# LIVIA HUMMEL AND THOMAS MARLEY

Size: px
Start display at page:

Transcription

1 THE AUSLANDER-BRIDGER FORMULA AND THE GORENSTEIN PROPERTY FOR COHERENT RINGS LIVIA HUMMEL AND THOMAS MARLEY Abstract. The concept of Gorenstein dimension, defined by Auslander and Bridger for finitely generated modules over a Noetherian ring, is studied in the context of finitely presented modules over a coherent ring. A generalization of the Auslander-Bridger formula is established and is used as a cornerstone in the development of a theory of coherent Gorenstein rings. 1. Introduction In addressing a problem posed by Glaz ([11], [12]), Hamilton and the second author give a definition of Cohen-Macaulay for commutative rings which agrees with the usual notion for Noetherian rings with the property that every coherent regular ring is Cohen- Macaulay [13]. (A quasi-local ring is defined to be regular if every finitely generated ideal has finite projective dimension.) A natural question is whether there is a reasonable concept of Gorenstein for commutative rings such that every coherent regular ring is Gorenstein and every coherent Gorenstein ring is Cohen-Macaulay. In this paper, we develop such a theory of coherent Gorenstein rings which mirrors much of the theory in the Noetherian case. Central to this development is the concept of Gorenstein dimension (G-dimension, for short), first introduced in the context of finitely generated modules over Noetherian rings by Auslander and Bridger [1]. In particular, we prove the following generalization of the Auslander-Bridger formula for coherent rings using a notion of depth for arbitrary quasi-local rings developed by Barger [2], Hochster [14], and Northcott [18]: Date: February 25, Mathematics Subject Classification. Primary 13D05, 13D07. Key words and phrases. Gorenstein dimension, coherent ring. 1

2 2 LIVIA HUMMEL AND THOMAS MARLEY Theorem 1.1. Let R be a quasi-local coherent ring and M a finitely presented R-module of finite G-dimension. Then depth M + Gdim R M = depth R. Since working with depth in the non-noetherian case typically requires passage to faithfully flat extensions (which does not generally preserve coherence), our development of G-dimension must be done with some care. Out of necessity, we prove a version of Theorem 1.1 which holds for modules over an arbitrary commutative ring which have free resolutions consisting of finitely generated free modules in each degree (so-called (F P ) modules). Auslander and Bridger prove that every ideal having finite G-dimension characterizes Noetherian local Gorenstein rings. It is natural, therefore, to make the following definition: Definition 1.2. A quasi-local ring R is said to be Gorenstein if every finitely generated ideal has finite G-dimension. We are able to prove several properties of coherent Gorenstein rings which are analogous to results in the Noetherian case. We summarize some of these results in the following: Theorem 1.3. Let R be a quasi-local coherent ring. (1) If x is a non-unit non-zero-divisor on R, then R is Gorenstein if and only if R/(x) is Gorenstein. (2) If x is an indeterminate over R, then R is Gorenstein if and only if R[x] is Gorenstein. (3) If depth R <, then R is Gorenstein if and only if R has finite FP-injective dimension. (4) Every coherent regular ring is Gorenstein. (5) Every coherent Gorenstein ring is Cohen-Macaulay. (6) If R is a coherent quasi-local Gorenstein ring with depth R = n, and x 1,..., x n is a regular sequence on R, then (x 1,..., x n ) is an irreducible ideal. The paper is structured in the following way: In section two, we summarize the main results we need concerning non-noetherian depth and (F P ) modules. Section three defines

3 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 3 G-dimension and restricted G-dimension and proves their basic properties. Section four is devoted to the proof of the Auslander-Bridger formula for modules of finite restricted G- dimension, while in section five the main results concerning Gorenstein rings are established. 2. Preliminaries Throughout R will always denote a commutative ring with identity. All modules are assumed to be unital. The term local ring will be used exclusively for commutative Noetherian rings with a unique maximal ideal. When the ring is not necessarily Noetherian, the term quasi-local will be used. A ring is coherent if every finitely generated ideal is finitely presented. See [10] for basic properties of coherent rings. The following definition is due to Bieri [4]: Definition 2.1. An R-module M is said to be (F P ) R n exact sequence F n F n 1 F 1 F 0 M 0 for some n 0 if there exists an where F 0,..., F n are finitely generated free R-modules. For brevity we often write (F P ) n for (F P ) R n when there is no ambiguity about the ring R. If M is (F P ) n for all n 0 then we say M is (F P ). We note that over a Noetherian ring the class of (F P ) -modules coincides with the class of of finitely generated modules. Over a coherent ring, (F P ) -modules are just the finitely presented modules (cf.[10, Corollary 2.5.2]). Lemma 2.2. Let M be an R-module and n 0 an integer. The following are equivalent: (1) M is (F P ) n+1. (2) The functors Ext i R (M, ) commute with direct limits for i = 0,..., n. Proof. See [4, Corollary 1.6]. Lemma 2.3. Let 0 L M N 0 be a short exact sequence of R-modules. Then the following hold for any n 0: (1) If L is (F P ) n and M is (F P ) n+1 then N is (F P ) n+1. (2) If M and N are (F P ) n+1 then L is (F P ) n.

4 4 LIVIA HUMMEL AND THOMAS MARLEY (3) If L and N are (F P ) n then M is (F P ) n. Consequently, if any two modules in a short exact sequence are (F P ) then so is the third. Proof. See [4, Proposition 1.4] or [10, Theorem 2.1.2]. The following remark is easily seen from the definition of (F P ) n and Lemma 2.3: Remark 2.4. Let S be a flat R-algebra, M an R-module, and n 0 an integer. (1) If M is (F P ) R n then M R S is (F P ) S n. (2) If S is faithfully flat then the converse to (1) holds. We also will need the following: Remark 2.5. Suppose M is (F P ) R and x R is a non-zero-divisor on R and M. Then M/xM is (F P ) R/(x). Proof. Let F be a free resolution of M consisting of finitely generated free R-modules. As x is a non-zero-divisor on R and M, Tor R i (M, R/(x)) = 0 for all i > 0. Hence, F R R/(x) is a resolution of M/xM consisting of finitely generated free R/(x)-modules. Depth plays an important role in the theory of G-dimension over local rings. The notion of depth was extended to quasi-local rings by Hochster [14]: Definition 2.6. Let M be an R-module and I an ideal of R such that IM M. Define the classical depth of M with respect to I, denoted Depth I M, to be the supremum of the lengths of regular sequences on M contained in I. Define the depth of M with respect to I, denoted depth I M, to be the supremum of Depth IS M R S over all faithfully flat extensions S of R. If R is quasi-local with maximal ideal m, we denote depth m M by depth R M or simply depth M. We note that if R is Noetherian and M is finitely generated then depth I M = Depth I M for all ideals I of R (see, for instance, Lemma 2.8 below). We summarize some basic properties of depth in the following Proposition. We refer the reader to [2], [14], [18, Chapter 5], [10, Chapter 7] or [5, Section 9.1] for details. Proposition 2.7. Let M be an R-module and I an ideal of R such that IM M.

5 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 5 (1) depth I M = sup{depth J M J I, J finitely generated}. (2) If I = (x 1,..., x n ) then depth I M = inf{i 0 H n i (x, M) 0}, where H j (x, M) denotes the jth Koszul homology of x = x 1,..., x n on M. (3) depth I M = depth IS (M R S) for any faithfully flat R-algebra S. (4) If depth I M > 0 then Depth IS M R S > 0 where S = R[X], a polynomial ring in one variable over R. (5) If I is generated by n elements then depth I M = Depth IS M R S where S = R[X 1,..., X n ], a polynomial ring in n variables over R. (6) depth I M = depth I M. (7) If x I is M-regular then depth I M = depth I M/xM + 1. (8) Suppose 0 L M N 0 is a short exact sequence of R-modules such that IL L and IN N. If depth I M > depth I N then depth I L = depth I N + 1. In the case I is (F P ) we have another characterization of depth I M: Lemma 2.8. Let M be an R-module and I an ideal such that IM M. Suppose R/I is (F P ) n. The following are equivalent: (1) depth I M n. (2) Ext i R (R/I, M) = 0 for 0 i < n. Proof. See Theorem or Theorem of [10]. Combining Lemma 2.8 with part (1) of Proposition 2.7, we have: Proposition 2.9. Let R be a coherent quasi-local ring with maximal ideal m and M a nonzero R-module such that mm M. Then depth M := sup{n 0 Ext i R(R/I, M) = 0 for all i < n for some f.g. ideal I m}. 3. Gorenstein dimension As noted in the introduction, the theory of Gorenstein dimension, or G-dimension, for finitely generated modules over a Noetherian ring was developed by Auslander and Bridger

6 6 LIVIA HUMMEL AND THOMAS MARLEY in [1]. This theory was later generalized for modules over arbitrary rings (Gorenstein projective dimension) by Enochs and Jenda [8]. However, for the notion of depth to be used effectively (e.g., as an inductive tool), one needs to restrict to a resolving class of finitely generated modules so that Nakayama s lemma can be applied. Hence, our development of G-dimension falls somewhere between Auslander-Bridger s and Enochs-Jenda s. While our main interest is the case of finitely presented modules over coherent rings, restricting the theory to such rings does not allow one to fully utilize the concept of depth, which requires passage to faithfully flat extensions to guarantee the existence of non-zero-divisors. (It is well-known that coherence is not preserved under faithfully flat extensions, even finitely generated ones. See [10, Example ], for example.) For an R-module M we let M denote the dual module Hom R (M, R). Following Auslander and Bridger [1], we make the following definition: Definition 3.1. A finitely generated R-module M is said to be a member of the G-class of R, denoted G(R), if the following hold: (1) Ext i R (M, R) = 0 for i > 0. (2) Ext i R (M, R) = 0 for i > 0. (3) The natural map M M is an isomorphism. The class G(R) is closed under extensions and summands. In particular, every finitely generated projective module is a member of G(R). However, if R is not Noetherian, G(R) may not be closed under duals (since the dual of a member of G(R) may not be finitely generated see Example 5.2). A much better behaved class for our purposes is the following: Definition 3.2. An R-module M is a member of the restricted G-class of R, denoted G(R), if the following hold: (1) M is in G(R). (2) M is (F P ). (3) M is (F P ). The class G(R) is easily seen to be closed under extensions, summands, and duals. Additionally, every finitely generated projective R-module is a member of G(R). Since any

7 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 7 reflexive module is isomorphic to a submodule of a free module, we note that modules in G(R) and G(R) are torsion-free. Let M be an R-module. A complex G is called a G-resolution (respectively, G-resolution) of M if each G i is a member of the G-class (respectively, G-class), Gn = 0 for n < 0, H i (G) = 0 for all i 0, and H 0 (G) = M. For a resolution G we set the length of G to be the infimum of the set of integers n such that G n 0. (By convention, we define the length of the zero complex to be zero.) For an integer n we define the nth syzygy of G to be the kernel of the map G n 1 G n 2 (here we let G 1 = M). If M has a G-resolution (respectively, G-resolution), we define the G-dimension of M, Gdim R M (respectively, the G-dimension of M, GdimR M), to be the infimum of the lengths of all G-resolutions (respectively, Gresolutions) of M. Otherwise, we say the G-dimension (respectively, G-dimension) of M is undefined. We remark that Gdim R M and Gdim R M are defined for any R-module M which is (F P ), since in this case M has a resolution consisting of finitely generated free R-modules. We also note that any syzygy module of a G-resolution (respectively, G-resolution) is finitely generated (respectively, (F P ) ). Since every module in G(R) is a member of G(R), it is clear that Gdim R M Gdim R M for every (F P ) R-module M. We ll show below that if Gdim R M < then Gdim R M = Gdim R M (Corollary 3.6). Furthermore, if R is coherent and M is finitely presented then Gdim R M = Gdim R M (Corollary 3.7). In the remainder of this section we summarize the basic properties of G-dimension and G-dimension. The proofs of many of these results are similar to the analogous results for finitely generated modules over a Noetherian ring, once one observes the relevant modules are finitely generated or (F P ). In a few instances alternative arguments must be made. We refer the reader to Chapter 1 of [6] for a clear and thorough treatment of G-dimension over Noetherian rings. More complete proofs of the results below can be found in [17]. Proposition 3.3. Let 0 L M N 0 be a short exact sequence of R-modules. Suppose N is a member of the G-class (respectively, G-class) and L and M are finitely generated (respectively, (F P ) ). Then L G(R) if and only if M G(R) (respectively, L G(R) if and only if M G(R)).

8 8 LIVIA HUMMEL AND THOMAS MARLEY Proof. See [6, Lemma (a)]. We add the observation that, as 0 N M L 0 is exact and N is (F P ), we have L is (F P ) if and only if M is (F P ) by Lemma 2.3. Lemma 3.4. Let M be an R-module of finite G-dimension (respectively, G-dimension) such that Ext i R (M, R) = 0 for all i > 0. Then Gdim R M = 0 (respectively, Gdim R M = 0). Proof. See [6, Lemma 1.2.6]. The following proposition will be used frequently: Proposition 3.5. Let M be an R-module which is (F P ) and n 0 an integer. The following are equivalent: (1) Gdim R M n. (2) Gdim R M < and Ext i R (M, R) = 0 for i > n. (3) The nth syzygy of any G-resolution of M is in the G-class of R. The same statement also holds if Gdim, G-resolution, and G-class are replaced with Gdim, G-resolution, and G-class, respectively. Proof. The proof is similar to that of [6, Theorem 1.2.7], substituting Proposition 3.3 and Lemma 3.4 where appropriate. We add the observation that, using the notation of the proof of [6, Theorem 1.2.7], H n is (F P ) if and only if K n is (F P ). Consequently, we have the following: Corollary 3.6. Let M be a nonzero R-module of finite G-dimension. Then Gdim R M = sup{i 0 Ext i R(M, R) 0}. The same statement also holds if G-dimension is replaced by G-dimension. Hence, if M has finite G-dimension then Gdim R M = Gdim R M. We note that if R is coherent and M is a finitely presented R-module then M is also finitely presented (and hence (F P ) ). This follows from the more general fact that, over a coherent ring, kernels of homomorphisms between finitely presented modules are again finitely presented (cf. Chapter 2 of [10]). In light of this observation, we have:

9 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 9 Corollary 3.7. Let R be a coherent ring and M a finitely presented R-module. Gdim R M = Gdim R M. Then Proof. Note that as R is coherent and M is finitely presented, M is (F P ). Hence Gdim R M and Gdim R M are defined. If Gdim R M = then clearly Gdim R M =. Suppose Gdim R M = n <. Let G be a G-resolution of M. By Proposition 3.5, the nth syzygy K n of G is in G(R) and is (F P ). By the remark above, Kn is (F P ). Hence K n is in G(R) and Gdim R M <. The result now follows from Corollary 3.6. As we are mainly interested in G-dimension in the context of coherent rings, and since G-dimension is equal to G-dimension for finitely presented modules over coherent rings, we will subsequently state results only in terms of G-dimension (although some results may also hold for G-dimension). For an R-module M we let pd R M denote the projective dimension of M. If M is (F P ) and pd R M = n it is easily seen that M has a projective resolution of length n consisting of finitely generated projective modules in each degree. Proposition 3.8. Suppose M is (F P ) and has finite projective dimension. Then Gdim R M = pd R M. Proof. Let n = pd R M and m = G dim R M. As remarked above, there exists a projective resolution P of M of length n where P i is finitely generated for each i. Hence, P is a G-resolution of M of length n and m n. Let K m be the mth syzygy of P. By Proposition 3.5, K m is in G(R). It is enough to show that K m is projective. Since 0 P n P n 1 P m+1 K m 0 is exact and Ext i R (K, R) = 0 for all i > 0, we have is also exact. As P i 0 K m P m+1 P n 0 is projective for all i, K m is projective. Hence, K m = K m is projective. Proposition 3.9. Suppose 0 K G M 0 is exact where Gdim R M > 0 and G is in G(R). Then Gdim R K = Gdim R M 1.

10 10 LIVIA HUMMEL AND THOMAS MARLEY Proof. By Lemma 2.3, K is (F P ) and Gdim R K is defined. It is clear that if Gdim R M = then Gdim R K =, so we may assume n = Gdim R M is positive and finite. Clearly, Gdim R K n 1. Let F be a resolution of K consisting of finitely generated free R- modules. Since the composition F G gives a G-resolution of M, the (n 1)st syzygy of F is in G(R) by Proposition 3.5. Hence, Gdim R K n 1. We will also need the following: Proposition Let 0 L M N 0 be an exact sequence of R-modules each of which is (F P ). Then the following hold: (1) If Gdim R L n and Gdim R N n, then Gdim R M n. (2) If Gdim R M n and Gdim R N n, then Gdim R L n. (3) If Gdim R L n and Gdim R M n, then Gdim R N n + 1. In particular, if any two of the modules has finite G-dimension, then so does the third. Proof. We prove (3). The remaining two parts are proved similarly. Let F and F be freeresolutions of L and N, respectively, which consist of finitely generated free R-modules. By the Horseshoe Lemma, there exists a free resolution F of M consisting of finitely generated free R-modules and chain maps F F and F F such that is an exact sequence of complexes. 0 F F F 0 Let K n, K n, and K n denote the nth syzygies of F, F, and F, respectively. Then the sequence 0 K n K n K n 0 is exact. Since L and M have G-dimension at most n, we have that K n and K n are in G(R) by Proposition 3.5. Hence, Gdim R K n 1, which implies that Gdim R N n + 1. The next result shows the restricted G-class is preserved by flat base change: Proposition Let S be a flat R-algebra and M an R-module. (1) If M is in G(R) then M R S is in G(S). (2) If S is faithfully flat then the converse of (1) holds.

11 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 11 Proof. For an S-module N, let N v denote Hom S (N, S). By Remark 2.4, if M is (F P ) R then M R S is (F P ) S, and the converse holds if S is faithfully flat. Hence, in the proof of (1) or its converse (under the assumption that S is faithfully flat), there are natural isomorphisms Ext i R (M, R) R S = Ext i S (M R S, S) for all i 0. In particular, M R S = (M R S) v. Again by Remark 2.4, we have that M is (F P ) R and (M R S) v is (F P ) S in both the proof of (1) and its converse; furthermore, we have natural isomorphisms Ext i R (M, R) R S = Ext i S ((M R S) v, S) for all i 0. Thus, if Ext i R (M, R) = Exti R (M, R) = 0 for i > 0, then Ext i S (M R S, S) = Ext i S ((M R S) v, S) = 0 for all i > 0, with the converse holding if S is faithfully flat. Let ρ : M M and ϕ : M R S (M R S) vv be the canonical maps. Let K and C be the kernel and cokernel of ρ, and K and C the kernel and cokernel of ϕ. Then by the flatness of S we have the following commutative diagram: 0 K R S M R S ρ 1 = M R S C R S 0 = 0 K M R S ϕ (M R S) vv C 0. If ρ is an isomorphism, then so is ρ 1 and thus ϕ also. Conversely, if ϕ is an isomorphism and S is faithfully flat, then K = C = 0 and thus ρ is an isomorphism. Membership in the restricted G-class is a local condition, as the next result demonstrates. Proposition Suppose M and M are (F P ). The following are equivalent: (1) M is in G(R) (2) M p is in G(R p ) for all prime ideals p. (3) M m is in G(R m ) for all maximal ideals m. Proof. By Proposition 3.11, it is enough to prove that (3) implies (1). As M and M are (F P ), Ext i R m (M m, R m ) = Ext i R (M, R) RR m = 0 and Ext i R m ((M m ), R m ) = Ext i R (M, R) R R m = 0 for all i > 0 and all maximal ideals m. Furthermore, there are natural isomorphisms (M ) m = (Mm ) v and (M ) m = (Mm ) vv, where ( ) v = Hom Rm (, R m ). Let ρ : M M be the canonical homomorphism and K = ker ρ and C = coker ρ. By hypothesis, for each maximal ideal m we have M m is in G(R m ), which implies K m = C m = 0 for all m. Hence, ρ is an isomorphism and M G(R).

12 12 LIVIA HUMMEL AND THOMAS MARLEY Corollary Suppose M is (F P ). Then (1) Gdim R M Gdim S M R S for all flat R-algebras S. (2) Gdim R M = Gdim S M R S for all faithfully flat R-algebras S. (3) If in addition M is (F P ) then Gdim R M = sup{ Gdim Rm M m m a maximal ideal of R}. Proof. Part (1) is clear from Proposition For part (2), suppose Gdim S M R S = n. Let G be a G-resolution of M and K the nth syzygy of G. Since G R S is a G(S)-resolution of M R S, we have that K R S is in G(S) by Proposition 3.5. By Proposition 3.11, K is in G(R) and hence Gdim R M n. Part (3) is proved similarly using Proposition The following is a basic change of rings result. Lemma Let M be in G(R) and x R a non-zero-divisor on R. Then M/xM is in G(R/(x)). Proof. Since both M and M are torsion-free, x is a non-zero-divisor on M and M. Thus, M/xM and M /xm are (F P ) R/(x) to Lemma of [6]. by Remark 2.5. The remainder of the proof is identical Lemma Let M be an R-module such that Gdim R M n. finitely presented. Then Ext n R (M, R) is Proof. Since Ext i R (M, R) = 0 for i > Gdim R M, we may assume n = Gdim R M. If n = 0 the result is clear. Suppose n > 0 and let 0 K G M 0 be exact where G is in G(R). By Proposition 3.9, GdimR K = n 1. By induction, Ext n 1 (K, R) is finitely presented. By the exactness of Ext n 1 R (G, R) Extn 1 (K, R) Extn R(M, R) 0 R R we see that Ext n R (M, R) is finitely presented (e.g., Lemma 2.3). Let J(R) denote the Jacobson radical of R. Proposition Let M be an R-module which is (F P ) and x J(R) a non-zerodivisor on M and R. Suppose M/xM is in G(R/(x)).

13 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 13 (1) If Gdim R M < then M is in G(R). (2) If R is coherent then M is in G(R). Proof. We prove part (1). The proof of (2) is similar to the proof of Lemma of [6], where one uses coherence to ensure the kernel and cokernel of the canonical map M M are finitely generated so that Nakayama s lemma can be applied. Let Gdim R M = n and assume n > 0. As x is a non-zero-divisor on M and R, Ext i R (M, R/(x)) = Ext i R/(x) (M/xM, R/(x)) = 0 for i > 0. Then Ext n R (M, R) is nonzero and finitely generated (Lemma 3.15). Applying Hom R (M, ) to the short exact sequence 0 R x R R/(x) 0 yields the exact sequence Ext n R (M, R) x Ext n R (M, R) 0, which implies by Nakayama s lemma that Ext n R (M, R) = 0, a contradiction. Hence, n = 0. Applying these results to restricted G-dimension, we have: Proposition Let M be an R-module which is (F P ) and x R a non-zero-divisor on M and R. Then (1) Gdim R/(x) M/xM Gdim R M. (2) If x J(R) and either M has finite G-dimension or R is coherent then Gdim R/(x) M/xM = Gdim R M. Proof. For part (1), we may assume Gdim R M = n <. If n = 0 the result follows by Lemma If n > 0 then there exists a short exact sequence 0 K G M 0 where G is in G(R) and Gdim R K = n 1. Since x is a non-zero-divisor on M and R, the sequence 0 K/xK G/xG M/xM 0 is exact. By the induction hypothesis, GdimR/(x) K/xK n 1 and G/xG is in G(R/(x)) by the n = 0 case. Hence, Gdim R/(x) M/xM n. For part (2), it is enough to show Gdim R M Gdim R/(x) M/xM. Again, we may assume Gdim R/(x) M/xM = n <. If n = 0 the result follows by Proposition Otherwise, consider a short exact sequence 0 K G M 0 where G is in G(R). Then 0 K/xK G/xG M/xM 0 is exact and G/xG is in G(R/(x)). By Proposition

14 14 LIVIA HUMMEL AND THOMAS MARLEY 3.9 we have Gdim R/(x) K/xK = n 1. Since x is a non-zero-divisor on K we have by the induction hypothesis that Gdim R K n 1. Hence, Gdim R M n. We will need the following lemma in the proof of Theorem 3.19: Lemma Let n be a nonnegative integer and consider an exact sequence of R-modules 0 M A 0 A 1 A n 0. Suppose that for 1 i n we have Ext j R (A i, R) = 0 for 1 j i. Then is exact. 0 A n A n 1 A 0 M 0 Proof. We use induction on n, the cases n = 0 and n = 1 being trivial. If n > 1 let K be the kernel of the map A n 1 A n. It follows easily from the short exact sequence 0 K A n 1 A n 0 that 0 A n A n 1 K 0 is exact and Ext i R (K, R) = 0 for 1 i n 1 By the induction hypothesis, we have that 0 K A n 2 A 0 M 0 is exact. Patching the two sequences we obtain the desired result. The proof of the following important change of rings theorem is similar to the argument of [1, Proposition 4.35] except for a simplification (using the lemma above) which replaces a tedious calculation [1, p. 137]. Theorem Let M be a nonzero R-module such that Gdim R M < and suppose x is a non-zero-divisor on R such that xm = 0. Then Gdim R/(x) M = Gdim R M 1. Proof. We use induction on Gdim R M. Note that since x is a non-zero-divisor on R and annihilates M, M cannot be reflexive. Hence Gdim R M 1. Suppose Gdim R M = 1. Then there exists an exact sequence 0 G 1 G 0 M 0 such that G 1 and G 0 are in G(R). Let R denote R/(x). For an R-module N let N v denote Hom R (N, R) (to distinguish this dual from N = Hom R (N, R)). As x is a non-zero-divisor on R and xm = 0, we have

15 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 15 that M = 0 and Ext i+1 R (M, R) = Ext i (M, R) for i 0 by [16, Lemma 18.2]. Hence, R M v = Ext 1 R (M, R) and Exti R (M, R) = 0 for all i > 0. Applying Hom R(, R) to the short exact sequence above, we have 0 G 0 G 1 M v 0. Thus, GdimR M v = 1. Repeating the above argument with M v in place of M, we get Ext i R (M v, R) = 0 for all i > 0. It remains to show that M is reflexive as an R-module. It is readily seen by the assumptions on x that Tor R 1 (M, R/(x)) = M and Tor 1 (G 0, R/(x)) = 0. Hence we have an exact sequence of R-modules 0 M G 1 /xg 1 G 0 /xg 0 M 0. Note that by Lemma 3.14, G 1 /xg 1 and G 0 /xg 0 are in G(R). Applying Hom R (, R) twice to this exact sequence and using Lemma 3.18, we obtain the exactness of 0 M vv (G 1 /xg 1 ) vv (G 0 /xg 0 ) vv M vv 0. Since G 0 /xg 0 and G 1 /xg 1 are reflexive (as R-modules), we see that M is reflexive by the five lemma. Hence, Gdim R M = 0. Suppose Gdim R M = n > 1. Let ϕ : G M be a surjective homomorphism where G G(R). Tensoring with R/(x) we get a surjection ϕ : G/xG M. Letting K be the kernel of ϕ we get a short exact sequence of R-modules 0 K G/xG M 0. As x is a non-zero-divisor on G, 0 G x G G/xG 0 is exact. Hence, Gdim R G/xG = 1. (Note that G/xG is nonzero as M is.) By Proposition 3.10, Gdim R K <. Moreover, as Gdim R M = n > 1, it readily follows from Proposition 3.5 that Ext i R (K, R) = 0 for i > n 1 and Ext n 1 R (K, R) 0. Hence Gdim R K = n 1. Since xk = 0 we have by the induction hypothesis that Gdim R/(x) K = n 2. Since G/xG is in G(R/(x)) (Lemma 3.14), we have Gdim R/(x) M = n 1 by Proposition 3.9. In the coherent case, we have the same result holding if the assumption Gdim R M < is replaced with Gdim R/(x) M/xM < and x J(R):

16 16 LIVIA HUMMEL AND THOMAS MARLEY Proposition Let R be coherent, M a nonzero finitely presented R-module, and x J(R) a non-zero-divisor on R such that xm = 0. If Gdim R/(x) M < then Gdim R/(x) M = Gdim R M 1. Proof. The proof is virtually identical to the proof of Lemma of [6], where we use part (2) of Proposition 3.17 in place of [6, Proposition 1.4.5]. Combining Theorem 3.19, Proposition 3.20, and Proposition 3.17 we obtain: Corollary Let R be coherent, M a nonzero finitely presented R-module, and x J(R) a non-zero-divisor on R. (1) If xm = 0 then Gdim R/(x) M = Gdim R M 1. (2) If x is a non-zero-divisor on M then Gdim R M/xM = Gdim R M The Auslander-Bridger formula for restricted Gorenstein dimension In this section we prove the Auslander-Bridger formula for modules of finite restricted G-dimension over a quasi-local ring. The proof here differs by necessity from those given in [1] and [6] to avoid arguments using the theory of associated primes. One also needs to pass to faithfully flat extensions to make the induction argument work. We begin with the following elementary result: Lemma 4.1. Let R be a quasi-local ring with depth R = 0 and M a finitely presented R-module. Then M = 0 if and only if M = 0. Proof. Let M be generated by n elements and suppose M = 0. We show by induction on n that M = 0. If n = 1 then M = R/I for some finitely generated ideal I. Since depth R = 0 we have Hom R (R/I, R) 0 unless I = R (Lemma 2.8). As M = Hom R (R/I, R) = 0 we must have I = R and hence M = 0. If n > 1 then there exists a submodule M of M generated by n 1 elements and such that M/M is cyclic. Clearly, M/M is finitely presented and (M/M ) = 0. By the n = 1 case, we have M/M = 0. Hence, M is generated by n 1 elements and M = 0. Lemma 4.2. Let R be a quasi-local ring with depth R = 0 and M a nonzero R-module of finite G-dimension. Then Gdim R M = 0 and depth M = 0.

17 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 17 Proof. By induction it suffices to prove the case when Gdim R M 1. Then there exists a short exact sequence 0 G 1 G 0 M 0 where G 0 and G 1 are in G(R). Applying Hom R (, R) twice, we get the exact sequence 0 Ext 1 R(M, R) G 1 G 0. Since G 0 and G 1 are reflexive and the map G 1 G 0 is injective, we obtain Ext 1 R (M, R) = 0. As Ext 1 R (M, R) is finitely presented by Lemma 3.15, we have Ext1 R (M, R) = 0 by Lemma 4.1. Hence, Gdim R M = 0 by Proposition 3.5. By way of contradiction, assume depth M > 0. Let m be the maximal ideal of R and S = R[x] mr[x] where x is an indeterminate over R. As S is faithfully flat over R, M R S is in G(S), depth S S = depth R R = 0, and depth S M R S = depth M (cf. Propositions 3.11 and 2.7, part (3)). Furthermore, Depth S M R S > 0 by part (4) of Proposition 2.7. Thus by passing to S and resetting notation, we may assume there exists an element x m such that x is a non-zero-divisor on M. Applying Hom R (, R) to the short exact sequence 0 M x M M/xM 0 we get the exact sequence 0 (M/xM) M x M Ext 1 R(M/xM, R) 0. In particular, this sequence shows that Ext 1 R (M/xM, R) is finitely presented. Dualizing again, we have 0 Ext 1 R(M/xM, R) M x M is exact. Since M = M and x is a non-zero-divisor on M we have Ext 1 R (M/xM, R) = 0. By Lemma 4.1, we see that Ext 1 R (M/xM, R) = 0. From the exact sequence above, this implies M = xm. By Nakayama s Lemma we have M = 0, a contradiction. Thus, depth M = 0. Lemma 4.3. Let R be a quasi-local ring and M a nonzero R-module in G(R). depth R = depth M. Then Proof. Let m denote the maximal ideal of R. Let n = depth R and suppose n <. If n = 0 the result holds by Lemma 4.2. If n > 0, as in the proof of Lemma 4.2 we may

18 18 LIVIA HUMMEL AND THOMAS MARLEY assume, by passing to R[x] mr[x] if necessary, that there exists x m which is a non-zerodivisor on R. Then x is also a non-zero-divisor on M (as reflexive modules are torsionfree) and M/xM is in G(R/(x)) by Lemma Hence depth R/(x) R/(x) = n 1 and depth R/(x) M/xM = depth M 1. By induction on n, we obtain depth R/(x) M/xM = n 1 and hence depth M = n. Suppose now that depth R =. We prove that depth M n for all n. The case n = 0 is trivial. Assume that for all quasi-local rings S with depth S = and modules N in G(S), that depth S N n. As above, we may assume there exists x m such that x is a nonzero-divisor on R (and hence M). Then M/xM is in G(R/(x)), depth R/(x) R/(x) =, and depth R/(x) M/xM = depth M 1. By assumption, depth R/(x) M/xM n which implies depth M n + 1. We now prove the Auslander-Bridger formula for restricted G-dimension: Theorem 4.4. Let R be a quasi-local ring and M a nonzero R-module of finite G-dimension. Then depth M + Gdim R M = depth R. Proof. First assume depth R =. If Gdim R M = 0 then depth M = by Lemma 4.3. Suppose Gdim R M = n > 0 and depth N = for all R-modules N such that Gdim R N < n. Let 0 K G M 0 be exact where G is in G(R). Then Gdim R K = n 1 and hence depth K =. Since depth G =, depth M = by part (8) of Proposition 2.7 and the equality holds. Assume now that depth R <. We proceed by induction on depth R. If depth R = 0 the formula holds by Lemma 4.2. Suppose depth R = n > 0. Let m denote the maximal ideal of R. If depth M > 0 we may assume, by passing to R[x] mr[x] if necessary, that there exists x m such that x is a non-zero-divisor on R and M. (In this case, one can show depth M R > 0 and we may apply part (4) of Proposition 2.7.) Then Gdim R/(x) M/xM = Gdim R M by Proposition Furthermore, depth R/(x) M/xM = depth M 1 and depth R/(x) R/(x) = depth R 1. By the induction hypothesis, we have depth R/(x) M/xM + Gdim R/(x) M/xM = depth R/(x) R/(x). Substituting, we see the formula holds.

19 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 19 Finally, assume depth R = n > 0 and depth M = 0. Then Gdim R M > 0 by Lemma 4.3. Let 0 K G M 0 be a short exact sequence where G is in G(R). Then Gdim R K = Gdim R M 1 (Proposition 3.9) and, since depth G = depth R > depth M = 0, depth K = 1 (part (8) of Proposition 2.7). By the depth M > 0 case applied to K, we have depth K + Gdim R K = depth R. Substituting, we again see the formula holds. As a corollary, we get the desired result for coherent rings: Corollary 4.5. Let R be a quasi-local coherent ring and M a finitely presented R-module of finite G-dimension. Then depth M + Gdim R M = depth R. Proof. As R is coherent and M is finitely presented, Gdim R M = Gdim R M by Corollary 3.7. We note that Theorem 4.4 also generalizes (and hence gives another proof of) [18, Theorem 2, Chapter 6]: Corollary 4.6. Let R be a quasi-local ring and M an (F P ) module of finite projective dimension. Then depth M + pd R M = depth R. Proof. By Proposition 3.8, pd R M = Gdim R M. 5. Coherent Gorenstein rings In [3], J. Bertin defines a quasi-local ring to be regular if every finitely generated ideal has finite projective dimension. A ring R is said to be regular if R m is regular for every maximal ideal m of R. It is clear that this definition of regular reduces to the usual definition for Noetherian local rings. In this spirit, we propose the following definition: Definition 5.1. A quasi-local ring R is called Gorenstein if every finitely generated ideal of R has finite G-dimension. In general, a ring R is called Gorenstein if R m is Gorenstein for every maximal ideal m of R.

20 20 LIVIA HUMMEL AND THOMAS MARLEY We remark that by [1, Theorem 4.20], this definition agrees with the usual notion of Gorenstein for Noetherian rings. While most of the results and examples in this section pertain to coherent Gorenstein rings, we point out that there exist non-coherent Gorenstein rings. We thank the referee for suggesting the following example. A valuation ring R is almost maximal if for every non-zero ideal I, R/I is linearly compact in the discrete topology. Example 5.2. Let V be an almost maximal valuation domain with value group R. (For instance, one could take V to be the ring of formal power series over an arbitrary field K with exponents in the nonnegative reals. See section II.6 of [9] for details.) Let P denote the maximal ideal of V and let a be a nonzero element of P. Then R = V/aP is a non-coherent Gorenstein ring. Proof. Since P has elements of arbitrarily small positive value, (0 : R a) = P/aP is not finitely generated. Thus, R is not coherent. Note that the ideals of V have the form xv or xp for x V. If xv ap, it is easily seen that (ap : V xv ) = x 1 ap. Similarly, if xp ap then (ap : V xp ) = x 1 av. It follows that for all ideals I of R one has (0 : R (0 : R I)) = I; i.e., every ideal of R is an annihilator ideal. We next show that R is injective as an R- module. To see this, let J be an arbitrary ideal of R generated by {r α α Λ} where Λ is an index set, and let ϕ : J R be an R-homomorphism. Since for any r R we have (0 : R (0 : R r)) = rr, it follows that for each α Λ there exists exists an s α R such that ϕ(r α ) = s α r α. Since every finitely generated ideal of R is principal, we get that for every finite subset Λ 0 of Λ there exists an s R such that s α r α = ϕ(r α ) = sr α for all α Λ 0. In other words, the set of cosets {r α + (0 : R r α ) α Λ} has the finite intersection property. By the linear compactness of R, there exists an s R such that ϕ(r α ) = sr α for all α Λ. Thus, R is injective. It now readily follows from the injectivity of R and the fact that every ideal is an annihilator ideal that every ideal is reflexive and every cyclic module is a member of G(R). By induction on the number of generators, one obtains that every finitely generated R-module has G-dimension zero. Hence, R is Gorenstein.

21 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 21 For the remainder of this section we restrict our attention to the properties of coherent Gorenstein rings. In particular, since every finitely generated projective module is a member of the restricted G-class, we have: Proposition 5.3. Let R be a coherent regular ring. Then R is Gorenstein. Thus, every valuation domain is Gorenstein, as are polynomial rings in an arbitrary number of variables over a field. We note that the property of being Gorenstein localizes: Proposition 5.4. Let R be a coherent Gorenstein ring and S a multiplicatively closed set. Then R S is Gorenstein. Proof. It is enough to show that if p is a prime ideal of R then R p is Gorenstein. Let I be a finitely generated ideal contained in p and m a maximal ideal containing p. As R is Gorenstein and coherent, we have by Corollary 3.13 and Corollary 3.7 Gdim Rp I p Gdim Rm I m = Gdim Rm I m <. Hence, R p is Gorenstein (and coherent). We also have: Proposition 5.5. Let R be a quasi-local coherent ring. The following are equivalent: (1) R is Gorenstein. (2) Every finitely presented R-module has finite G-dimension. Proof. An ideal I is finitely generated if and only if R/I is finitely presented. Hence, (2) implies (1) is clear. Suppose R is Gorenstein and let M be a finitely presented R-module generated by n elements. We use induction on n to prove Gdim R M <. This is clear if n 1. Suppose n > 1. Then there exists a submodule M of M generated by n 1 elements such that M/M is cyclic. As R is coherent and M is finitely presented, M is also finitely presented. Hence, GdimR M and Gdim R M/M are finite. By Proposition 3.10, Gdim R M <. In [13] a definition for an arbitrary commutative ring to be Cohen-Macaulay is given which agrees with the usual definition if the ring is Noetherian. We will show below that

22 22 LIVIA HUMMEL AND THOMAS MARLEY every coherent Gorenstein ring is Cohen-Macaulay. Let R be a ring and x denote a finite sequence x 1,..., x n of elements of R. For an R-module M, let Ȟi x(m) denote the ith Čech cohomology of M with respect to x and H i x(m) the ith local cohomology of M (i.e., the ith right derived functor of the (x)-torsion functor). The sequence x is called weakly proregular if for all i 0 and all R-modules M the natural map H i x(m) Ȟi x(m) is an isomorphism (cf. [19]). A sequence x of length n is called a parameter sequence if x is weakly proregular, (x)r R, and H n x(r) p 0 for all prime ideals containing (x). The sequence x is a strong parameter sequence if x 1,..., x i is a parameter sequence for all 1 i n. A ring R is called Cohen-Macaulay if every strong parameter sequence on R is a regular sequence. It is not known if this property localizes; thus, we say that R is locally Cohen-Macaulay if R p is Cohen-Macaulay for all prime ideals p of R. It is easily seen that locally Cohen-Macaulay rings are Cohen-Macaulay; see [13] for details. Proposition 5.6. Let R be a coherent Gorenstein ring. Then R is locally Cohen-Macaulay. Proof. Since R p is coherent Gorenstein for all primes p of R, it suffices to prove that a coherent quasi-local Gorenstein ring is Cohen-Macaulay. Let x = x 1,..., x n be a strong parameter sequence of R. We proceed by induction on n to show that x is a regular sequence. Let x denote the sequence x 1,..., x n 1. (In the case n = 1, x denotes the empty sequence, which generates the zero ideal.) By way of contradiction, we assume x is a regular sequence but x n is a zero-divisor on R/(x ). Then there exists a prime p of R such that x n p and depth R p /(x )R p = 0 [13, Lemma 2.8]. Hence, depth R p = n 1. By localizing at p, we can assume R is a quasi-local coherent Gorenstein ring with depth R = n 1 and x is a parameter sequence of length n. In particular, H n x(r) 0. As R is coherent Gorenstein and (x) t is finitely generated for all t, we have Gdim R R/(x) t < for all t. By Theorem 4.4, Gdim R R/(x) t depth R = n 1 for all t, and thus Ext n R (R/(x)t, R) = 0 for all t. But then H n x(r) = lim t Ext n R(R/(x) t, R) = 0, a contradiction. Hence, x is a regular sequence and R is Cohen-Macaulay. We next show that the Gorenstein property is preserved by passing to (and lifting from) a quotient modulo a non-zero-divisor. This certainly does not hold for regularity. We note also

23 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 23 that, in contrast to the Noetherian case, there are examples of quasi-local Cohen-Macaulay rings which do not remain Cohen-Macaulay modulo a non-zero-divisor. (See Example 4.9 of [13].) Theorem 5.7. Let R be a quasi-local coherent ring with maximal ideal m and x m a non-zero-divisor on R. Then R is Gorenstein if and only if R/(x) is Gorenstein. Proof. Assume R is Gorenstein. Let J be a finitely generated ideal of R/(x). Then J = I/(x) for some finitely generated ideal I of R containing (x). As R is (coherent) Gorenstein, Gdim R R/I <. By Theorem 3.19, Gdim R/(x) R/I <, which implies Gdim R/(x) J <. Thus, R/(x) is Gorenstein. Conversely, assume R/(x) is Gorenstein. Let M be a finitely presented R-module. Assume first that x is a non-zero-divisor on M. Then M/xM is a finitely presented R/(x)-module and by Proposition 5.5, GdimR/(x) M/xM <. By Proposition 3.17, Gdim R M <. If x is a zero-divisor on M let 0 K G M 0 be a short exact sequence where G is in G(R). As R is coherent, K is finitely presented. Furthermore, x is a non-zero-divisor on G and hence on K. Thus, Gdim R K <. Hence, Gdim R M <. Hence, for example, if V is a valuation domain and x V is a non-unit non-zero-divisor, then V/xV is Gorenstein. We now aim to prove that the Gorenstein property passes to finitely generated polynomial ring extensions, assuming coherence is preserved. To do this, we first need a folklore result which can be traced at least as far back as 1966 ([15]; see also [21]). The proof here is of a different style than the one found in the above references, although the argument is essentially the same. Lemma 5.8. Let R be a ring, x an indeterminate over R, and M an R[x]-module. There exists a short exact sequence of R[x]-modules 0 R[x] R M R[x] R M M 0. Proof. Let t be an indeterminate over R[x] and let t act on M via t m := xm for all m M. In this way one can consider M as an R[x, t]-module. Consider the short exact sequence of

24 24 LIVIA HUMMEL AND THOMAS MARLEY R[x, t]-modules 0 R[x, t] x t R[x, t] R[t] 0. Applying R[t] M, we get the short exact sequence of R[x, t]-modules 0 R[x, t] R[t] M R[x, t] R[t] M R[t] R[t] M 0. Now, R[x, t] R[t] M = (R[x] R R[t]) R[t] M = R[x] R M as R[x, t]-modules. Thus, we have a short exact sequence of R[x, t]-modules 0 R[x] R M R[x] R M M 0. Restricting scalars to R[x] we get the desired result. This leads to another change of rings result for restricted Gorenstein dimension: Corollary 5.9. Let R be a ring, x an indeterminate over R, and M an R[x]-module which is (F P ) R. Then Gdim R[x] M Gdim R M + 1. Proof. Clearly, we may assume Gdim R M <. Since R[x] is faithfully flat as an R-module, Gdim R M = Gdim R[x] R[x] R M by Proposition The result now follows from Lemma 5.8 and Proposition Theorem Let R[x] be a polynomial ring over R and assume R[x] is coherent. Then R is Gorenstein if and only if R[x] is Gorenstein. Proof. Suppose R[x] is Gorenstein. As x is a non-zero-divisor on R, R = R[x]/(x) is Gorenstein (and coherent) by Theorem 5.7. Conversely, suppose R is Gorenstein and let P be a maximal ideal of R[x]. By localizing at q = P R, we can assume R is quasilocal coherent Gorenstein (by Proposition 5.4) with maximal ideal m and P R = m. Now, P = (m, f)r[x] for some monic polynomial f R[x]. Then R[x]/fR[x] is a free R-module of finite rank. Let J be a finitely generated ideal of R[x] P. Then J = I P for some finitely generated (hence, finitely presented) ideal I of R[x]. Consequently, I/f I is a finitely presented R[x]/f R[x]-module, and hence finitely presented as an R-module as well. Thus, Gdim R I/fI <. By Corollary 5.9, Gdim R[x] I/fI <. Localizing, we have Gdim R[x]P J/fJ <. Since f P is a non-zero-divisor on both R[x] P and J, we have that Gdim R[x]P J < by Corollary Hence, R[x] P is Gorenstein.

25 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 25 It is known that the direct limit of a flat family of coherent regular rings is coherent regular (e.g., [10, Theorem 6.2.2]). The analogous result holds for Gorenstein rings: Proposition Let {R i } i Λ be a direct system of commutative rings over a directed index set Λ. Suppose each R i is a coherent Gorenstein ring and the maps R i R j are flat for all i j. Then lim R i is a coherent Gorenstein ring. Proof. Let S = lim R i and Q be a prime ideal of S. Since S Q = lim (R i )q i, where q i = Q R i, we may assume each R i is quasi-local. Thus, R i S is faithfully flat for all i. Let I be a finitely generated ideal of S. Then there exists a j Λ and a finitely generated ideal J of R j such that JS = I. Since R j is coherent Gorenstein and quasi-local, Gdim Rj R j /J <. Since S is faithfully flat over R j, by Corollary 3.13, Gdim S S/I <. Hence, S is Gorenstein. Note that the argument also shows that S is coherent. As an application, we have the following: Example Let S = k[x 1, x 2, x 3,... ]/(x 2 1, x2 2, x2 3,... ), where k is a field and the x i are variables. Then S = lim R i where for i N, R i = k[x 1, x 2,..., x i ]/(x 2 1, x2 2,..., x2 i ). Since each R i is a coherent Gorenstein ring and the maps R i R i+1 are flat for all i, S is coherent Gorenstein by Proposition It is well-known that a Noetherian local ring is Gorenstein if and only if the ring has finite injective dimension. This characterization can be generalized to quasi-local coherent rings using the notion of FP-injective dimension as defined by Stenström [20]: Definition Let M be an R-module. The FP-injective dimension of M is defined by FP-id R M := inf{n 0 Ext n+1 (N, M) = 0 for every finitely presented R-module N}, R where the infimum of the empty set is defined to be. If FP-id R M = 0 we say M is FP-injective. Proposition Let R be a coherent ring, M an R-module, and n a nonnegative integer. The following conditions are equivalent: (1) FP-id R M n.

26 26 LIVIA HUMMEL AND THOMAS MARLEY (2) Ext i R (N, M) = 0 for all i > n and all finitely presented R-modules N. (3) Ext n+1 R (R/I, M) = 0 for all finitely generated ideals I of R. (4) For every exact sequence 0 M E 0 E 1 E n 1 E n 0 such that E i is FP-injective for 0 i n 1, we have E n is FP-injective. Proof. See Lemma 3.1 of [20]. Combining this result with Proposition 2.9 we have: Corollary Let R be a coherent quasi-local ring with maximal ideal m and M an R-module such that mm M. Then depth M FP-id R M. We now give a characterization of quasi-local Gorenstein rings of finite depth in terms of FP-injective dimension. (For a proof of this result in the Noetherian case, see Theorem 4.20 of [1].) Theorem Let R be a quasi-local coherent ring, and n a nonnegative integer. The following conditions are equivalent: (1) FP-id R R = n. (2) R is Gorenstein and depth R = n. Proof. Suppose R is Gorenstein and depth R = n. Let M be a finitely presented R-module. By Theorem 4.4, Gdim R M n and hence Ext i R (M, R) = 0 for all i > n by Proposition 3.5. On the other hand, Ext n R (R/I, R) 0 for some finitely generated ideal I of R by Proposition 2.9. Therefore, FP-id R R = n. Conversely, suppose FP-id R R = n. By Corollary 5.15, depth R = m n. If R is Gorenstein then m = n as we have shown (2) implies (1). Let M be a finitely presented R-module and consider 0 K F n 1 F 1 F 0 M 0 where each F i is a finitely generated free R-module. Since Ext i R (M, R) = 0 for all i > n, it is easily seen that Ext i R (K, R) = 0 for all i > 0. It suffices to show that K is in G(R). In

27 THE AUSLANDER-BRIDGER FORMULA FOR COHERENT RINGS 27 fact, we will show that for any finitely presented R-module N such that Ext i R (N, R) = 0 for all i > 0, one has N G(R). Let 0 C G n 1 G n 2 G 1 G 0 N 0 be exact where G i is a finitely generated free R-module for all i. sequence By Lemma 3.18 the 0 N G 0 G 1 G n 1 C 0 is exact. As C is finitely presented (coherence) and FP-id R R = n, Ext i R (C, R) = 0 for i > n. Hence (by the same argument as for K above), Ext i R (N, R) = 0 for i > 0. Now consider the short exact sequence 0 L G 0 N 0. Then Ext i R (L, R) = 0 for all i > 0 and hence Ext i R (L, R) = 0 for i > 0 as well (using L in place of N). Hom R (, R) twice, one obtains the exactness of 0 L G 0 N 0. Applying Since the canonical map G 0 G 0 is an isomorphism, the canonical map N N is surjective. With L in place of N, we also obtain the map L L is surjective. Hence, the map N N is an isomorphism (by the snake lemma) and N is in G(R). Thus, R is Gorenstein. We remark that a version of Theorem 5.16 (without reference to depth) was proved in [7, Theorem 7] using very different methods. We also note that a polynomial ring over a field in infinitely many variables localized at a maximal ideal is a coherent quasi-local regular (hence Gorenstein) ring with infinite FP-injective dimension. Another equivalent characterization for a local ring R to be Gorenstein is that R be Cohen-Macaulay and some (equivalently, every) system of parameters generate an irreducible ideal. We have already seen that coherent Gorenstein rings are Cohen-Macaulay. Additionally, we have the following: Proposition Let R be a quasi-local coherent Gorenstein ring with depth R = n <. Then every n-generated ideal generated by a regular sequence is irreducible.

### THE AUSLANDER-BRIDGER FORMULA AND THE GORENSTEIN PROPERTY FOR COHERENT RINGS

University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln Faculty Publications, Department of Mathematics Mathematics, Department of 2009 THE AUSLANDER-BRIDGER FORMULA AND THE GORENSTEIN

### Cohen-Macaulay Dimension for Coherent Rings

University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln Dissertations, Theses, and Student Research Papers in Mathematics Mathematics, Department of 5-2016 Cohen-Macaulay Dimension

### COHEN-MACAULAY RINGS SELECTED EXERCISES. 1. Problem 1.1.9

COHEN-MACAULAY RINGS SELECTED EXERCISES KELLER VANDEBOGERT 1. Problem 1.1.9 Proceed by induction, and suppose x R is a U and N-regular element for the base case. Suppose now that xm = 0 for some m M. We

### Homological Methods in Commutative Algebra

Homological Methods in Commutative Algebra Olivier Haution Ludwig-Maximilians-Universität München Sommersemester 2017 1 Contents Chapter 1. Associated primes 3 1. Support of a module 3 2. Associated primes

### FROBENIUS AND HOMOLOGICAL DIMENSIONS OF COMPLEXES

FROBENIUS AND HOMOLOGICAL DIMENSIONS OF COMPLEXES TARAN FUNK AND THOMAS MARLEY Abstract. It is proved that a module M over a Noetherian local ring R of prime characteristic and positive dimension has finite

### Injective Modules and Matlis Duality

Appendix A Injective Modules and Matlis Duality Notes on 24 Hours of Local Cohomology William D. Taylor We take R to be a commutative ring, and will discuss the theory of injective R-modules. The following

### Homological Dimension

Homological Dimension David E V Rose April 17, 29 1 Introduction In this note, we explore the notion of homological dimension After introducing the basic concepts, our two main goals are to give a proof

### FILTERED RINGS AND MODULES. GRADINGS AND COMPLETIONS.

FILTERED RINGS AND MODULES. GRADINGS AND COMPLETIONS. Let A be a ring, for simplicity assumed commutative. A filtering, or filtration, of an A module M means a descending sequence of submodules M = M 0

### On the vanishing of Tor of the absolute integral closure

On the vanishing of Tor of the absolute integral closure Hans Schoutens Department of Mathematics NYC College of Technology City University of New York NY, NY 11201 (USA) Abstract Let R be an excellent

### Gorenstein homological dimensions

Journal of Pure and Applied Algebra 189 (24) 167 193 www.elsevier.com/locate/jpaa Gorenstein homological dimensions Henrik Holm Matematisk Afdeling, Universitetsparken 5, Copenhagen DK-21, Denmark Received

### THE RADIUS OF A SUBCATEGORY OF MODULES

THE RADIUS OF A SUBCATEGORY OF MODULES HAILONG DAO AND RYO TAKAHASHI Dedicated to Professor Craig Huneke on the occasion of his sixtieth birthday Abstract. We introduce a new invariant for subcategories

### ACYCLIC COMPLEXES OF FINITELY GENERATED FREE MODULES OVER LOCAL RINGS

ACYCLIC COMPLEXES OF FINITELY GENERATED FREE MODULES OVER LOCAL RINGS MERI T. HUGHES, DAVID A. JORGENSEN, AND LIANA M. ŞEGA Abstract We consider the question of how minimal acyclic complexes of finitely

### ALGEBRA HW 4. M 0 is an exact sequence of R-modules, then M is Noetherian if and only if M and M are.

ALGEBRA HW 4 CLAY SHONKWILER (a): Show that if 0 M f M g M 0 is an exact sequence of R-modules, then M is Noetherian if and only if M and M are. Proof. ( ) Suppose M is Noetherian. Then M injects into

### REMARKS ON REFLEXIVE MODULES, COVERS, AND ENVELOPES

REMARKS ON REFLEXIVE MODULES, COVERS, AND ENVELOPES RICHARD BELSHOFF Abstract. We present results on reflexive modules over Gorenstein rings which generalize results of Serre and Samuel on reflexive modules

### GENERALIZED MORPHIC RINGS AND THEIR APPLICATIONS. Haiyan Zhu and Nanqing Ding Department of Mathematics, Nanjing University, Nanjing, China

Communications in Algebra, 35: 2820 2837, 2007 Copyright Taylor & Francis Group, LLC ISSN: 0092-7872 print/1532-4125 online DOI: 10.1080/00927870701354017 GENERALIZED MORPHIC RINGS AND THEIR APPLICATIONS

### HOMOLOGICAL PROPERTIES OF MODULES OVER DING-CHEN RINGS

J. Korean Math. Soc. 49 (2012), No. 1, pp. 31 47 http://dx.doi.org/10.4134/jkms.2012.49.1.031 HOMOLOGICAL POPETIES OF MODULES OVE DING-CHEN INGS Gang Yang Abstract. The so-called Ding-Chen ring is an n-fc

### Math 210B. Artin Rees and completions

Math 210B. Artin Rees and completions 1. Definitions and an example Let A be a ring, I an ideal, and M an A-module. In class we defined the I-adic completion of M to be M = lim M/I n M. We will soon show

### The Depth Formula for Modules with Reducible Complexity

The Depth Formula for Modules with Reducible Complexity Petter Andreas Bergh David A Jorgensen Technical Report 2010-10 http://wwwutaedu/math/preprint/ THE DEPTH FORMULA FOR MODULES WITH REDUCIBLE COMPLEXITY

### GORENSTEIN DIMENSIONS OF UNBOUNDED COMPLEXES AND CHANGE OF BASE (WITH AN APPENDIX BY DRISS BENNIS)

GORENSTEIN DIMENSIONS OF UNBOUNDED COMPLEXES AND CHANGE OF BASE (WITH AN APPENDIX BY DRISS BENNIS) LARS WINTHER CHRISTENSEN, FATIH KÖKSAL, AND LI LIANG Abstract. For a commutative ring R and a faithfully

### VANISHING THEOREMS FOR COMPLETE INTERSECTIONS. Craig Huneke, David A. Jorgensen and Roger Wiegand. August 15, 2000

VANISHING THEOREMS FOR COMPLETE INTERSECTIONS Craig Huneke, David A. Jorgensen and Roger Wiegand August 15, 2000 The starting point for this work is Auslander s seminal paper [A]. The main focus of his

### Notes for Boot Camp II

Notes for Boot Camp II Mengyuan Zhang Last updated on September 7, 2016 1 The following are notes for Boot Camp II of the Commutative Algebra Student Seminar. No originality is claimed anywhere. The main

### REFLEXIVE MODULES OVER GORENSTEIN RINGS

REFLEXIVE MODULES OVER GORENSTEIN RINGS WOLMER V. VASCONCELOS1 Introduction. The aim of this paper is to show the relevance of a class of commutative noetherian rings to the study of reflexive modules.

### COURSE SUMMARY FOR MATH 508, WINTER QUARTER 2017: ADVANCED COMMUTATIVE ALGEBRA

COURSE SUMMARY FOR MATH 508, WINTER QUARTER 2017: ADVANCED COMMUTATIVE ALGEBRA JAROD ALPER WEEK 1, JAN 4, 6: DIMENSION Lecture 1: Introduction to dimension. Define Krull dimension of a ring A. Discuss

### THE FROBENIUS FUNCTOR AND INJECTIVE MODULES

THE FOBENIUS FUNCTO AND INJECTIVE MODULES THOMAS MALEY Abstract. We investigate commutative Noetherian rings of prime characteristic such that the Frobenius functor applied to any injective module is again

### Homological Aspects of the Dual Auslander Transpose II

Homological Aspects of the Dual Auslander Transpose II Xi Tang College of Science, Guilin University of Technology, Guilin 541004, Guangxi Province, P.R. China E-mail: tx5259@sina.com.cn Zhaoyong Huang

### DOUGLAS J. DAILEY AND THOMAS MARLEY

A CHANGE OF RINGS RESULT FOR MATLIS REFLEXIVITY DOUGLAS J. DAILEY AND THOMAS MARLEY Abstract. Let R be a commutative Noetherian ring and E the minimal injective cogenerator of the category of R-modules.

### THE ASSOCIATED PRIMES OF LOCAL COHOMOLOGY MODULES OVER RINGS OF SMALL DIMENSION. Thomas Marley

THE ASSOCATED PRMES OF LOCAL COHOMOLOGY MODULES OVER RNGS OF SMALL DMENSON Thomas Marley Abstract. Let R be a commutative Noetherian local ring of dimension d, an ideal of R, and M a finitely generated

### Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra

Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra D. R. Wilkins Contents 3 Topics in Commutative Algebra 2 3.1 Rings and Fields......................... 2 3.2 Ideals...............................

### arxiv:math/ v1 [math.ac] 4 Oct 2006

ON THE VNISHING OF (CO)HOMOLOGY OVER LOCL RINGS PETTER NDRES BERGH arxiv:math/0610150v1 [mathc] 4 Oct 2006 bstract Considering modules of finite complete intersection dimension over commutative Noetherian

### Ring Theory Problems. A σ

Ring Theory Problems 1. Given the commutative diagram α A σ B β A σ B show that α: ker σ ker σ and that β : coker σ coker σ. Here coker σ = B/σ(A). 2. Let K be a field, let V be an infinite dimensional

### TIGHT CLOSURE CONTENTS

TIGHT CLOSURE ZHAN JIANG CONTENTS 1. Preliminiary 1 1.1. Base change 1 1.2. Characteristic of a ring 2 1.3. Frobenius functor 2 2. Tight Closure 2 2.1. Definition 2 2.2. Basic properties 3 2.3. Behaviour

### Honors Algebra 4, MATH 371 Winter 2010 Assignment 4 Due Wednesday, February 17 at 08:35

Honors Algebra 4, MATH 371 Winter 2010 Assignment 4 Due Wednesday, February 17 at 08:35 1. Let R be a commutative ring with 1 0. (a) Prove that the nilradical of R is equal to the intersection of the prime

### Formal power series rings, inverse limits, and I-adic completions of rings

Formal power series rings, inverse limits, and I-adic completions of rings Formal semigroup rings and formal power series rings We next want to explore the notion of a (formal) power series ring in finitely

### Math 711: Lecture of September 7, Symbolic powers

Math 711: Lecture of September 7, 2007 Symbolic powers We want to make a number of comments about the behavior of symbolic powers of prime ideals in Noetherian rings, and to give at least one example of

### A NEW PROOF OF SERRE S HOMOLOGICAL CHARACTERIZATION OF REGULAR LOCAL RINGS

A NEW PROOF OF SERRE S HOMOLOGICAL CHARACTERIZATION OF REGULAR LOCAL RINGS RAVI JAGADEESAN AND AARON LANDESMAN Abstract. We give a new proof of Serre s result that a Noetherian local ring is regular if

### n-x -COHERENT RINGS Driss Bennis

International Electronic Journal of Algebra Volume 7 (2010) 128-139 n-x -COHERENT RINGS Driss Bennis Received: 24 September 2009; Revised: 31 December 2009 Communicated by A. Çiğdem Özcan Abstract. This

### Algebraic Geometry Spring 2009

MIT OpenCourseWare http://ocw.mit.edu 18.726 Algebraic Geometry Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.726: Algebraic Geometry

### 11. Finitely-generated modules

11. Finitely-generated modules 11.1 Free modules 11.2 Finitely-generated modules over domains 11.3 PIDs are UFDs 11.4 Structure theorem, again 11.5 Recovering the earlier structure theorem 11.6 Submodules

### EXT, TOR AND THE UCT

EXT, TOR AND THE UCT CHRIS KOTTKE Contents 1. Left/right exact functors 1 2. Projective resolutions 2 3. Two useful lemmas 3 4. Ext 6 5. Ext as a covariant derived functor 8 6. Universal Coefficient Theorem

### MATH 101B: ALGEBRA II PART A: HOMOLOGICAL ALGEBRA

MATH 101B: ALGEBRA II PART A: HOMOLOGICAL ALGEBRA These are notes for our first unit on the algebraic side of homological algebra. While this is the last topic (Chap XX) in the book, it makes sense to

### Semidualizing Modules. Sean Sather-Wagstaff

Semidualizing Modules Sean Sather-Wagstaff Department of Mathematics, North Dakota State University Department # 2750, PO Box 6050, Fargo, ND 58108-6050, USA E-mail address: sean.sather-wagstaff@ndsu.edu

### INJECTIVE MODULES: PREPARATORY MATERIAL FOR THE SNOWBIRD SUMMER SCHOOL ON COMMUTATIVE ALGEBRA

INJECTIVE MODULES: PREPARATORY MATERIAL FOR THE SNOWBIRD SUMMER SCHOOL ON COMMUTATIVE ALGEBRA These notes are intended to give the reader an idea what injective modules are, where they show up, and, to

### CRing Project, Chapter 16

Contents 16 Homological theory of local rings 3 1 Depth........................................... 3 1.1 Depth over local rings.............................. 3 1.2 Regular sequences................................

### Algebra Homework, Edition 2 9 September 2010

Algebra Homework, Edition 2 9 September 2010 Problem 6. (1) Let I and J be ideals of a commutative ring R with I + J = R. Prove that IJ = I J. (2) Let I, J, and K be ideals of a principal ideal domain.

### Integral Extensions. Chapter Integral Elements Definitions and Comments Lemma

Chapter 2 Integral Extensions 2.1 Integral Elements 2.1.1 Definitions and Comments Let R be a subring of the ring S, and let α S. We say that α is integral over R if α isarootofamonic polynomial with coefficients

### REFLEXIVITY AND RING HOMOMORPHISMS OF FINITE FLAT DIMENSION

Communications in Algebra, 35: 461 500, 2007 Copyright Taylor & Francis Group, C ISSN: 0092-7872 print/1532-4125 online DOI: 10.1080/00927870601052489 EFEXIVITY AND ING HOMOMOPHISMS OF FINITE FAT DIMENSION

### ALMOST GORENSTEIN RINGS TOWARDS A THEORY OF HIGHER DIMENSION

ALMOST GORENSTEIN RINGS TOWARDS A THEORY OF HIGHER DIMENSION SHIRO GOTO, RYO TAKAHASHI, AND NAOKI TANIGUCHI Abstract. The notion of almost Gorenstein local ring introduced by V. Barucci and R. Fröberg

### CONTRAVARIANTLY FINITE RESOLVING SUBCATEGORIES OVER A GORENSTEIN LOCAL RING

CONTRAVARIANTLY FINITE RESOLVING SUBCATEGORIES OVER A GORENSTEIN LOCAL RING RYO TAKAHASHI Introduction The notion of a contravariantly finite subcategory (of the category of finitely generated modules)

### Cohomology and Base Change

Cohomology and Base Change Let A and B be abelian categories and T : A B and additive functor. We say T is half-exact if whenever 0 M M M 0 is an exact sequence of A-modules, the sequence T (M ) T (M)

### Gorenstein Injective Modules

Georgia Southern University Digital Commons@Georgia Southern Electronic Theses & Dissertations Graduate Studies, Jack N. Averitt College of 2011 Gorenstein Injective Modules Emily McLean Georgia Southern

### STABLE HOMOLOGY OVER ASSOCIATIVE RINGS

STABLE HOMOLOGY OVER ASSOCIATIVE RINGS OLGUR CELIKBAS, LARS WINTHER CHRISTENSEN, LI LIANG, AND GREG PIEPMEYER Abstract. We analyze stable homology over associative rings and obtain results over Artin algebras

### ON sfp-injective AND sfp-flat MODULES

Gulf Journal of Mathematics Vol 5, Issue 3 (2017) 79-90 ON sfp-injective AND sfp-flat MODULES C. SELVARAJ 1 AND P. PRABAKARAN 2 Abstract. Let R be a ring. A left R-module M is said to be sfp-injective

### Structure of rings. Chapter Algebras

Chapter 5 Structure of rings 5.1 Algebras It is time to introduce the notion of an algebra over a commutative ring. So let R be a commutative ring. An R-algebra is a ring A (unital as always) together

### COFINITENESS AND COASSOCIATED PRIMES OF LOCAL COHOMOLOGY MODULES

MATH. SCAND. 105 (2009), 161 170 COFINITENESS AND COASSOCIATED PRIMES OF LOCAL COHOMOLOGY MODULES MOHARRAM AGHAPOURNAHR and LEIF MELKERSSON Abstract Let R be a noetherian ring, an ideal of R such that

### The most important result in this section is undoubtedly the following theorem.

28 COMMUTATIVE ALGEBRA 6.4. Examples of Noetherian rings. So far the only rings we can easily prove are Noetherian are principal ideal domains, like Z and k[x], or finite. Our goal now is to develop theorems

### Gorenstein modules, finite index, and finite Cohen Macaulay type

J. Pure and Appl. Algebra, (), 1 14 (2001) Gorenstein modules, finite index, and finite Cohen Macaulay type Graham J. Leuschke Department of Mathematics, University of Kansas, Lawrence, KS 66045 gleuschke@math.ukans.edu

### SEQUENCES FOR COMPLEXES II

SEQUENCES FOR COMPLEXES II LARS WINTHER CHRISTENSEN 1. Introduction and Notation This short paper elaborates on an example given in [4] to illustrate an application of sequences for complexes: Let R be

### Relative Left Derived Functors of Tensor Product Functors. Junfu Wang and Zhaoyong Huang

Relative Left Derived Functors of Tensor Product Functors Junfu Wang and Zhaoyong Huang Department of Mathematics, Nanjing University, Nanjing 210093, Jiangsu Province, China Abstract We introduce and

### A TALE OF TWO FUNCTORS. Marc Culler. 1. Hom and Tensor

A TALE OF TWO FUNCTORS Marc Culler 1. Hom and Tensor It was the best of times, it was the worst of times, it was the age of covariance, it was the age of contravariance, it was the epoch of homology, it

### Lecture 2. (1) Every P L A (M) has a maximal element, (2) Every ascending chain of submodules stabilizes (ACC).

Lecture 2 1. Noetherian and Artinian rings and modules Let A be a commutative ring with identity, A M a module, and φ : M N an A-linear map. Then ker φ = {m M : φ(m) = 0} is a submodule of M and im φ is

### Torq" (M, N) 0. In [1, Theorem 1.2] the formula. Krull dimension of R. We shall also give some sufficient conditions for an

MODULES OVER REGULAR LOCAL RINGS BY M. PAVAMAN MURTHY Introduction In [1] M. Auslander has proved the following: THEOREM. Let R be an unramified regular local ring, and M a torsion-free R-module of finite

### Extended Index. 89f depth (of a prime ideal) 121f Artin-Rees Lemma. 107f descending chain condition 74f Artinian module

Extended Index cokernel 19f for Atiyah and MacDonald's Introduction to Commutative Algebra colon operator 8f Key: comaximal ideals 7f - listings ending in f give the page where the term is defined commutative

### Notes on p-divisible Groups

Notes on p-divisible Groups March 24, 2006 This is a note for the talk in STAGE in MIT. The content is basically following the paper [T]. 1 Preliminaries and Notations Notation 1.1. Let R be a complete

### ON THE VANISHING OF HOMOLOGY WITH MODULES OF FINITE LENGTH

ON THE VANISHING OF HOMOLOGY WITH MODULES OF FINITE LENGTH PETTER ANDREAS BERGH Abstract We study the vanishing of homology and cohomology of a module of finite complete intersection dimension over a local

### (1) A frac = b : a, b A, b 0. We can define addition and multiplication of fractions as we normally would. a b + c d

The Algebraic Method 0.1. Integral Domains. Emmy Noether and others quickly realized that the classical algebraic number theory of Dedekind could be abstracted completely. In particular, rings of integers

### Infinite-Dimensional Triangularization

Infinite-Dimensional Triangularization Zachary Mesyan March 11, 2018 Abstract The goal of this paper is to generalize the theory of triangularizing matrices to linear transformations of an arbitrary vector

### Commutative Algebra. Timothy J. Ford

Commutative Algebra Timothy J. Ford DEPARTMENT OF MATHEMATICS, FLORIDA ATLANTIC UNIVERSITY, BOCA RA- TON, FL 33431 Email address: ford@fau.edu URL: http://math.fau.edu/ford Last modified January 9, 2018.

### Math 121 Homework 4: Notes on Selected Problems

Math 121 Homework 4: Notes on Selected Problems 11.2.9. If W is a subspace of the vector space V stable under the linear transformation (i.e., (W ) W ), show that induces linear transformations W on W

### Rings and groups. Ya. Sysak

Rings and groups. Ya. Sysak 1 Noetherian rings Let R be a ring. A (right) R -module M is called noetherian if it satisfies the maximum condition for its submodules. In other words, if M 1... M i M i+1...

### THE FROBENIUS STRUCTURE OF LOCAL COHOMOLOGY

THE FROBENIUS STRUCTURE OF LOCAL COHOMOLOGY by Florian Enescu and Melvin Hochster 1. INTRODUCTION All given rings in this paper are commutative, associative with identity, and Noetherian. Throughout, p

### On U-dominant dimension

Journal of Algebra 285 (2005) 669 68 www.elsevier.com/locate/jalgebra On U-dominant dimension Zhaoyong Huang Department of Mathematics, Nanjing University, Nanjing 20093, PR China Received 20 August 2003

### THE AUSLANDER BUCHSBAUM FORMULA. 0. Overview. This talk is about the Auslander-Buchsbaum formula:

THE AUSLANDER BUCHSBAUM FORMULA HANNO BECKER Abstract. This is the script for my talk about the Auslander-Buchsbaum formula [AB57, Theorem 3.7] at the Auslander Memorial Workshop, 15 th -18 th of November

### ALGEBRAIC GEOMETRY COURSE NOTES, LECTURE 2: HILBERT S NULLSTELLENSATZ.

ALGEBRAIC GEOMETRY COURSE NOTES, LECTURE 2: HILBERT S NULLSTELLENSATZ. ANDREW SALCH 1. Hilbert s Nullstellensatz. The last lecture left off with the claim that, if J k[x 1,..., x n ] is an ideal, then

### Lectures on Grothendieck Duality II: Derived Hom -Tensor adjointness. Local duality.

Lectures on Grothendieck Duality II: Derived Hom -Tensor adjointness. Local duality. Joseph Lipman Purdue University Department of Mathematics lipman@math.purdue.edu February 16, 2009 Joseph Lipman (Purdue

### Hilbert function, Betti numbers. Daniel Gromada

Hilbert function, Betti numbers 1 Daniel Gromada References 2 David Eisenbud: Commutative Algebra with a View Toward Algebraic Geometry 19, 110 David Eisenbud: The Geometry of Syzygies 1A, 1B My own notes

### 5 Dedekind extensions

18.785 Number theory I Fall 2016 Lecture #5 09/22/2016 5 Dedekind extensions In this lecture we prove that the integral closure of a Dedekind domain in a finite extension of its fraction field is also

### Rees Algebras of Modules

Rees Algebras of Modules ARON SIMIS Departamento de Matemática Universidade Federal de Pernambuco 50740-540 Recife, PE, Brazil e-mail: aron@dmat.ufpe.br BERND ULRICH Department of Mathematics Michigan

### 4.4 Noetherian Rings

4.4 Noetherian Rings Recall that a ring A is Noetherian if it satisfies the following three equivalent conditions: (1) Every nonempty set of ideals of A has a maximal element (the maximal condition); (2)

### TIGHT CLOSURE IN NON EQUIDIMENSIONAL RINGS ANURAG K. SINGH

TIGHT CLOSURE IN NON EQUIDIMENSIONAL RINGS ANURAG K. SINGH 1. Introduction Throughout our discussion, all rings are commutative, Noetherian and have an identity element. The notion of the tight closure

### LECTURE IV: PERFECT PRISMS AND PERFECTOID RINGS

LECTURE IV: PERFECT PRISMS AND PERFECTOID RINGS In this lecture, we study the commutative algebra properties of perfect prisms. These turn out to be equivalent to perfectoid rings, and most of the lecture

### arxiv:math/ v2 [math.ac] 25 Sep 2006

arxiv:math/0607315v2 [math.ac] 25 Sep 2006 ON THE NUMBER OF INDECOMPOSABLE TOTALLY REFLEXIVE MODULES RYO TAKAHASHI Abstract. In this note, it is proved that over a commutative noetherian henselian non-gorenstein

### COMMUNICATIONS IN ALGEBRA, 15(3), (1987) A NOTE ON PRIME IDEALS WHICH TEST INJECTIVITY. John A. Beachy and William D.

COMMUNICATIONS IN ALGEBRA, 15(3), 471 478 (1987) A NOTE ON PRIME IDEALS WHICH TEST INJECTIVITY John A. Beachy and William D. Weakley Department of Mathematical Sciences Northern Illinois University DeKalb,

### arxiv: v1 [math.ac] 4 Nov 2012

ON TEST MODULES FOR FLAT DIMENSION arxiv:1211.0734v1 [math.ac] 4 Nov 2012 JAVIER MAJADAS Abstract. A ring with a test module of finite upper complete intersection dimension is complete intersection. 1.

### CHARACTERIZING GORENSTEIN RINGS USING CONTRACTING ENDOMORPHISMS

CHARACTERIZING GORENSTEIN RINGS USING CONTRACTING ENDOMORPHISMS BRITTNEY FALAHOLA AND THOMAS MARLEY Dedicated to Craig Huneke on the occasion of his 65th birthday. Abstract. We prove several characterizations

### n P say, then (X A Y ) P

COMMUTATIVE ALGEBRA 35 7.2. The Picard group of a ring. Definition. A line bundle over a ring A is a finitely generated projective A-module such that the rank function Spec A N is constant with value 1.

LINKAGE CLASSES OF GRADE 3 PERFECT IDEALS LARS WINTHER CHRISTENSEN, OANA VELICHE, AND JERZY WEYMAN Abstract. While every grade 2 perfect ideal in a regular local ring is linked to a complete intersection

### Groups of Prime Power Order with Derived Subgroup of Prime Order

Journal of Algebra 219, 625 657 (1999) Article ID jabr.1998.7909, available online at http://www.idealibrary.com on Groups of Prime Power Order with Derived Subgroup of Prime Order Simon R. Blackburn*

### CHAPTER 1. AFFINE ALGEBRAIC VARIETIES

CHAPTER 1. AFFINE ALGEBRAIC VARIETIES During this first part of the course, we will establish a correspondence between various geometric notions and algebraic ones. Some references for this part of the

### Projective and Injective Modules

Projective and Injective Modules Push-outs and Pull-backs. Proposition. Let P be an R-module. The following conditions are equivalent: (1) P is projective. (2) Hom R (P, ) is an exact functor. (3) Every

### Gorenstein Homological Algebra of Artin Algebras. Xiao-Wu Chen

Gorenstein Homological Algebra of Artin Algebras Xiao-Wu Chen Department of Mathematics University of Science and Technology of China Hefei, 230026, People s Republic of China March 2010 Acknowledgements

### Relative FP-gr-injective and gr-flat modules

Relative FP-gr-injective and gr-flat modules Tiwei Zhao 1, Zenghui Gao 2, Zhaoyong Huang 1, 1 Department of Mathematics, Nanjing University, Nanjing 210093, Jiangsu Province, P.R. China 2 College of Applied

### On a Conjecture of Auslander and Reiten

Syracuse University SURFACE Mathematics Faculty Scholarship Mathematics 4-30-2003 On a Conjecture of Auslander and Reiten Craig Huneke University of Kansas Graham J. Leuschke Syracuse University Follow

### Reid 5.2. Describe the irreducible components of V (J) for J = (y 2 x 4, x 2 2x 3 x 2 y + 2xy + y 2 y) in k[x, y, z]. Here k is algebraically closed.

Reid 5.2. Describe the irreducible components of V (J) for J = (y 2 x 4, x 2 2x 3 x 2 y + 2xy + y 2 y) in k[x, y, z]. Here k is algebraically closed. Answer: Note that the first generator factors as (y

### Generalized tilting modules with finite injective dimension

Journal of Algebra 3 (2007) 69 634 www.elsevier.com/locate/jalgebra Generalized tilting modules with finite injective dimension Zhaoyong Huang Department of Mathematics, Nanjing University, Nanjing 20093,

### BENJAMIN LEVINE. 2. Principal Ideal Domains We will first investigate the properties of principal ideal domains and unique factorization domains.

FINITELY GENERATED MODULES OVER A PRINCIPAL IDEAL DOMAIN BENJAMIN LEVINE Abstract. We will explore classification theory concerning the structure theorem for finitely generated modules over a principal

### STABLE MODULE THEORY WITH KERNELS

Math. J. Okayama Univ. 43(21), 31 41 STABLE MODULE THEORY WITH KERNELS Kiriko KATO 1. Introduction Auslander and Bridger introduced the notion of projective stabilization mod R of a category of finite

### HOMOLOGICAL ALGEBRA MODULO EXACT ZERO-DIVISORS

HOMOLOGICAL ALGEBRA MODULO EXACT ZERO-DIVISORS PETTER ANDREAS BERGH, OLGUR CELIKBAS AND DAVID A. JORGENSEN Abstract. We study the homological behavior of modules over local rings modulo exact zero-divisors.