PROCEEDINGS OF THE

AMERICAN MATHEMATICAL SOCIETY

Volume 125, Number 1, January 1997, Pages 67–74
S 0002-9939(97)03663-0

PRIME IDEALS IN POLYNOMIAL RINGS

IN SEVERAL INDETERMINATES

MIGUEL FERRERO

(Communicated by Ken Goodearl)

Abstract. If P is a prime ideal of a polynomial ring K[x], where K is a field,

then P is determined by an irreducible polynomial in K[x]. The purpose of this
paper is to show that any prime ideal of a polynomial ring in n-indeterminates
over a not necessarily commutative ring R is determined by its intersection
with R plus n polynomials.

Introduction
Let K be a field and K[x] the polynomial ring over K in an indeterminate x. If
P is a prime ideal of K[x], then there exists an irreducible polynomial f in K[x]
such that P = K[x]f . This result is quite old and basic; however no corresponding
result seems to be known for a polynomial ring in n indeterminates x1 , ..., xn over
K. Actually, it seems to be very difficult to find some system of generators for a
prime ideal of K[x1 , ..., xn ].
Now, K[x1 , ..., xn ] is a Noetherian ring and by a converse of the principal ideal
theorem for every prime ideal P of K[x1 , ..., xn ] there exist n polynomials f1 , ..., fn
such that P is minimal over (f1 , ..., fn ), the ideal generated by {f1 , ..., fn } ([4],
Theorem 153). Also, as a consequence of ([1], Theorem 1) it follows that any prime
ideal of K[x1 , ..., xn ] is determined by n polynomials. However it is not clear in [1]
how to find these polynomials and no converse result is proved. The purpose of
this paper is to show a result which in particular implies that every prime ideal of
K[x1 , ..., xn ] is determined by a sequence of n polynomials which is in some sense
irreducible, and such that the converse is also true.
In general, let R be any prime ring. If P is a nonzero prime ideal of the polyno-
mial ring R[x] with P ∩ R = 0 (an R-disjoint prime ideal), then P = Q[x]f0 ∩ R[x]
where Q is a ring of right quotients of R and f0 ∈ C[x] is an irreducible polynomial,
C being the extended centroid of R ([2], Corollary 2.7). This characterization has
the problem that we have to compute Q to have a prime ideal determined. A better
way (an intrinsic one) to determine an R-disjoint prime ideal has been given in [3].
We proved that every R-disjoint prime ideal P of R[x] is determined by just one
polynomial in P which is in some sense irreducible ([3], Theorem 1.4).

Received by the editors March 15, 1995 and, in revised form, July 28, 1995.
1991 Mathematics Subject Classification. Primary 16D30, 16S36; Secondary 12E05.
This research was supported by a grant given by Conselho Nacional de Desenvolvimento
Cientı́fico e Tecnológico (CNPq), Brazil.

c 1997 American Mathematical Society

67
68 MIGUEL FERRERO

The purpose of this paper is to extend the above result. We show that any prime
ideal P in a polynomial ring in n indeterminates is determined by its intersection
with R plus n polynomials in P . Moreover, these polynomials define a sequence
which has some irreducibility property and will be called completely irreducible
(mod P ∩ R). The converse is also true, i.e., for any prime ideal Q of R and
any completely irreducible sequence (mod Q), there exists a unique prime ideal
determined by them. It turns out that there exists a one-to-one correspondence
between prime ideals of R[x1 , ..., xn ] and equivalence classes of tuples (Q, f1 , ..., fn ),
where Q is a prime ideal of R and (f1 , ..., fn ) is a completely irreducible sequence
(mod Q).
Throughout this paper R is any ring with an identity element. If f ∈ R[x], then
δf (resp. lc(f )) denotes the degree (resp. leading coefficient) of f . The leading
coefficient of the zero polynomial will be defined as the identity of R. Finally, the
notation ⊃ means strict inclusion.

1. Prime ideals in R[x]

Throughout this section Q denotes a prime ideal of R. We will describe all
the prime ideals P of R[x] such that P ∩ R = Q. The results can be obtained
by factoring out from R and R[x] the ideals Q and Q[x], respectively, and then
applying the results of [2] and [3]. However we will give here direct arguments for
the sake of completeness.
We define

ΓQ = {f ∈ R[x] : arf − f ra ∈ Q[x], for all r ∈ R, δf 6= 0, and a = lc(f ) 6∈ Q}.

The definition of ΓQ intends to give an extension of the definition of ΓR in [2],

Section 1. Thus they should coincide when Q = 0. However they are slightly differ-
ent. In fact, we exclude here nonzero constant polynomials. Also, the convention
lc(0) = 1 allows us to assume always that lc(f ) 6∈ Q. Note that the zero polynomial
belongs to ΓQ .
For f ∈ ΓQ with a = lc(f ), define

[Q, f ] = {g ∈ R[x] : there exists e ≥ 0 with g(Ra)e ⊆ R[x]f + Q[x]}.

We can easily see that [Q, f ] is an ideal of R[x] containing f and Q[x].
In the rest of the section we denote by a and b the leading coefficients of f and
g, respectively. We begin with the following.

Lemma 1.1 (cf. [2, Lemma 1.1]). Assume that f ∈ ΓQ . If g ∈ [Q, f ] and δg < δf ,
then g ∈ Q[x]. In particular, [Q, f ] ∩ R = Q.

Proof. Note that a 6∈ Q. So by assumption, there exists an ideal H 6⊆ Q of R such

that for every r ∈ H there is k = xm bm + ... + b0 ∈ R[x] with gra = kf + h, where
h ∈ Q[x]. We have bm a ∈ Q and for s ∈ R, grasa = kf sa + hsa = kasf + hsa + l,
for some l ∈ Q[x].
Assume, by induction, that bi a ∈ Q for i = m, ..., m − t + 1. Then from the above
relation we get bm−t asa ∈ Q, for every s ∈ R. It follows that bm−t a ∈ Q because
a 6∈ Q. Hence bi a ∈ Q for i = 1, ..., m, and so gHaRa ⊆ Q[x]. Thus g ∈ Q[x] since
Q[x] is a prime ideal of R[x]. The proof is complete.
 PRIME IDEALS IN POLYNOMIAL RINGS 69

Note that the definition [f ]R in [2] also looks different than our definition here
of [0, f ], for f ∈ Γ0 . However we see they coincide. In fact, put
I = {g ∈ R[x] : there exists H
R, H 6⊆ Q, such that gHa ⊆ R[x]f + Q[x]}.
Lemma 1.2. [Q, f ] = I.
Proof. If f = 0, then [Q, f ] = Q[x] as well as I = Q[x], since Q[x] is a prime ideal.
In general, it is clear that [Q, f ] ⊆ I. Assume that g ∈ I, δg = m. If m < δf = n
we prove as in Lemma 1.1 that g ∈ Q[x] ⊆ [Q, f ]. Suppose δg = n. We have
gra − brf ∈ I, for any r ∈ R, and δ(gra − brf ) < n. Hence gra − brf ∈ Q[x]
and it follows that gRa ⊆ R[x]f + Q[x]. Finally, assume by induction that m > n
and for any h ∈ I with δh < m we have h(Ra)m−n ⊆ R[x]f + Q[x]. For r ∈ R
put l = gra − xm−n rbf ∈ I. Applying the induction assumption to l we easily get
g(Ra)m−n+1 ⊆ R[x]f + Q[x]. The result follows.

Now we prove the following key result.

Proposition 1.3. Assume that f, g ∈ ΓQ and g ∈ [Q, f ]. Then [Q, g] ⊆ [Q, f ].
Also, [Q, g] = [Q, f ] if and only if δg = δf .
Proof. Take h ∈ [Q, g]. By Lemma 1.2 there exists an ideal H 0 of R with H 0 6⊆ Q
such that hH 0 b ⊆ R[x]g + Q[x]. Also, by assumption there exists H
R with
H 6⊆ Q and gHa ⊆ R[x]f + Q[x]. Thus hH 0 bHa ⊆ R[x]f + Q[x] and it follows that
h ∈ [Q, f ] since H 0 bH 6⊆ Q. The first part follows.
Assume δg = δf . For every r ∈ R, f rb − arg ∈ [Q, f ] and δ(f rb − arg) < n.
Then f rb − arg ∈ Q[x] by Lemma 1.1. Therefore f Rb ⊆ R[x]g + Q[x] and thus
f ∈ [Q, g]. Using the first part we get [Q, g] = [Q, f ].
The converse follows easily from Lemma 1.1, since f, g 6∈ Q[x].

Remark 1.4. We can define [Q, f ] in a dual way. In fact, denote by [Q, f ]0 the ideal
of R[x] defined as [Q, f ], but with the condition (aR)e g ⊆ f R[x] + Q[x] instead
of the condition g(Ra)e ⊆ R[x]f + Q[x]. As in Lemma 1.1 we can prove that if
g ∈ [Q, f ]0 and δg < δf , then g ∈ Q[x]. Actually, it is not hard to show that
[Q, f ] = [Q, f ]0 .
Corollary 1.5 (cf. [2, Corollary 1.3]). Assume that f, g ∈ ΓQ . Then [Q, f ] =
[Q, g] if and only if gra − brf ∈ Q[x] for every r ∈ R.
Proof. If [Q, f ] = [Q, g] we have δf = δg by Proposition 1.3. Hence we easily obtain
gra − brf ∈ Q[x] for all r ∈ R. Conversely, if gra − brf ∈ Q[x], for every r ∈ R, we
have gRa ⊆ R[x]f + Q[x] and so g ∈ [Q, f ]. In particular, δg ≥ δf since g 6∈ Q[x].
Also, bRf ⊆ gR[x] + Q[x] and it follows that f ∈ [Q, g]0 . Hence δf ≥ δg by Remark
1.4. It follows that δf = δg and Proposition 1.3 gives [Q, f ] = [Q, g].

The following is well-known (or can easily be obtained by factoring out Q and
Q[x] from R and R[x], respectively).
Lemma 1.6. Let P be an ideal of R[x]. Then P is prime if and only if P ∩ R is
a prime ideal of R and either P = (P ∩ R)[x] or P is maximal amongst ideals I of
R[x] such that I ∩ R = P ∩ R.
Now we extend the definition of Γ-completely irreducibility (see [3], Definition
1.2).
70 MIGUEL FERRERO

Definition 1.7. We say that a polynomial f ∈ ΓQ is ΓQ -completely irreducible if

the following condition is satisfied:
If there exist b ∈ R, g ∈ ΓQ , and h ∈ R[x] such that f b 6∈ Q[x] and f b−hg ∈ Q[x],
then δg = δf .

Note that the above definition is vacuously true for f = 0. So we may consider
the zero polynomial as a ΓQ -completely irreducible polynomial.

Remark 1.8. Assume that f ∈ ΓQ . Then we can easily see that if b 6∈ Q we have
f b ∈ ΓQ and [Q, f ] = [Q, f b] (use Proposition 1.3).

Now we can prove the main result of this section.

Theorem 1.9 (cf. [3, Theorem 1.4]). Let R be any ring. Then an ideal P of R[x]
is prime if and only if Q = P ∩ R is prime and there exists a ΓQ -completely
irreducible polynomial f ∈ R[x] such that P = [Q, f ]. Moreover, the polynomial f
can be chosen to be any polynomial in P which is of minimal degree with respect to
the condition lc(f ) 6∈ Q.

Proof. If P is a prime ideal, then Q = P ∩ R is prime. In case P = Q[x] we take

f = 0 and so P = [Q, f ] follows. Assume P ⊃ Q[x] and take any polynomial f ∈ P
of minimal degree n with respect to the condition a = lc(f ) 6∈ Q. For any r ∈ R,
f ra − arf ∈ P and δ(f ra − arf ) < n. It follows that f ∈ ΓQ . We show that
P = [Q, f ].
Suppose g ∈ P , δg = m. If m < n we have g ∈ Q[x] by minimality of n.
Repeating the arguments of the proof of Lemma 1.2 we show, by induction, that
g ∈ [Q, f ]. Consequently P ⊆ [Q, f ] and Lemma 1.6 gives P = [Q, f ].
Assume that for some b ∈ R, g ∈ ΓQ , and h ∈ R[x] we have f b 6∈ Q[x] and
f b − hg ∈ Q[x]. By Remark 1.8 and Proposition 1.3 we have P = [Q, f ] = [Q, f b] ⊆
[Q, g]. It follows that [Q, f ] = [Q, g] by maximality of P and hence δf = δg.
Consequently f if ΓQ -completely irreducible.
Conversely, assume that Q is prime and P = [Q, f ] for some f ∈ ΓQ which is
ΓQ -completely irreducible. If f = 0 we clearly have P = Q[x] and so P is prime.
Suppose f 6= 0 and take an ideal M of R[x] which is maximal with respect to the
conditions M ⊇ P and M ∩ R = Q. Then M is prime and it follows from the
first part that M = [Q, g] for some g ∈ ΓQ . Since f ∈ P there exists e ≥ 0 with
f (Rb)e ⊆ R[x]g + Q[x]. Using now the irreducibility property of f we easily obtain
that δf = δg and consequently P = M , by Proposition 1.3. Hence P is a prime
ideal and the proof is complete.

Let P be the set of all the pairs (Q, f ) where Q is a prime ideal of R and f ∈ ΓQ
is a ΓQ -completely irreducible polynomial. We define an equivalence relation by
saying that (Q, f ) ∼ (Q0 , g) if and only if Q = Q0 and f rb − arg ∈ Q[x], for every
r ∈ R, where a = lc(f ) and b = lc(g). By Corollary 1.5, (Q, f ) ∼ (Q0 , g) if and only
if [Q, f ] = [Q0 , g]. We clearly have

Theorem 1.10. Let R be any ring and let P be the set of all the pairs as above.
Then the mapping sending (Q, f ) to [Q, f ] defines a one-to-one correspondence
between the set of all the equivalence classes in P and the set of all the prime ideals
of R[x].
 PRIME IDEALS IN POLYNOMIAL RINGS 71

2. Prime ideals in R[x1 , ..., xn ]

The purpose of this section is to extend the results of the former one. We prove
similar results for a polynomial ring in n-indeterminates. We begin with a technical
lemma.
Let R be a ring, T = R[x1 , ..., xn ] the polynomial ring over R in n-indeterminates
x1 , ..., xn , and S = T [t] the polynomial ring over T in an indeterminate t.
If f ∈ S we consider f as a polynomial in t with coefficients in T and we denote
by δf (resp. a) the degree (resp. leading coefficient) of f with respect to t.
For a polynomial h ∈ T , ∆h denotes the total degree of h as a polynomial in the
indeterminates x1 , ..., xn . Also, for g = am tm + ... + a0 ∈ S, aj ∈ T for 0 ≤ j ≤ m,
∆g is defined as max{∆a0 , ..., ∆am }.
Let P be a prime ideal of S and Q = P ∩ T . Assume that P ⊃ Q[t] and f ∈ P
is a polynomial of minimal degree u with respect to the condition a = lc(f ) 6∈ Q.
Put ∆f = v.
In the next lemma b denotes the leading coefficient of g ∈ S and (f ) the ideal of
S generated by f . We have
Lemma 2.1. Under the same notation as above, given an integer m ≥ u, for every
g ∈ P with δg ≤ m, and for every r0 , r1 , ..., rm−n in R, there exists h ∈ (f ) such
that l = gr0 ar1 a...rm−n a − h ∈ Q[t] and ∆l ≤ w + (m − u + 1)v, where w = ∆g.
Proof. If m = u and δg < u we have g ∈ Q[t]. So it is enough to take h = 0 in this
case. If δg = u, then gra − brf ∈ Q[t], for every r ∈ R, since gra − brf ∈ P and
δ(gra − brf ) < u. Also ∆(gra − brf ) ≤ max{∆(gra), ∆(brf )} ≤ w + v.
By induction, assume that m > u and the result is true for any polynomial
p ∈ P with δp ≤ m − 1. Let g be a polynomial of degree m in P with ∆g =
w. For r0 ∈ R put gr0 = gr0 a − tm−u br0 f ∈ P . Then δgr0 < m and ∆gr0 ≤
max{∆(gr0 a), ∆(tm−u br0 f )} ≤ w + v.
Therefore, for any r1 , ..., rm−n ∈ R there exists h ∈ (f ) such that
l = gr0 r1 a...rm−n a − h ∈ Q[t]
and
∆l ≤ w + v + (m − u)v = w + (m − u + 1)v.
Hence gr0 ar1 a...rm−n a − h0 ∈ Q[t], where h0 = h + tm−u br0 f r1 a...rm−n a ∈ (f ), and
the proof is complete.
For a subset B ⊂ S, δB (resp. ∆B) denotes sup{δh : h ∈ B} (resp. sup{∆h :
h ∈ B}). The following is clear.
Corollary 2.2. Under the same notation as above, if B ⊆ P , δB ≤ m, and ∆B ≤
w, for every h ∈ B(Ra)m−u+1 there exists h0 ∈ (f ) such that h − h0 ∈ Q[t] and
∆(h − h0 ) ≤ w + (m − u + 1)v.
Now we change notations. Let R be any ring and S = R[x1 , ..., xn ] a polynomial
ring over R in n indeterminates x1 , ..., xn .
We fix the natural order in the variables. Denote by Ri the polynomial ring
R[x1 , ..., xi ]. If fi ∈ Ri we consider fi as a polynomial in xi with coefficients in
Ri−1 and denote by δfi (resp. ai ) its degree (resp. leading coefficient) with respect
to xi . Also, ∆i fi denotes the total degree of fi as a polynomial in x1 , ..., xi .
Let Q be a prime ideal of R and put Q[x] = Q[x1 , ..., xn ]. We denote by F
(resp. Fi ) the set of all the sequences (f1 , ..., fn ) (resp. (f1 , ..., fi )), where fj ∈ Rj ,
72 MIGUEL FERRERO

1 ≤ j ≤ n. If f = (f1 , ...) is a sequence in F (resp. Fi ), then (f ) (resp. (f )i )

denotes the ideal of S (resp. Ri ) generated by {f1 , ..., fn } (resp. {f1 , ..., fi }). For
f ∈ F we put
[Q, f1 , ..., fn ] = {g ∈ S : there exist ei ≥ 0, 1 ≤ i ≤ n, such that

g(Ran )en ...(Ra1 )e1 ⊆ Q[x] + (f )}.

Note that if g(Ran )en ...(Ra1 )e1 ⊆ Q[x] + (f ), then g(Ran )dn ...(Ra1 )d1 ⊆ Q[x] +
(f ), for every sequence of integers (dj )1≤j≤n with di ≥ ei for 1 ≤ i ≤ n. Then it
follows easily that [Q, f1 , ..., fn ] is an ideal of S.
Now we can prove the main result of this paper. It shows that every prime ideal
of S is completely determined by its intersection with R plus n polynomials in S.
Theorem 2.3. Let P be a prime ideal of S with P ∩ R = Q. Then there exist n
polynomials f1 , ..., fn in S such that (f1 , ..., fn ) ∈ F and P = [Q, f1 , ..., fn ].
Proof. Put Pi = P ∩ Ri , 1 ≤ i ≤ n. If P1 = Q[x1 ] we take f1 = 0 (so a1 = 1).
If P1 ⊃ Q[x1 ], take a polynomial f1 of minimal degree in P1 with respect to the
condition a1 6∈ Q. It follows that P1 = [Q, f1 ], by Theorem 1.9. Moreover, if
B ⊆ P1 and δB ≤ m for some integer m, then there exists e1 ≥ 0 such that
B(Ra1 )e1 ⊆ Q[x1 ] + (f )1 , by Corollary 2.2 (with n = 0 and t = x1 ).
We continue the construction by induction. Assume there exists a sequence
(f1 , ..., fi−1 ) ∈ Fi−1 such that Pi−1 = [Q, f1 , ..., fi−1 ]. Moreover, if B ⊆ Pi−1 and
∆i−1 B ≤ m for some integer m, there exist e1 , ..., ei−1 ≥ 0 such that
B(Rai−1 )ei−1 ...(Ra1 )e1 ⊆ Q[x1 , ..., xi−1 ] + (f )i−1 .
If Pi = Pi−1 [xi ] we take fi = 0. If Pi ⊃ Pi−1 [xi ] we take a polynomial fi ∈ Pi of
minimal degree with respect to the condition ai 6∈ Pi−1 . Applying again Theorem
1.9 we have Pi = [Pi−1 , fi ]. It remains to show that [Q, f1 , ..., fi ] = Pi and complete
the proof of the induction argument.
If g ∈ [Q, f1 , ..., fi ], then g(Rai )ei ...(Ra1 )e1 ⊆ Q[x1 , ..., xi ] + (f )i ⊆ Pi , for some
integers e1 , ..., ei ≥ 0. Since Pi is prime and aj 6∈ Pi , for 1 ≤ j ≤ i, it follows that
g ∈ Pi .
Conversely, assume that C ⊆ Pi and ∆i C ≤ m. Then we have δC ≤ m and
∆i−1 C ≤ m. By Corollary 2.2 there is an integer l ≥ 0 such that for every
h ∈ C(Rai )m−ui +1 there exists h0 ∈ (fi ) with h−h0 ∈ Pi−1 [xi ] and ∆i−1 (h−h0 ) ≤ l,
where (fi ) denotes the ideal of Ri generated by fi and ui = δfi . By the induction as-
sumption there exist integers e1 , ..., ei−1 ≥ 0 such that (h−h0 )(Rai−1 )ei−1 ...(Ra1 )e1
⊆ Q[x1 , ..., xi ] + (f )i , for every such a diference h − h0 . Since h0 ∈ (f )i we obtain
h(Rai−1 )ei−1 ...(Ra1 )e1 ⊆ Q[x1 , ..., xi ] + (f )i . Hence
C(Rai )m−ui +1 (Rai−1 )ei−1 ...(Ra1 )e1 ⊆ Q[x1 , ..., xi ] + (f )i ,
and the proof is complete.
Definition 2.4. Let f = (f1 , ..., fn ) ∈ F. We say that f is completely irreducible
(mod Q) if fi+1 is ΓPi -completely irreducible as a polynomial in xi+1 , for every
0 ≤ i ≤ n − 1, where P0 = Q and Pi = [Pi−1 , fi ].
Note that if f is completely irreducible (mod Q), then f1 ∈ ΓQ is ΓQ -completely
irreducible as a polynomial in R1 and so P1 = [Q, f1 ] is a prime ideal of R1 . It
follows by induction that Pi is prime, for every i ≤ n.
 PRIME IDEALS IN POLYNOMIAL RINGS 73

By the construction in Theorem 2.3 and the results of Section 1, the sequence
f ∈ F such that P = [Q, f1 , ..., fn ] is completely irreducible (mod Q). Therefore,
for every prime ideal P of S with P ∩ R = Q there exists a sequence in F which
is completely irreducible (mod Q) and which determines P . The converse is also
true.
Theorem 2.5. Let Q be a prime ideal of R and (f1 , ..., fn ) ∈ F a completely
irreducible sequence (mod Q). Then P = [Q, f1 , ..., fn ] is a prime ideal of S such
that P ∩ R = Q.
Proof. By Definition 2.4, Pi = [Pi−1 , fi ] is a prime ideal of Ri with Pi ∩Ri−1 = Pi−1 ,
for 1 ≤ i ≤ n, where R0 = R and P0 = Q. So P = Pn is a prime ideal of S with
P ∩ Ri = Pi , for 1 ≤ i ≤ n. Also, by Lemma 1.1, fi ∈ Pi is either zero or
a polynomial of minimal degree in Pi with respect to the condition ai 6∈ Pi−1 ,
according to Pi = Pi−1 [xi ] or Pi ⊃ Pi−1 [xi ], respectively. It follows as in Theorem
2.3 that P = [Q, f1 , ..., fn ].

Denote by I the set of all the sequences (Q, f1 , ..., fn ), where Q is a prime ideal of
R and (f1 , ..., fn ) ∈ F is a completely irreducible sequence (mod Q). Two sequences
(Q, f1 , ..., fn ) and (L, h1 , ..., hn ) of I are said to be equivalent if and only if Q = L
0
and (Pi−1 , fi ) ∼ (Pi−1 , hi ) in the sense defined in Section 1, for every 1 ≤ i ≤ n,
0 0 0
where P0 = P0 = Q, Pi−1 = [Pi−2 , fi−1 ], and Pi−1 = [Pi−2 , hi−1 ], for i ≥ 2. It is
clear that (Q, f1 , ..., fn ) ∼ (L, h1 , ..., hn ) if and only if [Q, f1 , ..., fn ] = [L, h1 , ..., hn ].
Therefore we have the following.
Theorem 2.6. There is a one-to-one correspondence between the set of all the
prime ideals of S and the set of all the equivalence classes of I module the equiva-
lence relation ∼.
Remark 2.7. Perhaps is useful to point out that if R = K is a field, then for every
prime ideal P of S = K[x1 , ..., xn ] there exists a sequence (f1 , ..., fn ) of polynomials
in P such that P = [f1 , ..., fn ], where
[f1 , ..., fn ] = {g ∈ S : there exist e1 , ..., en ≥ 0 with gae11 ...aenn ⊆ P }.
It should be convenient to find an algorithm to determine 0-completely irreducible
sequences in this case.
Example 2.8. Assume that R is a ring and Q is a prime ideal of R. Let f1 be
a polynomial of degree one in Z(R)[x1 ] with lc(f1 ) 6∈ Q, where Z(R) denotes the
center of R. Then f1 is a polynomial in ΓQ which is ΓQ -completely irreducible.
Thus P1 = [Q, f1 ] is a prime ideal of R[x1 ]. Take a polynomial f2 of degree one
in Z(R)[x1 , x2 ] such that its leading coefficient with respect to x2 is not in P1 .
Repeating the argument we determine a completely irreducible sequence (mod Q)
(f1 , ..., fn ) in R[x1 , ..., xn ] and a prime ideal P = [Q, f1 , ..., fn ].
We give a concrete example (cf., [3], Example 4.1). Let R be the integral domain
of all the power series of Q[[t]] having the coefficient of t equal to zero. Consider
the polynomials fi = t2 xi − ti+2 , 1 ≤ i ≤ n. Then P = [0, f1 , ..., fn ] is a prime ideal
of R[x1 , ..., xn ] which is not generated by the polynomials f1 , ..., fn (in general, if
every fi is monic, then [Q, f1 , ..., fn ] equals the ideal Q[x] + (f1 , ..., fn ) generated
by Q and {f1 , ..., fn }).
74 MIGUEL FERRERO

Remark 2.9. The method given in this paper can easily be extended to describe
prime ideals of polynomial rings in any number of indeterminates. In fact, choosing
a well order for the indeterminates we can proceed by transfinite induction. Thus
prime ideals will be described by sequences corresponding to the number of indeter-
minates. Finally, since every centred extension of a ring R which has a commuting
centralizing generator set is a factor ring of a polynomial ring over R, this gives
also a description of prime ideals in this kind of centred extensions.

Acknowledgement
This paper was written while the author visited the Department of Mathematics
of University of Southern California. The author is grateful for the hospitality given
by the members of the Department. He is also grateful to the referee for suggestions
that helped to improve the presentation of the paper.

References
1. D. Eisenbud and E. Graham Evans, Jr., Every Algebraic Set in n-Space is the Intersection of
n Hypersufaces, Inventiones Math. 19 (1973), 107-112. MR 48:6125
2. M. Ferrero, Prime and Principal Closed Ideals in Polynomial Rings, J. Algebra 134 (1990),
45-59. MR 91h:16008
3. M. Ferrero, Prime and Maximal Ideals in Polynomial Rings, Glasgow Math. J. 37 (1995),
351-362.
4. I. Kaplansky, Commutative Rings, The University of Chicago Press, Chicago (1974). MR
49:10674

Instituto de Matemática, Universidade Federal do Rio Grande do Sul, 91509-900,

Porto Alegre, Brazil
E-mail address: [email protected]