1 Introduction

Throughout only finite groups are considered. Terminologies and notations employed agree with standard usage, as in Robinson [8].

Two subgroups \(H\) and \(K\) of a group \(G\) are said to be permutable if \(HK = KH\). The subgroup \(H\) is said to be \(S\)-quasinormal in \(G\) if \(H\) permutes with every Sylow subgroups of \(G\), i.e., \(HP = PH\) for any Sylow subgroup \(P\) of \(G\). This concept was introduced by Kegel in [7] and has been studied widely by many authors, such as [2, 9]. Recently, There is a generalization of \(S\)-quasinormality in [14]. The subgroup \(H\) is called \(s\)-semipermutable in \(G\) if \(H\) permutes with every Sylow \(p\)-subgroup of \(G\) with \((|H |, p )=1\). An \(s\)-semipermutable subgroup is no need to be an \(S\)-quasinormal subgroup. \(S_3\) is a counter-example. On the other hand, Wang [10] introduced the concept of \(c\)-normal subgroups. The subgroup \(H\) is said to be \(c\)-normal in \(G\) if there exists a normal subgroup \(U\) of \(G\) such that \(G= HU\) and \(H\cap U\) is contained in \(H_G\), where \(H_G\) is the maximal normal subgroup of \(G\) which is contained in \(H\). The \(c\)-normality is a generalization of the normality. Applying the \(c\)-normality of subgroups, Wang obtained new criteria for supersolvability of groups. In 2007, Wei and Wang [11] introduced the concept of \(c^*\)-normal subgroups which is both \(c\)-normality and \(S\)-quasinormal embedding and used the \(c^*\)-normality of maximal subgroups to give some necessary and sufficient conditions for a group to be \(p\)-nilpotent, \(p\)-supersolvable or supersolvable. Based on the observation above concepts, we note that \(c^*\)-normal subgroups and \(s\)-semipermutable subgroups are two different concepts. There are examples to show that \(s\)-semipermutable subgroups are not \(c^*\)-normal subgroups and in general the converse is also false. In this paper, we investigate \(s\)-semipermutable and \(c^*\)-normal subgroups of \(G\) and give criteria for a group belonging to \(\fancyscript{F}\). Some interesting results are obtained and known results on this topic are generalized.

2 Preliminaries

Lemma 2.1

[11, Lemma 2.3] Let \(H\) be a subgroup of a group \(G\).

  1. (1)

    If \(H\) is \(c^*\)-normal in \(G\) and \(H\le M\le G\), then \(H\) is \(c^*\)-normal in \(M\).

  2. (2)

    Let \(N\lhd G\) and \(N \le H\). Then \(H\) is \(c^*\)-normal in \(G\) if and only if \(H{/}N\) is \(c^*\)-normal in \(G{/}N\).

  3. (3)

    Let \(\pi \) be a set of primes, \(H\) a \(\pi \)-subgroup of \(G\) and \(N\) a normal \(\pi ^\prime \)-subgroup of \(G\). If \(H\) is \(c^*\)-normal in \(G\), then \(HN{/}N\) is \(c^*\)-normal in \(G{/}N\).

Lemma 2.2

[14, Property] Suppose that \(H\) is an \(s\)-semipermutable subgroup of \(G\). Then

  1. (1)

    If \(H \le K\le G\), then \(H\) is \(s\)-semipermutable in \(K\).

  2. (2)

    Let \(N\) be a normal subgroup of \(G\). If \(H\) is a \(p\)-group for some prime \(p\in \pi (G)\), then \(HN{/}N\) is \(s\)-semipermutable in \(G{/}N\).

  3. (3)

    If \(H\le O_p(G)\), then \(H\) is \(S\)-quasinormal in \(G\).

Lemma 2.3

[9, 11] Suppose that \(U\) is \(S\)-quasinormally embedded in a group \(G\), and that \(H\le G\) and \(K \unlhd G\).

  1. (1)

    If \(U\le H\), then \(U\) is \(S\)-quasinormally embedded in \(H\).

  2. (2)

    \(UK\) is \(S\)-quasinormally embedded in \(G\) and \(UK{/}K\) is \(S\)-quasinormally embedded in \(G{/} K\).

  3. (3)

    If \(K\le H\) and \(H{/} K\) is \(S\)-quasinormally embedded in \(G{/}K\), then \(H\) is \(S\)-quasinormally embedded in \(G\).

  4. (4)

    A \(p\)-subgroup \(H\) of \(G\) is \(S\)-quasinormal in \(G\) if and only if \(N_G(H)\ge O^p(G)\) for some prime \(p\in \pi (G).\)

Lemma 2.4

[11, Lemma 2.8] Let \(G\) be a group and let \(p\) be a prime number dividing \(| G|\) with \(( | G|,p -1) = 1\).Then

  1. (1)

    If \(N\) is normal in \(G\) of order \(p\), then \(N\) lies in \(Z( G)\);

  2. (2)

    If \(G\) has cyclic Sylow \(p\)-subgroups, then \(G\) is \(p\)-nilpotent;

  3. (3)

    If \(M\) is a subgroup of \(G\) with index \(p\), then \(M\) is normal in \(G\).

Lemma 2.5

[1, A, Lemma 1.2] Let \(U , V\) and \(W\) be subgroups of a group \(G\). The following statements are equivalent.

  1. (1)

    \(U\cap V W =(U\cap V )( U\cap W )\).

  2. (2)

    \(U V\cap U W = U (V \cap W )\).

Lemma 2.6

[4, 6.4.8] Let \(H\), \(K\) be subgroups of the group \(G\) such that

$$\begin{aligned} (|G : H |, |G : K |)=1. \end{aligned}$$

Then \(G = HK\) and \(|G : H\cap K | = |G : H || G : K |\).

Lemma 2.7

[11, Lemma 2.5] Let \(G\) be a group, \(K\) an \(S\)-quasinormal subgroup of \(G\) and \(P\) a Sylow \(p\)-subgroup of \(K\), where \(p\) is a prime. If \(K_G=1\), then \(P\) is \(S\)-quasinormal in \(G\).

Lemma 2.8

[11, Theorem 4.1] Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G{/} H\in \fancyscript{F}\). If all maximal subgroups of any Sylow subgroup of \(H\) are \(c^*\)-normal in \(G\), then \(G\in \fancyscript{F}\).

Lemma 2.9

[11, Theorem 4.3] Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G{/} H\in \fancyscript{F}\). If all maxim al subgroups of any Sylow subgroup of \(\fancyscript{F}^*(H)\) are \(c^*\)-normal in \(G\), then \(G\in \fancyscript{F}\).

Lemma 2.10

[5, X, 13] Let \(G\) be a group. If \(F^*(G)\) is solvable, then \(F^*(G)= F (G).\)

Lemma 2.11

[5, IV, Satz 4.7] If \(P\) is a Sylow \(p\)-subgroup of \(G\) and \(N\unlhd G\) such that \(P\cap N\le \Phi (P )\), then \(N\) is \(p\)-nilpotent.

3 Main results

Theorem 3.1

Let \(G\) be a group and \(P\) a Sylow \(p\)-subgroup of \(G\), where \(p\) is the smallest prime dividing \(|G|\). If every maximal subgroup of \(P\) is either \(s\)-semipermutable or \(c^*\)-normal in \(G\), then \(G\) is \(p\)-nilpotent.

Proof

Asuume that the theorem is false and \(G\) is a counterexample with minimal order. We will consider the following steps.

  1. (1)

    \(G\) has a unique minimal normal subgroup \(N\) such that \(G{/}N\) is \(p\)-nilpotent and \(\Phi ( G)=1\).

    Let \(N\) be a minimal normal subgroup of \(G\). We have to show \(G{/}N\) satisfies the hypotheses of the theorem. Let \(M{/}N\) be a maximal subgroup of \(PN{/}N\). We can see \(M = P_1 N\) for some maximal subgroup \(P_1\) of \(P\). By the hypotheses, \(P_1\) is either \(s\)-semipermutable or \(c^*\)-normal in \(G\). If \(P_1\) is \(c^*\)-normal in \(G\), then there is a normal subgroup \(K_1\) of \(G\) such that \(G = P_1 K_1\) and \(P_1\cap K_1\) is \(S\)-quasinormally embedded in \(G\). Then \(G{/}N= M{/}N\cdot K_1 N{/}N= P_1 N{/}N\cdot K_1 N{/}N\). It is easy to see that \(K_1 N{/}N\) is normal in \(G{/}N\). Since \((|N : P_1\cap N |, |N : K_1\cap N |) = 1, (P_1\cap N )( K_1\cap N ) = N = N \cap G = N\cap (P_1 K_1)\) by Lemma 2.6. Now using Lemma 2.5, \(( P_1 N )\cap (K_1 N ) = ( P_1\cap K_1)N\). It follows that \(( P_1 N ){/}N\cap (K_1 N ){/}N =(P_1\cap K_1)N{/}N\) is \(S\)-quasinormally embedded in \(G{/}N\) by Lemma 2.3. Thus \(M{/}N\) is \(c^*\)-normal in \(G{/}N\). If \(P_1\) is \(s\)-semipermutable in \(G\), then \(M{/}N= P_1 N{/}N\) is \(s\)-semipermutable in \(G{/}N\) by Lemma 2.2. Consequently, \(G{/}N\) satisfies the hypotheses of the theorem. The choice of \(G\) yields that \(G{/}N\) is \(p\)-nilpotent. The uniqueness of \(N\) and \(\Phi (G)=1\) are obvious.

  2. (2)

    \(O_{p^\prime }(G)=1\).

    If \(O_{p^\prime }(G)\ne 1\), then \(N \le O_{p^\prime }(G)\) by (1). By Lemmas 2.1 and 2.2, \(G{/}N\) satisfies the hypotheses, hence \(G{/}N\) is \(p\)-nilpotent. Now the \(p\)-nilpotency of \(G{/}N\) implies the \(p\)-nilpotency of \(G\), a contradiction.

  3. (3)

    \(O_p(G)=1\) and \(G\) is not solvable. paraIf \(O_p(G)\ne 1\), (1) yields \(N\le O_p(G)\) and \(\Phi ( O_p(G))\le \Phi (G)=1\). Hence, \(G\) has a maximal subgroup \(M\) such that \(G = M N\) and \(M\cap N = 1\). Since \(O_p(G)\cap M\) is normalized by \(N\) and \(M\), \(O_p(G)\cap M\) is normal in \(G\). The uniqueness of \(N\) yields \(N = O_p(G)\). Obviously \(P = N M_p = N (P\cap M )\). Since \(P\cap M < P\), we can take a maximal subgroup \(P_1\) of \(P\) such that \(P\cap M\le P_1<\cdot P\). Then \(P = N P_1\) and \(P\cap M= P_1\cap M\). By the hypotheses of theorem, \(P_1\) is either \(s\)-semipermutable or \(c^*\)-normal in \(G\). If \(P_1\) is \(c^*\)-normal in \(G\), then there is a normal subgroup \(K_1\) such that \(G = P_1 K_1\) and \(P_1\cap K_1\) is \(S\)-quasinormally embedded in \(G\). Thus \(P_1\cap K_1\) is a Sylow \(p\)-subgroup of some \(S\)-quasinormal subgroup \(K\) of \(G\). If \(K_G \ne 1\), by (1) we have that \(N\le K_G\). Hence, \(P=NP_1\le P_1\), a contradiction. If \(K_G = 1\), by Lemma 2.7 we have that \(P_1\cap K_1\) is \(S\)-quasinormal in \(G\). It follows that \(P_1\cap K_1\) is normalized by \(P\) and \(O^p(G)\). Now we know \(P_1\cap K_1\lhd G\). If \((P_1\cap K_1)M=G\), then \(P_1M=PM=G\) and so \(P_1 = P\), a contradiction. Thus \((P_1\cap K_1)M=M\) and so \((P_1\cap K_1)\le M\). On the other hand, \(P_1\cap K_1\le N\). We know \(P_1\cap K_1\le N\cap M\). Hence, \(P_1\cap K_1=1\) and so \(|P\cap K_1|=p\). So \(|K|_p= p\). By the uniqueness of \(N\), we have that \(N\le K_1\), of course, \(N\) is a cyclic group of order \(p\). By Lemma 2.4, \(N\le Z(G)\). Since \(G{/}N\) is \(p\)-nilpotent, \(G\) is also \(p\)-nilpotent, a contradiction. Now suppose that \(P_1\) is \(s\)-semipermutable in \(G\). Then \(P_1 M_q\) is a group for \(q\ne p\). Therefore, \(P_1<M_p , M_q|q\in \pi (M ), q\ne p>= P_1 M\) is a group. Then \(P_1 M = M\) or \(G\) by maximality of \(M\). If \(P_1 M = G\), then \(P = P\cap P_1 M = P_1(P \cap M ) = P_1\), which is a contradiction. If \(P_1 M = M\), then \(P_1\le M\). Hence, \(P_1\cap N = 1\) and \(N\) is of prime order. Then the \(p\)-nilpotency of \(G{/}N\) implies the \(p\)-nilpotency of \(G\), a contradiction. Combining this with (2), it is easy to see that \(G\) is not solvable, now thus (3) holds.

  4. (4)

    For any \(q\ne p\), \(P G_q < G\), where \(G_q\) is a Sylow \(q\)-subgroup of \(G\). That is to say, \(P G_q\) is \(p\)-nilpotent.

    At first, we have \(NP=G\). In fact, if \(NP< G\), then \(NP\) is \(p\)-nilpotent since \(NP\) satisfies the hypotheses of theorem. Hence, \(N\) is \(p\)-nilpotent and by (1) we know \(N\) is a nontrivial \(p\)-group, but this is a contradiction with (3). So we have that \(NP=G\). If for all \(P_1<\cdot P\), we have that \(NP_1< G\). Then \((P\cap N)P_1<P\) and so \(P\cap N\le P_1\). Hence, \(P\cap N\le \Phi (P)\) and \(N\) is \(p\)-nilpotent by Lemma 2.11, a contradiction. So there exists \(P_1<\cdot P\) such that \(G=NP_1\). By the hypotheses, if \(P_1\) is \(c^*\)-normal in \(G\), then there is a normal subgroup \(K\) such that \(G = P_1 K_1\) and \(P_1\cap K_1\) is \(S\)-quasinormally embedded in \(G\). So \(P_1\cap K_1 \in Syl_p(K)\), where \(K\) is \(S\)-quasinormal in \(G\). If \(K_G\ne 1\), then \(N\le K_G\). It follows that \(P_1\cap K_1\cap N\in Syl_p(N)\). Now by \(G=NP_1\) we get \(P_1\in Syl_p(G)\), a contradiction. So \(K_G=1\). By Lemma 2.7 we have that \(P_1\cap K_1\) is \(S\)-quasinormal in \(G\), so \(P_1\cap K_1 \le O_p(G)=1\) and \(P_1\cap K_1=1\). Moreover, \(|P\cap K_1|=p\) and so \(|K_1|_p=p\). By Lemma 2.4 we know \(K_1\) is \(p\)-nilpotent. Of course, \(N\) is also \(p\)-nilpotent, a contradiction. From [5, IV, Satz 2.8], it follows that \(P\) is non-cyclic. We could take a maximal subgroup \(P_2\) of \(P\) satisfying \(G=NP_2\). By the same argument, we know that \(P_2\) cannot be \(c^*\)-normal in \(G\). Now suppose that \(P_i\) is \(s\)-semipermutable in \(G\) and \(P_iG_q\) is a group, \(i = 1, 2\), where \(G_q\) is a Sylow \(q\)-subgroup of \(G\). Thus we have \(P_1, P_2\) such that \(P= P_1 P_2\). Hence \(P G_q\) is a group. By (3) and the famous \(p^aq^b\)-theorem we infer \(P G_q\) is a proper subgroup of \(G\). Therefore, \(P G_q\) is \(p\)-nilpotent by the minimality of \(G\).

  5. (5)

    The final contradiction.

By (4) we have \([P , G_q]\le G_q\) for any \(q\ne p\). Suppose that \(S_1\) is an arbitrary subgroup of \(P\). Let \(N_G (S_1 ) = N_1\). Since \([ S_1, (N_1)_q]\le S_1\cap G_q= 1, S_1\) is centralized by \((N_1)_{p^\prime }\). Thus \(G\) is \(p\)-nilpotent by the famous Frobenius Theorem [8, 10.3.2], which is the final contradiction. \(\square \)

Corollary 3.2

Suppose that \(G\) is a group and \(P\) a Sylow subgroup of \(G\). If every maximal subgroup of \(P\) is either \(s\)-semipermutable or \(c^*\)-normal in \(G\), then \(G\) has a Sylow tower of supersolvable type.

Proof

Let \(p\) be the smallest prime dividing \(|G|\) and \(P\) a Sylow \(p\)-subgroup of \(G\). By hypothesis, every maximal subgroup of \(P\) is either \(s\)-semipermutable or \(c^*\)-normal in \(G\). In particular, \(G\) satisfies the condition of Theorem 3.1, so \(G\) is \(p\)-nilpotent. Let \(U\) be the normal \(p\)-complement of \(G\). By Lemmas 2.1 and 2.2, \(U\) satisfies the hypothesis. It follows by induction that \(U\), and hence \(G\) possesses the Sylow tower property of supersolvable type. \(\square \)

Corollary 3.3

[6, Theorem 3.1] Let \(G\) be a group and \(P = G_p\) a Sylow \(p\)-subgroup of \(G\), where \(p\) is the smallest prime dividing \(|G|\). If every maximal subgroup of \(P\) is either \(s\)-semipermutable or \(c\)-normal in \(G\), then \(G\) is \(p\)-nilpotent.

Corollary 3.4

Suppose that \(G\) is a group and \(P\) a Sylow subgroup of \(G\). If every maximal subgroup of \(P\) is either \(s\)-semipermutable or \(c\)-normal in \(G\), then \(G\) has a Sylow tower of supersolvable type.

We are now in a position to unify and generalize Theorem 4.1 and Theorem 4.3 in [11].

Theorem 3.5

Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G{/} H\in \fancyscript{F}\). If all maximal subgroups of any Sylow subgroup of \(H\) are either \(c^*\)-normal or \(s\)-semipermutable in \(G\), then \(G\in \fancyscript{F}\).

Proof

Suppose that \(P\) is a Sylow \(p\)-subgroup of \(H\) , \(\forall p\in \pi (H )\). Since every maximal subgroups of \(P\) are either \(c^*\)-normal or \(s\)-semipermutable in \(G\), thus in \(H\) by Lemmas 2.1 and 2.2. By Corollary 3.2 we konw that \(H\) has a Sylow tower of supersolvable type. Let \(q\) be the maximal prime divisor of \(|H |\) and \(Q\in Syl_q(H )\). Then \(Q\) char \(H\unlhd G\). Since \(( G{/}Q, H{/}Q)\) satisfies the hypotheses of the theorem, by induction, \(G{/}Q\in \fancyscript{F}.\) Since \(Q \le O_q(G)\), every maximal subgroups of \(Q\) are either \(S\)-quasinormal or \(c^*\)-normal in \(G\) by Lemma 2.2, in particular, \(c^*\)-normal in \(G\). So \(G\in \fancyscript{F}\) by Lemma 2.8. \(\square \)

Theorem 3.6

Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G{/} H\in \fancyscript{F}\). If all maxim al subgroups of any Sylow subgroup of \(F^*(H)\) are either \(c^*\)-normal or \(s\)-semipermutable in \(G\), then \(G\in \fancyscript{F}\).

Proof

Suppose that \(P\) is a Sylow \(p\)-subgroup of \(F^*(H)\) , \(\forall p\in \pi (F^*(H) )\). Since every maximal subgroups of \(P\) are either \(c^*\)-normal or \(s\)-semipermutable in \(G\), thus in \(F^*(H)\) by Lemmas 2.1 and 2.2. By Corollary 3.2 we konw that \(F^*(H)\) has a Sylow tower of supersolvable type, in particular, \(F^*(H)\) is Solvable. By Lemma 2.10 we have that \(F^*(H)=F(H)\). Since \(F^*(H)_p=F(H)_p\), \(\forall p\in (H)\), every maximal subgroups of \(P\) are either \(c^*\)-normal or \(S\)-quasinormal in \(G\) by Lemma 2.2, in particular, \(c^*\)-normal in \(G\). Applying Lemma 2.9 we get that \(G\in \fancyscript{F}\). \(\square \)

Corollary 3.7

[12, Theorem 3.1] Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G{/} H\in \fancyscript{F}\). If all maximal subgroups of every Sylow subgroup of \(F^*(H)\) are c-normal in \(G\), then \(G\in \fancyscript{F}\).

Corollary 3.8

[14, Theorem 1] Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G{/} H\in \fancyscript{F}\). If all maximal subgroups of any Sylow subgroup of \(H\) are \(s\)-semipermutable in \(G\), then \(G\in \fancyscript{F}\).

Corollary 3.9

[14, Theorem 2] Let \(\fancyscript{F}\) be a saturated formation containing \(\fancyscript{U}\). Suppose that \(G\) is a group with a normal subgroup \(H\) such that \(G{/} H\in \fancyscript{F}\). If all maximal subgroups of any Sylow subgroup of \(F^*(H)\) are \(s\)-semipermutable in \(G\), then \(G\in \fancyscript{F}\).