Let A/S be a ring extension with S commutative. We prove that ω⊗SAA is a generalized tilting module if ωS is a generalized tilting module. In this case, we obtain that ⊥ω-resol.dimS(M) and ⊥(ω⊗SA)-resol.dimA(M) are identical for any A-module M. As an application, we show that S satisfies gorenstein symmetric Conjecture if and only if so does A. Furthermore, we introduce the concept of ⊥ω-Gorenstein projective modules, and we obtain that the relative Gorenstein projectivity is invariant under Frobenius extensions.
Citation: Dongxing Fu, Xiaowei Xu, Zhibing Zhao. Generalized tilting modules and Frobenius extensions[J]. Electronic Research Archive, 2022, 30(9): 3337-3350. doi: 10.3934/era.2022169
[1] | Agustín Moreno Cañadas, Robinson-Julian Serna, Isaías David Marín Gaviria . Zavadskij modules over cluster-tilted algebras of type $ \mathbb{A} $. Electronic Research Archive, 2022, 30(9): 3435-3451. doi: 10.3934/era.2022175 |
[2] | Zhen Zhang, Shance Wang . Relative cluster tilting subcategories in an extriangulated category. Electronic Research Archive, 2023, 31(3): 1613-1624. doi: 10.3934/era.2023083 |
[3] | Hui Xu, Jun Kong, Mengyao Liang, Hui Sun, Miao Qi . Video behavior recognition based on actional-structural graph convolution and temporal extension module. Electronic Research Archive, 2022, 30(11): 4157-4177. doi: 10.3934/era.2022210 |
[4] | Dong Su, Shilin Yang . Representation rings of extensions of Hopf algebra of Kac-Paljutkin type. Electronic Research Archive, 2024, 32(9): 5201-5230. doi: 10.3934/era.2024240 |
[5] | Yaguo Guo, Shilin Yang . Projective class rings of the category of Yetter-Drinfeld modules over the $ 2 $-rank Taft algebra. Electronic Research Archive, 2023, 31(8): 5006-5024. doi: 10.3934/era.2023256 |
[6] | Agustín Moreno Cañadas, Isaías David Marín Gaviria, Pedro Fernando Fernández Espinosa . Brauer configuration algebras and Kronecker modules to categorify integer sequences. Electronic Research Archive, 2022, 30(2): 661-682. doi: 10.3934/era.2022035 |
[7] | Jiuda Huang, Chao Han, Wuju Wei, Chengjun Zhao . Analysis of long-term maintenance decision for asphalt pavement based on analytic hierarchy process and network level optimization decision. Electronic Research Archive, 2023, 31(9): 5894-5916. doi: 10.3934/era.2023299 |
[8] | Yajun Ma, Haiyu Liu, Yuxian Geng . A new method to construct model structures from left Frobenius pairs in extriangulated categories. Electronic Research Archive, 2022, 30(8): 2774-2787. doi: 10.3934/era.2022142 |
[9] | Jiange Liu, Yu Chen, Xin Dai, Li Cao, Qingwu Li . MFCEN: A lightweight multi-scale feature cooperative enhancement network for single-image super-resolution. Electronic Research Archive, 2024, 32(10): 5783-5803. doi: 10.3934/era.2024267 |
[10] | Zhaoyong Huang . On the $ C $-flatness and injectivity of character modules. Electronic Research Archive, 2022, 30(8): 2899-2910. doi: 10.3934/era.2022147 |
Let A/S be a ring extension with S commutative. We prove that ω⊗SAA is a generalized tilting module if ωS is a generalized tilting module. In this case, we obtain that ⊥ω-resol.dimS(M) and ⊥(ω⊗SA)-resol.dimA(M) are identical for any A-module M. As an application, we show that S satisfies gorenstein symmetric Conjecture if and only if so does A. Furthermore, we introduce the concept of ⊥ω-Gorenstein projective modules, and we obtain that the relative Gorenstein projectivity is invariant under Frobenius extensions.
Throughout this paper, all rings are associative with identity and all modules are unital right modules unless stated otherwise. For a ring S, we denote the category of all right S-modules (resp. finitely generated right S-modules) by Mod-S (resp. mod-S). We use pdS(M) (resp. idS(M)) to denote projective dimension (resp. injective dimension) of MS.
The generalized tilting modules were firstly introduced as a generalization of tilting modules by T. Wakamatsu in [1]. Sometimes, it is also called the Wakamatsu tilting module, see [2].
Definition 1.1. Let S be a ring. An S-module ωS∈mod-S is called a generalized tilting module (it is also called a Wakamatsu tilting module) if it is self-orthogonal, that is ExtiS(ωS,ωS)=0 for any i≥1, and there is an exact sequence
0→SSf0→T0f1→T1f2→⋯fi→Ti→⋯ |
such that: (1) Ti∈addωS for any i≥0, where addωS is the full subcategory of mod-S that consisting of all modules isomorphic to direct summands of finite direct sum of copies of ωS, and (2) after applying by HomS(−,ωS) the sequence is still exact.
About the generalized tilting module, there is a famous homological conjecture in the representation theory of Artin algebras, which is called the Wakamatsu tilting conjecture (WTC). This conjecture states that every generalized tilting module with finite projective dimension is tilting, or equivalently, every generalized tilting module with finite injective dimension is cotilting (see [3]). The homological conjecture is closely related to other homological conjectures. For example, the validity of finitistic dimension conjecture (FDC) implies the validity of WTC, and the validity of WTC implies the validity of the Gorenstein symmetric conjecture (GSC) and the Generalized Nakayama conjecture (GNC) (see [3,4]). Hence, the generalized tilting modules are studied widely, see [3,4,5,6].
The notion of Frobenius extensions was firstly introduced by Kasch in [7] as a generalization of Frobenius algebras. They play an important role in topological quantum field theories in dimension 2 and even 3 (see [8]) and in representation theory and knot theory (see [9,10,11]). Also, each Frobenius extension with base ring commutative provides us with a series of solutions to classical Yang-Baxter equation (see [10]). The fundamental example of Frobenius extensions is the group algebras induced by a finite index subgroup. There are other examples of Frobenius extensions include Hopf subalgebras, finite extensions of enveloping algebras of Lie super-algebras and finite extensions of enveloping algebras of Lie coloralgebras etc [12,13].
Separable extensions were firstly defined by Hirata and Sugano in [14] as a generalization of separable algebras, and they made a thorough study of these connection with Galois theory for noncommutative rings and generalizations of Azumaya algebras. If a ring extension is both separable extension and Frobenius extension, then it is called a separable Frobenius extension. Sugano proved that the central projective separable extensions are Frobenius extensions in [15]. More examples of separable Frobenius extensions can be found in Example Example 2.4. We refer to [10] for more details.
It is well-known that many homological properties are preserved under change of rings, especially excellent extension and Frobenius extension (see [16,17,18,19,20]). In this paper, we will consider some homological modules and homological dimension related a generalized tilting module under Frobenius (or separable Frobenius) extensions.
For a generalized tilting module ωS, we denote the left orthogonal class of ω by ⊥ωS={X∈Mod−S|ExtiS(XS,ωS)=0, for any i≥1}. The ⊥ω-resolution dimension of a module is defined as follows.
Definition 1.2. Let M be an S-module. The ⊥ω-resolution dimension of M, denoted by ⊥ω-resol.dimS(M), is defined as ⊥ω−resol.dimS(M)=inf{n∣∃⊥ωS−resolution 0→Tn→⋯T1→T0→M→0}. We set ⊥ω-resol.dimS(M)=∞ if no such integer exists.
For the homological dimension above, we have the following result.
Theorem A. Let A/S be a Frobenius extension with S commutative. For any A-module M, we have ⊥ω-resol.dimS(M)=⊥(ω⊗SA)-resol.dimA(M).
As an application, we get the following corollary.
Corollary B. Let S and A be both two-sided Noetherian rings and A/S be a Frobenius extension. Then S is a Gorenstein ring if and only if so is A. Furthermore, S satisfies GSC if and only if so does A.
Relating to a generalized tilting module ωS, ⊥ω-Gorenstein projective module is defined as a kind of relative Gorenstein homological module.
Definition 1.3. Let ω be a generalized tilting S-module. A right S-module G is called ⊥ω-Gorenstein projective if there exists an exact sequence
P:=⋯→P1→P0→P0→P1→⋯ |
of projective right S-modules which is still exact after applying HomS(−,X) for any module X∈⊥ωS and G=Ker(P0→P0). Furthermore, the exact sequence P is called an ⊥ω-complete projective resolution of G.
We denoted by ⊥ω-GP(S) the full subcategory of Mod-S consisting of all ⊥ω-Gorenstein projective modules.
For the relative Gorenstein projectivity, we prove it is preserved under Frobenius (or separable Frobenius) extensions.
Theorem C. Let A/S be a Frobenius extension with S commutative and ω a generalized tilting S-module. Then GS is ⊥ω-Gorenstein projective if and only if G⊗SAA is ⊥(ω⊗SA)-Gorenstein projective.
Theorem D. Let A/S be a Frobenius extension with S commutative, ω a generalized tilting S-module and G an A-module. If GA is ⊥(ω⊗SA)-Gorenstein projective, then GS is ⊥ω-Gorenstein projective. Furthermore, if the ring extension A/S is also separable, then GA is ⊥(ω⊗SA)-Gorenstein projective if and only if GS is ⊥ω-Gorenstein projective.
The paper is organized as follows. In Section 2, we give some notations and preliminaries. In Section 3, we will study the ⊥ω-resolution dimension under Frobenius extensions; the main result Theorem A is proved, see Theorem 3.5. As a corollary, we show that for a Frobenius extension of two-sided Noetherian rings, the base ring is Gorenstein if and only if so is the extension ring, see Corollary 3.12. In Section 4, ⊥ω-Gorenstein projectivity and ⊥ω-Gorenstein projective dimension are studied, the main results Theorem C and Theorem D are proved, see Theorems 4.4 and 4.5, respectively. Furthermore, we get some applications concerning some classical homological dimensions in this section.
A ring extension A/S is a ring homomorphism A ring extension is an algebra if S is commutative and l factors S→Z(A)↪A where Z(A) is the center of A. The natural bimodule structure of SAS is given by s⋅a⋅s′:=l(s)⋅a⋅l(s′). Similarly, we can get some other module structures, for example AS, SAA and AAS, etc.
For a ring extension A/S, there is a restriction functor R:Mod-A→Mod-S sending MA↦MS, given by m⋅s:=m⋅l(s). In the opposite direction, there are two natural functors as follows:
(1) T=−⊗SAA:Mod-S→Mod-A is given by MS↦M⊗SAA.
(2) H=HomS(AAS,−):Mod-S→Mod-A is given by MS↦HomS(AAS,MS).
It is easy to check that both (T,R) and (R,H) are adjoint pairs.
Definition 2.1. (see reference [10,Theorem 1.2]) A ring extension A/S is a Frobenius extension, provided that one of the following equivalent conditions holds:
(1) The functors T and H are naturally equivalent.
(2) SAA≅HomS(AAS,SSS) and AS is finitely generated projective.
(3) AAS≅HomSop(SAA,SSS) and SA is finitely generated projective.
(4) There exist E∈HomS−S(A,S), xi,yi∈A such that for ∀a∈A, one has ∑ixiE(yia)=a and ∑iE(axi)yi=a.
Definition 2.2. A ring extension A/S is a separable extension if and only if
μ:A⊗SA→A,a⊗b↦ab, |
is a split epimorphism of A-A-bimodules. If a ring extension A/S is both Frobenius extension and separable extension, then it is called a separable Frobenius extension.
Let A/S be a ring extension and M an A-module. Then MS is a right S-module. There is a natural surjective map π:M⊗SA→M given by m⊗a↦ma for any m∈M and a∈A. It is easy to check that π is split as an S-module homomorphism. However, π is not split as an A-homomorphism in general. The following lemma is a characterization of separable extensions.
Lemma 2.3. (see [21]) A ring extension A/S is separable if and only if for every module MA andA-homomorphism M→N, the natural epimorphism M⊗SAA→MA is a split A-epimorphism and natural with respect to M→N.
There are some other examples of Frobenius extensions or separable Frobenius extensions.
Example 2.4. (1) Let S be any ring and A=S[x]/(x2) be the quotient ring of S[x]. Then A/S is a Frobenius extension (see [22,Lemma 3.1]).
(2) Let A be a ring and n a positive integer. Then Mn(A) is a separable Frobenius extension of Sn(A), where Mn(A) is the full n×n matrix ring over A and Sn(A) is the centrosymmetric matrix ring over A, see [11,Theorem]
(3) Let S be a commutative algebra and A an Azumaya algebra over S. Then A/S is a separable Forbenius extension. See [10,Chapter 5] for more details.
(4) Every strongly separable extension is a separable Frobenius extension. Some examples of strongly separable extensions can be found in [9].
(5) Every excellent extension is a Frobenius extension. Furthermore, for an excellent extension, if the base ring is commutative, then it is also a separable extension. More examples of excellent extension can be found in [17,Example 2.2].
For an S-module W, we denote the left orthogonal class of W by ⊥W={X∈Mod−S|ExtiS(X,W)=0, for any i≥1}. Recall that a class of modules is called projectively resolving if it contains all projective modules and closed under extensions and kernel of epimorphisms. We have the following lemma.
Lemma 2.5. The class ⊥W is projectively resolving and closed under direct sums and summands.
Remark 2.6. The condition "(2)" in Definition 1.1 can be replaced by "Cokerfi∈⊥ωS for any i≥0" .
By Lemma 2.5, the following result is a direct consequence of [23,Lemma 3.12].
Proposition 2.7. Let M be an S-module and consider two exact sequences,
0→Kn→Tn−1→⋯→T0→M→0, |
0→K′n→T′n−1→⋯→T′0→M→0, |
where T0,⋯,Tn−1 and T′0,⋯,T′n−1 are in ⊥W. Then Kn is in ⊥W if and only if so is K′n.
In this section, ω is always a generalized tilting S-module. We prove that if A/S is a Frobenius extension with base ring S commutative, then ω⊗SAA is also a generalized tilting module over A and ⊥ω-resol.dimS(M) and ⊥(ω⊗SA)-resol.dimA(M) are identical for any A-module M.
The following two lemmas show that orthogonal classes with respect to a generalized tilting module are preserved under Frobenius extensions.
Lemma 3.1. Let A/S be a Frobenius extension of rings and X a right A-module. Then XA∈⊥(ω⊗SA)A if andonly if XS∈⊥ωS.
Proof. Since A/S is a Frobenius extension, we have T(MS)=M⊗SAA≅HomS(AAS,MS)=H(MS) for any MS∈Mod-S. By the adjoint isomorphism,
ExtiS(XS,ωS)≅ExtiS(X⊗AAS,ωS)≅ExtiA(XA,HomS(AAS,ωS))≅ExtiA(XA,ω⊗SAA) |
for any i≥1. Consequently, ExtiS(XS,ωS)=0 for any i≥1 if and only if ExtiA(XA,ω⊗SAA)=0 for any i≥1.
Lemma 3.2. Let A/S be a Frobenius extension with S commutative. For any S-module X, XS∈⊥ωS if andonly if X⊗SAA∈⊥(ω⊗SA)A.
Proof. By the adjoint isomorphism, for any i≥1,
ExtiA(X⊗SA,ω⊗SA)≅ExtiS(X,HomA(SA,ω⊗SA))≅ExtiS(X,ω⊗SA). |
Since A/S is a Frobenius extension, SA is a finitely generated projective S-module. And so ω⊗SAS∈addωS. If XS∈⊥ωS, then ExtiS(X,ω⊗SA)=0 for any i≥1. It follows that ExtiA(X⊗SA,ω⊗SA)=0 for any i≥1, that is, X⊗SAA∈⊥(ω⊗SA)A.
Conversely, if X⊗SAA∈⊥(ω⊗SA)A, then X⊗SAS∈⊥ωS by Lemma 3.1. Since XS is a direct summand of X⊗SAS, we know that XS is also in ⊥ωS.
Let C and D be two categories. Recall that a functor F:C→D is called a Frobenius functor if there exists a functor G:D→C such that both (F,G) and (G,F) are adjoint pairs (see [24]). By Definition 2.1, we know that the functors −⊗SAA≅HomS(AAS,−) induced by the Frobenius bimodule AAS are Frobenius functors.
Proposition 3.3. Let A/S be a Frobenius extension with S commutative. Then the Frobenius bimodule AAS induces a Frobenius functorfrom⊥ωS to ⊥(ω⊗SA)A.
Proof. By assumption, T(=−⊗SAA)≅H(=HomS(AAS,−)):Mod-S→Mod-A is a Frobenius functor with the restriction functor R as a left and right adjoint at same time.
By Lemmas 3.1 and 3.2, we get T∣⊥ωS⊆⊥(ω⊗SA)A and R∣⊥(ω⊗SA)A⊆⊥ωS, respectively. Hence T is a Frobenius functor from ⊥ωS to ⊥(ω⊗SA)A.
The following proposition shows that we can get generalized tilting modules over an extension ring from generalized tilting modules over a base ring when the ring extension is Frobenius.
Proposition 3.4. Let A/S be a Frobenius extension with S commutative. If ωS is a generalized tilting module, then ω⊗SAAis also a generalized tilting A-module.
Proof. Since ωS is finitely generated as an S-module, we know that ω⊗SAA is also finitely generated as an A-module. For any i≥1,
ExtiA(ω⊗SAA,ω⊗SAA)≅ExtiS(ωS,HomA(SAA,ω⊗SAA))≅ExtiS(ωS,ω⊗SAS). |
By assumption, SA is finitely generated projective. Then ω⊗SAS∈addωS. It follows that ExtiA(ω⊗SAA,ω⊗SAA)=0 for i≥1 from that ωS is self-orthogonal.
On the other hand, since ωS is a generalized tilting S-module, there exists an exact sequence
T:=0→SS→T0→T1→⋯→Ti→⋯ |
with Ti∈addωS for any i≥0 and HomS(T,ωS) is still exact. Applying by the functor −⊗SAA, we get the following sequence
T⊗SAA:=0→S⊗SAA≅AA→T0⊗SAA→T1⊗SAA→⋯→Ti⊗SAA→⋯ |
with each Ti⊗SAA∈ add(ω⊗SA)A, which is exact because SA is finitely generated projective. Since ω⊗SAS is in addωS, HomS(T,ω⊗SAS) is exact. Considering the following complex isomorphisms
HomA(T⊗SAA,ω⊗SAA)≅HomS(T,HomA(SAA,ω⊗SAA))≅HomS(T,ω⊗SAS), |
we know that HomA(T⊗SAA,ω⊗SAA) is also exact.
By Definition 1.1, ω⊗SAA is a generalized tilting A-module.
Let X be a subcategory of Mod-S and M an S-module. If there exists an exact sequence ⋯→Xn→⋯X1→X0→M→0 in Mod-S with Xi∈X for any i≥0, then we define the X-resolution dimension of M, denoted by X-resol.dimS(M), as X-resol.dimS(M) = inf{n∣∃ an exact sequence 0→Tn→⋯T1→T0→M→0 with each Ti∈X for 0≤i≤n }. We set X-resol.dimS(M)=∞ if no such integer exists.
Theorem 3.5. Let A/S be a Frobenius extension with S commutative. For any A-module M, we have ⊥ω−resol.dimS(M)=⊥(ω⊗SA)−resol.dimA(M).
Proof. Without loss of generality, we assume that ⊥(ω⊗SA)-resol.dimA(M)=n<∞. There is an exact sequence 0→Tn→⋯T1→T0→MA→0 with Ti∈⊥(ω⊗SA) in Mod-A for 0≤i≤n. Applying the restriction functor −⊗AAS, we get the following exact sequence 0→Tn→⋯T1→T0→MS→0 in Mod-S. By Lemma 3.1, Ti∈⊥ωS as S-modules for 0≤i≤n. Then ⊥ω-resol.dimS(M)≤n.
Conversely, we assume that ⊥ω-resol.dimS(M)=m<∞. As an A-module M, there is an exact sequence 0→Km→Gm−1→⋯→G0→MA→0 in Mod-A with Gi in ⊥(ω⊗SA)A for any 0≤i≤m−1. Applying by restriction functor, we get the following exact sequence 0→Km→Gm−1→⋯→G0→MS→0 in Mod-S with Gi in ⊥ωS for any 0≤i≤m−1 by Lemma 3.1. Since ⊥ω-resol.dimS(M)=m, Km is also in ⊥ωS as an S-module by Proposition 2.7. It follows from Lemma 3.1 that Km is in ⊥(ω⊗SA)A as an A-module. Hence ⊥(ω⊗SA)-resol.dimA(M)≤m=⊥ω-resol.dimS(M).
Similar to some classical homological dimensions, we have the following property of ⊥ω-resolution dimensions.
Proposition 3.6. If (Mi)i∈I is a family of S-modules, then
⊥ω-resol.dim(⨁i∈IMi)=sup{⊥ω-resol.dim(Mi)∣i∈I}. |
Proof. It is easy to see that ⊥ω-resol.dim(⊕i∈IMi)≤sup{⊥ω-resol.dim(Mi)∣i∈I} since ⊥ω is closed under direct sums.
For the converse inequality "≥", it suffices to show that ⊥ω-resol.dim(M′)≤⊥ω-resol.dim(M) for any direct summand M′ of M. Without loss of generality, we assume that ⊥ω-resol.dim(M)=n is finite. We use induction on n.
If n=0, M is in ⊥ω, then so is M′.
Now, we assume that n>0. We write that M=M′⊕M″ for some S-module M″. Taking exact sequences 0→K′→P′→M′→0 and 0→K″→P″→M″→0, where P′ and P″ are projective. By the Horseshoe Lemma, we get the following commutative diagram
![]() |
with exact columns and split exact rows. It follows from Proposition 2.7 that ⊥ω-resol.dim(K′⊕K″)=n−1. Hence the inductive hypothesis yields that ⊥ω-resol.dim(K′)≤n−1. Therefore, we have ⊥ω-resol.dim(M′)≤n by Proposition 2.7 again.
Theorem 3.7. Let A/S be a Frobenius extension with S commutative. For any S-module M, we have ⊥ω-resol.dimS(M)=⊥(ω⊗SA)-resol.dimA(M⊗SA).
Proof. Firstly, we claim that ⊥(ω⊗SA)-resol.dimA(M⊗SA)≤⊥ω-resol.dimS(M). We can assume that ⊥ω-resol.dimS(M)=n<∞.
By Definition 1.2, there is an exact sequence 0→Tn→⋯T1→T0→MS→0 with Ti∈⊥ωS in Mod-S for 0≤i≤n. By Lemma 3.2, Ti⊗SAA∈⊥(ω⊗SA)A for 0≤i≤n. Applying by the functor −⊗SAA, we get the following sequence
0→Tn⊗SAA→⋯→T1⊗SAA→T0⊗SAA→MS⊗SAA→0 |
in Mod-A, which is also exact because SA is finitely generated projective. Then ⊥(ω⊗SA)-resol.dimA(M⊗SA)≤n=⊥ω-resol.dimS(M).
Conversely, there is a natural surjective map π:M⊗SA→M given by π(m⊗a)=ma for any m∈M and a∈A which is split as an S-module homomorphism. Therefore, we have that MS is a direct summand of M⊗SAS, and ⊥ω-resol.dimS(M)≤⊥ω-resol.dimS(M⊗SA) by Proposition 3.6. It follows from Theorem 3.5 that ⊥ω-resol.dimS(M)≤⊥ω-resol.dimS(M⊗SA)≤⊥(ω⊗SA)-resol.dimA(M⊗SA).
Therefore, we get ⊥ω-resol.dimS(M)=⊥(ω⊗SA)-resol.dimA(M⊗SA).
We define r.Global⊥ω-resol.dim(S)=sup{⊥ω-resol.dim(M)∣M is any right S-module }, and call it right global ⊥ω-resolution dimension of S.
Proposition 3.8. Let A/S be a Frobenius extension with S commutative. Then r.Global⊥ω-resol.dim(S)=r.Global⊥(ω⊗SA)-resol.dim(A).
Proof. It follows from Theorem 3.5 that r.Global⊥(ω⊗SA)-resol.dim(A)≤r.Global⊥ω-resol.dim(S). And Theorem 3.7 shows that r.Global⊥ω-resol.dim(S)≤r.Global⊥(ω⊗SA)-resol.dim(A).
Lemma 3.9. (see [4,Propositon 3.1]) Let ωS be a generalized tilting S-module and n a non-negative integer. Then r.idS(ω)≤n if and only if ⊥ω-resol.dimS(M)≤n for any module M in mod-S.
Theorem 3.10. Let A/S be a Frobenius extension with S commutative. Then r.idS(ω)=r.idA(ω⊗SAA).
Proof. Using the fact that the functors R, T and H preserve the finiteness of modules, we get the assertion by Proposition 3.8 and Lemma 3.9.
Corollary 3.11. Let A/S be a Frobenius extension of rings. Then r.idS(S)=r.idA(A).
Proof. Put ωS=SS, and the corollary follows from the Theorem 3.10.
Recall that a two sided Noetherian ring S is called a Gorenstein ring if l.idS(S) and r.idS(S) are finite. A famous homological conjecture is called Gorenstein symmetric conjecture (GSC), which states that the left injective and right injective dimensions of a two sided Noetherian ring are identical. It is well-known that l.idS(S) and r.idS(S) are identical provided that both of them are finite (see [25]). By the corollary above, we have the following valuable corollary.
Corollary 3.12. Let S and A be two-sided Noetherian rings and A/S be a Frobenius extension. Then S is a Gorenstein ring if and only if so is A.Furthermore, S satisfies GSC if and only if so does A.
Proof. By the "symmetry" of Frobenius extension and Corollary 3.11 , we get l.idS(S)=l.idA(A). Thus, r.idS(S)=r.idA(A) and l.idS(S)=l.idA(A) if A/S is a Frobenius extension.
Example 3.13. (1) Let A be a two-sided Noetherian ring and Mn(A) the n×n matrix ring over A. Then Mn(A) is an excellent extension of A. It followsthat A is Gorenstein if and only if so is Mn(A) from Corollary 3.12.
(2) Let A be an Artin ring and Q=A[x]/(x2) is the quotient of the polynomialring, where x is a variable which is supposed to commute with all the elements of A. Then A is Gorenstein if and only if so is Q.
(3) Let A be a central separable Artin algebra over center C. Then A is a strong separable extension of C. By Corollary 3.12, A is Gorensteinif and only if so is C.
In this section, we will consider the ⊥ω-Gorenstein projectivity and ⊥ω-Gorenstein projective dimension under Frobenius extension, where ωS is a generalized tilting S-module. Furthermore, some corollaries related to classical homological dimensions are obtained.
For the ⊥ω-Gorenstein projective modules (see Definition 1.3), we have the following facts.
Remark 4.1. (1) The ⊥ω-Gorenstein projective module is a special case of X-Gorenstein projective module for X=⊥ω. (See [26,Definition 2.1]).
(2) Every projective module is ⊥ω-Gorenstein projective. And every ⊥ω-Gorenstein projective module is Gorenstein projective.
(3) The class ⊥ω-GP(S) is projectively resolving and closed under direct summands and direct sums.(See [26,Proposition 2.6]).
The following assertion is a direct consequence of [23,Lemma 3.12].
Proposition 4.2. Let M be an S-module. Suppose that
0→Kn→Gn−1→⋯→G0→M→0 |
and
0→K′n→G′n−1→⋯→G′0→M→0 |
are two exact sequences, where G0,⋯,Gn−1 and G′0,⋯,G′n−1 are in ⊥ω-GP(S). Then Kn is in ⊥ω-GP(S) if and only if so isK′n.
It is easy to get the following equivalent condition of ⊥ω-Gorenstein projectivity by Definition 1.3.
Proposition 4.3. (see [26,Propostion 2.4]) Let G be a right S-module. Then the followings areequivalent.
(1) G is ⊥ω-Gorenstein projective.
(2) i) ExtiS(G,X)=0 for any X∈⊥ω and i>0;
ii) There exists an exact sequence Q:=0→G→P0→P1→⋯ in Mod−S with Pi projective forevery i≥0such that Hom(Q,X) is still exact for any X∈⊥ω.
(3) There exists a short exact sequence of S-modules 0→G→P→G′→0, where P is projective and G′ is⊥ω-Gorenstein projective.
The following results show that ⊥ω-Gorenstein projectivity is preserved under Frobenius extensions.
Theorem 4.4. Let A/S be a Frobenius extension with S commutative. For an S-module G, G is ⊥ω-Gorensteinprojective if and only if G⊗SAA is ⊥(ω⊗SA)-Gorenstein projective.
Proof.(⇒): It suffices to show G⊗SAA satisfying the condition (2) in the Proposition 4.3 when GS is an ⊥ω-Gorenstein projective module. For any X∈⊥(ω⊗SA)A, it follows from Lemma 3.1 that XS∈⊥ωS. Then, for any i≥1,
ExtiA(G⊗SAA,XA)≅ExtiS(GS,HomA(SAA,XA))≅ExtiS(GS,XS)=0. |
Since GS is ⊥ω-Gorenstein projective, there is an exact sequence Q:=0→GS→P0→P1→⋯ in Mod-S with Pi projective for any i≥0 and HomS(Q,X) is still exact for any X∈⊥ωS. By the assumption, SA is finitely generated projective, and we get the following sequence
Q⊗SAA:=0→G⊗SAA→P0⊗SAA→P1⊗SAA→⋯ |
is still exact with Pi⊗SAA projective in Mod-A for any i≥0. For any XA∈⊥(ω⊗SA)A, then XS∈⊥ωS. Thus
HomA(Q⊗SAA,XA)≅HomS(Q,HomA(SAA,XA))≅HomS(Q,XS) |
is exact. Therefore, G⊗SAA is ⊥(ω⊗SA)-Gorenstein projective by Proposition 4.3.
(⇐): We claim that G⊗SAS is ⊥ω-Gorenstein projective when G⊗SAA is ⊥(ω⊗SA)A-Gorenstein projective. For any YS∈⊥ωS, Y⊗SAA∈⊥(ω⊗SA)A by Lemma 3.2. Then, for any i≥1,
ExtiS(G⊗SAS,YS)≅ExtiS(G⊗SA⊗AAS,YS)≅ExtiA(G⊗SAA,HomS(AAS,YS))≅ExtiA(G⊗SAA,Y⊗SAA)=0 |
because G⊗SAA is ⊥(ω⊗SA)A-Gorenstein projective.
By assumption, there is an exact sequence P:=0→G⊗SAA→P0→P1→⋯ in Mod-A with Pi projective for any i≥0 and HomA(P,X) is still exact for any X∈⊥(ω⊗SA)A. After applying the restriction functor R=−⊗AAS, we get the following exact sequence P:=0→G⊗SAS→P0→P1→⋯ with Pi projective in Mod-S for any i≥0. And, for any YS∈⊥ω, the complex
HomS(PS,YS)≅HomS(P⊗AAS,YS)≅HomA(P,HomS(AAS,YS))≅HomA(P,Y⊗SAA) |
is exact because Y⊗SAA∈⊥(ω⊗SA)A. Thus G⊗SAS is ⊥ω-Gorenstein projective.
It is well-known that GS is a direct summand of G⊗SAS and the class ⊥ωS-GP(S) is closed under direct summands. Therefore, GS is ⊥ω-Gorenstein projective.
Theorem 4.5. Let A/S be a Frobenius extension with S commutative. For any A-module G, if GA is⊥(ω⊗SA)-Gorenstein projective, then GS is ⊥ω-Gorenstein projective.Furthermore, if the ring extension A/S is also separable, then GA is ⊥(ω⊗SA)-Gorenstein projective if and only if GS is ⊥ω-Gorensteinprojective.
Proof. The first assertion follows from the proof of sufficiency of Theorem 4.4.
If the ring extension A/S is also separable, then GA is a direct summand of G⊗SAA by Lemma 2.3. By Theorem 4.4, G⊗SAA is ⊥(ω⊗SA)A-Gorenstein projective if GS is ⊥ωS-Gorenstein projective. And we have that GA is ⊥(ω⊗SA)A-Gorenstein projective since the class ⊥(ω⊗SA)A-GP(A) is closed under direct summands.
By Theorems 4.4 and 4.5, similar to the proof of Proposition 3.3, we have
Corollary 4.6. Let A/S be a Frobenius extension with S commutative. Then the Frobenius bimodule AAS induces a Frobenius functor from⊥ω-GP(S) to ⊥(ω⊗SA)-GP(A).
Similar to the classical homological dimensions, we define the ⊥ω-Gorenstein projective dimension of modules and the global ⊥ω-Gorenstein projective dimension of rings as follows.
Definition 4.7. Let M be an S-module. The ⊥ω-Gorenstein projective dimension of M, denoted by ⊥ω-GpdS(M), is defined as ⊥ω-GpdS(M)=inf{n∣∃⊥ω-Gorenstein projective resolution 0→Gn→⋯G1→G0→M→0}. We set ⊥ω-GpdS(M)=∞ if no such integer exists.
We define r.⊥ω-Ggldim(S)=sup{⊥ω-GpdS(M)∣M is any right S-module }, and call it right global ⊥ω-Gorenstein projective dimension of S.
Proposition 4.8. Let A/S be a Frobenius extension with S commutative. For each S-module M, we have ⊥ω-GpdS(M)=⊥(ω⊗SA)-GpdA(M⊗SA).
Proof. The proof is similar to that of Theorem 3.7, for the sake of completeness, we give the proof as follows.
Assume that ⊥ω-GpdS(M)=n<∞, there is an exact sequence 0→Gn→Gn−1→⋯→G0→M→0 in Mod-S with Gi being ⊥ω-Gorenstein projective for 0≤i≤n. Applying by the functor T=−⊗SAA, we get the following exact sequence
0→Gn⊗SAA→Gn−1⊗SAA→⋯→G0⊗SAA→M⊗SAA→0 |
in Mod-A with Gi⊗SAA being ⊥(ω⊗SA)-Gorenstein projective for 0≤i≤n by Theorem 4.4. Then ⊥(ω⊗SA)-GpdA(M⊗SA)≤n=⊥ω-GpdS(M).
Conversely, we can assume that ⊥(ω⊗SA)-GpdA(M⊗SA)=m<∞. As an S-module M, there is an exact sequence 0→Km→Gm−1→⋯→G0→M→0 in Mod-S with Gi projective for 0≤i≤m−1. Since SA is Projective as an S-module, applying the functor T=−⊗SAA, we obtain the following exact sequence
0→Km⊗SAA→Gm−1⊗SAA→⋯→G0⊗SAA→M⊗SAA→0, |
where Gi⊗SAA is ⊥(ω⊗SA)-Gorenstein projective by Theorem 4.4 for 0≤i≤m−1. Thus Km⊗SAA is also ⊥(ω⊗SA)-Gorenstein projective by Proposition 4.2. Again by Theorem 4.4, Km is an ⊥ω-Gorenstein projective S-module. Thus ⊥ω-GpdS(M)≤m=⊥(ω⊗SA)-GpdA(M⊗SA).
Therefore, ⊥ω-GpdS(M)=⊥(ω⊗SA)-GpdA(M⊗SA).
Corollary 4.9. Let S be a commutative Artin ring and A/S a Frobenius extension. For each S-module MS, we have pdS(M)=pdA(M⊗SA).
Proof. Since S is a commutative Artin ring, there exists some generalized tilting module ωS with idS(ω)=0 (in fact, an injective cogenerator is such generalized tilting module). For any MA∈Mod-A and i≥1,
ExtiA(MA,ω⊗SAA)≅ExtiA(MA,HomS(AAS,ωS))≅ExtiS(M⊗AAS,ωS)≅ExtiA(MS,ωS)=0 |
Hence ω⊗SAA is also a generalized tilting A-module with idA(ω⊗SAA)=0. In this case, the ⊥ω-Gorenstein projective S-module is same to projecitive S-module and the ⊥ω-Gorenstein projective dimension coincides with the classical projective dimension. The assertion follows from Proposition 4.8.
Proposition 4.10. Let A/S be a Frobenius extension with S commutative and ωS a generalized tilting S-module and M a right A-module. Then ⊥ω-GpdS(M)≤⊥(ω⊗SA)-GpdA(M).Furthermore, if the ring extension A/S is also separable, then ⊥ω-GpdS(M)=⊥(ω⊗SA)-GpdA(M).
Proof. It is trivial for the case of ⊥(ω⊗SA)-GpdA(M)=∞. We assume that ⊥(ω⊗SA)-GpdA(M)=n<∞, there is an exact sequence 0→Gn→Gn−1→⋯→G0→MA→0 in Mod-A with Gi being ⊥(ω⊗SA)A-Gorenstein projective for 0≤i≤n. Applying the restriction functor R=−⊗AAS, we have an exact sequence 0→Gn→Gn−1→⋯→G0→MS→0 in Mod-S with Gi being ⊥ω-Gorenstein projective for 0≤i≤n by Theorem 4.5. Therefore, ⊥ω-GpdS(M)≤⊥ω⊗SA-GpdA(M).
Conversely, we can assume that ⊥ω-GpdS(M)=m. There is an exact sequence 0→Gm→Gm−1→⋯→G0→M→0 in Mod-S with Gi being ⊥ωS-Gorenstein projective for 0≤i≤n. By Theorem 4.4, the following sequence
0→Gm⊗SAA→Gm−1⊗SAA→⋯→G0⊗SAA→M⊗SAA→0, |
in Mod-A is exact with Gi⊗SAA being ⊥(ω⊗SA)-Gorenstein projective. And so ⊥(ω⊗SA)-GpdA(M⊗SA)≤m=⊥ω-GpdS(M).
If the ring extension A/S is separable, then MA is a direct summand of M⊗SAA. It follows from [26,Proposition 3.4] that ⊥ω⊗SA-GpdA(M)≤⊥(ω⊗SA)-GpdA(M⊗SA)≤m.
The following result maybe is well-known. In fact, we have known that : for a Frobenius extension A/S and an A-module M, if pdA(M)<∞, then one has pdA(M)=pdS(M), see [27,Theorem 8].
Corollary 4.11. Let S be a commutative Artin ring and A/S be a Frobenius extension. For each right A-module M, pdS(M)≤pdA(M). Furthermore, if the ring extension A/S is also separable, then pdS(M)=pdA(M).
Proof. The proof is similar to that of the Corollary 4.9.
Corollary 4.12. Let A/S be a Frobenius extension with S commutative and ωS a generalized tilting S-module.Then r.⊥ω-Ggldim(S)≤r.⊥(ω⊗SA)-Ggldim(A). Furthermore, if the ring extension A/S is also separable, then r.⊥ω-Ggldim(S)=r.⊥ω⊗SA-Ggldim(A).
Proof. The first assertion follows from Proposition 4.8. Furthermore, if A/S is separable, MA is a direct summand of M⊗SAA for any M∈Mod-A by Lemma 2.3. And the second assertion follows from Proposition 4.10.
Corollary 4.13. Let S be a commutative Artin ring and A/S a Frobenius extension. Then gldim(S)≤gldim(A).Furthermore, if the ring extension A/S is also separable, then gldim(S)=gldim(A).
Proof. It follows from Corollarys 4.9 and 4.11.
The authors thank Professor Xiao-Wu Chen for his helpful suggestions. Xiao-Wei Xu was partially supported by NSFC (Grant No.11771176, 11971289). Zhi-Bing Zhao was supported by NFSC (No. 11871071) and Project of University Natural Science Research of Anhui Province (No. KJ2019A0007). The authors thank the referees for their helpful comments and suggestions.
This work does not have any conflicts of interest.
[1] |
T. Wakamatsu, Tilting modules and Auslander's Gorenstein property, J. Algebra, 275 (2004), 3–39. https://doi.org/10.1016/j.jalgebra.2003.12.008 doi: 10.1016/j.jalgebra.2003.12.008
![]() |
[2] |
F. Mantese, I. Reiten, Wakamatsu tilting modules, J. Algebra, 278 (2004), 532–552. https://doi.org/10.1016/j.jalgebra.2004.03.023 doi: 10.1016/j.jalgebra.2004.03.023
![]() |
[3] | A. Beligiannis, I. Reiten, Homological and Homotopical Aspects of Torsion Theories, Memoirs of the American Mathematical Society, 2007. https://doi.org/10.1090/memo/0883 |
[4] |
Z. Y. Huang, Generalized tilting modules with finite injective dimension, J. Algebra, 31 (2007), 619–634. https://doi.org/10.1016/j.jalgebra.2006.11.025 doi: 10.1016/j.jalgebra.2006.11.025
![]() |
[5] | Z. Y. Huang, Wakamatsu tilting modules, U-dominant dimension and k-Gorenstein modules, in Abelian Groups, Rings, Modules, and Homological Algebra, Chapman Hall/CRC, (2006), 183–202. |
[6] | Z. B. Zhao, X. N. Du, Generalized tilting modules with finite injective dimension, Adv. Math. (China), 43 (2014), 844–850. |
[7] |
F. Kasch, Grundlagen einer theorie der Frobenius-Erweiterungen, Math. Ann., 127 (1954), 453–474. https://doi.org/10.1007/BF01361137 doi: 10.1007/BF01361137
![]() |
[8] | J. Kock, Frobenius Algebras and 2D Topological Quantum Field Theories, Cambridge University Press, 2004. https://doi.org/10.1017/CBO9780511615443 |
[9] |
L. Kadison, The jones polynomial and certain separable Frobenius extensions, J. Algebra, 186 (1996), 461–475. https://doi.org/10.1006/jabr.1996.0383 doi: 10.1006/jabr.1996.0383
![]() |
[10] | L. Kadison, New Example of Frobenius Extension, University Lecture Series, 1999. https://doi.org/10.1090/ulect/014 |
[11] |
C. C. Xi, S. J. Yin, Cellularity of centrosymmetric martrix algebras and Frobenius extensions, Linear Algebra Appl., 590 (2020), 317–329. https://doi.org/10.1016/j.laa.2020.01.002 doi: 10.1016/j.laa.2020.01.002
![]() |
[12] |
D. Fischman, S. Montgomery, H. J. Schneider, Frobenius extensions of subalgebras of Hopf algebras, Trans. Amer. Math. Soc., 349 (1997), 4857–4895. https://doi.org/10.1090/S0002-9947-97-01814-X doi: 10.1090/S0002-9947-97-01814-X
![]() |
[13] |
H. J. Schneider, Normal basis and transitivity of crossed products for Hopf algebras, J. Algebra, 151 (1992), 289–312. https://doi.org/10.1016/0021-8693(92)90034-J doi: 10.1016/0021-8693(92)90034-J
![]() |
[14] |
K. Hirata, K. Sugano, On semisimple extensions and separable extensions over noncommutative rings, J. Math. Soc. Japan., 18 (1966), 360–373. https://doi.org/10.2969/jmsj/01840360 doi: 10.2969/jmsj/01840360
![]() |
[15] | K. Sugano, Separable extensions and Frobenius extensions, Osaka J. Math., 7, (1970), 291–299. |
[16] |
X. W. Chen, Totally reflexive extensions and modules, J. Algebra, 379 (2013), 322–332. https://doi.org/10.1016/j.jalgebra.2013.01.014 doi: 10.1016/j.jalgebra.2013.01.014
![]() |
[17] |
Z. Y. Huang, J. X. Sun, Invariant properties of represenations under excellent extensions, J. Algebra, 358 (2012), 87–101. https://doi.org/10.1016/j.jalgebra.2012.03.004 doi: 10.1016/j.jalgebra.2012.03.004
![]() |
[18] |
W. Ren, Gorenstein projective and injective dimensions over Frobenius extensions, Commun. Algebra, 46 (2018), 1–7. https://doi.org/10.1080/00927872.2017.1355714 doi: 10.1080/00927872.2017.1355714
![]() |
[19] |
C. C. Xi, Frobenius bimodules and flat-dominant dimensions, Sci. China Math., 64 (2020), 33–44. https://doi.org/10.1007/s11425-018-9519-2 doi: 10.1007/s11425-018-9519-2
![]() |
[20] |
Z. B. Zhao, Gorenstein homological invariant properties under Frobenius extensions, Sci. China Math., 62 (2019), 2487–2496. https://doi.org/10.1007/s11425-018-9432-2 doi: 10.1007/s11425-018-9432-2
![]() |
[21] | R. S. Pierce, Projective modules over Artinian algebras, in Graduate Texts in Mathematics, Springer, (1982), 88–107. https://doi.org/10.1007/978-1-4757-0163-0_6 |
[22] |
W. Ren, Gorenstein projective mdoules and Frobenius extensions, Sci. China Math., 61 (2018), 1175–1186. https://doi.org/10.1007/s11425-017-9138-y doi: 10.1007/s11425-017-9138-y
![]() |
[23] | M. Auslander, M. Bridger, Stable Module Theory, Memoirs of the American Mathematical Society, 1969. https://doi.org/10.1090/memo/0094 |
[24] | K. Morita, Adojint pairs of functors and Frobenius extension, Sci. Rep. T. K. D. Sect. A., 9 (1965), 40–71. |
[25] |
T. Wakamatsu, On modules with trivial self-extensions, J. Algebra, 114 (1988), 106–114. https://doi.org/10.1016/0021-8693(88)90215-3 doi: 10.1016/0021-8693(88)90215-3
![]() |
[26] | J. Wang, X. W. Xu, Z. B. Zhao, X-Gorenstein projective dimensions, preprint, arXiv: 1801.09127. |
[27] |
T. Nakamaya, T. Tsuzuku, On frobenius extension I, Nagoya. Math. J., (1960), 89–110. https://doi.org/10.1017/S0027763000002075 doi: 10.1017/S0027763000002075
![]() |