Skip to content
Publicly Available Published by De Gruyter September 9, 2020

Even character degrees and normal Sylow 2-subgroups

  • Hongfei Pan EMAIL logo , Nguyen Ngoc Hung and Shuqin Dong
From the journal Journal of Group Theory

Abstract

The Ito–Michler theorem on character degrees states that if a prime p does not divide the degree of any irreducible character of a finite group G, then G has a normal Sylow p-subgroup. We give some strengthened versions of this result for p=2 by considering linear characters and those of even degree.

1 Introduction

The degrees of irreducible characters of a finite group encode a lot of structural information of the group. One of the typical results in this line of research is the Ito–Michler theorem [6, 8], which states that if a prime p does not divide the degree of any irreducible character of a finite group G, then G has a normal Sylow p-subgroup.

For a finite group G, as usual, let Irr(G) be the set of all complex irreducible characters of G. We write

Irr2(G):={χIrr(G)χ(1)=1or 2χ(1)},
S2(G):=χIrr2(G)χ(1)2/χIrr2(G)χ(1).

In [2], the average degree of characters in Irr2(G) has been considered to obtain an improvement of the Ito–Michler theorem for the prime p=2. In this paper, we investigate the S2(G)-version of the Ito–Michler theorem. This is initiated from the obvious observation that every irreducible character degree of G is odd if and only if S2(G)=1.

Theorem A.

Let G be a finite group with S2(G)<32. Then G has a normal Sylow 2-subgroup.

We have no clue what the possible values of S2(G) (or Sp(G) in general) are. However, we do know that there are groups G such that 1<S2(G)<32, with D8 and Q8 being the examples of smallest order. Note that S2(D8)=S2(Q8)=43.

Recall that a character χIrr(G) is real if χ(g) for every element gG, and χIrr(G) is strongly real if χ is afforded by a real representation, or equivalently, its Frobenius–Schur indicator v2(χ) is 1. We furthermore write

Irr2,(G):={χIrr2(G)χis real},
S2,(G):=χIrr2,(G)χ(1)2/χIrr2,(G)χ(1),
Irr2,+(G):={χIrr2(G)v2(χ)=1},
S2,+(G):=χIrr2,+(G)χ(1)2/χIrr2,+(G)χ(1).

The following are variations of Theorem A for real and strongly real characters.

Theorem B.

Let G be a finite group with S2,+(G)<32. Then G has a normal Sylow 2-subgroup.

Observe that, for every χIrr2,(G)-Irr2,+(G), we have χ(1)2. Hence, if S2,+(G)2, then we have

χIrr2,+(G)χ(1)2χIrr2,+(G)χ(1)χIrr2,(G)-Irr2,+(G)χ(1)2χIrr2,(G)-Irr2,+(G)χ(1)=χIrr2,(G)χ(1)2-χIrr2,+(G)χ(1)2χIrr2,(G)χ(1)-χIrr2,+(G)χ(1),

and it follows that S2,+(G)S2,(G). Therefore, we obtain the following consequence of Theorem B.

Corollary C.

Let G be a finite group with S2,R(G)<32. Then G has a normal Sylow 2-subgroup.

In order to prove Theorem A, we have to prove a criterion for solvability of G in terms of S2(G).

Theorem D.

Let G be a finite group. Suppose that S2(G)<175. Then G is solvable.

To end this introduction, we remark that all the bounds in Theorems A, B, D and Corollary C are best possible, as shown by the symmetric group S3 and the alternating group A5.

2 Lemmas

We write nk(G) and nk,+(G) to denote the number of irreducible complex and strongly real characters of degree k, respectively, of G. Let N be a normal subgroup of G and λIrr(N); we write IG(λ) for the inertia subgroup of λ in G, Irr(G|λ) for the set of irreducible characters of G that lie over λ, and nk(G|λ) for the number of members in Irr(G|λ) of degree k. We refer to [4] for other standard notation.

We need the following to prove Theorems A and D.

Lemma 2.1.

Let N be a normal subgroup of a group G such that NG.

  1. If S2(G)2, then S2(G/N)S2(G).

  2. If S2,+(G)2, then S2,+(G/N)S2,+(G).

Proof.

(i) Write A:=S2(G). Then

A=n1(G)+2k,k2k2nk(G)n1(G)+2k,k2knk(G),

and thus

(A-1)n1(G)=2k,k2k(k-A)nk(G).

From the hypothesis, we know that n1(G)=n1(G/N). Also, A2, and for k2, we have nk(G)nk(G/N). It follows that

(A-1)n1(G/N)2k,k2k(k-A)nk(G/N),

implying that

n1(G/N)+2k,k2k2nk(G/N)A(n1(G/N)+2k,k2knk(G/N)).

Hence S2(G/N)A=S2(G).

(ii) Similar to (i). ∎

In order to prove Theorem B, we need the following lemma on strongly real characters.

Lemma 2.2.

Let N be a normal subgroup of a group G such that G/N has odd order. If S2,+(G)2, then S2,+(N)S2,+(G).

Proof.

Clearly, every strongly real linear character of G restricts to a strongly real linear character of N. Also, by [7, Lemma 2.1], every strongly real linear character of N lies under a unique strongly real linear character of G. Hence we have n1,+(N)=n1,+(G). We now have

2k,k2nk,+(G)2k,k2nk,+(N).

Write a:=S2,+(G). Then, from

a=n1,+(G)+2k,k2k2nk,+(G)n1,+(G)+2k,k2knk,+(G),

we get that

(a-1)n1,+(N)=(a-1)n1,+(G)=2k,k2(k2-ak)nk,+(G)2k,k2(k2-ak)nk,+(N).

It follows that S2,+(N)a=S2,+(G), as desired. ∎

3 Proof of Theorem D

We restate Theorem D here for the reader’s convenience.

Theorem 3.1.

Let G be a finite group. Suppose that S2(G)<175. Then G is solvable.

Proof.

Let G be a counterexample of minimal order. Let M be a normal subgroup of G such that M is minimal nonsolvable, and let N be a minimal normal subgroup of G contained in M. Then NM=MG. If further [M,Rad(M)]1, then we suppose N[M,Rad(M)], where Rad(M) is the solvable radical of M.

Suppose that N is abelian. Then G/N is nonsolvable. By the choice of G, we have S2(G)<175S2(G/N). Hence we have

2k,k4(5k-17)knk(G)<12n1(G)+14n2(G),
12n1(G/N)+14n2(G/N)2k,k4(5k-17)knk(G/N).

Observe that nk(G/N)nk(G) for k1 and n1(G/N)=n1(G), so we have n2(G/N)<n2(G). Therefore, we can find a χIrr(G)-Irr(G/N) such that χ(1)=2. Let L=ker(χ). Then NL, and thus ML/L is a nontrivial nonsolvable subgroup of G/L.

We claim that χ is a primitive character of G¯:=G/L. Suppose χ=ψG¯ for some character ψ of a proper subgroup K¯ of G¯. Then 2=χ(1)=|G¯:K¯|ψ(1), and thus G¯/K¯2 and ψ(1)=1. Now all the irreducible constituents of χ|K¯ are linear. Note that χ is a faithful character of G¯, and therefore [K¯,K¯]=1, implying that K¯ is abelian. We now have that G¯ is solvable, which is impossible. This proves the claim.

Now G/L is a nonsolvable primitive linear group of degree 2. Let

C/L:=𝐙(G/L).

By Lemma [4, Theorem 14.23], G/C𝖠5. By the proof of [5, Theorem 2.2], we can show that G=MC is a central product. Here we give this argument again for the reader’s convenience. Note that G/C is simple and MCG. If G>MC, then MC=C, and thus ML/LC/L=𝐙(G/L), a contradiction. Hence G=MC and MC. Now MC<M and MCG. By the choice of M, we have MCRad(M). Since M/(MC)G/C is simple, we have MC=Rad(M), and thus Rad(M)C. Since C/L=𝐙(G/L), we have [M,Rad(M)]L. Hence N[M,Rad(M)]. By the choice of N, we have [M,Rad(M)]=1, and thus Rad(M)𝐙(M). Clearly, 𝐙(M)Rad(M), and N is abelian, so NRad(M)=MC=𝐙(M). Observe that

[M,C]=[C,M]MC=𝐙(M),

and hence [M,C,M]=[C,M,M]=1. Using the three subgroups lemma, we then have [M,C]=[M,M,C]=1. It follows that G=MC is a central product. Here G/CM/(MC)A5, N=Z(M)=MC2 and MSL(2,5).

Let λIrr(N)-1N; we have a bijection

($*$)Irr(M|λ)×Irr(C|λ)Irr(G|λ).

Notice that χIrr(G|λ), so there are αIrr(M|λ) (=Irr(M)-Irr(M/N)) and βIrr(C|λ) such that χ(1)=α(1)β(1). Then α(1)=2, β(1)=1, and thus λ extends to C. By Gallagher’s theorem [4, Corollary 6.17], we get a bijection Irr(C/N)Irr(C|λ). Then n1(G)=n1(C/N)=n1(C|λ). By ($*$), we also have

n2(G)=n2(G/N)+n2(G|λ)=n1(M/N)n2(C/N)+n2(M/N)n1(C/N)+n1(M|λ)n2(C|λ)+n2(M|λ)n1(C|λ)=n2(C/N)+2n1(G)2n1(G),

where

n1(M/N)=n1(PSL(2,5))=1,n2(M/N)=0,
n1(M|λ)=n1(M)-n1(M/N)=n1(SL(2,5))-n1(PSL(2,5))=0,
n2(M|λ)=n2(M)-n2(M/N)=2.

Similarly,

n4(G)2n1(G),n6(G)n1(G)+2n2(C/N),n8(G)2n2(C/N).

Now

2k,k4(5k-17)knk(G)12n4(G)+78n6(G)102n1(G)+156n2(C/N)=102n1(G)+156(n2(G)-2n1(G))=-210n1(G)+142n2(G)+14n2(G)-210n1(G)+284n1(G)+14n2(G)12n1(G)+14n2(G).

Hence S2(G)175, a contradiction.

Suppose that N is not abelian and NA5. By [2, Theorem 2.2], there exists ϕIrr(N) of even degree such that ϕ(1)8 and ϕ is extendible to IG(ϕ). By [2, Proposition 2.3], we have

n1(G)nd(G)|G:IG(ϕ)|,
n2(G)n2d(G)|G:IG(ϕ)|+12nd(G)|G:IG(ϕ)|,

where d=ϕ(1)|G:IG(ϕ)|8|G:IG(ϕ)|. Now

12n1(G)+14n2(G)12nd(G)|G:IG(ϕ)|+14(n2d(G)|G:IG(ϕ)|+12nd(G)|G:IG(ϕ)|)=19nd(G)|G:IG(ϕ)|+14n2d(G)|G:IG(ϕ)|19nd(G)d8+14n2d(G)d8=19d8nd(G)+7d4n2d(G)(5d2-17d)nd(G)+(20d2-34d)n2d(G)2k,k4(5k2-17k)nk(G),

leading to a contradiction again.

Finally, assume that NA5. By [1], the irreducible character of N of degree 4 is extendible to G since GAut(N). Then, by [2, Proposition 2.3] again, we have n1(G)n4(G) and n2(G)n8(G). Now

12n1(G)+14n2(G)12n4(G)+14n8(G)2k,k4(5k2-17k)nk(G),

and thus S2(G)175. This final contradiction completes the proof. ∎

4 Proof of Theorem A

The key step in the proof of Theorem A is the following lemma.

Lemma 4.1.

Let G=MN, where NG is an abelian group. Assume that no nontrivial irreducible character of N is invariant under M. If S2(G)<32, then there is no orbit of even size in the action of M on the set of irreducible characters of N.

Proof.

Let {α0=1N,α1,,αt} be a set of representatives of the action M on Irr(N). Let Ii=IG(αi), 1it. Since no nontrivial irreducible character of N is invariant under M, we have Ii<G for i1. Suppose that there is some orbit of even size in the action of M on the set of irreducible characters of N. Then there is some j such that 2|G:Ij|. For 0it, we set

ni,1=n1(Ii/N),ni,even=2knk(Ii/N),
Ti,even=λIrr(Ii/N),2λ(1)λ(1),Ti,even*=λIrr(Ii/N),2λ(1)λ(1)2.

Let βi be an extension of αi to Ii. By Gallagher’s theorem and the Clifford correspondence, we have bijections

Irr(Ii/N)Irr(Ii|αi),λ(λβi)Ii,
Irr(Ii|αi)Irr(G|αi),(λβi)Ii(λβi)G.

Note that (λβi)G(1)=|G:Ii|λ(1) is even if and only if |G:Ii| is even or |G:Ii| is odd and λ(1) is even. Hence we have

χIrr2(G)χ(1)=n1(G/N)+2|G:Ii||G:Ii|T(Ii/N)+2|G:Ii||G:Ii|Ti,even,
χIrr2(G)χ(1)2=n1(G/N)+2|G:Ii||G:Ii|2λIrr(Ii/N)λ(1)2+2|G:Ii||G:Ii|2Ti,even*n1(G/N)+2|G:Ii||G:Ii|2T(Ii/N)+22|G:Ii||G:Ii|2Ti,even.

On the other hand,

χIrr2(G)χ(1)2=S2(G)(n1(G/N)+2|G:Ii||G:Ii|T(Ii/N)+2|G:Ii|Ti,even).

Therefore, we derive that

2|G:Ii|(|G:Ii|-S2(G))|G:Ii|T(Ii/N)+2|G:Ii|(2|G:Ii|-S2(G))|G:Ii|Ti,even(S2(G)-1)n1(G/N).

Since 2|G:Ij|, we deduce that

(|G:Ij|-S2(G))|G:Ij|T(Ij/N)(S2(G)-1)n1(G/N).

Observe that

n1(G/N)=|G/N:G/N|andnj,1=n1(Ij/N)=|Ij/N,(Ij/N)|,

so n1(G/N)|G:Ij|nj,1. Also, T(Ij/N)n1(Ij/N). Hence we get that

(|G:Ij|-S2(G))|G:Ij|n1(Ij/N)(S2(G)-1)|G:Ij|n1(Ij/N).

It then follows that

S2(G)|G:Ij|+1232,

which is a contradiction. ∎

We now prove Theorem A, which is restated.

Theorem 4.2.

Let G be a finite group. Suppose that S2(G)<32. Then G has a normal Sylow 2-subgroup.

Proof.

Let G be a counterexample of minimal order, and let N be a minimal normal subgroup of G with NG. Since S2(G)<32<175, by Theorem 3.1, G is solvable, and then N is an elementary abelian group. By Lemma 2.1, we have S2(G/N)S2(G). Due to the choice of G, we know that G/N has a normal Sylow 2-subgroup T/N. If N is a 2-group, then T is a normal Sylow p-subgroup of G, a contradiction. Hence we may assume that N is an elementary abelian 2-group. Using the Schur–Zassenhaus theorem [3, Chapter 1, Theorem 18.1], we deduce that T=PN, where P is a Sylow 2-subgroup of T. By the Frattini argument, we then have G=TNG(P)=NNG(P).

If NNG(P), then G=NG(P), and thus P is a normal Sylow 2-subgroup of G, a contradiction. Hence we have NNG(P). Now

NG(P)N<NandNG(P)NNG(P),N=G.

Then, by the minimality of N, NG(P)N=1, and thus G=NG(P)N. If NZ(G), then T=P×N, and thus PG, a contradiction. Hence NZ(G), and then [N,G]=N. This means that no nontrivial irreducible character of N is invariant under NG(P). By Lemma 4.1, there is no orbit of even size in the action of P on N, that is, P acts trivially on N, and hence T=P×N. It follows that PG, completing the proof. ∎

5 Proofs of Theorem B and Corollary C

Theorem 5.1.

Let G be a finite group. Suppose that S2,+(G)<32. Then G has a normal Sylow 2-subgroup.

Proof.

Let G be a counterexample of minimal order, and let N be a minimal normal subgroup of G with NG. According to the Cauchy–Schwarz inequality, we have

χIrr2,+(G)χ(1)|Irr2,+(G)|S2,+(G)<32,

so, by [2, Theorem 5.1], we know that G is solvable. By Lemma 2.1 and the choice of G, G/N has a normal Sylow 2-subgroup T/N, T=PN, where PSyl2(G) and N is an elementary abelian 2-group. If T is a proper subgroup of G, then, by Lemma 2.2, we have T=P×N, and thus PG, a contradiction.

Hence G=T=PN. Since N has odd order, 1N is the only real irreducible character of N. Then Nker(χ) for any strongly real linear character χ of G. So

n1,+(G)=n1,+(G/N)=n1,+(P)=|P:Φ(P)|.

By [2, Lemma 5.4], we have n1,+(G)|N|-1. Let Ω1,,Ωs be the orbits of the action of P on Irr(N)-1N. Write |Ωi|=di,i=1,,s. Then we have n1,+(G)d1++ds. Notice that P acts irreducibly on Irr(N). Let aZ(P) be of order 2, and let M={ϕIrr(N)ϕa=ϕ-1}. Then M is a P-invariant subgroup of Irr(N). Also, since o(a)=2, M is nontrivial. Then M=Irr(N), and thus, for ϕIrr(N)-1N, ϕ and ϕ¯ are P-conjugate. By [10, Lemma 2.5], there exists a strongly real character χIrr(G) lying above ϕ.

Now let ϕiΩi for 1is. There is a strongly real character χiIrr(G) with degree χi(1)=|G:IG(ψi)|ψi(1)=|G:IG(ψi)|, where ψi is an extension of ϕi to IG(ϕi). Then χi(1)|G:IG(ϕi)|=di, and thus

n1,+(G)|N|-1=d1++dsχ1(1)++χs(1).

Now

n1,+(G)+i=1sχi(1)2nχi(1),+n1,+(G)+i=1sχi(1)nχi(1),+n1,+(G)+i=1sχi(1)2n1,+(G)+i=1sχi(1)n1,+(G)+2i=1sχi(1)n1,+(G)+i=1sχi(1)32.

For any χIrr2,+(G) with χ(1)2, we have χ(1)2/χ(1)232, so, by [9, Lemma 2.1], we get that S2,+(G)32, and this contradiction completes the proof. ∎

Corollary C follows from Theorem B, as we discussed in the introduction.


Communicated by Britta Späth


Award Identifier / Grant number: 11801208

Funding statement: Project supported by NSF of China (No. 11801208) and Jiangsu Government Scholarship for Overseas Studies (2018).

Acknowledgements

The authors are grateful to the referee for his or her comments and corrections on an earlier version of the paper.

References

[1] J. H. Conway, R. T. Curtis, S. P. Norton, R. A. Parker and R. A. Wilson, Atlas of finite groups, Oxford University, Eynsham, 1985. Search in Google Scholar

[2] N. N. Hung and P. H. Tiep, Irreducible characters of even degree and normal Sylow 2-subgroups, Math. Proc. Cambridge Philos. Soc. 162 (2017), no. 2, 353–365. 10.1017/S0305004116000669Search in Google Scholar

[3] B. Huppert, Endliche Gruppen. I, Springer, Berlin, 1967. 10.1007/978-3-642-64981-3Search in Google Scholar

[4] I. M. Isaacs, Character Theory of Finite Groups, Academic Press, New York, 1976. Search in Google Scholar

[5] I. M. Isaacs, M. Loukaki and A. Moretó, The average degree of an irreducible character of a finite group, Israel J. Math. 197 (2013), no. 1, 55–67. 10.1007/s11856-013-0013-zSearch in Google Scholar

[6] N. Itô, Some studies on group characters, Nagoya Math. J. 2 (1951), 17–28. 10.1017/S002776300001000XSearch in Google Scholar

[7] S. Marinelli and P. H. Tiep, Zeros of real irreducible characters of finite groups, Algebra Number Theory 7 (2013), no. 3, 567–593. 10.2140/ant.2013.7.567Search in Google Scholar

[8] G. O. Michler, Brauer’s conjectures and the classification of finite simple groups, Representation Theory. II (Ottawa 1984), Lecture Notes in Math. 1178, Springer, Berlin (1986), 129–142. 10.1007/BFb0075293Search in Google Scholar

[9] G. Qian, On the average character degree and the average class size in finite groups, J. Algebra 423 (2015), 1191–1212. 10.1016/j.jalgebra.2014.11.005Search in Google Scholar

[10] P. H. Tiep, Real ordinary characters and real Brauer characters, Trans. Amer. Math. Soc. 367 (2015), no. 2, 1273–1312. 10.1090/S0002-9947-2014-06148-5Search in Google Scholar

Received: 2020-03-14
Revised: 2020-08-24
Published Online: 2020-09-09
Published in Print: 2021-01-01

© 2021 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 24.4.2024 from https://www.degruyter.com/document/doi/10.1515/jgth-2020-0038/html
Scroll to top button