Chapter 4

The Cayley graph of ideals of

commutative ring

In this Chapter, we characterize all commutative Artinian non-local

rings R for which Cayley graph IΓ(R) has crosscap two. The con-
tents of this chapter have been communicated as a paper in “ARS
Combinatoria”.

4.1 Introduction

In ring theory, the structure of a ring R is closely tied to ideal’s be-

havior more than elements, and so it is deserving to define a graph
with vertex set as ideals instead of elements. M. Afkhami et al. [2]
introduced and investigated the Cayley graph of ideals. Also, for a
semigroup H and a subset S of H, the Cayley graph Cay(H, S) of
65
H relative to S is defined as the digraph with vertex set H and edge
set E(H, S) consisting of those ordered pairs (x, y) such that y = sx
for some s ∈ S (see [34]). By the ordered pair (x, y), we mean that
x → y. In this note we introduce a digraph which can be considered
as a generalization of Cayley graphs of semigroups. Further, they
associate a digraph to the ideals of R whose vertex set is the set of
all nontrivial ideals of R and for every two distinct vertices I and J,
there is an arc from I to J, denoted by I → J, whenever there exists
a nontrivial ideal L such that J = IL. We call this graph the ideal
−→ −
→
digraph of R and denote it by IΓ(R). In fact the ideal digraph IΓ(R)
is isomorphic to the Cayley graph Cay(I∗ , I∗ ), where I∗ is the set of all
nontrivial ideals of R. The undirected ideal (simple) graph of R, de-
noted by IΓ(R), has an edge joining I and J whenever either J = IL
or I = JL, for some nontrivial ideal L of R.

4.2 Some properties of IΓ(R)

In this section, we study some properties of IΓ(R).

Lemma 4.2.1. Let (R, m) be a local ring. If dim(m/m2 ) = 1 and for

some positive integer t, mt = (0), then the set of all non-trivial ideals
of R is the set {mi : 1 ≤ i < t}.

Proof. Since dim(m/m2 ) = 1, by Nakayama’s lemma, m = Rx for

66
some x ∈ R. Now, let I be a non-trivial ideal of R. Since mt = (0),
there exists a natural number i ≤ t such that I ⊆ mi and I * mi+1 .
Let a ∈ I\mi+1 . We have a = bxi for some b ∈ R. If b ∈ m, then
a ∈ mi+1 , a contradiction. Thus b is an unit. Hence xi ∈ I. This
implies that I = xi = mi , as desired. Thus, the set of all non-trivial
ideals of R is the set {mi : 1 ≤ i < t}. ✷

The next Proposition has a crucial role in this chapter.

Proposition 4.2.2. If (R, m) is a local ring and there is an ideal I

of R such that I 6= mi for every i, then R has at least three distinct
non-trivial ideals J, K and L such that J, K, L 6= mi for every i.

Proof. Assume that R has an ideal I such that I 6= mi for ev-

ery i. Then by Lemma 4.2.1, dim(m/m2 ) = n ≥ 2. Therefore,
/ m2 such that
by Nakayama’s Lemma, we can find x1 , x2 , . . . , xn ∈
m = hx1 , x2 , . . . , xn i. Thus, Rx1 , Rx2 and R(x1 + x2 ) are the distinct
non-trivial ideals with desired properties. ✷

Remark 4.2.3. Let R be a ring. If |M ax(R)| ≥ 3, then R has a cycle

on six vertices.

Theorem 4.2.4. Let R = R1 × R2 × · · · × Rn be a commutative

Artinian non-local ring, where each (Ri , mi ) is a local ring and IΓ(R)
is connected. Then IΓ(R) is a tree if and only if R = R1 × R2 where

67
(R1 , m1 ) is a local ring with exactly one non-trivial ideal and R2 is a
field.

Proof. Suppose R = R1 × R2 where (R1 , m1 ) is a local ring with

exactly one non-trivial ideal and R2 is a field, then IΓ(R) ∼
= P4

Conversely, suppose IΓ(R) is tree. By Remark 4.2.3, |Max(R)| ≤ 2

and so n = 2. Assume that Ri be a local ring with mi 6= {0} for every
i. Then we have a cycle R1 × (0) − m1 × (0) − R1 × m2 − R1 × (0),
a contradiction. Hence Ri is a field for some i. Suppose Ri is a field
for every i, then Ω∗R ∼
= K 2 , a contradiction. Suppose R1 is a local ring
with m1 6= {0} and R2 is a field. Let I be any non-trivial ideal in R1
such that I 6= m1 . Then we have a cycle R1 × (0) − m1 × (0) − m1 ×
R2 − (0) × R2 − I × R2 − I × (0) − R1 × (0), a contradiction. Hence
m1 is the only non-trivial ideal in R1 . ✷

Theorem 4.2.5. Let R = R1 × R2 × · · · × Rn be a commutative

Artinian non-local ring with identity, where each (Ri , mi ) is a local
ring and |I(R)∗ | ≥ 3. Let ni be the nilpotency of mi . Then IΓ(R) is
unicyclic if and only if n = 3 and R ∼
= R1 × R2 × R3 where each Ri is
a field.

Proof. Assume that IΓ(R) is unicyclic. Suppose n ≥ 4. Then R1 ×

R2 × R3 × (0) × · · · × (0) − R1 × R2 × (0) × (0) × · · · × (0) − R1 ×
(0) × (0) × (0) × · · · × (0) − R1 × R2 × R3 × (0) × · · · × (0) and
68
(0) × R2 × R3 × R4 × · · · × (0) − (0) × (0) × R3 × R4 × · · · × (0) − (0) ×
(0) × (0) × R4 × · · · × (0) − (0) × R2 × R3 × R4 × · · · × (0) are two
distinct cycles in IΓ(R), a contradiction. Hence n ≤ 3.

Suppose n = 2. If Ri is a field for every i, then IΓ(R) ∼

= K 2, a
contradiction. Hence Ri is not a field for some i. Suppose mi 6= {0}
for every i. Then m1 × (0) − R1 × m2 − R1 × (0) − m1 × (0) and
(0)×m2 −m1 ×R2 −(0)×R2 −(0)×m2 are two distinct cycles in IΓ(R), a
contradiction. Hence we assume that R1 is a local ring with m1 6= {0}
and R2 is a field. If n1 ≥ 3, then R1 ×(0)−m1 ×(0)−m21 ×(0)−R1 ×(0)
and (0) × R2 − m1 × R2 − m21 × R2 − (0) × R2 are two distinct cycles
in IΓ(R), a contradiction. Hence n1 = 2.

Suppose there is an ideal I of R1 such that I 6= m1 . Then by

Proposition 4.2.2, R1 has at least three distinct non-trivial ideals I1 ,
I2 and I3 such that I1 , I2 , I3 6= m1 . Then R1 × (0) − m1 × (0) − m1 ×
R2 − (0) × R2 − I1 × R2 − I1 × (0) − R1 × (0) and R1 × (0) − m1 × (0) −
m1 × R2 − (0) × R2 − I2 × R2 − I2 × (0) − R1 × (0) are two distinct
cycles in IΓ(R), a contradiction. Hence R1 has exactly one non-trivial
ideal m1 and by Theorem 4.2.4, a contradiction.

R1 × R2 × (0) (0) × R2 × (0)

R1 × (0) × (0) (0) × R2 × R3

R1 × (0) × R3 (0) × (0) × R3

Figure 4.1: IΓ(R1 × R2 × R3 )

69
 Suppose n = 3. Suppose at least one of the Ri is not a field. Without
loss of generality, we assume that R1 is a local ring with m1 6= {0}.
Then m1 × (0) × (0) − R1 × R2 × (0) − R1 × (0) × (0) − m1 × (0) × (0)
and m1 × R2 × R3 − (0) × (0) × R3 − (0) × R2 × R3 − m1 × R2 × R3
are two distinct cycles in IΓ(R), a contradiction. Hence Ri is a field
for every i.

Converse follows from Figure. 4.1. ✷

Theorem 4.2.6. Let R = R1 × R2 × · · · × Rn be a commutative ring

with n ≥ 2. Then IΓ(R) is a split graph if and only if one of the
following conditions holds:

(i) R = R1 × R2 where each Ri is a field;

(ii) R = R1 × R2 where (R1 , m1 ) is a local ring with m1 is the only

non-trivial ideal and R2 is a field.

Proof. Suppose R = R1 ×R2 where each Ri is a field. Then IΓ(R) ∼

=
K 2 . Suppose R = R1 × R2 and (R1 , m1 ) is a local ring with m1 is the
only non-trivial ideal and R2 is a field. Then IΓ(R) ∼
= P4 .

Conversely, assume that IΓ(R) is a split graph. Suppose that n ≥ 3.

Consider the set Ω = {x1 , x2 , y1 , y2 } where x1 = R1 ×(0)×(0)×· · ·×(0),
x2 = R1 × R2 × (0) × · · · × (0), y1 = (0) × R2 × R3 × · · · × (0),
y2 = (0) × (0) × R3 × · · · × (0) are the non-trivial ideals in R. Then
the subgraph induced by Ω contains a graph isomorphic to 2K2 and
70
by Theorem 2.3.26, IΓ(R) is not a split graph, a contradiction. Hence
n = 2.

Suppose that Ri is a local ring with mi 6= {0} for every i. Consider

the set Ω1 = {x1 , x2 , x3 , x4 } where x1 = R1 × m2 , x2 = m1 × (0),
x3 = m1 × R2 , x4 = (0) × m2 are the non-trivial ideals in R. Then
the subgraph induced by Ω1 contains a graph isomorphic to C4 and
by Theorem 2.3.26, IΓ(R) is not a split graph, a contradiction. Hence
Ri is a field for some i.

Without loss of generality, we assume that R1 is a local ring with

m1 6= {0} and R2 is a field. Let I be any non-trivial ideal in R1 with
I 6= m1 . Consider the set Ω2 = {u1 , u2 , v1 , v2 } where u1 = R1 × (0),
u2 = m1 × (0), v1 = (0) × R2 , v2 = I × R2 are the non-trivial ideals
in R. Then the subgraph induced by Ω2 contains a graph isomorphic
to 2K2 . Therefore by Theorem 2.3.26, IΓ(R) is not a split graph, a
contradiction. Hence m1 is the only non-trivial ideal in R1 .

If Ri is field for all i, then IΓ(R) ∼

= K 2. ✷

Theorem 4.2.7. Let R = R1 ×R2 ×· · ·×Rn be a commutative Artinian

non-local ring with identity, where each (Ri , mi ) is a local ring. Let
ni be the nilpotency of mi . Then IΓ(R) is triangle-free if and only if
n ≤ 3 and one of the following conditions holds:

(i) R ∼
= R1 × R2 or R ∼
= R1 × R2 × R3 where each Ri is a field;

71
(ii) R ∼
= R1 × R2 where (R1 , m1 ) is a local ring with n1 = 2 and have
at least one non-trivial ideal and R2 is a field.

Proof. Assume that IΓ(R) is triangle-free. Suppose that n ≥ 4.

Then we have a triangle R1 × (0)× (0)× (0)× · · ·×(0) - R1 × R2 × R3 ×
(0)×· · ·×(0) - R1 ×R2 ×(0)×(0)×· · ·×(0) - R1 ×(0)×(0)×(0)×· · ·×(0),
a contradiction. Hence n ≤ 3.

Suppose that n = 3. Suppose at least one of the Ri is a local ring

with mi 6= {0}. Without loss of generality, assume that R1 is a local
ring with m1 6= {0}. Then we have a triangle R1 × (0) × (0) − R1 ×
R2 × (0) − m1 × (0) × (0) − R1 × (0) × (0), a contradiction. Hence Ri
is a field for every i.

Suppose that n = 2. Suppose Ri is a local ring with mi 6= {0} for

every i. Then we have a triangle m1 ×(0)−R1 ×m2 −m1 ×m2 −m1 ×(0),
a contradiction. Hence Ri is a field for some i.

Without loss of generality, assume that R1 is a local ring with m1 6=

{0} and R2 is a field. Let n1 ≥ 3. Then we have a triangle m1 × R2 −
m21 × R2 − (0) × R2 − m1 × R2 , a contradiction. Hence n1 = 2.

72
 Ii × (0) Ii × R2

I2 × (0) I2 × R2

R1 × (0) (0) × R2
I1 × (0) I1 × R2

m 1 × (0) m 1 × R2

Figure 4.2: IΓ(R1 × R2 ) with n1 = 2

Conversely, suppose R = R1 × R2 × R3 where each Ri is a field, then

IΓ(R) is isomorphic to Fig 4.1. Suppose R = R1 × R2 where each Ri
is a field, then IΓ(R) ∼
= K 2 . Suppose R = R1 × R2 where (R1 , m1 ) is
a local ring with n1 = 2 and R2 is a field, then IΓ(R) is isomorphic to
Fig 4.2. ✷

Theorem 4.2.8. Let R = R1 ×R2 ×· · ·×Rn be a commutative Artinian

non-local ring with identity, where each (Ri , mi ) is a local ring. Let ni
be the nilpotency of mi . Then IΓ(R) is claw-free if and only if n ≤ 3
and one of the following conditions holds:

(i) R = R1 × R2 or R = R1 × R2 × R3 where each Ri is a field;

(ii) R = R1 × R2 where (Ri , mi ) is a local ring with mi is the only

non-trivial ideal for every i;

(iii) R = R1 × R2 where (R1 , m1 ) is a local ring with {mi1 : 1 ≤ i < s}

where ms1 = 0 are the non-trivial ideals and R2 is a field;

73
Proof. Suppose R = R1 × R2 × R3 where each Ri is a field, then
IΓ(R) ∼
= C6 . Suppose R = R1 × R2 where each Ri is a field, then
IΓ(R) ∼
= K 2 . Suppose R = R1 × R2 where (Ri , mi ) is a local ring with
mi is the only non-trivial ideal for every i, then IΓ(R) is isomorphic
to Fig 4.3.

m 1 × R2 m 1 × (0)

(0) × R2 R1 × (0)
m1 × m2

(0) × m 2 R1 × m 2

Figure 4.3

Conversely, assume that IΓ(R) is a claw-free graph. Suppose that

n ≥ 4. Then we have a claw with vertices {u1 , v1 , v2 , v3 } where u1 =
R1 × (0) × (0) × (0) × · · · × (0), v1 = R1 × R2 × (0) × (0) × · · · × (0),
v2 = R1 × (0) × R3 × (0) × · · · × (0), v3 = R1 × (0) × (0) × R4 × · · · × (0),
which is a contradiction. Hence n ≤ 3.

Suppose that n = 3. Suppose at least one of the Ri is not a field.

Without loss of generality, we may assume that R1 is a local ring
with maximal ideal m1 6= {0}. Then we have a claw with vertices
{x1 , y1 , y2 , y3 } where x1 = m1 × (0) × (0), y1 = R1 × (0) × R3 , y2 =
R1 × R2 × (0), y3 = m1 × R2 × R3 , which is a contradiction. Hence Ri
is a field for every i = 1, 2, 3.

Suppose that n = 2. Suppose Ri is a local ring with mi 6= {0} for

every i. Let ni > 2 for some i. Without loss of generality, assume

74
that n1 > 2. Then we have a claw with vertices {u1 , v1 , v2 , v3 } where
u1 = m1 × R2 , v1 = m1 × (0), v2 = (0) × R2 , v3 = m21 × m2 , which is a
contradiction. Hence ni = 2 for every i. Suppose I is any non-trivial
ideal in R1 such that I 6= m1 . Then we have a claw with vertices
{x1 , y1 , y2 , y3 } where x1 = R1 × m2 , y1 = m1 × (0), y2 = I × (0),
y3 = (0)×m2 , which is a contradiction. Hence m1 is the only non-trivial
ideal in R1 . Similarly, one can prove that m2 is the only non-trivial
ideal in R2 .

Suppose R1 is a local ring with m1 6= {0} and R2 is a field. Let

n1 ≥ 2. Suppose there is an ideal I of R1 such that I 6= mi1 where
1 ≤ i < n1 . Then by Proposition 4.2.2, R1 has at least three distinct
non-trivial ideals I1 , I2 and I3 such that I1 , I2 , I3 6= mi1 . Then we have
a claw with vertices {u1 , v1 , v2 , v3 } where u1 = R1 × (0), v1 = I1 × (0),
v2 = I2 × (0), v3 = I3 × (0), which is a contradiction. Then the set
of all non-trivial ideals of Ri are the set {mt1 : 1 ≤ t < n1 } where
mn1 1 = 0. ✷

4.3 Isomorphism properties of a Cayley graph

If R ∼
= S, then IΓ(R) ∼
= IΓ(S). For the converse of this, consider
R = Z4 × Z2 and S = Z9 × Z3 . Then IΓ(R) ∼
= IΓ(S) ∼
= P4 , whereas
R and S are not isomorphic.

75
 n m n m
Ri′ Fj′ be
Q Q Q Q
Theorem 4.3.1. Let R = Ri × Fj and S = ×
i=1 i=1 i=1 i=1
a commutative rings with n ≥ 2 where (Ri , mi ) and (Ri′ , m′i ) are local
rings with mi 6= {0} and m′i 6= {0} and each Fj and Fj′ are fields. Let
ki be the number of ideals in Ri and ki′ be the number of ideals in Ri′ .
Then IΓ(R) ∼
= IΓ(S) if and only if ki = ki′ for all i, 1 ≤ i ≤ n.

Proof. Suppose R ∼
= S, then the result is obvious. Assume that
R ≇ S. Suppose that ki = ki′ for all i, 1 ≤ i ≤ n. Then |V (IΓ(R))| =
|V (IΓ(S))|. Let I(Ri ) = {I1i = (0), I2i = m1 , I3i , . . . , Iki i = Ri } be the
set of ideals in Ri and I(Ri′ ) = {I1i
′ ′
= (0), I2i ′
= m1 , I3i , . . . , Ik′ i i = Ri′ }
be the set of ideals in Si . Then the map Iti → Iti′ is bijection from I(Ri )
n m
onto I(Ri′ ). Define φ : V (IΓ(R)) −→ V (IΓ(S)) by φ( Iti × Jj ) =
Q Q
i=1 i=1
n m
Iti′ × Jj′ where
Q Q
i=1 i=1


Fj′

if Jj = Fj
Jj′ =
(0) if J = (0)

j

n
Q m
Q
Clearly φ is well defined and bijective. Let I = Ii × Jj and
i=1 i=1
n
Q m
Q
J = Ai × Bj be two non-zero ideals in R. Suppose I and J
i=1 i=1
are adjacent in IΓ(R). Then I = JL or J = IL and so Ii = Ai L
or Ai = Ii L, Jj = Bj L or Bj = Jj L for all i, j where L is any non-
n m n m
′ ′
Ai × Bj′ .
′
Q Q Q Q
trivial ideal in R. Let φ(I) = Ii × Jj and φ(J) =
i=1 i=1 i=1 i=1
By definition of φ, Ii′ = A′i L′ or A′i = Ii′ L′ , Jj′ = Bj′ L′ or Bj′ = Jj′ L′

76
for all i, j and so φ(I) = φ(J)L1 or φ(J) = φ(I)L1 . Hence φ(I) and
φ(J) are adjacent in IΓ(S). Similarly one can prove that φ preserves
non-adjacency also. Hence IΓ(R) ∼
= IΓ(S).

Conversely, assume that IΓ(R) ∼

= IΓ(S). Suppose ki 6= ki′ for some
6 |V (IΓ(S))|. Hence ki = ki′ for all i.
i. Then |V (IΓ(R))| = ✷

Using Theorem 4.3.1, we have the following corollary.

n n
Fi′ , where each Fi and
Q Q
Corollary 4.3.2. Let R = Fi and S =
i=1 i=1
Fi′ are fields and n ≥ 2. Then IΓ(R) ∼
= IΓ(S).

Corollary 4.3.3. Let R = R1 × · · · × Rn and S = S1 × · · · × Sn be

a commutative rings with n ≥ 2 where (Ri , mi ) and (Si , m′i ) are local
rings with mi 6= {0} and m′i 6= {0}. Let ki be the number of ideals in
Ri and ki′ be the number of ideals in Si . Then IΓ(R) ∼
= IΓ(S) if and
only if ki = ki′ for all i, 1 ≤ i ≤ n.

4.4 Crosscap two of IΓ(R)

The main goal of this section is to determine commutative Artinian

non-local rings R for which IΓ(R) has crosscap two.

The following results about the planarity are very useful in the sub-
sequent sections.

77
Theorem 4.4.1. Let R = R1 × R2 × · · · × Rn × F1 × F2 × · · · × Fm
be a commutative ring with identity where each (Ri , mi ) is a local ring
with mi 6= {0} and Fj is a field, n, m ≥ 1 and n + m ≥ 3. Let ni be the
nilpotency of mi . Then IΓ(R) is planar if and only if n = 1, m = 2,
n1 = 2 and m1 is the only non-trivial ideal in R1 .

Theorem 4.4.2. Let R = R1 ×F1 be a commutative ring with identity

where (R1 , m1 ) is a local ring with m1 6= {0} is a principal ideal and
F1 is a field. Let n1 be the nilpotency of m1 . Then IΓ(R) is planar if
and only if one of the following conditions holds:

(i) n1 = 2 and m1 is the only non-trivial ideal in R1 ;

(ii) n1 = 3 and m1 , m21 are the only non-trivial ideals in R1 ;

(iii) n1 = 4 and m1 , m21 , m31 are the only non-trivial ideals in R1 .

Theorem 4.4.3. Let R = F1 ×F2 ×· · ·×Fn be a commutative ring with

identity where each Fi is a field and n ≥ 2. If IΓ(R) is non-planar,
then γ(IΓ(R)) > 2.

Proof. Assume that IΓ(R) is non-planar. Then by Theorem 2.5.12,

n ≥ 5. Let G1 be the subgraph of IΓ(R) induced by
Ω1 = {x1 , x2 , x3 , y1 , y2 , y3 , z1 , z2 } where x1 = F1 × (0) × (0) × (0) ×
(0) × · · · × (0), x2 = (0) × F2 × (0) × (0) × (0) × · · · × (0), x3 =
F1 ×F2 ×(0)×(0)×(0)×· · ·×(0), y1 = F1 ×F2 ×(0)×F4 ×(0)×· · ·×(0),
78
y2 = F1 × F2 × (0) × F4 × F5 × · · · × (0), y3 = F1 × F2 × F3 ×
F4 × (0) × · · · × (0), z1 = F1 × F2 × F3 × (0) × (0) × · · · × (0),
z2 = (0) × (0) × (0) × F4 × (0) × · · · × (0) are the vertices of IΓ(R) and
G2 be the subgraph of IΓ(R) induced by Ω2 = {u1 , u2 , u3 , v1 , v2 , v3 }
where u1 = (0) × (0) × F3 × (0) × (0) × · · · × (0), u2 = (0) × (0) × (0) ×
(0) × F5 × · · · × (0), u3 = (0) × (0) × F3 × (0) × F5 × · · · × (0), v1 =
F1 ×(0)×F3 ×(0)×F5 ×· · ·×(0), v2 = F1 ×(0)×F3 ×F4 ×F5 ×· · ·×(0),
v3 = (0) × (0) × F3 × F4 × F5 × · · · × (0) are the vertices of IΓ(R).
Then G1 is isomorphic to Fig 4.4 and so by Theorem 6.5.1 [22, p.197],
γ(G1 ) > 1. Also G2 ∼
= K3,3 and so by Lemma 2.4.4, γ(G2 ) = 1. Hence
γ(IΓ(R)) ≥ γ(G1 ) + γ(G2 ) > 2. ✷

z1

x1 x2 x3

y1 y2 y3

z2
Figure 4.4: Forbidden subgraph for the projective plane

Theorem 4.4.4. Let R = R1 × R2 × · · · × Rn be a commutative ring

with identity where each (Ri , mi ) is a local ring with mi 6= {0} and
n ≥ 2. Let ni be the nilpotency of mi . Then γ(IΓ(R)) = 2 if and only
if n = 2 and one of the following conditions holds:

(i) n1 = 2, n2 = 3 and m1 and m2 , m22 are the only non-trivial ideals

79
 in R1 and R2 respectively.

(ii) n1 = 3, n2 = 2 and m1 , m21 and m2 are the only non-trivial ideals

in R1 and R2 respectively.

(iii) n1 = n2 = 2 and R1 has exactly one non-trivial ideal say m1 and

R2 has exactly four distinct non-trivial ideals or R1 has exactly
four distinct non-trivial ideals and R2 has exactly one non-trivial
ideal say m2 .

Proof. Assume that γ(IΓ(R)) = 2. Suppose that n ≥ 3. Consider

the set Ω = {x1 , x2 , x3 , x4 , y1 , y2 , y3 , y4 , y5 , y6 } where x1 = R1 × R2 ×
m3 × (0) × · · · × (0), x2 = m1 × R2 × R3 × (0) × · · · × (0), x3 =
R1 × m2 × R3 × (0) × · · · × (0), x4 = m1 × m2 × m3 × (0) × · · · × (0),
y1 = m1 × (0) × (0) × (0) × · · · × (0), y2 = (0)× m2 × (0) × (0) × · · · × (0),
y3 = (0) × (0) × m3 × (0) × · · · × (0), y4 = m1 × m2 × (0) × (0) × · · · × (0),
y5 = m1 × (0) × m3 × (0) × · · · × (0), y6 = (0) × m2 × m3 × (0) × · · · × (0)
are the non-trivial ideals in R. Then the subgraph induced by Ω
contains K4,6 as a subgraph of IΓ(R). Therefore by Lemma 2.4.4,
γ(IΓ(R)) > 2, a contradiction. Hence n = 2.

Suppose that ni ≥ 3 for every i. Consider the set

Ω1 = {x1 , x2 , x3 , y1 , y2 , y3 , y4 , y5 , y6 , y7 } where x1 = m1 × R2 , x2 =
R1 × m2 , x3 = m1 × m2 , y1 = m1 × (0), y2 = m21 × (0), y3 = (0) × m2 ,
y4 = (0) × m22 , y5 = m1 × m22 , y6 = m21 × m2 , y7 = m21 × m22 are the non-
trivial ideals in R. Then the subgraph induced by Ω1 contains K4,6

80
as a subgraph of IΓ(R). Therefore by Lemma 2.4.4, γ(IΓ(R)) > 2, a
contradiction. Hence ni = 2 for some i.

Without loss of generality, assume that n1 = 2. Suppose that n2 > 3.

Consider the set Ω2 = {u1 , u2 , u3 , u4 , v1 , v2 , v3 , v4 , v5 } where u1 = m1 ×
R2 , u2 = R1 × m2 , u3 = m1 × m2 , u4 = R1 × m22 , v1 = m1 × (0),
v2 = (0) × m22 , v3 = m1 × m22 , v4 = m1 × m32 , v5 = (0) × m32 are the
non-trivial ideals in R. Then the subgraph induced by Ω2 contains
K4,5 as a subgraph of IΓ(R) and so by Lemma 2.4.4, γ(IΓ(R)) > 2, a
contradiction. Hence n2 ≤ 3.

Suppose that n2 = 3. Suppose there is an ideal J of R2 such that

J 6= m2 . Then by Proposition 4.2.2, R2 has at least three distinct non-
trivial ideals J1 , J2 and J3 such that J1 , J2 , J3 6= m2 . Consider the set
Ω3 = {x1 , x2 , x3 , x4 , y1 , y2 , y3 , y4 , y5 } where x1 = (0)×R2 , x2 = m1 ×R2 ,
x3 = [R1 × (0), R1 × J1 , R1 × J2 , R1 × J3 , R1 × m2 ], x4 = [m1 × (0), m1 ×
m2 , m1 ×J1 , m1 ×J2 , m1 ×J3 ], y1 = (0)×m2 , y2 = (0)×m22 , y3 = (0)×J1 ,
y4 = (0) × J2 , y5 = (0) × J3 are the vertices of IΓ(R). Then the
subgraph induced by Ω3 contains K4,5 as a subgraph of IΓ(R) and so
by Lemma 2.4.4, γ(IΓ(R)) > 2, a contradiction. Hence m2 and m22 are
the only non-trivial ideals in R2 .

81
 m1 × R2 R1 × m2

m1 × m22
m1 × (0) (0) × m2

m1 × m2 R1 × m22

(0) × m2 m1 × (0)
(0) × m22 R1 × (0)
(0) × R2

m1 × R2 R1 × m2
Fig 4.5: Embedding of IΓ(R1 × R2 ) with n1 = 2, n2 = 3

Suppose there is an ideal I of R1 such that I 6= m1 . Then by

Proposition 4.2.2, R1 has at least three distinct non-trivial ideals
I1 , I2 and I3 such that I1 , I2 , I3 6= m1 . Consider the set Ω4 =
{u1 , u2 , u3 , u4 , v1 , v2 , v3 , v4 , v5 } where u1 = m1 × (0), u2 = I1 × (0),
u3 = I2 × (0), u4 = I3 × (0), v1 = R1 × (0), v2 = R1 × m2 , v3 = R1 × m22 ,
v4 = [(0) × R2 , m1 × R2 , I1 × R2 , I2 × R2 , I3 × R2 ], v5 = [(0) × m2 , m1 ×
m2 , I1 × m2 , I2 × m2 , I3 × m2 ] are the vertices of IΓ(R). Then the sub-
graph induced by Ω4 contains K4,5 as a subgraph of IΓ(R) and so by
Lemma 2.4.4, γ(IΓ(R)) > 2, a contradiction. Hence m1 is the only
non-trivial ideal in R1 .

Suppose that n2 = 2. Suppose R2 has at least four distinct non-

trivial ideals different from m2 . Let J1 , J2 , J3 , J4 be the distinct
non-trivial ideals such that Ji 6= m2 for all i. Consider the set Ω5 =
{x1 , x2 , x3 , x4 , y1 , y2 , y3 , y4 , y5 } where x1 = m1 × R2 , x2 = (0) × R2 ,

82
x3 = [R1 × (0), R1 × J1 , R1 × J2 , R1 × J3 , R1 × J4 , R1 × m2 ], x4 =
[m1 × (0), m1 × J1 , m1 × J2 , m1 × J3 , m1 × J4 , m1 × m2 ], y1 = (0) × J1 ,
y2 = (0) × J2 , y3 = (0) × J3 , y4 = (0) × J4 , y5 = (0) × m2 are
the vertices of IΓ(R). Then the subgraph induced by Ω5 contains
K4,5 as a subgraph of IΓ(R) and so by Lemma 2.4.4, γ(IΓ(R)) > 2,
a contradiction. Hence by Proposition 4.2.2, R2 has exactly three
distinct non-trivial ideals, say J1 , J2 , J3 different from m2 (or) R2 has
exactly one non-trivial ideal say m2 .

Suppose R2 has exactly three distinct non-trivial ideals, say J1 , J2 ,

J3 different from m2 . Let I be any non-trivial ideal in R1 such that
I 6= m1 . Consider the set Ω6 = {u1 , u2 , u3 , u4 , v1 , v2 , v3 , v4 , v5 } where
u1 = (0)×J1 , u2 = (0)×J2 , u3 = (0)×J3 , u4 = (0)×m2 , v1 = m1 ×R2 ,
v2 = (0) × R2 , v3 = [R1 × (0), R1 × J1 , R1 × J2 , R1 × J3 , R1 × m2 ],
v4 = [m1 × (0), m1 × J1 , m1 × J2 , m1 × J3 , m1 × m2 ], v5 = [I × (0), I ×
J1 , I × J2 , I × J3 , I × m2 ] are the vertices of IΓ(R). Then the subgraph
induced by Ω6 contains K4,5 as a subgraph of IΓ(R) and so by Lemma
2.4.4, γ(IΓ(R)) > 2, a contradiction. Hence m1 is the only non-trivial
ideal in R1 .

Similar argument for other possibilities also.

83
 m1 × R2 (0) × R2

(0) × J3

(0) × J1 m 1 × J1
m 1 × J3
R1 × J3
R1 × J2
(0) × J2
m 1 × (0)
R1 × (0)

m 1 × J2 R1 × J1
R1 × m 2
(0) × J2 m1 × m2 (0) × J1
(0) × m 2

m1 × R2 (0) × R2
Figure 4.6: Embedding of IΓ(R1 × R2 ) with n1 = n2 = 2

Converse follows from Figs 4.5 and 4.6. ✷

Lemma 4.4.5. Let R = R1 × F1 × F2 be a commutative ring with

identity where (R1 , m1 ) is a local ring with m1 6= {0} and each Fi is
a field. Let n1 be the nilpotency of m1 . If IΓ(R) is non-planar, then
γ(IΓ(R)) > 2.

Proof. Assume that IΓ(R) is non-planar. Suppose that n1 > 2.

Let G be the subgraph of IΓ(R) induced by Ω = {x1 , x2 , . . . , x14 }
where x1 = (0) × F1 × (0), x2 = m1 × F1 × (0), x3 = m21 × F1 × (0),
x4 = (0) × (0) × F2 , x5 = m1 × (0) × F2 , x6 = m21 × (0) × F2 , x7 =
(0)×F1 ×F2 , x8 = m1 ×F1 ×F2 , x9 = m21 ×F1 ×F2 , x10 = m1 ×(0)×(0),
x11 = m21 × (0) × (0), x12 = R1 × (0) × (0), x13 = R1 × F1 × (0),
x14 = R1 ×(0)×F2 are the vertices of IΓ(R). Then < Ω > is a subgraph

84
of IΓ(R) with |V (IΓ(Ω))| = 14, |E(IΓ(Ω))| = 44 and gr(IΓ(Ω)) = 3.
Then by Lemma 2.4.5, γ(IΓ(R)) > 2.

Suppose that n1 = 2. Then by Theorem 4.4.1, R1 has at least

two non-trivial ideals. Suppose there is an ideal I of R1 such that
I 6= m1 . Then by Proposition 4.2.2, R1 has at least three distinct
non-trivial ideals I1 , I2 and I3 such that I1 , I2 , I3 6= m1 . Consider
the set Ω1 = {u1 , u2 , u3 , u4 , v1 , v2 , v3 , v4 , v5 } where u1 = m1 × (0) ×
(0), u2 = I1 × (0) × (0), u3 = I2 × (0) × (0), u4 = I3 × (0) × (0),
v1 = R1 × F1 × (0), v2 = R1 × (0) × F2 , v3 = R1 × (0) × (0), v4 =
[(0) × F1 × (0), I1 × F1 × (0), I2 × F1 × (0), I3 × F1 × (0), m1 × F1 × (0)],
v5 = [(0)×(0)×F2 , I1 ×(0)×F2 , I2 ×(0)×F2 , I3 ×(0)×F2 , m1 ×(0)×F2 ]
are the vertices of IΓ(R). Then the subgraph induced by Ω1 contains
K4,5 as a subgraph of IΓ(R) and so by Lemma 2.4.4, γ(IΓ(R)) > 2.
✷

Theorem 4.4.6. Let R = R1 × R2 × · · · × Rn × F1 × F2 × · · · × Fm

be a commutative ring with identity where each (Ri , mi ) is a local ring
with mi 6= {0} and each Fi is a field, n, m ≥ 1 and n + m ≥ 3. Let ni
be the nilpotency of mi . If IΓ(R) is non-planar, then γ(IΓ(R)) > 2.

Proof. Assume that IΓ(R) is non-planar. It is clear that IΓ(R1 ×

F1 ×F2 ) is a subgraph of IΓ(R). Then by Lemma 4.4.5, γ(IΓ(R)) > 2.
✷

Theorem 4.4.7. Let R = R1 ×F1 be a commutative ring with identity

85
where (R1 , m1 ) is a local ring with m1 6= {0} is a principal ideal and
F1 is a field. Let n1 be the nilpotency of m1 . If IΓ(R) is non-planar,
then γ(IΓ(R)) > 2.

Proof. Assume that IΓ(R) is non-planar. Then by Theorem 4.4.2,

n1 ≥ 5. Let u1 = m21 × (0), u2 = m31 × (0), u3 = m41 × (0), v1 = R1 × (0),
v2 = m1 × (0), v3 = m1 × R2 , v4 = m21 × R2 , x1 = (0) × R2 ,
x2 = m31 × R2 , x3 = m41 × R2 be the vertices of Ω∗R . Let G be
the subgraph of IΓ(R) induced by {u1 , u2 , u3 , v1 , v2 , v3 , v4 , x1 , x2 , x3 },
G′ = G−{u1 u2 , u2 , u3 , u1 u3 , v1 v2 , v2 v3 , v3 v4 } and G′′ = G′ −{x1 , x2 , x3 }.
Then G′′ ∼
= K3,4 and so γ(G′′ ) = 1. Since γ(G) = 1 and γ(G′′ ) ≤
γ(G′ ) ≤ γ(G), we get γ(G′ ) = 1. Note that |V (G′ )| = 10, |E(G′ )| =
24. Then by Euler’s formula, there are 14 faces when drawing G′ on
a torus. Fix a representation of G′ and let {F1′ , . . . , F14
′
} be the set of
faces of G′ corresponding to the representation. Let {F1′′ , . . . , Fn′′ } be
the set of faces of G′′ obtained by deleting x1 , x2 , x3 and all the edges
incident with x1 , x2 , x3 from the representation of G′ . Notice that
G′′ ∼
= K3,4 . From the fact that n − m + f = 2 − 2g, K3,4 has 5 faces,
one octagonal faces and 4 rectangular faces, or two hexagonal faces and
3 rectangular faces. So n = 5. Moreover, for every i, each boundary
of Fi′′ cannot have consecutive repetition of a single edge. Therefore
in K3,4 , the only way to have a closed walk of length 6 without con-
secutive repetition of single edge is to have 6−cycle and the only way
to have a closed walk of length 8 without consecutive repetition of
86
single edge is to have 8−cycle. Then in K3,4 , all faces boundaries are
4−cycles but with two 6−cycle or one 8−cycle. Now {F1′ , . . . , F14
′
}
can be recovered by inserting x1 , x2 , x3 and all the edges incident with
x1 , x2 , x3 into the representation corresponds to {F1′′ , . . . , F5′′ }. Note
that x1 x2 , x2 x3 , x1 x3 ∈ E(G′ ). Hence x1 , x2 , x3 should be inserted
to the same faces say Fm′′ of G′′ to avoid crossing. Also note that
x1 v3 , x1 v4 , x2 v3 , x2 v4 , x2 u3 , x2 u2 , x3 v3 , x3 v4 , x3 u3 ∈ E(G′ ) and therefore
u2 , u3 , v3 , v4 are the boundary vertices of Fm′′ . Consider the following
edges of G. Let e1 = x2 x3 , e2 = x2 u3 , e3 = x2 v3 , e4 = x2 v4 , e5 = x2 u2 ,
e6 = x3 v3 , e7 = x3 v4 , e8 = x3 u3 . From this, it is clear that x2 , x3
should be inserted into the same face. Since the vertices x2 and x3
have three neighbors in common, they cannot be inserted into the
same face in the embedding of G′ without crossing. This implies that
γ(IΓ(R)) > 2. ✷

87
Chapter 5

The Strongly annihilating ideal

graph of ring

In [44], N. KH. Tohidi, M. J. Nikmehr and R. Nikandish have in-

troduced and investigated the strongly annihilating ideal graph of a
commutative ring. For a non-domain commutative ring R, let A∗ (R)
be the set of non-zero ideals with non-zero annihilators. The strongly
annihilating ideal graph of R is defined as the graph SAG(R) with the
vertex set A∗ (R) and two distinct vertices I1 and I2 are adjacent if and
only if I1 ∩ Ann(I2 ) 6= (0) and I2 ∩ Ann(I1 ) 6= (0). The contents of this
chapter have been communicated as a paper in “AKCE International
Journal of Graphs and Combinatorics”.

88