International Journal of Group Theory Vol. XX No. X (201X), pp XX-XX.
On the Order of the Schur Multiplier of a Pair of Finite p-Groups II
Fahimeh Mohammadzadeh, Azam Hokmabadi∗ and Behrooz Mashayekhy
Abstract.
Let G be a finite p-group and N be a normal subgroup of G, with
∣N∣=pn and ∣G/N∣=pm. A result of Ellis (1998) shows
that the order of the Schur multiplier of such a pair (G,N) of finite p-groups is bounded
by p21n(2m+n−1) and hence it is equal to p21n(2m+n−1)−t, for some non-negative integer t.
Recently the authors characterized the structure of (G,N) when N has a complement in G and
t≤3. This paper is devoted to classify the structure of
(G,N) when N has a normal complement in G and t=4,5.
MSC(2010): Primary: 20E34; Secondary: 20D15.
Keywords: pair of groups, Schur multiplier, finite p-groups.
∗Corresponding author
c◯ 2011 University of Isfahan
Communicated by
1. Introduction
By a pair of groups (G,N) we mean a group G with a normal
subgroup N. In 1998 Ellis [2] defined the Schur multiplier of a
pair (G,N) to be the abelian group M(G,N) appearing in a
natural exact sequence
[TABLE]
in which H3(G) is the third homology of G with integer
coefficients. In 1956, Green [4] showed that if G is a group of
order pn , then its Schur multiplier is of order at most
p2n(n−1), and hence equals to
p2n(n−1)−t, for some non-negative integer t. Berkovich [1], Zhou [11], Ellis [3] and Niroomand [7, 8] determined
the structure of G for t=0,1,2,3,4,5 by different methods.
In 1998, Ellis [2] gave
an upper bound for the order of the Schur multiplier of a pair of finite
p-groups. He proved that
if G is a finite p-group with normal subgroup N of order pn and quotient G/N of order pm, then the Schur multiplier of (G,N) is bounded by p21n(2m+n−1) and hence equals to p21n(2m+n−1)−t, for some nonnegative integer t.
Let (G,N) be a pair of
groups and K be the complement of N in G. In 2004, Sallemkar, Moghaddam and Saeedi [10] characterized the structure of such a pair (G,N) when t=0,1 with some conditions. Recently the authors [5] determined the structure of the pair (G,N), for t=0,1, without any condition and also gave the structure of (G,N) for t=2,3,
when K is normal. In this paper we are going to determine the structure of (G,N) for t=4,5 when K is a normal subgroup of G.
In this paper D and Q denote the dihedral and quaternion
group of order 8 and, E1 and E2 denote the extra special
p-groups of order p3 of odd exponent p
and p2, respectively. E4 denotes the unique central product of a cyclic group of order p2 and a non-abelian group of order p3.
Also Zn(m) denote the direct product of m
copies of Zn.
The following result is essential to prove the main theorems.
Theorem 1.1**.**
[2*]*Let (G,N) be a pair of groups and
K be the complement of N in G. Then
[TABLE]
In 1907 Schur [6] gave an structure for the Schur multiplier of a
direct product of finite groups. He showed that
[TABLE]
As a consequence of this fact we have the following important result.
Corollary 1.2**.**
Let (G,N) be a pair of groups and
K be the complement of N in G. Then
[TABLE]
The following theorems gave the structure of a finite p-group, in terms of the order of its Schur multiplier.
Theorem 1.3**.**
[3*]**
Let G be a group of prime-power order pn.
Suppose that ∣M(G)∣=p21n(n−1)−t. Then
i) t=0 if and only if G is elementary abelian;
ii) t=1 if and only if G≅Zp2 or G≅E1;
iii) t=2 if and only if G≅Zp×Zp2, G≅D
or G≅Zp×E1;
iv) t=3 if and only if G≅Zp3,
G≅Zp×Zp×Zp2, G≅Q, G≅E2,
G≅D×Z2 or G≅E1×Zp×Zp.*
Theorem 1.4**.**
[9*]** Let G be an abelian group of order pn.
Suppose that ∣M(G)∣=p21n(n−1)−4. Then G is isomorphic to
Zp2×Zp2 or Zp2×Zp(3).
The following result can be easily obtained by using a method similar to the proof of the above theorem in [9].
Theorem 1.5**.**
*Let G be an abelian group of order pn.
Suppose that ∣M(G)∣=p21n(n−1)−5. Then G is isomorphic to
Zp3×Zp or Zp2×Zp(4).
Theorem 1.6**.**
[7*]** Let G be a non-abelian group of order pn.
Suppose that ∣M(G)∣=p21n(n−1)−4. Then G is isomorphic to one of the following groups:
For p=2,
-
D×Zp(2);
-
Q×Z2;
-
⟨a,b∣a4=b4=1,[a,b,a]=[a,b,b]=1,[a,b]=a2b2⟩;
-
⟨a,b,c∣a2=b2=c2=1,abc=bca=cab⟩;
For p=2
-
E4;
-
E1×Zp(3);
-
Zp(4)>⊲θZp;
-
E2×Zp;
-
⟨a,b∣ap2=bp=1,[a,b,a]=[a,b,b]=1⟩;
-
⟨a,b∣a9=b3=1,[a,b,a]=1,[a,b,b]=a6,[a,b,b,b]=1⟩;
-
⟨a,b∣ap=bp=1,[a,b,a]=[a,b,b,a]=[a,b,b,b]=1⟩(p=3);
Theorem 1.7**.**
[8*]** Let G be a non-abelian group of order pn.
Suppose that ∣M(G)∣=p21n(n−1)−5. Then G is isomorphic to one of the following groups:
-
D×Z2(3);
-
E1×Zp(4);
-
E2×Zp(2);
-
E4×Zp;
-
extra special p-group of order p5;
-
⟨a,b∣ap2=bp2=1,[a,b,a]=[a,b,b]=1,[a,b]=ap⟩;
-
⟨a,b∣ap2=bp=1,[a,b,a]=[a,b,b]=ap,[a,b,b,b,]=1⟩;
-
⟨a,b∣ap2=bp=1,[a,b,a]=[a,b,b,b,]=1,[a,b,b]=anp⟩ where n is a fixed quadratic non-residue of p and p=3;
-
⟨a,b∣ap2=1,b3=a3,,[a,b,a]=[a,b,b,b,]=1,[a,b,b]=a6⟩;
-
⟨a,b∣ap=1,bp=[a,b,b],[a,b,a]=[a,b,b,b,]=[a,b,b,a]=1⟩;
-
D16;
-
⟨a,b∣a4=b4=1,a−1ba=b−1⟩;
-
Q×Z2(2);
-
(D×Z2)>⊲Z2;
-
(Q×Z2)>⊲Z2;
-
Z2×⟨a,b,c∣a2=b2=c2=1,abc=bca=cab⟩;
2. Main Results
In this section, let (G,N) be a pair of
groups such that G≅N×K, with ∣N∣=pn and ∣K∣=pm. It is obtained that ∣M(G,N)∣=p21n(2m+n−1)−t, for some nonnegative integer t. Recently, all these pairs of finite p-groups are listed in [5] by the authors, when t=0,1,2,3.
The aim of this paper is characterizing the structure of such pairs of finite p-groups, when t=4,5.
Theorem 2.1**.**
*By the above assumption, t=4 if and only if G is isomorphic to one of the following groups:
-
G≅N×K where N≅Zp and K is any
group with d(K)=m−4;
-
G≅N×K where N≅Zp×Zp and K is any group with d(K)=m−2;
-
G≅N×K where N≅Zp(4) and K
is any group with d(K)=m−1;
-
G=N≅D×Z2(2);
-
G=N≅Q×Z2;
-
G=N≅⟨a,b∣a4=b4=1,[a,b,a]=[a,b,b]=1,[a,b]=a2b2⟩;
-
G=N≅⟨a,b,c∣a2=b2=c2=1,abc=bca=cab⟩;
-
G=N≅E4;
-
G=N≅E1×Zp(3);
-
G=N≅Zp(4)>⊲θZp;
-
G=N≅E2×Zp;
-
G=N≅⟨a,b∣ap2=bp=1,[a,b,a]=[a,b,b]=1⟩;
-
G=N≅⟨a,b∣a9=b3=1,[a,b,a]=1,[a,b,b]=a6,[a,b,b,b]=1⟩;
-
G=N≅⟨a,b∣ap=bp=1,[a,b,a]=[a,b,b,a]=1,[a,b,b,b]=1⟩(p=3);
-
G=N≅Zp2×Zp2;
-
G=N≅Zp2×Zp(3);
-
G≅N×K where K=Zp and N≅Zp(2)×Zp2;
-
G≅N×K where K=Zp and N≅Q;
-
G≅N×K where K=Zp and N≅E2;
-
G≅N×K where K=Zp and N≅D×Z2;
-
G≅N×K where K=Zp and N≅E1×Zp(2);
-
G≅N×K where K=Zp(2) and N≅Zp2×Zp;
-
G≅N×K where K=Zp(2) and N≅D;
-
G≅N×K where K=Zp(2) and N≅E1×Zp;
-
G≅N×K where K=Zp2×Zp
and N≅Zp2;
-
G≅N×K where K=Zp(3)
and N≅E1;
-
G≅N×K where K=Zp(3)
and N≅Zp2.
Proof.
The necessity of theorem follows from the fact that G=N×K and Corollary 1.2.
For sufficiency first suppose that N is an elementary abelian p-group. Then nm−4=nd(K).
Since ∣N⊗Kab∣=pnm−4 by Corollary 1.2 and also it is of order pnd(K).
So n(m−d(K))=4, which implies that n=1,2 or 4. Therefore N≅Zp
and K is any group with d(K)=m−4 or N≅Zp(2) and
K is any group with d(K)=m−2, or N≅Zp(4) and
K is any group with d(K)=m−1.
Now suppose that N is not an elementary abelian p-group .
Then using Corollary 1.2, we have
∣Nab⊗Kab∣>pnm−4 and so md(N)>nm−4 which
implies that m(n−d(N))<4. Therefore m=0,1,2,3 .
If m=0,
then K=1 and N is one of the groups which are listed in
Theorems 1.4 and 1.6.
If m=1, then K=Zp and d(N)=n−1,n−2
or n−3, respectively. It follows that ∣Nab⊗K∣=pn−1,pn−2 or pn−3. So Corollary 1.2 implies that
∣M(N)∣=p2n2−n−3,p2n2−n−2, or
p2n2−n−1. In the first case N is Zp×Zp×Zp2,Q,E2,,D×Z2 or E1×Zp×Zp and other cases are impossible, by Theorem 1.3.
If m=2, then d(N)=n−1 and
K=Zp×Zp or K=Zp2. In the
first case ∣Nab⊗K∣=p2(n−1) and so ∣M(N)∣=p2n2−n−2. Therefore N is Zp×Zp2,D or Zp×E1. In the second
case Nab≅Zp(n−1) or Nab≅Zp2×Zp(n−2). If Nab is an
elementary abelian p-group, then ∣Nab⊗K∣=p(n−1)
and so ∣M(N)∣=p2n2+n−6 which is impossible and
if Nab≅Zp2×Zp(n−2),
then ∣M(N)∣=p2n2+n−8 which is impossible too.
If m=3,
then d(N)=n−1 and K is an abelian p-group of order p3
or an extra special p-group of order p3. In the first
case we have three possibility for K. The first possibility is
K≅Zp(3),. Then similar to the previous part, one
can see that ∣M(N)∣=p2n2−n−1 and so N≅E1
or Zp2. The second possibility is K≅Zp3. This implies that n=1 which is a
contradiction. The third possibility is K≅Zp2×Zp. which implies that n=2 and N≅Zp2.
In the second case, if K is an extra special p-group of order
p3, then Nab≅Zp(n−1) or Nab≅Zp2×Zp(n−2). This implies that
∣Nab⊗K∣=∣Nab⊗Zp(2)∣=p2n−2 and so n=1 which is a contradiction.
Hence the proof is complete.
∎
Theorem 2.2**.**
*By the previous assumption, t=5 if and only if G is isomorphic to one of the following groups:
-
G≅N×K where N≅Zp and K is any
group with d(K)=m−5;
-
G≅N×K where N≅Zp(5) and K
is any
group with d(K)=m−1;
-
G=N≅D×Z2(3);
-
G=N≅E1×Zp(4);
-
G=N≅E2×Zp(2);
-
G=N≅E4×Zp;
-
G=N≅ extra special p-group of order p5;
-
G=N≅⟨a,b∣ap2=bp2=1,[a,b,a]=[a,b,b]=1,[a,b]=ap⟩;
-
G=N≅⟨a,b∣ap2=bp=1,[a,b,a]=[a,b,b]=ap,[a,b,b,b,]=1⟩;
-
G=N≅⟨a,b∣ap2=bp=1,[a,b,a]=[a,b,b,b,]=1,[a,b,b]=anp⟩ where n is a fixed quadratic non-residue of p and p=3 ;
-
G=N≅⟨a,b∣ap2=1,b3=a3,[a,b,a]=[a,b,b,b,]=1,[a,b,b]=a6⟩;
-
G=N≅⟨a,b∣ap=1,bp=[a,b,b],[a,b,a]=[a,b,b,b,]=[a,b,b,a]=1⟩;
-
G=N≅D16;
-
G=N≅⟨a,b∣a4=b4=1,a−1ba=b−1⟩;
-
G=N≅Q×Z2(2);
-
G=N≅(D×Z2)>⊲Z2;
-
G=N≅(Q×Z2)>⊲Z2;
-
G=N≅Z2×<a,b,c∣a2=b2=c2=1,abc=bca=cab>;
-
G=N≅Zp3×Zp;
-
G=N≅Zp2×Zp(4);
-
G≅N×K where K=Zp and N≅D×Z2(2);
-
G≅N×K where K=Zp and N≅Q×Z2;
-
G≅N×K where K=Zp and N≅<a,b,c∣a2=b2=c2=1,abc=bca=cab>;
-
G≅N×K where K=Zp and N≅E4;
-
G≅N×K where K=Zp and N≅E1×Zp(3);
-
G≅N×K where K=Zp and N≅Zp(4)>⊲Zp;
-
G≅N×K where K=Zp and N≅E2×Zp;
-
G≅N×K where K=Zp and N≅E2×Zp;
-
G≅N×K where K=Zp(2) and N≅E1×Zp(2);
-
G≅N×K where K=Zp(2) and N≅Zp(2)×Zp2;
-
G≅N×K where K=Zp(2) and N≅Q;
-
G≅N×K where K=Zp(2) and N≅E2;
-
G≅N×K where K=Zp(2) and N≅D×Z2;
-
G≅N×K where K=Zp2 and N≅E1;
-
G≅N×K where K=Zp2 and N≅Zp×Zp2;
-
G≅N×K where K=Zp3 and N≅Zp2;
-
G≅N×K where K=Zp(3) and N≅D;
-
G≅N×K where K=Zp(3) and N≅E1×Zp;
-
G≅N×K where K=Zp(3) and N≅Zp×Zp2;
-
G≅N×K where K is an extra special P-group
and N≅Zp2;
-
G≅N×K where K=Zp(4) and N≅E1,
-
G≅N×K where K=Zp(2)×Zp and N≅Zp2.
Proof.
The proof of this theorem is similar to the proof
of the previous theorem so we left the details to the reader.
Necessity is straight forward. For sufficiency first suppose that N is an elementary abelian p-group . Then
n(m−d(K))=5, so n=1 or 5. If n=1, then N≅Zp and K is any group with d(K)=m−5 and n=5
implies that N≅Zp(5) and K is any group with d(K)=m−1.
Suppose that N is not an elementary abelian p-group. Then we have
∣Nab⊗Kab∣>pnm−5, by Corollary 1.2. It follows
that md(N)>nm−5 and thus m(n−d(N))<5, which implies that
m=0,1,2,3 or 4. If m=0, then K=1 and N is one of the
groups that are listed in Theorems 1.5 and 1.7.
If m=1, then K=Zp
and d(N)=n−i, for 1≤i≤4. Therefore by Corollary
1.2, ∣M(N)∣=p2n(n−1)−(5−i) for 1≤i≤4,
respectively. It follows that N≅Zp2×Zp(3), D×Z2(2), Q×Z2, E4, E1×Zp(3), E2×Zp, N≅Zp3 or ⟨a,b,c∣a2=b2=c2=1,abc=bca=cab⟩, by Theorems 1.6, 1.4 and 1.3.
If m=2, then
K=Zp2 or K=Zp×Zp and
d(N)=n−1
or d(N)=n−2. First suppose that K=Zp×Zp . If d(N)=n−1, then ∣M(N)∣=p2n2−n−3.
It follows that N≅Zp×Zp×Zp2, N≅E1×Zp×Zp,
N≅Q, N≅E2 or N≅D×Z2. If
d(N)=n−2, then ∣N⊗K∣=p2(n−2). This implies that
∣M(N)∣=p2n2+n−2 and so n<1 which is
impossible.
Now suppose that K=Zp2. If d(N)=n−1 and
Nab≅Zp(n−1), then ∣M(N)∣=p2n2+n−8 which implies that n=3 and ∣M(N)∣=p2. Therefore N≅E1. If d(N)=n−1 and Nab≅Zp2×Zp(n−2), then n=3 or n=4. For n=3 there is not any structure for N and n=4 implies that N≅Zp2×Zp.
If d(N)=n−2, then Nab≅Zp(n−2) or
Nab≅Zp3×Zp(n−3) or
Nab≅Zp2×Zp2×Zp(n−4). Therefore similar to the previous case one can see that n<5 which is
impossible.
If m=3, then d(N)=n−1 and K is an abelian p-group of
order p3 or is an extra special p-group of order p3. In
the first case we have three possibility for K. The first
possibility is K≅Zp3. Now if Nab≅Zp(n−1), then n=1 which is impossible and if Nab≅Zp2×Zp(n−2), then N≅Zp2. The second possibility is K≅Zp2×Zp. In this case there is no
structure for N. The third possibility is K≅Zp(3). Thus ∣M(N)∣=p2n2−n−2 and so
N≅D, N≅E1×Zp or N≅Zp2×Zp, by Theorem 1.3.
Now suppose that K is an extra special p-group of order
p3. Then Nab≅Zp(n−1) or Nab≅Zp2×Zp(n−2). If Nab is an elementary abelian, then
there is no structure for N. Otherwise N=Zp2.
If m=4, then d(N)=n−1. So Nab≅Zp(n−1)
or Nab≅Zp2×Zp(n−2). In
the first case ∣Nab⊗Kab∣=∣Zp(n−1)⊗Kab∣≤pd(K)(n−1).
Now suppose that d(K)<4. Then we have ∣Nab⊗Kab∣≤p3(n−1). Therefore ∣M(N)∣≥p−3(n−1)+(8n+n2−n−10)/2,
by Corollary 1.2. So n<2 which is impossible. If d(K)=4, then we have
K≅Zp(4). Hence ∣Nab⊗Kab∣=p4(n−1) which implies that ∣M(N)∣=p(n2−n)/2−1.
So N≅E1, by Theorem 1.3.
In the second case, suppose that K is not abelian. Then ∣Nab⊗Kab∣=∣Zp2×Zp(n−2)⊗Kab∣=∣Zp2⊗Kab∣∣Zp(n−2)⊗Kab∣≤p3pd(K)(n−2)≤p3(n−1) which implies that
n<2 and it is impossible.
If K is abelian, then K≅Zp2×Zp(2). thus ∣Nab⊗K∣=p3n−2. It
follows that N≅Zp2. This completes the proof.
∎