ON CERTAIN SUBGROUPS OF E8(2) AND THEIR BRAUER CHARACTER TABLES
A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Science and Engineering
2018
Peter J Neuhaus School of Mathematics Contents
Abstract 5
Declaration 6
Copyright Statement 7
Acknowledgements 8
1 Introduction 9
2 Background Material 16 2.1 MaximalSubgroupsofAlgebraicGroups ...... 19
3 Preliminary Results 23
3.1 E8(2)...... 23 3.1.1 Maximal Subgroups ...... 23
3.1.2 Elements of E8(2) ...... 26 3.2 Brauer Characters ...... 32 3.2.1 Conway Polynomials ...... 33 3.2.2 Brauer Character of L(G)...... 34 3.2.3 Feasible Decompositions ...... 35 3.3 Irreducible Modules of Groups of Lie Type ...... 36 3.3.1 DimensionsofHighestWeightModules ...... 38 3.4 Determining Maximality ...... 39
4 Proof of Theorems 1.2 and 1.3 45
4.1 L2(7)...... 45
2 4.2 L2(11) ...... 46
4.3 L2(13) ...... 46
4.4 L2(17) ...... 46
4.5 L2(25) ...... 47
4.6 L2(27) ...... 48
4.7 L4(3)...... 49
4.8 U3(3)...... 49
4.9 G2(3)...... 49
4.10 M11 ...... 49
4.11 M12 ...... 50
4.12 L3(8)...... 50
4.13 F4(2)...... 50 2 4.14 F4(2)0 ...... 50 3 4.15 D4(2)...... 51
4.16 L3(3)...... 51 4.16.1 Case (ii) ...... 51 4.16.2 Case (i) ...... 52
4.17 L3(4)...... 56
4.18 U3(8)...... 57 4.18.1 Case (ii) ...... 57 4.18.2 Case (i) ...... 58
4.19 U3(16) ...... 61
5 Proof of Theorem 1.4 62 5.1 Methodology ...... 62
5.2 L2(128) ...... 66
5.3 L2(32) ...... 67 5.3.1 P(1234) ...... 67 5.3.2 P(1345) ...... 68 5.3.3 P(2456) ...... 68 5.3.4 P(3456) ...... 69 5.3.5 P(4567) ...... 69
3 5.3.6 P(5678) ...... 70
6 Brauer Character Tables 71
7 Further Work 123
Bibliography 124
APrograms 130 A.1 FeasibleCharacterCode ...... 130 A.2 IndecomposableModuleCode ...... 135
N A.3 L2(2 )code...... 138
B Supplementary Code 149
B.1 E8(2) setup ...... 149
4 The University of Manchester
Peter J Neuhaus Doctor of Philosophy On Certain Subgroups of E8(2) and their Brauer Character Tables December 4, 2018
For the exceptional group of Lie type E8(2) a maximal subgroup is either one of a known set or it is almost simple. In this thesis we compile a complete list of almost simple groups that may have a maximal embedding in E8(2) and in many cases it is proved that such an embedding does not exist. For the groups L2(32) and L2(128) we go further and find all conjugacy classes of their embeddings in E8(2). Extensive use is made of the theory of Brauer characters and modular representation theory, and as such include Brauer character tables in characteristic 2 for many small rank simple groups. The work in this thesis relies heavily on the computer package Magma and includes a collection of useful procedures for computational group theory. The results presented are the author’s contribution to the ongoing attempt to classify the maximal subgroups of E8(2).
5 Declaration
No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.
6 Copyright Statement
i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes.
ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trade marks and other intel- lectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. iv. Further information on the conditions under which disclosure, publication and com- mercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=487), in any rele- vant Thesis restriction declarations deposited in the University Library, The Univer- sity Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regul- ations) and in The University’s Policy on Presentation of Theses.
7 Acknowledgements
First and foremost, I would like to thank Peter Rowley, for his guidance throughout my time at Manchester, his invaluable expertise and mathematical advice, as well as his patience and understanding. Without him this thesis would certainly not exist.
I also wish to thank the other members of my research group; John Ballantyne for sharing his expertise in Magma so clearly, Jamie Phillips and David Ward for showing me the ropes and making me feel so welcome when I was just starting, and Alexander McGaw, for being a constant friend throughout, willing to be a sounding board for new ideas and frustrations alike.
IwouldliketothankMatthewGwynneforhisfriendshipandhelpinnavigating the stresses that a PhD entails.
Finally, I would like to thank my parents, without whose unwavering support and belief I could not have achieved any of what I have done.
8 Chapter 1
Introduction
When studying finite groups and their properties there is a natural desire to obtain classifications of the objects in question, and in the case of simplicity this has been achieved by the classification of finite simple groups. This remarkable result was the work of countless mathematicians over more than a century, starting with Galois in 1830 and generally being considered to be finally complete with two volumes by As- chbacher and Smith in 2004 [5]. Whilst this classification provides constructions for all the groups, for example see Wilson [58], there is still much that is not known about their properties.
For almost as long as they have been studied, the structure of the maximal sub- groups of these finite simple groups has been of great interest, and whilst a complete classification has not been achieved much progress has been made. In 1901 the max-
a imal subgroups of L2(q), for q = p a power of a prime, were classified by Dickson
[20], and this was followed up by the classification of maximal subgroups for L3(q) by Mitchell [53], for q odd, and Hartley [29], for q even. Since then, after extensive study, the maximal subgroups of the finite simple groups have either been classified or categorised with the exception of the Monster group.
The categorising of maximal subgroups was first achieved by the O’Nan-Scott the- orem, initially proved independently by O’Nan and Scott; however, due to the impor- tance of the result, there are now many proofs, for example by Aschbacher [3]. In this theorem it is proved that any maximal subgroup of a symmetric group is one
9 CHAPTER 1. INTRODUCTION 10 of five well understood types, or is almost simple. This is a common theme in the classification of maximal subgroups and is mirrored by Liebeck, Praeger and Saxl in their classification of the maximal subgroups of the finite (simple) alternating groups [40]. In 1984 Aschbacher [2] proved a similar theorem for the finite simple groups of classical Lie type.
The sporadic simple groups can have full classifications as they are finite in number. They have been tackled in a more piecemeal fashion across many papers by several authors, although the majority can be attributed to Wilson. For the classifications of maximal subgroups of sporadic simple groups one can read the following papers; M24
[15] [19], J1 [32], J2 [25], J3 [26], J4 [38], Co1 [59], Co2 [60], Co3 and McL [24], Fi22
[61], Fi23 [37], Fi240 [47], HS [52], He [13], Ru [62], Suz [63], O0N [64], HN [54], Ly [65], Th [46] and B the baby monster group [66]. The last of these is of particular interest as it was achieved using a 4370-dimensional GF (2)B-module. There are five sporadic groups not yet covered, M11, M12, M22, M23 and M, the monster group. For the first four of these there seems to be no paper giving their maximal subgroups, although the information is well known and can be found easily using modern com- putational methods. The last, M, is the only sporadic group for which the maximal subgroups are as yet unknown, with the current state of a↵airs described by Wilson [67].
This leaves just the groups of exceptional Lie type to cover, and here progress is more recent. Extensive work by Liebeck and Seitz [44], among others, has yielded a categorisation of the maximal subgroups of the simple exceptional groups of Lie type.
Maximal subgroups for G2(q)wereclassifiedin1981byCooperstein[17]forq even and in 1988 by Kleidman [35] for q odd. The only other families of exceptional groups
3 whose maximal subgroups have been classified are D4(q), also by Kleidman [36], and
F4(q)forqnotapowerof2or3byMagaard[51].Thereareahandfulofcomparatively small exceptional groups outside of these families for which the maximal subgroups have been classified, the first of which is F4(2) by Norton and Wilson in 1989 [55].
Kleidman and Wilson classified the maximal subgroups of E6(2) in the same year [39] 2 and Wilson published a proof for E6(2) [68], although the classification had been CHAPTER 1. INTRODUCTION 11
known for some time. Finally, the maximal subgroups of E7(2) were classified in 2015 by Ballantyne, Bates and Rowley [8].
This list of finite simple groups whose maximal subgroups have been classified has aparticulargroupwhichisconspicuousbyitsabsence,namelyE8(2). E8(2) is the only exceptional group of Lie type defined over GF (2) without a classification of max- imal subgroups, a gap that will hopefully be filled by an as yet unpublished paper of Aubad, Ballantyne, McGaw, Neuhaus, Rowley and Ward [6]. This thesis details the contribution of the author as part of that work to determine non-maximality of cer- tain subgroups of E8(2), as well as determining the conjugacy classes of some of those subgroups. Much of the work done has been computer assisted, making extensive use of the functionality provided by the computer algebra package Magma.
In Chapter 2 we cover the background material required for the work described in this thesis. This starts with a general definition of algebraic groups and some of their properties, including defining the Lie algebra of an algebraic group, maximal tori and through them the root system and associated weights. We also include the Frobenius morphisms necessary for the definition of the Chevalley groups as our context for E8(2).
Chapter 2 continues by exploring work of Liebeck and Seitz in first classifying the maximal subgroups of positive dimension in exceptional algebraic groups and then in producing a list of groups in which any maximal subgroup of a simple exceptional group of Lie type must be contained. The chapter ends with a summary of this work in the form of Theorem 2.3, a very important result for all that follows as it guides our analysis and makes a case by case approach to the problem possible.
In the four sections of Chapter 3 we give some preliminary results which are to be of use in this thesis. In Section 3.1 we turn to E8(2) and collect a variety of lists and theorems detailing the information already known about E8(2) which we will use in our analysis. We use Theorem 2.3 to produce an exhaustive list of maximal closed -stable subgroups of positive dimension along with a list, Tables 3.1.1 and 3.1.2, of those almost simple groups whose potential embeddings into E8(2) forms the focus of CHAPTER 1. INTRODUCTION 12 our study. Also included here are several preliminary results that eliminate many of the groups within that list for which there are no embeddings.
We then turn to what is known about the conjugacy classes of E8(2). The cen- tralisers and fixed point dimensions of the involutions and semisimple elements were found by Rowley et al. [7], and we detail those results in Table 3.1.4. This leads to another set of groups from Tables 3.1.1 and 3.1.2 for which it is proven there can be no embeddings of them into E8(2) by consideration of element orders and centraliser sizes. A brief summary of some useful procedures is also included here.
In Section 3.2 we describe some of the theory of modular representations and Brauer characters. This will be of great importance in the calculations detailed in Chapter 6. To complement this we give some definitions of feasible decompositions and fusion patterns as well as much of the Brauer character of E8(2) on its Lie algebra L(G). In Section 3.3 we show how similar calculations can be done for the relevant groups from Tables 3.1.1 and 3.1.2. This requires the theory of dominant weights and is based on work by Steinberg [57] and L¨ubeck [49]. We also include here the final part of the proof of Theorem 1.1.
Finally, in Section 3.4 we give a method developed by Alistair Litterick and David Craven [48] [18] for determining maximality in groups of Lie type. This method builds from a very important result, Lemma 3.2, by Seitz [56], which allows us to often re- duce the problem to one of finding fixed vectors of L(G). Chapter 3 ends with a set of lemmas needed for the implementation of the method.
Chapters 4 and 5 contain the main results of the thesis. Chapter 4 details the results of Theorems 1.2 and 1.3 on a case-by-case basis. The methods used in Chapter 4 are mainly those detailed in Chapter 3 and where exceptions occur the method is described in situ. The groups in Theorem 1.4, are the focus of Chapter 5 in which we describe the method and Magma functionality required for their analysis. It then proceeds to detail the results of Theorem 1.4 also on a case by case basis. CHAPTER 1. INTRODUCTION 13
In Chapter 6 we give, for all groups we have encountered, the Brauer character tables along with all feasible decompositions of L(G) and the dimension of H1(S, V ) for their irreducible modules. The tables in this chapter are consistent with the atlas [16] except where noted and have been calculated using the methods from Chapter 3Section3.3.TherearealsotwoappendicescontainingMagma code used for the calculations in this thesis.
Our results can be described by the following four theorems, each summarising a collection of groups from Tables 3.1.1 and 3.1.2 and what is known of their embeddings in E8(2).
Theorem 1.1. Let G = E (2) and H G. Then H cannot be isomorphic to any of ⇠ 8 the following groups:
L2(29) L2(37) L2(41) L2(49) L2(61)
L2(256) L2(512) L2(1024) L3(16) L4(5)
L4(8) U4(8) U3(16) G2(8) F4(4) 2 3 F4(8) F4(8) D4(8) Sp6(8) Sp8(4) + + Sp8(8) ⌦8 (4) ⌦8 (8) ⌦8 (8) Sz(32) Sz(128) Sz(512).
Theorem 1.2. Let G = E (2) and H G such that H is maximal. Then H cannot ⇠ 8 be isomorphic to any of the following groups:
L2(7) L2(13) L2(17) L2(27) L4(3)
L3(8) U3(8) G2(3) U3(3) F4(2) 2 3 F4(2)0 D4(2) M11 M12.
Theorem 1.3. Let G = E (2) and H G such that either H = L (3) or H = L (4). ⇠ 8 ⇠ 3 ⇠ 3 Then up to isomorphism the restriction M = L(G) is known and has a direct sum |H decomposition of one of the following forms: CHAPTER 1. INTRODUCTION 14
H = L (3) M = 144 104 ⇠ 3 ⇠ d
H = L (4) M = 20 20 ⇠ 3 ⇠ 4 5 5 5 M = 20 36 ⇠ b 5 5 5 M = 56 . ⇠ b 5 5 5
The modules given here are denoted by their dimension if they are not irreducible and by their name in the relevant character table if they are. Their structure shall be explained in Chapter 4.
Theorem 1.4. Let G ⇠= E8(2) and M either L2(32) or L2(128). Then there is no H G with H maximal such that H = M. Furthermore there are the following upper ⇠ bounds for the number of conjugacy classes, nc, of subgroups H ⇠= M.
L (32) nc 5, 200, 000 2 L (128) nc 8. 2
We note that the upper bounds given in Theorem 1.4 are not very close upper bounds; however, it is not unreasonable to hope that future work will improve these significantly. We also note that all the results of Theorems 1.1-1.4 concern the simple groups in Tables 3.1.1 and 3.1.2 but take no account of their automorphism groups. This will be dealt with in [6] and so is beyond the scope of this thesis. CHAPTER 1. INTRODUCTION 15 Index of Symbols
Throughout this thesis we use atlas notation for the simple groups with the exception that we use ⌦n±(q)forthesimpleorthogonalgroups.
Symbol Meaning
VX ( ), Irreducible KX-module of high weight (X reductive) L(X) Lie algebra of X (X an arbitrary algebraic group) V W Direct sum of V and W V + W Module having the same factors as V W X/Y Quotient of X by Y V V ... V Module with socle series V , V /V , ..., V /.../V /V 1| 2| | n 1 2 1 n 2 1 nV V + V + + V (n times) ··· V ⇤ Dual module Hom(V,K) V Restriction of V to the subgroup H |H G Fixed points of in G Z(G) Center of G
p O 0 (G) Minimal normal subgroup of p0 index (G finite) G0 Connected component of the algebraic group G
F ⇤(H) Fitting subgroup of H GF (q) Finite field of q elements
P ( i) Projective cover of the module i Chapter 2
Background Material
Before we start looking at E8(2) specifically, we first recall some general theory of a ne algebraic groups; for detailed coverage of this material we suggest reference materials of Humphreys [30] and Borovik[9]. For this section K denotes an algebraically closed field of characteristic p 0.
An a ne algebraic group over K is a subgroup of GLn(K)forsomeintegern>0 defined by zeros of a collection of finitely many polynomials in the matrix coordinates x , i.e. closed in the Zariski topology. It inherits many properties from GL (K) { ij} n including the Zariski topology and a coordinate ring K[G] as a quotient of K[ x ]by { ij} the ideal of functions vanishing on G. If G is an irreducible variety the we can define the dimension of G to be the transcendence degree of K[G]:G.
We are frequently interested in closed subgroups of an algebraic group G and there are many obvious classes of these [30, 8]. The kernel and image of an algebraic § homomorphism are closed subgroups, where an algebraic homomorphism is a group homomorphism that is also a morphism of varieties. The stabiliser of a vector or sub- space in a rational G-module is a closed subgroup; a rational G-module, V , being a module such that the representation G GL(V ) is an algebraic homomorphism. An ! algebraic group is unipotent if all its elements are unipotent, which is equivalent to it being isomorphic to a group of upper triangular matrices. A semisimple (or reduc- tive) algebraic group is one for which the radical R(G)(orunipotentradicalRu(G)) is trivial. These radicals are the unique largest normal, connected, soluble, (unipotent)
16 CHAPTER 2. BACKGROUND MATERIAL 17 subgroup of G.
Another important property of an algebraic group is its Lie algebra L(G). Many algebraic groups were first studied as the automorphism groups of semisimple Lie alge- bras, see Carter [14]. The study of semisimple algebraic groups and their Lie algebras mirrors that of the semisimple Lie algebras themselves. We can view L(G)asthe space of left-invariant derivations of K[G], which leads to the characterisation of L(G) as the adjoint module, a rational G-module with the same dimension as G. Just as in Carter [14, p.34], this Lie algebra comes equipped with a symmetric bilinear form called the Killing form.
There are two further very important closed subgroups to introduce: a maximal torus T and a corresponding Borel subgroup B. A torus is a connected algebraic group consisting of semisimple elements and it is simultaneously diagonalisable [30, 15] and § therefore isomorphic to a product of the multiplicative group K⇤.Amaximaltorusis atorusthatismaximalinthesetofalltori.IfT is a fixed maximal torus in G then every semisimple element of G is conjugate to an element in T [28], and from this there are several useful consequences. We get that all maximal tori in G are conjugate and have the same dimension, which defines the rank of G. For a given choice of maximal torus T we also get the Weyl group W = NG(T )/T . A Borel subgroup is a maximal closed, connected, solvable algebraic subgroup of G.
In Carter’s exposition of groups of Lie type [14] the root system is inherited from the Lie algebra of a group. An abstract root system E, for E afinite-dimensional ⇢ Euclidian vector space, is a finite set of vectors satisfying the following four conditions: i) is a spanning set for E, ii) for ↵, ↵ then = 1, iii) for any ↵, the 2 ± 2 reflection ( ):= –2((↵, )/(↵, ↵))↵ and iv) for any ↵, the number ↵ 2 2 ↵, := 2(↵, )/(↵, ↵) Z. Sometimes condition iv) is not required and a root sys- h i 2 tem that does satisfy it is known as crystallographic. A root system can be defined for any reductive (and hence semisimple) algebraic group in such a way that in con- curs with the root system given by Carter wherever relevant. To do this we need the concept of a weight of a module. CHAPTER 2. BACKGROUND MATERIAL 18
Let G be a reductive algebraic group over K, T amaximaltorusofG and the character group of T ; X(T ) ⇠= Hom(T,K⇤). For any G-module V the restriction of V to T can be decomposed into subspaces called weight spaces. These weight spaces are defined for some X(T )asV := v V : t v = (t)v for all t T .A 2 { 2 · 2 } weight, ,ofV (with respect to T )isanelementofX(T )suchthatV = 0 and the 6 { } multiplicity of is the dimension of V .
We now get our required root system as the non-zero weights of the Lie algebra
L(G). For a root i and its weight space V i we can define a root subgroup, U i ,asthe unique connected unipotent subgroup of G normalised by T such that L(U i )=V i . For any maximal torus T and choice of simple roots there is a unique Borel subgroup generated by T and the root subgroups U for +. i i 2
A simple algebraic group is a semisimple algebraic group whose root system is ir- reducible. The irreducible root systems were classified by Dynkin [21] and when G is semisimple Z(G)isfiniteandG/Z(G) is determined up to isomorphism by the root system of G. There are two (sometimes isomorphic) important semisimple algebraic groups for each irreducible Dynkin diagram, the adjoint and simply-connected groups. These are defined by the character group of their torus X(T ). Given a root system for G we have the Euclidean space E = X(T ) R in which there exists a lattice ⌦Z of weights ⇤ defined as x E such that x, ↵ Z for all ↵ +. We then have 2 h i2 2 Z X(T ) ⇤ and ⇤/Z is finite. The simply-connected and adjoint subgroups are the extremes of these inclusions, where G is simply-connected when X(T )=⇤ and adjoint when X(T )=Z .
If G is a semisimple algebraic group and ' an algebraic homomorphism from G G ! that is an automorphism of G as an abstract group, then there are two types of pos-
1 sibilities for '. If ' is also an algebraic homomorphism then it is known as an algebraic automorphism. Otherwise by Steinburg [57] the fixed-point subgroup G' is finite and ' is known as a Frobenius morphism. Inner automorphisms are always al- gebraic automorphisms, as are graph automorphisms, defined as those automorphisms CHAPTER 2. BACKGROUND MATERIAL 19 induced from symmetries of the Dynkin diagram. Turning our attention to Frobenius morphisms, they exist, induced by field automorphisms in positive characteristic p for each power q of p defined by sending K x xk. When is a Frobenius morphism 3 ! the fixed-point subgroup G is a group of Lie type and when G is simple of adjoint
p0 type then O (G ) is usually simple. We denote by Aut(G)thegroupgeneratedby the inner automorphisms of G along with graph and field morphisms.
2.1 Maximal Subgroups of Algebraic Groups
The study of the subgroup structure of a finite algebraic group G G splits into two types; Lie primitive where H G is contained in no proper positive-dimensional subgroup of G, and non Lie primitive otherwise. Knowledge of the maximal positive dimensional subgroups of arbitrary algebraic groups is therefore useful to the study of finite subgroups.
A classification of the subgroups of positive dimension started with Dynkin who classified the maximal connected subgroups of a classical group [22]. He also gave a classification of the maximal Lie subalgebras of exceptional complex Lie algebras [23]. This, combined with Humphrey’s result [30, 13] that in characteristic zero there is a § 1-1 correspondence between the Lie subalgebras of L(G)andtheconnectedsubgroups of an algebraic group, gives an equivalent result for exceptional algebraic groups. Since then work by Seitz and Liebeck [56] [42] [43] has extended this classification to positive characteristic and we now have a classification of maximal (not necessarily connected) subgroups of algebraic groups in any characteristic.
Theorem 2.1 (Liebeck Seitz). Let G be an exceptional algebraic group and G G 1 Aut(G). Let M be a proper closed connected subgroup of G which is maximal among proper closed connected NG1 (M)-invarient subgroups of G. Then one of the following holds:
(i) M is either parabolic or of maximal rank;
(ii) G = E , p =2and M =(22 D ).Sym(3); 7 6 ⇥ 4 CHAPTER 2. BACKGROUND MATERIAL 20
(iii) G = E , p =2, 3, 5 and M = A Sym(5); 8 6 1 ⇥ (iv) The connected component of M, M 0, is as in Table 2.1.1 below.
The subgroups M in (ii), (iii) and (iv) exist, and are unique up to conjugacy in Aut(G) and are maximal in G.
Table 2.1.1
G M 0 simple M 0 not simple
G A (p 7) 2 1 F A (p 13), G (p =7), A G (p =2) 4 1 2 1 2 6 E A (p =2, 3), G (p =7), A G 6 2 6 2 6 2 2 C (p =2),F 4 6 4 E A (2 classes,p 17, 19 resp,), A A (p =2, 3), A G (p =2), 7 1 1 1 6 1 2 6 A (p 5) A F , G C 2 1 4 2 3 E A (3 classes,p 23, 29, 31 resp,), A A (p =2, 3), A G G (p =2), 8 1 1 2 6 1 2 2 6 B (p 5) G F 2 2 4
The maximal parabolic subgroups are well known and Liebeck, Seitz and Saxl [41] give the reductive subgroups of maximal rank, ie those that contain a maximal torus. As any Lie imprimitive subgroup is contained in a subgroup of positive dimension we now know all maximal Lie imprimitive subgroups and can turn our attention to Lie primitive groups. For Lie primitive subgroups less is known; however, Liebeck and Seitz [42], extending an earlier result of Borovik [10], give a reduction theorem that limits the possibilities.
Theorem 2.2 (Liebeck Seitz). Let G be an adjoint simple algebraic group of excep- tional type over K, algebraically closed of characteristic p>0. Let be a Frobenius morphism of G and G be the fixed point group of G under .LetH be a maximal subgroup of G then one of the following holds:
(i) H = M , where M is a maximal -stable closed subgroup of positive dimension in G;
(ii) H is an exotic local subgroup; CHAPTER 2. BACKGROUND MATERIAL 21
(iii) G = E ,p>5 and H =(Alt(5) Alt(6)).22. 8 ⇥ (iv) H is almost simple;
The groups on part (i) are exactly those found in Theorem 2.1 with G = G 1 h i and those in part (ii) are listed by Liebeck and Seitz [44]. They also give two theorems giving the possibilities for part (iv) for when H Lie(p)[42]andH/Lie(p)[45].We 2 2 can now summarise the results of Section 2.1 with the following theorem:
Theorem 2.3 (Liebeck Seitz). Let H be a maximal subgroup of the finite exceptional
a group G over Fq, q = p where p is a prime. Then one of the following holds:
(i) H = M where M is a maximal closed -stable subgroup of positive dimension in G; the possibilities are as follows;
(a) Both M and H are parabolic subgroups;
(b) M is a reductive group of maximal rank. The possibilities for M are deter- mined in [41].
(c) G = E , p>2 and H =(22 ⌦+(q).22).Sym(3) or 3D (q).3; 7 ⇥ 8 4 (d) G = E , p>5 and H = PGL (q) Sym(5); 8 2 ⇥
(e) M is as in Table 1 of [44], and H = M as in Table 3 of [44].
(ii) H is of the same type as G;
(iii) H is an exotic local subgroup (see [44]);
(iv) G is of type E , p>5 and H (Alt(5) Alt(6)).22; 8 ⇠ ⇥
(v) F ⇤(H)=H0 is simple, and not in Lie(p): the possibilities for H0 are given up to isomorphism by [45];
1 (vi) F ⇤(H)=H(q ) is simple and in Lie(p); moreover rk(H(q )) rk(G), and one 0 0 2 of the following holds:
(a) q 9; 0 2 (b) H(q0) ⇠= A2(16) or A2(16); CHAPTER 2. BACKGROUND MATERIAL 22
(c) q (2,p 1)u(G) and H(q ) = A (q ), 2B (q ) or 2G (q ), where the 0 0 ⇠ 1 0 2 0 2 0 values of u(G) for each type of exceptional group are as follows:
G G2 F4 E6 E7 E8 u(G) 12 68 124 388 1312
In cases (i) (iv), H is determined up to G -conjugacy. Chapter 3
Preliminary Results
3.1 E8(2)
We now turn to E8(2), the focus of this thesis, and outline the information necessary for the methods used in Chapters 4-6. First we use Theorem 2.1 and the results of [41] to determine the maximal subgroups in E8(2) from Theorem 2.3 part (i), which we shall now refer to as maximal restricted subgroups of positive dimension. We then list all the possibilities for F ⇤(H)fromTheorem2.3parts(v)and(vi).Wemakeextensive use of information about the semisimple elements of E8(2), i.e. those of odd order. We therefore include the results of Rowley et al. [7] on the classes of semisimple elements, their centraliser and fixed-point spaces on L(G). The order of E8(2) is as follows
G =2120.313.55.74.112.132.172.19.312.41.43.73.127.151.241.331 | |
3.1.1 Maximal Subgroups
We have the following list of maximal restricted subgroups of positive dimension
(i) Maximal Parabolic Subgroups:
[278]:⌦+ (2) [298] : (Sym(3) L (2)) 14 ⇥ 7 [2106] : (Sym(3) L (2) L (2)) [2104] : (Alt(8) L (2)) ⇥ 3 ⇥ 5 ⇥ 5 [297]:(L (2) ⌦+ (2)) [283] : (Sym(3) E (2)) 3 ⇥ 10 ⇥ 6 92 57 [2 ]:L8(2) [2 ]:E7(2)
23 CHAPTER 3. PRELIMINARY RESULTS 24
(ii) Reductive Subgroups of Maximal Rank:
⌦+ (2) Sym(3) E (2) 16 ⇥ 7 L (2) : 2 3 U (2) : 2 9 · 9 (L (2) E (2)) : 2 3 (U (2) 2 E (2)) : Sym(3) 3 ⇥ 6 · 3 ⇥ 6 2 2 (L5(2)) .4(U5(2)) .4
SU5(4).4 PGU5(4).4 (⌦+(2))2.(Sym(3) 2) ⌦+(4).(Sym(3) 2) 8 ⇥ 8 ⇥ 3 2 3 ( D4(2)) .6 D4(4).6 4 2 4 2 (L3(2)) .GL2(3) [3 ].(U3(2)) .[3 ].GL2(3) 2 (U3(4)) .8 U3(16).8 38.(2.⌦+(2).2) 54.((4 21+4).Alt(6).2) 8 ⇤ 74.(2.(3 U (2))) 112.(5 SL (5)) ⇥ 4 ⇥ 2 132.(12 GL (3)) 312.(5 SL (5)) ⇤ 2 ⇥ 2 151.30 331.30
We now need a list of possibilities for F ⇤(H)givenbyTheorem2.3parts(v)and (vi)
Alt(5 17) L (7) L (11) L (13) L (17) 2 2 2 2 L2(19) L2(25) L2(27) L2(29) L2(31)
L2(37) L2(41) L2(61) L3(3) L3(5)
L4(3) L4(5) PSp4(5) U3(3) G2(3)
M11 M12 J2 J3
Table 3.1.1: Simple groups / Lie(2) 2 CHAPTER 3. PRELIMINARY RESULTS 25
L2(8) L2(16) L2(32) L2(64) L2(128) L2(256) L2(512)
L2(1024) L3(4) L3(8) L3(16) L4(4) L4(8) L5(2)
L5(4) L5(8) U3(4) U3(8) U3(16) U4(2) U4(4)
U4(8) U5(2) U5(4) U5(8) Sp4(4) Sp4(8) Sp6(2)
Sp6(4) Sp6(8) Sp8(2) Sp8(4) Sp8(8) Sz(8) Sz(32) + + + Sz(128) Sz(512) ⌦8 (2) ⌦8 (4) ⌦8 (8) ⌦8 (2) ⌦8 (4) 3 3 3 ⌦8 (8) D4(2) D4(4) D4(8) F4(2) F4(4) F4(8) 2 2 F4(2)0 F4(8)
Table 3.1.2: Simple groups Lie(2) 2
The groups missing from Table 3.1.2 are those isomorphic to some group in Table
3.1.1, i.e. L2(4) ⇠= Alt5, L3(2) ⇠= L2(7) and L4(2) ⇠= Alt8.
Using Lagrange’s theorem we can immediately remove the following groups from consideration.
F ⇤(H) F ⇤(H) Invalid divisor | | 2 L2(29) 2 .3.5.7.29 29 2 2 L2(37) 2 .3 .19.37 37 2 L2(61) 2 .3.5.31.61 61 7 2 6 6 L4(5) 2 .3 .5 .13.31 5 32 5 4 2 Sp8(4) 2 .3 .5 .7.13.17 .257 257 24 4 3 ⌦8 (4) 2 .3 .5 .7.13.17.257 257 3 36 6 2 2 2 D4(8) 2 , 3 .7 .19 .37.73 .109 109 48 6 4 2 2 2 F4(4) 2 .3 .5 .7 .13 .17 .241.257 257 72 10 2 4 2 2 2 F4(8) 2 .3 .5 .7 .13 .17.19 .37.73 .109.241 109 2 36 5 2 2 2 F4(8) 2 .3 .5 .7 .13 .19.37.109 109 8 L2(256) 2 .3.5.17.257 257 Sz(128) 214.5.29.113.127 113 Sz(512) 218.5.7.13.37.73.109 37
Table 3.1.3: Invalid group orders CHAPTER 3. PRELIMINARY RESULTS 26
3.1.2 Elements of E8(2)
We now wish to detail various properties of the elements in E8(2) up to conjugacy. We start with the elements of order 2, of which there are four classes.
Proposition 3.1. Let G ⇠= E8(2), V = L(G)andt be an involution in G. Also let
U = Ru(CG(t)) and L be such that CG(t)=UL. Then the possibilities for t are as follows:
(i) If t 2A, then dim(C (t)) = 190, U = 21+56 and L = E (2), 2 V ⇠ ⇠ 7 (ii) If t 2B, then dim(C (t)) = 156, U [278]andL = Sp (2), 2 V ⇠ ⇠ 12 (iii) If t 2C, then dim(C (t)) = 138, U [281]andL = Sym(3) F (2), 2 V ⇠ ⇠ ⇥ 4 (iv) If t 2D, then dim(C (t)) = 128, U [284]andL = Sp (2). 2 V ⇠ ⇠ 8
Proof. For the shape of CG(t) see Aschbacher, Seitz [4]. The dimension of CV (t)can be calculated directly in Magma.
We now consider the semisimple elements of E8(2). Table 3.1.4 contains the results of [7] in which the semisimple conjugacy classes of E8(2) are identified with their corresponding number in L¨ubeck’s ordering from [50] where much of the data was determined. It also contains the shape and size of the centraliser of the element, as well as its fixed space dimension and the class of its powers. CHAPTER 3. PRELIMINARY RESULTS 27 ------3C 3C 3C 3C 3B,5B 3B,5A 3B,5A 3C,5A 3D,5B 3A,5A 3D,5A Powers )) x ( V 92 86 80 68 48 80 38 48 34 30 28 28 32 20 48 34 28 26 24 20 16 32 16 14 C 248 134 dim( 127 · 73 · 43 · 19 73 43 43 31 · · · · (2). 31 · 8 · 17 41 31 19 31 E 17 17 31 · · 2 · · · 17 · · · 2 17 2 · 2 17 2 13 · 7 17 13 13 11 17 17 19 13 17 11 19 13 13 11 17 · · 11 · 3 · · · · · · 13 7 · · · · · · 13 · · · 17 | · | 5 19 7 2 · · 4 2 2 ) · · 3 2 2 5 7 · 7 7 · · 5 11 7 7 5 5 2 13 2 x 13 11 5 5 7 17 · · 11 · · 5 · 6 · 4 · · · · ( (2) 5 · 5 5 · · · · · · 11 · · 8 7 3 8 2 3 3 6 · 2 G · · 5 7 5 6 5 4 2 2 · 4 7 3 3 · 5 5 · 3 7 7 3 3 3 3 3 E 5 6 4 7 3 3 3 C 3 · 13 · · 2 | · · · · 8 · | · 3 · · · · · 3 3 · · · · · · 7 2 2 8 6 2 9 4 2 4 · · 2 2 11 10 6 6 4 · 5 3 3 15 13 10 10 2 5 5 2 12 3 5 2 2 2 2 2 · 3 · · 2 2 2 2 11 · · 13 12 2 · · 5 · 2 2 6 12 6 · 20 20 13 3 3 20 3 3 2 2 3 7 · 2 · · 10 · · 3 30 2 36 36 · 39 2 5 2 2 2 · 42 2 12 3 · 63 2 3 . (2) (2) (8) (2) 4 (2) (2) (2) 5 3 4 4 4 (2) 8 D (2)) U U L 3 U 3 D 3 ⌦ (2) (2) 3 (2) (2) (4) (2) (4) (2) (2) (2) ) 4 ⇥ ⇥ ⇥ U (2) (2) (16) (16) (2) ⇥ 5 3 5 3 ⇥ 8 + 10 7 6 x 2 2 9 ⇥ (4) 14 12 10 ⇥ D ( U U U U (2) 5 ⇥ ⌦ ⌦ U L L 3 E E PGU 8 ⌦ ⌦ ⌦ G (2) · (4) :2 3 ⇥ ⇥ ⇥ ⇥ 2 ⇥ E 3 ⇥ (2) ⇥ ⇥ ⇥ C ⇥ ⇥ ⇥ SU ⇥ ⇥ ⇥ (2) 2 3 L 3 6 3 7 3 ⇥ 3 5 9 11 13 15 15 L 17 GU 15 Sym(3) Sym(3) 15 17 13 Sym(3) E ⇥ ⇥ 2 ⇥ 19 ⇥ ⇥ ⇥ ( ⇥ . 5 15 7 5 9 9 3 9 Table 3.1.4: Conjugacy classes of semisimple elements of 1 294 376 147 258 480 247 441 516 560 656 580 366 679 712 709 540 636 686 621 600 706 695 738 693 823 ¨bc Number L¨ubeck A A A A F C G E B B D A A A A A C C B B B B D D AB CD 1 3 5 7 9 3 9 3 5 7 9 3 9 11 13 15 19 15 15 15 15 13 15 15 17 17 Conjugacy Class CHAPTER 3. PRELIMINARY RESULTS 28 - - - - 3B,7B 3C,7B 3A,7B 3B,7A 3A,7B 3C,7A 3D,7B 3A,7A 5A,7A Powers 3B,13B 3B,11A 3C,11A 3C,13A 3A,11A 3A,13A 3D,11A 3B,17AB 3C,17AB 3A,17AB 3D,17CD 3C,9C,5A,15D 3C,9A,5A,15D 3C,9A,5A,15D )) x ( V 8 8 8 8 38 32 32 26 20 20 14 14 28 20 16 14 12 20 14 14 10 20 12 10 20 14 14 C dim( 2 31 17 17 31 2 · · 7 7 13 · · 7 11 13 17 · · 3 2 2 2 · 3 2 7 | · · · · 7 7 7 5 5 11 11 11 13 7 17 7 7 7 2 2 3 ) 43 · 13 17 · · · · · · · · · · · · · · 5 5 5 5 7 5 7 · x 2 5 · · 41 · · · 4 5 4 4 4 · · · 5 5 4 4 4 · ( 5 5 5 5 2 3 · 3 3 3 3 3 3 3 3 3 5 3 5 6 4 4 · G 5 4 3 · · · 5 31 3 · · · · · · · · · · · · · 3 3 3 3 3 5 3 3 · 3 5 5 2 C 3 · · · 6 6 3 2 | · · · 4 3 2 3 3 · 2 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 9 6 6 · 6 4 3 · · · 12 2 2 2 2 2 2 6 2 15 12 10 2 2 2 2 2 (8) 2 (8) (4)) : 3 : 2Alt(4) 2 3 L Alt(5) (8) : 2Alt(4) L L 2 (2) (2) (2) ⇥ (2) (2) (2) (2) (2) (2) 3 ⇥ ) 1+2 + 4 L ⇥ Alt(5) 6 5 4 4 4 4 8 ⇥ 3 Sym(3) x : 2Alt(4) : 2Alt(4) : 2Alt(4) 2 1+2 + Sym(3) D ( U U L L L L 8 ⌦ ⇥ Sym(3) 3 3 ⇥ GU ⇥ G ⇥ 31 195 205 255 (2) ⇥ 3 Q ⇥ ⇥ ⇥ ⇥ ⇥ ⇥ 3 3 ⇥ ⇥ ⇥ 1+2 + 1+2 + 1+2 + C 3 ⇥ ⇥ : 3 3 3 3 L ⇥ ⇥ 2 Sym(3) 33 35 21 31 45 51 ⇥ 21 Sym(3) ⇥ 21 17 (2) ⇥ ⇥ ⇥ ⇥ 129 (3 21 3 ⇥ 45 . ⇥ 45 33 3 L Sym(3) 21 33 13 21 13 51 ⇥ ⇥ ⇥ 7 7 11 610 720 728 469 594 697 760 826 672 857 768 748 811 790 778 762 820 872 864 837 773 798 853 783 764 832 870 ¨bc Number L¨ubeck A A A A F F C G C C E E B B B D D H AB AB AB EF CD CD GH ABC ABC 21 35 39 45 21 33 21 21 39 45 21 33 21 39 45 21 31 21 33 41 51 51 33 51 51 31 43 Conjugacy Class CHAPTER 3. PRELIMINARY RESULTS 29 - - Powers 5B,13B 5A,13B 7B,13A 3C,19A 5A,11A 3A,19A 7A,13A 3D,19A 5B,17CD 5A,17AB 7A,17AB 3B,31ABC 3A,31ABC 3C,9B,7B,21F 3C,9D,7B,21F 3C,9B,7A,21D 3C,9A,7A,21D 3C,9B,13A,39B 3C,9C,11A,33CD 3C,9A,11A,33CD 3B,5A,7A,15A,21C,35A 3A,5A,7A,15B,21A,35A 3C,5A,7A,15D,21D,35A )) x ( V 8 8 8 8 8 8 8 8 14 16 12 10 10 12 14 12 14 14 12 10 14 12 10 10 10 C dim( 7 7 7 7 13 17 13 31 31 2 73 · · · 2 | · · · · · · 7 7 19 · 7 ) 11 13 5 5 5 11 17 17 · · · 2 2 5 5 7 5 127 · · · x 19 19 13 7 13 11 · · · · · · 5 7 · 4 4 · · · 4 ( · 3 · · · · · · 3 3 3 4 3 · · 3 3 5 3 5 7 3 G 2 2 2 3 3 3 3 3 2 2 2 3 3 3 3 3 · · · · · · 3 3 3 3 3 3 · · · · 3 3 5 7 3 C · · · · · 4 | · · · · · 3 2 3 3 3 2 2 2 2 2 3 2 3 2 2 2 2 3 3 2 2 2 2 2 2 2 2 2 2 2 (2) 3 2 2 (2) L 3 9 (2) 2 ⇥ (2) (2) (2) 3 ) 3 7 3 ⇥ 5 Sym(3) 3 3 3 Sym(3) Sym(3) x ( L L L ⇥ ⇥ ⇥ Alt(5) Alt(5) Alt(5) Sym(3) Sym(3) ⇥ ⇥ Sym(3) GU PGU Sym(3) 19 G ⇥ ⇥ Sym(3) · 165 255 357 ⇥ ⇥ ⇥ 3 3 ⇥ ⇥ ⇥ ⇥ ⇥ C 7 3 57 91 99 ⇥ ⇥ ⇥ ⇥ 13 ⇥ ⇥ ⇥ 3 Sym(3) 73 91 93 ⇥ ⇥ 63 65 93 35 99 85 117 127 105 ⇥ 19 63 63 105 63 877 823 861 863 754 802 843 849 800 858 814 804 870 817 865 808 788 841 867 829 794 851 845 878 835 ¨bc Number L¨ubeck C D A C E D D AB AB AB AB AB EF CD DE ABC ABC ABC ABC DEF FGH 55 57 63 63 91 ABCD ABCD CDEF 105 105 57 85 99 65 99 57 105 119 63 91 93 93 63 117 65 73 85 ABCDEF GHI Conjugacy Class 127 CHAPTER 3. PRELIMINARY RESULTS 30 - - - Powers 5B,41AB 3A,43ABC 5A,31ABC 7A,31ABC 3D,43ABC 3B, 43ABC 3A,73ABCD 7A,73ABCD 3C,9D,19A,57AB 3C,9A,17AB,51EF 3A,127ABCDEFGHI 3D,5A,11A,15C,33AB,55A 3A,7B,13A,21B,39A,91ABC 3B,5B,13B,15F,39C,65ABCD 3B,5A,15A,17AB,51AB,85AB 3A,5A,15B,17AB,51CD,85AB 3B,7A,17AB,21C,51AB,119AB 3D,5B,15G,17CD,51GH,85CDEF 3B,5A,15A,31ABC,93DEF,155ABC 3A,7A,21A,31ABC,93ABC,217ABCDEF 3C,5A,7A,9A,15D,21D,35A,45A,63D,105AB )) x ( V 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 10 10 10 10 C dim( 17 31 | 7 · ) 43 19 17 · 31 11 13 17 13 17 31 31 · · · · x 43 43 17 5 · · · · · · · · 41 7 73 73 ( 5 127 · · · · 2 3 5 · · · · 5 5 5 5 7 7 5 7 · G · 3 3 151 241 331 · 2 2 2 2 · · · · · · · · 5 3 3 7 2 · · 3 3 3 3 C 3 2 | · 3 3 3 3 3 3 3 3 3 · 2 2 3 2 2 3 3 3 ) x ⇥ ⇥ ⇥ ( Sym(3) Sym(3) Sym(3) Sym(3) G 151 153 465 165 195 205 219 241 255 273 315 331 357 381 465 511 651 ⇥ ⇥ ⇥ ⇥ C 129 129 255 129 171 217 255 837 859 859 868 879 876 877 847 872 864 839 875 866 855 860 870 873 871 869 878 874 876 862 880 ¨bc Number L¨ubeck AB AB AB EF GHI ABC ABC ABCD ABCD ABCD ABCD ABCDE 153 165 315 255 ABCDEF ABCDEF ABCDEF ABCDEF ABCDEF JKLMNO 129 155 273 195 219 255 357 GHIJKLMN ABCDEF GH ABCDEF GH 151 ABCDEF GHI ABCDEF GHIJ 129 171 217 465 651 129 ABCDEF GHIJK Conjugacy Class 255 205 511 381 241 331 CHAPTER 3. PRELIMINARY RESULTS 31
One immediate consequence of the information in Table 3.1.4 is knowledge of the exponents of the Sylowp subgroups of G. The exponent of a Sylow3 subgroup is 9 and all other Sylowp subgroups are elementary abelian. This allows us to rule out another set of subgroups whose Sylowp subgroups are incompatible with those of G.
F ⇤(H) p prime Invalid Exponent of Sylowp
L2(49) 7 49
L2(512) 5 25
L2(1024) 3 27 Sz(32) 5 25
Sp6(8) 3 27
Sp8(8) 3 27 + ⌦8 (8) 3 27
⌦8 (8) 3 27
Table 3.1.5: Invalid Sylow Exponents
Finally, we can also rule out L4(8) and U4(8) by the order of the centraliser of an element of order 13. In L (8) there exists x of order 13 such that C (x) =32.5.13 4 | L4(8) | which does not divide either centraliser order in E8(2); likewise, in U4(8) there exists x of order 13 such that C (x) =5.7.13 which does not divide either centraliser | U4(8) | order in E8(2).
In the course of this thesis we often need to find the centraliser of a semisimple
2 t 1 element g or all inverting involutions for it, i.e. all t G such that t =1andg = g . 2 In smaller groups than E8(2) this is a fairly straightforward calculation; however, due to the size of the group normal methods do not work here. We therefore introduce methods for finding these two objects developed by Rowley and Ballantyne for use in [6].
Given a semisimple element x G = E (2) and a centraliser C (y)wherey xG, 2 ⇠ 8 G 2 the code in Appendix B returns a subgroup of CG(x)thathasaverygoodchanceof being the whole centraliser. As we know the orders of all semisimple centralisers we CHAPTER 3. PRELIMINARY RESULTS 32 can easily check whether we have indeed found the whole centraliser.
For inverting involutions suppose g G = E (2) is an element of odd order, then 2 ⇠ 8 all inverting involutions are contained in the subgroup
x x 1 C (g),t = C⇤ (g)= x G g = g or g = g . h G i G { 2 | }
We therefore only need to find a single inverting involution and CG(g)usingthecen- traliser code described earlier. To find an inverting involution for g we use the following procedure. First we find an involution t and y G such that h = tty gG. Then 2 2 t inverts h and we can use them as a guide for where to search for an inverting in- volution for g. We can calculate K = CG(h)anduseBray’smethodtofindCK (t).
Choosing an element z of odd order out of CK (t)insuchawaythatanequivalent element of CG(g)canbechosen,wenowknowthatCG(z)containsaninvertinginvo- lution for g. In some cases more must be done to find an inverting involution inside
CG(z). For our purposes this is easy, as we can employ LMGRadicalQuotient to get
CG(z)=CG(z)/ z in which the inverse image of N ( g )containsaninverting h i CG(z) h i involution, and is very small. Examples of the odd order element z that are picked are as follows
g 3C z C (g), z 13A 2 2 G 2 g 3B z C (g), z 11A, 13B, or 17AB 2 2 G 2 g 3D z Syl (C (g)), z 7B 2 h i2 7 G 2 g 7B z C (g), z 13A 2 2 G 2 g 31ABC z C (g), z 5A 2 2 G 2
3.2 Brauer Characters
An important tool in studying embeddings of groups is that of the Brauer charac- ters. Just as ordinary characters encode the information about the eigenvalues of an element of a complex representation, so too do the Brauer characters on a modular representation. For a GF (q)G-module V with q = pn we can define a complex valued function V on the set of p-regular conjugacy classes in the following way. For ordinary representations the value of 'V (x)isthesumoftheeigenvaluesoftheactionofx on CHAPTER 3. PRELIMINARY RESULTS 33
V ; however, in finite fields we lose information when taking such a sum so we first map the values into a field of characteristic 0, i.e. C. This map, known as a lifting, th takes the elements of the multiplicative group GF (q)⇤ and maps them to the q roots of unity. We can now sum the images of our eigenvalues and we have our required complex valued function.
An important fact of modular representation theory is that Maschke’s theorem frequently doesn’t hold as p - G is a requirement for it to do so. Whilst this makes | | finding the set of indecomposable modules for G a much harder task it also allows for deeper insight if they are well understood. In Section 3.2.3 we use the irreducible Brauer characters of H G to determine the feasible decompositions of the restriction L(G) . |H
3.2.1 Conway Polynomials
In order for the Brauer characters calculated for H G to be consistent with the restrictions of those of G we must fix our lifting of the finite fields. We do this in the way introduced in the atlas [33] via the Conway polynomials. Up to isomorphism there is a unique field GF (q)oforderq = pn, p prime n N.Forn =1,GF (p) 2 is the field Z/pZ and then GF (q)maybedefinedasGF (p)[X]/(fn)wherefn is an irreducible polynomial of degree n. For our consistent lifting to be possible, we must
th fix an fn for each n which will be called the n Conway polynomial.
This choice is made with certain conditions in mind; fn is a monic and primitive polynomial ensuring that zn = X +(fn) is a generator for the multiplicative group
GF (q)⇤. We also require fn to be consistent with fd where d is any divisor of n, i.e. if ↵ =(pn 1)/(pd 1) then z↵ is a root of f . A lexographic ordering is now defined n d on the set of all such polynomials as
n n 1 n n n 1 n X an 1X + +( 1) a0