Bourbaki

Proposition 1

Tag 04RD, permanent. Cite it as [Bourbaki, Tag 04RD].

Commutative Algebra, chapter V, § 3. Finitely generated algebras over a field

Let AA be a finitely generated algebra over a field kk and LL the algebraic closure of kk.

(i) If A≠{0}A \neq \{0\}, there exists a kk-homomorphism from AA to LL.

(ii) Let f1,f2f_1, f_2 be two kk-homomorphisms from AA to LL. For f1f_1 and f2f_2 to have the same kernel, it is necessary and sufficient that there exist a kk-automorphism ss of LL such that f2=s∘f1f_2 = s \circ f_1.

(iii) Let aa be an ideal of AA. For aa to be maximal, it is necessary and sufficient that it be the kernel of a k-homomorphism from AA to LL.

(iv) For an element xx of AA to be such that f(x)=0f(x) = 0 for every k-homomorphism ff from AA to LL, it is necessary and sufficient that xx be nilpotent.

Assertion (i) follows from no. 1, Corollary 3 to Theorem 1 applied replacing AA by kk, BB by AA, bb by the unit element of BB and ff by the canonical injection of kk into LL.

If ff is a k-homomorphism from AA to LL, f(A)f(A) is a subring of LL containing kk; as LL is an algebraic extension of kk, f(A)f(A) is a field (Algebra, Chapter V, § 3, no. 2, Proposition 3) and, if aa is the kernel off, A/aA/a, isomorphic to f(A)f(A), is therefore a field, which proves that aa is maximal. Conversely, if aa is a maximal ideal of AA, it follows from (i) that there exists a k-homomorphism from A/aA/a to LL and hence a k-homomorphism of AA to LL whose kernel bb contains aa; but as aa is maximal, b=ab = a; this proves (iii).

We now prove (ii). If ss is a k-automorphism of LL such that f2=s∘f1f_2 = s \circ f_1, clearly f1f_1 and f2f_2 have the same kernel. Conversely, suppose that f1f_1 and f2f_2 have the same kernel; then there exists a k-isomorphism s0s_0 of the field f1(A)f_1(A) onto the field f2(A)f_2(A) such that f2=s0∘f1f_2 = s_0 \circ f_1; but by Algebra, Chapter V, § 6, no. 3, Proposition 7, s0s_0 extends to a k-automorphism of LL and hence f2=s∘f1f_2 = s \circ f_1.

Finally, if x∈Ax \in A is such that xn=0x^n = 0, for every k-homomorphism ff from AA to LL, (f(x))n=f(xn)=0(f(x))^n = f(x^n) = 0 and hence f(x)=0f(x) = 0 since LL is a field. Conversely, suppose that x∈Ax \in A is not nilpotent; then A[x−1]A[x^{-1}] is a finitely generated AA-algebra (and therefore a finitely generated k-algebra) not reduced to 0 (Chapter 11, § 2, no. 1, Remark 3) and hence there exists a k-homomorphism gg from A[x−1]A[x^{-1}] to LL by (i). If j:A→A[x−1]j : A \to A[x^{-1}] is the canonical homomorphism, f=g∘jf = g \circ j is a k-homomorphism from AA to LL and f(x)g(1/x)=g(x/1)g(1/x)=g(1)=1f(x)g(1/x) = g(x/1)g(1/x) = g(1) = 1, whence f(x)≠0f(x) \neq 0.

Let kk be a field and LL an extension field of kk; an element x=(x1,…,xn)x = (x_1, \ldots, x_n) of LnL^n is called a zero in LnL^n of an ideal cc of the polynomial ring k[X1,…,Xn]k[X_1, \ldots, X_n] if

P(x)=P(x1,…,xn)=0 P(x) = P(x_1, \ldots, x_n) = 0

for all P∈rP \in r.

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Cites

Transcribed from the printed volume. Not translated.