,
there exists an element w R such that zw = 1 and wz = 1. Since z S, we have zr = rz
for all r R. We also have the implications
zr = rz = w(zr) = w(rz) = (wz)r = (wr)z = 1r = (wr)z = r = (wr)z
= rw =
_
(wr)z
_
w = rw = (wr)(zw) = rw = (wr) 1 = rw = wr .
Thus, if we assume that z S, then wr = rw for all r R. Therefore, w S. We have
proved that if z is a nonzero element of S, then there exists an element w S such that
zw = 1 and wz = 1. Hence S is a division ring.
Finally, if a S, then ar = ra for all r R. Since S R, we can say that ab = ba for
all b S. Hence S is a commutative ring. Since S has been proved to be a division ring, it
follows that S is a eld. We have proved that if R is a division ring, then the center of R is
a eld.
Page 231, problem 8. Let H denote the ring of quaternions. Suppose that a, b, c, d R
and that = a + bi + cj + dk is in the center of H. It follows that = for all H.
We will rst take = i and then we will take = j. We have
i = ai + b(1) + c(k) + dj = b + ai + dj + (c)k,
i = ai + b(1) + ck + (d)j = b + ai + (d)j + ck .
Therefore,
i = i d = d and c = c c = d = 0 .
Also,
j = aj + bk + c(1) + (d)i = c + (d)i + aj + bk,
j = aj + (b)k + c(1) + di = c + di + aj + (b)k .
Therefore,
j = j d = d and b = b b = d = 0 .
If is in the center of H, it follows that i = i and j = j, and therefore it follows that
b = c = d = 0. Thus, has the form = a+0i +0j +0k. In the denition of H, we identify
such a quaternion with the real number a, and thereby regard R as a subring of H. With
that identication, we have proved that if is in the center of H, then R. Conversely,
if R, then is in the center of H. This is part of the denition of multiplication in H.
Therefore, we have proved that the center of H is the subring R, which is explicitly given
as
R = { a + 0i + 0j + 0k | a R } .
Now let S = { a + bi + 0j + 0k | a, b R }. We will prove that S is a subring of H
and that S is isomorphic to C. First of all, note that 0
H
= 0 + 0i + 0j + 0k is clearly in
S. Suppose that a, b, a
, b
+ b
i + 0j + 0k) = (a + a
) + (b + b
)i + 0j + 0k,
(a + bi + 0j + 0k)(a
+ b
i + 0j + 0k) = (aa
bb
) + (ab
+ ba
)i + 0j + 0k .
Both of these elements of H are actually in S. Hence S is closed under the operations of
addition and multiplication for H. Furthermore, the additive inverse of a + bi + 0j + 0k
is (a) + (b)i + 0j + 0k, which is clearly in S. It follows that S is a subgroup of the
underlying additive group of H and that S is closed under multiplication. Therefore, S is
indeed a subring of H.
Dene a map : C S as follows. For all a, b R, dene
(a + bi) = a + bi + 0j + 0k .
Suppose that a, b, a
, b
R. Let = a + bi,
= a
+ b
i.Thus,
+
= (a + a
) + (b + b
)i,
= (aa
bb
) + (ab
+ a
b)i .
Using the above calculations, we see that
(+
) = (a+a
)+(b+b
)i+0j+0k = (a+bi+0j+0k)+(a
+b
i+0j+0k) = ()+(
)
and
(
) = (aa
bb
)+(ab
+a
b)i+0j+0k = (a+bi+0j+0k)(a
+b
i+0j+0k) = ()(
)
Note also that is a bijection from C to S. Therefore, is an isomorphism of the ring C to
the subring S of H.
Finally, if a, b R and b = 0, then a + bi + 0j + 0k is in S, but not in the center of H
(which we determined previously). Therefore, S is not contained in the center of H.
Page 231, problem 17. This problem concerns the direct product RS of two rings R
and S. As a set, R S = { (r, s) | r R, s S }. We dene addition and multiplication
in R S as follows. If (r, s) and (r
, s
, s
) = (r + r
, s + s
) , (r, s) (r
, s
) = (r r
, s s
) .
Under addition, R S is the direct product of the underlying additive groups of R and S.
Thus, RS is an abelian group under the above dened addition operation. Since r r
R
and s s
S, we do have (r, s) (r
, s
, s
)
_
= (u, v) (r r
, s s
) =
_
u (r r
), v (s s
)
_
=
_
(u r) r
, (v s) s
_
= (u r, v s) (r
, s
) =
_
(u, v) (r, s)
_
(r
, s
) .
To verify the left distributive law, note that
(u, v)
_
(r, s) +(r
, s
)
_
= (u, v) (r +r
, s +s
) =
_
u (r +r
), v (s +s
)
_
=
_
u r + u r
, v s + v s
_
= (u r, v s) + (u r
, v s
)
= (u, v) (r, s) + (u, v) (r
, s
) .
The right associative law is veried in a similar way.
Next we consider commutativity of multiplication. As above, suppose that and are
in R S. We can write = (r, s) and = (r
, s
), where r, r
R and s, s
S. Then
= (r r
, s s
) and = (r
r, s
s)
Therefore, = if and only if r r
= r
r and s s
= s
s.
In particular, if R and S are commutative rings, then r r
= r
r for all r, r
R and
s s
= s
R and s, s
S. Let = (r, s)
and = (r
, s
= r
= s
s and
so S is a commutative ring too.
Finally, we consider the existence of a multiplicative identity element. Assume that R
and S are rings with identity. Let 1
R
and 1
S
denote the identity elements of R and S,
respectively. Consider the element (1
R
, 1
S
) in R S. For all r R and s S, we have
(1
R
, 1
S
) (r, s) = (1
R
r, 1
S
s) = (r, s) , (r, s) (1
R
, 1
S
) = (r 1
R
, s 1
S
) = (r, s)
Therefore, R S has a multiplicative identity element, namely the element (1
R
, 1
S
).
Conversely, assume that RS has a multiplicative identity element. Denote that element
by . We can write = (a, b), where a R and b S. Suppose that r R and s S.
Let = (r, s). By assumption, we have = and = . Equivalently, these equations
mean that
(a r, b s) = (r, s) and (r a, s b) = (r, s) .
Therefore, we have a r = r and r a = r for all r R. Hence R is a ring with identity,
namely the element a of R. Furthermore, we have b s = s and s b = s for all s S. Hence
S is a ring with identity, namely the element b of S.
Page 232, problem 24. For D = 3, 6, and 7, the ring in question is Z[
D]. For D = 5,
the ring is Z[(1 +
5)/2]. In each case, we will give a unit which satises the inequality
> 1. The group of units will then contain all powers of . Since > 1, we have
n+1
>
n
for all positive integers n. Therefore, the powers of provide an innite number of units in
the ring in question.
For D = 3, let = 2 +
3. Let
= 2
6. Let
= 5 2
7. Let
= 8 3
5)/2. Let = .
Let
= (1
= 1 and so
=
_
(1 +
5)/2
__
(1
5)/2
_
= (1 5)/4 = 1 .
Hence (
) = 1. Note that
a, d R
_
Now suppose that A is an element of the center of the ring M
2
(R). Then AB = BA for
all B M
2
(R). In particular, we have AE
11
= E
11
A and AE
21
= E
21
A, where
E
21
=
_
0 0
1 0
_
.
As shown above, the fact that AE
11
= E
11
A implies that A has the form
A =
_
a 0
0 d
_
where a, d R. Now we use the fact that AE
21
= E
21
A. We have
AE
21
=
_
a 0
0 d
__
0 0
1 0
_
=
_
0 0
d 0
_
, E
21
A =
_
0 0
1 0
__
a 0
0 d
_
=
_
0 0
a 0
_
We have AE
21
= E
21
A if and only if a = d. Thus,
A =
_
a 0
0 a
_
= aI
2
,
where I
2
=
_
1 0
0 1
_
, a scalar multiple of the identity matrix I
2
. Note that I
2
is the multi-
plicative identity element in the ring M
2
(R). It is obvious that matrices of the form aI
2
do
indeed commute with all elements of M
2
(R). Thus,
{A M
2
(R) | AB = BA for all B M
2
(R) } = {aI
2
| a R }
That is, the center of the ring M
2
(R) is the subring {aI
2
| a R }.
Additional Problem D. We rst prove that the subset
S =
_ _
a b
b a
_
a, b R
_
.
is a subring of M
2
(R). We will then show that S
= C.
The additive identity element of M
2
(R) is
_
0 0
0 0
_
and this is clearly in S. For every
element A =
_
a b
b a
_
in S, its additive inverse is
A =
_
a b
(b) a
_
,
which is also in S. Furthermore, suppose that A
=
_
a
_
, where a
, b
R. Then
A + A
=
_
a b
b a
_
+
_
a
_
=
_
(a + a
b + b
(b + b
) (a + a
)
_
,
which is in S. We have proved that S is a subgroup of the underlying additive group of
M
2
(R).
To complete the verication that S is a subring of M
2
(R), it suces to show that S is
closed under the multiplication operation in M
2
(R). Let A and A
be as in the previous
paragraph. Then
AA
=
_
a b
b a
_ _
a
_
=
_
aa
bb
ab
+ ba
ba
+ a(b
) bb
+ aa
_
=
_
aa
bb
ab
+ ba
(ab
+ ba
) aa
bb
_
,
which is indeed in the subset S. We have proved that S is a subring of M
2
(R).
Now dene a map from C to S as follows.: For all a, b R, dene
(a + bi) =
_
a b
b a
_
.
The map is clearly a bijection from C to S. We will prove that is a ring homomorphism
and therefore that the subring S of M
2
(R) is isomorphic to C.
Consider z = a + bi, w = c + di C. We have
z + w = (a + c) + (b + d)i, zw = (ac bd) + (ad + bc)i
and so
(z + w) =
_
a + c b + d
(b + d) a + c
_
=
_
a b
b a
_
+
_
c d
d c
_
= (z) + (w)
and
(z)(w) =
_
a b
b a
__
c d
d c
_
=
_
ac bd ad + bc
bc ad bd + ac
_
=
_
ac bd ad + bc
(ad + bc) ac bd
_
= (zw) ,
showing that is indeed a ring homomorphism. Since is also a bijection, is an isomor-
phism of the ring C to the ring S.
Additional Problem E. Suppose that F is any eld. Suppose that we dene R and its
operations just as stated in the problem. The additive identity 0
R
of R is 0
F
+ 0
F
i. The
multiplicative identity 1
R
of R is 1
F
+0
F
i. Also, it will be convenient to identify an element
a F with the element a + 0
F
i in R. Thus, with this identication, we can regard F as a
subring of R.
We distinguish two cases.
Case 1: The equation a
2
+b
2
= 0
F
has a solution where a and b are nonzero elements of F.
In this case, consider = a + bi, which is a nonzero element of R. Let = a + (b)i,
which we write more simply as a bi. Then is also a nonzero element of R. Furthermore,
we have
= (a + bi)(a bi) = (a
2
+ b
2
) + 0
F
i = 0
F
+ 0
F
i = 0
R
Hence R has zero-divisors and so R is not an integral domain. We proved in class that every
eld is an integral domain. Hence R is not a eld.
Case 2: The only solution to the equation a
2
+ b
2
= 0
F
, where a, b F, is given by
a = b = 0
F
.
In this case, we will prove that R is a eld. First of all, note that R is a commutative
ring with identity 1
R
(as specied above). Also, 1
R
= 0
R
. Now suppose that = a + bi is
any nonzero element of R. This means that a and b are not both zero. Thus, a
2
+ b
2
= 0
F
.
Let c = a
2
+ b
2
, which is a nonzero element of F. Hence c is a unit in F. Hence, c has an
inverse in F under multiplication, which we write as c
1
. As stated above, we can regard
c
1
as an element of R.
Let = a bi, which is an element of R. Furthermore, we have
= (a
2
+ b
2
) + 0
F
i = (c + 0
F
i)(1
F
+ 0
F
i) = (c + 0
F
i)1
R
= c1
R
.
Multiplying both sides of the equation by c
1
, we obtain
(c
1
) = 1
R
.
Since c
1
is an element of R, it follows that is a unit in R. We have proved that every
nonzero element of R is a unit of R. Since R is a commutative ring with identity and 1
R
= 0
R
,
it follows that R is a eld.
Now consider F = Z/3Z. There are only two nonzero elements in F, namely 1 +3Z and
2 +3Z. Consider a
2
+b
2
, where a and b are nonzero elements of F. Then, checking the four
possibilities, we see that a
2
+ b
2
is always nonzero. Thus, the only solution to a
2
+ b
2
= 0
F
is a = b = 0
F
. Therefore, we are in case 2. Therefore, R is a eld.
Now consider F = Z/5Z. Let a = 1 + 5Z and b = 2 + 5Z, two nonzero elements of F.
We have a
2
+ b
2
= 5 + 5Z = 0 + 5Z = 0
F
. Hence we are in case 1. Hence R is not a eld.