() @ Appl. Gen. Topol. 14, no. 2 (2013), 159-169doi:10.4995/agt.2013.1586 c© AGT, UPV, 2013 Zariski topology on the spectrum of graded classical prime submodules Ahmad Yousefian Darani a and Shahram Motmaen b a Department of Mathematics and Applications, Faculty of Mathematical Sciences, University of Mohaghegh Ardabili, 56199-11367, Ardabil, Iran. (yousefian@uma.ac.ir, youseffian@gmail.com) b Young Researchers Club, Ardabil Branch Islamic Azad University, Ardabil, Iran. (sh.motmaen@gmail.com) Abstract Let R be a G-graded commutative ring with identity and let M be a graded R-module. A proper graded submodule N of M is called graded classical prime if for every a, b ∈ h(R), m ∈ h(M), whenever abm ∈ N, then either am ∈ N or bm ∈ N. The spectrum of graded classical prime submodules of M is denoted by Cl.Specg(M). We topologize Cl.Specg(M) with the quasi-Zariski topology, which is analogous to that for Specg(R). 2010 MSC: 13A02, 16W50. Keywords: Graded prime ideal, Zariski topology, quasi-Zariski topology. 1. Introduction Recently many authors have been interested in equip algebraic structures with Zariski topology (cf. [4, 11, 12]). A grading on a ring and its mod- ules usually aids computations by allowing one to focus on the homogeneous elements, which are presumably simpler or more controllable than random el- ements. However, for this to work one needs to know that the constructions being studied are graded. One approach to this issue is to redefine the con- structions entirely in terms of the category of graded modules and thus avoid any consideration of non-graded modules or non-homogeneous elements; Sharp gives such a treatment of attached primes in [15]. Unfortunately, while such an approach helps to understand the graded modules themselves, it will only help Received August 2012 – Accepted March 2013 http://dx.doi.org/10.4995/agt.2013.1586 A. Yousefian Darani and S. Motmaen to understand the original construction if the graded version of the concept happens to coincide with the original one. Therefore, notably, the study of graded modules is very important. Our main purpose is to study some new classes of graded submodules of graded modules and endow these classes of submodules with quasi-Zariski topology. Zariski topology on the prime spectrum of a module over a com- mutative ring have been already studied in [11, 12]. Moreover some topologies on the spectrum of graded prime submodules of a graded module have been studied in [16]. Therefore these results will be used in order to obtain the main aims of this paper. The organization of this paper is as follows: In section 2 we recollect the results concerning the topologies on the prime spectrum of a module over a commutative ring. Moreover we remind the notation and the elemental prop- erties about graded modules and rings that we will use in this paper. In section 3 we introduce and study the concept of graded classical prime submodules and define the quasi-Zariski topology on the spectrum of all graded classical prime submodules of a graded module. 2. preliminaries In this section, we recall some definitions and notations used throughout. Throughout this paper all rings are commutative with a nonzero identity and all modules are considered to be unitary. Prime submodules play an important role in the module theory over commutative rings. Let M be a module over a commutative ring R. A prime (resp. primary) submodule N of M is a proper submodule N of M with the property that for a ∈ R and m ∈ M, am ∈ N implies that m ∈ N or a ∈ (N :R M) (resp. a k ∈ (N :R M) for some positive integer k). In this case P = (N :R M) (resp. P = √ (N :R M)) is a prime ideal of R and we say that N is a P-prime (resp. P-primary) submodule of M. There are several ways to generalize the notion of prime submodules. We could restrict where am lies or we can restrict where a and/or b lie. We begin by mentioning some examples obtained by restricting where ab lies. Weakly prime submodules were introduced by Ebrahimi Atani and Farzalipour in [8]. A proper submodule N of M is weakly prime if for a ∈ R and m ∈ M with 0 6= am ∈ N, either m ∈ N or a ∈ (N :R M). Behboodi and Koohi in [3] defined another class of submodules and called it classical prime. A proper submodule N of M is said to be classical prime when for a, b ∈ R and m ∈ M, abm ∈ N implies that am ∈ N or bm ∈ N. Recently, M. Baziar and M. Behboodi [2] defined a classical primary submod- ule in the R-module M as a proper submodule Q of M such that if abm ∈ Q, where a, b ∈ R and m ∈ M, then either bm ∈ Q or akm ∈ Q for some positive integer k. Clearly, in case M = R, classical primary submodules coincide with primary ideals. Let G be an arbitrary group. A commutative ring R with a non-zero identity is G-graded if it has a direct sum decomposition R = ⊕ g∈G Rg such that for c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 160 Zariski topology on the spectrum of graded classical prime submodules all g, h ∈ G, RgRh ⊆ Rgh. The G-graded ring R is called a graded integral domain provided that ab = 0 implies that either a = 0 or b = 0 where a, b ∈ h(R) := ⋃ g∈G Rg. If R is G-graded, then an R-module M is said to be G- graded if it has a direct sum decomposition M = ⊕ g∈G Mg such that for all g, h ∈ G, RgMh ⊆ Mgh . For every g ∈ G, an element of Rg or Mg is said to be a homogeneous element. We denote by h(M) the set of all homogeneous elements of M, that is h(M) = ⋃ g∈G Mg. Let M be a G-graded R-module. A submodule N of M is called graded (or homogeneous) if N = ⊕ g∈G (N ∩ Mg) or equivalently N is generated by homogeneous elements. Moreover, M/N becomes a G-graded R-module with g-component (M/N)g = (Mg + N)/N for each g ∈ G. An ideal I of R is called a graded ideal if it is a graded submodule of R and a graded R-module. Let R be a G-graded ring. A proper graded ideal I of R is said to be a graded prime ideal if whenever ab ∈ I, we have a ∈ I or b ∈ I, where a, b ∈ h(R). The graded radical of I , denoted by Gr(I), is the set of all x ∈ R such that for each g ∈ G there exists ng > 0 with x ng ∈ I. A graded R-module M is called graded finitely generated if M = ∑n i=1 Rxgi, where xgi ∈ h(M) for every 1 ≤ i ≤ n. It is clear that a graded module is finitely generated if and only if it is graded finitely generated. For M, consider the subset T g(M) = {m ∈ M : rm = 0 for some nonzero r ∈ h(R)}. If R is a graded integral domain, then T g(M) is a graded submodule of M. M is called graded torsion-free (g-torsion-free for short) if T g(M) = 0, and it is called graded torsion (g-torsion for short) if T g(M) = M. It is clear that if M is torsion-free, then it is g-torsion-free. Moreover, if M is g-torsion, then it is torsion. Let R be a G-graded ring and M a graded R-module. We recall from [8] that a proper graded submodule N of M is called graded prime (resp. graded primary) if rm ∈ N, then m ∈ N or r ∈ (N :R M) = {r ∈ R|rM ⊆ N} (resp. rk ∈ (N :R M) for some positive integer k), where r ∈ h(R), m ∈ h(M). It is shown in [8, Proposition 2.7] that if N is a graded prime submodule of M, then P := (N :R M) is a graded prime ideal of R, and N is called graded P-prime submodule. Let N be a graded submodule of M. Then N is a graded prime submodule of M if and only if P := (N :R M) is a graded prime ideal of R and M/N is a g-torsion-free R/P-module. Note that some graded R-modules M have no graded prime submodules. We call such graded modules g-primeless. A submodule S of M will be called graded semiprime if S is an intersection of graded prime submodules of M. Let Specg(M) denote the set of all graded prime submodules of M. Let N be a graded submodule of M. The graded radical of N in M, denoted by GrM (N) is defined to be the intersection of all graded prime submodules of M containing N [10]. Hence GrM (N) is a graded semiprime submodule. A proper graded submodule N of M is called graded weakly prime if 0 6= rm ∈ N, then m ∈ N or r ∈ (N :R M). Hence every graded prime submodule is graded weakly prime. c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 161 A. Yousefian Darani and S. Motmaen From now on, R is a G-graded ring and M is a graded R-module unless otherwise stated. 3. Graded classical prime submodules A proper graded submodule N of M is called graded classical prime if for every a, b ∈ h(R), m ∈ h(M), whenever abm ∈ N, then either am ∈ N or bm ∈ N. Let N be a graded classical prime submodule of M. Then, it is easy to see that Ng is a classical prime submodule of the Re-module Mg for every g ∈ G. It is evident that every graded prime submodule is graded classical prime. However the next example shows that a graded classical prime submodule is not necessarily graded prime. Example 3.1. Assume that R is a graded integral domain and P is a non-zero graded prime ideal of R. In this case the ideal Q := P ⊕ 0 is a graded classical prime submodule of the graded R-module R ⊕ R while it is not graded prime. This example shows also that a graded classical prime submodule need not be classical prime. We denote by Cl.Specg(M), the set of all graded classical prime submodules of M. Obviously, some graded R-modules M have no graded classical prime submodules; such modules are called g-Cl.primeless. For example, the zero module is clearly g-Cl.primeless. A submodule S of M will be called graded classical semiprime if S is an intersection of graded classical prime submodules of M. Let N be a graded submodule of M. The graded classical radical of N in M, denoted by Cl.GrM (N), is defined to be the intersection of M and all graded classical prime submodules of M containing N. So if Cl.Specg(M) = ∅, then GrclM (N) = M, and if Cl.Specg(M) 6= ∅, then Gr cl M (N) is a graded classical semiprime submodule. If N = 0, then GrclM (0) is called the graded classical nil-radical of M. We know that Specg(M) ⊆ Cl.Specg(M). As it is mentioned in example 3.1, it happens sometimes that this containment is strict. We call M a graded compatible R-module if its graded classical prime submodules and graded prime submodules coincide, that is if Specg(M) = Cl.Specg(M). If R is a G-graded ring, then every graded classical prime ideal of R is a graded prime ideal. So, if we consider R as a graded R-module, it is graded compatible. The following lemma is obvious. Lemma 3.2. Let N be a proper graded submodule of M. Then N is a graded classical prime submodule if and only if for each x ∈ h(M) \ N, (N :R x) is a graded prime ideal of R. c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 162 Zariski topology on the spectrum of graded classical prime submodules Proposition 3.3. (1) Let N be proper graded submodule of M. Then N is a graded prime submodule of M if and only if N is graded primary and graded classical prime. (2) Assume that N and K are graded submodule of M with K ⊆ N. Then N is a graded classical prime submodule of M if and only if N/K is a graded classical prime submodule of the graded R-module M/K. Proof. (1) If N is a graded prime submodule of M, then it clearly is both graded primary and graded classical prime. Now assume that N is a graded primary and graded classical prime submodule of M. Let am ∈ N but m /∈ N, where a ∈ h(R) and m ∈ h(M). Since N is graded primary, there exists a positive integer k such that ak ∈ (N :R M). Therefore, for every y ∈ h(M) \ N, ak ∈ (N :R y) and (N :R y) is a prime ideal of R by Lemma 3.2. Hence a ∈ (N :R y). It follows that a ∈ (N :R M), i.e. N is graded prime in M. (2) Straightforward. � Let R be a G-graded R-module and consider Specg(R), the spectrum of all graded prime ideals of R. The Zariski topology on Specg(R) is defined in a similar way to that of Spec(R). For each graded ideal I of R, the graded variety of I is the set V g R (I) = {P ∈ Specg(R)|I ⊆ P}. Then the set {V g R (I)|I is a graded ideal of R} satisfies the axioms for the closed sets of a topology on Specg(R), called the Zariski topology on Specg(R) (see [14]). In [16], Specg(M) has endowed with quasi-Zariski topology. For each graded submodule N of M, let V g ∗ (N) = {P ∈ Specg(M)|N ⊆ P}. In this case, the set ζ g ∗ (M) = {V g ∗ (N)|N is a graded submodule of M} contains the empty set and Specg(M), and it is closed under arbitrary intersections, but it is not necessarily closed under finite unions. The graded R-module M is said to be a g-Top module if ζ g ∗ (M) is closed under finite unions. In this case ζ g ∗ (M) satisfies the axioms for the closed sets of a unique topology τ g ∗ on Specg(M). The topology τ g ∗ (M) on Specg(M) is called the quasi-Zariski topology. In the remainder of this section we use a similar method to define a topology on Cl.Specg(M). To this end, For each graded submodule N of M, set Vg∗(N) = {P ∈ Cl.Specg(M)|N ⊆ P}. Proposition 3.4. Let M be a graded R-module. Then (1) For each subset E ⊆ h(M), Vg∗(E) = V g ∗(N) = V g ∗ (Gr cl M (N)), where N is the graded submodule of M generated by E. (2) Vg∗(0) = Cl.Specg(M), and V g ∗(M) = ∅. (3) If {Nλ}λ∈Λ is a family of graded submodules of M, then ⋂ λ∈Λ V g ∗(Nλ) = Vg∗( ∑ λ∈Λ Nλ). (4) For every pair N and K of graded submodules of M, we have Vg∗(N) ∪ Vg∗(K) ⊆ V g ∗(N ∩ K). Proof. The proof of (2) − (4) is easy. So we just provide a proof for part (1). Assume that N is the graded submodule of M generated by E ⊆ h(M). Then from E ⊆ N ⊆ GrclM (N) we have c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 163 A. Yousefian Darani and S. Motmaen Vg∗(Gr cl M (N)) ⊆ V g ∗(N) ⊆ V g ∗(E). On the other hand, N is the smallest graded submodule of M containing E, so that if P ∈ Vg∗(E), then P ∈ V g ∗(N). Therefore V g ∗(E) = V g ∗(N). More- over GrclM (N) is the intersection of all graded classical prime submodules of M containing N; so Vg∗(N) = V g ∗(Gr cl M (N)). Therefore V g ∗(E) = V g ∗(N) = Vg∗(Gr cl M (N)). � Now if we set η g ∗(M) = {V g ∗(N)|N is a graded submodule of M} then η g ∗(M) contains the empty set and Cl.Specg(M). Moreover η g ∗(M) is closed under arbitrary intersections, but it is not necessarily closed under finite unions. Definition 3.5. Let M be a graded R-module. (1) We shall say that M is a g-Cl.Top module if η g ∗(M) is closed under finite unions, i.e. for any graded submodules N and L of M there exists a graded submodule K of M such that Vg∗(N) ∪ V g ∗(L) = V g ∗(K). (2) A graded classical prime submodule N of M will be called graded clas- sical extraordinary, or g-Cl.extraordinary for short, if whenever K and L are graded classical semiprime submodules of M with K ∩ L ⊆ N then K ⊆ N or L ⊆ N. Note that if M is a g-Cl.Top module, then η g ∗(M) satisfies the axioms for the closed sets of a unique topology ̺ g ∗ on Cl.Specg(M). In this case, the topology ̺ g ∗(M) on Cl.Specg(M) is called the quasi-Zariski topology. Note that we are not excluding the trivial case where Cl.Specg(M) is empty; that is every g- Cl.primeless modules is a g-Cl.Top module. The next result is a useful tool for characterizing g-Cl.Top modules. Theorem 3.6. Let M be a graded R-module. Then, the following statements are equivalent: (i) M is a g-Cl.Top module. (ii) Every graded classical prime submodule of M is g-Cl.extraordinary. (iii) Vg∗(N) ∪ V g ∗(L) = V g ∗(N ∩ L) for any graded classical semiprime sub- modules N and L of M. Proof. The result is clear when Cl.Specg(M) = ∅. So assume that Cl.Specg(M) 6= ∅. (i) ⇒ (ii) Let M be a g-Cl.Top module. Assume that P is a graded classical prime submodule of M and that N, L are graded classical semiprime submod- ules of M with N ∩L ⊆ P . By assumption, there exists a graded submodule K of M with Vg∗(N) ∪ V g ∗(L) = V g ∗(K). Since N is a graded classical semiprime submodule, N = ⋂ i∈I Pi in which {Pi}i∈I is a collection of graded classical prime submodules of M. For every i ∈ I, we have Pi ∈ V g ∗(N) ⊆ V g ∗(K) ⇒ K ⊆ Pi ⇒ K ⊆ ⋂ i∈I Pi = N c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 164 Zariski topology on the spectrum of graded classical prime submodules Similarly, K ⊆ L. So K ⊆ N ∩ L. Now we have Vg∗(N) ∪ V g ∗(L) ⊆ V g ∗(N ∩ L) ⊆ V g ∗(K) = V g ∗(N) ∪ V g ∗(L). Consequently, Vg∗(N) ∪ V g ∗(L) = V g ∗(N ∩ L). Now from N ∩ L ⊆ P we have P ∈ Vg∗(N ∩L) = V g ∗(N)∪V g ∗(L). Hence either P ∈ V g ∗(N) or P ∈ V g ∗(L), that is either N ⊆ P or L ⊆ P . So P is g-Cl.extraordinary. (ii) ⇒ (iii) Suppose that every graded classical prime submodule of M is g-Cl.extraordinary. Assume that N and L are two graded classical semiprime submodules of M. Clearly Vg∗(N) ∪ V g ∗(L) ⊆ V g ∗(N ∩ L). Now assume that P ∈ Vg∗(N∩L). Then N∩L ⊆ P . Since P is g-Cl.extraordinary, we have N ⊆ P or L ⊆ P , that is either P ∈ Vg∗(N) or P ∈ V g ∗(L). Therefore V g ∗(N ∩ L) ⊆ Vg∗(N) ∪ V g ∗(L), and so V g ∗(N) ∪ V g ∗(L) = V g ∗(N ∩ L). (iii) ⇒ (i) Let N, L be two graded submodules of M. We can assume that Vg∗(N) and V g ∗(L) are both nonempty, for otherwise V g ∗(N) ∪ V g ∗(L) = V g ∗(N) or Vg∗(N) ∪ V g ∗(L) = V g ∗(L). We know that Gr cl M (N) and Gr cl M (L) are both graded classical semiprime submodules of M. Setting K = GrclM (N) ∩ Gr cl M (L) we have: Vg∗(N) ∪ V g ∗(L) = V g ∗(Gr cl M (N)) ∪ V g ∗(Gr cl M (L)) = V g ∗(Gr cl M (N) ∩ Gr cl M (L)) = Vg∗(K) by (iii). Hence M is a g-Cl.Top module. � Corollary 3.7. Every g-Cl.Top module is a g-Top module. Proof. Assume that M is a g-Cl.Top module. Let P be a graded prime sub- module of M. Since every graded prime L-submodule is a graded classical prime submodule, P is g-Cl.extraordinary by Proposition 3.6. Hence it is g- extraordinary. Now the result follows from [16, Theorem 2.3]. � Theorem 3.8. Let M be a g-Cl.Top R-module. Then, (1) For every graded submodule K of M, the R-module M/K is a g-Cl.Top module. (2) The graded RP -module MP is a g-Cl.Top module for every graded prime ideal P of R. (3) If GrclM (N) = N for every graded submodule N of M, then M is a graded distributive module. Proof. There will be nothing to prove if M has no graded classical prime sub- modules. So assume that Cl.Specg(M) 6= ∅. (1) By Proposition 3.3, the graded classical prime submodules of M/K are just the submodules N/K where N is a graded classical prime submodule of M with K ⊆ N. Consequently, any graded classical semiprime submodule of M/K is of the form S/K in which S is a graded classical semiprime submodule of M with K ⊆ S. Assume that S1/K and S2/K are two graded classical semiprime submodules of M/K. Then, by Theorem 3.6, Vg∗(S1) ∪ V g ∗(S2) = V g ∗(S1 ∩ S2) since M is a g − Cl.T op module. Thus Vg∗(S1/K) ∪ V g ∗(S2/K) = V g ∗(S1/K ∩ S2/K). It follows from Theorem 3.6 that M/K is a g − Cl.T op module. c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 165 A. Yousefian Darani and S. Motmaen (2) By Theorem 3.6, it is enough to show that every graded classical prime submodule of MP is g-Cl.extraordinary. Let N be a graded classical prime submodule of MP , and let S1 ∩ S2 ⊆ N for some graded classical semiprime submodules S1, S2 of MP . Clearly, N ∩M is a proper graded submodule of M. Assume that a, b ∈ h(R) and m ∈ h(M) are such that abm ∈ N ∩ M. Then, a/1, b/1 ∈ h(RP ) and m/1 ∈ h(MP ) with (a/1)(b/1)(m/1) = (abm)/1 ∈ N. It follows that either (a/1)(m/1) ∈ N or (b/1)(m/1) ∈ N since N is graded classical prime. Therefore, either am ∈ N ∩ M or bm ∈ N ∩ M. This implies that N ∩ M is a graded classical prime submodule of M. Hence N is g- Cl.extraordinary by Theorem 3.6. As another consequence, S1 ∩M and S2 ∩M are graded classical semiprime submodules of M with (S1 ∩ M) ∩ (S2 ∩ M) ⊆ N ∩M. Therefore, S1 ∩M ⊆ N ∩M or S2 ∩M ⊆ N ∩M. It follows that either S1 = (S1 ∩M)RP ⊆ (N ∩M)RP = N or S2 = (S2 ∩M)RP ⊆ (N ∩M)RP = N. Hence N is a g-Cl.extraordinary submodule of MP . (3) For every graded submodules N, K and L of M we have: (K + L) ∩ N = GrclM ((K + L) ∩ N) = ⋂ {P |P ∈ Vg∗((K + L) ∩ N)} = ⋂ {P |P ∈ Vg∗(K + L) ∪ V g ∗(N)} = ⋂ {P |P ∈ (Vg∗(K) ∩ V g ∗(L)) ∪ V g ∗(N)} = ⋂ {P |P ∈ (Vg∗(K)∪V g ∗(N))∩(V g ∗(L)∪V g ∗(N))} = ⋂ {P |P ∈ (Vg∗(K ∩ N)) ∩ (V g ∗(L ∩ N))} = ⋂ {P |P ∈ Vg∗((K ∩ N) + (L ∩ N))} = GrclM ((K ∩N)+(L∩N)) = (K ∩N)+(L∩N). Thus M is graded distributive. � Let M be a g-Cl.Top module and let X = Cl.Specg(M). We know that any closed subset of X is of the form Vg∗(N) for some graded submodule N of M. But now the question arises as to what open subsets of X look like. To say that any open subset of X is of the form X − Vg∗(N) for some graded prime submodule N of M, though true, is not very helpful. For every subset S of h(M), define XS = X − V g ∗(S) In particular, if S = {f}, we denote XS be Xf . Proposition 3.9. The set {Xf|f ∈ h(M)} is a basis for the quasi-Zariski topology on X. Proof. Let U be a non-void open subset in X. Then U = X − Vg∗(N) for some graded submodule N of M. Assume that N is generated by some subset E ⊆ h(M). Then we have U = X − Vg∗(N) = X − V g ∗(E) = X − V g ∗( ⋃ f∈E {f}) = X − ⋂ f∈E Vg∗(f) = ⋃ f∈E (X − Vg∗(f)) = ⋃ f∈E Xf Therefore the set {Xf |f ∈ h(M)} is a basis for X. � c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 166 Zariski topology on the spectrum of graded classical prime submodules A topological space X is said to be irreducible if X 6= ∅ and if every pair of non-void open sets in X intersect. Let X be a topological space. A subset A ⊆ X is said to be dense in X if and only if A ∩ G 6= ∅ for every non-void open subset G ⊆ X. Therefore X is irreducible if and only if every non-void open subset of X is dense. Lemma 3.10. Let M be a graded R-module. Then, N := GrclM (0) is a graded classical prime submodule of M if and only if Cl.Specg(M) is irreducible. Proof. Set X = Cl.Specg(M). Assume first that N is a graded classical prime submodule of M. Let U, V ⊆ X be non-void open subsets. Pick P ∈ U. Now, U = X \ Vg∗(E) for some E ⊆ h(M). Then P ∈ U implies that E * P . Moreover, from N ⊆ P we have E * N, so that N ∈ U. Similarly, N ∈ V . Hence N ∈ U ∩V , and thus U ∩V 6= ∅. Therefore, X is irreducible. Conversely, assume that N is not a graded classical prime submodule of M. So there exist a, b ∈ h(R) and m ∈ h(M) such that am, bm /∈ N, but abm ∈ N. Both Xam and Xbm are open in X. Also am /∈ N ⇒ V g ∗(am) 6= X so Xam 6= ∅. Similarly, Xbm 6= ∅. Now we have, Xam ∩ Xbm = Xabm = X − V g ∗ (abm) ⊆ X − Vg ∗ (N) = ∅ Therefore, X is not irreducible. � 4. Homomorphisms and graded classical prime spectrum of modules In our discussion so far we have concerned ourselves with the graded classical prime spectrum of but one graded module at any given time. A natural question to ask is what relationships on their respective graded classical prime spectra are induced by a homomorphism between two rings. In this section, we address this question. Let M and M′ be two graded R-modules and let φ : M → M′ be a graded R-homomorphism. The inverse image of a graded classical prime submodule of M′ is a graded classical prime submodule of M. For every graded submodule of M′, we write φ−1(N′) = N′ c , the contraction of N′ to M. Also, if N is a graded submodule of M, then we write Rφ(N) = Ne, the graded submodule of M′ generated by φ(N), the extension of N to M′. Let X = Cl.Specg(M) and Y = Cl.Specg(M ′). Thus if P ∈ Y , then P c ∈ X. So we see that φ induces a map φ∗ : Y → X defined by φ∗(P) = P c, for all P ∈ Y . Before continuing, we introduce a more explicit notation: P ∈ Vg ∗M (E) means that E ⊆ h(M) and E ⊆ P ∈ X and Q ∈ Vg ∗M′ (F) means that F ⊆ h(M′) and F ⊆ P ∈ Y c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 167 A. Yousefian Darani and S. Motmaen Proposition 4.1. Let M and M′ be two graded R-modules and let φ : M → M′ be a graded R-homomorphism. Let X = Cl.Specg(M), Y = Cl.Specg(M ′), and let φ∗ : Y → X be the induced map. (1) φ∗ is continuous. (2) If N is a graded submodule of M, then φ∗ −1 (Vg ∗M (N)) = Vg ∗M′ (Ne). (3) If φ is an epimorphism, then φ∗ is a homeomorphism from Y onto the closed subset Vg ∗M (Ker(φ)) of X. Proof. (1) It is enough to show that if U is open in X, then φ∗ −1 (U) is open if Y . For every subset E ⊆ h(M) and Q ∈ Y , we have φ(E) ⊆ Q ⇔ E ⊆ φ−1(Q) ⇔ φ∗(Q) ∈ Vg ∗M (E) ⇔ Q ∈ φ∗ −1 (Vg ∗M (E)) Hence, if f ∈ h(M) and Q ∈ Y , then Q ∈ Yφ(f) ⇔ φ(f) /∈ Q ⇔ Q /∈ φ∗ −1 (Vg ∗M (f)) ⇔ Q ∈ φ∗ −1 (X) − φ∗ −1 (Vg ∗M (f)) ⇔ Q ∈ φ∗ −1 (X − Vg ∗M (f)) = φ∗ −1 (Xf ) Therefore, φ∗ −1 (Xf ) = Yφ(f). In particular, if U is open in X, then φ ∗ −1 (U) is open if Y . Hence φ∗ is continuous. (2) Assume that Q ∈ Y . Then, Q ∈ φ∗ −1 (Vg ∗M (N)) ⇔ φ(N) ⊆ Q ⇔ Q ∈ Vg ∗M′ (φ(N)) ⇔ Q ∈ Vg ∗M′ (Ne) Therefore, φ∗ −1 (Vg ∗M (N)) = Vg ∗M′ (Ne). (3) Suppose that φ is an epimorphism. Then, there exists a one-to-one cor- respondence between graded submodules of M′ and graded submodules of M containing Ker(φ). Under this correspondence, graded classical prime submod- ules of M′ correspond to the graded classical prime submodules of M containing Ker(φ). Therefore, φ∗ : Y → Vg ∗M (Ker(φ)) is bijective. As φ∗ is continuous by (1), it suffices to prove that φ∗ is an open map. Assume that U is an open subset of Y . Then, without loss of generality, we may assume that U = Yf for some f ∈ h(M′). In this case, φ∗(U) = φ∗(Yf ) = {φ ∗(Q)|Q ∈ Y and f /∈ Q} = {P ∈ Vg ∗M (Ker(φ))|φ−1(f) /∈ P} = Xφ−1(f) ∩ V g ∗M (Ker(φ)) This implies that φ∗(U) is an open subset of Vg ∗M (Ker(φ)), that is φ∗ is an open map. Consequently, φ∗ : Y → Vg ∗M (Ker(φ)) is a homeomorphism. � Corollary 4.2. Let M be a graded R-module, and let N be the graded clas- sical nil-radical of M. Then Cl.Specg(M) and Cl.Specg(M/N) are naturally homeomorphic. c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 168 Zariski topology on the spectrum of graded classical prime submodules Proof. By Proposition 4.1, the canonical graded R-epimorphism f : M → M/N induces the homeomorphism f∗ : Cl.Specg(M/N) → V g ∗M (Ker(f)). Now the result follows from Vg ∗M (Ker(f)) = Vg ∗M (N) = Cl.Specg(M). � References [1] M. F. Atiyah and I. G. Macdonald, Introduction to commutative algebra, Longman Higher Education, New York 1969. [2] M. Baziar and M. Behboodi, Classical primary submodules and decomposition theory of modules, J. Algebra Appl. 8, no. 3 (2009), 351–362. [3] M. Behboodi and H. Koohi, Weakly prime modules, Vietnam J. Math. 32, no. 2 (2004), 185–195. [4] M. Behboodi and M. J. Noori, Zariski-Like topology on the classical prime spectrum of a module, Bull. Iranian Math. Soc. 35, no. 1 (2009), 255–271. [5] M. Behboodi and S. H. Shojaee, On chains of classical prime submodules and dimension theory of modules, Bulletin of the Iranian Mathematical Society 36 (2010), 149–166. [6] J. Dauns, Prime modules, J. Reine Angew. Math. 298 (1978), 156–181. [7] S. Ebrahimi Atani, On graded prime submodules, Chiang Mai J. Sci. 33, no. 1 (2006), 3–7. [8] S. Ebrahimi Atani and F. Farzalipour, On weakly prime submodules, Tamkang Journal of Mathematics 38, no. 3 (2007), 247–252. [9] S. Ebrahimi Atani and F. Farzalipour, On graded multiplication modules, Chiang-Mai Journal of Science, to appear. [10] S. Ebrahimi Atani and F. E. K. Saraei, Graded modules which satisfy the Gr-Radical formola, Thai Journal of Mathematics 8, no. 1 (2010), 161–170. [11] C. P. Lu, The Zariski topology on the prime spectrum of a module, Houston J. Math. 25, no. 3 (1999), 417–425. [12] R. L. McCasland, M. E. Moore and P. F. Smith, On the spectrum of a module over a commutative ring, Comm. Algebra 25, no. 1 (1997), 79–103. [13] K. H. Oral, U. Tekir and A. G. Agargun, On graded prime and primary submodules, Turk. J. Math. 25, no. 3 (1999), 417–425. [14] P. C. Roberts, Multiplicities and Chern classes in local algebra, Cambridge University Press, 1998. [15] R. Y. Sharp, Asymptotic behaviour of certain sets of attached prime ideals, J. London Math. Soc. 34, no. 2 (1986), 212–218. [16] A. Yousefia Darani, Topologies on Specg(M), Buletinul Academiei de Stiinte a Republicii Moldova Matematica, to appear. c© AGT, UPV, 2013 Appl. Gen. Topol. 14, no. 2 169