will use the letter s to index these groups C_{s} . The space of conjugacy classes of subgroups \mathfrak{X}_{SG}=\mathrm{Sub}(SG)/SG with the Hausdorff metric topology is the one point compactification \mathcal{C}^{\#} , with SG as the compactifying point. 2.B. Representations The trivial representation is real, but all other representations z^{n} are complex, so we write U=a+\bigoplus_{n\neq 0}a_{n}z^{n} . In our case a=0 because U^{SG}=0 . We just need to remember the dimension function d_{U}:\mathcal{C}\longrightarrow\mathbb{Z} . Thus d_{U}(s)=\dim_{\mathbb{R}}(U^{C_{s}})=a+2\sum_{s|n}a_{n}. 2.C. Notation We write • \mathbb{Q}[c]=H^{*}(BSG) • \mathbb{Q}[c,c^{-1}] • 0\longrightarrow\mathbb{Q}[c]\longrightarrow\mathbb{Q}[c,c^{-1}]\longrightarrow I\longrightarrow 0 so that I=\Sigma^{2}\mathbb{Q}[c]^{*}. • \mathcal{O}_{\mathcal{C}}=\prod_{s}\mathbb{Q}[c] • \mathcal{E}=\langle c^{v}\;|\;v:\mathcal{C}\longrightarrow\mathbb{N}\mbox{ zero a.e. }\rangle • t=\mathcal{E}^{-1}\mathcal{O}_{\mathcal{C}} • \mathbb{I}=\bigoplus_{s}I_{s} • 0\longrightarrow\mathcal{O}_{\mathcal{C}}\longrightarrow t\longrightarrow\mathbb{I}\longrightarrow 0 2.D. The standard model The standard model \mathcal{A}(SG) is an abelian category which provides an algebraic model for rational SG -spectra in the sense that there is a Quillen equivalence [ 7 ] : \mbox{$SG$-spectra}\simeq\mbox{DG-$\mathcal{A}(SG)$} Since all objects are formal we’ll work with \mathcal{A}(SG) , but use the notation familiar from SG -spectra. The objects of \mathcal{A}(SG) are (N\longrightarrow t\otimes V) where N is a \mathcal{O}_{\mathcal{C}} -module, V is a graded rational vector space and the map N is inverting \mathcal{E} . This should be thought of as N (the ‘nub’) together with the rigidification that the vector space V gives a ‘basis’ of \mathcal{E}^{-1}N . Morphisms (N\longrightarrow t\otimes V)\longrightarrow(N^{\prime}\longrightarrow t\otimes V^{\prime}) are given by compatible pairs \theta:N\longrightarrow N^{\prime} and \phi:V\longrightarrow V^{\prime} . The category \mathcal{A}(G) has injective dimension 1 and there is a short exact sequence 0\longrightarrow\mathrm{Ext}(\Sigma H_{*}(X),H_{*}(Y))\longrightarrow[X,Y]^{SG}\longrightarrow\mathrm{Hom}(H_{*}(X),H_{*}(Y))\longrightarrow 0. This is a symmetric monoidal category with tensor unit S^{0}=(\mathcal{O}_{\mathcal{C}}\longrightarrow t\otimes\mathbb{Q}) One checks that if T is a torsion \mathcal{O}_{\mathcal{C}} -module in the sense that \mathcal{E}^{-1}T=0 then T=\bigoplus_{s}T_{s} and T_{s} is a torsion \mathbb{Q}[c] -module fore all s . We then define \Sigma^{U}T=\bigoplus_{s}\Sigma^{d_{U}(s)}T_{s}. More generally we may define \Sigma^{U}N for any \mathcal{O}_{\mathcal{C}} -module. This relies on the fact that d_{U} is zero almost everywhere. Thus if \mathcal{C}_{d} denotes the subgroups on which d_{U} takes the value d , \mathcal{C}_{0} is cofinite and the partition is finite. We may define \Sigma^{U}N=\prod_{d}\Sigma^{d}e_{\mathcal{C}_{d}}N. The objects e(V)=(t\otimes V\longrightarrow t\otimes V) and f(T)=(T\longrightarrow 0) are useful ( T must be torsion so that f(T) lies in \mathcal{A}(SG) ). If X=(M\longrightarrow t\otimes U) then \mathrm{Hom}(X,e(V))=\mathrm{Hom}(U,V),\mathrm{Hom}(X,f(T))=\mathrm{Hom}(M,T) Thus e(V) is injective and if T is a torsion injective f(T) is injective. Thus we have an injective resolution 0\longrightarrow S^{0}\longrightarrow e(\mathbb{Q})\stackrel{{\scriptstyle i}}{{\longrightarrow}}f(\mathbb{I})\longrightarrow 0. By suspension, we have S^{U}=(\Sigma^{U}\mathcal{O}_{\mathcal{C}}\longrightarrow t\otimes\mathbb{Q}) and an injective resolution 0\longrightarrow S^{U}\longrightarrow e(\mathbb{Q})\stackrel{{\scriptstyle i}}{{\longrightarrow}}f(\Sigma^{U}\mathbb{I})\longrightarrow 0. We thus have \textstyle{\mathrm{Hom}(S^{0},e(\mathbb{Q}))\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{i_{*}} \textstyle{\mathrm{Hom}(S^{0},f(\Sigma^{U}\mathbb{I}))\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\mathrm{Hom}(\mathbb{Q},\mathbb{Q})\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\mathrm{Hom}(\mathcal{O}_{\mathcal{C}},\Sigma^{U}\mathbb{I})\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\mathbb{Q}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\Sigma^{U}\mathbb{I}} Lemma 2.1 . The vector space (\Sigma^{U}\mathbb{I})_{0} is finite dimensional: it is a sum with one copy of \mathbb{Q} for each s with d_{U}(s)<0 . The connecting map \mathbb{Q}\longrightarrow\Sigma^{U}\mathbb{I} is the diagonal. In particular it is zero if d_{U}(s)\geq 0 for all s and it is nonzero if there is some s with d_{U}(s)<0 . Corollary 2.2 . In even degrees [S^{0},S^{U}]^{SG}_{*} is \mathbb{Q} if d_{U}\geq 0 and it is 0 if d_{U} takes a negative value. In odd degrees [S^{0},S^{U}]^{SG}_{*} is \Sigma^{U-1}\mathbb{I} if d_{U}\geq 0 and it is the quotient of this by the diagonal \Delta_{-1}\cong\mathbb{Q} in degree -1 if d_{U} takes a negative value. 3. The group O(2) In Sections 3 , 4 and 5 , the ambient group is G=O(2) . In this section we introduce the subgroups and representations and in Sections 4 and 5 we describe the two summands of [S^{0},S^{U}]^{G}_{*} . All the calculations in this section can be found made with [ 2 ] , since the Adams spectral sequence collapses. 3.A. Subgroups in blocks We describe the space \mathfrak{X}_{G}=\mathrm{Sub}(G)/G of subgroups of G=O(2) up to

Read the complete text