Home Mathematics On the triple tensor product of prime-power groups
Article Publicly Available

On the triple tensor product of prime-power groups

  • EMAIL logo and
Published/Copyright: May 12, 2020

Abstract

Let G be a finite p-group, and let 3G be its triple tensor product. In this paper, we obtain an upper bound for the order of 3G, which sharpens the bound given by G. Ellis and A. McDermott, [Tensor products of prime-power groups, J. Pure Appl. Algebra 132 1998, 2, 119–128]. In particular, when G has a derived subgroup of order at most p, we classify those groups G for which the bound is attained. Furthermore, by improvement of a result about the exponent of 3G determined by G. Ellis [On the relation between upper central quotients and lower central series of a group, Trans. Amer. Math. Soc. 353 2001, 10, 4219–4234], we show that, when G is a nilpotent group of class at most 4, exp(3G) divides exp(G).

1 Introduction and motivation

Let G and H be groups equipped with an action (g,h)hg of G on H and an action (h,g)gh of H on G. The group G is assumed to act on itself by conjugation (g,g)gg=ggg-1, and H is assumed to act on itself similarly. Suppose that the various actions are compatible (see [6]). The non-abelian tensor product GH was defined in [6] as the group generated by symbols gh (for gG, hH), subject to the relations

ggh=(gghg)(gh)andghh=(gh)(ghhh)

for all g,gG and h,hH. It was shown in [11] that GH is a finite group (or a p-group) when G and H are finite groups (or p-groups).

If G=H and all actions are taken to be conjugation, then GG is called the non-abelian tensor square of G. Since conjugation yields compatible actions, there is a diagonal action of G on GG. Also, the tensor square GG acts on G by conjugation in G via the homomorphism GGG induced by the commutator map. It is readily seen that these actions are compatible, and we can thus construct the triple tensor product 3(G)=(GG)G (see [7]). This construction can be iterated to form the c-fold tensor product c(G) defined recursively by 1(G)=G and c+1(G)=c(G)G for c1. We shall write cG instead of c(G).

The following notation will be used throughout this paper:

  1. Gab:=G/G is the abelianization of G;

  2. (Cpi)k is the direct product of k copies of the cyclic group of order pi;

  3. Q8 is the quaternion group of order 8;

  4. E1 is the extraspecial group of order p3 and odd prime exponent p;

  5. Gp(α,β,γ;ρ,σ) is the 2-generator p-group of class 2 introduced in [16] with the following presentation:

    a,b[a,b]pγ=[a,b,a]=[a,b,b]=1,apα=[a,b]pρ,bpβ=[a,b]pσ.

An important aspect of the research into the tensor products of groups includes either establishing their structure or estimating their size and exponent. In this paper, our aim is to obtain some bounds for the size and exponent of 3G which yield a sharpening of the estimates given in [10, 9]. Focusing on the first, let us recall a result from [1] which indicates that |3E1|p12. On the other hand, the assertion |3G|pm2d+m1d2 established in [10, Corollary 3] implies that |3E1|p14, in which G is a d-generator group of order pm1 with derived subgroup of order pm2. Furthermore, we know |E1E1|=p6 (see [3] or [15]) which shows that the upper bound given in [13] for |GG| is attained by E1. So one would hope that the above bound for 3G could be sharpened to the point where E1 attains it. This motivates us to give a suitable estimate as follows.

Theorem 1.

Let G be a finite d-generator group of prime-power order pn with Gab of order pm and exponent pe. Then |3G|p(n-e)d2+md. Moreover, when |G|p, the bound is attained if and only if G is isomorphic to (Cpnd)d, E1 or Gp(α,α,1;1,1) with α>1.

In view of Theorem 1 together with the main result of [13], it might seem reasonable to consider the following conjecture.

Conjecture.

Let G be a finite d-generator group of prime-power order pn with Gab of order pm and exponent pe. Then |cG|p(n-e)dc-1+mdc-2.

P. Moravec [17] showed that, for locally finite groups of finite exponent, the exponent of their non-abelian tensor product can be bounded by a function depending only on the exponent of the group. In particular, if G is a finite nilpotent group of class at most 3, then exp(GG) divides exp(G). It is shown by an example that this result cannot be extended to groups of class 4 or more. In view of the above result and some extensive computational evidence, we pose the following conjecture.

Conjecture.

Let G be a finite nilpotent group of class at most c. Then exp(c-1G) divides exp(G).

In order to prove this conjecture for c=4, we establish the following result which improves the bound given by G. Ellis [9].

Theorem 2.

Let G be a finite nilpotent group of class k3. Then exp(3G) divides exp(G)[k2]-1, where [k2] is the smallest integer greater than or equal to k2. In particular, when G has nilpotency class at most 4, exp(3G) divides exp(G).

2 Basic results

At first, we mention some properties of non-abelian tensor products of groups.

Proposition 2.1 ([5]).

Let G and H be groups. The following relations hold in GH for all g,gG and h,hH:

  1. (g-1h)g=(gh)-1=(gh-1)h,

  2. g(hgh-1)=(gh)g(gh)-1,

  3. (ghg-1)h=(gh)(gh)-1h,

  4. (gh)(gh)(gh)-1=(gh)[g,h],

  5. [gh,gh]=(ghg-1)(hgh-1),

in which [g,h] may be interpreted as ghg-1G or hgh-1H.

Lemma 2.2.

Suppose G is a nilpotent group of class 3 and x,y,z,tG. Then the actions of G and GG on GG reduce to

(xy)[z,t]=(xy)(x[z,t,y])([z,t,x]y),
(2.1)(xy)z=(xy)([z,x]y)(x[z,y])([z,x,x]y)(x[z,y,y])×(x[z,x,y])([z,y,x]y)([z,x][z,y]).

Proof.

Apply the relations of GG, which are obtained by putting G=H in Proposition 2.1. ∎

It is shown in [13] that if G is a finite d-generator group of prime-power order pn whose abelianization is of order pm and exponent pe, then

|GG|p(n-e)d+m.

The following result shows that this estimate is sharp.

Proposition 2.3.

By the above assumptions and assuming |G|p, we have

|GG|=p(n-e)d+m

if and only if G is isomorphic with one of the following groups:

(Cpnd)d,Q8,E1,Gp(α,α,1;1,1)𝑤𝑖𝑡ℎα>1.

Proof.

First suppose G is abelian and

GCpn1×Cpn2××Cpnd,

where n1n2nd and n1+n2++nd=n=m. Then

GGCpnd×(Cpnd-1)3×(Cpnd-2)5××(Cpn1)2d-1.

Clearly, |GG|p(n1+n2++nd-1)d+n1+n2++nd=p(n-e)d+n and equality holds if and only if n1=n2==nd, as desired.

Assume

GCpandGabCpn1×Cpn2××Cpnd,

where n1n2nd and n1+n2++nd=n-1=m. Then, from [5, Proposition 9], we have |GG||GabGab||GG|. On the other hand, as above, we have |GabGab|p(n-e-1)d+n-1, and equality holds if and only if n1=n2==nd. Also, since GZ(G), G and G act on each other trivially. So, from [6, Proposition 2.4], we have GGGGab(Cp)d. Hence |GG|p(n-e)d+n-1, and obviously we must have n1=n2==nd. By this hypothesis, it follows that |GG|=pnd. But, from [14, Theorem 2.3], we know |GG|p(n-1)d+2, which implies that only 2-generator groups may satisfy our condition.

Thanks to [16, Theorem 36], when G is a 2-generator group of order p2α+1 with GabCpα×Cpα, and p>2,

|GG|={p4α+2ρifρσ,p4α+2σifρ>σ.

So if |GG|=p2n, the equality 2n=4α+2ρ implies that ρ=1, and consequently σ=1 for γ=1. It is also readily seen that the second case does not satisfy our condition because it implies ρ>1, which contradicts the fact that ργ=1. Thus, when p>2, only the group Gp(α,α,1;1,1) has tensor square of order p2n. As above, one can conclude from [16, Theorem 47] that, when p=2, only the groups G2(1,1,1;0,0) and G2(α,α,1;1,1) with α>1 have tensor square of order 22n. From the presentation of the group Gp(α,β,γ;ρ,σ), it follows that G2(1,1,1;0,0)Q8 and Gp(1,1,1;1,1)E1 for p>2. ∎

Let

3G=(GG)G

be the triple tensor product of a group G and x,y,zG. To simplify the notation, we shall identify xyz with (xy)z. Proposition 2.1 (iv) implies that, for all a,b,cG,

(xyz)abc=(xyz)[ab,c]=(xyz)[a,b]c=(xyz)[a,b,c].

So it follows that, when G is a nilpotent group of class at most 3, 3G is an abelian group, which makes the computations more convenient.

Lemma 2.4.

Let G be a group and x,y,zG, and let n be a positive integer.

  1. There are homomorphisms λ:3GGG and λ:3GG given by

    λ(xyz)=(xz)z(xy)-1=[x,y]z,
    λ(xyz)=[x,y,z],

    respectively, such that the actions of G on Kerλ and GG on Kerλ are trivial. Furthermore, Kerλ and Kerλ are central subgroups of 3G (see [6]).

  2. If [x,y]z=1, then xyzn=(xyz)n.

  3. If [x,y,z]=1, then (xy)nz=(xyz)n.

  4. If [x,y,z]Z(G), then x[y,z]n=(x[y,z])n.

Proof.

The proof is straightforward. ∎

The following well known identities are helpful in the discussion of nilpotent groups of class 3.

Lemma 2.5.

Let G be a nilpotent group of class 3 and x,y,zG. Then, for all positive integers m,n,

  1. (xy)n=xnyn[y,x](n2)[y,x,x](n3)[y,x,y](n2)+2(n3),

  2. [xm,yn]=[x,y]mn[x,y,x]n(m2)[x,y,y]m(n2).

3 Proof of the main theorems

Let G be a group, and let γk(G) be the k-th term of the lower central series of G. The following result plays a fundamental role in the proof of Theorem 1. It implies that, when the nilpotency class of G is k, the image of the canonical homomorphism (GG)γk(G)3G is contained in the image of the canonical homomorphism (γk(G)G)G3G.

Lemma 3.1.

For any nilpotent group G of class k, we have

Im((GG)γk(G)3G)Im((γk(G)G)G3G).

Proof.

Obviously, the result is true for k=2 by [1, Corollary 2.7]. Assume k=3 and a,b,x1,x2,x3G. The defining relations of GG and Lemma 2.1 imply that

(3.1)ab[x1,x2,x3]=abx3[x1,x2]x3-1=(x1x2x3)[a,b](x1x2x3)-1=((x1x2)(x1x2)-1x3[a,b])-1=([x1,x2]x3[a,b])-1.

As ab=[a,b]ba, we expand the element [x1,x2]x3ab in two ways. First, by using Lemma 2.4 (i), we have

[x1,x2]x3ab=([x1,x2]x3a)([x1,x2]x3b)a=([x1,x2]x3a)×((a[x1,x2,x3])([x1,x2]x3)[a,b]b)=([x1,x2]x3a)([x1,x2]x3[a,b])×([x1,x2]x3b)(a[x1,x2,x3]b)

Note that, by Lemma 2.4 (i),

a[x1,x2,x3][a,b]=(([x1,x2]x3)([x1,x2]x3)-1a[a,b])-1=(([x1,x2]x3a)([x1,x2]x3a)-1[a,b])-1=(([x1,x2]x3a)([x1,x2]x3a)-1)-1=1.

On the other hand,

[x1,x2]x3[a,b]ba=([x1,x2]x3[a,b])([x1,x2]x3b)×([x1,x2]x3[b,a])([x1,x2]x3a)×(b[x1,x2,x3]a).

Equating the terms from the two expansions yields

a[x1,x2,x3]b=([x1,x2]x3[b,a])(b[x1,x2,x3]a).

From here, we conclude that

[x1,x2]x3[b,a]=([x1,x2,x3]ab)-1([x1,x2,x3]ba).

So the result follows directly from equation (3.1). A similar argument to the one just presented generalizes the assertion for k>3. ∎

Proposition 3.2.

If G is a nilpotent group of class k, then the following sequence is exact:

(γk(G)G)G3G3Gγk(G)1.

Proof.

This follows by [9, Lemma 3] and Lemma 3.1. ∎

Lemma 3.3.

Let G be a finite d-generator nilpotent group of class at most 2. Then |3G||GG|d.

Proof.

Since the group G is finite, it is the direct product of its Sylow subgroups, G=S1××Sr. We can apply the distributive law in [12] to compute the tensor product 3G. When ij, the groups Si and Sj act trivially upon each other by conjugation; therefore, for any k, the cross term SiabSjabSkab is trivial. Also, when ik, the term (SiSi)abSkab would be trivial. So 3Gi=1r3Si, and it suffices to consider the case where G is a finite p-group. First suppose that G is abelian. Then GCpn1×Cpn2××Cpnd, where n1n2nd. Therefore,

GGCpnd×(Cpnd-1)3×(Cpnd-2)5××(Cpn1)2d-1=i=0d-1(Cpnd-i)2i+1,

and

3G[Cpnd×(Cpnd-1)3×(Cpnd-2)5××(Cpn1)2d-1]
  ×[Cpnd-1×(Cpnd-1)3×(Cpnd-2)5××(Cpn1)2d-1]
  ×[Cpnd-2×(Cpnd-2)3×(Cpnd-2)5××(Cpn1)2d-1]
  ××[Cpn1×(Cpn1)3×(Cpn1)5××(Cpn1)2d-1]
Cpnd×(Cpnd-1)2.3+1×(Cpnd-2)3.5+3+1
  ××(Cpn1)d(2d-1)+(2d-3)++3+1
=i=0d-1(Cpnd-i)3i2+3i+1.

Hence

|3G|=pnd+(2.3+1)nd-1+(3.5+3+1)nd-2++(d(2d-1)+(2d-3)++3+1)n1pdnd+3dnd-1+5dnd-2++(2d-1)dn1=|GG|d.

Now assume G has nilpotency class 2. Recall that GGt1××tl is an abelian group, where the ti’s are the generators of GG (see [2]). So, for any t=t1m1t2m2tlmlGG and any gG, we have

tg=i=1l(tig)mi3G.

Suppose G=x1,x2,,xd. Then, by applying the defining relations of 3G and [1], we conclude that any element tkg may be presented by products of elements of the form tixj for some 1il and 1jd. Thus

3G=tixj1il, 1jd.

On the other hand, observe that |tixj||ti| by Lemma 2.4 (iii). Since 3G is abelian, it follows that |3G||t1|d|t2|d|tr|d=|GG|d, as desired. ∎

We are now in a position to undertake the proof of Theorem 1.

Proof of Theorem 1.

We prove the theorem by induction on n. The case n=1 is certainly true. So assume the assertion holds for all p-groups of order less than pn, and let |G|=pn. If m=0, then G is abelian, and the result holds by Lemma 3.3 and [13]. Let m0, |γk(G)|=pr and γk+1(G)=1. Since

|(γk(G)G)G|=|(γk(G)G)abGab|=|(γk(G)Gab)Gab||γk(G)|d2,

the exact sequence given in Proposition 3.2 implies that

|3G||3Gγk(G)||(γk(G)G)G|p(n-r-e)d2+md+rd2=p(n-e)d2+md.

Now assume |G|p. By Lemma 3.3, we know that if |3G|=p(n-e)d2+md, then |GG|=p(n-e)d+m. So it suffices to verify whether the groups that appear in Proposition 2.3 attain the bound or not. Evidently, the bound is attained for G=(Cpnd)d. Assume G=Gp(α,α,1;1,1)=a,b with α1. If p>2, by [16, Theorem 36], we have

GG=aa,bb,(ab)(ba),ab,a[a,b],b[a,b](Cpα)4×(Cp)2.

Set

T=aa,bb,(ab)(ba)(Cpα)3,
S=ab,a[a,b],b[a,b]Cpα×(Cp)2.

Thanks to [5, Proposition 10] and [1], we get 3G(TG)×(SG). As T and G act on each other trivially, then TGTabGab(Cpα)6. In addition, we have (ab)pαa=(aba)pα=1, which shows that pα divides |aba|. Similarly, pα divides |abb| and p divides |a[a,b]a|, |a[a,b]b|, |b[a,b]a| and |b[a,b]b|. Moreover, one can verify that the map

θ:S×G(pα)2×(p)4
((ab)r(a[a,b])s(b[a,b])t,aibj[b,a]k)
(ir,jr,is+r(i2),jt+r(j2),js+rk,it+r(ij-k))

is a crossed pairing. Therefore, it determines a unique homomorphism

θ¯:SG(pα)2×(p)4

(see [5, Remark 3]), which implies that |aba|=|abb|=pα and |a[a,b]a|=|a[a,b]b|=|b[a,b]a|=|b[a,b]b|=p. Hence 3G(Cpα)8×(Cp)4 which tells us that the bound is attained by G. One can analogously show that the group G2(α,α,1;1,1) with α>1 attains the bound too.

Now it remains to consider the group G=Q8 with the presentation

a,b[a,b]2=[a,b,a]=[a,b,b]=1,a2=b2=[a,b].

From [16, Theorem 47], we know

GG=aa,bb,(ab)(ba),(aa)(bb)(ab)(C4)2×(C2)2.

Set T=aa,bb,(ab)(ba). As TGG, there is a homomorphism TG3G which tells us |aaa| divides 2. On the other hand, the natural homomorphism 3G3Gab implies that |aaa| is divisible by 2. So |aaa|=2. By the same manner, one can conclude that

|aab|=|bba|=|bbb|=|(ab)(ba)a|=|(ab)(ba)b|=2.

In addition, by applying the facts

1=((aa)(bb)(ab))2a=((aa)(bb)(ab)a)2

and 3Gab(C2)8, we get |(aa)(bb)(ab)a|=2. Similarly, we get |(aa)(bb)(ab)b|=2. Thus 3G(C2)8, whence the group Q8 does not attain the bound. ∎

Let G be a group given by a free presentation GF/R. The 2-nilpotent multiplier of G is the abelian group (2)(G)=(Rγ3(F))/[R,F,F] which is invariant of G (see [4]). It is well known that if G is finite, then (2)(G) is finite too. There are some known bounds for the order of (2)(G) (see for instance [8, Theorem 3] and [18, Theorem 3.9]). The following result, thanks to [7, Theorem 2.9], sharpens the latter for those groups with small enough number of generators whose abelianizations are not elementary abelian (in particular, for any 2-generator [or 3-generator] p-group G of order p5 [or of order p8] with exp(Gab)>p).

Corollary 3.4.

Let G be a finite d-generator group of prime-power order pn with Gab of order pm and exponent pe. Then |M(2)(G)||γ3(G)|p(n-e)d2+md.

In the rest of this paper, we work on the exponent of the triple tensor products of nilpotent groups of class 3. To facilitate the proof of Theorem 2, we shall take note of some useful identities.

Proposition 3.5.

Let G be a nilpotent group of class 3. Then, for any x,y,zG and any positive integer n, the following hold:

(3.2)xyzn=(xyz)n([z,x]yz)(n2)×(x[z,y]z)(n2)([z,x,x]yz)(n2)×(x[z,y,y]z)(n2)(x[z,x,y]z)(n2)×([z,y,x]yz)(n2)([z,x][z,y]z)(n2)×(z[z,x,y]z)(n3)([z,y,x]zz)(n3),
(3.3)(xy)nz=(xyz)n(xy[x,y,z])(n2).

Proof.

To prove (3.2), assume n=2. Then

xyz2=(xyz)z(xyz)=(xyz)((xy)zz).

By applying (2.1) of Lemma 2.2 and Lemma 2.4 (i), we obtain

xyz2=(xyz)2([z,x]yz)(x[z,y]z)×([z,x,x]yz)(x[z,y,y]z)(x[z,x,y]z)×([z,y,x]yz)([z,x][z,y]z).

When n=3, we have, by above, that

xyz3=(xyz)z(xyz2)=(xyz)3([z,x]yz)2(x[z,y]z)2×([z,x,x]yz)2(x[z,y,y]z)2(x[z,x,y]z)2×([z,y,x]yz)2([z,x][z,y]z)2×([z,x]yz)z(x[z,y]z)z([z,x,x]yz)×(x[z,y,y]z)(x[z,x,y]z)([z,y,x]yz)×([z,x][z,y]z).

But (2.1) of Lemma 2.2 implies that

([z,x]yz)z=([z,x]yz)([z,[z,x]]yz)([z,x][z,y]z).

On the other hand, it follows from Lemma 2.1 that

[z,x][z,y]z=(zy)[z,x](zy)-1z=(z[z,x,y]z)([z,x,z]yz).

So ([z,x]yz)z=([z,x]yz)(z[z,x,y]z). A similar argument shows that (x[z,y]z)z=(x[z,y]z)([z,y,x]zz). Therefore,

xyz3=(xyz)3([z,x]yz)3(x[z,y]z)3×([z,x,x]yz)3(x[z,y,y]z)3×(x[z,x,y]z)3([z,y,x]yz)3×([z,x][z,y]z)3(z[z,x,y]z)([z,y,x]zz).

Now the assertion follows by proceeding in the same manner.

(ii) If n=2, then

(xy)2z=(xyz)[x,y](xyz)=(xy[x,y,z]z)(xyz)=(xyz)2(xy[x,y,z]).

One can easily complete the proof by induction on n. ∎

P. Moravec [17] showed that, when G is a finite nilpotent group of class at most 3, exp(GG) divides exp(G). Using this fact and Proposition 3.5, we derive the following result.

Proposition 3.6.

Let G be a finite nilpotent group of class less than or equal to 3. Then exp(3G) divides exp(G).

Proof.

If G is a nilpotent group of class at most 2, the assertion holds by [9, Proposition 5]. Assume G has nilpotency class 3. By the same argument as in the proof of Lemma 3.3, we can assume that G is a finite p-group with exp(G)=pe. If n=2e, for any x,yG, by Lemma 2.5 (i), the identity

[y,x](n2)[y,x,y](n2)=1

holds. On the other hand, it follows by Lemma 2.5 (ii) that [x,y,x](n2)=1. Hence [y,x](n2)=1. Consequently, Lemma 2.4 (iii) and (iv) imply that

([z,x]yz)(n2)=(x[z,y]z)(n2)=([z,x][z,y]z)(n2)=1.

As above,

([z,x,x]yz)(n2)=(x[z,y,y]z)(n2)=(x[z,x,y]z)(n2)=([z,y,x]yz)(n2)=1.

Also, since (n3) is divisible by n, we have

(z[z,x,y]z)(n3)=([z,y,x]zz)(n3)=1.

Thus, by (3.2) of Proposition 3.5, we conclude that (xyz)n=1, as desired. Assume p>2 and n=pe. As [x,y][x,y,z]=1, then

(xy[x,y,z])(n2)=1.

Now (3.3) of Proposition 3.5 and [17, Theorem 2] give the result at once. ∎

Proof of Theorem 2.

By Proposition 3.6, the assertion holds for k=3. Assume k>3. Clearly, it follows from Lemma 2.4 (iii) and (iv) that

exp((γk-1(G)G)G)exp(γk-1(G)G)exp(G).

So, by invoking the exact sequence

(γk-1(G)G)G3G3Gγk-1(G)1

given in [9, proof of Proposition 5], as well as the induction hypotheses, we obtain

exp(3G)exp(G)[k-22]-1×exp(G)=exp(G)[k2]-1.

The following practical implication follows directly from [7, Theorem 2.9] and Theorem 2. As computation of 2-nilpotent multipliers is generally a difficult problem, this result ensures that, when considering a class of nilpotent groups of class at most 4, one may speed up computations of (2)(G) to a certain extent, at least for large groups.

Corollary 3.7.

For any finite nilpotent group G of class k3, we have

exp((2)(G))exp(G)[k2]-1,

where [k2] is the smallest integer greater than or equal to k2. In particular, if G is nilpotent of class at most 4, then exp(M(2)(G)) divides exp(G).


Communicated by Evgenii I. Khukhro


References

[1] E. Asheghi and S. Hadi Jafari, Triple tensor product of a nilpotent group of class two, submitted. Search in Google Scholar

[2] M. R. Bacon, On the nonabelian tensor square of a nilpotent group of class two, Glasgow Math. J. 36 (1994), no. 3, 291–296. 10.1017/S0017089500030883Search in Google Scholar

[3] M. R. Bacon and L.-C. Kappe, The nonabelian tensor square of a 2-generator p-group of class 2, Arch. Math. (Basel) 61 (1993), no. 6, 508–516. 10.1007/BF01196588Search in Google Scholar

[4] R. Baer, Representations of groups as quotient groups. I, II and III, Trans. Amer. Math. Soc. 58 (1945), 295–419. 10.1090/S0002-9947-1945-0015108-3Search in Google Scholar

[5] R. Brown, D. L. Johnson and E. F. Robertson, Some computations of nonabelian tensor products of groups, J. Algebra 111 (1987), no. 1, 177–202. 10.1016/0021-8693(87)90248-1Search in Google Scholar

[6] R. Brown and J.-L. Loday, Van Kampen theorems for diagrams of spaces, Topology 26 (1987), no. 3, 311–335. 10.1016/0040-9383(87)90004-8Search in Google Scholar

[7] J. Burns and G. Ellis, On the nilpotent multipliers of a group, Math. Z. 226 (1997), 405–428. 10.1007/PL00004348Search in Google Scholar

[8] J. Burns and G. Ellis, Inequalities for Baer invariants of finite groups, Canad. Math. Bull. 41 (1998), 385–391. 10.4153/CMB-1998-051-3Search in Google Scholar

[9] G. Ellis, On the relation between upper central quotients and lower central series of a group, Trans. Amer. Math. Soc. 353 (2001), no. 10, 4219–4234. 10.1090/S0002-9947-01-02812-4Search in Google Scholar

[10] G. Ellis and A. McDermott, Tensor products of prime-power groups, J. Pure Appl. Algebra 132 (1998), no. 2, 119–128. 10.1016/S0022-4049(97)00112-6Search in Google Scholar

[11] G. J. Ellis, The nonabelian tensor product of finite groups is finite, J. Algebra 111 (1987), no. 1, 203–205. 10.1016/0021-8693(87)90249-3Search in Google Scholar

[12] S. Hadi Jafari and S. M. Davarpanah, On the triple tensor product of groups of order p4, submitted. 10.1080/00927872.2021.2016787Search in Google Scholar

[13] S. Hadi Jafari, A bound on the order of non-abelian tensor square of a prime-power group, Comm. Algebra 40 (2012), no. 2, 528–530. 10.1080/00927872.2010.532845Search in Google Scholar

[14] S. Hadi Jafari, Categorizing finite p-groups by the order of their non-abelian tensor squares, J. Algebra Appl. 15 (2016), no. 5, Article ID 1650095. 10.1142/S021949881650095XSearch in Google Scholar

[15] S. Hadi Jafari, F. Saeedi and E. Khamseh, Characterization of finite p-groups by their non-abelian tensor square, Comm. Algebra 41 (2013), no. 5, 1954–1963. 10.1080/00927872.2011.645178Search in Google Scholar

[16] A. Magidin and R. F. Morse, Certain homological functors of 2-generator p-groups of class 2, Computational Group Theory and the Theory of Groups. II, Contemp. Math. 511, American Mathematical Society, Providence (2010), 127–166. 10.1090/conm/511/10046Search in Google Scholar

[17] P. Moravec, The exponents of nonabelian tensor products of groups, J. Pure Appl. Algebra 212 (2008), no. 7, 1840–1848. 10.1016/j.jpaa.2007.11.003Search in Google Scholar

[18] P. Niroomand and M. Parvizi, On the 2-nilpotent multiplier of finite p-groups, Glasg. Math. J. 57 (2015), no. 1, 201–210. 10.1017/S0017089514000263Search in Google Scholar

Received: 2019-11-07
Revised: 2020-03-03
Published Online: 2020-05-12
Published in Print: 2020-09-01

© 2020 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 27.3.2026 from https://www.degruyterbrill.com/document/doi/10.1515/jgth-2019-0161/html
Scroll to top button