Home The role of w-tilting modules in relative Gorenstein (co)homology
Article Open Access

The role of w-tilting modules in relative Gorenstein (co)homology

  • Driss Bennis , Enrique Duarte , Juan R. García Rozas EMAIL logo and Luis Oyonarte
Published/Copyright: December 9, 2021

Abstract

Let R be a ring, C be a left R -module and S = End R ( C ) . When C is semidualizing, the Auslander class A C ( S ) and the Bass class C ( R ) associated with C have been the subject of extensive investigations. It has been proved that these classes, also known as Foxby classes, are one of the central concepts of (relative) Gorenstein homological algebra. In this paper, we answer several natural questions which arise when we weaken the condition of C being semidualizing: if we let C be w-tilting (see Definition 2.1), we establish the conditions for the pair ( A C ( S ) , A C ( S ) 1 ) to be a perfect cotorsion theory and for the pair ( B C 1 ( R ) , B C ( R ) ) to be a complete hereditary cotorsion theory. This tells us when the classes of Auslander and Bass are preenveloping and precovering, which generalizes a number of results disseminated in the literature. We investigate Gorenstein flat modules relative to a not necessarily semidualizing module C and we find conditions for the class of G C -projective modules to be special precovering, the class of G C -flat modules to be covering, the one of Gorenstein C -projective modules to be precovering and that of Gorenstein C -injective modules to be preenveloping. We also find how to recover Foxby classes from Add R ( C ) -resolutions of R .

MSC 2010: 16E30; 18G25

1 Introduction

Foxby classes (Auslander and Bass classes) have proven to be very useful when studying Gorenstein injective and projective dimensions: over Gorenstein rings, every module has finite Gorenstein projective and finite Gorenstein injective dimensions [1]. However, when the ring is no longer Gorenstein, modules of finite Gorenstein injective or projective dimensions constitute two new full subcategories. These were identified in the case of a commutative local Cohen-Macaulay ring with a dualizing module, at least for finitely generated modules. Foxby proved that the Auslander and Bass classes behave in a duality and that finitely generated modules in the Auslander class are precisely those of finite Gorenstein projective dimension, while finitely generated modules in the Bass class are those of finite Gorenstein injective dimension [2].

This result has been much appreciated and later on many authors have studied and extended the situation given by Foxby (see [3,4,5] or [6] for instance).

On the other hand, the existence of Gorenstein projective precovers and Gorenstein injective preenvelopes is an active subject of study nowadays, not only in the category of modules but also in any category where the knowledge of homological properties is interesting, as for example the categories of sheaves over a topological space, the ones of complexes of modules, those of representations of quivers by modules, etc. (read for instance Enochs et al. in [7], Enochs et al. in [8], Gillespie in [9] or Krause in [10]). At the same time, generalizations of these concepts by means of a semidualizing or even a weakly Wakamatsu tilting (cotilting), module (see Definition 2.1) have been given ( G C -projective, G C -injective, Gorenstein C -projective and Gorenstein C -injective modules), with the corresponding study of the dimensions relative to them [11,12, 13,14]. However, little is yet known about the existence of precovers or preenvelopes relative to these new classes of modules: it is known that if C is weakly Wakamatsu tilting (cotilting), then every module of finite G C -projective (injective) dimension has a special G C -precover (preenvelope) [11, Corollary 3.6 and Theorem 4.11]. It is then clear at this point how important it would be to know when the existence of G C -projective, C G P ( R ) and even G C -flat precovers, and G C -injective and C G I ( R ) preenvelopes, can be guaranteed, or, from another point of view, when all modules have finite G C -projective (injective) dimension.

In this paper, we will address that problem, and we will do so by a systematic use of the duality between the Auslander and Bass classes. We will study when the pairs of classes ( A C ( S ) , A C ( S ) 1 ) and ( C 1 ( R ) , C ( R ) ) are complete cotorsion theories (in fact, ( A C ( S ) , A C ( S ) 1 ) is not only complete but also perfect) in Theorem 3.16 and Corollary 3.12 (two of the main results of the paper), respectively.

Moreover, we will see (Theorem 3.10 and Corollary 4.8) that, with the extra assumption that C is self-small, the classes A C ( S ) and C ( R ) can both be defined as the right orthogonal classes of some given modules. This is a big step in understanding their structure.

And finally we will be able to prove the last two main results of the paper: Corollary 5.7, which says that, under certain circumstances, the class of all G C -flat modules is covering in R -Mod, and Corollary 5.14, which states that G C P ( R ) is special precovering in R -Mod.

The paper is organized as follows. Section 2 is devoted to setting the conditions assumed in all this work, to introducing the definitions of all the concepts related to this problem, and to stating and, sometimes proving, the basic results needed throughout the paper.

In Section 3, we study the problems addressed above regarding the pairs of classes ( A C ( S ) , A C ( S ) 1 ) and ( C 1 ( R ) , C ( R ) ) . We give conditions for the first pair to be a perfect cotorsion theory (Theorem 3.16) and conditions for the second pair to be a complete cotorsion theory (Corollary 3.12). In particular, we find conditions for A C ( S ) and C ( R ) to be precovering and preenveloping. We also study when the classes Add R ( C ) and Prod S ( Hom R ( C , E ) ) ( R E is an injective cogenerator) are covering (Corollary 3.5) and enveloping (Corollary 3.4), respectively, with precovers (preenvelopes) of modules in C ( R ) ( A C ( S ) ) being epimorphisms (monomorphisms) as can be checked in Propositions 3.8 and 3.9, respectively. This fact will lead us to prove that the Bass class C ( R ) can be recovered from a right Add R ( C ) -resolution of R , as a direct sum of C and the right orthogonal class of all cokernels in such resolution (Theorem 3.10). The dual of this result relative to the Auslander class will need a little more work and will be proved in Section 4.

In Section 4, we study when the class C G P ( R ) of Gorenstein C -projective modules and the class C G I ( R ) of Gorenstein C -injective modules are precovering (Proposition 4.7) and preenveloping (Corollary 4.4), respectively. Using this, we get, as applications, how to recover the Auslander class from a right Add R ( C ) -resolution of R on one side (Corollary 4.8), and on the other, what is the common part of the Bass class and its left 1-orthogonal class, and the common part of the Auslander class and its first right orthogonal class (Proposition 4.9).

Finally, in Section 5 we study the existence of G C -projective precovers. We find the conditions for the class of G C -flat modules to be covering (Corollary 5.7) and to get the inclusion of the class of G C -projective modules in the one of G C -flat modules (Theorem 5.9). From this fact we deduce the existence of G C -projective precovers (indeed special G C -projective precovers) in Corollary 5.14.

2 Preliminaries

Throughout this paper R will be an associative (not necessarily commutative) ring with identity, and all modules will, unless otherwise specified, be unital left R -modules. When it is necessary to use right R -modules, they will be denoted by M R , while in these cases left R -modules will be denoted by R M . Following this terminology, R M S will mean that M is an ( R , S ) -bimodule. The category of all left (right if necessary) R -modules will be denoted by R -Mod (Mod- R ). For an R -module C we use Add R ( C ) to denote the class of all R -modules which are isomorphic to direct summands of direct sums of copies of C , and Prod R ( C ) will denote the class of all R -modules which are isomorphic to direct summands of direct products of copies of C . We will denote by Gen R ( C ) the class of left R -modules which are generated by C , that is, modules isomorphic to quotients of modules in Add R ( C ) and by Cogen R ( C ) the class of R -modules which are cogenerated by C , that is, they are isomorphic to submodules of modules in Prod R ( C ) .

By σ R [ C ] we mean the full subcategory of R -Mod consisting of all modules subgenerated by C , that is, modules isomorphic to submodules of quotients of direct sums of arbitrary copies of C . Dually, π R [ C ] will be the class of left R -modules which are isomorphic to quotients of submodules of direct products of copies of C (see [15]). We then see that Gen R ( C ) σ R [ C ] π R [ C ] and that Cogen R ( C ) π R [ C ] .

From now on, and unless otherwise specified, C will be a left R -module and S its endomorphism ring ( S = End R ( C ) ).

The character module of any module M will be represented by M + , and P ( R ) , ( R ) and ( R ) will be the classes of all projective, injective and flat R -modules, respectively. As usual, the projective and injective dimensions of M will be denoted as p d ( M ) and i d ( M ) , respectively.

If A is a class of modules, a left A -resolution of a left R -module M is a Hom ( A , ) -exact complex (so not necessarily exact) with components in A , that is, a complex

A 1 A 0 M 0

such that A i A i and such that the resulting complex after applying Hom R ( A , ) for any A A is exact. Right A -resolutions (or A -coresolutions) of R M are defined dually.

Given two classes of modules C and D , a complete ( C , D ) -resolution of M is an exact, Hom ( C , ) -exact and Hom ( , D ) -exact sequence of modules

C 1 C 0 D 0 D 1

with C i C , D i D and M = Im ( C 0 D 0 ) . A complete C -resolution will mean a complete ( C , C ) -resolution.

A degreewise finite projective resolution is a projective resolution (that is, an exact left P ( R ) -resolution) in which all the components are finitely generated.

For any class of left R -modules C we use the following standard notation:

C 1 = { M R -Mod ; Ext R 1 ( C , M ) = 0 , C C } , C = { M R -Mod ; Ext R i ( C , M ) = 0 , C C , i 1 } .

Similarly, 1 C and C are defined.

Recall that if is any class of modules, an -precover of a module M is a morphism φ : F M with F in the class , such that for any other module F in the class , the map Hom ( F , φ ) : Hom ( F , F ) Hom ( F , M ) is surjective. The morphism φ is said to be a special -precover provided that φ is surjective and ker ( φ ) holds in the class 1 . And φ is an -cover if Hom ( F , φ ) ( f ) = φ f is an automorphism. Preenvelopes, special preenvelopes and envelopes are defined dually.

A class of modules is said to be precovering if every module has an -precover. Analogously, special precovering, covering, preenveloping, special preenveloping and enveloping classes are defined.

A cotorsion theory is defined as a pair of classes ( C , D ) such that C 1 = D and 1 D = C . The cotorsion theory is as follows:

  • Cogenerated by a set if there is a set X such that X 1 = D .

  • Complete if any left R -module has a special D -preenvelope (or equivalently, any left R -module has a special C -precover).

  • Perfect if any left R -module has a C -cover and a D -envelope.

  • Hereditary if C 1 = C [16, Definition 1.2.10].

Associated with every bimodule R C S there are two classes: the Auslander ( A C ( S ) ) and the Bass ( C ( R ) ) classes (also called Foxby classes), defined as follows:

A C ( S ) is the class of all left S -modules M satisfying:

  1. Tor 1 S ( C , M ) = 0 .

  2. Ext R 1 ( C , C S M ) = 0 .

  3. The canonical map μ M : M Hom R ( C , C S M ) is an isomorphism of S -modules.

C ( R ) consists of all left R -modules N satisfying:

  1. Ext R 1 ( C , N ) = 0 .

  2. Tor 1 S ( C , Hom R ( C , N ) ) = 0 .

  3. The canonical map ν N : C S Hom R ( C , N ) N is an isomorphism of R -modules.

So we see that the functors Hom R ( C , ) and C S establish an equivalence between the subcategories A C ( S ) and C ( R ) .

Of course the classes A C ( R ) and C ( S ) in Mod- R and in Mod- S , respectively, can also be considered and they are in a situation of equivalence as well:

A C ( R ) is the class of all right R -modules M satisfying:

Tor 1 R ( M , C ) = 0 , Ext S 1 ( C , M R C ) = 0 and the canonical map

μ M : M Hom S ( C , M R C )

is an isomorphism of R -modules.

C ( S ) consists of all right S -modules N satisfying:

Ext S 1 ( C , N ) = 0 , Tor 1 R ( Hom S ( C , N ) , C ) = 0 and the canonical map

ν N : Hom S ( C , N ) R C N

is an isomorphism of S -modules.

We now recall some concepts that will be needed throughout the paper.

Definition 2.1

[11] A left R -module C is said to be weakly Wakamatsu tilting (w-tilting for short) if it has the following two properties:

  1. Ext R 1 ( C , C ( I ) ) = 0 for every set I (that is, C ( I ) Add R ( C ) for every set I ).

  2. R has an exact right Add R ( C ) -resolution 0 R C 0 C 1 C 2

If C satisfies (1) but perhaps not (2), then C will be said to be Σ -self-orthogonal. If Ext R 1 ( C , C ) = 0 , then C is self-orthogonal.

Dually, a left R -module U is said to be weakly cotilting (w-cotilting for short) if it satisfies the following two conditions:

  1. Ext R 1 ( U X , U ) = 0 for every set X .

  2. There is an injective cogenerator D in R -Mod with an exact left Prod R ( U ) -resolution  A 1 A 0 D 0 .

If U satisfies (1) but possibly not (2), then U will be said to be -self-orthogonal.

We will say that a bimodule R C S is a weakly Wakamatsu (co)tilting bimodule if R C and C S are weakly Wakamatsu (co)tilting modules.

Definition 2.2

[3, Definition 2.1] Let C be a class of left R -modules and M be any left R -module. M is said to be Hom -faithful relative to C if:

C C and Hom R ( M , C ) = 0 C = 0 .

When M is Hom -faithful relative to R -Mod then we just say that M is Hom -faithful, and M is self- Hom -faithful if it is Hom -faithful relative to σ [ M ] .

On the other hand, M is said to be Hom -cofaithful relative to C if:

C C and Hom R ( C , M ) = 0 C = 0 .

Again, M is Hom -cofaithful if M is Hom -cofaithful relative to R -Mod and M is self- Hom -cofaithful if it is Hom -cofaithful relative to π [ M ] .

A right S -module N is said to be S -faithful relative to a class D of left S -modules if:

D D and N S D = 0 D = 0 .

N is S -faithful if N is S -faithful relative to S -Mod.

A module M is self-small if Hom R ( M , M ( I ) ) Hom R ( M , M ) ( I ) for every set I .

In what follows, E will be a fixed injective cogenerator of R -Mod and C = Hom R ( C , E ) .

Proposition 2.3

For any bimodule R C S we have that A C ( S ) Cogen S ( C ) and that C ( R ) Gen R ( C ) .

Proof

Take any M A C ( S ) . Since E is an injective cogenerator in R -Mod, we know that there is a monomorphism 0 C S M E I for some index set I . Apply Hom R ( C , ) to get a monomorphism of left S -modules

0 Hom R ( C , C S M ) Hom R ( C , E I ) ( C ) I .

But we know that M Hom R ( C , C S M ) by the properties of A C ( S ) so we get M Cogen S ( C ) .

On the other hand, suppose that N holds in C ( R ) , take a free S -presentation of Hom R ( C , N ) , S ( I ) Hom R ( C , N ) 0 , and apply C S . We get an epimorphism

C ( I ) C S S ( I ) C S Hom R ( C , N ) 0 ,

and since N C S Hom R ( C , N ) we have that N Gen R ( C ) .□

3 Cotorsion theories via Foxby classes

In this section, we study the properties of the different cotorsion theories arising from the Auslander and Bass classes. Namely, we will see under what conditions the pairs of classes ( A C ( S ) , A C ( S ) 1 ) and ( C 1 ( R ) , C ( R ) ) are complete cotorsion theories (Theorem 3.16 and Corollary 3.12, respectively).

Moreover, we will prove that when C is w-tilting and self-small, the Bass class C ( R ) can be seen as the right orthogonal class of certain modules constructed from a right Add R ( C ) -resolution of R (Theorem 3.10).

We start by showing that the classes Add R ( C ) and Prod R ( C ) can be thought of, under our standard conditions on the module C , as other classes from which several interesting properties can be easily deduced. This identification will be very useful throughout the paper.

It is known (see [17]) that when R is commutative and C is semidualizing, the classes Add R ( C ) and C S P ( S ) coincide. We start by extending this result to the noncommutative case and with significantly lower restrictions on C .

Proposition 3.1

If R C is self-small, then Add R ( C ) = C S P ( S ) .

Proof

The class C S P ( S ) clearly lies inside Add R ( C ) .

Let then N be any module of Add R ( C ) . We know there is a module T such that N T = C ( I ) , so

Hom R ( C , N ) Hom R ( C , T ) Hom R ( C , C ( I ) ) Hom R ( C , C ) ( I ) = S ( I ) .

This implies both that Hom R ( C , N ) P ( S ) and that

( C S Hom R ( C , N ) ) ( C S Hom R ( C , T ) ) C ( I ) ,

so we have two commutative diagrams and which show that N C S Hom R ( C , N ) .□

As for the classes Hom R ( C , ( R ) ) and Prod S ( C ) we have the following.

Proposition 3.2

Let S be any ring, C be a self-orthogonal right S -module admitting a degreewise finite projective resolution and R = End S ( C ) . Then, Prod S ( C ) = Hom R ( C , ( R ) ) .

Proof

The inclusion Hom R ( C , ( R ) ) Prod S ( C ) clearly holds without any assumption on C .

On the other hand, C S admitting a degreewise finite projective resolution makes A C ( S ) closed under direct products, and by [3, Theorem 2.5] we have ( R ) C ( R ) . Hence, C A C ( S ) and so Prod S ( C ) A C ( S ) .

If we now let L Prod S ( C ) , then L K = ( C ) J Hom R ( C , E J ) for some K S -Mod, and then we have that

( C S L ) ( C S K ) C S Hom R ( C , E J ) E J

( ( R ) C ( R ) ). Hence, Hom R ( C , C S L ) Hom R ( C , C S K ) Hom R ( C , E J ) and therefore L Hom R ( C , ( R ) ) since L Hom R ( C , C S L ) .□

There exist in the literature a number of results providing information about when the image of a (pre)cover or a (pre)envelope by means of a functor in an adjoint situation is again a (pre)cover or a (pre)envelope. And surely there are solutions to the question of whether or not the “image” of a (pre)covering or (pre)enveloping class is (pre)covering or (pre)enveloping. However, we have not found them and so we give below two results that may shed some light on this problem and that will be useful later in the paper.

Theorem 3.3

Let C and D be two abelian categories and F : C D be a left adjoint functor of G : D C . Then, the following statements hold:

  1. If Y is a (pre)enveloping subcategory of D , then G ( Y ) is a (pre)enveloping subcategory of C .

  2. If X is a (pre)covering subcategory of C , then F ( X ) is a (pre)covering subcategory of D .

Proof

We only prove 1. since 2. is just a dualization.

Assume that Y is preenveloping. Given any object A of C let φ : F ( A ) Y be a Y -preenvelope. Let us prove that if μ A : A G F ( A ) is the unit of the adjunction in A , then G ( φ ) μ A : A G ( Y ) is a G ( Y ) -preenvelope.

Take any morphism f : A G ( Y ) for any object Y of Y , call ν : F G 1 C the co-unit of ( F , G ) and let η : Hom C ( , G ( ) ) Hom D ( F ( ) , ) be the natural equivalence provided by the adjunction. We have the situation where η A , Y 1 ( φ ) = G ( φ ) μ A by the properties of the unit of the adjunction ( F , G ) .

Now, using once more the properties of the unit and the co-unit of the adjunction we have

ν Y F ( η A , Y 1 ( φ ) ) = ν Y F ( G ( φ ) μ A ) = η A , Y ( G ( φ ) μ A ) = η A , Y ( η A , Y 1 ( φ ) ) = φ

and since φ is a Y -preenvelope, there is ω : Y Y with ω φ = ν Y F ( f ) . Applying now G we obtain G ( ω φ ) = G ( ω ) G ( φ ) = G ( ν Y F ( f ) ) . But again, using the properties of μ A and ν Y we get

G ( ν Y F ( f ) ) μ A = η A , Y 1 ( ν Y F ( f ) ) = η A , Y 1 ( η A , Y ( f ) ) = f

so we see that G ( ω ) : G Y G Y is such that G ( ω ) G ( φ ) μ A = f , that is, G ( φ ) μ A is a G ( Y ) -preenvelope.

If Y is now enveloping and φ : F ( A ) Y is a Y -envelope, we already know that G ( φ ) μ A : A G Y is a G ( Y ) -preenvelope, so the diagram can be completed commutatively by some α . Then we have α G ( φ ) μ A = G ( φ ) μ A = η A , Y 1 ( φ ) .

On the other hand, since η A , Y ( α G ( φ ) μ A ) : F ( A ) Y and φ is a Y -preenvelope, there exists β : Y Y such that β φ = η A , Y ( α G ( φ ) μ A ) . Therefore, we have

β φ = η A , Y ( α G ( φ ) μ A ) = η A , Y η A , Y 1 ( φ ) = φ .

But φ is a Y -envelope so the latter implies that β must be an automorphism and then G ( β ) = α must be an automorphism too.□

We know (see [18]) that the classes Add R ( C ) and Prod S ( C ) are always precovering and preenveloping respectively. Now, we can immediately find conditions for Prod S ( C ) to be enveloping and for Add R ( C ) to be covering.

Corollary 3.4

Let S be any ring, C be a self-orthogonal right S -module admitting a degreewise finite projective resolution and R = End S ( C ) . Then, Prod S ( C ) is enveloping.

Proof

By Proposition 3.2, we know that Prod S ( C ) = Hom R ( C , ( R ) ) , so apply Theorem 3.3 to the class ( S ) with the adjunction ( C S , Hom R ( C , ) ) .□

Following the same type of arguments we also get the following.

Corollary 3.5

Assume that C is Σ -self-orthogonal as a left R -module and that S is left perfect. Then, Add R ( C ) is covering in R -Mod.

We now turn to study conditions which guarantee when a module is w-(co)tilting.

Recall that given a subcategory C of R -Mod, a module Q in C is said to be Ext -projective in C if Q C . The module Q is a generator of C if for any M C there is an exact sequence 0 M Q M 0 with M C and Q Add R ( Q ) . Ext -injective modules and cogenerator modules can be defined dually.

The next result follows somewhat similar arguments to those of [19, Proposition 3.6].

Proposition 3.6

Let C be a preenveloping class of left R -modules containing the class of injective modules and with an Ext -projective generator Q of C such that Add R ( Q ) C . Then Q is w-tilting.

Proof

Since Q is Ext -projective in C and Add ( Q ) C we immediately have that Q is Σ -self-orthogonal, so we only have to prove that R has an exact right Add ( Q ) -resolution.

Start by choosing a C -preenvelope g 0 : R T 0 of R . Since C contains the injective modules, the injective envelope E ( R ) of R holds in C and then the diagram can be completed commutatively. This shows that every C -preenvelope (in particular g 0 ) is monic.

Now, Q is a generator in C so there is an exact sequence

0 Y 0 Q 0 α T 0 0

with Q 0 Add R ( Q ) and Y 0 C . Then, any f 0 Hom R ( R , Q 0 ) with α f 0 = g 0 is a monic C -preenvelope since for any morphism h : R X with X C we have the situation

But g 0 is a C -preenvelope so there is an h 0 : T 0 X completing the diagram commutatively and then h 0 α : Q 0 X shows that f 0 is a preenvelope.

Thus, the exact sequence

0 R f 0 Q 0 K 0 0

is also Hom R ( , C ) -exact, and since Ext R 1 ( Q 0 , C ) = 0 ( Q is Ext -projective in C and so Q 0 is too) we get that Ext R 1 ( K 0 , C ) = 0 . This means that we can substitute R by K 0 and apply again the same argument to K 0 , and then to K 1 and so on, getting at the end an exact right Add R ( Q ) -resolution

0 R Q 0 Q 1

Dually, the following result can be easily proved.

Proposition 3.7

Let C be a precovering class in R -Mod containing the class of projective modules and with an Ext -injective cogenerator D of C such that Prod R ( D ) C . Then, D is w-cotilting.

We now give results which find sufficient conditions on C for Add R ( C ) to be precovering in C ( R ) and Prod S ( C ) to be preenveloping in A C ( S ) .

Proposition 3.8

Let C be a self-small and Σ -self-orthogonal left R -module. Any module in C ( R ) has an epic Add R ( C ) -precover with kernel in C ( R ) .

Proof

By [11, Proposition 5.4] we have that P ( S ) A C ( S ) so we can apply [20, Theorem 2.14] to get, for any M C ( R ) , an epic C S P ( S ) -precover γ : L M . But Proposition 3.1 says that Add R ( C ) = C S P ( S ) so we have indeed found and epic Add R ( C ) -precover γ : L M of M .

Now, R C being self-small and Σ -self-orthogonal means that Add R ( C ) C ( R ) , so indeed L C ( R ) . And L , M C ( R ) implies Ext R 1 ( C , L ) = Ext R 1 ( C , M ) = 0 , so applying Hom R ( C , ) to the exact sequence

0 ker γ L γ M 0

we immediately get that Ext R 1 ( C , ker γ ) = 0 .

Similarly, the fact that M is in C ( R ) implies Tor 1 S ( C , Hom R ( C , M ) ) = 0 , so applying C S to the exact sequence

0 Hom R ( C , ker γ ) Hom R ( C , L ) Hom R ( C , M ) 0

we get that Tor 1 S ( C , Hom R ( C , ker γ ) ) Tor 1 S ( C , Hom R ( C , L ) ) = 0 ( L C ( R ) ).

Finally, the commutative diagram with exact rows shows that ker γ C S Hom R ( C , ker γ ) canonically.□

Proposition 3.9

Suppose C is self-orthogonal and has a degreewise finite projective resolution as a right S -module, and End S ( C ) R canonically. Then, any N A C ( S ) has a monic Prod S ( C ) -preenvelope 0 N α T with coker α A C ( S ) .

Proof

By Proposition 3.2 we have that Prod S ( C ) = Hom R ( C , ( R ) ) , and by [3, Theorem 2.5] that ( R ) C ( R ) , so finally we get Prod S ( C ) = Hom R ( C , ( R ) ) A C ( S ) . Then, applying [20, Theorem 2.11] we find a monic Prod S ( C ) -preenvelope 0 N α T .

Now, applying Hom S ( , C ) to the exact sequence

0 N T A 0

we immediately get that Ext S 1 ( A , C ) = 0 since N , T A C ( S ) . Thus, by [1, Theorem 3.2.1],

Tor 1 S ( C , A ) Ext S 1 ( A , C ) = 0

and then Tor 1 S ( C , A ) = 0 , that is, A = coker α satisfies A1.

But then the sequence

0 C S N C S T C S A 0

is also exact, and applying Hom R ( C , ) we get another short exact sequence which associated long exact sequence gives Ext R 1 ( C , C S A ) = 0 (so A satisfies A2 as well), and a commutative diagram with exact rows which shows that A Hom R ( C , C S A ) canonically, that is, A satisfies A3.□

We will now see that if R C is w -tilting, then the class C ( R ) can be recovered from C itself and from the cokernels of any of the exact right Add R ( C ) -resolutions of R , at least when R C is self-small.

Theorem 3.10

Let C be a self-small and w -tilting left R -module and

0 R t 0 C 0 t 1 C 1 t 2

be an exact right Add R ( C ) -resolution of R . Then, C ( R ) = ( C ( i = 0 coker t i ) ) .

Proof

Call K i = coker t i , i 0 , K = C ( i = 0 K i ) and choose any M C ( R ) . By Proposition 3.8, we know that M has an epic Add R ( C ) -precover γ : L M with ker γ C ( R ) , so the exact sequence

0 ker γ L γ M 0

is Hom R ( C , ) -exact.

Applying both Hom R ( , M ) and Hom R ( , L ) to the exact sequence

0 R t 0 C 0 K 0 0

we get a commutative diagram with exact rows with Ext R 1 ( C 0 , M ) = 0 (and so Ext R 2 ( K 0 , M ) = 0 ) since M C ( R ) .

Now, the commutativity of the diagram shows that Hom R ( t 0 , M ) is an epimorphism so we see that Ext R 1 ( K 0 , M ) = 0 and so that Ext R 1 ( K 0 , H ) = 0 for all H in C ( R ) .

Now, for any j 0 we have that Ext R 1 ( C j + 1 , ker γ ) = 0 , so the resulting long exact sequence after applying Hom R ( , ker γ ) to the exact sequence

0 K j C j + 1 K j + 1 0

shows that Ext R i ( K j , ker γ ) Ext R i + 1 ( K j + 1 , ker γ ) i 1 .

On the other hand, applying Hom R ( K j , ) to the exact sequence

0 ker γ L γ M 0 ,

and noting that Ext R 1 ( K j , L ) = 0 ( 0 R C 0 C 1 is a right Add R ( C ) -resolution) we immediately see that Ext R i ( K j , M ) Ext R i + 1 ( K j , ker γ ) for every i 1 and every j 0 .

Therefore ,we get

Ext R i ( K j , ker γ ) Ext R i ( K j + 1 , M ) i 1 , j 0 .

But we know that Ext R 1 ( K 0 , ker γ ) = 0 so we have that Ext R 1 ( K 1 , M ) = 0 and then, by induction, we see that Ext R 1 ( K j , M ) = 0 j 0 and so that Ext R 1 ( K , M ) = 0 .

Conversely, P ( S ) A C ( S ) [11, Proposition 5.4] and any M K is in C , so by [20, Theorem 2.14] it suffices to find an exact left C S P ( S ) -resolution of M , which, by Proposition 3.1, is equivalent to finding an exact left Add R ( C ) -resolution of M .

Let R ( X ) M be an epimorphism, consider the morphism t 0 : R ( X ) C 0 ( X ) and construct the pushout Since Ext R 1 ( K 0 ( X ) , M ) Ext R 1 ( K 0 , M ) X = 0 , the right-hand column splits and hence there is an epimorphism C 0 ( X ) M and this means that every Add R ( C ) -precover C M of M (which exists by [18, Corollary 3.7]) is epic.

Now, applying Hom R ( K i , ) to the exact sequence

0 B C M 0

we get that 0 = Ext R j ( K i , M ) Ext R j + 1 ( K i , B ) j 1 , i 0 , and applying Hom R ( C , ) to the same sequence we obtain that B C . But then, the long exact sequence associated with

0 K i C i + 1 K i + 1 0

after applying Hom R ( , B ) shows that Ext R j ( K i , B ) Ext R j + 1 ( K i + 1 , B ) = 0 j 1 , i 0 , that is, B K . Repeating this procedure we find an exact left Add R ( C ) -resolution of M .□

Now let us show that when C satisfies the conditions of Theorem 3.10, the class C ( R ) is the right-hand class of a cotorsion theory which is in fact complete and hereditary. We recall first what a (co)resolving class of modules is.

Definition 3.11

A class of modules C is said to be resolving provided that it contains the class of projective modules and is closed under extensions and kernels of epimorphisms, that is, if

0 K M N 0

is a short exact sequence and K C , then M C K C .

Dually, C is coresolving if it contains the class of injective modules and is closed under extensions and cokernels of monomorphisms (if K C , then M C N C ).

Corollary 3.12

Let R C be self-small and w -tilting. Then, ( C 1 ( R ) , C ( R ) ) is a complete and hereditary cotorsion theory. In particular, C ( R ) is special preenveloping and 1 C ( R ) is special precovering.

Proof

Call K = C ( i = 0 coker t i ) the module of Theorem 3.10, choose any projective representation of K ,

P 1 P 0 K 0 ,

call L i = ker ( P i 1 P i 2 ) (where P 1 = K ) and consider the module X = K ( L i ) . We claim that C ( R ) = X 1 .

For any left R -module A we have

Ext R 1 ( X , A ) Ext R 1 ( K , A ) × ( Ext R 1 ( L i , A ) ) ,

and of course Ext R 1 ( L i , A ) Ext R i + 1 ( K , A ) , so we get

Ext R 1 ( X , A ) i 1 Ext R i ( K , A ) .

Therefore, we see that X 1 = K = C ( R ) ( K = C ( R ) by Theorem 3.10), that is, the pair ( C 1 ( R ) , C ( R ) ) is cogenerated by the set X . This clearly implies that it is indeed a cotorsion theory since

( C 1 ( R ) ) 1 = ( 1 ( X 1 ) ) 1 = X 1 = C ( R ) .

Now, it is well known that when a cotorsion theory is cogenerated by a set, then it is complete (see for instance [1, Theorem 7.4.1]), and C ( R ) is closed under extensions and cokernels of monomorphisms [11, Proposition 5.6] so it is indeed a coresolving class, and applying [16, Theorem 1.2.10] we get that ( C 1 ( R ) , C ( R ) ) is also hereditary.□

We now recall the definition of duality pairs.

Definition 3.13

[21, Definition 2.1] A left duality pair over R is a pair ( , C ) , where is a class of left R -modules and C is a class of right R -modules, subject to the following conditions:

  1. For a left R -module M , one has M if and only if M + C .

  2. C is closed under direct summands and finite direct sums.

Analogously, right duality pairs with in Mod- R and C is R -Mod are defined.

A left duality pair ( , C ) is called (co)product-closed if the class is closed under (co)products in the category of all left R -modules.

A left duality pair ( , C ) is called perfect if it is coproduct-closed, is closed under extensions, and R belongs to .

We now want to check when A C ( R ) , A C ( S ) , C ( R ) and C ( S ) are covering, and indeed, which of these pairs behave as a perfect cotorsion theory situation.

We start our way with the next result, whose proof follows the arguments of [6, Proposition 7.2].

Proposition 3.14

Assume that R C has a degreewise finite projective resolution. The following statements hold:

  1. M A C ( S ) M + C ( S ) .

  2. M C ( R ) M + A C ( R ) .

Now, let S be any ring, C be a right S -module admitting a degreewise finite projective resolution and R = End S ( C ) . The following statements hold:
  1. M A C ( R ) M + C ( R ) .

  2. M C ( S ) M + A C ( S ) .

Corollary 3.15

Let R C have a degreewise finite projective resolution. The following statements hold:

  1. ( A C ( S ) , C ( S ) ) is a coproduct-closed left duality pair over S and A C ( S ) is closed under extensions. If, in addition, R C is Σ -self-orthogonal, then the left duality pair is perfect.

  2. ( C ( R ) , A C ( R ) ) is a coproduct-closed left duality pair over R .

If, in addition, C S has a degreewise finite projective resolution, then the pairs ( A C ( S ) , C ( S ) ) and ( C ( R ) , A C ( R ) ) are also product-closed.

On the other hand, if C S has a degreewise finite projective resolution, then we have:

  1. ( A C ( R ) , C ( R ) ) is a coproduct-closed right duality pair over R and A C ( R ) is closed under extensions. Furthermore, if C S is Σ -self-orthogonal, then the pair is a perfect right duality pair.

  2. ( C ( S ) , A C ( S ) ) is a coproduct-closed right duality pair over S .

If, in addition, R C has a degreewise finite projective resolution, then the pairs ( A C ( R ) , C ( R ) ) and ( C ( S ) , A C ( S ) ) are also product-closed.

Proof

Items (a) and (b): by Proposition 3.14 we see that the two pairs are left duality pairs. Of course both are coproduct-closed as well (and product closed when C S has a degreewise finite projective resolution), and by [11, Proposition 5.4]  A C ( S ) is closed under extensions. If R C is Σ -self-orthogonal, then A C ( S ) contains all projective S -modules again by [11, Proposition 5.4].

Items (c) and (d) are proved dually.□

With the tools developed so far we can answer several of the questions raised in Section 1. This answer generalizes [21, Theorem 3.2] and improves [3, Theorem 4.4].

Theorem 3.16

  1. Let R C has a degreewise finite projective resolution. The following assertions hold.

    1. The classes A C ( S ) and C ( R ) are both covering.

    2. If, furthermore, R C is Σ -self-orthogonal, then ( A C ( S ) , A C ( S ) 1 ) is a perfect cotorsion theory.

  2. Assume now that C S has a degreewise finite projective resolution. The following statements are true:

    1. The classes A C ( R ) and C ( S ) are covering.

    2. If, in addition, C S is Σ -self-orthogonal, then ( A C ( R ) , A C ( R ) 1 ) is a perfect cotorsion theory.

  3. If both R C and C S have degreewise finite projective resolutions, then the classes A C ( S ) , C ( R ) , A C ( R ) and C ( S ) are all preenveloping.

Proof

  1. By Corollary 3.15 ( A C ( S ) , C ( S ) ) and ( C ( R ) , A C ( R ) ) are coproduct-closed left duality pairs, so by [21, Theorem 3.1, (b)] both A C ( S ) and C ( R ) are covering.

    Now, if R C is Σ -self-orthogonal, then ( A C ( S ) , C ( S ) ) is a perfect duality pair [11, Proposition 5.4] so, again by [21, Theorem 3.1], ( A C ( S ) , A C ( S ) 1 ) is a perfect cotorsion theory.

  2. Dual to (i).

  3. If both R C and C S have degreewise finite projective resolutions, then all the classes A C ( S ) , C ( R ) , A C ( R ) and C ( S ) are closed under products and then [21, Theorem 3.1, (a)] says that all of them are preenveloping.□

Corollary 3.17

Suppose that R C has a degreewise finite projective resolution. Then, the pair ( A C ( S ) , A C ( S ) 1 ) is a hereditary cotorsion theory if and only if R C is Σ -self-orthogonal.

Proof

If ( A C ( S ) , A C ( S ) 1 ) is a cotorsion theory, then A C ( S ) = 1 ( A C ( S ) 1 ) and so P ( S ) A C ( S ) . Then, by [11, Proposition 5.4, 1.] we get that R C is Σ -self-orthogonal.

Conversely, if R C is Σ -self-orthogonal, then ( A C ( S ) , A C ( S ) 1 ) is a cotorsion theory by Theorem 3.16 and so again P ( S ) A C ( S ) , and if A A C ( S ) is arbitrary and

0 K P A 0

is exact with P a projective left S -module, then K A C ( S ) [11, Proposition 5.4, 4.]. Therefore, for any M A C ( S ) 1 , the exactness of the sequence

Ext S i + 1 ( A , M ) Ext S i + 1 ( P , M ) Ext S i + 1 ( K , M )

shows, using induction, that Ext S 1 ( A , M ) = 0 and so that M A C ( S ) .□

Using Theorem 3.16 and Proposition 3.6 we can see what conditions we need on C and C ( R ) to ensure that R C is w-tilting.

Corollary 3.18

Suppose that R C is Σ -self-orthogonal, that C S is self-orthogonal, that R End S ( C ) canonically and that both R C and C S have degreewise finite projective resolutions. Then, C ( R ) is a coresolving, preenveloping and covering class, R C is an Ext -projective generator in C ( R ) and Add R ( C ) C ( R ) . As a consequence, R C is w-tilting.

Proof

The class C ( R ) is preenveloping and covering (Theorem 3.16), coresolving [3, Theorem 2.5], Add R ( C ) C ( R ) [11, Proposition 5.6] and of course R C is Ext -projective in C ( R ) by definition.

Now, any B C ( R ) has an epic Add R ( C ) -precover with kernel in C ( R ) (Proposition 3.8), so C is a generator in C ( R ) .

Finally, applying Proposition 3.6 we get that R C is w-tilting.□

In a similar way we have the following result.

Corollary 3.19

Suppose that R C is Σ -self-orthogonal and Hom -faithful, that C S is self-orthogonal, that R End S ( C ) canonically and that both R C and C S have degreewise finite projective resolutions. Then, A C ( S ) is a resolving, covering and preenveloping class, C is an Ext -injective cogenerator in A C ( S ) and Prod S ( C ) A C ( S ) . As a consequence, C is w-cotilting.

Proof

Theorem 3.16 shows that A C ( S ) is both covering and preenveloping, A C ( S ) is resolving by [11, Proposition 5.4], C is a cogenerator of A C ( S ) by Proposition 3.9, and it is Ext-injective by [1, Theorem 3.2.1], and Prod S ( C ) A C ( S ) by [3, Theorem 2.5]. Therefore, C is w-cotilting by Proposition 3.7.□

4 Gorenstein C -injective preenvelopes and Gorenstein C -projective precovers. Applications

In this section, we address another interesting question: when are the classes of syzygies of complete Add R ( C ) -resolutions and complete Prod S ( C ) -resolutions, that is the classes C - G P ( R ) and C - G I ( S ) of Gorenstein C -projective modules and Gorenstein C -injective modules, precovering and preenveloping respectively? The answer to this question will be given in Proposition 4.7 and Corollary 4.4.

As applications we will see that, dually to Theorem 3.10, the Auslander class can also be recovered from a right Add R ( C ) resolution of R when C is w-tilting (Corollary 4.8), and we will also be able to compute the common parts of the Bass class with its first left-orthogonal class, and of the Auslander class with its first right-orthogonal class, obtaining that these intersections are precisely the classes of P C -projective and C -injective modules, respectively (Proposition 4.9).

We start by defining all these C -relative concepts.

Definition 4.1

A left R -module is said to be P C -projective if it is isomorphic to some C S P ( P projective in S -Mod).

A module M is G C -projective if it has a complete ( P ( R ) , C S P ( S ) ) -complete resolution. The class of all G C -projective modules will be denoted by G C P ( R ) .

M is said to have G C -projective dimension ( P C -projective dimension) less than or equal to n , G C –pd ( M ) n ( P C –pd ( M ) n ), if there is an exact sequence

0 G n G 1 G 0 M 0

with G i G C P ( R ) ( G i P C ( R ) ) for every i { 0 , , n } . If n is the least nonnegative integer for which such a sequence exists, then G C –pd ( M ) = n ( P C –pd ( M ) = n ), and if there is no such n , then G C –pd ( M ) = ( P C –pd ( M ) = ).

A right R -module M is said to be C -injective if it is isomorphic to Hom S ( C , E ) for some injective right S -module E , and it is G C -injective if it admits a complete ( Hom S ( C , ( S ) ) , ( R ) ) -resolution (where ( R ) denotes the class of injective right R -modules). We use G C I ( R ) to denote the class of all G C -injective R -modules.

The concepts of C -injective and of G C -injective dimensions can be defined in a natural way now.

We denote by C G P ( R ) the class of all left R -modules with a complete Add R ( C ) -resolution, and by C G I ( R ) that of all left S -modules having a complete Prod S ( C ) -resolution. We call these modules Gorenstein C -projective and Gorenstein C -injective, respectively.

It is worth noting that when R C is self-small, the class of P C -projective modules is nothing but Add R ( C ) (Proposition 3.1), so G C -projective modules are zero syzygies of exact and Hom R ( Add R ( C ) ) -exact complexes

P 1 P 0 A 0 A 1

with all P i projective and all A i in Add R ( C ) .

Similarly, if C S is self-orthogonal with a degreewise finite projective resolution and R = End S ( C ) , then C -injective and G C -injective right modules can be defined in terms of Prod S ( C ) .

We start with the case of complete Prod S ( C ) -resolutions. We will need the following facts.

Lemma 4.2

Let F : C D be a left adjoint functor of a functor G : D C , and μ , ν be the unit and the co-unit of the adjunction, respectively. If A , B O b ( D ) are two objects such that ν A and μ G ( B ) are both isomorphisms, then, for any morphism f : G ( A ) G ( B ) , it holds that G ( ν B F ( f ) ν A 1 ) = f .

Proof

We know that G ( ν A ) μ G ( A ) = 1 G ( A ) , so ν A being an isomorphism implies that μ G ( A ) is also an isomorphism and G ( ν A ) = μ G ( A ) 1 . Similarly, G ( ν B ) = μ G ( B ) 1 .

Thus, G ( ν B F ( f ) ν A 1 ) = μ G ( B ) 1 G F ( f ) μ G ( A ) .

Now, the naturality of μ says that the diagram is commutative, that is, μ G ( B ) 1 G F ( f ) μ G ( A ) = f and we are done.□

Proposition 4.3

Let R C be Hom-faithful and suppose that C S is S -faithful, has a degreewise finite projective resolution and C A C ( S ) . Then

C G I ( S ) = Hom R ( C , G I ( R ) C ( R ) ) .

Proof

We start by noting that by [11, Theorem 5.9] we have that R End S ( C ) and that C S is self-orthogonal. Then, by Proposition 3.2, Prod S ( C ) = Hom R ( C , ( R ) ) . Therefore, if M C G I ( S ) we can find a complete Hom R ( C , ( R ) ) -resolution of M

X : Hom R ( C , E 1 ) Hom R ( C , E 0 ) Hom R ( C , E 1 ) .

Call K i = ker ( X i X i + 1 ) (so M = K 1 ). It is clear that K i C G I ( S ) i so, again by [11, Theorem 5.9], K i A C ( S ) i and then C S K i C ( R ) i . In particular, C S M C ( R ) . But since M A C ( S ) we also have that M Hom R ( C , C S M ) , so if we prove that C S M G I ( R ) , we will have the left to right inclusion.

Applying C S to the resolution X we get the complex

C S X : E 1 E 0 E 1

since ( R ) C ( R ) [11, Theorem 5.9].

Now, for every i we have that the sequence

0 C S K i E i C S K i + 1 0

is exact since K i + 1 A C ( S ) (so Tor 1 S ( C , K i + 1 ) = 0 ), and this means that C S X is exact.

We then see that to have C S M G I ( R ) we only need to check that C S X is Hom R ( ( R ) , ) -exact.

Choose any injective left R -module E and any morphism f : E C S K i + 1 . We want to complete the diagram

If we apply Hom R ( C , ) we get the following diagram with our original sequence since K i A C ( S ) i .

But X is Hom S ( Hom R ( C , ( R ) ) , ) -exact, so this last diagram can be completed commutatively. Finally, apply C S to it to get the desired completion of our first diagram.

Conversely, let N G I ( R ) C ( R ) , choose a complete injective resolution of N and call L i = ker ( E i E i + 1 ) . Then, L i G I ( R ) i , but also L i C ( R ) i by [11, Proposition 5.6], so applying Hom R ( C , ) to the complex above we get another exact complex ( L i C ( R ) i )

It only remains to prove that Hom R ( C , E ) is Hom S ( Hom R ( C , ( R ) ) , ) -exact and Hom S ( , Hom R ( C , ( R ) ) ) -exact.

Let then E be any left R -module and f : Hom R ( C , E ) Hom R ( C , L i + 1 ) be any morphism. We want to prove that the diagram can be completed commutatively.

Applying C S to the diagram we get since all L i and all E i are in C ( R ) , where

ν E : C S Hom R ( C , E ) E

is the natural homomorphism given in the definition of the Bass class, that is, the co-unit of the adjunction ( C S , Hom R ( C , ) ) .

But E is Hom R ( ( R ) , ) -exact so this last diagram can be completed commutatively by some g : E E i . Applying again Hom R ( C , ) we get Hom R ( C , g ) : Hom R ( C , E ) Hom R ( C , E i ) such that

Hom R ( C , δ i ) Hom R ( C , g ) = Hom R ( C , ν L i + 1 ( 1 C f ) ν E 1 ) ,

and Hom R ( C , ν L i + 1 ( 1 C f ) ν E 1 ) , = f by Lemma 4.2, so we are done.

The Hom S ( , Hom R ( C , ( R ) ) ) -exactness of Hom R ( C , E ) is completely dual.□

Corollary 4.4

Suppose that R C is Hom -faithful and that C S is -faithful, has a degreewise finite projective resolution and C A C ( S ) . If i d R ( C ) < , then C G I ( S ) = Hom R ( C , G I ( R ) ) .

If, in addition, R is left noetherian, then C G I ( S ) is preenveloping.

Proof

By [11, Theorem 5.9] we have ( R ) C ( R ) , so [3, Proposition 3.2] says that G I ( R ) C ( R ) . Then, an application of Proposition 4.3 gives the first part.

Now, if R is left noetherian, then G I ( R ) is preenveloping in R -Mod, so by Theorem 3.3 Hom R ( C , G I ( R ) ) is preenveloping in S -Mod.□

We now give the dual results concerning complete Add R ( C ) -resolutions. We omit their proofs since they are dual to those of Proposition 4.3 and Corollary 4.4, respectively.

Proposition 4.5

Let R C be Hom -faithful, self-small and Σ -self-orthogonal. Then,

C G P ( R ) = C S ( G P ( S ) A C ( S ) ) .

Corollary 4.6

Let R C be self-small, Σ -self-orthogonal, Hom -faithful and suppose that E is an injective cogenerator in R -Mod such that p d S ( Hom R ( C , E ) ) < . Then, C G P ( R ) = C S G P ( S ) .

As a consequence of [17, Proposition 2.12], we have that the class C G P ( R ) is precovering in the subcategory of all modules having finite C G P ( R ) -dimension. Now we can give conditions for C G P ( R ) to be precovering in the whole category R -Mod.

Proposition 4.7

Suppose that both i d R ( C ) and i d S ( C ) are finite, that R C is self- Hom -faithful and Σ -self-orthogonal with a degreewise finite projective resolution, and that S is right noetherian and left n -perfect. Then, C G P ( R ) is precovering in R -Mod.

Proof

The class C ( R ) and that of all left R -modules having a finite left C G P ( R ) -resolution ( C G P ( R ) ¯ ) coincide by [3, Theorem 3.4]. Then, this class C G P ( R ) ¯ is covering in R -Mod by Theorem 3.16.

Let then M be any R -module and φ : G ¯ M be a C G P ( R ) ¯ -precover. By definition, G ¯ being in C G P ( R ) ¯ implies that G ¯ has G C P ( R ) -precover ψ : G G ¯ . It is immediate to verify that φ ψ : G M is a G C P ( R ) -precover of M .□

In Theorem 3.10 we showed how, under certain circumstances, the Bass class C ( R ) can be recovered from C and the cokernels of an exact right Add R ( C ) -resolution of R . We now show that, when C satisfies some additional conditions, the Auslander class A C ( R ) can also be recovered, in a different way, from the same modules.

Corollary 4.8

Suppose that R C is self-small and w-tilting (so let

0 R t 0 C 0 t 1 C 1 t 2

be an exact right Add R ( C ) -resolution of R and call K = ( C ( i = 0 coker t i ) ) ), that R = End S ( C ) and that C S has a degreewise finite projective resolution. Then,

A C ( R ) = { M Tor i 1 R ( M , K ) = 0 } = ( K + ) .

Proof

Using Proposition 3.14 we get that M A C ( R ) M + C ( R ) , and by Theorem 3.10 we know that C ( R ) = K , so we have

M A C ( R ) Ext R i ( K , M + ) = 0 i 1 .

But Ext R i ( K , M + ) Tor i R ( M , K ) + (see for instance [1, Theorem 3.2.1]) so indeed M A C ( R ) Tor i R ( M , K ) + = 0 Tor i R ( M , K ) = 0 i 1 .

But at the same time, we have Ext R i ( M , K + ) Tor i R ( M , K ) + , so again

Ext R i ( M , K + ) = 0 i 1 Tor i R ( M , K ) = 0 i 1

and we finally get M A C ( R ) Ext R i ( M , K + ) = 0 i 1 .□

We finish this section by proving that, in some circumstances, the common part of the two components of the pairs ( A C ( S ) , A C ( S ) 1 ) and ( C 1 ( R ) , C ( R ) ) is precisely the class of C -injective left S -modules and the class of P C -projective left R -modules, respectively.

Proposition 4.9

The following statements hold.

  1. R C is self-small and Σ -self-orthogonal if and only if C ( R ) 1 C ( R ) = C S P ( S ) .

  2. Suppose that R C is Hom -faithful and that C S has a degreewise finite projective resolution. Then, R End S ( C ) (canonically) and C S is self-orthogonal if and only if A C ( S ) A C ( S ) 1 = Hom R ( C , ( R ) ) .

Proof

(1) (2) On one hand, we know by Proposition 3.1 that Add R ( C ) = C S P ( S ) and by [11, Proposition 5.6] that Add R ( C ) C ( R ) . On the other hand, for any projective left S -module P and any left R -module B C ( R ) we have that Ext R 1 ( C S P , B ) is a direct summand of Ext R 1 ( C S S ( I ) , B ) for some index set I . But

Ext R 1 ( C S S ( I ) , B ) Ext R 1 ( C ( I ) , B ) Ext R 1 ( C , B ) I = 0

so Ext R 1 ( C S P , B ) = 0 .

We then have C S P ( S ) C ( R ) 1 C ( R ) .

Conversely, choose any M C ( R ) 1 C ( R ) . By Proposition 3.8 we know that M has an epic Add R ( C ) -precover 0 T L M 0 with kernel T C ( R ) . But then Ext R 1 ( M , T ) = 0 and so the sequence splits, which proves that M Add R ( C ) = C S P ( S ) .

( ) For any set I we have C ( I ) C S S ( I ) C S P ( S ) = C ( R ) C 1 ( R ) , so C ( I ) C ( R ) . Then, Add R ( C ) C ( R ) and then R C is Σ -self-orthogonal and self-small by [11, Proposition 5.6].

(2) (1) By [11, Theorem 5.9] we know that ( R ) C ( R ) so Hom R ( C , ( R ) ) A C ( S ) .

On the other hand, for any injective E ( R ) and any A A C ( S ) we have

Ext S 1 ( A , Hom R ( C , E ) ) Hom R ( Tor 1 S ( C , A ) , E )

by [1, Theorem 3.2.1]. But Tor 1 S ( C , A ) = 0 since A A C ( S ) , so Hom R ( C , E ) A C ( S ) 1 and then Hom R ( C , ( R ) ) A C ( S ) A C ( S ) 1 .

Conversely, if M A C ( S ) A C ( S ) 1 , [20, Theorem 2.11] guarantees the existence of a monic Hom R ( C , ( R ) ) -preenvelope 0 M X . Now, since Hom R ( C , ( R ) ) A C ( S ) , we have that M , X A C ( S ) , so by [11, Proposition 5.4, 2] the cokernel X / M is also in A C ( S ) . But M A C ( S ) 1 , so applying Hom S ( X / M , ) to the preenvelope we see that the sequence

0 M X X / M 0

splits, and since Hom R ( C , ( R ) ) = Prod S ( C ) , we see that M Hom R ( C , ( R ) ) .

( ) Hom R ( C , ( R ) ) A C ( S ) C A C ( S ) , so the result follows by [11, Theorem 5.9].□

5 G C -flat and G C -projective covers and dimensions

The information existing in the literature concerning the existence of relative Gorenstein flat precovers is very limited. In this section, we want to address two natural problems which arise when studying this class of modules: when do relative Gorenstein flat (pre)covers exist? and, when is the class of relative Gorenstein projective modules included in the class of relative Gorenstein flats? And, related to this last problem, can we compare relative Gorenstein projective and flat dimensions? Answers to the first two questions will be given in Corollary 5.7 and Theorem 5.9, respectively, and a partial answer to the third question will be given in Theorem 5.13.

To better know how the class of relative Gorenstein flat modules look like, we will study its relations with the class of relative Gorenstein injective modules in Proposition 5.6.

As a consequence of the results to be obtained in this section, we will also give sufficient conditions for the class of relative projective modules to be special precovering in R -Mod (Corollary 5.14).

We start by recalling the concept of relative Gorenstein flat module.

Definition 5.1

A left R -module M is said to be G C -flat if there exists an exact and ( Hom S ( C , ( S ) ) R ) -exact sequence of left R -modules

X = F 1 F 0 C S F 0 C S F 1 ,

where the F i ’s are all flat left R -modules, the F i ’s are all flat left S -modules and M = Im ( F 0 C S F 0 ) .

The class of all G C -flat left R -modules will be denoted by G C F ( R ) .

It is immediate to see that the class G C F ( R ) is closed under direct sums.

Proposition 5.2

Suppose that R C is Σ -self-orthogonal and has a degreewise finite projective resolution. Then, ( S ) A C ( S ) and so we have C S ( S ) C ( R ) and Hom R ( C , C S ( S ) ) ( S ) .

Proof

By [11, Proposition 5.4] we know that P ( S ) A C ( S ) , and flat modules are direct limits of projective modules, so R C having a degreewise finite projective resolution (which means that both Hom R ( C , ) and Ext R 1 ( C , ) preserve direct limits) proves that ( S ) A C ( S ) . But then C S ( S ) C ( R ) and so Hom R ( C , C S ( S ) ) A C ( S ) . Therefore, for any F ( S ) we have

F Hom R ( C , C S F )

so Hom R ( C , C S F ) is a flat left S -module.□

From Proposition 5.2, we can give conditions for the class C S ( S ) to be closed under direct products and pure submodules, which, as a consequence, implies that C S ( S ) is preenveloping.

Proposition 5.3

Let R C be Hom -faithful, Σ -self-orthogonal and have a degreewise finite projective resolution. Then, C S ( S ) is closed under pure submodules.

If S is right coherent and C S is finitely presented, then C S ( S ) is closed under direct products.

As a consequence, when all conditions above are satisfied, C S ( S ) is preenveloping in R -Mod.

Proof

To prove that C S ( S ) is closed under pure submodules one can follow [6, Lemma 5.2] and use Proposition 5.2 when necessary (since [6, Lemma 5.2] is proved when C is semidualizing).

The statement about C S ( S ) being closed under products is clear.

Finally, we know that if a Kaplansky class is closed under direct products, then it is preenveloping [1, Theorem 11.9.4] so we only have to prove that C S ( S ) is Kaplansky.

For let F ( S ) be any flat left S -module and x C S F be any element. Then, we can write x as x = i = 1 n c i x i , and by [1, Lemma 5.3.12] we know that there is a cardinal number α (dependent only on R ) such that for any i there exists a pure submodule A i F containing x i such that A i < α . But then A = i = 1 n A i F is a pure submodule of F so it is indeed a flat module, and of course A < α . If we choose any cardinal number > Max { C , α } we get that C S A C × A < . Therefore, we have x C S A C S ( S ) and C S A < .

Now, A F being pure means that the sequence

0 C S A C S F C S F A 0

is exact, so since F / A is flat we immediately see that C S F C S A holds in C S ( S ) and so that C S ( S ) is Kaplansky.□

It is a very interesting question to know whether or not the class of I C -injective right R -modules, that is, the class Hom S ( C , ( S ) ) is embedded in the class of all G C -injective modules. It is known that if C is semidualizing then the answer is yes, but what happens in more general cases? We will need an answer to this question in the proof of our Proposition 5.5.

Proposition 5.4

Suppose that R C is self-orthogonal and that it has a degreewise finite projective resolution. Then, the class Hom S ( C , ( S ) ) is inside the class G C I ( R ) of all G C -injective right R -modules.

Proof

For any injective right S -module I , let

0 Hom S ( C , I ) E 0 E 1 E 2

be an injective resolution of the left R -module Hom S ( C , I ) . Let us prove that the sequence is Hom R ( Hom S ( C , ( S ) ) , ) -exact.

Call K i = ker ( E i E i + 1 ) for every i 1 , choose any injective right S -module I and apply the functor Hom R ( Hom S ( C , I ) , ) to the exact sequence

0 Hom S ( C , I ) E 0 K 1 0 .

We get that Ext R i ( Hom S ( C , I ) , K 1 ) Ext R i + 1 ( Hom S ( C , I ) , Hom S ( C , I ) ) for every i .

But

Ext R i ( Hom S ( C , I ) , Hom S ( C , I ) ) Hom S ( Tor i R ( Hom S ( C , I ) , C ) , I )

by [1, Theorem 3.2.1], and

Hom S ( Tor i R ( Hom S ( C , I ) , C ) , I ) Hom S ( Hom S ( Ext R i ( C , C ) , I ) , I )

by [1, Theorem 3.2.13], so we get Ext R i ( Hom ( C , I ) , K 1 ) = 0 i since C is self-orthogonal.

Now, Ext R i ( Hom S ( C , I ) , K j ) Ext R i + j ( Hom S ( C , I ) , Hom S ( C , I ) ) = 0 so our original injective resolution is Hom R ( Hom S ( C , ( S ) ) , ) -exact.□

Now we give a slight modification of the dual of Holm’s Corollary 2.11 in [22] for Gorenstein projective modules, extended by White in [14, Corollary 3.8] to the case of G C -projective modules for a semidualizing module C .

Proposition 5.5

Suppose that R C is self-orthogonal and has a degreewise finite projective resolution. If 0 M Hom S ( C , I ) G 0 is a short exact sequence of right R -modules where I is an injective right S -module, G is G C -injective and Ext R 1 ( A , M ) = 0 for every C -injective right R -module A , then M is G C -injective as well.

Proof

Since G is G C -injective, there is a Hom R ( Hom S ( C , ( S ) ) , ) -exact short exact sequence

0 K Hom S ( C , I 0 ) G 0

for some injective right S -module I 0 . We claim that the right R -module K is G C -injective.

Of course, K has an exact left Hom S ( C , ( S ) ) -resolution since G is G C -injective. Let then

0 K E 0 E 1

be an injective resolution of K . If this resolution is Hom R ( Hom S ( C , ( S ) ) , ) -exact, then pasting together both resolutions of K we get that K is G C -injective.

Call then K i = ker ( E i E i + 1 ) i 1 and choose any I ( S ) .

Applying Hom R ( Hom S ( C , I ) , ) to the exact sequence

0 K j E j K j + 1 0

we get that

Ext R i ( Hom S ( C , I ) , K j + 1 ) Ext R i + 1 ( Hom S ( C , I ) , K j ) ,

so we actually have

Ext R i ( Hom S ( C , I ) , K j ) Ext R i + j ( Hom S ( C , I ) , K ) i , j 1 .

But, as in the proof of Proposition 5.4, Ext R i ( Hom S ( C , I ) , Hom S ( C , I 0 ) ) = 0 , so from the exact sequence

0 K Hom S ( C , I 0 ) G 0

we see that Ext R i + 1 ( Hom S ( C , I ) , K ) Ext R i ( Hom S ( C , I ) , G ) , and of course, since G is G C -injective, Ext R i ( Hom S ( C , I ) , G ) = 0 , so we get Ext R i 2 ( Hom S ( C , I ) , K ) = 0 .

As for Ext R 1 ( Hom S ( C , I ) , K ) = 0 , this follows from the Hom R ( Hom S ( C , ( S ) ) , ) -exactness of

0 K Hom S ( C , I 0 ) G 0 .

Therefore, K is G C -injective.

Now, constructing the pullback of Hom S ( C , I ) G and Hom S ( C , I 0 ) G we get a commutative diagram with exact rows

Since K is G C -injective we have Ext R 1 ( Hom S ( C , I ) , K ) = 0 and so we get P K Hom S ( C , I ) , that is, P is G C -injective (see Proposition 5.4). But we know by the hypotheses Ext R 1 ( Hom S ( C , I 0 ) , M ) = 0 so P M Hom S ( C , I 0 ) .

Finally, G C -injective modules are closed under direct summands as a consequence of [23, Proposition 4.8]: though the authors assume that their classes V and W are closed under finite direct sums and direct summands throughout the paper, their arguments are similar to and based on those of [24, Theorem 3.6], and the only necessary assumption is that the classes are closed under finite direct sums. But both Hom S ( C , ( S ) ) and ( R ) are closed under finite direct sums, so G C ( R ) -injective modules are closed under direct summands.

Therefore, M is G C -injective.□

The following result has been proved when C is semidualizing [25, Proposition 2.5]. Here we give the following extension whose proof follows that of [25] with the only modification of using Propositions 5.3, 5.4 and 5.5, which are adaptations to our context of the analogous cases when C is semidualizing.

Proposition 5.6

Let M be any left R -module. The following statements hold:

  1. If M is G C -flat, then M + is a G C -injective right R -module.

  2. Suppose now that S is right coherent, that R C is Σ -self-orthogonal, Hom -faithful and that it has a degreewise finite projective resolution, and finally that C S is finitely presented. If M + G C I ( R ) , then M G C F ( R ) .

Corollary 5.7

Suppose that C S finitely presented, that R C is Σ -self-orthogonal and Hom -faithful, and that it has a degreewise finite projective resolution. If S is right coherent, then the class G C F ( R ) is covering in R -Mod.

Proof

Since G C I ( R ) is closed under products and direct summands, Proposition 5.6 shows that the pair ( G C F ( R ) , G C I ( R ) ) is a left duality pair, which, in addition, is coproduct closed, so [21, Theorem 3.1] guarantees the existence of a G C F ( R ) -cover for every left R -module.□

Lemma 5.8

Let R C be Σ -self-orthogonal and X be a complete ( P ( R ) , C S P ( S ) ) -resolution of certain module. If

0 C S Q n C S Q n 1 C S Q 0 L 0

is an exact complex and all Q i are projective left S -modules, then the sequence of complexes of abelian groups

0 Hom R ( X , C S Q n ) Hom R ( X , C S Q 0 ) Hom R ( X , L ) 0

is also exact.

Proof

We have to prove that for any integer m , the sequence

0 Hom R ( X m , C S Q n ) Hom R ( X m , C S Q 0 ) Hom R ( X m , L ) 0

is exact.

If m 0 , then X m is projective and so the corresponding sequence is exact.

On the other hand, if m 1 , then X m = C S K for some projective left S -module K . Then, K is a direct summand of S ( J ) for some set J , and likewise, Q n is a direct summand of some S ( I ) . Therefore, Ext R i ( C S K , C S Q n ) is (isomorphic to) a direct summand of Ext R i ( C S S ( J ) , C S S ( I ) ) Ext R i ( C , C ( I ) ) J .

But R C is Σ -self-orthogonal so Ext R i ( C , C ( I ) ) J = 0 and then Ext R 1 ( C S K , C S Q n ) = 0 .

The same argument shows that Ext R 1 ( C S K , C S Q i ) = 0 for every i , so making use of the long exact sequence associated with

0 C S Q n C S Q n 1 A n 1 0

we see that Ext R 1 ( X m , A n 1 ) = 0 .

An inductive process gives now that indeed Ext R 1 ( X m , A i ) = 0 for every i so the sequences

0 Hom R ( X m , C S Q n ) Hom R ( X m , C S Q 0 ) Hom R ( X m , L ) 0

are all exact.□

Theorem 5.9

Suppose that S is right coherent, that C S is finitely presented and that R C is Σ -self-orthogonal. If P C –pd ( C S F ) < for every flat and cotorsion left S -module F , then G C P ( R ) G C F ( R ) .

Proof

Let M be any G C -projective left R -module and X a complete ( P(ℛ) , C S P(S) ) -resolution of M .

If we prove that the complex Hom S ( C , I ) R X is exact for any injective right S -module I , then we will have that M is G C -flat.

Of course Hom S ( C , I ) R X is exact if and only if ( Hom S ( C , I ) R X ) + is exact, and

( Hom S ( C , I ) R X ) + Hom R ( X , Hom S ( C , I ) + ) Hom R ( X , C S I + ) .

Now, I + is cotorsion, and since S is right coherent, I + is also flat, so, by the hypotheses, P C –pd ( C S I + ) < . This means that there is an exact sequence

0 C S Q n C S Q n 1 C S Q 0 C S I + 0

with all Q i projective left S -modules.

Then, by Lemma 5.8 we get that the sequence of complexes

0 Hom R ( X , C S Q n ) Hom R ( X , C S Q 0 ) Hom R ( X , C S I + ) 0

is exact. Since each Hom R ( X , C S Q i ) is an exact complex we get by induction that Hom R ( X , C S I + ) is exact too.□

Our next goal will be to find conditions for every G C -flat module to have finite G C -projective dimension. We will need the following result, inspired by the proof of [26, Proposition 3.1].

Recall that a module M is said to have flat cotorsion dimension less than or equal to a nonnegative integer number n whenever there exists an exact sequence

0 M F 0 F 1 F n 0

with all F i flat and cotorsion modules.

The following result gives an interesting property of flat modules over n -perfect rings. We believe it is surely known but we have not found its proof so we state and prove it here.

Proposition 5.10

The flat cotorsion dimension of any flat left R -module over a left n -perfect ring R is always less than or equal to n .

Proof

Let F be any flat left R -module and

0 F C 0 C 1

be a cotorsion resolution of F . If we call K i = ker ( C i C i + 1 ) for every i 1 , then we know that K i is flat by Wakamatsu’s Lemma (see for instance [27, Lemma 2.1.2]), and this implies that every C i is flat and cotorsion. Therefore, we see that we only need to prove that K n is cotorsion. But for every flat module F ¯ we have Ext R 1 ( F ¯ , C i ) = 0 i , so

Ext R i ( F ¯ , K j ) Ext R i + j ( F ¯ , F ) i , j 1 .

Now, since R is left n -perfect we have Ext R i ( F ¯ , F ) = 0 i n + 1 so

Ext R 1 ( F ¯ , K n ) Ext n + 1 ( F ¯ , F ) = 0

and K n is cotorsion.□

Lemma 5.11

Suppose that S is left n -perfect, that C S is finitely presented and that R C is Σ -self-orthogonal. Then, given any exact complex X in R -Mod with X i G C -projective for every i , we have: if Hom S ( C , I ) R X is exact for any injective right S -module I , then Hom R ( X , C S P ) is exact for every projective left S -module P .

Proof

In first place, we note that the natural morphism of left S -modules F F + + is a pure injection, so if F is flat and cotorsion then F is a direct summand of F + + .

Now, C S is finitely presented so there is a natural isomorphism

( Hom S ( C , F + ) R X ) + Hom R ( X , C S F + + ) .

Thus, since F + is an injective right S -module, the complex ( Hom S ( C , F + ) R X ) + (and so the complex Hom R ( X , C S F + + ) ) is exact by the hypotheses, and therefore its direct summand Hom R ( X , C S F ) is exact. Hence, Hom R ( X , C S F ) is an exact complex for every flat and cotorsion left S -module F .

Now, since S left n -perfect, by Proposition 5.10 we have that if P is any projective left S -module, then we can find an exact sequence

0 P F 0 F 1 F n 0

in which every F i is a flat and cotorsion left S -module. Then, the sequence of left R -modules

0 C S P C S F 0 C S F n 0

is exact and, if we prove that

0 Hom R ( X , C S P ) Hom R ( X , C S F 0 ) Hom R ( X , C S F n ) 0

is exact, we will have that the complex Hom R ( X , C S P ) is exact since we have already proved that all the complexes Hom R ( X , C S F i ) are exact.

But each S F i is flat so it has finite projective dimension and then each C S F i has finite P C -projective dimension. Thus, Ext R i ( X j , C S F i ) = 0 i 1 j since every X j is G C -projective (see [28, Corollary 3.5]).□

We now give an extension of the concept of n -perfect ring, adapted to this new setting of C -relativity.

Definition 5.12

Given a nonnegative integer number n , a ring R will be said to be left ( C , n ) -perfect if every flat left R -module has P C -projective dimension less than or equal to n .

Theorem 5.13

Suppose that R is left ( C , n ) -perfect, that S is left n -perfect and left coherent and that R C is weakly Wakamatsu tilting and has a degreewise finite projective resolution. Then, every G C -flat left R -module has finite G C -projective dimension.

Proof

Let M be a G C -flat left R -module. Then, there is a ( Hom S ( C , ( S ) ) R ) -exact complex

X : X 1 X 0 X 1

in which M = Im ( X 0 X 1 ) , X i is a flat left R -module for every i 0 and X i C S F i where F i is a flat left S -module for every i > 0 .

Therefore, P C –pd ( X i ) n i 0 since R is left ( C , n ) -perfect, and p d - ( F i ) n i 1 since S left n -perfect. Then, P C –pd ( C S F i ) n i 1 and then, by [28, Theorem 3.6], we get that G C –pd ( X i ) n i .

Now, for any i 0 let

0 K i P i X i 0

be a short exact sequence in R -Mod with P i projective, and if i 1 consider an exact sequence

0 K i P i F i 0

in S -Mod with P i projective. Then, the sequences of R -modules

0 C S K i C S P i C S F i = X i 0

are also exact.

For any i 1 we have a commutative diagram with exact rows and columns. Similarly, if i 1 we have and if i = 0 we have

Therefore, if for any module N we call N ¯ [ i ] the complex with N at the i th and ( i + 1 )th positions, we get an exact sequence of complexes

0 W P 0 = ( i 0 P i ¯ [ i ] ) ( i 1 C S P i ¯ [ i ] ) X 0

in which every W i is obtained in the following way: for i 0 as the pullback For i 2 as the pullback For i = 1 as the pullback

But P 0 and X are exact complexes so W is exact too. Then, since P 0 and X are Hom S ( C , ( S ) ) R -exact, so is W .

Moreover, we see from the pullbacks above that

W i K i P i 1 if i 0 , ( C S K 1 ) P 0 if i = 1 .

Now, for any i > 1 we know that C S P i 1 is a direct summand of C S S ( I ) C ( I ) , so Ext R 1 ( C S P i 1 , C S K i ) is a direct summand of

Ext R 1 ( C ( I ) , C S K i ) Ext R 1 ( C , C S K i ) I .

Since R C has a degreewise finite projective resolution by [1, Theorem 3.2.15] we have

Ext R 1 ( C , C S K i ) Ext R 1 ( C , C ) S K i = 0

so we see that

W i ( C S K i ) ( C S P i 1 ) i > 1 .

Therefore, the complex W is

K 1 P 2 K 0 P 1 ( C S K 1 ) P 0 ( C S K 2 ) ( C S P 1 )

Moreover, for every i 0 , P 0 i is a projective left R -module so it is G C -projective; for any i 2 , P 0 i = ( C S P i 1 ) ( C S P i ) and G C –pd ( C S P i ) P C –pd ( C S P i ) = 0 ; and finally G C –pd ( P 0 ( C S P 1 ) ) = 0 , so in the exact sequence

0 W P 0 X 0

we have

Sup { G C –pd ( P 0 i ) , G C –pd ( X i ) 1 } = G C –pd ( X i ) 1 n 1

(unless X i is G C -projective, in which case W i will also be G C -projective by [11, Proposition 2.9]).

Hence, by [11, Proposition 3.11] we get that G C –pd ( W i ) n 1 i .

This means that we can apply the same argument to the complex W and we can start an inductive procedure to get an exact sequence of exact and ( Hom S ( C , ( S ) ) R ) -exact complexes

0 W t P t 1 P 0 X 0

with t n and G C –pd ( W t i ) = 0 i .

Now, Add R ( C ) = C S P ( S ) since R C is self-small (it has a degreewise finite projective resolution). Then, by [11, Proposition 2.16] we see that G C P 2 ( R ) = G C P ( R ) , and since every W t i has G C -projective dimension zero, we get that Z 0 ( W t ) is a G C -projective left R -module. But we have already seen that the left R -modules Z 0 ( P i ) are all G C -projective, so the exact sequence

0 Z 0 ( W t ) Z 0 ( P t 1 ) Z 0 ( P 0 ) Z 0 ( X ) = M 0

shows that G C –pd ( M ) n .□

We now give conditions for the class G C P ( R ) to be special precovering.

Corollary 5.14

Let R be a left ( C , n ) -perfect ring, S be left n -perfect and right coherent, R C be a Hom -faithful weakly Wakamatsu tilting module with a degreewise finite projective resolution and C S be finitely presented. Then, G C P ( R ) is special precovering in R -Mod.

Proof

By Theorems 5.13 and 5.7, we have the inclusions

G C P ( R ) G C F ( R ) G C P ( R ) ¯ ,

where G C P ( R ) ¯ denotes the class of left R -modules of finite G C -projective dimension.

Now, given any left R -module M , by Corollary 5.7 we know that we have a G C F ( R ) -cover φ : F M , and then, by [11, Theorem 3.5], there is a special G C P ( R ) -precover ψ : P F . We claim that ψ φ : P M is a special G C P ( R ) -precover of M .

Since φ is a cover and ψ is a special precover, we get that ker ( φ ) holds in G C F ( R ) 1 and ker ( ψ ) holds in G C P ( R ) 1 . But G C F ( R ) 1 G C P ( R ) 1 so indeed both kernels are in G C P ( R ) 1 . Then, from the left-hand column of the pullback diagram we see that D is in G C P ( R ) 1 , and this immediately implies that φ ψ is a special G C P ( R ) -precover.□



  1. Funding information: Juan R. García Rozas and Luis Oyonarte were partially supported by Ministerio de Economía y Competitividad, grant reference 2017MTM2017-86987-P.

  2. Conflict of interest: Authors state no conflict of interest.

References

[1] E. Enochs and O. Jenda , Relative Homological Algebra , 2nd ed., De Gruyter Expositions in Mathematics , vol. 30, Walter de Gruyter, Berlin-New York, 2011. 10.1515/9783110215236Search in Google Scholar

[2] H. B. Foxby , Quasi-perfect modules over Cohen-Macaulay rings, Math. Nachr. 66 (1975), 103–110. 10.1002/mana.19750660111Search in Google Scholar

[3] D. Bennis , J. R. García Rozas , and L. Oyonarte , When do Foxby classes coincide with the classes of modules of finite Gorenstein dimension?, Kyoto J. Math. 56 (2016), no. 4, 785–802, https://doi.org/10.1215/21562261-3664914. Search in Google Scholar

[4] E. Enochs , O. Jenda , and J. A. López-Ramos , Dualizing modules and n -perfect rings, Proc. Edinburgh Math. Soc. 48 (2005), no. 2, 75–90, https://doi.org/10.1017/S0013091503001056. Search in Google Scholar

[5] E. Enochs , O. Jenda , and J. Xu , Foxby duality and Gorenstein injective and projective modules, Trans. Amer. Math. Soc. 348 (1996), no. 8, 3223–3234. 10.1090/S0002-9947-96-01624-8Search in Google Scholar

[6] H. Holm and D. White , Foxby equivalence over associative rings, J. Math. Kyoto Univ. 47 (2007), 781–808, https://doi.org/10.1215/kjm/1250692289. Search in Google Scholar

[7] E. Enochs , S. Estrada , and J. R. García Rozas , Gorenstein categories and Tate cohomology on projective schemes, Math. Nachr. 281 (2008), 525–540, https://doi.org/10.1002/mana.200510622. Search in Google Scholar

[8] E. Enochs , J. R. García Rozas , L. Oyonarte , and B. Torrecillas , Gorenstein flat precovers and Gorenstein injective preenvelopes in Grothendieck categories, Ark. Mat. 57 (2019), 55–83, https://doi.org/10.4310/ARKIV.2019.v57.n1.a4. Search in Google Scholar

[9] J. Gillespie , Models for homotopy categories of injectives and Gorenstein injectives, Comm. Alg. 45 (2017), no. 6, 2520–2545, https://doi.org/10.1080/00927872.2016.1233215. Search in Google Scholar

[10] H. Krause , The stable derived category of a noetherian scheme, Compos. Math. 141 (2005), 1128–1162, https://doi.org/10.1112/S0010437X05001375. Search in Google Scholar

[11] D. Bennis , J. R. García Rozas , and L. Oyonarte , Relative Gorenstein dimensions, Mediterr. J. Math. 13 (2016), 65–91, https://doi.org/10.1007/s00009-014-0489-8. Search in Google Scholar

[12] H. Holm and P. Jørgensen , Semi-dualizing modules and related Gorenstein homological dimensions, J. Pure Appl. Algebra 205 (2006), 423–445, https://doi.org/10.1016/j.jpaa.2005.07.010. Search in Google Scholar

[13] Z. Liu , Z. Huang , and A. Xu , Gorenstein projective dimension relative to a semidualizing bimodule, Comm. Algebra 41 (2013), 1–18, https://doi.org/10.1080/00927872.2011.602782. Search in Google Scholar

[14] D. White , Gorenstein projective dimension with respect to a semidualizing module, J. Commut. Algebra 2 (2010), no. 1, 111–137, https://doi.org/10.1216/JCA-2010-2-1-111. Search in Google Scholar

[15] R. Wisbauer , Cotilting objects and dualities , in: Representations of Algebras, Notes in Pure and Appl. Math. , vol. 224, Dekker, New York, 2002, pp. 215–233. 10.1201/9780429187759-20Search in Google Scholar

[16] J. R. García Rozas , Covers and envelopes in the category of complexes of modules , Research Notes in Mathematics , vol. 407, Chapman & Hall/CRC, Boca-Ratón, FL, 1999. 10.1201/9781315140438Search in Google Scholar

[17] Y. Geng and N. Ding , W-Gorenstein modules, J. Algebra 325 (2011), 132–146, https://doi.org/10.1016/j.jalgebra.2010.09.040 . 10.1016/j.jalgebra.2010.09.040Search in Google Scholar

[18] J. Rada and M. Saorín , Rings characterized by (pre)envelopes and (pre)covers of their modules, Comm. Algebra 26 (1998), 899–912, https://doi.org/10.1080/00927879808826172. Search in Google Scholar

[19] J. Sun , A characterization of Auslander category, Turkish J. Math. 37 (2013), no. 5, 793–805, https://doi.org/10.3906/mat-1110-29 . 10.3906/mat-1110-29Search in Google Scholar

[20] J. R. García Rozas , J. A. López-Ramos , and B. Torrecillas , Semidualizing and tilting adjoint pairs, applications to comodules, Bull. Malays Math. Sci. Soc. 38 (2015), no. 1, 197–218, https://doi.org/10.1007/s40840-014-0013-0. Search in Google Scholar

[21] H. Holm and P. Jørgensen , Cotorsion pairs induced by duality pairs, J. Commut. Algebra 1 (2009), no. 4, 621–633, https://doi.org/10.1216/JCA-2009-1-4-621. Search in Google Scholar

[22] H. Holm , Gorenstein homological dimensions, J. Pure App. Alg. 189 (2004), 167–193, https://doi.org/10.1016/j.jpaa.2003.11.007 . 10.1016/j.jpaa.2003.11.007Search in Google Scholar

[23] G. Zhao and J. Sun , V W -Gorenstein categories, Turk. J. Math. 40 (2016), 365–375, https://doi.org/10.3906/MAT-1502-37. Search in Google Scholar

[24] Z. Huang , Proper resolutions and Gorenstein categories, J. Algebra 393 (2013), 142–169, https://doi.org/10.1016/j.jalgebra.2013.07.008 . 10.1016/j.jalgebra.2013.07.008Search in Google Scholar

[25] A. Xu and N. Ding , Semidualizing bimodules and related Gorenstein homological dimensions, J. Algebra Appl. 15 (2016), no. 10, 165–193, https://doi.org/10.1142/S0219498816501930. Search in Google Scholar

[26] J. Asadollahi , T. Dehghanpour , and R. Hafezi , On the existence of Gorenstein projective precovers, Rend. Sem. Mat. Univ. Padova 136 (2016), 257–264, https://doi.org/10.4171/rsmup/136-16. Search in Google Scholar

[27] J. Xu , Flat Covers of Modules , Lecture Notes in Mathematics , vol. 1634, Springer-Verlag, Berlin, 1996. 10.1007/BFb0094173Search in Google Scholar

[28] D. Bennis , J. R. García Rozas , and L. Oyonarte , Relative projective and injective dimensions, Comm. Algebra 44 (2016), no. 8, 3383–3396, https://doi.org/10.1080/00927872.2015.1065852. Search in Google Scholar

Received: 2020-09-08
Revised: 2021-07-07
Accepted: 2021-07-07
Published Online: 2021-12-09

© 2021 Driss Bennis et al., published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

  1. Regular Articles
  2. Sharp conditions for the convergence of greedy expansions with prescribed coefficients
  3. Range-kernel weak orthogonality of some elementary operators
  4. Stability analysis for Selkov-Schnakenberg reaction-diffusion system
  5. On non-normal cyclic subgroups of prime order or order 4 of finite groups
  6. Some results on semigroups of transformations with restricted range
  7. Quasi-ideal Ehresmann transversals: The spined product structure
  8. On the regulator problem for linear systems over rings and algebras
  9. Solvability of the abstract evolution equations in Ls-spaces with critical temporal weights
  10. Resolving resolution dimensions in triangulated categories
  11. Entire functions that share two pairs of small functions
  12. On stochastic inverse problem of construction of stable program motion
  13. Pentagonal quasigroups, their translatability and parastrophes
  14. Counting certain quadratic partitions of zero modulo a prime number
  15. Global attractors for a class of semilinear degenerate parabolic equations
  16. A new implicit symmetric method of sixth algebraic order with vanished phase-lag and its first derivative for solving Schrödinger's equation
  17. On sub-class sizes of mutually permutable products
  18. Asymptotic solution of the Cauchy problem for the singularly perturbed partial integro-differential equation with rapidly oscillating coefficients and with rapidly oscillating heterogeneity
  19. Existence and asymptotical behavior of solutions for a quasilinear Choquard equation with singularity
  20. On kernels by rainbow paths in arc-coloured digraphs
  21. Fully degenerate Bell polynomials associated with degenerate Poisson random variables
  22. Multiple solutions and ground state solutions for a class of generalized Kadomtsev-Petviashvili equation
  23. A note on maximal operators related to Laplace-Bessel differential operators on variable exponent Lebesgue spaces
  24. Weak and strong estimates for linear and multilinear fractional Hausdorff operators on the Heisenberg group
  25. Partial sums and inclusion relations for analytic functions involving (p, q)-differential operator
  26. Hodge-Deligne polynomials of character varieties of free abelian groups
  27. Diophantine approximation with one prime, two squares of primes and one kth power of a prime
  28. The equivalent parameter conditions for constructing multiple integral half-discrete Hilbert-type inequalities with a class of nonhomogeneous kernels and their applications
  29. Boundedness of vector-valued sublinear operators on weighted Herz-Morrey spaces with variable exponents
  30. On some new quantum midpoint-type inequalities for twice quantum differentiable convex functions
  31. Quantum Ostrowski-type inequalities for twice quantum differentiable functions in quantum calculus
  32. Asymptotic measure-expansiveness for generic diffeomorphisms
  33. Infinitesimals via Cauchy sequences: Refining the classical equivalence
  34. The (1, 2)-step competition graph of a hypertournament
  35. Properties of multiplication operators on the space of functions of bounded φ-variation
  36. Disproving a conjecture of Thornton on Bohemian matrices
  37. Some estimates for the commutators of multilinear maximal function on Morrey-type space
  38. Inviscid, zero Froude number limit of the viscous shallow water system
  39. Inequalities between height and deviation of polynomials
  40. New criteria-based ℋ-tensors for identifying the positive definiteness of multivariate homogeneous forms
  41. Determinantal inequalities of Hua-Marcus-Zhang type for quaternion matrices
  42. On a new generalization of some Hilbert-type inequalities
  43. On split quaternion equivalents for Quaternaccis, shortly Split Quaternaccis
  44. On split regular BiHom-Poisson color algebras
  45. Asymptotic stability of the time-changed stochastic delay differential equations with Markovian switching
  46. The mixed metric dimension of flower snarks and wheels
  47. Oscillatory bifurcation problems for ODEs with logarithmic nonlinearity
  48. The B-topology on S-doubly quasicontinuous posets
  49. Hyers-Ulam stability of isometries on bounded domains
  50. Inhomogeneous conformable abstract Cauchy problem
  51. Path homology theory of edge-colored graphs
  52. Refinements of quantum Hermite-Hadamard-type inequalities
  53. Symmetric graphs of valency seven and their basic normal quotient graphs
  54. Mean oscillation and boundedness of multilinear operator related to multiplier operator
  55. Numerical methods for time-fractional convection-diffusion problems with high-order accuracy
  56. Several explicit formulas for (degenerate) Narumi and Cauchy polynomials and numbers
  57. Finite groups whose intersection power graphs are toroidal and projective-planar
  58. On primitive solutions of the Diophantine equation x2 + y2 = M
  59. A note on polyexponential and unipoly Bernoulli polynomials of the second kind
  60. On the type 2 poly-Bernoulli polynomials associated with umbral calculus
  61. Some estimates for commutators of Littlewood-Paley g-functions
  62. Construction of a family of non-stationary combined ternary subdivision schemes reproducing exponential polynomials
  63. On the evolutionary bifurcation curves for the one-dimensional prescribed mean curvature equation with logistic type
  64. On intersections of two non-incident subgroups of finite p-groups
  65. Global existence and boundedness in a two-species chemotaxis system with nonlinear diffusion
  66. Finite groups with 4p2q elements of maximal order
  67. Positive solutions of a discrete nonlinear third-order three-point eigenvalue problem with sign-changing Green's function
  68. Power moments of automorphic L-functions related to Maass forms for SL3(ℤ)
  69. Entire solutions for several general quadratic trinomial differential difference equations
  70. Strong consistency of regression function estimator with martingale difference errors
  71. Fractional Hermite-Hadamard-type inequalities for interval-valued co-ordinated convex functions
  72. Montgomery identity and Ostrowski-type inequalities via quantum calculus
  73. Universal inequalities of the poly-drifting Laplacian on smooth metric measure spaces
  74. On reducible non-Weierstrass semigroups
  75. so-metrizable spaces and images of metric spaces
  76. Some new parameterized inequalities for co-ordinated convex functions involving generalized fractional integrals
  77. The concept of cone b-Banach space and fixed point theorems
  78. Complete consistency for the estimator of nonparametric regression model based on m-END errors
  79. A posteriori error estimates based on superconvergence of FEM for fractional evolution equations
  80. Solution of integral equations via coupled fixed point theorems in 𝔉-complete metric spaces
  81. Symmetric pairs and pseudosymmetry of Θ-Yetter-Drinfeld categories for Hom-Hopf algebras
  82. A new characterization of the automorphism groups of Mathieu groups
  83. The role of w-tilting modules in relative Gorenstein (co)homology
  84. Primitive and decomposable elements in homology of ΩΣℂP
  85. The G-sequence shadowing property and G-equicontinuity of the inverse limit spaces under group action
  86. Classification of f-biharmonic submanifolds in Lorentz space forms
  87. Some new results on the weaving of K-g-frames in Hilbert spaces
  88. Matrix representation of a cross product and related curl-based differential operators in all space dimensions
  89. Global optimization and applications to a variational inequality problem
  90. Functional equations related to higher derivations in semiprime rings
  91. A partial order on transformation semigroups with restricted range that preserve double direction equivalence
  92. On multi-step methods for singular fractional q-integro-differential equations
  93. Compact perturbations of operators with property (t)
  94. Entire solutions for several complex partial differential-difference equations of Fermat type in ℂ2
  95. Random attractors for stochastic plate equations with memory in unbounded domains
  96. On the convergence of two-step modulus-based matrix splitting iteration method
  97. On the separation method in stochastic reconstruction problem
  98. Robust estimation for partial functional linear regression models based on FPCA and weighted composite quantile regression
  99. Structure of coincidence isometry groups
  100. Sharp function estimates and boundedness for Toeplitz-type operators associated with general fractional integral operators
  101. Oscillatory hyper-Hilbert transform on Wiener amalgam spaces
  102. Euler-type sums involving multiple harmonic sums and binomial coefficients
  103. Poly-falling factorial sequences and poly-rising factorial sequences
  104. Geometric approximations to transition densities of Jump-type Markov processes
  105. Multiple solutions for a quasilinear Choquard equation with critical nonlinearity
  106. Bifurcations and exact traveling wave solutions for the regularized Schamel equation
  107. Almost factorizable weakly type B semigroups
  108. The finite spectrum of Sturm-Liouville problems with n transmission conditions and quadratic eigenparameter-dependent boundary conditions
  109. Ground state sign-changing solutions for a class of quasilinear Schrödinger equations
  110. Epi-quasi normality
  111. Derivative and higher-order Cauchy integral formula of matrix functions
  112. Commutators of multilinear strongly singular integrals on nonhomogeneous metric measure spaces
  113. Solutions to a multi-phase model of sea ice growth
  114. Existence and simulation of positive solutions for m-point fractional differential equations with derivative terms
  115. Bernstein-Walsh type inequalities for derivatives of algebraic polynomials in quasidisks
  116. Review Article
  117. Semiprimeness of semigroup algebras
  118. Special Issue on Problems, Methods and Applications of Nonlinear Analysis (Part II)
  119. Third-order differential equations with three-point boundary conditions
  120. Fractional calculus, zeta functions and Shannon entropy
  121. Uniqueness of positive solutions for boundary value problems associated with indefinite ϕ-Laplacian-type equations
  122. Synchronization of Caputo fractional neural networks with bounded time variable delays
  123. On quasilinear elliptic problems with finite or infinite potential wells
  124. Deterministic and random approximation by the combination of algebraic polynomials and trigonometric polynomials
  125. On a fractional Schrödinger-Poisson system with strong singularity
  126. Parabolic inequalities in Orlicz spaces with data in L1
  127. Special Issue on Evolution Equations, Theory and Applications (Part II)
  128. Impulsive Caputo-Fabrizio fractional differential equations in b-metric spaces
  129. Existence of a solution of Hilfer fractional hybrid problems via new Krasnoselskii-type fixed point theorems
  130. On a nonlinear system of Riemann-Liouville fractional differential equations with semi-coupled integro-multipoint boundary conditions
  131. Blow-up results of the positive solution for a class of degenerate parabolic equations
  132. Long time decay for 3D Navier-Stokes equations in Fourier-Lei-Lin spaces
  133. On the extinction problem for a p-Laplacian equation with a nonlinear gradient source
  134. General decay rate for a viscoelastic wave equation with distributed delay and Balakrishnan-Taylor damping
  135. On hyponormality on a weighted annulus
  136. Exponential stability of Timoshenko system in thermoelasticity of second sound with a memory and distributed delay term
  137. Convergence results on Picard-Krasnoselskii hybrid iterative process in CAT(0) spaces
  138. Special Issue on Boundary Value Problems and their Applications on Biosciences and Engineering (Part I)
  139. Marangoni convection in layers of water-based nanofluids under the effect of rotation
  140. A transient analysis to the M(τ)/M(τ)/k queue with time-dependent parameters
  141. Existence of random attractors and the upper semicontinuity for small random perturbations of 2D Navier-Stokes equations with linear damping
  142. Degenerate binomial and Poisson random variables associated with degenerate Lah-Bell polynomials
  143. Special Issue on Fractional Problems with Variable-Order or Variable Exponents (Part I)
  144. On the mixed fractional quantum and Hadamard derivatives for impulsive boundary value problems
  145. The Lp dual Minkowski problem about 0 < p < 1 and q > 0
Downloaded on 13.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/math-2021-0101/html
Scroll to top button