1 Introduction

There are many papers on assigning a graph to a ring \(R\), see [14, 9, 10]. In this paper, we introduce the M-principal graph of \(R\), denoted by \(M-PG(R)\), where \(M\) is an \(R\)-module. Throughout the paper all rings are commutative with non-zero identity and all modules are non-zero unitary.

Let \(R\) be a ring and \(M\) be an \(R\)-module. The annihilator of \(M\) is denoted by \(ann(M)\). The module \(M\) is called a faithful \(R\)-module if \(ann(M)=0\). Also, \(M\) is called an simple \(R\)-module if \(M\ne 0\), and \(M\) has no submodules other than \(0\) and \(M\). We denote the characteristic of \(R\) by char \(R\). Also, \(J(R)\) denotes the Jacobson radical of \(R\) and \(U(R)\) denotes the group of units of \(R\). A ring having just one maximal ideal is called a local ring and a ring having only finitely many maximal ideals is said to be a semilocal ring. The direct product of a family of rings \(\{R_i \mid i\in I \}\) is denoted by \(\prod _{i\in I}{R_i}\). As usual, \(\mathbb {Z}\), \(\mathbb {R}, \mathbb {C}\), and \(\mathbb {Z}_n\) will denote the integers, real numbers, complex numbers and integers modulo \(n\), respectively.

A graph in which each pair of distinct vertices is joined by an edge is called a complete graph. The complete graph on \(n\) vertices is denoted by \(K_n\). A graph \(G\) is called regular if each vertex has the same number of neighbors. An empty graph is one whose edge set is empty. Let \(G\) be a graph. The set of vertices and the set of edges of \(G\) are denoted by \(V(G)\) and \(E(G)\), respectively. “A subgraph \(H\) of \(G\) is said to be an induced subgraph of \(G\) if it is a subgraph i.e. \(V(H)\subset V(G)\) and \(E(H)\subset E(G)\) and it has exactly the same edges that appear in \(G\) over the vertices \(V(H)\) i.e. \(\forall u,v\in V(H)\) an edge \(e=uv\in E(H)\) if and only if \(e\in E(G)\).”

Also a subgraph \(H\) of \(G\) is called a spanning subgraph if \(V(H)=V(G)\). We say that \(G\) is connected if there is a path between any two distinct vertices of \(G\). A cycle of \(G\) is a path such that the start and end vertices are the same. The girth of \(G\), denoted by \(gr(G)\), is the length of a shortest cycle in \(G\) (\(gr(G)=\infty \) if \(G\) contains no cycles). A Hamiltonian cycle is a spanning cycle in a graph. A graph \(G\) is called Hamiltonian if \(G\) has a Hamiltonian cycle. A forest is a graph with no cycles. A clique in \(G\) is a set of pairwise adjacent vertices and a set in \(G\) whose no two vertices are adjacent is called an independent set. The clique number and the independence number of \(G\), denoted by \(\omega (G)\) and \(\alpha (G)\), are the largest orders of a clique and an independent set of \(G\), respectively. Also, the chromatic number of \(G\), denoted by \(\chi (G)\), is the smallest number of colors which can be assigned to the vertices of \(G\) in such a way that every two adjacent vertices have different colors.

In this article, we introduce and investigate the \(M\)-principal graph of \(R\), denoted by \(M-PG(R)\), where \(R\) is a commutative ring and \(M\) is a non-zero \(R\)-module. If \(R\) is regarded as a module over itself, that is, \(M=R\), then the \(M\)-principal graph of \(R\) is denoted by \(PG(R)\). Also, we study some properties of \(PG(R)\), in particular we consider the graph \(PG(\mathbb {Z}_n)\) for each positive integer \(n>1\). Finally, some relations between \(PG(R)\) and \(M-PG(R)\) are established.

2 The \(M\)-principal graph of \(R\)

In this section, we introduce the \(M\)-principal graph of \(R\) and study its basic properties.

Definition 2.1

Let \(R\) be a ring and \(M\) be a non-zero \(R\)-module. The \(M\) -principal graph of \(R\), denoted by \(M-PG(R)\), is the (undirected) graph whose vertex set is \(R\backslash \{0\}\) and two distinct vertices \(x\) and \(y\) are adjacent if and only if \(xM=yM\).

It is clear that if \(M\) and \(N\) are isomorphic modules over \(R\), then \(M-PG(R)\) is the same as \(N-PG(R)\), but the converse is not true in general. For instance, consider \(\mathbb {R}-PG(\mathbb {Z})\) and \(\mathbb {C}-PG(\mathbb {Z})\).

Example 2.2

Let \(R=\mathbb {Z}_6\). So we have the following graphs.

figure a

Remark 2.3

Clearly, \(M-PG(R)\) is a disjoint union of complete graphs. Hence \(gr(M-PG(R))\in \{3,\infty \}\) and also \(|U(R)|\le \omega (M-PG(R))=\chi (M-PG(R))\).

By the previous remark, we have the next immediate result.

Corollary 2.4

Let \(R\) be a ring with \(|R|>3\), and let \(M\) be a non-zero \(R\)-module. Then the following conditions are equivalent:

  1. (1)

    \(M-PG(R)\) is a complete graph.

  2. (2)

    \(M-PG(R)\) is a Hamiltonian graph.

  3. (3)

    \(M-PG(R)\) is a connected graph.

Remark 2.5

Let \(x\) be a non-zero element of \(R\). Then each coset \(x+ann(M)\) is a clique in \(M-PG(R)\) and so \(|ann(M)|\le \omega (M-PG(R))\). Also, \(\alpha (M-PG(R)) \le |R/ ann(M)|\). Moreover, if \(ann(M)\ne 0\), then \(M-PG(R)\) is disconnected.

Theorem 2.6

Let \(R\) be a ring and \(M\) be a non-zero \(R\)-module. Then \(M\) is a faithful simple \(R\)-module if and only if \(M-PG(R)\) is a complete graph and \(M\) is a cyclic \(R\)-module.

Proof

Suppose that \(M-PG(R)\) is a complete graph and there exists a non-zero element \(m\in M\) such that \(M=Rm\). If \(0\ne x\in ann(M)\), then \(x\) is not adjacent to \(1\), a contradiction. So \(ann(M)=0\). Let \(m'\) be a non-zero element of \(M\). We have \(m'=xm\) for some \(x\in R\). Since \(M-PG(R)\) is a complete graph, \(M=xM=xRm=Rm'\). Thus \(M\) is simple. For the other direction, assume that \(x\in R\) and \(x\ne 0\). Since \(xM\) is a non-zero submodule of \(M\) and \(M\) is simple, we have \(xM=M\). The proof is complete. \(\square \)

Theorem 2.7

Let \(R\) be a ring and \(M\) be a (non-zero) finitely generated \(R\)-module. If \(M-PG(R)\) is a complete graph, then \(J(R)=0\).

Proof

By contradiction, assume that \(x\in J(R)\) and \(x\ne 0\). Since \(M-PG(R)\) is a complete graph, \(xM=M\). So \(J(R)M=M\) and by Nakayama’s Lemma [7, Proposition 2.6], we have \(M=0\), a contradiction. \(\square \)

Theorem 2.8

Let \(R\) be a ring and \(M\) be a non-zero \(R\)-module. If \(M-PG(R)\) is empty, then \(M\) is a faithful \(R\)-module and we have \(|U(R)|=1\). Moreover, if \(R\) is an Artinian ring, then \(R\cong {(\mathbb {Z}_2)}^k\) for some positive integer \(k\).

Proof

By Remark 2.5, \(ann(M)=0\) and since \(xM=M\) for every \(x\in U(R)\), we have \(|U(R)|=1\) (so char \(R=2\) and \(J(R)=0\)). If \(R\) is Artinian, then by [7, Theorem 8.7], \(R\cong R_1\times \cdots \times R_k\), where \(R_i\) is a local ring and \(k\) is a positive integer. Suppose that \(m_i\) is a maximal ideal of \(R_i\), for each integer \(i, 1\le i \le k\). Since \(|R_i\backslash m_i|=|U(R_i)|=1\), we have \(|m_i|=1\) and hence \(R_i\cong \mathbb {Z}_2\). Thus \(R\cong {(\mathbb {Z}_2)}^k\). \(\square \)

Theorem 2.9

Let \(R\) be an integral domain and \(M\) be an Artinian module. If \(R\) contains a non-zero element \(x\) such that \(x^k\ne 1\) for each positive integer \(k\), then \(\omega (M-PG(R))=\infty \).

Proof

Since \(M\) is Artinian, there exists a positive integer \(n\), such that \(x^kM=x^nM\) for each positive integer \(k\ge n\). Now, \(\{x^k \ | \ k\ge n\) } is a clique in \(M-PG(R)\) and hence \(\omega (M-PG(R))=\infty \). \(\square \)

Let \(R\) be a commutative ring. An \(R\)-module \(M\) is said to be secondary (see [8, p. 42]), if \(M\ne 0\) and, for each \(x\in R, xM =M\) or \(x\in \sqrt{ann(M)}\). (\(\sqrt{ann(M)}\) denotes the radical of \(ann(M)\).)

Theorem 2.10

Let \(R\) be a ring and \(M\) be a secondary \(R\)-module. Then \(|R\backslash P|\le \omega (M-PG(R))\), where \(P=\sqrt{ann(M)}\).

Proof

Since \(M\) is a secondary \(R\)-module, \(xM=M\) for every \(x\in R\backslash P\). So \(R\backslash P\) is a clique in \(M-PG(R)\) and hence \(|R\backslash P|\le \omega (M-PG(R))\). \(\square \)

Theorem 2.11

Let \(R\) be a ring and \(M\) be a non-zero \(R\)-module. If \(N\) is a proper submodule of \(M\), then \(M-PG(R)\) is a spanning subgraph of \(M/N-PG(R)\).

Proof

Suppose that \(x,y\in R\) and \(xM=yM\). So \(x(M/N)=(xM+N)/N=(yM+N)/N=y(M/N)\). Hence if \(x\) and \(y\) are adjacent vertices of \(M-PG(R)\), then \(x\) is adjacent to \(y\) in \(M/N-PG(R)\). Therefore \(M-PG(R)\) is a spanning subgraph of \(M/N-PG(R)\). \(\square \)

Suppose that \(R\) is a commutative ring and \(M\) is a module over \(R\). It is well known (see [7, p. 19]) that if \(I\) is an ideal of \(R\) such that \(IM=0, M\) can be regarded as an \(R/I\)-module, as follows: if \(\overline{x}\in R/I\) is represented by \(x\in R\), define \(\overline{x}m\) to be \(xm\) for every \(m\in M\). So we can deduce the next theorem.

Theorem 2.12

Let \(R\) be a ring and \(M\) be a non-zero \(R\)-module, and let \(I\) be an ideal of \(R\) such that \(IM=0\). Then \(M-PG(R/I)\) is an induced subgraph of \(M-PG(R)\). Moreover, if \(I=ann(M)\ne 0\) and \(M-PG(R/I)\) is a complete graph, then \(M-PG(R)\) is a disjoint union of two complete graphs.

Proof

Suppose that \(x,y\in R\). Clearly, \(xM=yM\) if and only if \(\overline{x}M=\overline{y}M\). So (by assigning \(\overline{x}\) to \(x\)), \(M-PG(R/I)\) is an induced subgraph of \(M-PG(R)\). Now, assume that \(M-PG(R/ann(M))\) is a complete graph and let \(x,y\in R\backslash \{0\}\). If \(x,y\in ann(M)\), then \(xM=yM=0\). Suppose that \(x,y\in R\backslash ann(M)\). If \(\overline{x}=\overline{y}\), then \(x-y\in ann(M)\) implies that \(xM=yM\). Otherwise, \(\overline{x}\) and \(\overline{y}\) are two distinct vertices of \(M-PG(R/ann(M))\) and hence \(\overline{x}\) is adjacent to \(\overline{y}\). So \(x\) is adjacent to \(y\). It is clear that if \(x\in ann(M)\) and \(y\in R\backslash ann(M)\), then \(xM\ne yM\). Thus \(M-PG(R)\) is a disjoint union of two complete graphs. \(\square \)

Let \(S\) be a multiplicatively closed subset of a commutative ring \(R\), and let \(M\) be an \(R\)-module. We denote the ring of fractions of \(R\) and the module of fractions of \(M\) (with respect to \(S\)) by \(R_S\) and \(M_S\), respectively. Now, we have the following theorem.

Theorem 2.13

Let \(R\) be a ring and \(M\) be a non-zero \(R\)-module. If \(S=Reg(R)\), then \(M-PG(R)\) is a subgraph of \(M_S-PG(R_S)\). Moreover, if \(M-PG(R)\) is complete, then \(M_S-PG(R_S)\) is also a complete graph.

Proof

Suppose that \(x,y\in R\) and \(xM=yM\). So \((x/1)M_S=(xM)_S=(yM)_S=(y/1)M_S\). Thus by assigning \(x\) to \(x/1\), \(M-PG(R)\) is a subgraph of \(M_S-PG(R_S)\). Note that if \(x\ne 0\), then \(x/1\ne 0\). Now, assume that \(M-PG(R)\) is a complete graph and \(x/r,y/t\) are two distinct vertices of \(M_S-PG(R_S)\). Then \(tx\) and \(ry\) are two distinct vertices of \(M-PG(R)\) and so \(txM=ryM\). Thus \((x/r)M_S=(tx/tr)M_S=(txM)_S=(ryM)_S=(ry/tr)M_S=(y/t)M_S\) and hence \(x/r\) is adjacent to \(y/t\). \(\square \)

3 The principal graph of \(R\)

If \(R\) is regarded as a module over itself, that is, \(M=R\), then the \(M\)-principal graph of \(R\) is denoted by \(PG(R)\) (the principal graph of \(R\)). In this section we study some properties of \(PG(R)\), in particular we consider the graph \(PG(\mathbb {Z}_n)\) for each positive integer \(n>1\).

Note that for two rings \(R\) and \(S\) if \(R\cong S\), then \(PG(R)\cong PG(S)\), but the converse is not true. For instance, consider \(PG(\mathbb {Z}_4)\) and \(PG(\mathbb {Z}_2[x]/(x^2))\).

Example 3.1

figure b

Corollary 3.2

Let \(R\) be a ring with \(|R|>3\). Then the following conditions are equivalent:

  1. (1)

    \(R\) is a field.

  2. (2)

    \(PG(R)\) is a complete graph.

  3. (3)

    \(PG(R)\) is a Hamiltonian graph.

  4. (4)

    \(PG(R)\) is a connected graph.

Notice that \(\alpha (PG(R))\) is equal to the number of non-zero principal ideals of \(R\). So we have the next theorem.

Theorem 3.3

Let \(R\) be a ring. If \(\alpha (PG(R))\) is finite, then \(R\) is Artinian.

Proof

\(R\) has only a finite number of distinct principal ideals, since \(\alpha (PG(R))\) is finite. Now, suppose that \(I\) is a non-zero ideal of \(R\). We have \(I=\sum _{x\in I} xR\), so \(R\) contains only finitely many ideals and hence \(R\) is Artinian. \(\square \)

Remark 3.4

Note that there exist some infinite rings which have only a finite number of ideals, for example the ring \(F[x]/(x^2)\), where \(F\) is an infinite field.

Theorem 3.5

Let \(R\) be a semilocal ring. If \(\omega (PG(R))\) is finite, then \(R\) is finite.

Proof

Suppose that \(m_1,\ldots ,m_k\) are all the maximal ideals of \(R\). We have \(U(R)=R\backslash \bigcup ^{k}_{i=1}m_i\). Since \(\omega (PG(R))\) is finite, \(U(R)\) is finite and so by [4, Theorem 2], \(R\) is finite. \(\square \)

Corollary 3.6

Let \(R\) be a semilocal ring. If \(PG(R)\) is empty, then \(R\cong {(\mathbb {Z}_2)}^k\) for some positive integer \(k\).

Proof

By the previous theorem, \(R\) is finite and so by Theorem 2.8, the result holds. \(\square \)

Theorem 3.7

Let \(R\) be an Artinian ring. If \(PG(R)\) is a forest, then \(R\) is isomorphic to one of the rings:

$$\begin{aligned} \mathbb {Z}_3,\ \mathbb {Z}_4,\ \mathbb {Z}_2[x]/{(x^2)},\ {(\mathbb {Z}_2)}^k,\ {(\mathbb {Z}_2)}^k \times \mathbb {Z}_3,\ {(\mathbb {Z}_2)}^k\times \mathbb {Z}_4,\ {(\mathbb {Z}_2)}^k \times \mathbb {Z}_2[x]/{(x^2)}, \end{aligned}$$

for some positive integer \(k\).

Proof

Since \(gr(PG(R))=\infty \), so \(|U(R)|\le 2\) and by [4, Lemma 1], \(R\) is isomorphic to one of the rings:

$$\begin{aligned} \mathbb {Z}_3,\ \mathbb {Z}_4,\ \mathbb {Z}_2[x]/{(x^2)},\ {(\mathbb {Z}_2)}^k,\ {(\mathbb {Z}_2)}^k \times \mathbb {Z}_3,\ {(\mathbb {Z}_2)}^k\times \mathbb {Z}_4,\ {(\mathbb {Z}_2)}^k \times \mathbb {Z}_2[x]/{(x^2)}, \end{aligned}$$

for some positive integer \(k\). \(\square \)

The strong product \(G \boxtimes H\) of graphs \(G\) and \(H\) is a graph such that the vertex set of \(G \boxtimes H\) is the Cartesian product \(V(G) \times V(H)\); and any two distinct vertices \((u_1,u_2)\) and \((v_1,v_2)\) are adjacent in \(G \boxtimes H\) if and only if (\(u_1=v_1\) and \(u_2\) adj \(v_2\)) or (\(u_1\) adj \(v_1\) and \(u_2=v_2\)) or (\(u_1\) adj \(v_1\) and \(u_2\) adj \(v_2\)). Now, suppose that \(R_1\) and \(R_2\) are two rings. Let \(x_1,y_1\in R_1\) and \(x_2,y_2\in R_2\). We know that \((x_1,x_2)(R_1\times R_2)=(y_1,y_2)(R_1\times R_2)\) if and only if \(x_1R_1=y_1R_1\) and \(x_2R_2=y_2R_2\). So \(PG(R_1\times R_2)=PG(R_1)\boxtimes PG(R_2)\) and also \(\omega (PG(R_1\times R_2))=\omega (PG(R_1))\times \omega (PG(R_2))\).

Theorem 3.8

Let \(R_1\) and \(R_2\) be two rings, and let \(\alpha (PG(R_1))=\alpha _1\) and \(\alpha (PG(R_2))=\alpha _2\). Then \(\alpha (PG(R_1\times R_2))=\alpha _1+\alpha _2+\alpha _1\alpha _2\).

Proof

It is obvious that \(PG(R_1)\) is the union of \(\alpha _1\) disjoint complete graphs and similarly, \(PG(R_2)\) is the union of \(\alpha _2\) disjoint complete graphs. Suppose that \(A_1,\ldots ,A_{\alpha _1}\) are all the components of \(PG(R_1)\) and \(B_1,\ldots ,B_{\alpha _2}\) are all the components of \(PG(R_2)\). Then \(A_1\times 0,\ldots ,A_{\alpha _1}\times 0\) and \(0\times B_1,\ldots ,0\times B_{\alpha _2}\) and \(A_i\times B_j\), where \(1\le i\le \alpha _1\) and \(1\le j \le \alpha _2\) are all the components of \(PG(R_1\times R_2)\), each of which is a complete graph. Hence \(\alpha (PG(R_1\times R_2))=\alpha _1+\alpha _2+\alpha _1\alpha _2\). \(\square \)

Let \(R_1,\ldots ,R_n\) be rings. By induction, one can easily prove that \(PG(R_1\times \cdots \times R_n)=PG(R_1)\boxtimes \cdots \boxtimes PG(R_n)\).

Now, we consider the graph \(PG(\mathbb {Z}_n)\) for each positive integer \(n>1\). Let \(d(n)\) denote the number of positive divisors of \(n\). Clearly, \(\alpha (PG(\mathbb {Z}_n))=d(n)-1\). Suppose that \(\varphi \) is the Euler phi function. By [6, Theorem 2.5], if \(d\) is a positive divisor of \(n\), then \(\varphi (d)\le \varphi (n)\). So \(\chi (PG(\mathbb {Z}_n))=\omega (PG(\mathbb {Z}_n))=\varphi (n)\).

The chromatic index of \(G\), denoted by \(\chi '(G)\), is the smallest number of colors which can be assigned to the edges of \(G\) such that no two edges incident on the same vertex have the same color. By Vizing’s Theorem, if \(G\) is a graph whose maximum vertex degree is \(\Delta \), then \(\Delta \le \chi '(G) \le \Delta +1\). Vizing’s Theorem divides the graphs into two classes according to their chromatic index; graphs satisfying \(\chi '(G)=\Delta \) are called class 1, those with \(\chi '(G)=\Delta + 1\) are class 2.

We now state the following result which shows that the graph \(PG(\mathbb {Z}_n)\) is class 1.

Theorem 3.9

The principal graph \(PG(\mathbb {Z}_n)\) is class 1, for each positive integer \(n>1\).

Proof

By [6, Theorem 2.5], \(\varphi (d)\) is even for each positive integer \(d\ge 3\). Thus each connected component of \(PG(\mathbb {Z}_n)\) with two or more vertices is a complete graph which contains an even number of vertices. So by [5, Theorem 5.11], \(PG(\mathbb {Z}_n)\) is class 1. \(\square \)

Remark 3.10

By [6, Theorem 2.5], \(\varphi (d)\ge 2\) for each positive integer \(d\ge 3\). So if \(n\) is even, then \(PG(\mathbb {Z}_n)\) contains exactly one isolated vertex and otherwise it contains no isolated vertex.

Theorem 3.11

If \(G = PG(\mathbb {Z}_n)\), then the following conditions are equivalent:

  1. (1)

    \(G\) is a regular graph.

  2. (2)

    \(n\) is a prime number.

  3. (3)

    \(G\) is a complete graph.

  4. (4)

    \(G\) is a Hamiltonian graph or \(G\cong PG(\mathbb {Z}_2)\) or \(G\cong PG(\mathbb {Z}_3)\).

  5. (5)

    \(G\) is a connected graph.

Proof

Suppose that \(PG(\mathbb {Z}_n)\) is a regular graph. So if \(d\ne 1\) is a positive divisor of \(n\), then \(\varphi (d)=\varphi (n)\). Assume that \(p\) and \(q\) are prime divisors of \(n\). Then \(\varphi (p)=\varphi (q)\) implies that \(p=q\) and thus \(n\) is a prime number (note that \(\varphi (p^2)\ne \varphi (p)\)). By Corollary 3.2, the proof is complete. \(\square \)

Let \(m,n>1\) be positive integers. We know that \(PG(\mathbb {Z}_m\times \mathbb {Z}_n)=PG(\mathbb {Z}_m)\boxtimes PG(\mathbb {Z}_n)\) and so \(\omega (PG(\mathbb {Z}_m\times \mathbb {Z}_n))=\varphi (m)\varphi (n)\). Then \(\omega (PG(\mathbb {Z}_m\times \mathbb {Z}_n))\le \omega (PG(\mathbb {Z}_{mn}))\).

We close this article by considering some relations between \(PG(R)\) and \(M-PG(R)\), where \(R\) is a ring and \(M\) is a non-zero \(R\)-module.

Theorem 3.12

Let \(R\) be a ring and \(M\) be a non-zero \(R\)-module. Then \(PG(R/ann(M))\) is a subgraph of \(M-PG(R)\). Moreover, \(|ann(M)|\omega (PG(R/ann(M)))\le \omega (\)M-PG(R)\()\).

Proof

Let \(\overline{R}=R/ann(M)\) and suppose that \(\overline{x}\) and \(\overline{y}\) are two adjacent vertices of \(PG(\overline{R})\). Now, assume that \(m\in M\). Since \(\overline{x}\overline{R}=\overline{y}\overline{R}\), there exists \(r\in R\) such that \(\overline{x}=\overline{yr}\). Hence \((x-yr)m=0\) which implies that \(xM\subseteq yM\). Similarly, \(yM\subseteq xM\). So \(x\) and \(y\) are two adjacent vertices of \(M-PG(R)\). Thus by assigning \(\overline{x}\) to \(x, PG(\overline{R})\) is a subgraph of \(M-PG(R)\). By Remark 2.5, the last part is clear. \(\square \)

Theorem 3.13

Let \(R\) be a ring and \(M\) be a non-zero \(R\)-module. Then \(PG(R)\) is a spanning subgraph of \(M-PG(R)\). Furthermore, if \(M\) is a faithful cyclic \(R\)-module, then the graph \(PG(R)\) is exactly the same as the graph \(M-PG(R)\).

Proof

Suppose that \(x\) and \(y\) are two adjacent vertices of \(PG(R)\). Clearly, \(xR=yR\) implies that \(xM=yM\). So \(x\) and \(y\) are two adjacent vertices of \(M-PG(R)\). Thus \(PG(R)\) is a spanning subgraph of \(M-PG(R)\). Now, assume that \(ann(M)=0\) and \(M=Rm\) for some \(m\in M\). If \(x\) and \(y\) are two adjacent vertices of \(M-PG(R)\), then \(xM=yM\) and so there exists \(r\in R\) such that \(xm=yrm\). Hence \(x-yr\in ann(M)=0\) which implies that \(xR\subseteq yR\). Similarly, \(yR\subseteq xR\). Therefore \(x\) and \(y\) are two adjacent vertices of \(PG(R)\). So \(PG(R)\) is exactly the same as \(M-PG(R)\). \(\square \)

Theorem 3.14

Let \(R\) be a ring. If \(M=\prod _{i\in I}{R_i}\), where \(R_i\cong R\) for each \(i\in I\), then \(PG(R)\) is exactly the same as \(M-PG(R)\).

Proof

Suppose that \(x,y\in R\backslash \{0\}\). By the previous theorem, \(xR=yR\) implies that \(x\prod _{i\in I}{R_i}=y\prod _{i\in I}{R_i}\). Now, assume that \(x\prod _{i\in I}{R_i}=y\prod _{i\in I}{R_i}\). Let \(e=(e_i)\), where \(e_1=1\) and \(e_i=0\), for each \(i\ne 1\). So there exists \(r=(r_i)\in \prod _{i\in I}{R_i}\) such that \(xe=y(r_i)\) and hence \(x=yr_1\). Thus \(xR\subseteq yR\). Similarly, \(yR\subseteq xR\) and so \(xR=yR\). \(\square \)

Corollary 3.15

Let \(R\) be a ring and \(F\) be a free \(R\)-module. Then \(PG(R)\) is exactly the same as \(F-PG(R)\).

Proof

Since \(F\) is isomorphic to a direct sum of copies of \(R\), the result holds by the previous theorem. \(\square \)