Introduction

Group rings are a special class of rings. They are built by combining two of the most
important structures in Algebra: groups and commutative rings. Interestingly, the
resultant group rings are not necessarily commutative. So group rings provide an
additional example of noncommutative rings outside of the usual example of matrix
rings. Further, we’ll see that group rings can be viewed as modules over the component
ring, with an additional way to multiply elements of the module. In particular, if the ring
is a field, its group rings can be viewed as vector spaces over the field, with the additional
multiplicative operation. This paper will provide an overview of group rings, explore
conditions under which group rings have zero divisors, and discuss ways to view group
rings in terms of simple components.

Section 1. Basic Definitions and Properties

We begin our exploration of group rings by looking at some useful facts about group
rings. These properties will not only provide some support for some of the later theorems,
but also show that group rings are different enough from generic rings to make them
worth studying. Throughout this paper we assume that all rings are commutative with
1  0 . We begin with the definition of group ring.

Definition 1.1. Let R be a ring, and let G   g1 , , g n  be a finite group whose operation

is written multiplicatively. Then the group ring of G with coefficients in R is the set of all
formal sums:

RG  a1 g1  a2 g2   an g n : ai  R for 1  i  n ,

with operations as defined below. This definition is extended to group rings of infinite

1
groups by only considering finite sums (i.e., sums in which only finitely many
coefficients are nonzero). The ordering of terms in the sum is irrelevant. So two elements
of a group ring are considered equal if and only if the coefficients of each group element
are equal. That is, if a1 g1  a2 g 2   an g n and b1 g1  b2 g 2   bn g n are elements of

RG , then a1 g1  a2 g 2   an g n  b1 g1  b2 g 2   bn g n if and only if ai  bi for each i

between 1 and n. Based on the context, we will usually write elements of RG as a g

i i ,
n
representing the element a g
i 1
i i . Similarly, we may write an element as  a x , and
xG
x

again without the index as a x. x

Addition is defined componentwise:

 a1 g1  a2 g2   an g n    b1 g1  b2 g 2   bn g n 
  a1  b1  g1   a2  b2  g 2    an  bn  g n .

That is,  a g   b g   a  b  g .

i i i i i i i

Multiplication is defined by

 a1 g1  a2 g2   an gn  b1 g1  b2 g 2   bn g n   c1 g1  c2 g 2   cn g n ,

where ck , the coefficient of g k , is  gi g j  g k

ai b j . That is,

n  
  i i    j j     aib j  gk .
a g b g 
k 1  gi g j  gk 

2
Since elements of RG are finite sums, this characterization of multiplication is valid
even in the case that G is an infinite group.

Proposition 1.1. With the operations defined above, RG is a ring with unity.

Proof: First, we show that  RG,   is an abelian group. It is easy to check that  0g is

the additive identity. It is also clear that for an element  a g , the additive inverse is
g

  a  g . The remaining group properties for RG are inherited from the ring R. As an
g

example, we check the associativity of addition in RG . Let

gG
 a g ,  b g , and  c g
g
gG
g
gG
g

be elements of RG . Then, since addition is associative in R,

 
  ag g   bg g    cg g    ag  bg  g   cg g
 gG gG  gG gG gG


   ag  bg   cg g
gG


  ag   bg  cg  g
gG

  ag g    bg  cg  g
gG gG

 
  ag g    bg g   cg g  ,
gG  gG gG 

and hence addition is associative in RG .

The associativity of multiplication likewise follows from the corresponding properties in

the ring R and the group G.

3
Note that for any i and k, there is a unique j with g i g j  g k . By the definition of addition

in RG , and because addition is commutative and associative in RG , for any a g

i i and

b g j j in RG ,

 
  a g    b g      aib j  g k    aib j  gi g j .
n

i i j j
k 1  gi g j  g k

 i , j 

Notice that for a particular pair  i, j  , the definition of multiplication in RG implies that

 a b  g g    a g  b g  . Therefore,
i j i j i i j j

  a g    b g     a b  g g    a g  b g  .
i i j j i j i j i i j j
    i, j i, j

It is now straightforward to check the distributive properties in RG .

The unity element of RG is 1R1G since for every  a x  RG

x we have

1R1G   ax x    1R ax 1G x    ax x    ax1R  x1G     ax x 1R1G .

Example. If G  D8  s, r | s 2  r 4  1, rs  sr 3 is the dihedral group of order 8, and

R , then   3r  2sr 2 and   2s  sr 2 are elements of D8 . Then

    3r   2  1 sr 2  2 s
 3r  sr 2  2s
and

4
    3r  2 sr 2  2 s  sr 2 
 6rs  3rsr 2  4 sr 2 s  2 sr 2 sr 2
 6 sr 3  3sr  4r 2  2 1G.

In the case where F is a field, an alternate characterization of the group ring FG is as an

F-vector space with basis G and the additional multiplicative operation defined above. As
we have already seen, FG is an additive abelian group. The remaining parts of the
definition of vector space are straightforward to verify.

There are a few quick facts worth pointing out regarding group rings. Note that we can
identify the ring R with the subring r1G : r  R of RG using the injective ring

homomorphism  : R  RG by   r   r1G . Also, we can identify the group G with the

subgroup 1g : g  G of  RG  , the multiplicative group of invertible elements in RG ,



using the injective group homomorphism  : G   RG  by   g   1R g .



In addition to the group G and the ring R both residing within the group ring RG , group
rings also inherit some other properties from the group and the ring.

Proposition 1.2. Let G be a group and R be a ring. Then RG is a commutative ring if

and only if G is abelian.

Proof: First suppose that RG is a commutative ring. Then G is abelian as a corollary to

the fact that G   RG  .



Now, suppose that G is abelian, and let  a g ,  b h  RG . Then, since R is a

gG
g
hG
h

commutative ring,

5
   
  ag g    bh h     ag bh   gh 
 gG   hG   g ,h 
   b a   hg 
h g
 g ,h

  
   bh h    ag g 
 hG   gG 

and hence RG is a commutative ring.

The next proposition is straightforward to check, but shows that subrings of the ring R
and subgroups of the group G translate into subrings of the group ring RG as one would
expect.

Proposition 1.3. Let G be a group and let R be a ring.

(i) If S is a subring of R, then SG is a subring of RG ; and
(ii) If H  G , then RH is a subring of RG .

Note that if 1G is the identity in G, then for every a  R we write a1G as a . If 1R is the

identity in R, we write 1R g as g for every g  G . In particular, we simply write 1R1G as

1.

The following propositions show that group rings of finite groups truly differ from
arbitrary rings in the general sense. In particular, most group rings have nontrivial centers
and, in the case that the ring is a field, group rings have maximal ideals.

Proposition 1.4. Let G be a group with a nontrivial finite normal subgroup H, and let R
be a ring. Then RG has a nontrivial center; specifically,    hH h is in the center of

the group ring RG . In particular, if G is finite, then G is a nontrivial finite normal

subgroup of itself, and hence RG has nontrivial center.

6
Proof: Since H G , gH  Hg for every g  G . Since  is the sum of all of the
elements of H, for every g  G , g is the sum of all of the elements of gH . Similarly,
 g is the sum of all of the elements of Hg . Since gH  Hg , it follows that g   g .

That is,  commutes with every element of G. Now let  a g  RG . Then

g

  a g       a g 
g g

 a  g  g

     a g  . g

Thus,    h is in the center of RG .

hH

Definition 1.2. Let R be a ring and let G   g1 , , g n  be a finite group. Then the map

f : RG  R by f  a g    a
i i i is called the augmentation map. Proposition 1.5

shows that this map is a homomorphism. Its kernel is called the augmentation ideal of
RG .

Proposition 1.5. Let R be a ring and let G   g1 , , g n  be a finite group. Then the

augmentation map is a homomorphism. Also, the augmentation ideal is generated by

g  1: g  G . Further, if R is a field then the augmentation ideal is maximal.

Proof: Let f : RG  R be the augmentation map. Let x   ai gi and y   b j g j be

elements of RG . Then

f  x  y  f    a  b g     a  b    a   b  f  x   f  y  .
i i i i i i i

Also, using the fact that f preserves sums,

7
  
f  xy   f    ai bi  gi g j    ai b j    ai    b j   f  x  f  y  .
 i , j   i , j 
 

The augmentation ideal is I   ai gi  RG :  ai  0 . Let I 0  g  1: g  G . Suppose

that x  I . If x   ai gi , then x  x   ai   ai  gi  1 . So x  I 0 . On the other hand,

note that for every g  G , f  g  1  f 1R g  1R1G   1R  1R  0 . So g  1  I for every

g  G . Thus, I  I 0 , and the augmentation ideal is generated by  g  1: g  G .

The augmentation map is clearly onto, so RG I  R . So if R is a field, then the quotient

RG I is a field, and thus I must be maximal [2, p. 224].

Section 2. Zero Divisors

We’ve already seen conditions under which a group ring is commutative. Taking our
investigation a step further, we can try to decide whether a group ring has zero divisors,
which will help in classifying group rings in the hierarchy of ring structures. For
example, in the case that the group ring is commutative and has no zero divisors, the
group ring is an integral domain.

Recall that a group element is called a torsion element when its order is finite.

Proposition 2.1. Let G be any group containing a torsion element g, and let R be a ring.
Then the group ring RG has zero divisors. So in particular, if G is a finite group, then the
group ring RG has zero divisors.

8
Proof. Set m  ord  g  . Then 1  g and 1  g   g m1 are nonzero elements of RG ,

and 1  g  1  g   g m1   1  g m  1  1  0 .

Whether a group ring RG must have zero divisors when G is infinite and torsion-free is
an open problem. The current conjecture is that G is torsion-free if and only if RG has
no zero divisors. When R is a field, we know the conjecture to be true in the case of
abelian groups, free groups, and supersolvable groups [4, p. 174]. Proposition 2.2 shows
the conjecture is true in the special case where G is an abelian group and R is a field.

Proposition 2.2. Let G be an abelian group and let F be a field. Then G is torsion-free if
and only if FG has no zero divisors.

Proof. The backwards direction follows from Proposition 2.1.

For the forwards direction, suppose that G is torsion-free and let  ,   FG be such that
  0 . Since there are only finitely many group elements with nonzero coefficients in
each of the group ring elements, there is a finitely generated subgroup H of G such that
FH contains  and  . By the Fundamental Theorem of Finitely Generated Abelian

Groups, H is the direct product of cyclic groups x1 , , xm . Since G is torsion-free,

each xi is infinite. Then FH is contained in F  x1 , , xm  , the quotient field of the

polynomial ring F  x1 , , xm  . Since F  x1 , , xm  is a field, FH is an integral domain.

Hence,   0 or   0 , so FG has no zero divisors.

There is a related result that sheds some additional light on the zero divisor problem, but
first we need a definition.

Definition 2.1. A ring R is said to be prime if for all  ,   R ,  R  0 implies that

  0 or   0 .