Commutative Algebra
Andrew Kobin
Spring 2016 / 2019 Contents Contents
Contents
1 Preliminaries 1 1.1 Radicals ...... 1 1.2 Nakayama’s Lemma and Consequences ...... 4 1.3 Localization ...... 5 1.4 Transcendence Degree ...... 10
2 Integral Dependence 14 2.1 Integral Extensions of Rings ...... 14 2.2 Integrality and Field Extensions ...... 18 2.3 Integrality, Ideals and Localization ...... 21 2.4 Normalization ...... 28 2.5 Valuation Rings ...... 32 2.6 Dimension and Transcendence Degree ...... 33
3 Noetherian and Artinian Rings 37 3.1 Ascending and Descending Chains ...... 37 3.2 Composition Series ...... 40 3.3 Noetherian Rings ...... 42 3.4 Primary Decomposition ...... 46 3.5 Artinian Rings ...... 53 3.6 Associated Primes ...... 56
4 Discrete Valuations and Dedekind Domains 60 4.1 Discrete Valuation Rings ...... 60 4.2 Dedekind Domains ...... 64 4.3 Fractional and Invertible Ideals ...... 65 4.4 The Class Group ...... 70 4.5 Dedekind Domains in Extensions ...... 72
5 Completion and Filtration 76 5.1 Topological Abelian Groups and Completion ...... 76 5.2 Inverse Limits ...... 78 5.3 Topological Rings and Module Filtrations ...... 82 5.4 Graded Rings and Modules ...... 84
6 Dimension Theory 89 6.1 Hilbert Functions ...... 89 6.2 Local Noetherian Rings ...... 94 6.3 Complete Local Rings ...... 98
7 Singularities 106 7.1 Derived Functors ...... 106 7.2 Regular Sequences and the Koszul Complex ...... 109 7.3 Projective Dimension ...... 114
i Contents Contents
7.4 Depth and Cohen-Macauley Rings ...... 118 7.5 Gorenstein Rings ...... 127
8 Algebraic Geometry 133 8.1 Affine Algebraic Varieties ...... 133 8.2 Morphisms of Affine Varieties ...... 142 8.3 Sheaves of Functions ...... 146 8.4 Finite Morphisms and the Fibre Theorem ...... 147 8.5 Projective Varieties ...... 150 8.6 Nonsingular Varieties ...... 153 8.7 Abstract Curves ...... 156 8.8 The Spectrum of a Ring ...... 159
ii 1 Preliminaries
1 Preliminaries
The contents of this document come from an Algebra IV course taught by Dr. Peter Abra- menko at the University of Virginia in Spring 2016. The main theory of commutative rings are covered, along with applications to algebraic geometry towards the end. Topics include: Integral dependence Noetherian, Artinian and Dedekind rings Valuations and discrete valuation rings Dimension theory Affine and projective varieties Affine algebras Hilbert’s Nullstellensatz Morphisms of varieties The commutative algebra may be found in Atiyah and MacDonald’s Introduction to Commu- tative Algebra, while the algebraic geometry generally follows the first chapter of Hartshorne’s Algebraic Geometry. Several topics in these notes also come from a course on commutative algebra taught by Dr. Craig Huneke at UVA in Spring 2019. The main topics I’ve included from this course are: associated primes, Cohen’s structure theory for complete local rings, some dimension theory and notions of singularity in rings. We will adhere to several conventions throughout these notes: All rings are commutative with unity 1. A will always denote such a ring. All subrings contain 1. In particular, proper ideals are never subrings. For any ring homomorphism f : A → B, in addition to the homomorphism axioms, we stipulate that f(1A) = 1B. In any integral domain, we assume 1 6= 0.
1.1 Radicals
There are three important types of radicals that we may define in a commutative ring A. The first two are defined below. Definition. The nil radical of A is defined as the intersection of all prime ideals of A, written \ N(A) = P. prime ideals P ⊂A We say A is reduced if N(A) = 0.
1 1.1 Radicals 1 Preliminaries
Definition. The Jacobson radical of A is the intersection of all maximal ideals of A, written \ J(A) = M. maximal ideals M⊂A Remark. Clearly N(A) ⊆ J(A). Lemma 1.1.1. For any x ∈ A, x ∈ J(A) if and only if 1 − xy is a unit in A for all y ∈ A. Proof. ( =⇒ ) Suppose A(1 − xy) 6= A for some y ∈ A. Then A(1 − xy) ⊂ M for a maximal ideal M. By definition, x ∈ J(A) ⊆ M and since J(A) is an ideal, xy ∈ M as well. This means 1 = xy + (1 − xy) ∈ M, a contradiction. Therefore A(1 − xy) = A so there is some a ∈ A such that a(1 − xy) = 1; that is, 1 − xy is a unit. ( ⇒ = ) Conversely, suppose x 6∈ M for some maximal ideal M ⊂ A. Then M + Ax = A by maximality of M, so b + xy = 1 for some b ∈ M, or 1 − xy = b ∈ M, meaning 1 − xy cannot be a unit in A. Proposition 1.1.2. N(A) = {x ∈ A | xn = 0 for some n ≥ 1}. That is, the nil radical consists of all nilpotent elements in A. Proof. (⊇) Suppose xn = 0. Then for any prime ideal P , xn = 0 ∈ P which implies x ∈ P . Thus x ∈ N(A). (⊆) If x ∈ A is not nilpotent, we will produce a prime ideal not containing x. Let Σ denote the set of ideals I ⊂ A such that xn 6∈ I for any n ≥ 1. Notice that 0 is an element of Σ since we are assuming x is not nilpotent; thus Σ is nonempty. Therefore we may apply Zorn’s Lemma to choose an ideal P ∈ Σ which is maximal among this collection of ideals. We claim that P is prime. Clearly P 6= A since A is not an element of Σ. Suppose y, z ∈ A but y, z 6∈ P . Since P is maximal in Σ, the ideals P + Ay and P + Az are both strictly larger ideals than P . Thus P + Ay, P + Az 6∈ Σ so there exist m, n ∈ N such that xm ∈ P + Ay and xn ∈ P + Az. Then xm+n ∈ (P + Ay)(P + Az) ⊆ P + Ayz, but xm+n 6∈ P by assumption, so we must have yz 6∈ P . This proves P is prime. n Definition. For an ideal I ⊂ A, the radical of I√is r(I) = {x ∈ A | x ∈ I for some n ≥ 1}. Alternate notations for the radical ideal include I and rad(I). Corollary 1.1.3. For any ideal I ⊂ A, r(I) equals the intersection of all prime ideals of A containing I. In particular, r(I) is an ideal. Proof. Consider the quotient ring A¯ = A/I. For x ∈ A we have the following equivalent statements:
n x ∈ r(I) ⇐⇒ there exists an n ∈ N such that x ∈ I n ⇐⇒ there exists an n ∈ N such that (x + I) = 0 ∈ A¯ ⇐⇒ x + I is nilpotent, i.e. x + I ⊂ N(A¯) ⇐⇒ x + I is contained in every prime of A¯ ⇐⇒ x lies in every prime ideal of A containing I. \ Therefore r(I) = P as desired. prime ideals P ⊇I
2 1.1 Radicals 1 Preliminaries
Remark. By Corollary 1.1.3, N(A) = r(0). Lemma 1.1.4. Suppose I,J ⊂ A are ideals. Then (a) I ⊆ r(I). (b) I = r(I) if and only if A/I is reduced. In particular, every prime ideal is radical. (c) r(r(I)) = r(I).
(d) r(Im) = r(I) for all m ∈ N. (e) r(IJ) ⊇ r(I)r(J). (f) r(I + J) = r(r(I) + r(J)). (g) r(I) = A if and only if I = A. Proof. (a) is clear from the definition. (b) By Corollary 1.1.3, r(I) equals the intersection of all primes lying over I. Let π : A → A/I be the quotient map. Then N(A/I) = 0 ⇐⇒ π(r(I)) = 0 ⇐⇒ r(I) ⊆ ker π = I. By part (a), we always have I ⊆ r(I) so this shows that N(A/I) = 0 ⇐⇒ r(I) = I. (c) By part (a), it suffices to show r(r(I)) ⊆ r(I). If x ∈ r(r(I)) then xn ∈ r(I) for some n ≥ 1, but then (xn)m = xnm ∈ I for some m ≥ 1. Thus x ∈ r(I). (d) If x ∈ r(Im) then xn ∈ Im for some n ≥ 1, but Im ⊆ I so this shows x ∈ r(I). On the other hand, if x ∈ r(I) then xn ∈ I for some n ≥ 1. Letting k = max{n, m}, we see that xk ∈ Im. Hence r(I) = r(Im). (e) It suffices to prove this for elements of the form xy ∈ r(I)r(J), where x ∈ r(I) and y ∈ r(J). Then xn ∈ I and ym ∈ J for some n, m ≥ 1. Without loss of generality, assume n ≥ m. Then (xy)n = xnyn ∈ IJ so xy ∈ r(IJ). (f) First take z ∈ r(I + J), so that zn ∈ I + J for some n ≥ 1. Then zn = x + y for x ∈ I, y ∈ J. Since I ⊆ r(I) and J ⊆ r(J) by part (a), we have zn = x + y ∈ I + J ⊆ r(I) + r(J). Therefore z ∈ r(r(I) + r(J)). On the other hand, if w ∈ r(r(I) + r(J)) then wm ∈ r(I) + r(J) for some m ≥ 1. This means wm = x + y for some x ∈ r(I), y ∈ r(J). In turn this says that xk ∈ I and y` ∈ J for some k, ` ≥ 1. Using the binomial formula, write k + ` wm(k+`) = (x + y)k+` = xk+` + (k + `)xk+`−1y + ... + xk+`−qyq + ... + yk+`. q Note that q ranges from 0 to k + `. When 0 ≤ q ≤ `, the qth term (the first term in the sum being the 0th term) is divisible by xk and so it lies in I. Otherwise when 1 + ` ≤ q ≤ k + `, the qth term is divisible by y` and so it lies in J. Therefore wm(k+`) is of the form a + b where a ∈ I and b ∈ J. Hence w ∈ r(I + J). (g) Suppose r(I) = A. Then in particular 1 ∈ r(I) so for some n ≥ 1, 1 = 1n ∈ I. Thus I = A. On the other hand, if I = A then part (a) says A ⊆ r(A), which implies r(A) = A.
Example 1.1.5. Let A = Z and I = J = 2Z. Then IJ = 4Z so r(IJ) = 2Z, but I and J are each prime and thus radical, so r(I)r(J) = IJ = 4Z. Hence r(IJ) ) r(I)r(J), showing that the containment in (e) may be strict.
3 1.2 Nakayama’s Lemma and Consequences 1 Preliminaries
Example 1.1.6. Let k be an algebraically closed field and consider A = k[x, y]. Take I = (y) and J = (y − x2). Then I = r(I) since A/I = k[x] which is a PID and hence reduced. Likewise, k[x, y]/(y − x2) ∼= k[x] by the first isomorphism theorem applied to the map k[x, y] → k[x], f(x, y) 7→ f(x, x2). Thus J = r(J), so r(I) + r(J) = I + J. However, I + J = (y) + (y − x2) = (y, x2) and the radical of this ideal is r((y, x2)) = (y, x), a strictly larger ideal than I + J. This shows the containment r(I + J) ⊂ r(I) + r(J) is sometimes strict, so the second radical on the right side of property (f) is necessary for equality. Tn Tn Lemma 1.1.7. Let J1,...,Jn be ideals of a ring A and J = i=1 Ji. Then r(J) = i=1 r(Ji). n Proof. It suffices to prove the case when n = 2. Suppose x ∈ r(J1 ∩J2). Then x ∈ J1 ∩J2 for n n some n ∈ N, so in particular x ∈ J1, implying x ∈ r(J1), and x ∈ J2, implying x ∈ r(J1). Hence x ∈ r(J1) ∩ r(J2). k On the other hand, if y ∈ r(J1) ∩ r(J2), then y ∈ r(J1) so y ∈ J1 for some k ∈ N. ` m Similarly, y ∈ r(J2) implies y ∈ J2 for some ` ∈ N. Set m = max(k, `). Then y = m−k k m m−` ` m y y ∈ J1 and y = y y ∈ J2 by the setup, so y ∈ J1 ∩ J2. Hence y ∈ r(J1 ∩ J2). This proves r(J1 ∩ J2) = r(J1) ∩ r(J2). Example 1.1.8. The conclusion of Lemma 1.1.7 does not hold for arbitrary intersections n of ideals. Let k be a field and A = k[t] and consider the family of ideals Jn = (t ) for n ∈ . Then T J = 0 since if t` | f(t) for all ` ≥ 1 then necessarily f = 0. So N n∈N n r T J = r(0) = N(A) = 0 since A is a PID. However, for each n ∈ , r((tn)) = (t) so n∈N n N T r(J ) = (t) 6= 0. n∈N n Lemma 1.1.9. If f : A → B is a ring homomorphism and J is an ideal of B, then f −1(r(J)) = r(f −1(J)). Proof. Let x ∈ A. Then x ∈ f −1(r(J)) ⇐⇒ f(x) ∈ r(J) n ⇐⇒ f(x) ∈ J for some n ∈ N n ⇐⇒ f(x ) ∈ J for some n ∈ N since f is a homomorphism n −1 ⇐⇒ x ∈ f (J) for some n ∈ N ⇐⇒ x ∈ r(f −1(J)). Hence f −1(r(J)) = r(f −1(J)).
1.2 Nakayama’s Lemma and Consequences
A classic result in ring theory is Nakayama’s Lemma, which will be useful later on. Theorem 1.2.1 (Nakayama’s Lemma). If I ⊂ A is an ideal such that I ⊆ J(A), and M is a finitely generated A-module such that IM = M, then M = 0. Proof. Assume IM = M but M 6= 0. Since M is finitely generated, pick some minimal finite set of generators {m1, . . . , mn} of M. Then mn ∈ M = IM = Im1 + ... + Imn so there exist elements a1, . . . , an ∈ I such that mn = a1m1 + ... + anmn. This can be written
(1 − an)mn = a1m1 + ... + an−1mn−1.
4 1.3 Localization 1 Preliminaries
Since an ∈ I ⊆ J(A), Lemma 1.1.1 shows 1−an is a unit in A. Then mn ∈ Am1+...+Amn−1. This shows that {m1, . . . , mn−1} generates M as an A-module, contradicting minimality. Therefore we must have M = 0. One can find examples of ideals not lying in the Jacobson radical for which IM = M for nontrivial modules M.
Example 1.2.2. Let A be a local ring which is an integral domain but not a field. As an a example, for any prime p, the ring of p-adic integers Z(p) = b : a, b ∈ Z, p - b is a local subring of Q which is not a field. Let M be the unique maximal ideal of A and take F to be its field of fractions (see below for details). Since A is not a field, M 6= 0. In fact, in the local ring situation, M = J(A). Then we have MF = F for the A-module F , but F 6= 0. This provides a counterexample to Nakayama’s Lemma when the module is not finitely generated, as is the case with the residue field of a local ring.
The following is another version of Nakayama’s Lemma which may be useful in some contexts.
Corollary 1.2.3. Let M be a finitely generated A-module and take N ⊆ M to be a submod- ule. If I ⊂ A is an ideal such that I ⊆ J(A) and M = IM + N, then M = N.
Proof. Consider the finitely generated A-module M = M/N. Then
IM = (IM + N)/N = M/N = M.
Applying Nakayama’s Lemma (1.2.1) shows that M = 0, that is, M = N. Let A be a local ring with maximal ideal M = J(A). We call the quotient field k = A/M the residue field of A. If N is a finitely generated A-module then N/MN is a finite dimensional k-vector space.
Proposition 1.2.4. If x1, . . . , xn are elements of N such that x¯1,..., x¯n ∈ N/MN span N/MN as a k-vector space, then N is generated by the xi as an A-module.
0 Pn Proof. Set N = i=1 Axi, which is a submodule of N. Consider the canonical projection π : N → N/MN. By assumption, π(N 0) = N/MN. Thus N = N 0 + MN but since M = J(A) and N is finitely generated as an A[x]-module, the conditions of Corollary 1.2.3 are satisfied. Hence N 0 = N.
1.3 Localization
One of the most important constructions in commutative algebra is the localization of a ring at a multiplicatively closed subset.
Definition. A subset S ⊆ A is said to be multiplicatively closed provided that 1A ∈ S and for all x, y ∈ S, xy ∈ S as well.
Examples. Two of the most important examples of multiplicatively closed subsets are:
5 1.3 Localization 1 Preliminaries
1 For any prime ideal P ⊂ A, the set S = A r P is multiplicatively closed. 2 For a fixed element f ∈ A (usually a function in a ring of functions), the set S = {f n | 0 n ≥ 0} is multiplicatively closed, where by convention f = 1A. Definition. Let S ⊆ A be multiplicatively closed and suppose M is a finitely generated A- module. The module of fractions S−1M is defined as S−1M = M × S/ ∼ where ∼ is the equivalence relation (m, s) ∼ (n, t) iff u(tm − sn) = 0 for some u ∈ S. It is common to write −1 m −1 elements of S M as s instead of (m, s). Addition and the A-module action of S M are given by m n tm + sn m am + = and a = . s t st s s Lemma 1.3.1. The above operations make S−1M into an A-module. −1 a b Proposition 1.3.2. For the A-module A, S A is a ring under the multiplication law s · t = ab 1A , with unity 1S−1A = . st 1A Proof. We will verify that multiplication is well-defined. The other ring axioms are easily checked. If (a, s) ∼ (a0, s0) then u(s0a − sa0) = 0 for some u ∈ S. Thus we have u(s0tab − sta0b) = tbu(s0a − sa0) = 0 so (a0b, s0t) ∼ (ab, st). Example. 3 If A is an integral domain then S = A r {0} is multiplicatively closed. The ring S−1A is called the field of fractions of A. −1 a Lemma 1.3.3. There is a homomorphism j : A → S A defined by a 7→ 1 which is one-to- one if and only if S contains no zero divisors. Proposition 1.3.4 (Universal Property of Localization). Let S ⊆ A be a multiplicatively closed subset. For any ring homomorphism ϕ : A → B such that ϕ(S) ⊆ B∗, there exists a unique ring homomorphism ϕ0 : S−1A → B making the following diagram commute: ϕ A B
j ϕ0
S−1A
a −1 0 0 a 0 a 0 1 Proof. First, if s ∈ S A and such a ϕ is defined, we must have ϕ s = ϕ 1 ϕ s = −1 0 0 a −1 ϕ(a)ϕ(s) . This implies uniqueness. Now define ϕ in this way: ϕ s = ϕ(a)ϕ(s) for 0 a b −1 any a ∈ A, s ∈ S. To see that ϕ is well-defined, suppose s = t in S A. Then there is an element u ∈ S such that u(at − bs) = 0. Applying ϕ, we have ϕ(u)ϕ(at − bs) = 0. Since ϕ(u) is a unit in B, it follows that ϕ(at − bs) = 0 =⇒ ϕ(a)ϕ(t) = ϕ(b)ϕ(s) =⇒ ϕ(a)ϕ(s)−1 = ϕ(b)ϕ(t)−1. Hence ϕ0 is well-defined. The rest follows easily from this definition of ϕ0.
6 1.3 Localization 1 Preliminaries
Examples.
−1 4 If S = ArP for a prime ideal P ⊂ A, the localization of A at P is the ring AP := S A.
5 If S = {f n | n ≥ 0} for some element f ∈ A then the localization of A at f is −1 Af := S A. ∼ Proposition 1.3.5. If f, g ∈ A are elements such that r((f)) = r((g)) then Af = Ag. Proof. The assumption r((f)) = r((g)) means that f n = xg and gm = yf for some n, m ≥ 1 and x, y ∈ A. Let jf : A → Af and jg : A → Ag be the canonical homomorphisms sending a g n a 7→ 1 in each localization. Then jf (g) = 1 is invertible in Af , since f − xg = 0 and so g x 1 f 1 · f n = 1 . Symmetrically, jg(f) = 1 is invertible in Ag. Therefore by the universal property of localization, there are unique homomorphisms ϕ1 : Ag → Af and ϕ2 : Af → Ag such that ϕ1 ◦ jg = jf and ϕ2 ◦ jf = jg: A
jf jg
ϕ2 Af Ag ϕ1 ∼ Then it is clear that ϕ1 and ϕ2 are inverses of each other, so Af = Ag. Proposition 1.3.6. If M is an A-module and S ⊆ A is a multiplicatively closed subset then there exists a well-defined isomorphism of S−1A-modules (or of A-modules)
−1 −1 f : S A ⊗A M −→ S M a am s ⊗ m 7−→ s . Proof. Define a map
S−1A × M −→ S−1M a am s , m 7−→ s . It is easy to check that is A-bilinear, so by the universal property of tensor products, we get a −1 −1 well-defined homomorphism of A-modules f : S A ⊗A M → S M. Clearly f is surjective. P ai −1 Q Q For an arbitrary element ⊗ mi ∈ S A ⊗ M, set s = si ∈ S and ti = sj ∈ S. i si i j6=i Then X ai X aiti X 1 1 X ⊗ mi = ⊗ mi = ⊗ aitimi = ⊗ aitimi. si s s s i i i i −1 1 Therefore every element of S A ⊗ M has the form s ⊗ m for s ∈ S and m ∈ M. Now 1 m suppose f s ⊗ m = 0. Then s = 0 by definition of the map, so tm = 0 for some t ∈ S. 1 t 1 1 Thus s ⊗ m = st ⊗ m = st ⊗ tm = st ⊗ 0 = 0. This proves f is injective and thus we have −1 ∼ −1 an isomorphism S A ⊗A M = S M as A-modules. Finally, it’s easy to see that the same map is also an isomorphism of S−1A-modules.
7 1.3 Localization 1 Preliminaries
Given an ideal I ⊂ A, we can define the subset S−1I ⊆ S−1A by viewing I as an −1 a A-module. Specifically, S I = s : a ∈ I, s ∈ S . Lemma 1.3.7. Let I ⊂ A be an ideal and let S ⊂ A be a multiplicatively closed subset. Then
(a) S−1I ⊆ S−1A is an ideal.
(b) S−1I = S−1A if and only if S ∩ I 6= ∅.
a −1 b −1 a b ab Proof. (a) Let s ∈ S A and t ∈ S I, meaning s, t ∈ S, a ∈ A and b ∈ I. Then s · t = st lies in S−1I since ab ∈ I by absorption and st ∈ S by the multiplicatively closed property. Hence S−1I is an ideal of S−1A. s 1 −1 −1 −1 (b) On one hand, if s ∈ S ∩ I then s = 1 ∈ S I so S I = S A. On the other hand, −1 −1 1 −1 a 1 S I = S A implies 1 ∈ S I so there exist elements a ∈ I and s ∈ S such that s = 1 . Then there is a t ∈ S such that t(a − s) = 0, that is, ts = ta ∈ S ∩ I. Thus S ∩ I is nonempty.
Proposition 1.3.8. Let S be a multiplicatively closed subset of A. Then
(a) Every proper ideal J ⊂ S−1A is of the form S−1I for some ideal I not intersecting S.
(b) For any prime ideal P ⊂ A not intersecting S, j−1(S−1P ) = P .
(c) The map
−1 {P ⊂ A | P is a prime ideal,S ∩ P = ∅} −→ {prime ideals of S A} P 7−→ S−1P
is an inclusion-preserving bijection.
−1 −1 a −1 Proof. (a) Let J ⊂ S A be an ideal. Then I = j (J) is an ideal of A. If s ∈ S I for a a a 1 a a a s a ∈ I, s ∈ S, then 1 ∈ J so s = 1 · s ∈ J. On the other hand, if s ∈ J then 1 = s · 1 ∈ J. −1 a −1 −1 Then a ∈ I since I = j (J). Hence s ∈ S I, which proves that S I = J. Also, S ∩I = ∅ since S−1I = J ( S−1A. (b) By definition of the map j, P ⊆ j−1(S−1P ). For the other containment, suppose −1 −1 a −1 a b a ∈ j (S P ). Then 1 ∈ S P so there exist b ∈ P and s ∈ S such that 1 = s . Thus for some t ∈ S we have t(sa − b) = 0, or tsa = tb which lies in P since b ∈ P . Since s, t ∈ S and S ∩ P = ∅, s, t 6∈ P but then tsa ∈ P implies a ∈ P since P is prime. Thus P = j−1(S−1P ) as claimed. x y −1 (c) Suppose P ⊂ A is a prime ideal with S ∩ P = ∅. Let s , t ∈ S A such that x y xy −1 xy a s · t = st ∈ S P . Then there are elements a ∈ P and u ∈ S such that st = u . So for some v ∈ S, v(uxy − sta) = 0; that is, vuxy = vsta which lies in P since a ∈ P . Now x −1 y −1 v, u ∈ S so v, u 6∈ P , meaning x ∈ P or y ∈ P . Thus s ∈ S P or t ∈ S P . This proves that S−1P is a prime ideal of S−1A. Finally, denote the map P 7→ S−1P by Φ. By (b), Φ is injective. For surjectivity, assume J ⊂ S−1A is prime. Then P = j−1(J) is also prime. By (a), J = S−1P = Φ(P ). Also, S ∩ P = ∅ since J is prime and therefore J 6= S−1A. Hence Φ is surjective, so it is a bijection.
8 1.3 Localization 1 Preliminaries
Corollary 1.3.9. If P ⊂ A is prime then AP is a local ring with unique maximal ideal S−1P , where S = A r P . Moreover, there is an inclusion-preserving bijection
{Q ⊂ A | Q is a prime ideal,Q ⊆ P } ←→ {prime ideals of AP }.
Proof. We know that S−1P ⊂ S−1A is a prime ideal. Observe that
−1 −1 a a ∗ S A r S P = s : a ∈ A r P, s ∈ S = s : a ∈ S, s ∈ S = AP ,
−1 −1 the set of units in AP . This implies S P is the unique maximal ideal of S A. The bijection follows from (c) of Proposition 1.3.8, using the fact that S ∩Q = ∅ if and only if Q ⊆ P . Lemma 1.3.10. Let S be a multiplicatively closed subset and I,J ideals of the ring A. Then
(a) S−1(IJ) = (S−1I)(S−1J).
(b) S−1(I ∩ J) = (S−1I) ∩ (S−1J).
(c) S−1r(I) = r(S−1I).
x −1 Pn Proof. (a) Let s ∈ S (IJ), with x ∈ IJ and s ∈ S. Then x = k=1 akbk for ak ∈ I and bk ∈ J. Moreover, we assume 1 ∈ S by convention so
n n x X akbk X ak bk = = · . s s s 1 k=1 k=1
−1 −1 x Then each summand is a product of an element in S I and an element in S J, so s ∈ −1 −1 −1 −1 −1 y −1 −1 (S I)(S J). Thus S (IJ) ⊆ (S I)(S J). On the other hand, if t ∈ (S I)(S J) then there are elements ak ∈ I, bk ∈ J, sk, tk ∈ S such that
n Pn y X akbk s1t1 ··· sˆktˆk ··· sntnakbk = = k=1 . t sktk s1t1 ··· sntn k=1
The numerator lies in IJ since ak ∈ I, bk ∈ J so each summand is a product of an element of I and an element of J. The denominator is an element of S since this set is multiplicative. y −1 −1 −1 −1 Therefore t ∈ S (IJ) by definition. Hence (S I)(S J) ⊆ S (IJ). x −1 (b) Observe that for any s ∈ S A, x ∈ S−1(I ∩ J) ⇐⇒ x ∈ I ∩ J and s ∈ S s x x ⇐⇒ ∈ S−1I and ∈ S−1J s s x ⇐⇒ ∈ (S−1I) ∩ (S−1J). s So S−1(I ∩ J) = (S−1I) ∩ (S−1J). (c) If I ∩ S 6= ∅, I ⊆ r(I) implies r(I) ∩ S 6= ∅ as well so we have S−1r(I) = S−1A = r(S−1A) = r(S−1I).
9 1.4 Transcendence Degree 1 Preliminaries
Now assume I ∩ S = ∅. Since S is multiplicative, if s were in r(I) for any s ∈ S then sn ∈ I ∩ S for some n ∈ N, which is impossible. Therefore r(I) ∩ S = ∅. By Corollary 1.1.3, T r(I) = p where p runs over all prime ideals of A that contain I. Then by Proposition 1.3.8, we have ! \ \ \ S−1r(I) = S−1 p = S−1p = q = r(S−1I) p⊇I p⊇I q⊇S−1I where p runs over the primes of A containing I and q runs over the primes of S−1A containing S−1I. Hence we have S−1r(I) = r(S−1I) in all cases.
−1 Lemma 1.3.11. For a prime ideal P ⊂ A, the quotient field AP /S P is isomorphic to the field of fractions of A/P .
Proof. Let ϕ : A −→π A/P −→i Frac(A/P ) be the canonical quotient map composed with a¯ i :a ¯ 7→ 1 . Then P ⊆ ker ϕ, so S = ArP gets mapped to nonzero elements in Frac(A/P ), but these are all units since Frac(A/P ) is a field. Therefore the universal property of localization gives us a unique homomorphism f : AP → Frac(A/P ) such that f ◦ j = ϕ. Notice that if a −1 a −1 s ∈ S P then a ∈ P so ϕ(a) = 0 and therefore f s = 0. So S P ⊆ ker f. Clearly f is −1 ∼ onto, so by the first isomorphism theorem, we have AP /S P = Frac(A/P ). Example.
6 Let A = Z. Since Z is an integral domain, (0) is a prime ideal and its localization is A(0) = Q, the field of fractions of Z. For any nonzero prime p ∈ Z, the localization a Z(p) = s : a, s ∈ Z, p - s is called the ring of p-adic integers. Each Z(p) contains only −1 the prime ideals (0) and pZ(p) := S (p).
1.4 Transcendence Degree
Let K/k be an extension of fields and let E ⊆ K be any subset.
Definition. We say E is algebraically independent over k if for any finite subset {x1, . . . , xn} ⊆ E and any function F ∈ k[t1, . . . , tn] for which F (x1, . . . , xn) = 0, it must be that F = 0. When E = {x}, we say the element x is transcendental (or transcendent) over k.
Clearly x ∈ K is transcendental over k if and only if it is not algebraic over k.
Definition. A set E ⊆ K is a transcendence basis for K/k if E is a maximal algebraically independent subset of K. In the case that E is a transcendence basis for K/k and K = k(E), we say the extension is purely transcendental.
Note that the empty set E = ∅ is a transcendence basis if and only if K/k is an algebraic extension.
Example 1.4.1. Let K = k(t1, . . . , tn) be the field of fractions of the polynomial ring k[t1, . . . , tn]. Then K is a purely transcendental extension of K, with transcendence basis E = {t1, . . . , tn}.
10 1.4 Transcendence Degree 1 Preliminaries
∼ Lemma 1.4.2. If {x1, . . . , xn} ⊂ K is algebraically independent then k(x1, . . . , xn) = k(t1, . . . , tn). Proof. Consider the surjection
ϕ : k[t1, . . . , tn] −→ k[x1, . . . , xn]
ti 7−→ xi.
Then ϕ is injective because {x1, . . . , xn} is algebraically independent. Moreover, ϕ is a k- algebra homomorphism so this extends to an isomorphism k(t1, . . . , tn) → k(x1, . . . , xn). Lemma 1.4.3. Let K/k be a field extension and suppose E ⊆ K with |E| = n. Then
(a) If E is algebraically independent over k then there exists a transcendence basis E0 of K/k such that E0 ⊇ E.
(b) If E is finite and K = k(E) then the elements of E may be numbered such that k(x1, . . . , xr)/k is purely transcendental for some 0 ≤ r ≤ n and K/k(x1, . . . , xr) is algebraic.
(c) E is a transcendence basis for K/k if and only if K/k(E) is algebraic and K/k(E1) is not algebraic for any proper subset E1 ( E. Proof. (a) Apply Zorn’s Lemma (exercise). (b) Let E be numbered so that {x1, . . . , xr} is a maximal algebraically independent subset of E. Clearly {x1, . . . , xr} is a transcendence basis for k(x1, . . . , xr) so k(x1, . . . , xr)/k is purely transcendental. We will show that for each r + 1 ≤ j ≤ n, xj is algebraic over k(x1, . . . , xr). We know for each j, {x1, . . . , xr, xj} is algebraically dependent over k, so there exists a nonzero polynomial F ∈ k[t1, . . . , tr+1] such that F (x1, . . . , xr, xj) = 0. Write
d F = Fd(t1, . . . , tr)tr+1 + ... + F1(t1, . . . , tr)tr+1 + F0(t1, . . . , tr)
where Fi ∈ k[t1, . . . , tr] and Fd 6= 0. Then
d 0 = Fd(x1, . . . , xr)xj + ... + F1(x1, . . . , xr)xj + F0(x1, . . . , xr)
and Fd(x1, . . . , xr) 6= 0 since {x1, . . . , xr} is algebraically independent. We must therefore have d ≥ 1. This shows xj is algebraic over k(x1, . . . , xr) for each r + 1 ≤ j ≤ n, as we had claimed. Finally, it follows that K/k(x1, . . . , xr) is algebraic. (c) Pick any x ∈ K r E. Then since E is a maximal algebraically independent subset of K, E ∪ {x} is algebraically dependent over k. This means there exist x1, . . . , xr ∈ E such that {x1, . . . , xr, x} is algebraically dependent over k. Hence x must be algebraic over k(x1, . . . , xr) as in (b). In particular, x is algebraic over k(E). To show the second property, we may suppose without loss of generality that E1 = {x1, . . . , xr−1} so that K/k(x1, . . . , xr−1) is algebraic. Consider the tower K ⊃ k(E) ⊃ k(x1, . . . , xr−1). Then k(E) = k(x1, . . . , xr) is algebraic over k(x1, . . . , xr−1), which implies xr is algebraic over k. This clearly contradicts the algebraic independence of {x1, . . . , xr}. Hence K/k(E1) cannot be algebraic for any proper subset E1 ( E.
11 1.4 Transcendence Degree 1 Preliminaries
Going the other direction, suppose E satisfies the properties that K/k(E) is algebraic and K/k(E1) is not algebraic for any E1 ( E. By (b), we may number E = {x1, . . . , xr, . . . , xn} so that k(x1, . . . , xr)/k is purely transcendental and K/k(x1, . . . , xr) is algebraic. By as- sumption {x1, . . . , xr} cannot be a proper subset of E, so we get r = n and k(E)/k is purely transcendental. This implies E is algebraically independent over k. Finally, since K/k(E) is algebraic, for any x ∈ K r k(E), x is algebraic. So E ∪ {x} is not algebraically independent. This proves maximality of E, hence E is a transcendence basis.
Proposition 1.4.4. Assume E1,E2 ⊆ K are finite subsets with K/k(E1) algebraic and E2 algebraically independent over k. Then |E1| ≥ |E2|.
Proof. Let E1 = {x1, . . . , xm} and E2 = {y1, . . . , yn}, so that |E1| = m and |E2| = n. We claim that the elements of E1 can be numbered so that K/k(y1, . . . , y`, x`+1, . . . , xm) is algebraic for all 0 ≤ ` ≤ min{m, n}. We prove this by induction on `. For ` = 0, K/k(E1) is already algebraic by assumption. Now suppose we can number E1 so that K/k(y1, . . . , y`−1, x`, . . . , xm) is algebraic; set K`−1 = k(y1, . . . , y`−1, x`, . . . , xm). Now y` is algebraic over K`−1 since y` ∈ K. This implies there exist polynomials G0,...,Ge in m variables such that
e Ge(y1, . . . , y`−1, x`, . . . , xm)y` + ... + G0(y1, . . . , y`−1, x`, . . . , xm) = 0 and Ge(y1, . . . , y`−1, x`, . . . , xm) 6= 0. In particular, Ge must be nonzero. It follows that there exists a nonzero polynomial F in m + 1 variables such that
F (y1, . . . , y`, x`, . . . , xm) = 0 (∗)
We may assume F is such a polynomial of minimal degree. Now {y1, . . . , y`} is algebraically independent over k so one of x`, . . . , xm must occur in (∗). After renumbering, we may assume x` occurs in (∗), which then becomes
d Fd(y1, . . . , y`, x`+1, . . . , xm)x` + ... + F0(y1, . . . , y`, x`+1, . . . , xm) = 0 for polynomials Fi in m variables over k with Fd 6= 0. Then Fd(y1, . . . , y`, x`+1, . . . , xm) 6= 0 due to the minimality of deg F . This also implies d ≥ 1. Then deg F ≥ deg Fd + d. The above equation shows that x` is algebraic over K` := k(y1, . . . , y`, x`+1, . . . , xm). By induction, K/K`−1 is algebraic but since K`−1 ⊆ K`, we have now shown that K/K`(x`) and K`(x`)/K` are both algebraic extensions. By the tower property of algebraic extensions, K/K` is algebraic, proving the claim. If min{m, n} = n then n ≤ m and we’re done. On the other hand, if min{m, n} = m then we can use the above induction to conclude that K/k(y1, . . . , ym) is algebraic. But then we must have n = m since if n > m, ym+1 will be algebraic over k(y1, . . . , ym), in which case there is a nontrivial polynomial equation H(y1, . . . , ym+1) = 0. This however contradicts {y1, . . . , ym+1} being algebraically independent. In every case we have n ≤ m as desired.
Corollary 1.4.5. If E1 and E2 are transcendence bases for K/k then |E1| = |E2| or they are both infinite. Proof. Apply Proposition 1.4.4 and (c) of Lemma 1.4.3.
12 1.4 Transcendence Degree 1 Preliminaries
This allows us to define the following:
Definition. The transcendence degree of a field extension K/k is defined as tr degk K = |E| if K/k admits a finite transcendence basis E, and tr degk K = ∞ otherwise.
Remark. tr degk K = 0 if and only if K/k is algebraic.
Corollary 1.4.6. The number r in Lemma 1.4.3(b) such that k(x1, . . . , xr)/k is purely tran- scendental and K/k(x1, . . . , xr) is algebraic is equal to tr degk K.
Proof. We proved that {x1, . . . , xr} is algebraically independent over k and {x1, . . . , xr, x} is algebraically dependent for any x ∈ K r {x1, . . . , xr}. Hence {x1, . . . , xr} is a maximal algebraically independent set in K, and thus a transcendence basis for K/k. Therefore
r = #{x1, . . . , xr} = tr degk K.
Example 1.4.7. For the purely transcendental extension k(t1, . . . , tn), its transcendence degree is tr degk k(t1, . . . , tn) = n. Proposition 1.4.8. For a tower of fields L ⊃ K ⊃ k in which each extension is finitely generated, we have
tr degk L = tr degK L + tr degk K.
Proof. Suppose E1 = {x1, . . . , xn} is a transcendence basis for K/k and E2 = {y1, . . . , ym} is a transcendence basis for L/K. Set E = E1 ∪ E2 = {x1, . . . , xn, y1, . . . , ym}. We first show E is algebraically independent over k. Suppose f ∈ k[t1, . . . , tn+m] such that
f(x1, . . . , xn, y1, . . . , ym) = 0.
Then we may write X an+1 an+m f = fi1,...,im (t1, . . . , tn)tn+1 ··· tn+m i1,...,im where each fi1,...,im ∈ k[t1, . . . , tn], aj ≥ 0 and the sum is over all tuples (i1, . . . , im) ∈ m N . Notice that f(E1, tn+1, . . . , tn+m) ∈ K[tn+1, . . . , tn+m] so because E2 is algebraically independent over K and f(E1,E2) = 0, we must have f(E1, tn+1, . . . , tn+m) ≡ 0. Therefore each fi1,...,im (E1) = 0 but because E1 is algebraically independent over k, each fi1,...,im ≡ 0. Hence f ≡ 0, showing E is algebraically independent over k. To finish the proof, note that K/k(E1) and L/K(E2) are algebraic by Lemma 1.4.3(c), and we have a tower k(E) ⊂ K(E2) ⊂ L in which each intermediate extension is algebraic, so L/k(E) is algebraic by transitivity. Hence E is a transcendence basis for L/k. It follows that tr degk L = |E| = |E1| + |E2| = tr degk K + tr degK L.
Proposition 1.4.9. If k is a perfect field, K = k(E) and tr degk K = r, then the elements of E can be numbered such that k(x1, . . . , xr)/k is purely transcendental and K/k(x1, . . . , xr) is a separable algebraic extension.
Proof. If char k = 0 then this follows from Corollary 1.4.6 and (b) of Lemma 1.4.3. If char k = p > 0 then k = kp and the result follows from Proposition 4.1 in Lang.
13 2 Integral Dependence
2 Integral Dependence
A critical concept in algebraic number theory and algebraic geometry is the idea of integrality: the property that an element (or a collection of elements) satisfies a polynomial identity with coefficients in a ring. In this chapter, we define integrality and explore the basic properties of integral extensions of rings, eventually proving the ‘going up’ and ‘going down’ theorems (Section 2.3) and Noether’s Normalization Lemma (Section 2.4). These results have vital consequences in later chapters.
2.1 Integral Extensions of Rings
In this section we define and outline the main results about integrality over a ring A. Let A ⊆ B be a ring extension, i.e. A is a subring of B.
Lemma 2.1.1. If M is a B-module which is finitely generated as an A-module, then for any x ∈ B, there exists a monic polynomial f(t) ∈ A[t] of degree at least 1 such that f(x)M = 0. Pn Proof. Let M = i=1 Ami. Since M is a B-module, there exist elements aij ∈ A for 1 ≤ i, j ≤ n, such that x satisfies the equations
n X xmi = aijmj for each 1 ≤ i ≤ n. j=1
Pn On the other hand, x also satisfies xmi = j=1 xδijmj for each i, so subtracting these equations gives us a system
n X (xδij − aij)mj = 0 for each 1 ≤ i ≤ n. j=1
Set cij = xδij − aij ∈ B and consider the matrix C = (cij) ∈ Mn(B). Then the system of equations above can be written m1 0 . . C . = . (∗) mn 0
From linear algebra, there exists an adjoint matrix adj(C) ∈ Mn(B) such that adj(C)C = (det C)In. Multiplying through (∗) on the left by adj(C) gives us m1 0 . . (det C) . = . mn 0
14 2.1 Integral Extensions of Rings 2 Integral Dependence
Since the mi generated M, this proves that (det C)M = 0. Now, expanding det C using the Leibniz formula for a determinant yields X det C = sgn(σ)cσ(1)1 ··· cσ(n)n
σ∈Sn n Y 0 n−2 0 0 = (x − aii) + an−2x + ... + a0 for some ai ∈ A i=1 n n−1 = x + an−1x + ... + a0 for some ai ∈ A, n n−1 after combining coefficients. Set f(t) = t + an−1t + ... + a0. Then f(x)M = 0. Definition. An element x ∈ B is integral over A if there exists a monic polynomial f(t) ∈ A[t] of degree at least 1 such that f(x) = 0. If every element in B is integral over A then we say B is integral over A, or B is an integral extension of A. Lemma 2.1.2. For x ∈ B the following are equivalent: (i) x is integral over A. (ii) A[x] is a finitely generated A-module. (iii) There exists a subring C ⊆ B containing A[x] such that C is finitely generated as an A-module. (iv) There exists a faithful A[x]-module M which is finitely generated as an A-module. n n−1 n Pn−1 i Proof. (i) =⇒ (ii) If x +an−1x +...+a1x+a0 = 0 for ai ∈ A then x ∈ M := i=0 Ax which is a finitely generated A-module. By induction, for all m ≥ n, xm ∈ M which proves A[x] = M. Hence A[x] is finitely generated. (ii) =⇒ (iii) Let C = A[x]. Then C is finitely generated as an A-module by assumption. (iii) =⇒ (iv) Let M = C. Then since C is a subring, 1A ∈ C and hence C is a faithful A-module. (iv) =⇒ (i) Lemma 2.1.1 implies that f(x)M = 0 for some monic polynomial f(t) ∈ A[t], but if M is faithful, f(x) = 0. Hence x is integral over A.
Corollary 2.1.3. Let B/A be a ring extension and suppose x1, . . . , xn ∈ B are integral over A. Then A[x1, . . . , xn] is a finitely generated A-module. Proof. We prove this by induction on n. The base case for n = 1 follows from (ii) of Lemma 2.1.2. For n ≥ 2, suppose C := A[x1, . . . , xn] is finitely generated over A. Then there exist elements y1, . . . , y` ∈ C such that ` X C = A[x1, . . . , xn−1] = Ayi. i=1 Pm−1 j By hypothesis, xn is integral over A and therefore over C as well. Then C[xn] = j=0 Cxn for some m ∈ N. Thus m−1 ` X X j A[x1, . . . , xn] = C[xn] = Ayixn, j=0 i=1
so in particular, A[x1, . . . , xn] is a finitely generated A-module.
15 2.1 Integral Extensions of Rings 2 Integral Dependence
Definition. For a ring extension B/A, the set A0 of elements x ∈ B that are integral over A is called the integral closure of A in B. If A0 = A, we say A is integrally closed in B. Further, if A0 = B then B/A is called an integral extension.
Example 2.1.4. By Corollary 2.1.3, A[x1, . . . , xn] is an integral extension of A for any set of algebraic integers x1, . . . , xn over A. Lemma 2.1.5. Let B/A be a ring extension. Then A0, the integral closure of A in B, is a subring of B which contains A. Proof. It is trivial that A ⊆ A0, for every a ∈ A is a root of t − a ∈ A[t]. Suppose x, y ∈ A0. Then by Corollary 2.1.3, A[x, y] is a finitely generated A-module which is a subring of B. Clearly x + y, xy ∈ A[x, y] so applying Lemma 2.1.2(iii) shows that x + y and xy are integral over A. Proposition 2.1.6. If B/A and C/B are integral extensions of rings, then C/A is also an integral extension. Proof. Take x ∈ C. We must show that x is integral over A. First, x is integral over B so there exist elements b0, . . . , bn−1 ∈ B such that
n n−1 x + bn−1x + ... + b0 = 0.
0 0 Then since B/A is integral, x is integral over B := A[b0, . . . , bn−1]. By Corollary 2.1.3, B is finitely generated as an A-module since all the bi are integral over A. Then there are 0 0 0 0 elements y1, . . . , y` ∈ B which generate B as an A-module. Also, B [x] is a B -module which is finitely generated by xj, 0 ≤ j ≤ m − 1. Then we have
m−1 m−1 ` 0 X 0 j X X j A[x] ⊆ B [x] = B x = Ayix . j=0 j=0 i=1 Thus by Lemma 2.1.2(iii), x is integral over A. This proves C/A is an integral extension. Corollary 2.1.7. If A0 is the integral closure of A in B then A0 is integrally closed in B. Proof. Take x ∈ B which is integral over A0. Then A0[x]/A0 is integral by Corollary 2.1.3 and A0/A is integral by definition of A0. By Proposition 2.1.6, this implies A0[x]/A is integral, so in particular x is integral over A. Thus x ∈ A0. Definition. An integral domain A is called integrally closed (or alternatively, normal) if A is integrally closed in its field of fractions. Proposition 2.1.8. Every UFD is integrally closed. Proof. Let A be a UFD and denote the field of fractions of A by K. Take x ∈ K and write a x = b for a, b ∈ A and b 6= 0; we may assume a and b are relatively prime, that is, there is no prime p ∈ A that divides both a and b. Suppose x is integral over A. Then there are elements a0, . . . , an−1 ∈ A such that an an−1 + a + ... + a = 0. b n−1 b 0
16 2.1 Integral Extensions of Rings 2 Integral Dependence
Multiplying through by bn gives us
n n a + an−1b + ... + a0b = 0.
Subtracting an to the other side shows that b | an in A, but since a and b are relatively prime, and therefore an and b are also relatively prime, this is only possible when b is a unit. −1 a −1 Hence b exists in A, so x = b = ab ∈ A, proving A is integrally closed. Example 2.1.9. Both the integers Z and the Gaussian integers Z[i] are Euclidean domains, so in particular they are UFDs and therefore integrally closed by Proposition 2.1.8.
Example 2.1.10. Let K be√ a quadratic extension of Q, i.e. a field extension such that [K : Q] = 2. Then K = Q( d) for some squarefree integer d ∈ Z r {0, 1}, meaning d has prime factorization d = ±p1p2 ··· pr for primes pi. We want to compute the integral closure of Z in K. Since K/Q is Galois, there is one nontrivial Q-automorphism σ ∈ Gal(K/Q): σ : K −→ K √ √ a + b d 7−→ a − b d.
Using the norm and trace maps for K/Q, we have that
α ∈ K is integral over Z ⇐⇒ N(α) = ασ(α) and Tr(α) = α + σ(α) ∈ Z.
To verify this, suppose f(α) = 0 for a monic polynomial f(t) ∈ Z[t]. Then 0 = σ(f(α)) = f(σ(α)), so σ(α) is integral over Z. Since the integral elements form a ring, N(α) = ασ(α) and Tr(α) = α + σ(α) are also integral over Z. But the norm and trace maps take values in Q, so N(α), Tr(α) ∈ Q and they’re integral, so both lie in Z. Conversely, if ασ(α), α + σ(α) ∈ Z then α is a root of the monic polynomial
2 (t − α)(t − σ(α)) = t − (α + σ(α))t + ασ(α) ∈ Z[t]. Therefore α is integral. For any number field K/Q, we let OK denote the integral closure of Z in K, called the ring of integers of K. In the quadratic case, our work above shows that √ 2 2 OK = {a + b d | a, b ∈ Q and a − b d, 2a ∈ Z} √ ( [ d] if d ≡ 2, 3 (mod 4) Z √ = h 1+ d i Z 2 if d ≡ 1 (mod 4).
√ 1+ d To see how the second case arises, suppose d ≡ 1 (mod 4). Then ω = 2 satisfies the 2 1−d integral expression ω − ω + 4 = 0.
17 2.2 Integrality and Field Extensions 2 Integral Dependence
2.2 Integrality and Field Extensions
In this section, let A and B be integral domains and let K be the field of fractions of A.
Lemma 2.2.1. If B/A is an integral extension then A is a field if and only if B is a field.
Proof. ( =⇒ ) Assume A is a field and choose x ∈ B r {0}. Then x is integral over A so n n−1 there exist elements a0, . . . , an−1 ∈ A such that x + an−1x + ... + a0 = 0. This can be rewritten as n−1 n−2 x(x + an−1x + ... + a1) = −a0 (∗) Choose such an expression for x with n minimal. By assumption, x 6= 0 and since n was n−1 n−2 minimal, x + an−1x + ... + a1 6= 0. Then since B is an integral domain, (∗) implies −1 a0 6= 0, and since A is a field, a0 ∈ A exists. Divide through (∗) by −a0 to obtain
−1 −1 n−1 x := −a0 (x + ... + a1).
This shows x is invertible in B. ( ⇒ = ) Conversely, suppose B is a field and take a ∈ A r {0}. Then a−1 ∈ B exists. Since B/A is integral, there exist elements ai ∈ A such that
−n −(n−1) −1 n−1 a + an−1a + ... + a0 = 0 ⇐⇒ a = −(an−1 + ... + a0a ) ∈ A.
This shows that a is invertible in A.
Lemma 2.2.2. Let A be integrally closed in its fraction field K, let L/K be a finite extension and suppose α ∈ L is algebraic over K. Then α is integral over A if and only if α has a minimal polynomial f(t) ∈ A[t].
Proof. ( ⇒ = ) is the definition of integrality. ( =⇒ ) Take a splitting field E of f over K. Then f splits into linear factors:
n Y f(t) = (t − αi) for αi ∈ E, α1 = α. i=1 Since α is integral, there are K-isomorphisms
σi : K(α) −→ K(αi), α 7→ αi, ∼ using the fact that K(αi) = K[t]/(f(t)) for each root αi. By assumption, there exists a monic polynomial g(t) ∈ A[t] such that g(α) = 0. Applying σi to this equation gives
0 = σi(g(α)) = g(σi(α)) = g(αi).
Qn Thus αi is integral over A for all 1 ≤ i ≤ n. It follows that f(t) = i=1(t − αi) ∈ K[t] has coefficients which are integral over A. Finally, since A is integrally closed, this proves f ∈ A[t].
18 2.2 Integrality and Field Extensions 2 Integral Dependence
Remark. The most important rings in algebraic number theory are rings of integers: for a finite extension K/Q, OK denotes the integral closure of Z in K. By Lemma 2.2.2, we can write OK = {α ∈ K | fα ∈ Z[t]} where fα denotes the minimal polynomial of α over Q. Definition. For any finite field extension L/K of degree n = [L : K], there is a K-algebra homomorphism ∼ ϕ : L,→ EndK (L) = Mn(K)
α 7→ (ϕα : x 7→ αx).
The trace form of the extension L/K is defined as
Tr = TrL/K : L −→ K
α 7−→ tr(ϕα).
By properties of the matrix trace tr, the trace form TrL/K is K-linear.
Lemma 2.2.3. A finite extension L/K is separable if and only if TrL/K 6= 0. Proof. We will only show the forward direction. Suppose L/K is separable. Then L = K(α) for some α ∈ L by the primitive element theorem. We claim that for some 0 ≤ i ≤ n − 1, Tr(αi) 6= 0. Let n Y f(t) = (t − αj) j=1
be the minimal polynomial of α over K. By separability, f has distinct roots α = α1, . . . , αn in some splitting field E of f over K. By the Cayley-Hamilton theorem, the characteristic polynomial χ of ϕα satisfies 0 = χ(ϕ(α)) = ϕ(χ(α)). Thus χ(α) = 0 by injectivity of ϕ. Hence f(t) | χ(t) in K[t] but deg f = n = deg χ and both polynomials are monic, so we have f = χ. This implies that α1, . . . , αn are the distinct eigenvalues of ϕα in E. Let Mα be the Pn matrix representing ϕα in Mn(K). Then Tr(α) = tr(Mα) = j=1 αj. Taking powers of Mα, we have n i i X i Tr(α ) = tr(Mα) = αj. j=1 i Consider the n × n matrix C = (αj) where 0 ≤ i ≤ n − 1 and 1 ≤ j ≤ n. Then C ∈ Mn(E) and its determinant is a Vandermonde determinant:
Y 2 det C = (αj − α`) ` which is nonzero because the αj are distinct. In particular, 0 1 1 + ... + 1 Tr(1) . . . . . 6= C . = . = . . n−1 n−1 n−1 0 1 α1 + ... + αn Tr(α ) So Tr(αi) 6= 0 for some i. Therefore the trace form on L/K is nondegenerate. 19 2.2 Integrality and Field Extensions 2 Integral Dependence When Tr 6= 0, one may define a nondegenerate, symmetric, K-bilinear form b : L×L → K by b(x, y) = TrL/K (xy). Lemma 2.2.4. For any β ∈ L, the characteristic polynomial χ of ϕβ satisfies χ(t) = (f(t))[L:K(β)] where f(t) is the minimal polynomial of β over K. Therefore TrL/K (β) = [L : K(β)] TrK(β)/K (β). Proof. Let ϕβ be the map L → L given by ϕβ(x) = βx. Set m = [L : K(β)] and n = [K(β): K]. If m = 1, then L = K(β) so deg f(t) = [L : K] = n. By Cayley-Hamilton, f(t) divides χ(t), but both are by definition monic polynomials and deg χ(t) = n, so we have χ(t) = f(t). In the general case, let {x1, . . . , xr} be a basis of K(β)/K and let {y1, . . . , ym} be a basis of L/K(β). We know the set {xiyj : 1 ≤ i ≤ r, 1 ≤ j ≤ m} is a basis for L/K. Let B = (bk`) Pr be the standard matrix for ϕβ in the extension K(β)/K. Then βxi = k=1 bkixk and thus r X β(xiyj) = bki(xkyj). k=1 If we write the basis {xiyj} as {x1y1, x2y1, . . . , xry1, x1y2, . . . , xrym} then ϕβ has the following standard matrix in L/K: B 0 B .. 0 . B There are m blocks, each of which is B, so χ(t) = [det(B − tI)]m but by the m = 1 case, det(B − tI) = f(t). Therefore χ(t) = (f(t))m as claimed. Proposition 2.2.5. If A is an integrally closed integral domain with K its fraction field, and L/K is a finite separable extension, then for the integral closure B of A in L, there Pn exists a K-basis {v1, . . . , vn} of L such that B ⊆ i=1 Avi. Proof. If x ∈ L then we claim there is some a ∈ A r {0} such that ax ∈ B. Since x is algebraic, there exist q0, . . . , qn−1 ∈ K such that n n−1 x + qn−1x + ... + q0 = 0. Then a may be chosen to be the common denominator of the qi, for which we have n n−1 n (ax) + aqn−1(ax) + ... + a q0 = 0. Each coefficient of this expression now lies in A, so ax is integral over A and thus lies in B. Now this implies that there is a K-basis {u1, . . . , un} of L with ui ∈ B for each 1 ≤ i ≤ n. Since L/K is separable, Lemma 2.2.3 says that TrL/K 6= 0 and so b(·, ·) is nondegenerate. Then there exists a dual basis {v1, . . . , vn} of L with respect to b(·, ·). Let x ∈ B. Then we Pn can write x = j=1 cjvj for cj ∈ K. We finish by showing each cj ∈ A. Consider n ! n X X Tr(uix) = b(ui, x) = b ui, cjvj = cjb(ui, vj) = ci. j=1 j=1 20 2.3 Integrality, Ideals and Localization 2 Integral Dependence Now each ui, x ∈ B so uix ∈ B and thus uix is integral over A. Up to sign, Tr(uix) is a coefficient of the characteristic polynomial χ of ϕ(uix). By Lemma 2.2.4, χ is a power of the minimal polynomial of uix, which according to Lemma 2.2.2 has coefficients in A. Hence Pn χ ∈ A[t] so Tr(uix) = ci ∈ A. This shows that x ∈ i=1 Avi as desired. Corollary 2.2.6. If A is a PID and L/K is a finite separable extension, where K is the fraction field of A, then the integral closure B of A in L is a finitely generated free A-module of rank n = [L : K]. Proof. First, by Proposition 2.1.8, every PID is integrally closed so we may apply Proposi- tion 2.2.5 and its proof to produce K-bases {u1, . . . , un} and {v1, . . . , vn} of L such that n n X X Aui ⊆ B ⊆ Avi. i=1 i=1 The vi form a K-basis, so in particular they are linearly independent over A and thus Pn i=1 Avi is a free A-module of rank n. By the theory of modules over a PID, we have that B is free of rank at most n over A, but reversing the roles of the ui and vi, we see that the rank of B equals n. Example 2.2.7. If K/Q is finite, then there exists a Q-basis of K, say {α1, . . . , αn}, which is a Z-basis of OK . Such a basis is called an integral basis of K (or of OK ). In particular, Corollary 2.2.6 shows that OK is a free abelian group of rank√ n = [K : Q]. In the special case of a quadratic extension, K = Q( d), the ring of integers is OK = ⊕ ω, with Z Z (√ d if d ≡ 2, 3 (mod 4) √ ω = 1+ d 2 if d ≡ 1 (mod 4). Thus an integral basis of OK is {1, ω}. Remark. Proposition 2.2.5 and Corollary 2.2.6 are not true in general if L/K is not a separable extension, as we will show in Section 4.5. 2.3 Integrality, Ideals and Localization In this section we develop the theory of ideals in integral extensions and explore in particular how these notions interact with localization. We will prove the famous ‘going up’ and ‘going down’ theorems. Lemma 2.3.1. Assume B/A is a ring extension with I ⊂ A and J ⊂ B ideals such that J ∩ A = I. Then (a) B/J is integral over A/I. (b) If I and J are prime ideals then I is maximal in A if and only if J is maximal in B. 21 2.3 Integrality, Ideals and Localization 2 Integral Dependence Proof. (a) We have a sequence A,→ B B/J where the kernel of A → B/J is A ∩ J = I. Therefore A/I injects into B/J so the ring extension (B/J)/(A/I) makes sense. Take x¯ ∈ B/J withx ¯ = x + J for some x ∈ B. Then there exist ai ∈ A such that n n−1 x + an−1x + ... + a0 = 0 mod J n n−1 =⇒ x¯ +a ¯n−1x¯ + ... +a ¯0 = 0. The coefficients of the second equation are in A/I, so we see thatx ¯ is integral over A/I. (b) If I and J are prime ideals then A/I and B/J are integral domains. By (a), B/J is integral over A/I, and by Lemma 2.2.1, B/J is a field exactly when A/I is a field. Hence J is maximal in B if and only if I is maximal in A. Lemma 2.3.2. Let B/A be a ring extension and S ⊆ A be a multiplicatively closed subset. (a) If B/A is integral then S−1B/S−1A is integral. (b) If C is the integral closure of A in B then S−1C is the integral closure of S−1A in S−1B. −1 −1 a a Proof. First, there is a canonical map S A → S B, s 7→ s which is a well-defined, one- to-one ring homomorphism. As a consequence, we can regard S−1A as a subring of S−1B. x −1 (a) Let s ∈ S B where x ∈ B and s ∈ S. Given that x is integral over A, there exist a0, . . . , an−1 ∈ A such that n n−1 x + an−1x + ... + a0 = 0 in B. Applying the map j : B → S−1B to this expression and dividing by sn gives us xn a xn−1 a + n−1 + ... + 0 = 0 in S−1B. s s s sn −1 x −1 Since each coefficient lies in S A, this shows that s is integral over S A and thus the extension S−1B/S−1A is integral. (b) We have A ⊆ C ⊆ B so by the preliminary remark, we may consider the tower of ring extensions S−1A ⊆ S−1C ⊆ S−1B. Since C/A is integral, part (a) implies S−1C/S−1A is integral. If x ∈ S−1B is integral over S−1A, there exist fractions a0 ,..., an−1 ∈ S−1A such s s0 sn−1 that xn a xn−1 a + n−1 + ... + 0 = 0. s sn−1 s s0 Multiplying through by sn and the common denominator, we can write 0 n 0 n−1 0 −1 t x + an−1x + ... + a0 = 0 in S B 0 0 00 where t ∈ S and ai ∈ A. Then there is some t ∈ S such that 00 0 n 0 n−1 0 t (t x + an−1x + ... + a0) = 0 in B. Setting t = t0t00 ∈ S, we see that n 00 n−1 00 00 tx + an−1x + ... + a0 = 0 in B, where ai ∈ A. 22 2.3 Integrality, Ideals and Localization 2 Integral Dependence Multiplying through by tn−1 gives us n 00 n−1 00 n−1 (tx) + an−1(tx) + ... + a0t = 0. x tx −1 Therefore tx is integral over A, so tx ∈ C. This means s = ts ∈ S C, so we conclude that S−1C is the integral closure of S−1A in S−1B. Note that Lemma 2.3.2(b) does not imply that if B/A is an integral extension of rings with Q ⊂ B prime and Q ∩ B = p, then BQ/Ap is integral. Lemma 2.3.3. Suppose B/A is an integral extension of rings. If Q, Q0 ⊂ B are prime ideals with Q ⊆ Q0 and Q ∩ A = Q0 ∩ A, then Q = Q0. Proof. Set p = Q ∩ A = Q0 ∩ A. Then p is a prime ideal of A. If S = A r p then by −1 −1 Lemma 2.3.2(a), S B/Ap is an integral extension. By Corollary 1.3.9, S p is the unique 0 0 maximal ideal of the local ring Ap. Notice that S ∩ Q = S ∩ A ∩ Q = S ∩ p = ∅, and likewise S ∩Q = ∅. Therefore S−1Q ⊆ S−1Q0 are prime ideals of S−1B by Proposition 1.3.8. 1 −1 0 −1 0 −1 −1 −1 0 −1 −1 Also, 1 6∈ S Q so S Q ∩ S A 6= S A. Thus S Q ∩ S A = S p, and likewise S−1Q ∩ S−1A = S−1p. Since S−1p is maximal in S−1A, Lemma 2.3.1(b) implies S−1Q and S−1Q0 are both maximal ideals in S−1B. Therefore S−1Q = S−1Q0 so Proposition 1.3.8(c) shows that we must have Q = Q0. Lemma 2.3.4. Suppose B/A is an integral extension of rings and p ⊂ A is a prime ideal. Then there exists a prime ideal Q ⊂ B with Q ∩ A = p. Proof. Let S = A r p and consider the commutative diagram A B i j −1 −1 Ap = S A S B Choose any maximal ideal m ⊂ S−1B and set Q = j−1(m) which is a prime ideal of B by Proposition 1.3.8. We claim that Q ∩ A = p. Since B is integral over A, Lemma 2.3.2(a) −1 −1 −1 implies S B is integral over S A = Ap. Then by Lemma 2.3.1(b), m maximal in S B −1 implies m ∩ Ap is maximal in Ap. But Ap is a local ring, so m ∩ Ap = S p. Since the first diagram commutes, we have another commutative diagram: Q ∩ A Q i j −1 S p = m ∩ Ap m −1 −1 −1 −1 Therefore Q ∩ A = j (m) ∩ A = i (m ∩ Ap) = i (S p). Finally, Proposition 1.3.8(b) shows that Q ∩ A = p as desired. 23 2.3 Integrality, Ideals and Localization 2 Integral Dependence Lemmas 2.3.3 and 2.3.4 together show that in an integral extension B/A, any prime ideal p ⊂ A lifted to B is contained in a prime ideal of B which is unique in any chain of ideals of B. This generalizes in an important way, in what is known as the going up theorem. Theorem 2.3.5 (Going Up). Suppose B/A is an integral extension of rings and p0 ( p1 ( ··· ( pn is a chain of prime ideals in A, in which each inclusion is strict. Then for any chain of prime ideals Q0 ( Q1 ( ··· ( Qm−1 in B such that 0 ≤ m ≤ n and Qi ∩ A = pi for each 0 ≤ i ≤ m − 1, there exists a prime ideal Qm ⊂ B such that Qm−1 ( Qm and Qm ∩ A = pm. Proof. When m = 0, Lemma 2.3.4 says there exists a prime ideal Q0 ⊂ B with Q0 ∩ A = p0. Now to induct, suppose the theorem holds for some m, 1 ≤ m ≤ n. Then Qm−1 ∩ A = pm−1 so consider the extension of quotient rings ¯ ¯ A = A/pm−1 ,−→ B = B/Qm−1. ¯ ¯ By Lemma 2.3.1(a), B/A is an integral extension. By the correspondence theorem, p¯m := ¯ ¯ pm/pm−1 is a prime ideal of A. Then by Lemma 2.3.4, there exists a prime ideal Qm ⊂ B ¯ −1 ¯ such that Qm ∩A = p¯m. Let Qm be the preimage π (Qm) where π : B → B is the canonical projection. Clearly Qm ) Qm−1 since p¯m 6= 0. Finally, the above work implies Qm ∩A = pm. Hence the statement holds by induction. Definition. If A is a ring and p ⊂ A is prime, the height of p is defined as ht(p) = sup{` ≥ 0 | there is a chain of prime ideals p0 ( p1 ( ··· ( p`−1 ( p}. The Krull dimension of A is then defined by dim A = sup{ht(p) | p ⊂ A is prime}. Remark. If p is not maximal then p ( m for some maximal ideal m ⊂ A, so we can reformulate the Krull dimension definition in one of the following ways: dim A = sup{ht(m) | m ⊂ A is maximal} = sup{` ≥ 0 | p0 ( p1 ( ··· ( p` is a chain of prime ideals}. Examples. 1 If K is a field then dim K = 0 since (0) is the only prime ideal. 2 If A is an integral domain, then dim A = 0 if and only if A is a field. However, there exist non-domains which have dimension 0. 3 If A is a PID but not a field, then all nonzero prime ideals are of the form (π) for a prime element π ∈ A. This shows that dim A = 1. 24 2.3 Integrality, Ideals and Localization 2 Integral Dependence Lemma 2.3.6. Suppose p ⊂ A is a prime ideal. Then (a) ht(p) = dim Ap. (b) ht(p) + dim A/p ≤ dim A. Proof. (a) By Proposition 1.3.8(c), prime ideals of Ap are in bijective, inclusion-preserving correspondence with primes Q ⊂ A intersecting S = A r p trivially, that is, primes Q ⊆ p. Therefore any chain p0 ( p1 ( ··· ( pn−1 ( p of prime ideals in A corresponds to a −1 −1 −1 −1 −1 chain S p0 ( S p1 ( ··· ( S pn−1 ( S p of prime ideals in Ap = S p; this proves −1 dim Ap ≥ ht(p). Conversely, any chain Q0 ( Q1 ( ··· ( Q`−1 ( S p of prime ideals in −1 −1 −1 Ap corresponds to a chain j (Q0) ( j (Q1) ( ··· ( j (Q`−1) ( p in A; this establishes ht(p) ≥ dim Ap. Hence ht(p) = dim Ap. (b) Suppose dim A/p = ` and let Q0 ( Q1 ( ··· ( Q` be a maximal chain of prime ideals in A/p. By the correspondence theorem, this lifts to a chain of prime ideals Q0 ( Q1 ( ··· ( Q` (∗) in A such that each Qi ⊇ p. Of course we can always add p to the bottom of such a chain, so if ` is maximal, we must have Q0 = p. Now for any chain of prime ideals p0 ( p1 ( ··· ( pn−1 ( p in A, we can extend (∗) by this new chain, giving p0 ( p1 ( ··· ( pn−1 ( p = Q0 ( Q1 ( ··· ( Q`, a chain of prime ideals in A. This shows dim A ≥ n + `. In particular, this holds for any such n ≤ ht(p), so we have dim A ≥ ht(p) + ` = ht(p) + dim A/p. Theorem 2.3.7. If B is an integral extension of A then dim A = dim B. Proof. If Q0 ( Q1 ( ··· ( Q` is a chain of prime ideals in B, then Q0 ∩ A ( Q1 ∩ A ( ··· ( Q` ∩ A is a chain of prime ideals in A – the strictness of each containment comes from Lemma 2.3.3. This implies dim B ≤ dim A. On the other hand, if p0 ( p1 ( ··· ( p` is a chain of prime ideals in A, then by the going up theorem, we get a chain of prime ideals Q0 ( Q1 ( ··· ( Q` in B, with Qi ∩ A = pi for each i. Hence dim A ≤ dim B so we have dim A = dim B. There is an analog of the going up theorem for descending chains of prime ideals, but it requires additional conditions. First we prove two important lemmas. Lemma 2.3.8. For a ring extension B/A and any prime ideal p ⊂ A, there exists a prime ideal Q ⊂ B with the property that Q ∩ A = p if and only if pB ∩ A ⊆ p. 25 2.3 Integrality, Ideals and Localization 2 Integral Dependence Proof. ( =⇒ ) If Q ⊂ B exists and satisfies Q ∩ A = p, then pB ∩ A = (Q ∩ A)B ∩ A ⊆ Q ∩ A = p. ( ⇒ = ) Set T = A r p. By assumption, pB ∩ A ⊆ p implies (pB ∩ A) ∩ T = ∅. Therefore T −1(pB) 6= T −1B so there is a maximal ideal m ⊂ T −1B such that T −1(pB) ⊆ m. By Proposition 1.3.8, there is a prime ideal Q ⊂ B such that Q ∩ T = ∅ and m = T −1Q. Then p ⊆ pB ∩ A ⊆ Q ∩ A = Q ∩ p ⊆ p so it follows that p = Q ∩ A as required. Lemma 2.3.9. Suppose A and B are integral domains with B/A an integral extension of rings, K the field of fractions of A and A integrally closed (in K). If p ⊂ A is a prime ideal of A and α ∈ pB is any element, then the nonleading coefficients of the minimal polynomial f of α over K lie in p. Proof. By Lemma 2.2.2, the coefficients of f must lie in A. Let α = b1p1 + ... + bkpk for bi ∈ B, pi ∈ p. Replacing B with A[b1, . . . , bk], we may assume B is a finitely generated Pn A-module generated by x1, . . . , xn. For each 1 ≤ i ≤ n, write αxi = j=1 aijxj with aij ∈ p. Pn Then j=1(δijαxi −aij) = 0. Multiplying by the classical adjoint of the matrix (δijαxi −aij), we get det(δijαxi − aij)xj = 0 for all 1 ≤ j ≤ n. Therefore αxi satisfies a monic polynomial with coefficients in p. Since this holds for every generator xi, it follows that α itself satisfies such a polynomial, say g ∈ p[t]. If f is the minimal polynomial of α over K, f must divide g, say g = fh for h ∈ A[t]. Reducing modulo p, we get tdeg g =g ¯ = f¯h¯ ∈ (A/p)[t]. Since p is prime, A/p is a domain, so f¯ and h¯ must also be powers of t. Thus all nonleading terms of f lie in p. Theorem 2.3.10 (Going Down). Suppose B/A is an integral extension of integral domains with A integrally closed. If p0 ) p1 ) ··· ) pn is a chain of prime ideals in A and Q0 ) Q1 ) ··· ) Qm−1 is a chain of prime ideals in B such that 0 ≤ m ≤ n and Qi ∩ A = p for each i ≤ m − 1, then there is a prime ideal Qm ⊂ B such that Qm ( Qm−1 and Qm ∩ A = pm. Proof. If m = 0, this is given to us by Lemma 2.3.4. Otherwise assume 1 ≤ m ≤ n. Then 0 −1 since B is an integral domain, the map B → BQm−1 is injective. Set B = BQm−1 = S B, 0 where S = B r Qm−1. Then we get a tower of ring extensions A ⊆ B ⊆ B . It suffices to 0 0 0 show there exists a nonzero prime ideal Q ⊂ B such that Q ∩A = pm, for then we will have 0 0 pm = Q ∩ A = (Q ∩ B) ∩ A = Qm ∩ A for some prime ideal Qm ⊂ B, by Proposition 1.3.8. 0 0 0 Set p = pm. By Lemma 2.3.8, showing the existence of Q ⊂ B satisfying Q ∩ A = p is equivalent to showing pB0 ∩ A ⊆ p. If pB0 ∩ A = 0, we are done. Otherwise, take a nonzero element a ∈ pB0 and write a = αs for some α ∈ pB and s ∈ S. Also let f be the minimal 26 2.3 Integrality, Ideals and Localization 2 Integral Dependence n n−1 polynomial of α over K. Then Lemma 2.3.9 implies f = t + cn−1t + ... + c0 for ci ∈ p. Since a ∈ A, we may write 1 c c f(at) = tn + n−1 tn−1 + ... + 0 = tn + c0 tn−1 + ... + c0 , an a an n−1 0 0 ci 1 with ci = an−i . Then an f(at) is also irreducible, and plugging in t = s gives 1 1 f(as) = f(α) = 0. an an 1 0 It follows that an f(at) is the minimal polynomial of s over K. Thus each ci lies in A for 0 i 0 0 ≤ i ≤ n − 1. Suppose that a 6∈ p. Then since p is prime, cn−ia = cn−i ∈ p implies cn−i ∈ p for each 0 ≤ i ≤ n − 1. Since s ∈ B, we get n 0 n−1 0 0 s = −cn−1s − ... − c0 ∈ pB ⊆ pm−1B ⊆ Qm−1 by hypothesis. Then s ∈ Qm−1 since Qm−1 is a prime ideal, but this contradicts s ∈ S = 0 B r Qm−1. Hence a ∈ p, implying pB ∩ A ⊆ p after all. For the remainder of the section, let A be an integral domain with field of fractions K. Whenever S ⊆ Ar{0}, we can consider the local ring S−1A as a subring of K. In particular, Ap ⊆ K for every prime p ⊂ A. \ Lemma 2.3.11. Any integral domain A can be written A = Am. maximal ideals m⊂A Proof. Let D be the intersection of Am over all maximal ideals m ⊂ A. Then D ⊆ K. For a fixed x ∈ D, consider the ideal I = {a ∈ A | ax ∈ A} ⊆ A. If m is a fixed maximal ideal b of A, by assumption x ∈ Am so there exist elements a, b ∈ A with a 6∈ m such that x = a . It follows that ax = b ∈ A, so a ∈ I. Since we took a 6∈ m, this implies I 6⊂ m. Now m was arbitrary, so we must have I = A. Thus 1 ∈ I, implying x = 1 · x ∈ A and therefore A = D as required. Proposition 2.3.12. For an integral domain A, the following are equivalent: (i) A is integrally closed. (ii) Ap is integrally closed for every prime ideal p ⊂ A. (iii) Am is integrally closed for every maximal ideal m ⊂ A. Proof. (i) =⇒ (ii) Take p ⊂ A to be a prime ideal and set S = A r p. By assumption, A −1 is the integral closure of A in K, and by Lemma 2.3.2(b), S A = Ap is integrally closed in K, which is also the field of fractions of Ap. (ii) =⇒ (iii) is trivial. (iii) =⇒ (i) If m ⊂ A is a maximal ideal, we have A ⊆ Am ⊆ K. Then the assumption that Am is integrally closed implies that the integral closure of A in K lies in Am for any such m. Therefore by Lemma 2.3.11, A is integrally closed. 27 2.4 Normalization 2 Integral Dependence 2.4 Normalization Let k be a field and A a finitely generated k-algebra, but not necessarily an integral domain. Definition. Elements a1, . . . , an ∈ A are said to be algebraically dependent over k if there exists a nonzero polynomial f ∈ k[t1, . . . , tn] such that f(a1, . . . , an) = 0. Otherwise the ai are said to be algebraically independent over k. If a1, . . . , an ∈ A are algebraically independent, then the k-algebra homomorphism ϕ : k[t1, . . . , tn] −→ A ti 7−→ ai ∼ is injective. As a consequence, k[a1, . . . , an] = k[t1, . . . , tn] as k-algebras. Theorem 2.4.1 (Noether’s Normalization Lemma). If A is a finitely generated k-algebra with A = k[x1, . . . , xn] for xi ∈ A, then there exist algebraically independent elements y1, . . . , yr ∈ A, 0 ≤ r ≤ n, such that A/k[y1, . . . , yr] is an integral extension. Moreover, Pn if k is infinite, we may choose the yi in j=1 kxj. Proof. We prove the theorem by induction on n. Starting with n = 1, there are two possi- bilities: x1 is transcendental over k, in which case we may take r = 1 and y1 = x1. x1 is algebraic over k, in which case we may take r = 0, so that k[x1]/k is integral. Now suppose n ≥ 2. If x1, . . . , xn are already algebraically independent, take r = n and yi = xi for each 1 ≤ i ≤ n. If not, we claim that there exist elements z1, . . . , zn−1 ∈ A (and Pn zi ∈ j=1 kxj if k is infinite) such that the following hold: (i) k[z1, . . . , zn−1, xn] = A. (ii) xn is integral over k[z1, . . . , zn−1]. Once we prove the claim, the proof finishes as follows. Set B = k[z1, . . . , zn−1] ⊆ A. Then A = B[xn] and A/B is integral by (i) and (ii). By induction, there exist y1, . . . , yr ∈ B (and Pn−1 Pn if k is infinite, yi ∈ j=1 kzj ⊆ j=1 kxj) such that the yi are algebraically independent over k and B is integral over k[y1, . . . , yr]. It will then follow that A/k[y1, . . . , yr] is an integral extension by Proposition 2.1.6. To prove the existence of z1, . . . , zn−1 satisfying (i) and (ii), assume k is infinite. By assumption there is a nonzero polynomial f ∈ k[t1, . . . , tn] such that f(x1, . . . , xn) = 0. Pd Write f = i=0 fi(t1, . . . , tn) where each fi is homogeneous of (total) degree i and fd 6= 0. (Since f0 ∈ k, we know d ≥ 1.) We can write fd in two ways: X j1 jn fd(t1, . . . , tn) = cJ t1 ··· tn where cJ ∈ k n J=(j1,...,jn)∈N0 j1+...+jn=d d X d−j fd(t1, . . . , tn) = Gj(t1, . . . , tn−1)tn where each Gj is homogeneous of degree i. j=0 28 2.4 Normalization 2 Integral Dependence Since fd 6= 0, not all of the Gj are 0. This implies fd(t1, . . . , tn−1, 1) 6= 0. Since k is infinite, there exist scalars λ1, . . . , λn−1 ∈ k such that fd(λ1, . . . , λn−1, 1) 6= 0. Plugging these into the top expression for fd, we get X j1 jn−1 j1+...+jn fd(λ1xn, . . . , λn−1xn, xn) = cJ λ1 ··· λn−1 xn J X j1 jn−1 d d = cJ λ1 ··· λn−1 xn = fd(λ1, . . . , λn−1, 1)xn. J Now set zi = xi − λixn for each 1 ≤ i ≤ n − 1. Then k[z1, . . . , zn−1, xn] = k[x1, . . . , xn] and d X 0 = f(x1, . . . , xn) = fj(x1, . . . , xn) j=0 d X = fj(z1 + λ1xn, . . . , zn−1 + λn−1xn, xn) j=0 = fd(z1 + λ1xn, . . . , zn−1 + λn−1xn, xn) + lower degree terms X j1 jn−1 jn = cJ (z1 + λ1xn) ··· (zn−1 + λn−1xn) xn + lower degree terms J = fd(λ1xn, . . . , λn−1xn, xn) + lower terms in xn d−1 d X j = fd(λ1, . . . , λn−1, 1)xn + Hj(z1, . . . , zn−1)xn j=1 d × where fd(λ1, . . . , λn−1, 1)xn ∈ k and each Hj ∈ k[t1, . . . , tn−1]. This shows xn satisfies a nontrivial monic polynomial equation with coefficients in B = k[z1, . . . , zn−1], since in particular we can divide out by fd(λ1, . . . , λn−1, 1) 6= 0. Thus xn is integral over B. If k is finite, we need a slightly different argument. By the assumption on x1, . . . , xn, there is a nonzero polynomial F ∈ k[t1, . . . , tn] such that F (x1, . . . , xn) = 0. Write X j1 jn F (t1, . . . , tn) = cjt1 ··· tn n j∈N0 n for cj ∈ k for all j = (j1, . . . , jn) ∈ N0 . Set J = {j | cj 6= 0}, which is a finite set since F is a n polynomial. We claim there exists a tuple m = (m1, . . . , mn) ∈ N0 such that for any distinct j, j0 ∈ J, jm 6= j0m. To justify such a claim, let N ∈ N be greater than any component n−1 n−2 ji of any tuple j = (j1, . . . , jn) ∈ J. Then m = (N ,N ,...,N, 1) satisfies the desired conclusion, since for N large enough, every base N expansion of a tuple j ∈ J is unique. mi For each 1 ≤ i ≤ n − 1, set zi = xi − xn . Then k[z1, . . . , zn−1, xn] = k[x1, . . . , xn] so (i) 29 2.4 Normalization 2 Integral Dependence is satisfied. Moreover, we have X j1 jn−1 jn 0 = F (x1, . . . , xn) = cjx1 ··· xn−1 xn j∈J X m1 j1 mn−1 jn−1 jn = cj(z1 + xn ) ··· (zn−1 + xn ) xn j∈J X m1j1 mn−1jn−1 jn = cjxn ··· xn xn + lower terms in xn j∈J X m1j1+...+mn−1jn−1+mnjn = cjxn + lower terms in xn j∈J X jm = cjxn + lower terms, j∈J where the lower terms at the end have total degree strictly bounded by m. By our choice jm of m, we have that the powers xn over all j ∈ J are distinct, so in particular there is a ∗ ∗ ∗ maximum exponent j m, j ∈ J and for this j , cj∗ 6= 0. This proves xn satisfies a nonzero polynomial in k[z1, . . . , zn−1][t] which may be made monic by dividing out by cj∗ 6= 0. Hence xn is integral over k[z1, . . . , zn−1]. Remark. If A is an integral domain with field of fractions K = k(x1, . . . , xn), the r in the normalization lemma equals tr degk K. We use Noether’s normalization lemma to prove the fundamental theorem in algebraic ge- ometry, Hilbert’s Nullstellensatz. This first version is stated in terms of algebraic extensions of fields; we will prove other versions later. Theorem 2.4.2 (Hilbert’s Nullstellensatz, Algebraic Version). If K/k is a field extension such that K is a finitely generated k-algebra, then K/k is a finite algebraic extension. Proof. Noether’s normalization lemma says that there are yi ∈ K such that K/k[y1, . . . , yr] is integral. Then by Lemma 2.2.1, K a field implies k[y1, . . . , yr] is also a field. However, ∼ k[y1, . . . , yr] = k[t1, . . . , tr] which is not a field unless r = 0. Thus no such yi exist, so K/k is an integral extension, which is obviously equivalent to K/k being algebraic. Finally, the extension is of finite degree because K is finitely generated as a k-algebra. n Definition. For k a field and a = (a1, . . . , an) ∈ k , the ideal associated to a is ma := (t1 − a1, . . . , tn − an) ⊂ k[t1, . . . , tn]. Remark. Notice that for each 1 ≤ i ≤ n, ti ≡ ai (mod ma), so for any polynomial F ∈ k[t1, . . . , tn], F (t1, . . . , tn) ≡ F (a1, . . . , an) (mod ma). Therefore k[t1, . . . , tn] = k + ma. This defines an evaluation homomorphism ϕa : k[t1, . . . , tn] −→ k F (t1, . . . , tn) 7−→ F (a1, . . . , an). 30 2.4 Normalization 2 Integral Dependence One can check that ϕa is a k-algebra homomorphism. Clearly ma ⊆ ker ϕa, and by k[t1, . . . , tn] = n k + ma, we get ma = ker ϕa and this is a maximal ideal of k[t1, . . . , tn]. Also, if a 6= b ∈ k then ma 6= mb. This defines an injection n k ,−→ MaxSpec(k[t1, . . . , tn]) a 7−→ ma where MaxSpec(A) = {maximal ideals of A} for any ring A. This mapping is not surjective in general, but in the case that k is algebraically closed, the Nullstellensatz (2.4.2) implies that it is surjective. Corollary 2.4.3. If k is algebraically closed, then the mapping n k −→ MaxSpec(k[t1, . . . , tn]) is bijective. Proof. We saw that if a 6= b then ma 6= mb, so it remains to show surjectivity. Let m be a maximal ideal of A := k[t1, . . . , tn]. Then the quotient ring K = A/m is a field which is finitely generated as a k-algebra by images of the monomials t¯i = ti + m in A/m. By the Nullstellensatz (2.4.2), K/k is algebraic, but since k was assumed to be algebraically closed, this implies K = k. Therefore for all 1 ≤ i ≤ n, there exists an ai ∈ k such that t¯i = ai =a ¯i in K. Equivalently, ti ≡ ai (mod m) so ti − ai ∈ m for all i. This proves ma ⊆ m for a = (a1, . . . , an), but we saw that ma is always maximal. Thus ma = m. We conclude that kn → A is both surjective and injective. By Corollary 2.4.3, if k is algebraically closed, every proper ideal I ⊂ k[t1, . . . , tn] has n a common zero in k , called a Nullstelle in German. Explicitly, I ⊆ ma for some maximal n ideal ma defined by a point a ∈ k . Lemma 2.4.4. If A is a ring in which every prime ideal P which is not maximal is an intersection of prime ideals properly containing P , then every prime ideal is an intersection of maximal ideals. Proof. Suppose to the contrary that p ⊂ A is a prime ideal that is not an intersection of maximal ideals. In particular, p is not maximal so by hypothesis, \ p = {Q ⊂ A | Q prime, Q ) p}. Consider the integral domain B = A/p. Then J(B) 6= 0, where J(B) denotes the Jacobson radical of B, so there is a nonzero element x ∈ J(B). Let S = {xn | n ≥ 0} be the −1 multiplicatively closed subset of powers of x in B, and consider the localization Bx = S B. Since p is prime, A/p is reduced, i.e. N(A/p) = 0, but this means x 6∈ N(B). In particular, x is not nilpotent, so Bx is nonzero. Let m ⊂ Bx be a maximal ideal and let Q be its corresponding prime ideal in B. If Q were itself maximal, we would have x ∈ Q but then Q ∩ S 6= ∅. Therefore Q is not maximal, but for any prime ideal Q0 ) Q, there is no 0 0 corresponding prime in Bx, so Q ∩S must be nonempty. In particular, x ∈ Q for every prime ideal Q0 ) Q. However, the assumption on p implies every non-maximal prime in B = A/p is T 0 0 an intersection of prime ideals properly containing it. Thus Q = {Q ⊂ B prime | Q ) Q}, but then x ∈ Q, a contradiction. Hence no such p exists in A, so every prime ideal in A is an intersection of maximal ideals. 31 2.5 Valuation Rings 2 Integral Dependence Corollary 2.4.5. If A is a finitely generated k-algebra, for k a field, every prime ideal of A is an intersection of maximal ideals. Proof. By Lemma 2.4.4, it suffices to show that for every prime ideal p ⊂ A which is not T maximal, p = {Q ⊂ A | Q prime, Q ) p}. Suppose to the contrary that there exists such T a prime p with p ( {Q ⊂ A | Q prime, Q ) p}. Let B = A/p. Then B is not a field, and by assumption the intersection of all nonzero prime ideals in B is nonzero. In particular this intersection contains a nonzero element f, which as in Lemma 2.4.4 is not nilpotent. n b o Consider the (nonzero) localization Bf = f n : b ∈ B, n ≥ 0 . Then Bf is finitely gener- ated as a k-algebra since A and B are. Suppose Bf is a field. Then by Nullstellensatz (2.4.2), Bf is an algebraic extension of k. In particular, Bf /k is integral, so Bf is integral over B as well. Then Lemma 2.2.1 implies Bf and B are fields simultaneously, but this contradicts the fact that B is not a field. Therefore Bf cannot be a field, so it has a nonzero prime ideal 0 0 m ⊂ Bf . This pulls back along j : B → Bf to a prime ideal m ⊂ B. Then f 6∈ m but clearly this is impossible, since f was chosen to lie in the intersection of all prime ideals of B. Thus no such prime p ⊂ A can exist, so every non-maximal prime ideal of A is equal to the intersection of all prime ideals properly containing it. 2.5 Valuation Rings Definition. An integral domain A with field of fractions K is called a valuation ring if x ∈ A or x−1 ∈ A for every x ∈ K×. Example 2.5.1. If A is a UFD and p ∈ A is a prime element, then the localization A(p) = a c × b : a, b ∈ A, p - b is a valuation ring. To see this, take d ∈ K with c, d ∈ A r {0}. By c unique factorization, we may assume d is in lowest terms, i.e. gcd(c, d) = 1. Then if p - d, c c −1 d we have d ∈ A(p). On the other hand, if p | d, we must have p - c so d = c ∈ A(p). Theorem 2.5.2. Every valuation ring A is a local ring and is integrally closed. Proof. We first prove A is a local ring. Let A× be the set of units in A and let m = A r A×. One can prove that m is an ideal of A, and since every element outside of m is a unit, m must be the unique maximal ideal of A. To prove A is integrally closed, assume x ∈ K× is integral over A. Then there exist n n−1 −1 a0, . . . , an−1 ∈ A such that x + an−1x + ... + a0 = 0. If x ∈ A, we can multiply the polynomial expression through by x−(n−1) to obtain −1 −(n−1) x = −(an−1 + an−2x + ... + a0x ). Since x−1 ∈ A, the right side lies in A and thus x ∈ A. This shows A is integrally closed. Theorem 2.5.3. For K a field and A ⊆ K a nonzero subring, the integral closure of A in K is equal to \ A = B. valuation rings B A⊆B⊆K We will see in Section 4.1 how valuation rings arise naturally as subrings of a field K. 32 2.6 Dimension and Transcendence Degree 2 Integral Dependence 2.6 Dimension and Transcendence Degree Question. How does one determine the dimension of a finitely generated k-algebra? Example 2.6.1. Consider A = k[t1, . . . , tn]. We would like to conclude that dim A = n, but this is not obvious from the definition of Krull dimension. Clearly dim A ≥ n, since 0 ( (t1) ( (t1, t2) ( ··· ( (t1, . . . , tn) is a chain of prime ideals in A. We will show the opposite inequality shortly. Assume C and D are integral domains, with D/C a ring extension and K = Frac(C) and L = Frac(D) their fields of fractions. Then L/K is a field extension. Define the transcendence degree of D over C to be tr degC D := tr degK L. If C = k is a field, we have tr degk D = tr degk L = sup{|E| : E ⊂ D,E is algebraically independent over k}. Lemma 2.6.2. Assume A and B are finitely generated k-algebras which are integral do- mains, and there is a surjective k-algebra homomorphism π : B → A with ker π 6= 0. Then tr degk A < tr degk B. Proof. Set r = tr degk B and s = tr degk A. Since A and B are finitely generated, r and s are each finite. If a1, . . . , as ∈ A are algebraically independent over k, choose b1, . . . , bs ∈ B such that π(bi) = ai for each i. Then for any polynomial F ∈ k[t1, . . . , ts] vanishing at (b1, . . . , bs), we have 0 = π(F (b1, . . . , bs)) = F (π(b1), . . . , π(bs)) = F (a1, . . . , as). Since the ai are algebraically independent over k, we must have F = 0, so s ≤ r. To show s < r strictly, we demonstrate that any set of r elements in A is algebraically dependent over k. Let a1, . . . , ar ∈ A be given. Again choose bi ∈ B such that π(bi) = ai for 1 ≤ i ≤ r and take b0 ∈ ker π r {0}. Since tr degk B = r, there is a nonzero polynomial F ∈ k[t0, t1, . . . , tr] such that F (b0, b1, . . . , br) = 0. Choose F of minimal degree in t0. Then d X i F (t0, t1, . . . , tr) = Fi(t1, . . . , tr)t0 for Fi ∈ k[t1, . . . , tr]. i=1 Pd i−1 We claim F0 6= 0; otherwise, F = t0 i=1 Fi(t1, . . . , tr)t0 and evaluating in the integral domain B implies d d X i−1 X i−1 0 = b0 Fi(b1, . . . , br)b0 =⇒ Fi(b1 . . . , br)b0 = 0. i=1 i=1 This is a nonzero polynomial of strictly smaller degree in t0 that vanishes on (b0, . . . , br), so we in fact must have F0 6= 0. Now plug in the bj and apply π to obtain d ! d X i X i 0 = π Fi(b1, . . . , br)b0 = Fi(a1, . . . , ar)π(b0) = F0(a1, . . . , ar). i=1 i=1 Since F0 6= 0, we see that a1, . . . , ar are algebraically dependent over k. 33 2.6 Dimension and Transcendence Degree 2 Integral Dependence Corollary 2.6.3. For any n ≥ 1, dim k[t1, . . . , tn] = n. Proof. Set A = k[t1, . . . , tn]. We showed in Example 2.6.1 that dim A ≥ n. Let p0 ( p1 ( ··· ( p` be a chain of prime ideals in A. Then we have a well-defined sequence of surjections A/p0 A/p1 ··· A/p` with all kernels nonzero since pi ( pi+1 for all i. By Lemma 2.6.2, n ≥ tr degk A/p0 > tr degk A/p1 > ··· > tr degk A/p` ≥ 0. In this sequence, there are ` strict inequalities, so ` ≤ n. Further, ` was arbitrary so dim A = n. Theorem 2.6.4. If A is a finitely generated k-algebra which is an integral domain, then dim A = tr degk A. Proof. By Noether’s normalization lemma, we can pick y1, . . . , yr ∈ A which are algebraically independent over k such that A/k[y1, . . . , yr] is an integral extension of rings. Then dim A = dim k[y1, . . . , yr] = r by Theorem 2.3.7. Now for all x ∈ A, x is integral over k[y1, . . . , yr]. If K is the field of fractions of A, then every z ∈ K is algebraic over the field k(y1, . . . , yr), but then Lemma 1.4.3(c) shows that {y1, . . . , yr} is a transcendence basis of K/k. Hence tr degk A = tr degk K = r = dim A. Definition. We say a ring A of finite Krull dimension has the strong dimension property (SDP) if whenever p0 ( ··· ( p` is a maximal (non-refinable) chain of prime ideals in A, we have ` = dim A. Lemma 2.6.5. Assume A has SDP and n = dim A. Then (1) ht(p) + dim A/p = dim A for any prime ideal p ⊂ A. (2) ht(m) = n for all maximal ideals m ⊂ A. (3) If p0 ( p1 ( ··· ( p` is a chain of prime ideals in A which cannot be refined, then ` = ht(p`) − ht(p0). (4) If p ⊂ A is a prime ideal then A/p has SDP. Proof. (1) Given a prime ideal p ⊂ A with ht(p) = m, choose a chain p0 ( ··· ( pm−1 ( p of prime ideals in A. If dim A/p = d then there is a maximal chain 0 = Q0 ( Q1 ( ··· ( Qd 34 2.6 Dimension and Transcendence Degree 2 Integral Dependence of prime ideals in A/p. Then we may form a chain p0 ( ··· ( pm−1 ( p = Q0 ( Q1 ( ··· ( Qd where Qi ⊂ A are the prime ideals such that Qi = Qi/p. This is a maximal, non-refinable chain of primes in A, so by SDP, dim A = m + d = ht(p) + dim A/p. (2) follows immediately from (1) since A/m is a field and dim F = 0 for any field F . (3) follows from a similar argument to the proof of (1). Given such a chain of primes p0 ( ··· ( p` we can extend this to a maximal chain of primes in A: Q0 ( ··· ( Qj = p0 ( ··· ( P` ( p`+1 ( ··· ( pm. Then SDP implies n = j + m. When such a chain is chosen so that j = ht(p0) and m − ` = dim A/p`, we must have n = j + m = j + ` + (m − `) = ht(p0) + ` + dim A/p`. However by (1), n = ht(p`) + dim A/p` so we see that ht(p`) = ht(p0) + `. (4) Any maximal prime ideal chain in A/p is of the form 0 ( p1/p ( ··· ( p`/p where p ( p1 ( ··· ( p` is a non-refinable prime ideal chain in A. By (3), ` = ht(p`) − ht(p) but p` must be maximal, so by (2), ht(p`) = n. Finally, applying (1) gives ` = n − ht(p) = dim A/p. Hence SDP holds for A/p. An important example of a ring with the strong dimension property is the polynomial algebra k[t1, . . . , tn]. Before proving this has SDP, we need a preliminary result which relies on the going down theorem. Proposition 2.6.6. Let A be a finitely generated k-algebra that is an integral domain. Then for any nonzero prime ideal p ⊂ A such that ht(p) = 1, we have dim A/p = dim A − 1. Proof. By Noether’s normalization lemma, there exist elements y1, . . . , yn ∈ A which are algebraically independent over k and have the property that A/k[y1, . . . , yn] is an integral extension. Then by Theorems 2.3.7 and 2.6.4, dim A = dim k[y1, . . . , yn] = n. Since ht(p) = 1, there is no nonzero prime ideal of A properly contained in p. Set q = p ∩ k[y1, . . . , yn]. If q = 0, then we would have 0 = p ∩ k[y1, . . . , yn] = 0 ∩ k[y1, . . . , yn], which would imply p = 0 by Lemma 2.3.3, but this contradicts ht(p) = 1. Suppose there exists a nonzero 0 0 prime ideal q ⊂ k[y1, . . . , yn] such that q ( q. Then the going down theorem would give us a nonzero prime ideal p0 ( p, contradicting ht(p) = 1. This shows that ht(q) = 1 also. Now by Lemma 2.3.1(a), A/p is integral over k[y1, . . . , yn]/q, so it suffices to prove dim k[y1, . . . , yn]/q = n − 1. Set B := k[y1, . . . , yn]. In light of Lemma 1.4.2 and Corollary 2.6.3, we have that dim B = n. We first show that q is a principal ideal. Since q 6= 0, there exists a nonzero element 35 2.6 Dimension and Transcendence Degree 2 Integral Dependence f 0 ∈ q. Then because B is a UFD, f 0 is divisible by some irreducible element f ∈ q, but then f is prime and 0 6= (f) ⊆ (f 0) ⊆ q. Since q is a minimal prime, we must have (f) = q. Finally, B/q = B/(f) is an integral extension of k[z2, . . . , zn] where {1, z2, . . . , zn} is a change of basis of {y1, . . . , yn} such that f(1, z2, . . . , zn) is monic. It follows from Theorem 2.3.7 and Corollary 2.6.3 that dim B/q = dim k[z2, . . . , zn] = n − 1. Theorem 2.6.7. If A is a finitely generated k-algebra which is an integral domain, then A has the strong dimension property. Proof. We induct on n = dim A. If n = 0, A is a field so the property holds trivially. Now assume that every k-algebra which is a domain and has dimension less than n has the strong dimension property. Take a maximal chain of prime ideals p0 ( ··· ( p` in A. Since A is a domain, we must have p0 = 0 and p` maximal. Consider the integral domain B = A/p1. By the correspondence theorem, 0 = p1/p1 ( p2/p1 ( ··· ( p`/p1 is a maximal chain of prime ideals in B of length ` − 1. Moreover, ht(p1) = 1 so by Proposition 2.6.6, dim B = dim A/p1 = dim A − 1 = n − 1. Thus by induction, n − 1 = ` − 1 and so we get n = ` as required. Corollary 2.6.8. For any n ≥ 1, k[t1, . . . , tn] has the strong dimension property. In partic- ular, dim k[t1, . . . , tn]/p = n − ht(p) for any prime ideal p ⊂ k[t1, . . . , tn]. 36 3 Noetherian and Artinian Rings 3 Noetherian and Artinian Rings 3.1 Ascending and Descending Chains Definition. A partially ordered set (Σ, ≤) satisfies the ascending chain condition (ACC) if any ascending chain x1 ≤ x2 ≤ · · · in Σ is stationary, i.e. there is some N ∈ N such that xn = xN for all n ≥ N. Definition. A partially ordered set (Σ, ≤) satisfies the descending chain condition (DCC) if any descending chain x1 ≥ x2 ≥ · · · in Σ is stationary, i.e. there is some N ∈ N such that xn = xN for all n ≥ N. In ring theory, we study chain conditions on the collection ΣM of submodules of a fixed A-module M, partially ordered by inclusion. The modules M that satisfy the above chain conditions have special names. Definition. An A-module M is noetherian if the collection of submodules of M satisfies the ascending chain condition. On the other hand, M is artinian if the collection of submodules of M satisfies the descending chain condition. Lemma 3.1.1. Let (Σ, ≤) be a poset. Then (i) Σ satisfies ACC if and only if every nonempty subset of Σ has a maximal element. (ii) Σ satisfies DCC if and only if every nonempty subset of Σ has a minimal element. In particular, M is noetherian iff every nonempty collection of submodules has a maximal element, and M is artinian iff every nonempty collection of submodules has a minimal element. Proposition 3.1.2. An A-module M is noetherian if and only if every submodule of M is finitely generated (as an A-module). f g Lemma 3.1.3. Let 0 → M 0 −→ M −→ M 00 → 0 be a short exact sequence of A-modules. Then (a) M is noetherian if and only if M 0 and M 00 are noetherian. (b) M is artinian if and only if M 0 and M 00 are artinian. Proof. ( =⇒ ) Consider the maps induced by f and g: ∗ ∗ f :ΣM 0 −→ ΣM g :ΣM 00 −→ ΣM N 0 7−→ f(N 0) N 00 7−→ g−1(N 00). ∗ ∗ Then f and g are both order-preserving and injective maps, so the posets ΣM 0 and ΣM 00 inherit the chain conditions from ΣM . 37 3.1 Ascending and Descending Chains 3 Noetherian and Artinian Rings ( ⇒ = ) We prove this for the artinian property; the proof for noetherian modules is similar. Let C : M1 ⊇ M2 ⊇ M3 ⊇ · · · be a descending chain of submodules Mi ⊂ M. These correspond to chains of submodules −1 −1 −1 −1 0 f (C):f (M1) ⊇ f (M2) ⊇ f (M3) ⊇ · · · in M 00 g(C):g(M1) ⊆ g(M2) ⊇ g(M3) ⊇ · · · in M . −1 −1 By the DCC on ΣM 0 and ΣM 00 , there exists a single N ∈ N such that f (Mn) = f (MN ) and g(Mn) = g(MN ) for all n ≥ N. For all n ≥ N, we already have Mn ⊇ MN . On the other hand, given x ∈ MN , there exists a y ∈ Mn such that g(x) = g(y). This implies g(x) − g(y) = g(x − y) = 0 by A-linearity. So x − y ∈ ker g which equals im f by exactness. 0 −1 −1 Thus there is some z ∈ M such that f(z) = x − y ∈ MN . Then z ∈ f (MN ) = f (Mn) so x − y = f(z) ∈ Mn. Hence x = y + f(z) ∈ Mn so MN ⊆ Mn. This proves the DCC on ΣM , so M is artinian. Ln Corollary 3.1.4. If M1,...,Mn are noetherian (resp. artinian) A-modules then i=1 Mi is also noetherian (resp. artinian). Proof. This is proved by induction on n, considering short exact sequences n n−1 M M 0 → Mn → Mi → Mi → 0 i=1 i=1 and applying Lemma 3.1.3. We distinguish the special case of A as a module over itself. Definition. A (commutative) ring A is a noetherian ring if it is noetherian as a module over itself. Similarly, A is an artinian ring if it is artinian as a module over itself. Let A be a ring. Then the A-submodules of A are exactly the ideals of A. Then by Proposition 3.1.2, A is noetherian if and only if all ideals of A are finitely generated. Corollary 3.1.5. If A is noetherian (resp. artinian), then any finitely generated A-module M is also noetherian (resp. artinian). Proof. Write M = Am1 + ... + Amn for mi ∈ M. Consider the projection An −→ M n X (a1, . . . , an) 7−→ aimi. i=1 Then An is noetherian (resp. artinian) by Corollary 3.1.4 so M is noetherian (resp. artinian) by Lemma 3.1.3. 38 3.1 Ascending and Descending Chains 3 Noetherian and Artinian Rings Lemma 3.1.6. For a field k and a k-vector space V , the following are equivalent: (i) dimk V < ∞. (ii) V is a noetherian k-module. (iii) V is an artinian k-module. Proof. (i) =⇒ (ii), (iii) follow from the linear algebra fact that every subspace of a finite dimensional vector space is finite dimensional, and thus has a finite basis. (ii), (iii) =⇒ (i) If dimk V = ∞ then there is a countable linear independent subset ∞ {vj}j=1 ⊂ V . Let Vi = Span{vj | j ≤ i} for each i ∈ N. Then V1 ( V2 ( ··· is an ascending chain in which each subspace is finitely generated, so it cannot stabilize. Likewise, let Wi = Span{vj | j ≥ i} for each i ∈ N. Then W1 ) W2 ) ··· is a descending chain which doesn’t stabilize. Hence dimk V = ∞ implies is V is neither noetherian nor artinian. Proposition 3.1.7. Suppose A has maximal ideals m1,..., mn, not necessarily distinct, with m1 ··· mn = 0. Then A is noetherian if and only if A is artinian. Proof. Say N is an A-module and m ⊂ A is a maximal ideal. Then N/mN becomes a k- vector space for k = A/m via (a + m)(n + mN) = an + mN. Then A-submodules of N/mN are the same as k-subspaces of N/mN. Thus N/mN is noetherian (resp. artinian) as an A-module if and only if N/mN is noetherian (resp. artinian) as a k-vector space; this in turn is equivalent to dimk N/mN < ∞ by Lemma 3.1.6. Hence N/mN is noetherian if and only if N/mN is artinian. Now let m1,..., mn ⊂ A be maximal ideals satisfying m1 ··· mn = 0. Set Ai = A/m1 ··· mi for each 1 ≤ i ≤ n. In particular, An = A/{0} = A by hypothesis. If A is noetherian (resp. artinian) then each Ai is noetherian (resp. artinian) by Lemma 3.1.3. Note that A1 = A/m1 is a field, so A1 is both noetherian and artinian. We claim that Ai is noetherian for each 1 ≤ i ≤ n if and only if Ai is artinian for each 1 ≤ i ≤ n. Indeed, A1 is the base case so for n ≥ 2, consider the short exact sequence 0 → m1 ··· mi−1/m1 ··· mi → Ai → Ai−1 → 0. Set Ni = m1 ··· mi−1/m1 ··· mi. Then by Lemma 3.1.3, Ai is noetherian ⇐⇒ Ni and Ai−1 are noetherian ⇐⇒ Ni is noetherian and Ai−1 is artinian (by induction) ⇐⇒ Ni is artinian and Ai−1 is artinian (by preliminary remarks) ⇐⇒ Ai is artinian. Finally, since A = An, this implies that A is noetherian if and only if A is artinian. 39 3.2 Composition Series 3 Noetherian and Artinian Rings 3.2 Composition Series Let M be an A-module. Definition. A chain in M is a strictly descending chain of submodules of M of the form M = M0 ) M1 ) ··· ) Mn = 0. The integer n is called the length of the chain. Definition. A composition series of M is a chain in M which cannot be refined, or equivalently, the quotient Mi/Mi+1 is simple for all 0 ≤ i ≤ n − 1. Definition. The length of a module M, denoted `(M), is by convention `(M) = ∞ if M does not have a composition series. Otherwise, `(M) = min{n ∈ N | M has a composition series of length n}. One can view the length function `(·) as a generalization of the dimension of a vector space. Proposition 3.2.1. Assume M is an A-module that has a composition series. Then (a) For every submodule N ⊂ M, `(N) ≤ `(M) and if N is a proper submodule, `(N) < `(M). (b) Any chain in M has length at most `(M). (c) Every composition series of M has length `(M). (d) Any chain in M can be extended to a composition series. Proof. (a) Choose a composition series of minimal length, M = M0 ) M1 ) ··· ) M` = 0. Then ` = `(M). Intersecting with N, setting Ni = N ∩ Mi, gives a sequence of containments that are not necessarily strict: N = N0 ⊇ N1 ⊇ · · · ⊇ N` = 0. However for each i, we have an inclusion Ni/Ni+1 ,→ Mi/Mi+1 since Ni+1 = N ∩ Mi+1 = N ∩ Mi ∩ Mi+1 = Ni ∩ Mi+1. ∼ By assumption, each Mi/Mi+1 is simple, so either Ni/Ni+1 = 0 or Ni/Ni+1 = Mi/Mi+1, in which case Ni/Ni+1 is also simple. After removing repetitions, i.e. where Ni = Ni+1, we obtain a composition series of N of length at most `. Hence `(N) ≤ `(M). Now suppose N ( M but `(N) = `(M). Then Ni ) Ni+1 for all 0 ≤ i ≤ ` − 1 in the above chain. By construction, M`−1 is a simple A-module and 0 6= N`−1 ⊆ M`−1, so we have N`−1 = M`−1. 40 3.2 Composition Series 3 Noetherian and Artinian Rings By induction, Ni = Mi for all 0 ≤ i ≤ ` − 1. Hence N = N0 = M0 = M, a contradiction. Therefore `(N) < `(M) for every proper submodule N. (b) Consider any chain M = M0 ) M1 ) ··· ) Mn = 0. Then by (a), ` = `(M0) > `(M1) > ··· > `(Mn) > 0. In this string we have n strict inequalities so it must be that n ≤ `. (c) By definition of `(M), any compositions series of M has length at least `(M), but by (b), every composition series of M has length at most `(M). (d) Given any chain, refine it as many times as possible. By (b), we cannot get a chain of length strictly greater than `(M), so the process will eventually terminate when the length of the chain is `(M). At this point we will have a composition series for M. Corollary 3.2.2. An A-module M has a composition series if and only if M is both noethe- rian and artinian. Proof. ( =⇒ ) Assume `(M) < ∞. Then by Proposition 3.2.1, all chains in M have length at most `(M). In particular, any ascending or descending chain must be finite, which implies M has satisfies ACC and DCC. ( ⇒ = ) Assume M 6= 0. Then the set of all proper submodules of M has a maximal element by Lemma 3.1.1; call it M1. If M1 = 0, then M ) M1 = 0 is a composition series for M. Otherwise, the set of proper submodules of M1 has a maximal element M2, by Lemma 3.1.1 again, using the fact that submodules of noetherian modules are noetherian. Continue in this fashion to construct a chain M = M0 ) M1 ) M2 ) ··· By maximality in each step, Mi/Mi+1 is simple for all i ≥ 0. Since M is artinian, this process must terminate, else the DCC is violated. Then at some point we have M = M0 ) M1 ) M2 ) ··· ) Mn ) 0 which is a composition series for M. Lemma 3.2.3. If M is an A-module that admits a composition series, then for any short exact sequence 0 → M 0 → M → M 00 → 0 of A-modules, `(M) = `(M 0) + `(M 00). That is, length is additive on short exact sequences. f g Proof. Label the morphisms 0 → M 0 −→ M −→ M 00 → 0. First, M is both noetherian and artinian by Corollary 3.2.2. Note that M 0 and M 00 are also noetherian and artinian by Lemma 3.1.3, so each admits a composition series. Consequently, `(M 0) and `(M 00) are defined. Let 0 0 0 0 00 00 00 00 M = M0 ) M1 ) ··· ) M` = 0 and M = M0 ) M1 ) ··· ) Mn = 0 be composition series of M 0 and M 00. Then by Proposition 3.2.1(c), `(M 0) = ` and `(M 00) = n. By exactness, we get a sequence of inclusions −1 00 −1 00 0 0 0 M = g (M0 ) ⊃ · · · ⊃ g (Mn ) = ker g = im f = f(M0) ⊃ f(M1) ⊃ · · · ⊃ f(M`) = 0. (∗) 41 3.3 Noetherian Rings 3 Noetherian and Artinian Rings −1 00 00 −1 00 Now since g is surjective, g(g (Mi )) = Mi so the inclusions are strict between the g (Mi ), −1 0 0 0 ≤ i ≤ n. Similarly, since f is injective, f (f(Mi )) = Mi so the inclusions between the 0 f(Mj), 0 ≤ j ≤ `, are strict. Therefore (∗) is a strictly descending chain of submodules of M of total length ` + n. Consider each quotient in the first half; since g is surjective, we get an isomorphism −1 00 −1 00 ' 00 00 g (Mi )/g (Mi+1) −→ Mi /Mi+1 for 0 ≤ i ≤ n. −1 00 −1 00 Then each g (Mi )/g (Mi+1) is simple. Similarly, injectivity of f gives us an isomorphism 0 0 ' 0 0 Mi /Mi+1 −→ f(Mi )/f(Mi+1) for 0 ≤ j ≤ `, 0 0 which shows the f(Mi )/f(Mi+1) are all simple. Hence (∗) is a composition series for M, so by Proposition 3.2.1(c), `(M) = ` + n = `(M 0) + `(M 00). 3.3 Noetherian Rings In this section we present the key details regarding noetherian rings and their modules in commutative algebra. Lemma 3.3.1. If A is a noetherian ring, then (a) For any ideal I ⊂ A, A/I is a noetherian ring. (b) For any multiplicatively closed subset S ⊆ A, S−1A is a noetherian ring. In particular, Ap is a local noetherian ring for any prime ideal p ⊂ A. Proof. (a) By Lemma 3.1.3, we know A/I is noetherian as an A-module. It follows from the isomorphism theorems that the A-submodules of A/I are exactly the ideals of A/I, so it follows that A/I is noetherian as a ring. −1 −1 (b) If J1 ⊆ J2 ⊆ · · · is a chain of ideals in S A then applying j gives a chain of −1 −1 ideals j (J1) ⊆ j (J2) ⊆ · · · in A. Since A is noetherian, the chain stabilizes, i.e. there −1 −1 is some n ∈ N such that j (Ji) = j (Jn) for all i ≥ n. Applying the localization functor −1 −1 −1 S to these modules returns the original Ji, since S (j (Ji)) = Ji for each i ≥ 1 by Proposition 1.3.8: J1 ⊆ J2 ⊆ · · · ⊆ Jn = Jn+1 = ··· . −1 −1 Hence the ACC holds for S A, so S A is noetherian. The final remark about Ap follows from Corollary 1.3.9. Remark. A similar argument as in Lemma 3.3.1(b) shows that for every noetherian A- module M, any localization S−1M is also a noetherian A-module. One of the fundamental results in the theory of noetherian rings is known as Hilbert’s ba- sis theorem. We will see that it implies every finitely generated ring extension of a noetherian ring is also noetherian. Theorem 3.3.2 (Hilbert’s Basis Theorem). If A is a noetherian ring, then A[t] is also noetherian. 42 3.3 Noetherian Rings 3 Noetherian and Artinian Rings Proof. Let I ⊂ A[t] be an ideal and let L denote the set of all leading coefficients of elements of I. That is, ( n ) X i L = α ∈ A : there is some ait ∈ I with an = α . i=0 We claim L is an ideal. Clearly 0 ∈ L since I contains the zero polynomial. Suppose f, g ∈ I n m such that f = ant + ... + a0 and g = bmt + ... + b0. We must show an + bm ∈ L. If n = m then the leading coefficient of f + g is an + bm, and since f + g ∈ I, we have an + bn ∈ L. If n 6= m, without loss of generality assume n < m. Then m−n m m−n m t f + g = (ant + ... + a0t ) + (bmt + ... + b0) m m−n = (an + bm)t + ... + (a0 + bm−n)t + ... + b0 which lies in I, showing an+bm ∈ L. Finally, if c ∈ A then can is the leading term of cf ∈ I, so L is an ideal of A. Now, since A is noetherian, L is finitely generated, say by a1, . . . , an ∈ A. For each 1 ≤ i ≤ n, let fi ∈ I be a polynomial having ai as its leading coefficient. For each ei of these, write fi = ait + ... so that deg fi = ei. Set E = max{e1, . . . , en}. For each 0 ≤ d ≤ E − 1, let Ld be the set of all leading coefficients of polynomials in I of degree d, together with 0: ( d ) X i Ld = α ∈ A : there is some ait ∈ I with ad = α . i=0 The proof that each Ld is an ideal of A is similar to the proof above for L. Then each Ld is finitely generated, say by bd1, . . . , bdnd ∈ A. Let fd1, . . . , fdnd ∈ I be polynomials such that the leading coefficient of fdr is bdr, 1 ≤ r ≤ nd. We claim that f1, . . . , fn and fdr, 0 ≤ d ≤ E − 1, 1 ≤ r ≤ nd, are a generating set for I. Let Q denote the ideal generated by these polynomials. By construction, Q ⊆ I since the generators of Q are elements of I. If Q ( I, choose f ∈ I r Q of minimum degree d with leading coefficient a. First suppose d ≥ E. Then a ∈ L so there exist elements c1, . . . , cn ∈ A such that d−e1 d−en a = c1a1 + ... + cnan. Observe that g = c1t f1 + ... + cnt fn is a polynomial in Q of degree d and leading coefficient a, so f − g ∈ I has strictly smaller degree than f. By minimality, we must have f − g = 0, that is, f = g ∈ Q, a contradiction. On the other hand, if d < E then f ∈ Ld so we have f = c1bd1 + ... + cnd bdnd for some cr ∈ A. Then g = c1fd1 + ... + cnd fnd is a polynomial in Q of degree d and leading coefficient a, producing the same contradiction. It follows that Q = I, so I is finitely generated. This proves that A[t] is noetherian. Remark. The converse of Hilbert’s basis theorem is trivially true: A[t] noetherian implies A noetherian since A ∼= A[t]/(t) as rings. Corollary 3.3.3. If A is noetherian, then (a) A[t1, . . . , tn] is noetherian for any n ≥ 1. (b) B is noetherian for any finitely generated ring extension B/A. 43 3.3 Noetherian Rings 3 Noetherian and Artinian Rings Proof. (a) Induct on the number of generators. The base case is Hilbert’s basis theorem. (b) We can write B = A[b1, . . . , bn] for some bi ∈ B. Then there is a (unique) surjective ring homomorphism ϕ : A[t1, . . . , tn] −→ B ti 7−→ bi. ∼ So B = A[t1, . . . , tn]/ ker ϕ and we can apply (a) and Lemma 3.3.1. Corollary 3.3.4. For a field k, any finitely generated k-algebra is noetherian. Proof. Every field is noetherian (it admits a trivial composition series) so Lemma 3.3.3(b) may be applied. Corollary 3.3.5. Every finitely generated commutative ring is noetherian. Proof. Let A be such a ring. There is a unique ring homomorphism f : Z → A given by f(n) = n · 1A for all n. Set A0 = f(Z). Then A0 is noetherian by Lemma 3.3.1(a) and the hypothesis that A is finitely generated as a ring means that A/A0 is a finitely generated ring extension. Then the statement follows from Corollary 3.3.3(b). Proposition 3.3.6. Suppose A is noetherian and C ⊃ B ⊃ A are extensions of rings such that C/A is a finitely generated extension and C is a finitely generated B-module. Then B/A is also finitely generated. Proof. Choose x1, . . . , xm ∈ C and y1, . . . , yn ∈ C with y1 = 1 such that n X C = A[x1, . . . , xm] = Byj. j=1 Then there exist bij` ∈ B such that for all 1 ≤ i ≤ m, 1 ≤ j ≤ n, n X xiyj = bij`y`. (∗) `=1 Define the subring B0 = A[bij` | 1 ≤ i ≤ m, 1 ≤ j, ` ≤ n]. Then B0 is finitely generated as an A-module by construction, so by Corollary 3.3.3, B0 is noetherian. Consider the B0-submodule n X M = B0y` ⊆ C. `=1 Pn By (∗), xiyj ∈ M for all i, j. Thus xi j=1 B0yj ⊆ M, or in other words xiM ⊆ M for each i. In particular, C = A[x1, . . . , xm] ⊆ M ⊆ C, so M = C. Now we have A ⊆ B0 ⊆ B ⊆ C with B0 noetherian and C a finitely generated B0-module. Thus by Corollary 3.1.5, C is a noetherian B0-module, and Proposition 3.1.2 implies that B ⊆ C is finitely generated as a B0-module. Hence B/B0 is a finitely generated ring extension, and by construction B0/A is a finitely generated ring extension, so we conclude that B/A is finitely generated as well. 44 3.3 Noetherian Rings 3 Noetherian and Artinian Rings As a result, we obtain the following interesting fact. Theorem 3.3.7. If a field K is finitely generated as a ring then K is a finite field. Proof. Consider the map f : Z → K. Since K is a field, f(Z) is an integral domain. We have ∼ ∼ two cases: (1) when char K = p > 0, f(Z) = Fp; and (2) when char K = 0, f(Z) = Z. In case (1), K is a finitely generated Fp-algebra. By Hilbert’s Nullstellensatz (2.4.2), K/Fp is an algebraic extension, so in fact it is an extension of finite fields. Therefore K is finite. On the other hand, in (2), K is finitely generated over Z and K is a field, so we have Z ⊂ Q ⊂ K. This means K is a finitely generated Q-algebra, so once again the Nullstellensatz says that K/Q is a finite extension. However, one sees that the conditions of Proposition 3.1.2 are satisfied, so Q is a finitely generated ring extension of Z, which is false. Therefore the only possibility is that K is a finite field. Corollary 3.3.8. Any matrix ring GLn(K), n ≥ 2, over an infinite field K cannot be finitely generated. The proof of Hilbert’s basis theorem can be generalized to prove that power series rings over a noetherian ring are also noetherian. Theorem 3.3.9. If A is noetherian, then the power series ring A[[t]] is also noetherian. P∞ i Proof. For a power series f ∈ A[[t]], where f = i=0 ait , let j be the smallest index such that aj 6= 0. We say aj is the coefficient of lowest degree of f. If ai = 0 for all i ≥ 0, i.e. f is the zero power series, we will say its coefficient of lowest degree is 0. Let I ⊂ A[[t]] be an ideal and define L ⊂ A to be the set of all coefficients of lowest degree of elements in I: ( ∞ ) X i L = aj : ait ∈ I for some j ∈ N0 and ai, i > j . i=j We first show L is an ideal. Since 0 ∈ I and 0 has lowest degree coefficient 0, we have 0 ∈ L. P∞ i P∞ k Let f, g ∈ I, with f = i=j ait and g = k=` bkt . We must show aj + b` ∈ L. If j = `, the lowest degree coefficient of f + g is aj + bj, and since f + g ∈ I, we have aj + bj ∈ L. If j 6= `, without loss of generality assume j < `. Then ∞ ∞ ∞ ∞ ∞ `−j `−j X i X k X i+`−j X k X k t f + g = t ait + bkt = ait + bkt = (ak+j−` + bk)t i=j k=` i=j k=` k=` so we see that the lowest degree coefficient is aj + b`. Therefore aj + b` ∈ L. Now, if α ∈ A, the lowest degree coefficient of αf is αai, and αf ∈ I since I is an ideal, so αai ∈ L. This proves L is an ideal of A. By the noetherian hypothesis, L is finitely generated by some α1, . . . , αn ∈ A. Now for each k = 1, . . . , n, take fk ∈ I such that the lowest degree term of fk is αk; let ∞ e be the exponent of t in f having α as its coefficient, i.e. f = α tek + P a ti. Set k k k k k i=ek+1 i E = max{ek | 1 ≤ k ≤ n}. For all 0 ≤ d ≤ E − 1, define Jd ⊂ A by ( ∞ ) X i Jd = ad : ait ∈ I for some ai, i > d ∪ {0}. i=d 45 3.4 Primary Decomposition 3 Noetherian and Artinian Rings The proof that each Jd is an ideal of A is similar to the proof for L. In particular, since A is noetherian we get that Jd is finitely generated for each 0 ≤ d ≤ E − 1. Say Jd is generated by βd1, . . . , βdnd ∈ A. For each 0 ≤ d ≤ E − 1 and for each 1 ≤ r ≤ nd, let fdr ∈ A[[t]] be d P∞ i such that fdr = βdrt + i=d+1 ait ∈ I. Let Q be the ideal of A[[t]] finitely generated by all fk, 1 ≤ k ≤ n and fdr, 0 ≤ d ≤ E − 1, 1 ≤ r ≤ nd. We claim that Q = I. Since all of the generators of Q lie in I, it is clear that Q ⊆ I. On the other hand, take f ∈ I and suppose d the lowest degree term of f is at , for some a ∈ L and d ≥ 0. If d ≤ E − 1, then a ∈ Jd so Pnd there exist cr ∈ A such that a = r=1 crβdr, which implies nd nd ∞ ! nd ∞ nd X X d X i X d X X i f − crfdr = f − cr βdrt + ait = f − crβdrt − crait r=1 r=1 i=d+1 r=1 i=d+1 r=1 is a power series in A[[t]] of lowest degree at least d + 1. (We can switch the sums in the last Pnd step since one is finite.) Since f − r=1 crfdr ∈ I, we may assume that the lowest degree of f is at least E. If the lowest degree of f is d > E, then nE X d−E f − crfdrt r=1 which is a power series in A[[t]] of lowest degree at least d + 1. Notice that the ideals Jd are ascending: J0 ⊆ J1 ⊆ J2 ⊆ · · · . Since A is noetherian, this chain stabilizes, or in other words, we can always find power series ∞ X `−E gr = c`rt for 1 ≤ r ≤ nE `=d such that f can be written n XE f = grfdr. r=1 Hence f ∈ Q so Q = I and it follows that I is finitely generated. 3.4 Primary Decomposition For this section, we assume A is a noetherian ring. Our goal is to prove an analog of primary decomposition (of modules over a PID) for modules over A. Lemma 3.4.1. For any ideal I ⊂ A, there exists n ∈ N such that r(I)n ⊆ I. Proof. Since A is noetherian, r(I) is finitely generated; say x1, . . . , xm ∈ r(I) such that ei (x1, . . . , xm) = r(I). Then there exist ei ∈ N such that xi ∈ I for each 1 ≤ i ≤ m. Set j1 jm n = e1 + ... + em. Then x1 ··· xm ∈ I whenever j1 + ... + jm ≥ n, but the products j1 jm n n x1 ··· xm generate r(I) , so we see that for this choice of n, r(I) ⊆ I. We obtain the following fact about the nil radical of a noetherian ring. Corollary 3.4.2. There exists an n ∈ N such that N(A)n = 0. 46 3.4 Primary Decomposition 3 Noetherian and Artinian Rings Proof. For the ideal I = (0), r(0) = N(A). The statement follows from Lemma 3.4.1. Definition. A proper ideal I ⊂ A is irreducible if for any ideals J, K ⊂ A with J ∩K = I, we have J = I or K = I. Example 3.4.3. Prime ideals are always irreducible. Indeed, if J ∩K = p is prime but p ( J and p ( K, then there are some elements x ∈ J rp and y ∈ Krp. But xy ∈ JK ⊆ J ∩K = p which is a contradiction since p is prime. Thus p is irreducible. Lemma 3.4.4. Any proper ideal I ⊂ A is a finite intersection of irreducible ideals. Proof. Consider the collection Σ = {proper ideals I ⊂ A | I is not an intersection of irreducible ideals}. If we assume Σ is nonempty, then since A is noetherian, Lemma 3.1.1 guarantees that Σ has a maximal element I0. In particular, I0 is not irreducible. Then there exist ideals J, K ⊂ A such that I0 = J ∩ K but I0 ( J and I0 ( K. Since I0 is maximal in Σ, neither J nor K lie in Σ. Thus J and K are each a finite intersection of irreducible ideals, but then I0 = J ∩K is also a finite intersection of irreducible ideals, a contradiction. Hence Σ must be empty. Definition. A proper ideal I ⊂ A is primary if for every x, y ∈ A, xy ∈ I implies x ∈ I or y ∈ r(I). Equivalently, I is primary if xy ∈ I implies that x ∈ I, y ∈ I or x, y ∈ r(I). Lemma 3.4.5. If a proper ideal I ⊂ A is irreducible, then I is primary. Proof. Consider B = A/I. Then if I is irreducible, (0) is irreducible in B, i.e. there do not exist two nonzero ideals J, K ⊂ B with J ∩ K = 0. Assume x, y ∈ B such that xy = 0 and x 6= 0. We will show that yn = 0 for some n ∈ N, i.e. y is nilpotent. Consider the ascending chain Ann(y) ⊆ Ann(y2) ⊆ Ann(y3) ⊆ · · · By the ACC, there is some n ∈ N such that Ann(yi) = Ann(yn) for all i ≥ n. Now we show (x) ∩ (yn) = 0. If b ∈ (x) ∩ (yn) then b = rx for some r ∈ B, so by = rxy = 0. Also, b = cyn for some c ∈ B, so cyn+1 = cyny = by = 0. This shows c ∈ Ann(yn+1) = Ann(yn), and thus 0 = cyn = b. Hence (x) ∩ (yn) = 0 but since (0) is irreducible and x 6= 0, we must have (yn) = 0. Finally, this implies I ⊂ A is primary, since if x0, y0 ∈ A such that x0y0 ∈ I but x0 6∈ I, we have x0 + I 6= I and (x0 + I)(y0 + I) = x0y0 + I = I. Then by the first part of the proof, x0 + I 6= I means (y0 + I)n = (y0)n + I = I for some n ∈ N, and thus (y0)n ∈ I. Theorem 3.4.6 (Primary Decomposition). For any proper ideal I ⊂ A, there exist finitely many primary ideals Q1,...,Qn such that I = Q1 ∩ · · · ∩ Qn. Proof. Apply Lemmas 3.4.4 and 3.4.5. Lemma 3.4.7. If Q ⊂ A is a primary ideal, then r(Q) is the smallest prime ideal of A containing Q. 47 3.4 Primary Decomposition 3 Noetherian and Artinian Rings Proof. Assume x, y ∈ A with xy ∈ r(Q). Then (xy)n = xnyn ∈ Q for some n ∈ N, so xn ∈ Q or yn ∈ r(Q). In the first case, we see immediately that x ∈ r(Q). Otherwise, yn ∈ r(Q) implies that y ∈ r(r(Q)) = r(Q) by Lemma 1.1.4(c). Hence r(Q) is a prime ideal. Now, by Corollary 1.1.3, r(Q) = T{p0prime | Q ⊆ p0 ⊂ A} but since r(Q) is prime, it is the smallest prime containing Q. Definition. For a prime ideal p ⊂ A, any proper ideal Q ⊂ A is called p-primary if p = r(Q). Examples. 1 If A is a UFD and p ∈ A is a prime element, then (pn) is p-primary (the usual notation for (p)-primary) for any n ∈ N. 2 In general, a p-primary ideal need not be a power of p. Let A = k[x, y] and Q = (x, y2). Then A/Q ∼= k[y](y2) and in this quotient, the zero divisors are the multiplies of y, hence nilpotent. This implies Q is primary, and clearly its radical is p = (x, y). However, p2 ( Q ( p are strict inclusions, so Q is not a power of a prime ideal. 3 If p is a prime ideal, pn need not always be primary. Let A = k[x, y, z]/(xy − z2) and letx, ¯ y,¯ z¯ ∈ A be the respective images of x, y, z in the quotient. Set p = (¯x, z¯). Then p is a prime ideal of A since A/p ∼= k[y] is an integral domain. However,x ¯y¯ =z ¯2 ∈ p2, butx ¯ 6∈ p2 andy ¯ 6∈ r(p2) = p. So p2 is not primary. The next result shows that the counterexamples in 2 and 3 do not occur when we consider maximal ideals. Lemma 3.4.8. Let A be a noetherian ring. (a) If I ⊂ A is a proper ideal such that r(I) is maximal, then I is primary. (b) If m ⊂ A is a maximal ideal, then mn is primary for all n ∈ N. Proof. (a) Assume r(I) is maximal. Then by Corollary 1.1.3, the set of all prime ideals in A containing I just consists of r(I). Thus B = A/I is a local ring with unique maximal ideal r(I)/I. If x, y ∈ B such that xy = 0 but y 6∈ r(I)/I, then since r(I)/I is the unique maximal ideal, y ∈ B×. So xy = 0 =⇒ xyy−1 = 0 =⇒ x = 0. Now if x0, y0 ∈ A such that x0y0 ∈ I but y0 6∈ r(I), then by the previous statement, (x0 + I)(y0 + I) = I =⇒ x0 + I = I =⇒ x0 ∈ I. This shows I is primary. (b) By Lemma 1.1.4, r(mn) = r(m) = m. By (a), this shows that mn is m-primary. Corollary 3.4.9. If m ⊂ A is a maximal ideal and I ⊂ A is any proper ideal, then I is m-primary if and only if there is some n ∈ N such that mn ⊆ I ⊆ m. Proof. ( =⇒ ) By Lemma 1.1.4, I ⊆ r(I) = m. Then by Lemma 3.4.1, r(I)n = mn ⊆ I for some n. ( ⇒ = ) Conversely, by Lemma 3.4.8(a), m = r(mn) ⊆ r(I) ⊆ r(m) = m so we have r(I) = m. By definition, this means I is m-primary. 48 3.4 Primary Decomposition 3 Noetherian and Artinian Rings Tn Proposition 3.4.10. Assume I ⊂ A is a proper ideal and I = i=1 Qi for primary ideals Qi. If p1,..., pn are (not necessarily distinct) prime ideals such that Qi is pi-primary for each 1 ≤ i ≤ n, then (a) For every prime ideal p ⊂ A such that p ⊇ I, p ⊇ pi for some 1 ≤ i ≤ n. (b) The minimal elements of the collection {p ⊂ A prime | p ⊇ I} are exactly the minimal n elements of {pi}i=1. (c) If A is noetherian, it only has finitely many minimal prime ideals. Tn Proof. (a) If p ⊇ I = i=1 Qi then by Lemma 1.1.4, n n \ \ p = r(p) ⊇ r(I) ⊇ r(Qi) = pi. i=1 i=1 Thus p contains one of the pi. (b) If p ⊂ A is a minimal prime ideal containing I, then by (a), p ⊇ pi for some pi. But by minimality, p = pi. (c) Apply (b) and primary decomposition (3.4.6) to the ideal I = (0). Lemma 3.4.11. If Q1,...,Qn are p-primary ideals for a fixed prime ideal p ⊂ A, then Tn i=1 Qi is also p-primary. Tn Tn Tn Proof. By assumption r(Qi) = p for all 1 ≤ i ≤ n, so r ( i=1 Qi) = i=1 r(Qi) = i=1 p = p. Tn Tn Tn It remains to show i=1 Qi is primary. Assume x, y ∈ A with xy ∈ i=1 Qi but x 6∈ i=1 Qi. Then for each i, xy ∈ Qi but there is some j such that x 6∈ Qj. Since Qj is p-primary, we Tn Tn must have y ∈ r(Qi) = p = r ( i=1). Therefore i=1 Qi is primary. Tn Definition. A primary decomposition I = Qi is called minimal (or reduced) if for T i=1 every 1 ≤ i ≤ n, Qi 6⊃ j6=i Qj and r(Qi) 6= r(Qj) whenever i 6= j. By Theorem 3.4.6 and Lemma 3.4.11, minimal primary decompositions exist for all ideals I ⊂ A. Example 3.4.12. Let A = k[x, y, z] and consider the ideals p1 = (x, y), p2 = (x, z) and ∼ ∼ m = (x, y, z). Then A/p1 = k[z] and A/p2 = k[y] which are integral domains, so p1 and p2 are prime. Also, A/m ∼= k which is a field, so m is a maximal ideal of A. We claim that p1 ∩ p2 = (x, yz). Clearly x and yz both lie in (x, y) ∩ (x, z), so (x, yz) ⊆ p1 ∩ p2. On the other hand, for any f ∈ p1 ∩ p2, f ∈ (x, y) implies f = gx + hy for g, h ∈ A. But f ∈ (x, z) is possible if and only if h = h0z for some h0 ∈ A, in which case we have f = gx+h0yz ∈ (x, yz). Hence p1 ∩ p2 = (x, yz). 2 Now we show that p1p2 = p1 ∩ p2 ∩ m and this is a minimal primary decomposition 2 2 of p1p2. The left can be written p1p2 = (x , xy, xz, yz). Notice that x , xy, xz and yz all 2 2 2 lie in m , and since p1p2 ⊆ p1 ∩ p2, this shows that x , xy, xz, yz ∈ p1 ∩ p2 ∩ m . Hence 2 2 2 p1p2 ⊆ p1 ∩ p2 ∩ m . Going the other way, suppose F ∈ p1 ∩ p2 ∩ m . Then F ∈ m so F = ax2 + bxy + cxz + dy2 + eyz + fz2 for some a, b, c, d, e, f ∈ A. Notice that by the 2 paragraph above, ax + bxy + cxz + eyz ∈ p1p2 so to show that F ∈ p1p2, it will suffice to 49 3.4 Primary Decomposition 3 Noetherian and Artinian Rings 2 2 2 show that dy + fz ∈ p1p2. By the previous statement, ax + bxy + cxz + eyz ∈ p1 ∩ p2 z 2 2 so because F ∈ p1 ∩ p2, we must have dy + fz ∈ p1 ∩ p2. Observe that fz ∈ p2 so it 2 2 follows that dy ∈ p2. But y 6∈ p2 and this ideal is prime, so we must have d ∈ p2. A similar 2 2 2 2 argument shows that f ∈ p1. Thus we have dy ∈ p2p1 and fz ∈ p1p2, so dy + fz ∈ p1p2. This proves F ∈ p1p2, so p1 ∩ p2 ∩ m ⊆ p1p2. 2 Since p1 and p2 are prime, they are primary (3.4.5). Also, since m is maximal, m is primary (3.4.8(b)). Each term in the primary decomposition has a distinct radical, so we need only check that none of them contains the intersection of the remaining two. Notice: 2 x ∈ p1 ∩ p2 but x 6∈ m , 2 2 2 y ∈ p1 ∩ m but y 6∈ p2, 2 2 2 and z ∈ p2 ∩ m but z 6∈ p1. 2 This shows p1p2 = p1 ∩ p2 ∩ m is a minimal primary decomposition. Definition. Let I ⊂ A be an ideal and take x ∈ A. The colon ideal, or ideal quotient of I by x is defined as (I : x) = {y ∈ A | xy ∈ I}. If J ⊂ A is another ideal, we can define the ideal quotient of I by J as (I : J) = {y ∈ A | yJ ⊆ I}. The following lemma collects some basic facts about ideal quotients. Lemma 3.4.13. Let I,J ⊂ A be ideals and x ∈ A. Then (a) (I : J) is an ideal of A and I ⊆ (I : J). (b) x ∈ I if and only if (I : x) = A. T T (c) For a collection of ideals {Ij} of A, j Ij : x = j(Ij : x). (d) AnnA(x) = (0 : x). (e) Let D = {z ∈ A | yz = 0 for some y ∈ A r {0}} be the set of zero divisors of A. Then [ [ [ D = (0 : x) = AnnA(x) = r(AnnA(x)). x6=0 x6=0 x6=0 Proof. (a) Since IJ ⊆ I, I ⊆ (I : J) is clear. To prove (I : J) is an ideal, let y, z ∈ (I : J) and r ∈ A. Then for any j ∈ J, yj, zj ∈ I and so (ry − z)j = ryj − zj = r(yj) − zj ∈ I. Hence (I : J) is an ideal. (b) This is the definition of an ideal I. 50 3.4 Primary Decomposition 3 Noetherian and Artinian Rings (c) Observe that ! \ \ y ∈ Ij : x ⇐⇒ xy ∈ Ij j j ⇐⇒ xy ∈ Ij for each Ij ⇐⇒ y ∈ (Ij : x) for each j \ ⇐⇒ y ∈ (Ij : x). j (d) This is the definition of the annihilator. (e) The first equality is trivial and the second is immediate from (d). To prove the third, we have AnnA(x) ⊆ r(AnnA(x)) by Lemma 1.1.4(a). On the other hand, for any n n y ∈ r(AnnA(x)), there is some n ≥ 1 such that y ∈ AnnA(x), i.e. xy = 0. Then y ∈ D so all equalities hold. Lemma 3.4.14. If Q ⊂ A is a p-primary ideal and x ∈ ArQ then (Q : x) is also p-primary. Proof. Fix x ∈ A r Q. Then for any y ∈ (Q : x), xy ∈ Q and since Q is primary, this implies y ∈ r(Q) = p. So (Q : x) ⊆ p. By Lemma 3.4.13(a), we have p = r(Q) ⊆ r((Q : x)) ⊆ r(p) = p. Therefore r((Q : x)) = p. It remains to show (Q : x) is primary. Suppose y, z ∈ A such that yz ∈ (Q : x) but z 6∈ r((Q : x)) = p. Then xyz ∈ Q but z 6∈ r(Q) = p, so xy ∈ Q since Q is primary. By definition this implies y ∈ (Q : x). Hence (Q : x) is p-primary. Tn Proposition 3.4.15. If i=1 Qi = 0 is a minimal primary decomposition of the zero ideal, Sn with Qi a pi-primary ideal for each 1 ≤ i ≤ n, then i=1 pi ⊆ D. T Proof. By minimality, for each 1 ≤ i ≤ n there is some xi ∈ j6=i Qj such that xi 6∈ Qi. Then by Lemma 3.4.13(c) and (d), n ! n \ \ AnnA(xi) = (0 : xi) = Qj : xi = (Qj : xi) = (Qi : xi). j=1 j=1 By Lemma 3.4.14, (Qi : xi) is pi-primary for each 1 ≤ i ≤ n. Moreover, pi = r((Qi : xi)) = Sn r(AnnA(xi)) ⊆ D by Lemma 3.4.13(e), so i=1 pi ⊆ D as claimed. Corollary 3.4.16. If A is noetherian and p ⊂ A is a minimal prime ideal, then p ⊆ D. Proof. By Theorem 3.4.6, (0) has a primary decomposition and we may assume it is minimal Tn by Lemma 3.4.11. Set 0 = i=1 Qi where Qi ⊂ A is a pi-primary ideal for some prime ideal pi. By Proposition 3.4.10, the minimal prime ideals of A are elements of {pi | 1 ≤ i ≤ n} and by Proposition 3.4.15, each pi ⊆ D. Thus all minimal prime ideals of A are contained in D. 51 3.4 Primary Decomposition 3 Noetherian and Artinian Rings Lemma 3.4.17. If Qi is a pi-primary ideal for 1 ≤ i ≤ n and pi + pj = A whenever i 6= j, then n n Y \ Qi = Qi. i=1 i=1 Qn Tn Proof. We always have i=1 Qi ⊆ i=1 Qi so it suffices to show the reverse inclusion. First note that by Lemma 1.1.4(f), whenever j 6= i, we have r(Qi + Qj) = r(r(Qi) + r(Qj)) = r(pi + pj) = r(A) = A. So 1 ∈ r(Qi + Qj) which implies Qi + Qj = A. Now if n = 2 and x ∈ Q1 ∩ Q2 then the previous statement tells us that there exist q1 ∈ Q1, q2 ∈ Q2 such that x = x · 1 = x(q1 + q2) = xq1 + xq2 ∈ Q1Q2. Therefore the statement holds when n = 2. To induct, we need to show that Q1 ··· Qn−1 + Qn = A. By hypothesis, for each 1 ≤ i ≤ n − 1 there are elements xi ∈ Qi, yi ∈ Qn−1 such that xi + yi = 1. Then n−1 Y 1 = (xi + yi) ∈ x1 ··· xn−1 + Qn ⊆ Q1 ··· Qn−1 + Qn. i=1 Therefore Q1 ··· Qn−1 + Qn = A. Finally, using the base case and inductive hypothesis, we have n n−1 ! n−1 ! n−1 ! \ \ \ Y Qi = Qi ∩ Qn = Qi ∩ Qn = Qi Qn. i=1 i=1 i=1 i=1 Hence the statement holds for any n. In the next proposition, we generalize the correspondence of Proposition 1.3.8 to a bijec- tion on primary ideals. Proposition 3.4.18. Let S ⊆ A be a multiplicatively closed subset and suppose Q ⊂ A is a p-primary ideal. Then (a) If S ∩ p 6= ∅ then S−1Q = S−1p. (b) If S ∩ p = ∅ then j−1(S−1Q) = Q and S−1Q is S−1p-primary. (c) The correspondence Q0 ⊂ A is primary and Φ: ←→ {primary ideals of S−1A} r(Q0) ∩ S = ∅ given by Q 7→ S−1Q is bijective and inclusion-preserving. 52 3.5 Artinian Rings 3 Noetherian and Artinian Rings n sn −1 Proof. (a) Take s ∈ S ∩ p. Then for some n ∈ N, s ∈ Q ∩ S which implies that 1 ∈ S Q, sn −1 −1 −1 but 1 is a unit in S A, so we must have S Q = S A. −1 −1 q −1 −1 −1 (b) The inclusion Q ⊆ j (S Q) is trivial: if q ∈ Q then 1 ∈ S Q, so q ∈ j (S Q) by −1 −1 x −1 definition of the map j. For the reverse inclusion, suppose x ∈ j (S Q). Then 1 ∈ S Q x q so there exist elements q ∈ Q, s ∈ S such that 1 = s . Hence for some t ∈ S, t(xs − q) = 0, i.e. stx = tq ∈ Q. Now stx ∈ Q but st ∈ S, so since S ∩ p = ∅, we must have st 6∈ p. Since Q is p-primary, this implies x ∈ Q. This proves Q = j−1(S−1Q). Next, r(Q) = p by assumption and r(S−1Q) = S−1(r(Q)) = S−1p, so it remains to −1 x y −1 xy −1 show S Q is a primary ideal. Suppose s , t ∈ S A such that st ∈ S Q. Then there are xy q elements u ∈ S, q ∈ Q such that st = u . Thus for some v ∈ S, v(uxy − stq) = 0, that is, vuxy = stq ∈ Q. Since vu ∈ S by multiplicative closure, vu 6∈ p as before, so xy ∈ Q because x −1 y −1 −1 Q is primary. Thus x ∈ Q or y ∈ r(Q), which implies s ∈ S Q or t ∈ S (r(Q)) = S p. Hence S−1Q is S−1p-primary. (c) By part (b), Φ is injective and for any primary ideal Q0 ⊂ A, S−1Q0 is primary in S−1A. Inclusion-preserving is as in Proposition 1.3.8. Suppose J ⊂ S−1A is a primary ideal. To show Φ is surjective, it’s enough to show I = j−1(J) is a primary ideal of A, since then Proposition 1.3.8(a) says that S−1I = S−1(j−1(J)) = J. Take x, y ∈ A such that xy ∈ I xy −1 x y but x 6∈ I. Then 1 ∈ S I = J so since J is primary, 1 ∈ J or 1 ∈ r(J). The former x −1 y is impossible, since 1 ∈ J = S I would imply x ∈ I. So we must have 1 ∈ r(J). Then yn n for some n ≥ 1, 1 ∈ J, and thus y ∈ I. Thus I is primary. We conclude that Φ is an inclusion-preserving bijection. 3.5 Artinian Rings Recall from Section 3.1 that a (commutative) ring A is artinian if the descending chain condition holds on the set of ideals of A. We have seen many examples of noetherian rings, but it turns out that the artinian condition is much more restrictive. Assume that A 6= 0. Example 3.5.1. If k is a field and A is a finite dimensional k-algebra, then A is artinian. Proposition 3.5.2. Let A be artinian. Then (a) If A is an integral domain then A is a field. (b) dim A = 0. (c) J(A) = N(A). Proof. (a) Let x ∈ A, x 6= 0, be given. Then the ideals generated by powers of x form a descending chain in A: (x) ⊇ (x2) ⊇ (x3) ⊇ · · · By the descending chain condition, there is some n ∈ N such that (xn) = (xn+1). So there is a y ∈ A such that xn = xn+1y, that is, xn(1 − xy) = 0. Since A is a domain, xn 6= 0 so we must have 1 − xy = 0. Hence x is a unit. (b) If A is any artinian ring with a prime ideal p ⊂ A, then A/p is also artinian by Lemma 3.1.3. Since A/p is a domain, by (a) it is a field, so p is maximal. Thus all prime ideals in A are maximal, and thus also minimal, so dim A = 0. 53 3.5 Artinian Rings 3 Noetherian and Artinian Rings (c) follows directly from the definitions of J(A) and N(A) as intersections of maximal and prime ideals. Lemma 3.5.3. If A is artinian then Spec(A) is finite. Proof. Consider the collection Σ = {p1 ∩ · · · ∩ pn | pi ⊂ is a prime ideal and n ≥ 1}. Since A 6= 0, Σ is nonempty. By Lemma 3.1.1, Σ has a minimal element, say I = p1 ∩ · · · pn. For any prime ideal p ∈ Spec(A), I ∩ p ∈ Σ but by minimality, I = I ∩ p, so p ⊇ I. Since p is prime, we must have p ⊇ pi for some 1 ≤ i ≤ n. Now all prime ideals are maximal by Proposition 3.5.2(b), so p = pi. Hence Spec(A) = {p1,..., pn}. Proposition 3.5.4. If A is artinian, then N(A)n = 0 for some n ∈ N. Proof. Consider the descending chain N(A) ⊇ N(A)2 ⊇ N(A)3 ⊇ · · · By DCC, there is an n ∈ N such that N(A)m = N(A)n for all m ≥ n. Set I = N(A)n and assume I 6= 0. Then the collection Σ = {ideals J ⊂ A | IJ 6= 0} is nonempty. By Lemma 3.1.1, Σ contains a minimal element J0. Then IJ0 6= 0 so there exists an element x ∈ J0 so that xI 6= 0. This means (x) ∈ Σ, but by minimality of J0, we must have (x) = J0. Now I2 = N(A)2n = N(A) = I so we have 0 6= xI = xI2 = (xI)I. Thus xI ∈ Σ and xI ⊆ (x) = J0. By minimality, this implies xI = J0 = (x), and thus there is some y ∈ I so that x = xy. By induction, x = xym for all m ∈ N, but y ∈ I ⊆ N(A) so y is nilpotent. Thus for some m, ym = 0, giving us x = xym = 0, contradicting the initial assumption that x 6= 0. Hence I = N(A)n = 0. Theorem 3.5.5 (Hopkins). For a (commutative) ring A, the following are equivalent: (i) A is artinian. (ii) A is noetherian and dim A = 0. Proof. By Proposition 3.5.2(b), A artinian directly implies dim A = 0. Assuming (i), Lemma 3.5.3 implies that A has finitely many maximal (and prime) ideals m1,..., mn. Then N(A) = J(A) = m1 ∩ · · · mn ⊇ m1 ··· mn. k k k By Proposition 3.5.4, there is some k ∈ N such that N(A) = 0, but (m1 ··· mn) ⊆ N(A) = 0 so by Proposition 3.1.7, A is noetherian. On the other hand, assuming (ii) implies that A has finitely many minimal prime ideals, by Proposition 3.4.10. But if dim A = 0, each minimal prime ideal is maximal as well. Thus by Corollary 3.4.2, N(A)k = 0 for some k ∈ N. Then once again the hypothesis of Proposition 3.1.7 is satisfied, so A is artinian. Theorem 3.5.6. Let A be a local noetherian ring with unique maximal ideal m. Then one of the following conditions holds: (i) mn 6= mn+1 for all n ∈ N, and thus A is not artinian. (ii) mn = 0 for some n ∈ N, in which case A is artinian. 54 3.5 Artinian Rings 3 Noetherian and Artinian Rings Proof. Consider the descending chain m ⊇ m2 ⊇ m3 ⊇ If mn 6= mn+1 for all n ∈ N, the chain is not stationary so A is not artinian. Clearly in this case mn 6= 0 for any n ∈ N, or else the chain would stabilize. Alternatively, if mn = mn+1 for some n, the locality condition implies mn+1 = J(A)mn. Since A is noetherian, mn is a finitely generated submodule of A, so Nakayama’s Lemma (1.2.1) implies mn = 0. Now if p ⊂ A is a prime ideal, 0 = mn ⊆ p so we must have m ⊆ p; otherwise, if x ∈ m r p then xn ∈ p, which is impossible. By maximality, m = p so m is in fact the unique prime ideal of A. Hence dim A = 0, so by Theorem 3.5.5, A is artinian. Examples. 1 If k is a field, A = k[[t]] is noetherian by Theorem 3.3.9. Also, A is local with maximal ideal m = (t). In this ring, (0) is prime so dim A = 1. Therefore A cannot be artinian, so by Theorem 3.5.6, mn = (tn) 6= 0 for all n ∈ N. 2 Choose a prime p ∈ Z, an integer n ∈ N and set A = Z/pnZ. Since A is finite, it is local artinian with maximal ideal pZ/pnZ. 3 Let k be a field and consider the polynomial ring B = k[t]. If f ∈ B is a prime element, i.e. an irreducible polynomial in t, then A = B/(f n) is a local artinian ring with maximal ideal (f)/(f n). Proposition 3.5.7. Let A be a local artinian ring with maximal ideal m and residue field k = A/m. Then the following are equivalent: (a) Every ideal of A is principal. (b) m is principal. 2 2 (c) dimk m/m ≤ 1, where m/m is viewed as a k-vector space. Proof. (a) =⇒ (b) is trivial and (b) =⇒ (c) follows from the fact that if m = Ax, then m/m2 = k(x + m2) which has dimension at most 1. 2 To prove (c) =⇒ (a), note that if dimk m/m ≤ 1 then there is some x ∈ m such that m/m2 = k(x+m2). By Theorem 3.5.5, A is noetherian so in particular m is finitely generated as an A-module. Thus by Nakayama’s Lemma (1.2.1), with M = N = J(A) = m, we get N/MN = m/m2 =⇒ m = (x). Now Proposition 3.5.4 implies there exists an n ∈ N such that mn = N(A)n = 0. Take a nonzero ideal I ⊂ A. Then I ⊆ m and there is an ` ∈ N such that I ⊆ m` = (x`) but I 6⊆ m`+1 = (x`+1). So there is a y ∈ I such that y = ax` for some a ∈ A r m. Since A is local, this a must be a unit in A. Thus we have (y) ⊆ I ⊆ m` = (x`) = (y), so I = (y) and I = m` as claimed. Proposition 3.5.8. Let k be a field and A a finitely generated k-algebra which is also a local artinian ring. Then dimk A is finite. 55 3.6 Associated Primes 3 Noetherian and Artinian Rings Proof. For any local ring A which is finitely generated as a k-algebra, we have surjections k[t1, . . . , tr] → A → A/m so that A/m is also a finitely generated k-algebra. By the algebraic Nullstellensatz (2.4.2), A/m is a finite (algebraic) extension. Since A is artinian, A is also noetherian by Theorem 3.5.5. Then Theorem 3.5.6 shows that mn = 0 for some n ∈ N, in which case the chain of submodules m` ⊇ m`+1 eventually stabilizes: A ⊇ m ⊇ m2 ⊇ · · · ⊇ mn−1 ⊇ mn = 0. For each 0 ≤ ` ≤ n − 1, m` is a finitely generated A-module (since A is noetherian), so each quotient m`/m`+1 is a finite dimensional A/m-vector space. By the first paragraph and transitivity of dimension, we then have that ` `+1 ` `+1 dimk m /m = (dimk A/m)(dimA/m m /m ) < ∞ for each 0 ≤ ` ≤ n − 1. In particular, taking ` = n − 1 shows that mn−1/mn = mn−1 is finite dimensional over A/m. From linear algebra, (∗) if W ⊆ V is a vector subspace over a field F such that W and V/W are both finite dimensional F -vector spaces, then V is finite dimensional. Taking V = mn−2 and W = mn−1, we now have that mn−2 is a finite dimensional A/m-vector space. By induction, m` is finite dimensional for all 1 ≤ ` ≤ n−1. In particular, m1 = m is finite dimensional over A/m, and of course A/m is finite dimensional over itself, so by (∗), A is a finite dimensional A/m-vector space. This proves the statement. The structure of commutative artinian rings is completely described in the next theorem. Theorem 3.5.9. Any artinian ring A is the finite direct product of local artinian rings, and the factors are uniquely determined by A up to isomorphism. Corollary 3.5.10. If A is a finitely generated k-algebra which is artinian, then dimk A is finite. Proof. By Theorem 3.5.9, A is the direct product of finitely many local artinian rings, and each of these has finite dimension by Proposition 3.5.8. 3.6 Associated Primes Definition. A prime filtration of an A-module M is a sequence of submodules 0 = M0 ⊆ M1 ⊆ · · · ⊆ Mn = M ∼ such that for each 0 ≤ i ≤ n − 1, Mi+1/Mi = R/pi+1 for some prime ideals p1,..., pn ⊂ A. Definition. Let M be an A-module. A prime ideal p ⊂ A is an associated prime of M if there is an embedding of A-modules A/p ,→ M. Let Ass(M) denote the set of all associated primes of M. Lemma 3.6.1. For any A-module M, Ass(M) = {p ⊂ A prime | there is some nonzero x ∈ M with AnnA(x) = p}. 56 3.6 Associated Primes 3 Noetherian and Artinian Rings Proof. If p ∈ Ass(M) then x = 1 + p ∈ A/p ,→ M is a nonzero element of M and by definition AnnA(x) = p. Conversely, if p = AnnA(x) for some x ∈ M, then the natural map ∼ ϕx : A → M, a 7→ ax has kernel p so it factors through A/p = im ϕx ⊆ M. We will prove that every A-module has a prime filtration, but first we need a lemma. Lemma 3.6.2. For all nonzero A-modules M, Ass(M) is nonempty. Proof. Consider the set Λ = {I = AnnA(x) | x ∈ M, x 6= 0}. Then since M 6= 0, Λ is nonempty. Let I = AnnA(x) be a maximal element of Λ; we claim I is a prime ideal of A. Suppose a, b ∈ A such that ab ∈ I, so that abx = 0. If a, b 6∈ I then ax 6= 0, but then b ∈ AnnA(ax) which shows I ( AnnA(ax). This contradicts maximality of I, so either a or b lies in I, showing I is prime. Theorem 3.6.3. If A is noetherian, then every finitely generated A-module has a prime filtration. Proof. Let M be a finitely generated A-module and consider the collection of submodules Γ = {N ⊆ M | N has a prime filtration}. By Lemma 3.6.2, there exists a prime p ∈ Ass(M) so the submodule A/p ,→ M has trivial prime filtration 0 ⊆ A/p. Thus Γ is nonempty. Let N be a maximal element of Γ, say with prime filtration 0 ⊆ N1 ⊆ · · · ⊆ Nk = N. If N 6= M, use Lemma 3.6.2 again to choose p ∈ Ass(M/N) with p = AnnA(x + N) for some x ∈ M r N. Then N + Ax has a prime filtration: 0 ⊆ N1 ⊆ · · · ⊆ Nk = N ( N + Ax ∼ with (N + Ax)/N = AnnA(x + N) = p, contradicting maximality of N. Hence M = N. Lemma 3.6.4. For every short exact sequence of A-modules 0 → K → M → N → 0, Ass(M) ⊆ Ass(N) ∪ Ass(K). Proof. Suppose p ∈ Ass(M) so that by Lemma 3.6.1, p = AnnA(y) for some nonzero y ∈ M. If y ∈ N then p ∈ Ass(N) already. Otherwise, u = y + N ∈ M/N is a nonzero element of K. Suppose p 6= AnnA(u). Then there is some s ∈ AnnA(u) such that s 6∈ p and 0 = su = sy+N in K, i.e. sy ∈ N. Then p = AnnA(y) ⊆ AnnA(sy). On the other hand, if t ∈ AnnA(sy) then sty = 0, so st ∈ p but since s 6∈ p and p is prime, this means t ∈ p. Hence p = AnnA(y) or AnnA(sy), so p lies in either Ass(K) or Ass(N). Corollary 3.6.5. For any A-module M, Ass(M n) = Ass(M) for all n ≥ 1. Example 3.6.6. If A is an integral domain, then Ass(A) = {0} (by Lemma 3.6.1 for example). 57 3.6 Associated Primes 3 Noetherian and Artinian Rings Theorem 3.6.7. Let A be a noetherian ring and M a finitely generated A-module. Then (1) Let p ⊂ A be a prime ideal. Then p ∈ Ass(M) if and only if pAp ∈ Ass(Mp). (2) | Ass(M)| < ∞. (3) If p is minimal among prime ideals containing AnnA(M) then p ∈ Ass(M). Proof. (1) On one hand, if p ∈ Ass(M) then A/p ,→ M by definition, but since localization is exact, (A/p)p ,→ Mp as well so pAp ∈ Ass(Mp). Conversely, suppose pAp ∈ Ass(Mp). By Lemma 3.6.1, pAp = AnnAp (α) for some nonzero α ∈ Mp. Write p = (x1, . . . , xn) so that xiα = 0 for all 1 ≤ i ≤ n. Then there exist elements si ∈ A r p such that sixiα = 0. Set s = s1 ··· sn. Since p is prime, s 6∈ p and we have sα ∈ M and xi(sα) = sxiα = 0 for each i. Thus p ⊆ AnnA(sα). If t ∈ AnnA(sα) r p then we would have (ts)α = 0 but this implies α = 0 in Mp, a contradiction. Hence p = AnnA(sα) so p ∈ Ass(M). (2) By Theorem 3.6.3, M has a prime filtration 0 = M0 ⊆ M1 ⊆ · · · ⊆ Mn = M ∼ ∼ with Mi+1/Mi = A/pi+1 for some primes p1,..., pn ⊂ A. If n = 1, notice that M1 = A/p1 is a domain so Ass(M1) = {p1} by Example 3.6.6. For n ≥ 2, we have a short exact sequence 0 → Mn−1 → M → M/Mn−1 → 0 so Lemma 3.6.4 says that Ass(M) ⊆ Ass(Mn−1) ∪ Ass(M/Mn−1). By induction, Ass(Mn−1) is finite, and Ass(M/Mn−1) = Ass(A/pn) = {pn} so Ass(M) is finite as well. (3) Suppose p is minimal among all prime ideals of A containing AnnA(M). Then Mp 6= 0 and Mp is an (A/ AnnA(M))p-module, so we may assume AnnA(M) = 0. Now p is minimal, so Spec(Ap) = {pAp}. By (1), we may replace A by Ap and M by Mp. Consider the filtration M ⊇ pM ⊇ p2M ⊇ · · · Since A is noetherian and Spec(A) = {p}, Theorem 3.5.5 shows that A is artinian and thus pn+1M = pnM for some smallest n ≥ 1. Then p(pnM) = pnM so Nakayama’s lemma (1.2.1) n n−1 says that p M = 0. Take y ∈ p M, y 6= 0. Then p ⊆ AnnA(y) but since Spec(A) = {p}, p is a maximal ideal and hence p = AnnA(y). So by Lemma 3.6.1, p ∈ Ass(M). Remark. The proof of (2) shows that if p1,..., pn are the primes corresponding to a prime filtration of M, then Ass(M) ⊆ {p1,..., pn}. Example 3.6.8. Let A = k[x, y](x2 − y3) and define an A-module M by the free resolution 2x −3y2 0 → A −−−−−−→ A2 → M → 0. (M is called the module of K¨ahlerdifferentials of A). We claim the only associated primes of M are 0 and (x, y). Let p be a prime ideal of A and consider the localized sequence 2 0 → Ap → Ap → Mp → 0 58 3.6 Associated Primes 3 Noetherian and Artinian Rings which remains exact since localization is exact. If p 6= (x, y) then x 6= 0 or y 6= 0, so 2x 1 0 the matrix can be put into the form or after a change of basis. Thus −3y2 0 1 ∼ Mp = Ap so Ass(Mp) = {0} and thus by Theorem 3.6.7(1), p 6∈ Ass(M). On the other hand, if p = (x, y) then consider the element 2x3 m = ∈ M. −3y5 Notice that 2x4 2x4 2x2 xm = = = x2 = 0 −3xy5 −3x2y2 −3y2 2x3y 2x3y 2x while ym = = = x2y = 0. −3y6 −3x2y3 −3y2 Hence AnnA(m) = (x, y) so (x, y) is an associated prime of M. Proposition 3.6.9. For a noetherian ring A and a finitely generated A-module M, [ p = {a ∈ A | ax = 0 for some nonzero x ∈ M}. p∈Ass(M) Proof. Let Z be the set of all a ∈ A for which ax = 0 for some x 6= 0 in M. It is clear that any p ∈ Ass(M) is contained in Z, so one inclusion is true. On the other hand, suppose a ∈ Z with ax = 0 for some nonzero x ∈ M. The set Λ = {AnnA(bx) | b ∈ A, b 6= 0} is nonempty since AnnA(ax) ∈ Λ, so it has a maximal element, say q = AnnA(bx) for some b 6= 0. If we can show q is prime, then we are done by Lemma 3.6.1. Suppose uv ∈ q but u 6∈ q. Then ubx 6= 0 and q ⊆ AnnA(ubx), so by maximality, q = AnnA(ubx). Hence v ∈ AnnA(ubx) = q which shows q is prime as desired. 59 4 Discrete Valuations and Dedekind Domains 4 Discrete Valuations and Dedekind Domains 4.1 Discrete Valuation Rings Definition. A discrete valuation on a field K is a surjective group homomorphism v : K× → Z such that v(x + y) ≥ min(v(x), v(y)) for all x, y ∈ K×. We often extend v to all of K by setting v(0) = ∞. 0 Lemma 4.1.1. Let A be an integral domain with field of fractions K. If v : A r {0} → N0 is a function satisfying (i) v0(xy) = v0(x) + v0(y) (ii) v0(x + y) ≥ min(v0(x), v0(y)) for all x, y ∈ A r {0} and v0(0) = ∞ when necessary, then there exists a unique discrete × 0 0 valuation v : K → Z which extends v , i.e. v|A = v . × a 0 0 Proof. Define v : K → Z by v b = v (a)−v (b). Then one can check that v is well-defined, a surjective homomorphism and satisfies v(x + y) ≥ min(v(x), v(y)). Corollary 4.1.2. If A is a UFD with field of fractions K and p ∈ A is a prime element, × n there is a discrete valuation vp : K → Z such that for any a ∈ A, vp(a) = n if p | a and pn+1 - a. Proof. It’s enough to show that vp satisfies conditions (i) and (ii) of Lemma 4.1.1. Take x, y ∈ A r {0} and write x = pnx0 and y = pmy0 where n, m ≥ 0 and p - x0, y0. Then n+m 0 0 0 0 xy = p x y and since p is prime, p - x y . Thus vp(xy) = n + m, so (i) is satisfied. On the other hand, suppose without loss of generality that n ≥ m, so that min(v(x), v(y)) = m. Then x + y = pnx0 + pmy0 = pm(pn−mx0 + y0) but since p | pn−mx0 and p - y0, we have n−m 0 0 p - (p x + y ). Thus vp(x + y) = m and condition (ii) is satisfied. × Definition. For a prime element p in a UFD A, the valuation vp : K → Z is called the p-adic valuation on A (or on K). Examples. 1 For A = Z and K = Q, and p a prime integer, the classical p-adic valuation is defined a n by vp b = vp(a) − vp(b), where vp(a) = n if a = p m, p - m. 2 Let k be a field and consider the field of rational polynomials K = k(t). The so- × called “infinite place” of K is a discrete valuation v∞ : K → Z defined by the degree f f × function: v∞(f) = − deg f for f ∈ k[t] and v∞ g = deg g − deg f for g ∈ K . For A = k[t−1], the prime element t−1 induces the same valuation on Frac(A) = k(t): vt−1 = v∞. 60 4.1 Discrete Valuation Rings 4 Discrete Valuations and Dedekind Domains 3 For k a field, consider the field of rational power series −n K = k((t)) = {t f | n ∈ N0, f ∈ k[[t]]}. P∞ i Define a function v on K by v(x) = m ∈ Z if x = i=m ait for ai ∈ k and am 6= 0. Formally let v(0) = ∞ so that v : K → Z is a discrete valuation. As in 2 , the prime element t in the UFD A[[t]] induces this valuation. Definition. If v : K× → Z is a discrete valuation, then O(v) = {v ∈ K | v(x) ≥ 0} is called the valuation ring of v. Lemma 4.1.3. For any discrete valuation v : K× → Z, O(v) is a subring of K. Proof. By convention, v(0) = ∞ so 0 ∈ O(v). If x, y ∈ O(v) then by definition v(xy) = v(x)+v(y) ≥ 0 so xy ∈ O(v). Likewise, v(x+y) ≥ min(v(x), v(y)) ≥ 0 shows that xy ∈ O(v). Hence O(v) ⊆ K is a subring. Proposition 4.1.4. An integral domain A is a DVR if there exists a discrete valuation v on K = Frac(A) such that O(v) = A. Proof. Suppose v : K× → Z is a discrete valuation with A = O(v). By Lemma 4.1.3, A is a ring so it is enough to show that A is a DVR. If x ∈ K× such that x−1 6∈ A, then v(x−1) < 0. Note that the definition of a valuation forces v(1) = 0 since v(1) = v(1 · 1) = v(1) + v(1), so for such an x we have 0 = v(1) = v(xx−1) = v(x) + v(x−1) < v(x) + 0 and so v(x) > 0. Thus x ∈ A, so A is a DVR. Examples, continued. 1 For a prime p and the p-adic valuation vp : Q → Z, the valuation ring O(vp) = a b : a, b ∈ Z, b 6= 0, p - b coincides with Z(p), the p-adic integers described in Exam- ple 1.2.2. 2 For v∞ on k(t), the valuation ring is similar to the p-adic integers: f O(v ) = : f, g ∈ k[t], g 6= 0, deg g ≥ deg f . ∞ g 3 For the valuation v on k((t)), O(v) = k[[t]]. Proposition 4.1.5. Let v : K× → Z be a discrete valuation with valuation ring O. Then (1) O× = {x ∈ K | v(x) = 0}. (2) O is a local ring with unique maximal ideal m = O r O× = {x ∈ K | v(x) > 0}. (3) O is a DVR. 61 4.1 Discrete Valuation Rings 4 Discrete Valuations and Dedekind Domains (4) If π ∈ O satisfies v(π) = 1, then m = (π). (5) Every proper ideal I ⊂ O is of the form I = (πm) = {x ∈ K | v(x) ≥ m}, where m = min(v(x) | x ∈ I). (6) O is a PID with unique prime elements of the form uπ, where u is a unit. (7) Spec(O) = {(0), m} and hence dim O = 1. Proof. Exercise. Definition. For a valuation ring O = O(v), an element π ∈ O such that m = (π) is the unique maximal ideal of O is called a uniformizer of O. Remark. Suppose A is a local noetherian domain. (1) If dim A = 1 then the fact that A is local and a domain implies Spec(A) = {0, m}, where m is the maximal ideal of A. (2) Further, if I ⊂ A is a nonzero proper ideal then Spec(A/I) = {m/I} by the corre- spondence theorem. Thus dim A/I = 0. By Lemma 3.1.3, A noetherian implies A/I is noetherian as well. Thus by Theorem 3.5.5, A/I is also artinian. (3) Now m/I = N(A/I) so by Proposition 3.5.4, (m/I)` = 0 for some ` ∈ N, and hence m` ⊆ I ⊆ m. This shows that r(I) = m, i.e. that I is m-primary by Lemma 3.4.8. (4) Finally, since dim A = 1, Theorem 3.5.5 says that A is not artinian, so by Theorem 3.5.6, ` `+1 m 6= m for any ` ∈ N0. The next result completely characterizes local noetherian DVRs. Theorem 4.1.6. Let A be a local noetherian domain with maximal ideal m and residue field k = A/m, such that dim A = 1. Then the following are equivalent: (i) A is a DVR. (ii) A is integrally closed. (iii) m is a principal ideal. 2 (iv) dimk m/m = 1. (v) For all nonzero proper ideals I ⊂ A, there exists an n ∈ N such that I = mn. (vi) There exists an x ∈ A such that for any nonzero proper ideal I ⊂ A, I = (xn) for some n ∈ N. 62 4.1 Discrete Valuation Rings 4 Discrete Valuations and Dedekind Domains Proof. (i) =⇒ (ii) If A is a DVR then A is a PID by (6) of Proposition 4.1.5, so A is a UFD. By Proposition 2.1.8, this proves A is integrally closed. (ii) =⇒ (iii) Choose a nonzero element a ∈ m. By Remark (3), there is some ` ∈ N with m` ⊆ (a) and m`−1 6⊂ (a). Choose b ∈ m`−1 with b 6∈ (a) and set x = ab ∈ k. Then −1 b −1 x = a 6∈ A since b 6∈ (a). Since A is integrally closed, this means x is not integral over A. Suppose x−1m ⊂ m. Then m is an A[x−1]-module, m is faithful since m 6= 0 and A[x−1] ⊆ k, and m is finitely generated as an A-module because A is noetherian. Therefore the conditions of Lemma 2.1.2(iv) are satisfied, showing x−1 is integral over A, a contradiction. Thus we must have x−1m 6⊂ m. Now b 1 1 1 x−1m = m = (bm) ⊆ m` ⊆ (a) = A, a a a a showing x−1m is an ideal of A. But x−1m 6⊂ m so by maximality we must have x−1m = A, i.e. m = (x). 2 2 2 (iii) =⇒ (iv) If m = (x) then m/m = k(x + m ) as a k-vector space, so dimk m/m . By 2 2 (4) of the Remark, m 6= m , so we must have dimk m/m = 1. (iv) =⇒ (v) By (3) of the Remark, there exists an ` ∈ N such that m` ⊆ I ⊆ m. Without loss of generality, assume ` ≥ 2 and consider the quotient A/m`. Then by (2) of the Remark, A/m` is local artinian. Thus ` 2 ` 2 dimk(m/m )/(m /m ) = dimk m/m = 1 by the isomorphism theorems. By Proposition 3.5.7, this implies I/m` = (m/m`)n = mn/m` for some n ≤ `. However, I ⊇ m` and mn ⊇ m` so we must have I = mn. (v) =⇒ (vi) By Remark (4), m 6= m2 so we can find some element x ∈ m r m2. By assumption, there is some n ∈ N such that (x) = mn. Since x 6∈ m2, n has to equal 1, so m = (x). Hence for any nonzero ideal I ⊂ A, I = m` = (x`) for some ` ∈ N. (vi) =⇒ (i) Suppose m = (xn) for some x ∈ A and n ∈ N. Since m is maximal, x ∈ m so we have (x) ⊆ m = (xn) ⊆ (x) and thus m = (x), that is, m is principal. By Remark (4), ` `+1 (x ) 6= (x ) for any ` ∈ N0. This means for any nonzero element a ∈ A, there is a unique ` ` ∈ N0 such that (a) = (x ) by assumption. Define v : A r {0} → N0 by v(a) = ` where ` is ` chosen as described. This map is surjective because x 7→ ` for any ` ∈ N0. If a, b ∈ A such that (a) = (x`) and (b) = (xm), then (ab) = (x`)(xm) = (x`+m) =⇒ v(ab) = v(a) + v(b). Moreover, suppose v(a) = ` and v(b) = m. Then (a) = (x`) and (b) = (xm) so there exist u, w ∈ A× such that a = ux` and b = wxm. Without loss of generality assume ` ≤ m. Then a + b = ux` + wxm = x`(u + vxm−`) and u + vxm−` ∈ A, so this shows a + b ∈ (x`) and hence (a + b) ⊆ (x`). If (a + b) = (xr), we have (xr) ⊆ (x`) so in particular r ≥ ` and thus v(a + b) ≥ ` = v(a). Now by Lemma 4.1.1, v extends uniquely to a discrete valuation v : K× → Z. It remains to verify that O(v) = A and we will be finished. By construction A ⊆ O(v) since a v|Ar{0} : A r {0} → N0. On the other hand, if y ∈ O(v) is nonzero then y = b for some a, b ∈ A r {0}. Then 0 ≤ v(y) = v(a) − v(b), that is, v(a) ≥ v(b). If v(a) = ` and v(b) = m as above, then (a) = (x`) and (b) = (xm), and since ` ≥ m, we have (a) ⊆ (b). Hence a ∈ (b) a so y = b ∈ A. This completes the proof that A = O(v), i.e. A is a valuation ring. 63 4.2 Dedekind Domains 4 Discrete Valuations and Dedekind Domains 4.2 Dedekind Domains Definition. An integral domain A is called a Dedekind domain if (i) A is noetherian. (ii) dim A = 1, i.e. every prime ideal is maximal. (iii) A is integrally closed. Notice that the Krull dimension requirement means that a field is not a Dedekind domain. Example 4.2.1. Every PID that is not a field is a Dedekind domain. Dedekind domains are the central object in algebraic number theory, as one can construct a canonical Dedekind domain OK , called the ring of integers, for any number field K ⊇ Q. In this way the axioms and properties of Dedekind domains were devised in an attempt to generalize the algebraic properties of Z. Proposition 4.2.2. If A is a noetherian domain of Krull dimension 1, the following are equivalent: (i) A is integrally closed, and hence a Dedekind domain. (ii) Ap is a DVR for every nonzero prime ideal p ⊂ A. (iii) Every primary ideal in A is a prime power. Proof. (i) =⇒ (ii) It follows from Proposition 2.3.12 that Ap is integrally closed for every prime ideal p in A. Also, by Lemma 2.3.6, dim Ap = ht(p) = 1 and by Lemma 3.3.1, Ap is noetherian, so Ap is a DVR by Lemma 3.3.1. (ii) =⇒ (i) If Ap is a DVR then it is a PID, hence a UFD, hence integrally closed by Proposition 2.1.8. Since this holds for all nonzero prime ideals p, Proposition 2.3.12 implies that A is also integrally closed. (ii) =⇒ (iii) Take a p-primary ideal 0 6= Q ⊂ A so that p = r(Q). Then p is maximal −1 −1 since dim A = 1. For S = Arp, S Q is a proper ideal in S A = Ap. By Proposition 3.4.18, −1 −1 −1 −1 n S Q is S p-primary in Ap. Now Ap is a DVR, so S Q = (S p) for some n ∈ N according to Theorem 4.1.6. Since p is maximal, Lemma 3.4.8 says that pn is p-primary, and by Lemma 1.3.10, (S−1p)n = S−1pn, so Proposition 3.4.18 implies Q = pn. (iii) =⇒ (ii) We are assuming A is a noetherian domain, so Ap is a local noetherian domain for all nonzero prime ideals p ∈ Spec(A) = MaxSpec(A) by Lemma 3.3.1. Then dim A = 1 implies dim Ap = 1. We want to show Ap is a DVR, so by Theorem 4.1.6, it’s −1 enough to show that all nonzero ideals in Ap are powers of S p, where S = A r p. Take −1 some nonzero ideal I ⊂ Ap. By Remark (3) preceding Theorem 4.1.6, I is S p-primary so Proposition 3.4.18 says there exists a primary ideal Q ⊂ A, Q ∩ S = ∅, with S−1Q = I. Since Q ∩ S = ∅, Q ⊆ p and thus r(Q) = p since p is the only prime ideal of A containing Q. By assumption, Q = pn for some n ∈ N, and thus Lemma 1.3.10 gives us I = S−1Q = S−1pn = (S−1p)n. −1 Hence I is a power of S p and we conclude that Ap is a DVR. 64 4.3 Fractional and Invertible Ideals 4 Discrete Valuations and Dedekind Domains Corollary 4.2.3. If A is a Dedekind domain, then any nonzero ideal of A is a product of prime ideals in A. T` Proof. Suppose I ⊂ A is a nonzero ideal. Since A is noetherian, I = i=1 Qi for primary ideals Qi ⊂ A with r(Qi) = pi prime, by Theorem 3.4.6. By Lemma 3.4.11, we may take this to be a minimal primary decomposition of I, so pi 6= pj for i 6= j. Since dim A = 1, each T` Q` pi is maximal so pi + pj = A for all i 6= j. Hence I = i=1 Qi = i=1 Qi by Lemma 3.4.17. Now by Proposition 4.2.2(iii), each primary ideal Qi is a power of a prime ideal, namely its Q` ni radical r(Qi) = pi. So there exist n1, . . . , n` ∈ N such that I = i=1 pi . We will prove in the next section that nonzero ideals in a Dedekind domain factor into a unique product of prime ideals. This is the most important feature of Dedekind domains, as it allows us to generalize the unique factorization properties of Z to similar objects in number theory which turn out to be Dedekind domains. For example, taking the integral closure of Z in a finite separable extension of Q preserves the Dedekind property: Theorem 4.2.4. If A is a Dedekind domain with field of fractions K and L/K is a finite separable field extension, then the integral closure B of A in L is a Dedekind domain. Proof. First, B is integrally closed by Corollary 2.1.7. Moreover, since B/A is an integral extension, dim B = dim A = 1 by Theorem 2.3.7. Finally, since L/K is separable, Proposi- Pn tion 2.2.5 says that there is a K-basis {v1, . . . , vn} of L such that B ⊆ i=1 Avi. Since A is noetherian, we get that B is noetherian as well. Hence B is a Dedekind domain. 4.3 Fractional and Invertible Ideals In this section we prove that every nonzero ideal in a Dedekind domain factors uniquely into a product of prime ideals. In order to prove this, we need to expand the notion of an ideal to include other submodules of the fraction field. Let A be an integral domain with field of fractions K. Then A-ideals are also A-submodules of K. Definition. For two A-submodules I,J ⊆ K, the generalized ideal quotient of I by J with respect to K is (I :K J) = {x ∈ K | xJ ⊆ I}. Lemma 4.3.1. For every pair of submodules I,J ⊆ K, (I :K J) is an A-submodule of K. Lemma 4.3.2. Let I,J ⊆ K be submodules. If S is any multiplicatively closed subset of A, −1 −1 −1 then S (I :K J) = (S I :K S J) if J is a finitely generated A-module. Definition. A nonzero A-submodule I ⊆ K is called a fractional ideal of A if there is some x ∈ K× such that xI ⊆ A. Equivalently, I is fractional if there is an ideal J ⊂ A such that I = yJ for some y ∈ K×. Definition. A nonzero A-submodule I ⊆ K is an invertible ideal of A if there is an A-submodule J ⊆ K such that IJ = A. Proposition 4.3.3. Let I ⊆ K be a nonzero A-submodule. Then 65 4.3 Fractional and Invertible Ideals 4 Discrete Valuations and Dedekind Domains (a) If I is finitely generated then I is a fractional ideal. (b) If A is noetherian and I is fractional, then I is finitely generated. (c) If I is invertible with IJ = A, then J = (A :K I) and J is unique. (d) If I is invertible then I is finitely generated and therefore fractional. Pn ai Proof. (a) Let I = Axi where xi = for ai, bi ∈ A {0}. Then b1 ··· bnI ⊆ A so I is i=1 bi r fractional. (b) If A is noetherian, then any A-ideal J is finitely generated, so if I = yJ for J ⊂ A and y ∈ K× then I is clearly finitely generated as an A-module. (c) Suppose IJ = A for some submodule J ⊆ K. Then J ⊆ (A :K I) by definition of the generalized ideal quotient. On the other hand, (A :K I) = (A :K I)A = (A :K I)IJ ⊆ AJ = J. Therefore we have J = (A :K I) and this implies uniqueness since J was arbitrary. (d) Suppose IJ = A for some submodules J ⊆ A. Since 1 ∈ A = IJ, there are xi ∈ I Pn and yi ∈ J such that 1 = i=1 xiyi. For each x ∈ I, we have n n X X x = x · 1 = x xiyi = xi(xyi). i=1 i=1 Pn Pn Since each xyi ∈ IJ = A, we have i=1 xi(xyi) ∈ i=1 Axi ⊆ I. Therefore x ∈ I, so Pn I = i=1 Axi, meaning I is finitely generated. Finally, (a) implies that I is a fractional A-ideal. −1 If I is an invertible ideal, we will write I = (A :K I). When I ⊂ A is an ideal, we will call it an integral ideal to distinguish from fractional ideals that are not submodules of A. −1 Take a nonzero prime ideal p ∈ Spec(A) and consider the local ring Ap = S A, where −1 S = A r p. If I ⊆ K is an A-submodule, we will write Ip = S I. Proposition 4.3.4. For an A-submodule I ⊆ K, the following are equivalent: (i) I is an invertible ideal of A. (ii) I is finitely generated as an A-module and Ip is an invertible Ap-ideal for all prime ideals p ∈ Spec(A). (iii) I is a finitely generated A-module and Im is an invertible Am-ideal for all maximal ideals m ∈ MaxSpec(A). Proof. (i) =⇒ (ii) By Proposition 4.3.3(d), I is finitely generated as an A-module. If IJ = A for a submodule J ⊆ K, then by Lemma 1.3.10(a), we have Ap = IpJp so in particular Ip is an invertible ideal of Ap. (ii) =⇒ (iii) is trivial. 66 4.3 Fractional and Invertible Ideals 4 Discrete Valuations and Dedekind Domains (iii) =⇒ (i)) Set L = I(A :K I). Then L ⊆ A is an ideal. For any maximal ideal m ⊂ A, set S = A r m. Then −1 −1 −1 Lm = ImS (A :K I) = Im(S A :K S I) by Lemma 4.3.2 −1 = Im(Am :K Im) = ImIm = Am. −1 In particular Lm is not contained in the maximal ideal S m, so either L ∩ S 6= ∅ or L 6⊂ m. However, if Lm ⊂ Am is an ideal, then L ∩ S = ∅, so we must have L 6⊂ m. Since m is maximal, it follows that L = A, i.e. I is invertible. Proposition 4.3.5. For a local domain A which is not a field, A is a DVR if and only if every fractional ideal of A is invertible. Proof. (i) =⇒ (ii) Suppose A is a DVR. By Proposition 4.1.5, A is a PID so every fractional ideal of A is of the form Ax for some x ∈ K×. The inverse of Ax is Ax−1, so all fractional ideals of A are also invertible. (ii) =⇒ (i) Every nonzero (integral) ideal of A is fractional, so by assumption they are all invertible. Then Proposition 4.3.3(d) implies every ideal of A is finitely generated, so by Proposition 3.1.2, A is noetherian. Let m be the unique maximal ideal of A. Then there exists a fractional ideal m ⊆ K such that mm−1 = A. We claim that every ideal of A is of the form mn for some n ∈ N. Let J ⊂ A be a nonzero integral ideal. Then J ⊆ m so Jm−1 ⊆ mm−1 = A, which shows that Jm−1 is an integral ideal of A. Moreover, A ( m implies J ( Jm−1, for if J = Jm−1 then we would have Jm = JA = J, and thus m = 0 by Nakayama’s Lemma (1.2.1). Now starting with any ideal J ⊂ A, we get a strictly ascending chain of ideals −1 −2 J ( Jm ( Jm ( Since A is noetherian, the chain stabilizes, i.e. Im−n = A for some n ∈ N. Therefore I = mn for this n. This implies Spec(A) = {0, m} so dim A = 1. Finally, Theorem 4.1.6 shows that A is a DVR. Our results in the previous two sections combine to give us the following characterization of Dedekind domains. Theorem 4.3.6. For an integral domain A which is not a field, the following are equivalent: (i) A is a Dedekind domain. (ii) Every fractional ideal of A is invertible. Proof. (i) =⇒ (ii) Suppose A is Dedekind. Then by Proposition 4.2.2, Ap is a DVR for every prime ideal p ∈ Spec(A) = MaxSpec(A). Thus Proposition 4.3.5 implies every fractional ideal Ip is invertible (as an Ap-module) for any such prime ideal p. Finally, since A is noetherian, every ideal I is finitely generated (3.1.2) and hence Proposition 4.3.4 implies I is an invertible ideal. (ii) =⇒ (i) If every fractional ideal of A is invertible, then each is finitely generated by Proposition 4.3.3(d), so A is noetherian. We want to show Ap is a DVR for all maximal ideals m ∈ MaxSpec(A). By Proposition 4.3.5, it’s enough to show that every fractional 67 4.3 Fractional and Invertible Ideals 4 Discrete Valuations and Dedekind Domains 0 0 −1 ideal of Am is invertible. Let I ⊂ Am be a nonzero ideal. Then I = S I = Im for some nonzero ideal I ⊂ A. By assumption I is invertible, so Im is invertible. It follows that every fractional ideal of Am is invertible, so we have that ht(m) = dim Am = 1 by Lemma 2.3.6(a) and this holds for any maximal ideal m of A. Hence dim A = 1 so Proposition 4.2.2 tells us that A is Dedekind. Dedekind domains have the important property that every nonzero ideal factors uniquely into a product of prime ideals. Theorem 4.3.7. If A is a Dedekind domain then every nonzero ideal I ⊂ A factors uniquely Qm into a product of prime ideals I = i=1 pi, with repetitions allowed. Qm Proof. By Corollary 4.2.3, I factors as I = i=1 pi for nonzero prime ideals pi ⊂ A that Qn need not be distinct. Thus it remains to check uniqueness. Suppose that I = j=1 qj for prime ideals qj ⊂ A. Then q1 ··· qn ⊆ p1 so there exists a 1 ≤ j ≤ n such that qj ⊆ p1. Since primes are maximal in a Dedekind domain, qj = p1. We may relabel the qj so that q1 = p1. −1 By Theorem 4.3.6, p1 exists so we have p1p2 ··· pm = p1q2 ··· qn −1 −1 =⇒ p1 p1p2 ··· pm = p1 q2 ··· qn =⇒ p2 ··· pm = q2 ··· qn. By induction, m = n and after relabelling, qi = pi for all 1 ≤ i ≤ m. Hence prime factorizations are unique up to reordering. Qm ei Qm fi Proposition 4.3.8. Take nonzero ideals I,J ⊂ A with I = i=1 pi and J = i=1 pi for distinct primes p1,..., pm ⊂ A and exponents ei, fi ≥ 0. Prove the following: (a) I ⊆ J if and only if ei ≥ fi for all i. Qm max(ei,fi) (b) I ∩ J = i=1 pi . Qm min(ei,fi) ei fj (c) I + J = i=1 pi . In particular, pi + pj = A for all ei and fj when i 6= j. ei fi Proof. (a) One direction is clear: if ei ≥ fi for all i then pi ⊆ pi for all i, and hence I ⊆ J. Qm fi f1 On the other hand, suppose I ⊆ J. Then I ⊆ i=1 pi ⊆ p1 . By Theorem 4.3.6, p1 is −1 invertible, so we may multiply by p1 to obtain: f1 e1 e2 em f1 I ⊆ p1 =⇒ p1 p2 ··· pm ⊆ p1 −e1 e1 e2 em −e1 f1 =⇒ p1 p1 p2 ··· pm ⊆ p1 p1 e2 em f1−e1 =⇒ p2 ··· pm ⊆ p1 . f1−e1 If f1 − e1 > 0 then p1 ⊆ p1. Since p1 is prime, the above implies p1 ⊇ pj for some 2 ≤ j ≤ m. But prime ideals are maximal in a Dedekind domain, so p1 = pj, contradicting the fact that p1,..., pm are distinct primes. Therefore f1 − e1 ≤ 0. Repeating the proof for each pi, 2 ≤ i ≤ m shows that ei ≥ fi for all i. 68 4.3 Fractional and Invertible Ideals 4 Discrete Valuations and Dedekind Domains (b) Using unique factorization of ideals, write n n n n Y ei Y fi Y gi Y hi I = pi ,J = pi ,I ∩ J = pi ,I + J = pi i=1 i=1 i=1 i=1 for distinct prime ideals p1,..., pm, pm+1,..., pn, n ≥ m, and unique exponents ei, fi, gi, hi ≥ 0, with ei = fi = 0 for m < i ≤ n. Then by (a), we get the following implications: I ∩ J ⊆ I =⇒ gi ≥ ei for all 1 ≤ i ≤ n and I ∩ J ⊆ J =⇒ gi ≥ fi for all 1 ≤ i ≤ n. So gi ≥ max(ei, fi) for all 1 ≤ i ≤ n. This shows gi ≥ max(ei, fi) ≥ ei and gi ≥ max(ei, fi) ≥ fi for all 1 ≤ i ≤ n, so because I ∩ J is the largest ideal of A which is contained in both I and J, part (a) implies n n Y max(ei,fi) Y max(ei,fi) I ∩ J ⊆ pi ⊆ I ∩ J =⇒ I ∩ J = pi . i=1 i=1 (c) As in part (b), we have: I ⊆ I + J =⇒ ei ≥ hi for all 1 ≤ i ≤ n and J ⊆ I + J =⇒ fi ≥ hi for all 1 ≤ i ≤ n. So min(ei, fi) ≥ hi for all 1 ≤ i ≤ n, meaning ei ≥ min(ei, fi) ≥ hi and fi ≥ min(ei, fi) ≥ hi. Since I + J is the smallest ideal of A containing both I and J, part (a) means that n n Y min(ei,fi) Y min(ei,fi) I ⊆ pi ⊆ I + J and J ⊆ pi ⊆ I + J i=1 i=1 n Y min(ei,fi) imply I + J ⊆ pi ⊆ I + J i=1 n Y min(ei,fi) so I + J = pi . i=1 Qm ei Theorem 4.3.9 (Chinese Remainder Theorem). If I = i=1 pi for distinct prime ideals pi ⊂ A and exponents ei ≥ 0, then there is a canonical isomorphism m Y ei A/I −→ A/pi . i=1 Remark. Let A be a Dedekind domain and take a nonzero prime (maximal) ideal p ∈ Qm ei Spec(A), so that Ap is a DVR. For 0 6= x ∈ A, we have a prime factorization (x) = i=1 pi with p1 = p and ei ≥ 0. The ei are determined by (x), so define a function vp : A r {0} −→ N0 x 7−→ e1. 69 4.4 The Class Group 4 Discrete Valuations and Dedekind Domains Then vp extends to a valuation on K = Frac(A) such that O(vp) = Ap. As this generalizes the construction of the p-adic valuation on Q, the function vp is called the p-adic valuation on K. Proposition 4.3.10. If A is a Dedekind domain with Spec(A) finite, then A is a PID. Proof. Let p1,..., pn be the prime ideals of A. By the Chinese remainder theorem, 2 ∼ 2 A/p1p2 ··· pn = A/p1 × A/p2 × · · · × A/pn. 2 2 2 In particular, for any fixed x1 ∈ p1 r p1, there exists y ∈ A that satisfies y + p1 = x1 + p1 2 and y + pi = 1 + pi for all 2 ≤ i ≤ n. Thus y − x1 ∈ p1 ⊂ p1 so it follows that y ∈ p1. We 2 have (y) ⊆ p1 but (y) 6⊂ p1 and (y) 6⊂ pi for any 2 ≤ i ≤ n. Since A is a Dedekind domain, (y) has a prime factorization n Y ei (y) = pi for ei ≥ 0. i=1 By Proposition 4.3.8(a) and the above work, we conclude that 1 ≤ e1 < 2 and 0 ≤ ei < 1 for each 2 ≤ i ≤ n. In other words, e1 = 1 and e2 = ... = en = 0, so (y) = p1. Repeating the proof for each pi, 1 ≤ i ≤ n shows that every prime ideal of A is principal. Suppose pi = (yi). Then any ideal I ⊂ A has a prime factorization n n Y ai Y ai a1 an I = pi = (yi ) = (y1 ··· yn ). i=1 i=1 Therefore I is principal. Corollary 4.3.11. Let I be a nonzero ideal in a Dedekind domain A. Then for any nonzero a ∈ I, there is some b ∈ I such that I = (a, b). Proof. We first prove A/I is a PID. By the correspondence theorem, every prime ideal of A/I pulls back along the quotient map A → A/I to a unique prime ideal of A which contains I. Qr aj Write I = j=1 pj for prime ideals pj ⊂ A and integers aj ≥ 1. Then Proposition 3.4.10(a) and the fact that dim A = 1 imply that the only prime ideals containing I are p1,..., pr. It follows that A/I has finitely many prime ideals. Therefore by Proposition 4.3.10, A/I is a PID. Now take 0 6= a ∈ I. Then by the first paragraph, A/(a) is principal and therefore the ideal I/(a) can be generated by a single element b+(a), b ∈ I. (If b ∈ (a), this means I = (a) so I was principal to begin with.) On one hand, it’s immediate that (a, b) ⊆ I. On the other hand, every y ∈ Ir(a) has reduction y+(a) ⊆ I/(a) in A/(a), so y+(a) = r(b+(a)) = rb+(a) for some r ∈ A, and thus y − rb ∈ (a), so y = rb + sa ∈ (a, b). This shows I = (a, b). 4.4 The Class Group Definition. For a Dedekind domain A with field of fractions K, the ideal group of A is the free abelian group IA = {I ⊂ K | I is a fractional ideal of A} = {I ⊆ K | I is an invertible ideal}. 70 4.4 The Class Group 4 Discrete Valuations and Dedekind Domains × The subgroup PA = {Ax | x ∈ K } ≤ IA is called the subgroup of principal fractional ideals of A. The class group of A is then defined as the quotient group CA = IA/PA. The order of this group, denoted h(A), is called the class number of A. In general, class numbers can be arbitrarily large, but a highly important fact for number fields is that their rings of integers have finite class number: Theorem 4.4.1. For an algebraic extension K/Q, with ring of integers OK , the class number hK := h(OK ) is finite. Lemma 4.4.2. For a Dedekind domain A, the following are equivalent: (i) h(A) = 1. (ii) A is a PID. (iii) A is a UFD. Proof. (i) ⇐⇒ (ii) Suppose h(A) = 1. Then for any ideal I ⊂ A, I is a fractional ideal (e.g. by Proposition 4.3.3), but since the class group of A is trivial, I must be principal. Hence A is a PID. Conversely, suppose every ideal I ⊂ A is principal. Any fractional ideal Qr ei J ∈ IA may be written as a formal product of prime ideals J = i=1 pi with ei ∈ Z for each i. Then each prime ideal is principal: pi = (xi) for an element xi ∈ A, 1 ≤ i ≤ r, so we have r r Y ei Y ei e1 er J = pi = (xi) = Ax1 ··· xr . i=1 i=1 Therefore J is a principal fractional ideal, and it follows that PA = IA, that is, h(A) = |IA/PA| = 1. (ii) ⇐⇒ (iii) It is well-known that every PID is a UFD. For the converse, suppose p ⊂ A is a (proper) nonzero prime ideal. Choose 0 6= x ∈ p (which is not a unit since p 6= A) and e1 er write x = p1 ··· pr for distinct irreducible elements pi ∈ A and exponents ei ∈ N. Since p is a prime ideal, pi ∈ p for some 1 ≤ i ≤ r. Then (pi) ⊆ p. Since pi is irreducible, (pi) is a prime ideal, so the fact that dim A = 1 implies (pi) = p. Thus every nonzero prime ideal of A is principal. This implies the result by unique factorization of ideals. √ Example√ 4.4.3. Let K = Q( −5) be the quadratic number field with ring of integers OK = Z[ −5]. Then OK is a Dedekind domain, but it is not a UFD, since for example √ √ 6 = 2 · 3 = (1 + −5)(1 − −5) and each of the elements in the two factorizations is irreducible but not prime. However, since OK is Dedekind, the ideal (6) factors uniquely into a product of prime ideals: 2 (6) = p p2p3, 1 √ √ where p1 = (2, 1 + −5) = (2, 1 − −5) √ p2 = (3, 1 + −5) √ p3 = (3, 1 − −5). 71 4.5 Dedekind Domains in Extensions 4 Discrete Valuations and Dedekind Domains √ √ 2 In particular,√ (2) = p1, (3)√ = p2p3, (1 +√ −5) = p1p2 and√ (1 − −5) = p1p3. We have (2) ( p1 (Z[ −5] and [Z[ −5] : 2Z[ −5]] = 4, so [Z[ −5] : p1] = 2 and hence p1 is a maximal ideal. Similar arguments show that p2 and p3 are also√ maximal (thus prime) ideals. Moreover, notice that each of p1, p2, p3 is not principal, so h(Z[ −5]) > 1 (in fact, the class number is 2 in this case). 4.5 Dedekind Domains in Extensions Theorem 4.2.4 says that extending a Dedekind domain in a finite separable extension of its fraction field (i.e. taking its integral closure) again yields a Dedekind domain. The critical step in the proof of this result is to prove the integral closure is noetherian, which requires Proposition 2.2.5. The issue in generalizing Theorem 4.2.4 is that Proposition 2.2.5 does not hold if L/K is not separable – in fact, we will demonstrate in a moment several counterexamples to 2.2.5 in the inseparable case, even when A and B are DVRs. Recall the following definition: Definition. A field extension L/K is purely inseparable if for each α ∈ L, the minimal polynomial of α over K has a single distinct root. Lemma 4.5.1. For a finite extension L/K, the following are equivalent: (i) L/K is purely inseparable. (ii) If α ∈ L is separable over K then α ∈ K. n (iii) There is some n ∈ N such that for all α ∈ L, αp ∈ K. Lemma 4.5.2. Let L/K be any algebraic extension of fields. Then there is a unique subfield K ⊆ Ksep ⊆ L such that Ksep/K is separable and L/Ksep is purely inseparable. We now prove the general version of 4.2.4. Theorem 4.5.3. Let A be a Dedekind domain with field of fractions K, L/K a finite ex- tension and B the integral closure of A in L. Then B is a Dedekind domain. Proof. It remains to prove the theorem when L/K is inseparable. Assume char K = p > 0. By Lemma 4.5.2, there is a unique subextension K ⊂ L0 ⊂ L such that L0/K is separable and L/L0 is purely inseparable. Let B0 denote the integral closure of A in L0. Since L0/K is separable, Theorem 4.2.4 shows B0 is Dedekind. Thus it suffices to consider the case when K = L0 and A = B0. After this reduction, L/K is purely inseparable, so by Lemma 4.5.1, there is some n ∈ N so that for every α ∈ L, αpn ∈ K. Set q = pn. Let M be the splitting field of all polynomials of the form tq − α, with α ∈ K. Alternatively, M = {α1/q ∈ K | α ∈ K} where K is an algebraic closure of K. Let C be the integral closure of A in M. 72 4.5 Dedekind Domains in Extensions 4 Discrete Valuations and Dedekind Domains purely insep. K L M A B C Define the Frobenius map ϕ : M → K by ϕ(x) = xq = xpn . Notice that ϕ is well-defined by Lemma 2; injective since char K = p; and surjective by the construction of M. It is clear that ϕ is a homomorphism, so it is an isomorphism (of rings) M → K. Moreover, we claim A = ϕ(C). Indeed, if y ∈ A then any root of tq − y lies in C and maps to y via ϕ. Conversely, if x ∈ C then xq − α = 0 for some α ∈ K, i.e. xq ∈ K, but xq ∈ C as well, so ϕ(x) = xq ∈ C ∩ K = A. We have thus established that A ∼= C as rings, so C is in particular a Dedekind domain. By Theorem 4.3.6, B is a Dedekind domain if and only if every ideal of B is invertible. For an ideal I ⊂ B, let J = IC be the corresponding ideal of C. Then J is invertible Pm −1 since C is Dedekind, so i=1 aibi = 1 for some elements ai ∈ I, bi ∈ J . Since we are in Pm q q q q−1 q characteristic p, i=1 ai bi = 1 and bi ∈ K. For each 1 ≤ i ≤ m, set ci = ai bi so that Pm q−1 −q q −q q −q i=1 aici = 1 and ci ∈ I J ⊆ L. Now ciI ⊆ I J ⊆ J J = C but the previous statement shows that ciI ⊆ L. So ciI ⊆ C ∩ L = B, and thus ci ∈ (B : I). In summary, Pm we have i=1 aici = 1 for elements ai ∈ I and ci ∈ (B : I), so it follows that I(B : I) = B. Hence every ideal of B is invertible, so B is a Dedekind domain. In fact, Theorem 4.5.3 can be obtained as a corollary to the much more powerful Krull- Akizuki Theorem, which we don’t quite have the tools to prove yet. Theorem (Krull-Akizuki). Let A be a noetherian integral domain and suppose dim A = 1,K = Frac(A), L/K is a finite algebraic extension and B ⊂ L is any subring containing A. Then B is a noetherian domain with dim B ≤ 1. For the remainder of the section, we discuss counterexamples to Proposition 2.2.5 and discuss a generalization of this result for affine algebras. Example 4.5.4. (Hochster) Consider a perfect field k of characteristic p > 0 (e.g. k = Fp p if you like) and let K = k(tn | n ∈ N). Also, for each n ∈ N, set Kn = k(t1, . . . , tn) and S∞ p S∞ L = n=1 Kn = k(tn | n ∈ N) so that L = K. Define a ring A = n=1 Kn[[x]] where x is an indeterminate. It is a fact that A is a local noetherian domain with maximal ideal (x), and any ideal of A is of the form I = (xm) for some m ∈ N. In particular, Theorem 4.1.6 P∞ i shows that A is a DVR. Observe that K[[x]] ⊆ A ⊆ L[[x]]. Set u = i=1 tix ∈ L[[x]] so that up ∈ A. If F = Frac(A), the extension F (u)/F is purely inseparable, but the integral closure of A in F (u) is not a finitely generated A-module. Example 4.5.5. (Kaplansky) Let k be a field of characteristic 2 and set C = k[[t]]. Take some u ∈ C r k and let K = k(t, u2) and L = k(t, u). If u is chosen so that u and t are algebraically independent over k, for example, then we will have [L : K] = 2. Consider the rings A = C ∩ K,B = C ∩ L and the field of fractions M = Frac(C) = k((t)). 73 4.5 Dedekind Domains in Extensions 4 Discrete Valuations and Dedekind Domains k K L M A B C Since C is a DVR (Example 3 in Section 4.1), A and B are also DVRs with Frac(A) = K and Frac(B) = L. In particular, B is the integral closure of A in L – if B0 is this integral closure, then B ⊆ B0 because B is integral over A, but on the other hand B is integrally closed (using the fact that char k = 2), so it must be that B = B0. Suppose B were a finitely generated A-module. Then we would have B ⊆ Aα1 + ... + Aαr + Aβ1u + ... + Aβsu for m some αi, βj ∈ K. By clearing denominators, there is some m ∈ N such that t αi ∈ A and m m P∞ ` t βj ∈ A for all 1 ≤ i ≤ r, 1 ≤ j ≤ s. So t B ⊆ A[u]. Write u = `=0 a`t with a` ∈ k. Set ∞ m −m−1 −m−1 X ` v = (u − a0 − a1t − ... − amt )t = t a`t . `=m+1 By the first description, v ∈ L and by the second, v ∈ C, so v ∈ C ∩L = B ⊆ A[u]. However, the second expression for v shows the coefficient of u in the expansion of tmv must be t−1, contradicting the fact that v ∈ A[u]. Thus B cannot be finitely generated over A. To close, we generalize Proposition 2.2.5 to the context of affine algebras. Theorem 4.5.6 (Noether). Let k be a field and suppose A is a finitely generated k-algebra that is also an integral domain with field of fractions K. If L/K is any finite extension and B is the integral closure of A in L, then B is a finitely generated A-module. Proof. (Eisenbud) By Noether normalization (2.4.1), there exist y1, . . . , yr ∈ A such that A is integral over the polynomial ring k[y1, . . . , yr]. Then the result will follow if we prove B/k[y1, . . . , yr] is a finite extension of rings, so we may replace A with k[y1, . . . , yr], and K 00 00 with k(y1, . . . , yr), to begin with. Moreover, if L is the normal closure of L over K and B is the integral closure of A in L00, then demonstrating B00/A is finite also implies the result. So we make the reduction L = L00, so that L/K is a normal extension. Set G = Gal(L/K) and let L0 = LG be the subfield fixed by G. It is a general fact from Galois theory that L/L0 is Galois and L0/K is a purely inseparable extension. purely insep. Galois K L0 L A B0 B Since L/L0 is finite and separable, Proposition 2.2.5 gives us that B is a finitely generated B0-module, so it’s enough to prove B0 is a finitely generated A-module to obtain the result. If L0 = K, we’re done so assume L0 6= K. This a nontrivial purely inseparable extension, so char K = p > 0 and by Lemma 4.5.1, there is some n ∈ N with q = pn and αq ∈ K for every 74 4.5 Dedekind Domains in Extensions 4 Discrete Valuations and Dedekind Domains 0 0 0 α ∈ L . In particular, if L = K(z1, . . . , zn) for z1, . . . , zn ∈ L , then for each 1 ≤ i ≤ n, q q fi z ∈ K, so there exist fi, gi ∈ A such that z = . Since A = k[y1, . . . , yr], we have i i gi X j1 jr X j1 jr fi = cJ,iy1 ··· yr and gi = dJ,iy1 ··· yr r r J=(j1,...,jr)∈N0 J=(j1,...,jr)∈N0 for cJ,i, dJ,i ∈ k, finitely many of which are nonzero. Let S be the finite set of all nonzero r cJ,i and dJ,i, where J ranges over all tuples (j1, . . . , jr) ∈ N0 and 1 ≤ i ≤ n. Then the field 0 1/q 0 1/q 1/q k := k(s | s ∈ S) is a finite extension of k, and consequently Le := k (y1 , . . . , yr ) is a finite extension of L0. If Be is the integral closure of A in Le, then Be contains B0 so it’s enough to prove the result when L0 = Le and B0 = Be. 0 1/q 1/q 0 Now consider C := k [y1 , . . . , yr ] ⊆ L . Then C is integrally closed (e.g. it’s a UFD) 0 0 1/q 0 1/q 1/q 0 and Frac(C) = L , so B ⊆ C. But each yj is integral over A, so C = k [y1 , . . . , yr ] ⊆ B since B0 is the integral closure of A in L0. Hence B0 = C, which is finitely generated over A by definition. This finishes the proof. 75 5 Completion and Filtration 5 Completion and Filtration 5.1 Topological Abelian Groups and Completion Definition. A topological abelian group is a topological space G with an abelian group structure such that the addition and inversion maps α : G × G −→ G η : G −→ G (x, y) 7−→ x + y x 7−→ −x are continuous, where G × G is endowed with the product topology. Lemma 5.1.1. The set {0} is closed in G if and only if G is Hausdorff as a topological space. Proof. Since addition is continuous, {0} is closed in G if and only if the diagonal ∆ = {(x, x) | x ∈ G} is closed in G × G. It is well known that the latter condition is equivalent to G being Hausdorff. Lemma 5.1.2. For a fixed a ∈ G, the translation map ta : G → G, x 7→ x + a, is continuous with inverse t−a, and hence a homeomorphism. Lemma 5.1.3. For any subgroup N ≤ G of a topological abelian group G, N is a topological abelian group with the subspace topology and G/N is a topological abelian group with the quotient topology. If U = U(0) is a neighborhood of 0, then U + a = ta(U) is also open, and thus a neighborhood of a. Since ta is a homeomorphism, every neighborhood of any point of G is of the form U + a for some U = U(0). Thus the topology of G is uniquely determined by the collection {U(0)} of neighborhoods of the identity. T Lemma 5.1.4. Let H = U(0)30 U(0) be the intersection of all neighborhoods of 0. Then (i) H is a subgroup of G. (ii) H = {0}. (iii) G/H is Hausdorff. (iv) G is Hausdorff if and only if H = {0}. Proof. (i) Take x, y ∈ H. Then x, y ∈ U(0) for every neighborhood U(0) of the identity. In particular, 0 ∈ −x + U(0) and this is an open set since t−x is a homeomorphism. Hence H ⊆ −x + U(0) so y ∈ −x + U(0). Thus x + y ∈ U(0). Since U(0) was arbitrary, we have that x + y ∈ H. This argument also shows that −x ∈ U(0) for any neighborhood U(0), so −x ∈ H. Hence H is a subgroup. (ii) For any x ∈ G, x ∈ H ⇐⇒ −x ∈ H ⇐⇒ −x ∈ U(0) for all neighborhoods U(0) ⇐⇒ 0 ∈ x + U(0) for all U(0) ⇐⇒ x is a limit point of {0}. Therefore H = {0}. (iii) For a fixed a ∈ G, a + H = ta(H) = ta({0}) is closed in G/H since ta is continuous. Thus cosets, aka the ‘points’ in G/H, are closed, so G/H is Hausdorff. (iv) If H = {0} then (iii) implies G is Hausdorff. On the other hand, for all x 6= 0, there is a neighborhood Ux of x such that 0 6∈ Ux. Thus x 6∈ H, which implies H = {0}. 76 5.1 Topological Abelian Groups and Completion 5 Completion and Filtration Assume 0 ∈ G has a countable basis of subgroups U(0) ⊆ G. ∞ Definition. A Cauchy sequence in G is a sequence of elements (xn)n=0, where xn ∈ G for all n ∈ N0, such that for all neighborhoods U(0), there is some N = N(U(0)) ∈ N such that xn − xm ∈ U(0) for all n, m ≥ N. Definition. We say two Cauchy sequences (xn) and (yn) are equivalent, written (xn) ∼ (yn), if for any neighborhood U(0), there is some N = N(U(0)) ∈ N such that xn −yn ∈ U(0) for all n ≥ N. Lemma 5.1.5. ∼ is an equivalence relation on the set of Cauchy sequences in G. Proof. Let (xn), (yn) and (zn) be Cauchy sequences in G. Then for any U(0), xn − xn = 0 ∈ U(0) for every n ∈ N, so (xn) ∼ (xn). If (xn) ∼ (yn) then there is an N ∈ N such that for all n ≥ N, xn − yn ∈ U(0). Since U(0) is assumed to be a subgroup of G, yn − xn = −(xn − yn) ∈ U(0) as well, so the same N shows that (yn) ∼ (xn). Finally, if (xn) ∼ (yn) and (yn ∼ zn) and U(0) is a neighborhood of 0, there are N1,N2 ∈ N such that for all n ≥ N1, xn − yn ∈ U(0) and for all n ≥ N2, yn − zn ∈ U(0). Thus for all n ≥ max{N1,N2}, we have xn − zn = xn − yn + yn − zn ∈ U(0) since U(0) is a subgroup. Therefore (xn) ∼ (zn), so ∼ is an equivalence relation. Definition. The completion of a topological group G is the quotient space Gb of all equiva- lence classes under ∼ of Cauchy sequences in G. For a Cauchy sequence (xn), we let [(xn)] denote its equivalence class in Gb. Proposition 5.1.6. For any topological group G, Gb is a topological group via pointwise addition and inversion: [(xn)] + [(yn)] = [(xn + yn)] and −[(xn)] = [(−xn)]. Proof. Consider Cauchy sequences (xn) and (yn) in G. We first prove pointwise addition 0 and inversion of Cauchy sequences are invariant on equivalences classes. If (xn) ∼ (xn) and 0 0 (yn) ∼ (yn) then for all U(0), there are numbers N1,N2 ∈ N such that xn − xn ∈ U(0) for 0 all n ≥ N1 and yn − yn ∈ U(0) for all n ≥ N2. Taking N = max{N1,N2}, we have that for 0 0 0 0 all n ≥ N,(xn + yn) − (xn + yn) = (xn − xn) + (yn − yn) ∈ U(0) since U(0) is a subgroup. 0 0 0 Thus (xn + yn) ∼ (xn + yn), so pointwise addition is well-defined. Similarly, (−xn) ∼ (−xn) so inversion is well-defined. To prove that Gb is a group under these actions, consider two Cauchy sequences (xn) and (yn) in G. Then for a neighborhood U(0), there are N1,N2 ∈ N such that xm − xn ∈ U(0) for all m, n ≥ N1 and ym − yn ∈ U(0) for all m, n ≥ N2. Then for all n ≥ N = max{N1,N2}, (xm + ym) − (xn − yn) = (xm − xn) + (ym − yn) ∈ U(0) since U(0) is a subgroup. Hence (xn + yn) is a Cauchy sequence. The zero sequence (0) represents the identity class in Gb and it’s easy to check that [(xn)] + [(−xn)] = [(0)] for all Cauchy sequences (xn), so we have that Gb is an abelian group. Finally, for each neighborhood U(0) in G, define a subgroup Ub(0) in Gb by Ub(0) = {[(xn)] : there exists an N ∈ N such that xn ∈ U(0) for all n ≥ N}. Then {Ub(0) : U(0) is a neighborhood of 0 in G} forms a basis for a topology on Gb which is compatible with the group structure on Gb. Hence Gb is a topological group. 77 5.2 Inverse Limits 5 Completion and Filtration Proposition 5.1.7. There is a homomorphism ϕ : G → Gb which is injective precisely when G is Hausdorff. T Proof. Define ϕ(x) = [(x, x, x, . . .)] = [(x)]. Then ker ϕ = H where H = U(0) U(0) so ϕ is injective if and only if H = 0, which by Lemma 5.1.4 is equivalent to G being Hausdorff. Definition. A homomorphism of topological abelian groups is a continuous homo- morphism f : G → H. Such a map f is an isomorphism of topological abelian groups if f is a homeomorphism and an isomorphism, that is, f is a continuous homomorphism with a continuous inverse f −1 : H → G which is also a homomorphism. Lemma 5.1.8. Given a homomorphism of topological abelian groups f : G → H, if (xn) is a Cauchy sequence in G then (f(xn)) is a Cauchy sequence in H. Proof. Suppose V (0) is a neighborhood of 0 ∈ H. Since f is continuous, U(0) := f −1(V (0)) is open in G, and since f is a homomorphism, f(0) = 0 and hence U(0) is a neighborhood of 0 ∈ G. Thus there is some N ∈ N such that for all m, n ≥ N, xm − xn ∈ U(0). Applying f and using the fact that it is a homomorphism, we get f(xm) − f(xn) = f(xm − fn) ∈ f(U(0)) = V (0) for all m, n ≥ N. Therefore (f(xn)) is Cauchy. Corollary 5.1.9. Every homomorphism of topological abelian groups f : G → H induces a ˆ ˆ homomorphism f : Gb → Hb such that f[(xn)] = [f(xn)]. In particular, this says that the completion Gb a solution to a universal mapping property for G. In addition, the completion operation (c·) is a covariant functor on the category TopAbGps of topological abelian groups: f g Proposition 5.1.10. For every sequence G −→ H −→ K of homomorphisms of topological abelian groups, we have (\g ◦ f) =g ˆ ◦ fˆ. 5.2 Inverse Limits In this section we formulate an alternative construction of the completion Gb from Section 5.1 using inverse limits. Definition. An inverse system is a family of groups (Am)m∈I , for some directed index set I, together with homomorphisms θm` : Am → A` for every m ≥ ` such that when m ≥ ` ≥ k, the following diagram commutes: θm` Am A` θmk θ`k Ak 78 5.2 Inverse Limits 5 Completion and Filtration Definition. Given an inverse system (Am, θm`) over a directed set I, we define a coherent Q sequence to be a sequence (ym) ∈ m∈I Am such that θm`(ym) = y` for all m ≥ `. The set of all coherent sequences of (Am, θm`) is called the inverse limit of the inverse system: ( ) Y lim Am = (ym) ∈ Am : θm`(ym) = y` for all m ≥ ` . ←− m∈I Now let G be a topological group possessing a countable basis of open subgroups Gn = Gn(0) such that G = G0 ⊇ G1 ⊇ G2 ⊇ · · · Lemma 5.2.1. Each Gn is both open and closed. Proof. For all a ∈ G, a + Gn = ta(Gn) is open since translation is a homeomorphism. Since the a + G partition the complement of G in G, we have G G = S (a + G ) which n n r n a6∈Gn n is a union of open sets and therefore open. Hence Gn is closed. For each n ≥ 1, Gn ⊆ Gn−1 so we have projections θn : G/Gn → G/Gn−1. Lemma 5.2.2. The system (G/Gn, θn)n∈N is an inverse system. Lemma 5.2.3. lim G/Gn is a topological group. ←− Q Proof. Consider G/Gn with the product topology. Then lim G/Gn has the subspace n∈N ←− topology from Q G/G . Indeed, if p : Q G/G → G/G is the mth projection map, n≥0 n m n∈N n m −1 then the sets S(x, m) := pm (x + Gm) = {(yn) | ym = x + Gm} over all m form a subbasis Q for the topology on G/Gn, so the sets Se(x, m) := S(x, m) ∩ lim G/Gn form a subbasis n∈N ←− for the (subspace) topology on lim G/Gn. In fact, Se(0, m) form a basis of this topology ←− because the collection of Se(0, m) are closed under intersections to begin with. The axioms of a topological abelian group are then easily checked for lim G/Gn. ←− When (G/Gn, θn)n∈N is the inverse system of quotients in G, the inverse limit coincides with the completion of G from Section 5.1. Proposition 5.2.4. For a topological abelian group G with inverse system (G/Gn, θn)n∈N, ∼ there is a canonical isomorphism of topological abelian groups Gb = lim G/Gn. ←− Proof. For each n ∈ N, define a map ζn : Gb −→ G/Gn [(xk)] 7−→ xkn + Gn where kn is the smallest N (or any N) such that xk + Gn = xN + Gn for all k ≥ N. Such a choice is possible because (xk) is Cauchy. Clearly the ζn are homomorphisms, and one can verify that they are well-defined on equivalence classes of Cauchy sequences. Now define a map into the inverse limit by ζ : Gb −→ lim G/Gn ←− [(xk)] 7−→ (ζn([(xk)]))n∈N. 79 5.2 Inverse Limits 5 Completion and Filtration The image of ζ consists of coherent sequences: for all n ∈ N, we have θn+1(xkn+1 + Gn+1) = xkn+1 + Gn = xkn + Gn since kn+1 ≥ kn, and therefore θn+1(ζn+1[(xk)]) = ζn[(xk)]. Now given (yn) ∈ lim G/Gn, ←− construct a Cauchy sequence mapping to (yn) by taking a coset representative xn for each n ∈ N such that yn = xn + Gn. Since (yn) is a coherent sequence, for each n ∈ N we have θn+1(yn+1) = yn =⇒ θn+1(xn+1 + Gn+1) = xn+1 + Gn = xn + Gn. So xn+1 − xn ∈ Gn and hence (xn) is a Cauchy sequence. Clearly ζ([(xn)]) = (yn) so ζ is surjective. To see that ζ is continuous, suppose [(xk)] ∈ Gb. Then using the notation from the proof of Lemma 5.2.3, for each m ∈ N, −1 [(xk)] ∈ ζ (Se(0, m)) ⇐⇒ xkm + Gm = 0 + Gm in G/Gm ⇐⇒ [(xk)] ∈ Gbm. Since Gbm is open for all m, we have that ζ is continuous. Finally, to show ζ is an isomorphism, we define an inverse function by η : lim G/Gn −→ Gb ←− xn + Gn 7−→ [(xn)]. We showed above that this map is well-defined and a homomorphism of topological abelian groups, and clearly ζ ◦ η = id by surjectivity of ζ. Finally, for any [(xk)] ∈ Gb, we lim G/Gn ←− have η ◦ ζ([(xk)]) = η(xkn + Gn) = [(xkn )] = [(xk)] since subsequences of Cauchy sequences are equivalent to their parent Cauchy sequence. This proves that ζ is an isomorphism of topological abelian groups. Definition. When (Gn)n∈N is the system of all subgroups of finite index in G, the inverse limit Gb = lim G/Gn is called the profinite completion of G. ←− Definition. A surjective inverse system is one in which every map θm` is surjective. Definition. Let (Am, αm`) and (Bm, βm`) be two inverse systems indexed over the same set I.A morphism of inverse systems is a set of maps (ϕ•) such that ϕm : Am → Bm is a homomorphism for each m ∈ I that is compatible with the αm` and βm`, i.e. the following diagrams commute for all m ≥ ` ∈ I: ϕm Am Bm αm` βm` A` B` ϕ` 80 5.2 Inverse Limits 5 Completion and Filtration Definition. Let (Am, αm`), (Bm, βm`) and (Cm, γm`) be inverse systems with the same index set I.A short exact sequence of inverse systems is a sequence of maps of inverse (im) (pm) im pm systems 0 → (Am) −−→ (Bm) −−→ (Cm) → 0 such that 0 → Am −→ Bm −→ Cm → 0 is an exact sequence for all m ∈ I. Proposition 5.2.5. For any short exact sequence of inverse systems 0 → (Am) → (Bm) → (Cm) → 0, there is an exact sequence of topological abelian groups 0 → lim Am → lim Bm → lim Cm. ←− ←− ←− In other words, inverse limit is a left exact functor from the category of inverse systems to the category of topological abelian groups. Furthermore, if (Am) is a surjective inverse system then 0 may be added on the right, i.e. lim(−) is an exact functor. ←− Q A Proof. Let A = Am and define a map d : A → A by (am) 7→ (am −θm+1,m(am+1)). Define B C A B B, d ,C and d similarly for (Bm) and (Cm). Then ker d = lim Am, ker d = lim Bm and ←− ←− C kerd = lim Cm. We have a commutative diagram with exact rows: ←− 0 A B C 0 dA dB dC 0 A B C 0 so by the Snake Lemma, we get an exact sequence 0 → ker dA → ker dB → ker dC → coker dA → coker dB → coker dC → 0. The first three terms of this are 0 → lim Am → lim Bm → lim Cm so we have our exact ←− ←− ←− sequence. For the final statement, suppose (Am, αm`) is a surjective system. We need to prove A A coker d = 0, i.e. for each (am) ∈ A, we want to find (xm) ∈ A such that d (xm) = (am). Let x0 = 0. Then a0 = x0 − θ1(x1) so θ1(x1) = −a0 and since θ1 is surjective, this allows us to choose some x1 ∈ A1 so that this holds. Proceed to inductively define (xm). This shows dA is surjective, so we are done. p Corollary 5.2.6. Let 0 → G0 → G −→ G00 → 0 be a short exact sequence of topological 0 00 0 0 abelian groups having systems of open subgroups {Gn}, {Gn} and {Gn}, where Gn = G ∩ Gn 00 and Gn = p(Gn) for each n ∈ N. Then the sequence of completions 0 → Gb0 → Gb → Gb00 → 0 is exact. Proof. This is done by considering the short exact sequence of surjective inverse systems 0 0 00 00 0 → (G /Gn) → (G/Gn) → (G /Gn) → 0 and applying Proposition 5.2.5. Corollary 5.2.7. For a topological abelian group G with system of open subgroups Gn and ∼ completion Gb, for each n ∈ N, Gbn is a topological subgroup of Gb and G/b Gbn = G/Gn. 81 5.3 Topological Rings and Module Filtrations 5 Completion and Filtration Proof. Apply Corollary 5.4.9 to the sequence 0 → Gn → G → G/Gn → 0. Notice that ∼ ∼ G/Gn = G/G\n = G/b Gbn because each has the discrete topology. Corollary 5.2.8. For any topological abelian group G, Gb ∼= Gb. Definition. A topological abelian group G is complete if the canonical map ϕ : G → G,b x 7→ [(x)], is an isomorphism. In this language, Corollary 5.2.8 says that completions are complete. T∞ Remark. Notice that if G is complete, ϕ : G → Gb is an isomorphism, so n=1 Gn = ker ϕ = 0 and hence by Lemma 5.1.4, G is Hausdorff. 5.3 Topological Rings and Module Filtrations One of the most important concepts in commutative algebra is the completion of a ring with respect to a certain topology induced by one of its ideals. We first define the notion of a topological ring. Definition. A topological ring is a ring A which is a topological abelian group such that the ring multiplication A × A → A, (a, b) 7→ ab, is a continuous map. Topological fields are defined analagously. Given an ideal J ⊂ A, one can always give A the structure of a topological ring: Definition. Let A be a ring and J ⊂ A an ideal. Then the chain A = J 0 ⊇ J 1 ⊇ J 2 ⊇ J 3 ⊇ · · · forms a basis of additive subgroups of (A, +) as a topological abelian group. This topology, called the J-adic topology, makes A into a topological ring. Lemma 5.3.1. The completion Ab of a topological ring A is a topological ring. Examples. 1 Let A = k[t] be the polynomial ring in a single variable and take J = (t). Then Ab = k[[t]], the power series ring. 2 When A = Z and J = (p) for a prime integer p, the completion Zb is called the p-adic n completion of , denoted p. Alternatively, p = lim /p . Z Z Z ←− Z Z Definition. Let A be a topological ring. An A-module M is called a topological A-module if M is a topological abeliain group and the A-action map A × M → M, (a, m) 7→ am, is a continuous map. 82 5.3 Topological Rings and Module Filtrations 5 Completion and Filtration If J ⊂ A is an ideal of A, so that A is a topological ring with the J-adic topology, n for an A-module M one can define open sets Mn = J M that make M into a topological A-module. Then the topological completion Mc of M with respect to this topology is a topological Ab-module. Rather than using J nM to define the J-adic topology on M, we instead use systems of additive subgroups of M called filtrations. Definition. A filtration of an A-module M is an infinite chain of submodules M = M0 ⊇ M1 ⊇ M2 ⊇ · · · If J ⊂ A is an ideal, a J-filtration of M is a filtration such that JMn ⊆ Mn+1 for all n ≥ 0. If in addition JMn = Mn+1 for all n larger than some n0 ∈ N, then the filtration is called a stable J-filtration. 0 Lemma 5.3.2. If (Mn) and (Mn) are stable J-filtrations of M, then there exists an n0 ∈ N 0 0 such that Mn+n0 ⊆ Mn and Mn+n0 ⊆ Mn for all n ≥ 0. 0 0 Proof. Let n0 ∈ N such that JMn = Mn+1 and JMn = Mn+1 for all n ≥ n0. Then 0 0 n n 0 0 JMn0 = Mn0+1 and JMn0 = Mn0+1. By induction, J Mn0 = Mn0+n and J Mn0 = Mn0+n. Then we have J nM ⊆ J n−1M ⊆ · · · ⊆ JM ⊆ M n 0 n−1 0 0 0 and J M ⊆ J M ⊆ · · · ⊆ JM ⊆ M for all n ≥ n0. n n 0 0 From this we obtain Mn+n0 = J Mn0 ⊆ J M ⊆ Mn. Reversing the roles of the Mn and Mn gives the other containment. Remark. The completion Ab along an ideal J can also be described in terms of a metric, as T∞ n long as n=1 J = 0: if x, y ∈ A, define 1 d(x, y) = where v(x, y) = sup{k ≥ 1 | x − y ∈ J k}. 2v(x,y) Formally, we set d(x, y) = 0 if v(x, y) = ∞. Completion of the induced metric topology on A, that is, by formally adding in limits of Cauchy sequences in this topology, yields a topological ring isomorphic to Ab. Moreover, this description makes it easier to see that the ring operations + and · on A are continuous with respect to the metric topology and extend to continuous operations on Ab, and that the canonical embedding A,→ Ab has dense image. T∞ n Proposition 5.3.3. Suppose A is a ring and J ⊂ A is an ideal such that n=1 J = 0. Let Ab be the completion of A along J. Then JAb ⊆ J(Ab), the Jacobson radical of Ab. Proof. By Lemma 1.1.1 it suffices to show 1 − x is a unit in Ab for all x ∈ J. Consider the formal expansion 1 = 1 + x + x2 + ... 1 − x n n+1 n+1 Set sn = 1 + x + ... + x ∈ A. Then for each n ≥ 1, sn − sn+1 = −x ∈ J so (sn) is a 1 Cauchy sequence with respect to the J-adic metric and hence its limit 1−x exists in Ab. 83 5.4 Graded Rings and Modules 5 Completion and Filtration T∞ n n ∼ n Lemma 5.3.4. If J ⊂ A is an ideal with n=1 J = 0, then A/Jb Ab = A/J for all n ≥ 1. In particular, if m ⊂ A is a maximal ideal, then Ab is a local ring with maximal ideal mb = mAb. Example 5.3.5. The completion of a polynomial ring A = k[t1, . . . , tn] along the maximal ideal m = (t1, . . . , tn) is a power series ring Ab = k[[t1, . . . , tn]], which is local with maximal ideal (t1, . . . , tn). More generally: T∞ n Proposition 5.3.6. If J = (a1, . . . , an) is a finitely generated ideal of A with n=1 J = 0, then Ab = A[[t1, . . . , tn]]/(t1 − a1, . . . , tn − an). Corollary 5.3.7. If A is noetherian then any completion Ab is also noetherian. Example 5.3.8. Be aware that localization and completion do not commute. For example, consider the ideal J = (x) in A = k[x, y]. Then A(x) = (k(y)[x])(x) and the completion of this ring along (x) is k(y)[[x]]. Meanwhile, the completion Ab of A along (x) is k[[x]][y], whose localization at (x)Ab is the proper subring (k[[x]][y])(x) ( k(y)[[x]]. 5.4 Graded Rings and Modules L∞ Definition. A ring A is said to be graded if as an abelian group, A = n=0 An for sub- groups An ≤ A with the property that AmAn ≤ Am+n for all m, n ∈ N0. Notice that A0 is always a subring of A, so each An is an A0-module. We also define L∞ A+ = n=1 An, which is an ideal of A. Definition. If A is a graded ring, an abelian group M is a graded A-module if M is an L∞ A-module and M = n=0 Mn for subgroups Mn ≤ M such that AmMn ⊆ Mm+n for all m, n ∈ N0. As above, each Mn is not just an A-module but also an A0-module. L∞ Definition. Given a graded A-module M = n=0 Mn, an element x ∈ Mn is said to be homogeneous of degree n in M. Definition. Let A and B be graded rings. A homomorphism of graded rings is a ring map f : A → B such that f(An) ⊆ Bn for all n ∈ N0. Likewise, if M and N are two graded A-modules, a homomorphism of graded modules is an A-module map f : M → N such that f(Mn) ⊆ Nn for all n ∈ N0. Example 5.4.1. If A = k[t1, . . . , tn] then A is a graded ring with respect to total degree: ∞ M X i1 ir A = An where A0 = k and An = kt1 ··· tr r n=0 (i1,...,ir)∈N0 i1+...+ir=n This is called the standard grading on the polynomial ring A. The ideal A+ here is (t1, . . . , tn). 84 5.4 Graded Rings and Modules 5 Completion and Filtration Lemma 5.4.2. For a graded ring A, the following are equivalent: (i) A is noetherian. (ii) A0 is noetherian and A/A0 is a finitely generated ring extension. Proof. (ii) =⇒ (i) follows directly from Corollary 3.3.3. ∼ (i) =⇒ (ii) If A is noetherian, then A0 = A/A+ is noetherian and A+ is finitely generated. Let A+ = (x1, . . . , xs) for xi ∈ A. We may assume each xi ∈ Aki is homogeneous of degree ki. We will show An ⊆ A0[x1, . . . , xs] for all n ∈ N. The n = 0 case is trivial so assume Pn n ≥ 1. Then An ⊂ A+ so for any x ∈ An there are elements ci ∈ A such that x = i=1 cixi. Since x is homogeneous of degree n, we may assume ci ∈ An−ki where by convention we treat A` = 0 when ` < 0. Since all ki > 0, then ci ∈ A0[x1, . . . , xs] by induction, and therefore x ∈ A0[x1, . . . , xs]. It follows that A ⊆ A0[x1, . . . , xs], so A = A0[x1, . . . , xs] and A/A0 is a finitely generated extension of rings. Remark. If A is a ring and I ⊂ A is an ideal, there are two canonical graded rings we can associate to I: ∞ ∞ ∗ M n M n n+1 A := I and GI (A) := I /I . n=0 n=0 0 0 ∗ ∗ By convention, we let I = A. Notice that I = A,→ A , so that A0 = A. If A is noetherian, then I is finitely generated as a submodule, that is, I = (x1, . . . , xr) for some xi ∈ I. Then powers of I are generated by homogeneous products of the generators of I: r ! n Y ei I = xi : ei ∈ N0, e1 + ... + er = n . i=1 ∗ ∗ In particular, this shows that A = A[x1, . . . , xs], so A is a noetherian ring by Corol- lary 3.3.3(b). Suppose I is an ideal of a ring A. Let M be an A-module with I-filtration (Mn)n∈N0 ; recall that this means there is a chain of submodules M = M0 ⊇ M1 ⊇ M2 ⊇ · · · n such that IMn ⊆ Mn+1 for all n ∈ N0. The standard I-filtration on M is when Mn = I M, which defines the I-topology (or I-adic topology) on M. For any filtration (Mn)n∈N0 , we define a graded module from M by ∞ ∗ M M = Mn. n=0 Lemma 5.4.3. M ∗ is a graded A∗-module. Lemma 5.4.4. Let A be a noetherian ring, I ⊂ A and ideal and M a finitely generated A-module. Then for an I-filtration (Mn)n∈N0 of M, the following are equivalent: 85 5.4 Graded Rings and Modules 5 Completion and Filtration ∗ L∞ ∗ (i) M = n=0 Mn is a finitely generated A -module. (ii) (Mn)n∈N0 is a stable I-filtration. Proof. Note that by Corollary 3.1.5, the assumptions that A is noetherian and M is finitely generated imply that M is a noetherian A-module. Thus each submodule Mn ⊆ M is finitely generated, so the modules n M Qn := Mi n=0 are finitely generated for each n ∈ N, by Corollary 3.1.4. For each n, Qn generates the following A∗-submodule of M ∗: ∞ ∗ M m Qn := Qn ⊕ I Mn. m=1 ∗ ∗ m (That Qn is an A -submodule follows from the fact that I Mn ⊆ Mm+n for all m ∈ N0.) ∗ ∗ Moreover, Qn is a finitely generated A -module generated by the A-module generators of ∗ ∗ m m−1 Qn. By construction, Qn ⊆ Qn+1 for all n, since I Mn ⊆ I Mn+1 for all m ≥ 1. Also, S∞ ∗ ∗ n=0 Qn = M . With these observations, we now proceed to the main proof. (i) =⇒ (ii) If M ∗ is a finitely generated A∗-module then it follows (3.1.5) that M ∗ is noetherian as an A∗-module, i.e. M ∗ satisfies the ACC on A∗-submodules. In particular, ∗ ∗ ∗ ∗ the ascending chain Qn ⊆ Qn+1 stabilizes, so there is some n0 ∈ N0 such that Qn = Qn0 for ∗ S∞ ∗ ∗ ∗ ∗ all n ≥ n0. By the above, M = n=0 Qn = Qn0 so Qn0 is a finitely generated A -module. Then ∞ n0 ∞ M M M m m Mn = Mn ⊕ I Mn0 =⇒ Mn0+m = I Mn0 for all m ∈ N0. n=0 n=0 m=n0 This is equivalent by Lemma 5.3.2 to (Mn) ∈n∈N0 being a stable I-filtration. m (ii) =⇒ (i) If (Mn) is a stable filtration, there is some n0 ∈ N0 such that Mn0+m = I Mn0 ∗ ∗ ∗ for all m ∈ N0. This implies M = Qn0 so by the preliminary remarks, M is a finitely generated A∗-module. Proposition 5.4.5. Assume A is a noetherian ring, I ⊂ A is an ideal, M is a finitely 0 generated A-module with stable I-filtration (Mn)n∈N0 and M ⊆ M is a submodule. Then the 0 0 0 filtration Mn = M ∩ Mn induced on M is a stable I-filtration. 0 Proof. First, (Mn) is an I filtration, since for each n, 0 0 0 0 0 IMn = I(M ∩ Mn) ⊆ IM ∩ IMn ⊆ M ∩ Mn+1 ⊆ M ∩ Mn. ∗ L∞ ∗ If (Mn) is stable then by Lemma 5.4.4, M = n=0 Mn is a finitely generated A -module, and thus by Lemma 5.4.2 a noetherian A∗-module. Then (M 0)∗ exists and is an A∗-submodule of M ∗. Since M ∗ is noetherian, (M 0)∗ is finitely generated as an A∗-submodule. Finally, 0 Lemma 5.4.4 implies that (Mn) is a stable I-filtration. We now obtain the useful Artin-Rees Lemma, which has a number of important conse- quences that follow. 86 5.4 Graded Rings and Modules 5 Completion and Filtration Corollary 5.4.6 (Artin-Rees Lemma). Assume A is a noetherian ring, I ⊂ A is an ideal, M 0 is a finitely generated A-module and M ⊆ M is a submodule. Then there exists an n0 ∈ N0 0 n+n0 n 0 n0 such that M ∩ I M = I (M ∩ I M) for all n ∈ N0. n Proof. Apply Proposition 5.4.5 to the stable I-filtration Mn = I M of M. Corollary 5.4.7. Assume A is a noetherian ring, I ⊂ A is an ideal, M is a finitely generated A-module and M 0 ⊆ M is a submodule. Then the subspace topology induced on M 0 by the I-topology on M is precisely the I-topology on M 0. Corollary 5.4.8 (Krull’s Intersection Theorem). Suppose A is a noetherian ring. Then (a) If I ⊂ A is an ideal such that I ⊆ J(A) and M is a finitely generated A-module, then T∞ n n=0 I M = 0. In particular, the I-topology on M is Hausdorff. T∞ n (b) If A is an integral domain and I is a proper ideal of A, then n=0 I = 0. 0 T∞ n 0 Proof. (a) Set M = n=0 I M so that M ⊆ M is a finitely generated A-submodule of M (A is noetherian and M is finitely generated). By the Artin-Rees lemma, there exists an n0 ∈ N0 0 n+n0 n 0 n0 0 n+n0 0 so that for all n ∈ N0, M ∩ I M = I (M ∩ I M). But notice that M ∩ I M = M and M 0 ∩ In0 M = M 0, so we have M 0 = InM 0. Since I ⊆ J(A), Nakayama’s lemma (1.2.1) implies M 0 = 0. T∞ n (b) Now suppose A is an integral domain. Set J = n=0 I and suppose x ∈ J. Then k k x ∈ I for all k ∈ N0, so in particular (x) ⊆ I for every k. By the Artin-Rees lemma, there n+n0 n n0 k is an n0 ∈ N0 such that for all n ∈ N0,(x) ∩ I = I ((x) ∩ I ). Since (x) ⊆ I for all k, this becomes (x) = In(x) for all n, even (x) = I(x). Thus there is some a ∈ I such that x = ax, which can be written x(1 − a) = 0. If a were 1, I would be the unit ideal, but this contradicts the assumption that I is proper. So a 6= 1, and since A is a domain, it follows that x = 0. Corollary 5.4.9. Suppose A is a noetherian ring and 0 → M 0 → M → M 00 → 0 is a short exact sequence of A-modules. For a fixed ideal I ⊂ A, consider the I-topologies on M 0,M,M 00 and the corresponding completions Mc0, M,c Mc00. Then the sequence 0 → Mc0 → Mc → Mc00 → 0 is also exact. Proof. This follows from Corollary 5.4.7 and Lemma 5.3.2. Now recall the other canonical graded ring associated to an ideal I ⊂ A, namely ∞ M n n+1 GI (A) = I /I . n=0 The multiplication law is given by Im/Im+1 × In/In+1 −→ Im+n/Im+n+1 (x + Im+1, y + In+1) 7−→ xy + Im+n+1. 87 5.4 Graded Rings and Modules 5 Completion and Filtration This is well-defined since In ⊇ In+1 is a chain. If M is an A-module with I-filtration (Mn)n∈N0 , we can define a GI (A)-module by ∞ M GI (M) = Mn/Mn+1. n=0 Then GI (M) is a graded GI (A)-module via the multiplication law m m+1 I /I × Mn/Mn+1 −→ Mm+n/Mm+n+1 m+1 (x + I , z + Mn+1) 7−→ xz + Mm+n+1 m which is well-defined since xz ∈ I Mn ⊆ Mm+n. Lemma 5.4.10. If A is a noetherian ring, then for every ideal I ⊂ A, (a) GI (A) is noetherian. (b) For any finitely generated A-module M with stable I-filtration (Mn)n∈N0 , GI (M) is finitely generated as a GI (A)-module. Proof. (a) Since A is noetherian, I is finitely generated, say by x1, . . . , xr ∈ I. Setx ¯i = 2 2 n n+1 xi + I ∈ I/I . Then as before, it can be checked that each I /I is generated by products of powers of thex ¯i. By this, GI (A) = A/I[¯x1,..., x¯r] which is noetherian by Corollary 3.3.3(b). n (b) If (Mn) is a stable I-filtration of M, there is some n0 ∈ N0 such that Mn0+n = I Mn0 n n+1 for all n0 ∈ N0. This implies Mn0+n/Mn0+n+1 = (I /I )Mn0 /Mn0+1 for all n ∈ N0, and thus n M0 GI = G(A) Mi/Mi+1. i=0 Now each Mi/Mi+1 is a finitely generated A-module and even an A/I-module, and A/I ⊆ GI (A). It follows that GI (M) is a finitely generated GI (A)-module. 88 6 Dimension Theory 6 Dimension Theory 6.1 Hilbert Functions L∞ Let A = n=0 An be a graded noetherian ring. Recall from Lemma 5.4.2 that A0 is noethe- L∞ rian and A = A0[x1, . . . , xs] for some xi ∈ Aki , ki ≥ 1, 1 ≤ i ≤ s. Also let M = n=0 Mn be a finitely generated graded A-module, generated by homogeneous elements m1, . . . , mt such that mj ∈ Mrj for each 1 ≤ j ≤ t. Lemma 6.1.1. Each Mn is a finitely generated A0-module. Proof. It is clear that each Mn is an A0-module by the definition of a graded A-module. Pt Take z ∈ Mn. Then there exist aj ∈ A for 1 ≤ j ≤ t such that z = j=1 ajmj. Further, we can choose polynomials fj ∈ A0[t1, . . . , ts] such that aj = fj(x1, . . . , xs) for each 1 ≤ j ≤ t. e1 es Note that monomials in the xi are homogeneous: x1 ··· xs ∈ Ak1e1+...+kses . Since z ∈ Mn, we can write t X z = gj(x1, . . . , xs)mj j=1 e1 es where for each j, gj(x1, . . . , xs) is an A0-linear combination of monomials x1 ··· xs such that k1e1 + ... + kses = n − rj. This shows Mn is finitely generated as an A0-module: the e1 es explicit generators are all elements of the form x1 ··· xs mj for 1 ≤ j ≤ t and ei ≥ 0 such that k1e1 + ... + kses = n − rj. Let C be the class of finitely generated A0-modules. Definition. A function λ : C → Z is said to be additive if for any short exact sequence 0 → M 0 → M → M 00 → 0 in C, we have λ(M) = λ(M 0) + λ(M 00). Examples. 1 If A0 is noetherian, the length function `(M) is additive by Lemma 3.2.3. 2 In the special case that A0 = k is a field, C is the class of all finite dimensional k-vector spaces. Here, the dimension function λ(M) = dimk M is an additive function. f1 f2 fn fn+1 Lemma 6.1.2. Suppose 0 → M0 −→ M1 −→· · · −→ Mn −−→ 0 is an exact sequence in C. Then for any additive function λ : C → Z, n X i (−1) λ(Mi) = 0. i=0 Proof. For each 1 ≤ i ≤ n + 1, set Ni = im fi which also equals ker fi+1 by exactness. Then we have short exact sequences fi+1 0 → Ni → Mi −−→ Ni+1 → 0 89 6.1 Hilbert Functions 6 Dimension Theory for every 1 ≤ i ≤ n + 1. By additivity, λ(Ni) − λ(Mi) + λ(Ni+1) = 0 for all i, so we get a telescoping sum: n n X i X i (−1) λ(Mi) = (−1) (λ(Ni) + λ(Ni+1)) = λ(N0) ± λ(Nn+1). i=0 i=0 Finally, additivity implies λ(0) = 0 so we have λ(N0) = λ(Nn+1) = 0. The result then follows. Definition. Let A be a graded ring, C the class of finitely generated A0-modules and λ an additive function on C. Then for any module M ∈ C, we define the Poincar´eseries of M with respect to λ as ∞ X n PM (t) = λ(Mn)t ∈ Z[[t]]. n=0 Formally, the Poincar´eseries of any module is a power series with coefficients in Z.A surprising result is that for finitely generated A0-module M, PM (t) is in fact a rational function. Theorem 6.1.3 (Hilbert, Serre). For any M ∈ C and additive function λ, there is some polynomial f(t) ∈ Z[t] such that f(t) PM (t) = s ∈ (t). Q ki Q i=1(1 − t ) Proof. We prove this by inducting on s, the number of generators of A0-generators of A. If s = 0, A = A0 and M is a finitely generated A-module, so Mn = 0 for all n > max{rj | 1 ≤ j ≤ t}. Thus PM (t) is a polynomial in this case, so f(t) = PM (T ) ∈ Z[t] works. Now assume the result holds for all modules in C generated as an A0-module by s − 1 elements or fewer. Consider the A-module homomorphism f : M → M, z 7→ xsz. This gives us A-modules K = ker f and L = coker f = M/xsM. Since A is noetherian, K and L are finitely generated (by Corollary 3.1.5 and Proposition 3.1.2). Note that xsK = f(K) = 0 and ¯ xsL = xsM/xsM = 0, so K and L are in fact graded modules over the quotient A = A/xsA: L∞ L∞ K = n=0 Kn and L = n=0 Ln. Also notice that Lr = Mr whenever r < ks. Moreover, by Lemma 3.1.3, A¯ is noetherian so K and L are finitely generated as A¯-modules. We can view A¯ as a graded noetherian ring: ∞ ¯ M ¯ ¯ A = An = A0[¯x1,..., x¯s] n=0 ¯ where An = An/xsAn−ks andx ¯i = xi + xsAki−ks for all n, i. Since ki ≥ 1 for all i, we actually ¯ ∼ ¯ have A0 = A0, so the finitely generated modules over A0 are precisely the elements of C. In a moment, this will allow us to apply the induction hypothesis to K and L. For n ∈ N0, let fn denote the restriction of f to Mn. We get the following exact sequences: fn 0 → Kn → Mn −→ Mn+ks → Ln+ks → 0 for n ∈ N0. 90 6.1 Hilbert Functions 6 Dimension Theory n+ks By Lemma 6.1.2, λ(Kn) − λ(Mn) + λ(Mn+ks ) − λ(Ln+ks ) = 0. Multiplying through by t and rearranging terms yields ks n n n+ks t (t λ(Kn) − t λ(Mn)) = t (λ(Ln+ks ) − λ(Mn+ks )) ks =⇒ t (PK (t) − PM (t))n+ks = (PL(t) − PM (t))n+ks for all n ∈ N0 ks =⇒ t (PK (t) − PM (t)) = PL(t) − PM (t) since both power series have 0 coefficients for indices less than ks, and the second line implies that the power series agree in every remaining term. This can be rearranged to read ks ks (1 − t )PM (t) = PL(t) − t PK (t). By the induction hypothesis, there exist polynomials g(t), h(t) ∈ Z[t] such that s−1 s−1 Y ki Y ki (1 − t )PL(t) = g(t) and (1 − t )PK (t) = h(t). i=1 i=1 It follows that s Y ki ks (1 − t )PM (t) = g(t) − t h(t) ∈ Z[t]. i=1 Thus the result holds for PM (t) for all M ∈ C by induction. Definition. The dimension of a graded A-module M is defined to be the order of the pole of PM (t) at t = 1, denoted d(M). Formally, we set d(M) = 0 if there is no pole at t = 1. For a graded A-module M of dimension d = d(M), Theorem 6.1.3 allows us to write the Poincar´eseries M as g(t) P (t) = M h(t)(1 − td) where h(1) 6= 0, and if d ≥ 1, g(1) 6= 0. Corollary 6.1.4. If d(M) ≥ 1 and x ∈ Ak such that xz 6= 0 for all z ∈ M r {0}, then d(M/xM) = d(M) − 1. Proof. Consider the map f : M → M, z 7→ xz. As in Theorem 6.1.3, we have modules K = ker f and L = coker f = M/xM. Since xz 6= 0 for any z ∈ M r {0}, K = 0. Then the k proof of Theorem 6.1.3 shows that (1−t )PM (t) = PL(t). It follows that d(L) = d(M)−1. Lemma 6.1.5. For all n, s ∈ N0 with n ≥ s, n + s − 1 #{(i , . . . , i ) ∈ s : i + ... + i = n} = . 1 s N0 1 s s − 1 Example 6.1.6. Let k be a field and consider the polynomial ring A = k[t1, . . . , ts]. Then by Example 5.4.1, A is a graded ring with M i1 is A0 = k and An = kt1 ··· ts for each n ≥ 1. ij ∈N0 i1+...+is=n 91 6.1 Hilbert Functions 6 Dimension Theory Let λ(M) = dimk M be the additive dimension function for k-vector spaces M. Notice that n+s−1 by Lemma 6.1.5, dimk An = s−1 for each n ≥ 1, so the Poincar´eseries for A as a module over itself is ∞ ∞ X X n + s − 1 1 P (t) = (dim A )tn = tn = . A k n s − 1 (1 − t)s n=0 n=0 Hence d(A) = s so the notion of vector space dimension coincides with the dimension of a graded ring. Notice that in this example, k1 = k2 = ··· = ks, i.e. the generators t1, . . . , ts are all homogeneous of degree 1. The following proposition generalizes this and introduces the Hilbert polynomial of an A-module. Proposition 6.1.7. Suppose A = A0[x1, . . . , xs] is a graded noetherian ring such that xi ∈ A1 for all 1 ≤ i ≤ s. Then for any additive function λ and graded A-module M, there exists n0 ∈ N0 and a polynomial g(t) ∈ Q[t] with deg g = d(M) − 1 such that λ(Mn) = g(n) for all n ≥ n0.