Math 103B HW 8 Solutions to Selected Problems

15.8. Prove that every ring homomorphism φ from Zn to itself has the form
φ(x) = axwhere a2 = a.
Solution: Let φ : Zn → Zn be a ring homomorphism. Set a = φ(1). Then for any m ∈ Zn ,
we have (using the fact that φ(x + y) = φ(x) + φ(y) for any x, y ∈ Zn )
φ(m) = φ(m · 1) = m · φ(1) = m · a.

Now, using the fact that φ(xy) = φ(x)φ(y) for any x, y ∈ Zn , we see that
a = φ(1) = φ(1 · 1) = φ(1)φ(1) = a2 .

Thus, we have shown that φ(x) = ax for all x ∈ Zn , with a2 = a.

√ √
15.14. Let Z[ 2] = {a + b 2|a, b ∈ Z} and
  
a 2b
H= a, b ∈ Z .
b a
√
Show that Z[ 2] and H are isomorphic as rings.
Solution: (As an exercise,
√ you may want to show that H is a subring of M2 (Z)). Dene a
function φ : H → Z[ 2] by
√
 
a 2b
φ = a + b 2.
b a
   
a 2b c 2d
Then, for , ∈ H , we have
b a d c

√
     
a 2b c 2d a + c 2b + 2d
φ + =φ = (a + c) + (b + d) 2
b a d c b+d a+c
√ √
= (a + b 2) + (c + d 2)
   
a 2b c 2d
=φ +φ .
b a d c

1
and
√
    
a 2b c 2d ac + 2bd 2ad + 2bc
φ =φ = (ac + 2bd) + (ad + bc) 2
b a d c ad + bc ac + 2bd
√ √
= (a + b 2)(c + d 2)
   
a 2b c 2d
=φ φ .
b a d c
√
Hence, φ is a ring homomorphism. Now, φis surjective since, given
 anyx ∈ Z[ 2], we have
√ √

a 2b a 2b
that x = a + b 2 for some a, b ∈ Z, so ∈ H and φ = a + b 2 = x.
b a   b a  
a 2b c 2d a 2b
Finally, to see that φ is injective, suppose that , ∈ H and φ =
b a d c b a
√ √
 
c 2d
φ . This means that a + b 2 = c + d 2. If b 6= d (so that d − b 6= 0), then this
d c
√
would imply that 2 = a−c d−b
is rational, a contradiction. Hence, we must have that b = d.
Hence, we have √ √ √
a + b 2 = c + d 2 = c + b 2.
√
 
a 2b
Subtracting b 2 from both sides of the above equation shows that a = c. Thus, =
  b a
c 2d
, so it follows that φ is injective. Hence, φ is a ring isomorphism.
d c

15.29. Determine all ring homomorphisms from Z × Z to Z.

Solution: We claim that the only ring homomorphisms from Z × Z to Z are the functions
φ0 , φ1 , φ2 : Z × Z → Z dened by

φ0 (m, n) = 0
φ1 (m, n) = m
φ2 (m, n) = n.

(as an exercise, you can show that these functions are indeed ring homomorphisms). To see
this, suppose φ : Z × Z → Z is a ring homomorphism. Note that for a ∈ {(0, 1), (1, 0), (1, 1)}
we have that a2 = a, which implies that (φ(a))2 = φ(a) since φ is a ring homomorphism.
The only n ∈ Z such that n2 = n are 0 and 1. This implies that φ(a) ∈ {0, 1} for each
a ∈ {(0, 1), (1, 0), (1, 1)}. We consider the following cases:
Case 1 - φ(1, 0) = 1 and φ(0, 1) = 0. Then for any (m, n) ∈ Z × Z we have
φ(m, n) = φ(m · (1, 0) + n · (0, 1)) = m · φ(1, 0) + n · φ(0, 1) = m · 1 + n · 0 = m.

Hence, in this case φ = φ1 .

Case 2 - φ(1, 0) = 0 and φ(0, 1) = 1. An argument similar to the one in case 1 shows that
in this case φ = φ2 .
Case 3 - φ(1, 0) = 0 and φ(0, 1) = 0. An argument similar to the one in case 1 shows that

Page 2
in this case φ is the trivial homomorphism φ0 .
Case 4 - φ(1, 0) = 1 and φ(0, 1) = 1. In this case, we have
φ(1, 1) = φ((1, 0) + (0, 1)) = φ(1, 0) + φ(0, 1) = 1 + 1 = 2,

contradicting our observation above that φ(1, 1) ∈ {0, 1}. Hence, this case cannot happen.

By the above work, we see that φ0 , φ1 , φ2 (as dened above) are the only ring homomor-
phisms from Z × Z to Z.
15.37. For any integer n > 1, prove that Zn [x]/hxi is isomorphic to Zn .
Solution: Let n be an integer which is greater than 1. Dene a function φ : Zn [x] → Zn by
φ(an xn + an−1 xn−1 + · · · + a0 ) = a0 .

As an exercise, check that φ is a ring homomorphism. Now, we claim that ker φ = hxi. To
see this, rst suppose that f ∈ hxi. Then, by denition of hxi, we have that f = xg for some
g ∈ Zn [x]. Then φ(f ) = φ(x)φ(g) = 0 · φ(g) = 0, so f ∈ ker φ. Hence, hxi ⊆ ker φ. Now,
suppose that an xn + an−1 xn−1 + · · · + a0 ∈ ker φ. Then
0 = φ(an xn + an−1 xn−1 + · · · + a0 ) = a0

so that
an xn +an−1 xn−1 +· · ·+a0 = an xn +an−1 xn−1 +· · ·+a1 x = (an xn−1 +an−1 xn−2 +· · ·+a1 )x ∈ hxi.

Hence, ker φ ⊆ hxi, so it follows that ker φ = hxi.

Now, clearly φ is surjective since for any a ∈ Zn we have that a ∈ Zn [x] as well and
φ(a) = a. Hence, by the First Isomorphism Theorem, we have Zn [x]/hxi = Zn [x]/ ker φ ∼=
Im(φ) = Zn .

15.38. For any integer n > 1, prove that hxi is a maximal ideal of Zn [x] if and
only if n is prime.
Solution: Let n be an integer greater than 1. We have
hxi is maximal ⇔ Zn [x]/hxi is a eld by Theorem 14.4 (since Zn [x] is a commutative ring with unity)
⇔ Zn is a eld by problem 37
⇔ n is prime

15.44. Let R be a commutative ring of prime characteristic p. Show that the

Frobenius map x 7→ xp is a ring homomorphism from R to R.

Page 3
Solution: Let φ : R → R denote the Frobenius map, so φ(x) = xp for all x ∈ R. Let
x, y ∈ R. By problem 13.49, we know that φ(x + y) = (x + y)p = xp + y p = φ(x) + φ(y). Also
φ(xy) = (xy)p = xp y p = φ(x)φ(y) since R is commutative. Hence, φ is a ring homomorphism.

15.46. Show that a homomorphism from a eld onto a ring with more than one
element must be an isomorphism.
Solution: Let F be a eld, R a ring with more than one element, and φ : F → R a surjective
homomorphism. We will show that this implies that φ is injective. We know that ker φ is
an ideal of F , and we know that the only ideals of F are {0F } and F (any ideal which is not
equal to {0F } contains a unit and must therefore be equal to F ). Since R has more than
one element, we know there exists r ∈ R \ {0R }. Since φ is surjective, we know that there
exists a ∈ F such that φ(a) = r 6= 0R . Hence, a ∈/ ker φ, so we must have ker φ 6= F . Hence
ker φ = {0F }, i.e. φ is injective. Thus, φ is an isomorphism.

15.54. Suppose that n divides m and that a is an idempotent of Zn (that is,

a2 = a). Show that the mapping x 7→ ax is a ring homomorphism from Zm to Zn .
Show that the same correspondence need not yield a ring homomorphism if n
does not divide m.
Solution: Recall that Zn ∼
= Z/nZ and Zm ∼
= Z/mZ as rings. We will translate the problem
statement in terms of these two rings since the statment will be much more convenient
to prove in this setting. The problem then becomes: suppose that n divides m and that
(a + nZ) is an idempotent of Z/nZ (that is, (a + nZ)2 = (a + nZ)). Show that the map
φa : Z/mZ → Z/nZ dened by φa (x + mZ) = (a + nZ)(x + nZ) = (ax + nZ) is a ring
homomorphism.
We rst need to show that φa is well-dened. Suppose that x1 , x2 ∈ Z and (x1 + mZ) =
(x2 + mZ), which means that x1 − x2 = mk for some k ∈ Z. Since n divides m, we know
that m = n` for some ` ∈ Z. Hence, we see that (ax1 + nZ) = (ax2 + nZ) since
ax1 − ax2 = a(x1 − x2 ) = a(mk) = a(n`)k ∈ nZ.

This shows that φa is well-dened.

Now, let (y1 + mZ), (y2 + mZ) ∈ Z/mZ. Then we have
φa ((y1 + mZ) + (y2 + mZ)) = φa ((y1 + y2 ) + mZ) = (a(y1 + y2 ) + nZ) = (ay1 + nZ) + (ay2 + nZ)
= φa (y1 + mZ) + φa (y2 + mZ)

and
φa ((y1 + mZ)(y2 + mZ)) = φa (y1 y2 + mZ)
= (ay1 y2 + nZ)
= (a + nZ)(y1 + nZ)(y2 + nZ)
= (a + nZ)2 (y1 + nZ)(y2 + nZ) since (a + nZ)2 = (a + nZ)
= φa (y1 + mZ)φa (y2 + mZ).

Page 4
Hence, φa is a ring homomorphism.
To see that the above process does not necessarily yield a ring homomorphism if n does
not divide m, consider the case where m = 3 and n = 2. Clearly 1 ∈ Z2 is an idempotent,
but the map φ : Z3 → Z2 dened by φ(x) = 1 · x is not a homomorphism because
φ(1 +3 2) = φ(0) = 0

but
φ(1) +2 φ(2) = 1 +2 0 = 1.
(in the context of the rings Z/3Z and Z/2Z, the above function φ becomes a map φ0 : Z/3Z →
Z/2Z given by φ0 (x + 3Z) = x + 2Z. The fact that φ is not a homomorphism translates to
the map φ0 not being well-dened: we have that 0 + 3Z = 3 + 3Z but 0 + 2Z 6= 3 + 2Z (so
φ0 is trying to map 0 + 3Z to two dierent elements of Z/2Z).)

15.67. Show that the prime subeld of a eld of characteristic pis ring-isomorphic
to Zp and the prime subeld of a eld of characteristic 0 is ring-isomorphic to Q.
Solution: First, let F be a eld of characteristic p. Let K be the prime subeld of F . By
Corollary 3 to Theorem 15.5, we know that F has a subeld L such that L is isomorphic to
Zp . By denition of the prime subeld, we have that K ⊆ L. On the other hand, Zp has no
subelds other than Zp itself : Any subeld of Zp must contain 1, and must therefore also
contain 1 + 1 = 2, 1 + 1 + 1 = 3, . . . , (p − 1) · 1 = p − 1, p · 1 = 0, which exhausts all of Zp .
Hence, L has no subelds other than L itself, which, since K is a subeld of L, implies that
L = K . Hence, the prime subeld of F is isomorphic to Zp in this case.
Now, suppose F is a eld of characteristic 0. Again let K denote the prime subeld of F .
By Corollary 3 to Theorem 15.5, we know that F has a subeld L such that L is isomorphic
to Q. By denition of the prime subeld, K ⊆ L. Following our argument above, we will
show that Q has no subelds other than Q itself, which will imply that L = K . So suppose
that k is a subeld of Q. Then 1 ∈ k. Since k must be closed under addition, we get that
Z+ ⊆ k . Since k must contain additive inverses of all its elements, we see that Z ⊆ k . Since
k must contain multiplicative inverses of all its nonzero elements, we see that n1 ∈ k for all
n ∈ Z \ {0}. Finally, since k is closed under multiplication, this implies that Q ⊆ k . Hence,
Q has no subelds other than Q itself, which implies that L has no subelds other than L
itself, which implies that L = K . Hence, the prime subeld of F is isomorphic to Q.

Page 5