Você está na página 1de 6

Hungerford: Algebra

III.3. Factorization in Commutative Rings

1. A nonzero ideal in a principal ideal domain is maximal if and only if it is prime.


Proof: Let R be a PID and let I be a proper ideal in R. Then R has an identity 1 6= 0. Thus if I is
a maximal ideal, then R/I is a field (by Theorem III-2.20: For a commutative ring R with identity,
an ideal I is maximal if and only if R/I is a field). Since a field is always an integral domain, and
so I must be prime (by Theorem III-2.16: For a commutative ring R with identity, an ideal I is
prime if and only if R/I is an integral domain). Note that in this direction, we only need that R is
a commutative ring with identity.
Now suppose that I is a prime ideal. Since R is a PID, we may assume that I = (a) for some
a R. If I is not maximal. then b R such that (a) (b) 6= R and (a) 6= (b). Thus a (b) and so
for some r R, a = br. Therefore, br = a (a). Since (a) is prime, we either have r (a) or b (a)
(by Theorem III-2.15: In a commutative ring R, a proper ideal P is prime if and only if a, b R,
that ab P implies either a P or b P ). If b (a), then (a) = (b), contrary to the assumption
that (a) 6= (b). Thus r (a), and so for some element s R, r = sa. It follows that a = br = bsa.
Since R is a PID, R has no zero divisors and so 1 = bs. This implies that 1 (b) and so (b) = R,
contrary to the assumption that (b) 6= R.

3. Let R = {a + b 10 : a, b Z} be a subring of the field of the reals.

(a) The map N : R 7 Z given by N (a + b 10) = a2 10b2 is multiplicative (that is, u, v R,


N (uv) = N (u)N (v)) and N (u) = 0 if and only if u = 0.
(b) u is a unit if and only if N (u) = 1.

(c) 2, 3, 4 + 10 and 4 10 are irreducibles in R.

(d) 2, 3, 4 + 10 and 4 10 are not primes of R.

Proof: (a) Let u = a + b 10 and v = c + d 10. Then


N (uv)

= N ((a + b 10)(c + d 10) = N (ac + 10bd + (ad + bc) 10)


=

(ac + 10bd)2 10(ad + bc)2 = a2 c2 + 20abcd + 100b2 d2 (10a2 d2 + 20abcd + 10b2 c2 )

= a2 c2 + 100b2 d2 10a2 d2 10b2 c2


= a2 (c2 10d2 ) 10b2 (c2 10d2 ) = (a2 10b2 )(c2 10d2 ) = N (u)N (v).

Note that if for some positive d, a = da1 and b = db1 , then N (u) = N (a + b 10) = a2 10b2 =

d2 (N (a1 + b1 10). Therefore, if 0 = N (u) = a2 10b2 , for some non zero a and b, then we may
assume that (a, b) = 1, (that is, a and b are relatively prime).

Suppose that for some non zero a and b such that N (a + b 10) = 0, and such that (a, b) = 1.
Now a2 = 10b2 . Then 2|a2 implies that 2|a. Therefore, we can write a = 2a1 . Thus 4a21 = 10b2 , and
so 2a21 = 5b2 . Then 2|b and so 2 is a common factor of a and b, contrary to the assumption that

(a, b) = 1. Therefore, we must have a = b = 0.

(b) Let u = a + b 10. Suppose that u is a unit. Then there must be a v R, such that uv = 1.
Since N is multiplicative, N (u)N (v) = N (1) = 1. Since N (u) and N (v) are both integers, we must
have N (u) = 1.

Conversely, we assume that N (u) = 1. Then u(a b 10 = a2 10b2 = 1. Therefore,

(a b| 10) is the inverse of u.


(c) To prove the statements in (c), we first note that the following fact holds in number theory.
(3c-1) For any n Z, both n2 6 2 (mod 5) and n2 6 2 (mod 5).
In fact, we can write n = 5k + r with 0 r 4. Then n2 r2 (mod 5). But 02 = 0, 22 32 4
(mod 5) and 12 42 1 (mod 5). This proves (3c-1).
Suppose that u 6= 0 is not an irreducible element. The for some non unit nonzero x and y R,
u = xy. It follows that N (u) = N (x)N (y), where both N (x) and N (y) are integers other than 1.
Now consider u = 2. Note that N (2) = 4, and that 4 = (2)(2) = (2)(2) are the only
factorizations. If u = xy, for some non units x and y, then we must have N (x) = N (y) = 2.

Let u = a + b 10. Then 2 = N (u) = a2 10b2 . Then we have 2 a2 10b2 = a2 (mod 5),
contrary to (3c-1). Therefore, 2 must be irreducible. Similarly, 3 must be irreducible.

Now suppose that u = 4 + 10 = xy is a product of two non unit nonzero elements. Then
6 = N (u) = N (x)N (y). Therefore, N (x) {2, 3}. This contradicts to (3c-1), as shown above.

Therefore, 4 + 10 must be irreducible. Similarly, 4 10 is also irreducible.

(d) As (2)(3) = 6 = (4 + 10)(4 10), we have 2|(4 + 10)(4 10) but neither 2 = 4 + 10 nor

2 = 4 10. Therefore, 2 is not a prime in R. Similarly, the other three elements are not primes in R.

4. Show that in the integral domain of Exercise 3, every element can be factored into a a product
of irreducibles, but this factorization need not unique.
Proof: That the factorization need not be unique has been shown in (d) of Problem III-3.3.
For any nonzero nonunit element x R, we argue by induction on |N (x)| to show that x can be
factored into a product of finitely many irreducible elements. If |N (x)| is a prime number, then by
Problem III-3.3(a), x itself is irreducible.
Suppose that x is not irreducible. Then there exist non unit elements y and z such that x = yz.
By Problem III-3.3, N (x) = N (y)N (z), and both |N (y)| < |N (x)| and |N (z)| < |N (x)|. Therefore,
by induction, each of y and z can be factored into a product of finitely many irreducible elements,
and so this holds also for x.
5. Let R be a PID.
(a) Every proper ideal is a product P1 P2 Pn of maximal ideals, which are uniquely determined
2

up to order.
(b) An ideal P in R is said to be primary if ab P and a 6 P imply bn P for some n. Show that
P is primary if and only if for some n, P = (pn ), where p R is prime (- irreducible) or p = 0.
(c) If P1 , P2 , , Pn are primary ideals such that Pi = (pini ) and the pi are distinct primes, then
P1 P2 Pn = P1 P2 Pn .
(d) Every proper ideal in R can be expressed (uniquely up to order) as the intersection of a finite
number of primary ideals.
Proof: (a) Let I be an ideal. Since R is a PID, a R such that I = (a). Since every PID is also
a UFD, a = p1 p2 pn is a product of irreducible elements in R, and this factorization is unique up
to permutation of the factors.
We have these observations: In a PID, an element is irreducible if and only if it is a prime
(Theorem III-3.4); in a PID, an element p is prime if and only if (p) is prime(Theorem III-3.4); in a
PID, a proper ideal is prime if and only if it is maximal (Exercise III-3.1).
Let Pi = (pi ), i = 1, 2, , n. Since pi is irreducible, Pi is prime and so is maximal, by the
observations above.
Note that
I = (a) = {ra : r R} = {rp1 p2 pn : r R} P1 P2 Pn .
Conversely,
P1 P2 Pn = {(r1 p1 ) (rn pn ) : r1 , , rn R} {rp1 p2 pn : r R} = (a) = I.
Hence I = P1 P2 Pn is a product of maximal ideals of R.
Now we assume that (a) = Q1 Q2 Qm is a product of maximal ideals of R. Since R is a PID,
qi R such that Qi = (qi ). By the observation made earlier, Qi is prime, and so qi is rime, and so
qi is irreducible.
The fact that (a) = Q1 Q2 Qm implies that a is associate with q1 q2 qm , and so we may
assume that a = q1 q2 qm . Since R is a UFD, we must have m = n and relabelling the qi s if
necessary, we may assume that pi = qi for i = 1, 2, ..., n. Therefore, Pi = Qi , i.
(b) (Sufficiency) Suppose that for some prime element p and integer n > 0, P = (pn ). Let a, b R
be such that ab P . Then pn |ab, and so for some r R, pn r = ab.
Suppose further that a 6 P . Then pn 6 |a. We want to show that p|b. Since R is a UFD, factor
a = p1 p2 pl and b = q1 q2 qm into irreducibles. Then by
pn r = ab = p1 p2 pl q1 q2 qm ,
and since pn 6 |a, we must have p|b. Therefore, pn |bn , implying b = (pn ) = P .
When p = 0, and P = (pn ) = 0. Then since R is an integral domain, ab = 0 and a 6= 0 would
imply that b = 0.
(Necessity) Suppose that a proper ideal P satisfies the hypothesis. Since R is a PID, r R such
that P = (r). If r is irreducible, then the necessity is proved with n = 1. Suppose that r is not

an irreducible element. Then as R is a UFD, there exist positive integers m, n1 , n2 , , nm and


irreducible elements p1 , p2 , , pm , such that i 6= j, pi and pj are not associates and such that
r = pn1 1 pn2 2 pnmm .
Let a = pn1 1 and b = pn2 2 pnmm . Then ab P . Since P is primary, either a P or for some
0
integer n > 0, bn P .
If p1n1 = a P = (r), then for some s R,
pn1 1 = sr = spn1 1 pn2 2 pnmm and so 1 = spn2 2 pnmm ,
which implies that p2 is a unit, contrary to the assumption that p2 is an irreducible. Hence we must
have in integer n > 0, such that (pn2 2 pnmm )n = bn P . It follows that s R such that
(pn2 2 pnmm )n = sr = spn1 1 pn2 2 pnmm .
Thus p1 |(pn2 2 pnmm )n . Since R is a PID, in which an element is irreducible if and only if it is also
a prime, we must have p1 |pj , for some 2 j m. Since pj is an irreducible, we conclude that p1
and pj are associates, contrary to the assumption that the pi s are not associates.
(c) Suppose that, for 1 i n, Pi = (pni i ), where ni > 0 is a positive integer, each pi R is
an irreducible element, and all the the pi s are distinct primes (non associates) in R. Since R is a
PID, for any a R, (a) = {ra : r R}, and so
P1 P2 Pn = {rpn1 1 pn2 2 pnnn : r R} P1 P2 Pn .
Conversely, suppose that a P1 P2 Pn is an arbitrary element. Then for each i, a Pi =
{ri pni i : ri R}. Thus for each i 2,
r1 pn1 1 = a = ri pni i .
Since p1 and pi are not associates, and since pi is a prime, we have pni i |r1 . Since the pi s are distinct
primes (non associates), we conclude that
r1 = rpn2 2 pn3 3 pnnn .
It follows that a = r1 = rpn1 1 pn2 2 pnnn P1 P2 Pn .
(d) Let I be a proper ideal of R. Since R is a PID, a R such that I = (a). Since a PID is
also a UFD, there exist positive integers m, n1 , n2 , , nm and irreducible elements p1 , p2 , , pm ,
such that i 6= j, pi and pj are not associates and such that
a = pn1 1 pn2 2 pnmm .
Let Pi = (pni i ). where 1 i m. Since R is a PID and by (c),
I = (a) = {ra : r R} = {rpn1 1 pn2 2 pnmm : r R} = P1 P2 Pm = P1 P2 Pm .
4

6. (a) If a and n are integers, n > 0, then integers q and r such that a = qn + r, where |r| n/2.
(b) The Gaussian integers Z(i) form a Euclidean domain with (a + bi) = a2 + b2 .
Proof: (a) By division algorithm, we can find integers q 0 and r0 such that a = q 0 n + r0 , where
0 r0 < n. If r0 n/2, then q = q 0 and r = r0 .
Assume that r0 > n/2. Then n r0 n/2. Write a = q 0 n + r0 = (q 0 + 1)n + (r0 n). Then
q = q 0 + 1 and r = r0 n satisfy the conclusion of (a).
(b) Define : Z(i) {0} 7 Z as in (b). We shall show the following.
(6b-1) y = a + bi, x = c + di Z(i) with xy 6= 0, (x) (xy).
In fact, xy = (ac bd) + (ad + bc)i,
(xy) = (ac bd)2 + (ad + bc)2 = a2 c2 + b2 d2 + a2 d2 + b2 c2 = (a2 + b2 )(c2 + d2 ) = (y)(x).
As y 6= 0, (y) 1, and so (xy) (x).
(6b-2) If y 6= 0, then there exist q, r Z(i) such that either r = 0 or r 6= 0 and (r) < (y).
Write y = a + bi and assume first that x Z and x > 0. By (a), we can find integers q1 , q2 , r1 , r2
such that a = q1 x + r1 , b = q2 x + r2 with |r1 | x/2 and |r2 | x/2. Let q = q1 + q2 i and r = r1 + r2 i.
Then y = qx + r. If r = 0, we are done. If r 6= 0, then (r) = r12 + r22 x2 /2 < x2 = (x). Thus
(6b-2) holds in this case.
In the general case (that is, x = c+di 6= 0), let x = cdi. Then xx = c2 +d2 > 0. Replacing x by
xx in the paragraph above, we can find q, r0 Z(i) such that yx = q(xx) + r0 , with r0 = 0 or r0 6= 0
but (r0 ) < (xx). Set r = y qx. Then rx = q(xx) + r0 q(xx) = r0 . By (a), (r) (r0 ) < (x).
7. What are the units in the ring of Z[i]?
Solution: We use the notation in Problem III-3.6(b) above and let be the function in Problem
III-3.6(b).
(7-1) Imitate the proof in Problem III-3.3(a), we can show
(uv) = (u)(v), u, v Z[i].
Let u = a + bi and v = c + di. Then
(uv)

= ((a + bi)(c + di)) = (ac bd + (ad + bc)i)


=

(ac bd)2 + (ad + bc)2 = a2 c2 2abcd + b2 d2 + (a2 d2 + 2abcd + b2 c2 )

= a2 c2 + b2 d2 + a2 d2 + b2 c2
= a2 (c2 + d2 ) + b2 (c2 + d2 ) = (a2 + b2 )(c2 + d2 ) = (u)(v).
Let U (Z[i]) denote the set of all the units in Z[i]. Let a U (Z[i]). Then b U (Z[i]), such that
ab = 1. Thus by (7-1),
(a)(b) = (ab) = (1) = 1.

Since (a) and (b) are positive integers, we conclude that if a U (Z[i]), then (a) = 1. Conversely,
assume that a = x + yi Z[i] such that (a) = 1. Then a(x yi) = (x + yi)(x yi) = (a) = 1,
and so a U (Z[i]). Since x2 + y 2 = 1 if and only is either x = 1 and y = 0 or y = 1 and x = 0.
Thus we have shown that
U (Z[i]) = {1, 1, i, i}.

Você também pode gostar