In this paper, we investigated the generalized lower triangular matrix algebra, and gave the sufficient and necessary condition for the generalized lower triangular matrix algebra to be quasi-tilted.
Citation: Xiu-Jian Wang, Jia-Bao Liu. Quasi-tilted property of generalized lower triangular matrix algebras[J]. Electronic Research Archive, 2025, 33(5): 3065-3073. doi: 10.3934/era.2025134
[1] | Yongjie Wang, Nan Gao . Some properties for almost cellular algebras. Electronic Research Archive, 2021, 29(1): 1681-1689. doi: 10.3934/era.2020086 |
[2] | Xiuhai Fei, Haifang Zhang . Additivity of nonlinear higher anti-derivable mappings on generalized matrix algebras. Electronic Research Archive, 2023, 31(11): 6898-6912. doi: 10.3934/era.2023349 |
[3] | 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 |
[4] | Rongmin Zhu, Tiwei Zhao . The construction of tilting cotorsion pairs for hereditary abelian categories. Electronic Research Archive, 2025, 33(5): 2719-2735. doi: 10.3934/era.2025120 |
[5] | Xiuhai Fei, Cuixian Lu, Haifang Zhang . Nonlinear Jordan triple derivable mapping on $ * $-type trivial extension algebras. Electronic Research Archive, 2024, 32(3): 1425-1438. doi: 10.3934/era.2024066 |
[6] | Kailash C. Misra, Sutida Patlertsin, Suchada Pongprasert, Thitarie Rungratgasame . On derivations of Leibniz algebras. Electronic Research Archive, 2024, 32(7): 4715-4722. doi: 10.3934/era.2024214 |
[7] | Ming Ding, Zhiqi Chen, Jifu Li . The properties on F-manifold color algebras and pre-F-manifold color algebras. Electronic Research Archive, 2025, 33(1): 87-101. doi: 10.3934/era.2025005 |
[8] | Hanpeng Gao, Yunlong Zhou, Yuanfeng Zhang . Sincere wide $ \tau $-tilting modules. Electronic Research Archive, 2025, 33(4): 2275-2284. doi: 10.3934/era.2025099 |
[9] | Quanguo Chen, Yong Deng . Hopf algebra structures on generalized quaternion algebras. Electronic Research Archive, 2024, 32(5): 3334-3362. doi: 10.3934/era.2024154 |
[10] | Hongliang Chang, Yin Chen, Runxuan Zhang . A generalization on derivations of Lie algebras. Electronic Research Archive, 2021, 29(3): 2457-2473. doi: 10.3934/era.2020124 |
In this paper, we investigated the generalized lower triangular matrix algebra, and gave the sufficient and necessary condition for the generalized lower triangular matrix algebra to be quasi-tilted.
Throughout this paper, we denote a field by k, an algebra R means a finite dimensional algebra over k, modR denotes the category of all finitely generated right R-modules, while indR denotes a full subcategory of modR containing one representative of each isomorphism class of indecomposable R-modules. We will use freely Auslander-Reiten translation τ=DTr, irreducible maps and properties of the Auslander-Reiten sequences. For an R-module X, the injective hull of X will be denoted by E(X), the projective (resp. injective) dimension of X will be denoted by pdRX (resp. idRX), and the global dimension of R will be denoted by gl.dimR.
As a proper generalization of tilted algebras, Happel has introduced quasi-tilted algebras in [1]. An algebra R is called quasi-tilted if it satisfies the two conditions: (ⅰ) If U is an indecomposable R-module, then pdRU≤1 or idRU≤1, and (ⅱ) gl.dimR≤2.
Let R and S be algebras and N be a finite dimensional left-S, right-R bimodule over k, and set
T=(R0NS)={(r0ns)∣r∈R,n∈N,s∈S}. |
Meanwhile, the addition and multiplication are defined as follows:
(r10n1s1)+(r20n2s2)=(r1+r20n1+n2s1+s2), |
(r10n1s1)(r20n2s2)=(r1r20n1r2+s1n2s1s2). |
Then, T is a finite-dimensional k-algebra, and T is called a generalized lower triangular matrix algebra.
It follows from a result of [1] that every quasi-tilted artin algebra is isomorphic to a generalized lower triangular matrix algebra (R0NS), where R and S are hereditary. So to find all quasi-tilted algebras, it is enough to investigate when such generalized lower triangular matrix algebras are quasi-tilted.
The conditions for the one-point extension algebra of a quasi-tilted algebra to be again quasi-tilted were given by Huard [2] and Coelho [3,4]. Recently, [5] has obtained a new proof concerning quasi-tilted algebras. In [6], Assem characterized cluster-tilted and quasi-tilted algebras. For more results, one can see [7,8,9].
The construction of the generalized lower triangular matrix ring or algebra is a major issue in ring and algebraic representation theory, and it has been often used to construct counterexamples in rings and modules. Following [10], Gorenstein conditions over formal triangular matrix rings are studied by Enochs, and Gorenstein-projective modules over triangular matrix Artin algebras are also studied by Xiong and Zhang in [8]. As for more examples of generalized lower triangular matrix algebras, we refer to [7,11,12] and so on.
Based on the above discussion, it is meaningful to discuss quasi-tilted property of generalized lower triangular matrix algebra. In the present paper, we consider the problem of characterizing when a generalized lower triangular matrix algebra is quasi-tilted.
It is widely known that modT is equivalent to a category Ω of triples (U,W,φ), where U∈modR,W∈modS and φ:W⊗SN→U is a map in modR. The right T-module which corresponds to (U,W,φ) is the additive group U⊕W with right T-action given by:
(u,w)(r0ns)=(ur+φ(w⊗n),ws). |
We consider the right T-modules as triples (U,W,φ). Given (U1,W1,φ1) and (U2,W2,φ2), a morphism f:(U1,W1,φ1)⟶(U2,W2,φ2) is a pair (f1,f2), where f1:U1→U2 is homomorphism in modR and f2:W1→W2 is homomorphism in modS; furthermore, we can get that
W1⊗SNφ1⟶U1f2⊗1↓ ↓f1W2⊗SNφ2⟶U2 |
is a commutative diagram. If f=(f1,f2):(U1,W1,φ1)⟶(U2,W2,φ2) is a homomorphism map, then we have f is surjective (resp. injective) if and only if f1:U1→U2 and f2:W1→W2 are surjective (resp., injective).
It is well-known that the right T-module (U,W,φ) is a projective module if, and only if, (U/φ(W⊗N))R and WS are projective modules and φ:W⊗N⟶U is one-one. So, we can get the following.
The T-module (U,W,φ) is a projective module if, and only if, WS is a projective module, φ:W⊗N⟶U is monic, and U=P⊕φ(W⊗N), where PR is projective.
For a right T-module (U,W,φ), define
˜φ:W⟶HomR(N,U) |
given by ˜φ(w)(n)=φ(w⊗n) for n∈N,w∈W. Then, we can get ˜φ is an S-homomorphism.
The right T-module (U,W,φ) is an injective module if, and only if, UR is an injective module and the map Φ:W⟶HomR(N,U)⊕E(ker˜φ) given by Φ(w)=(˜φ(w),τ(w)) is an isomorphism of right S-modules, where τ:W→E(ker˜φ) is an extension of the inclusion ker˜φ↪E(ker˜φ).
We know that a right T-module (UR,0,0) is a projective module if, and only if, UR is projective, hence, pdT(UR,0,0)=pdRUR. A right T-module (UR,0,0) is an injective module if, and only if, UR is an injective module and HomR(N,U)=0. A right T-module (0,WS,0) is a projective module if, and only if, WS is projective and W⊗N=0. A right T-module (0,WS,0) is injective if, and only if, WS is injective, hence, idT(0,WS,0)=idSWS.
In this section, we consider the problem of characterizing when a generalized lower triangular matrix algebra is quasi-tilted.
Suppose T=(R0NS) is a generalized lower triangular matrix algebra, while Goodearl [2] gave a condition for T to be hereditary.
Theorem 1 [13]. T is a hereditary ring if, and only if,
(1) R and S are both hereditary.
(2) SN is a projective module.
(3) (N/IN)R is projective for all I≤SS.
We know HomR(W⊗SN,U)σ≅HomS(W,HomR(N,U)); for φ:W⊗N⟶U, let φ′:W⟶HomR(N,U) be the map corresponding to φ, that is, σ(φ)=φ′. So, we can get the following result.
Theorem 2. Suppose T=(R0NS),(U,W,φ) is a right T-module such that φ′ is a monomorphism. Then, the following conditions are equivalent.
(1) idT(U,W,φ)≤1.
(2) There exists an exact sequence 0⟶W⟶HomR(N,I0)⟶HomR(N,I1)⊕I2⟶0 with I0,I1,I2 injective R-modules and idRU≤1.
Proof: Since HomR(W⊗SN,U)σ≅HomS(W,HomR(N,U)), we can view (U,W,φ) as the triple (U,W,φ′).
Suppose
0⟶Ui⟶I0(U)⟶I1(U) |
is a minimal injective resolution of U; thus, we can get
HomR(N,U)HomR(N,i)⟶HomR(N,I0(U)). |
Set φ″=HomR(N,i)∘φ′, then φ″ is a monomorphism because φ′ is a monomorphism. We have that
(U,W,φ′)(i,φ″)⟶(I0(U),HomR(N,I0(U)),1) |
is an injective envelope of T-module (U,W,φ′).
Then, we can get the following diagram:
0⟶Wφ″⟶HomR(N,I0(U))→HomR(N,I0(U))/φ″(W)→0φ′↓||↓g0→HomR(N,U)→HomR(N,I0(U))→HomR(N,I0(U)/i(U)). |
So, Coker(i,φ″)=(I0(U)/i(U),HomR(N,I0(U))/φ″(W),g). The injective envelope of Coker(i,φ″) is
(I1(U),HomR(N,I1(U)),1)⊕(0,E(I),0), |
where I=kerg≅Cokerφ′. Set F=HomR(N,−), hence, idT(U,W,φ)≤1 if, and only if,
0→(U,W,φ′)(i,φ″)⟶(I0(U),F(I0(U)),1)→(I1(U),F(I1(U)),1)⊕(0,E(I),0)→0 |
is exact if, and only if,
0⟶U⟶I0(U)⟶I1(U)⟶0 |
and
0⟶W⟶F(I0(W))⟶F(I1(U))⊕E(I)⟶0 |
are exact.
Given an algebra C, suppose LC denotes the subset of indC defined by
LC={U∈indC|pdCW≤1for each predecessor Wof U}. |
Also, we denote the full subcategory of LC of direct summands of finite direct sums of copies of LC by addLC.
We now give the following necessary conditions for generalized lower triangular matrix algebra R to be quasi-tilted.
Theorem 3. Let T=(R0NS). If T is quasi-tilted, then
(1) Both R and S are quasi-tilted.
(2) NR∈addLR.
Proof: Since T is quasi-tilted, we have gl.dimT≤2, and it follows from [3] that gl.dimR≤2,gl.dimS≤2.
(1) If S is not quasi-tilted, then there is an indecomposable R-module WS with pdSW=2=idSW. Since idT(0,WS,0)=idSW=2 and T is quasi-tilted, we know that pdT(0,WS,0)≤1. Assume that π:P→WS is a projective cover of WS, let ZS be the kernel of π, then (0,π):(P⊗N,P,1)→(0,WS,0) is a projective cover of (0,WS,0) with kernel (P⊗N,ZS,g). Since pdT(0,WS,0)≤1, we know that (P⊗N,ZW,g) is a projective T-module, and then ZW is a projective S-module, thus, pdSW≤1. This contradicts that pdSW=2, so S is quasi-tilted.
If R is not quasi-tilted, then there exists an indecomposable R-module UR with pdRU=2=idRU. Since pdT(UR,0,0)=pdRU=2 and T is quasi-tilted, we know idT(UR,0,0)≤1, and by Theorem 2, idRU≤1. This contradicts that idRU=2, so R is quasi-tilted.
(2) By (1), R is quasi-tilted. If some indecomposable summand N1 of N is not in LR, then, by definition of LR, there exists some indecomposable predecessor UR of N1 such that pdRU=2. Then, pdT(U,0,0)=2, and by [4], there exists an indecomposable injective T-module L such that HomT(L,τ(U,0,0))≠0. Thus, we have a sequence
L→τ(U,0,0)→∗→(U,0,0)→⋯→(N1,0,0)→(N1,S,1) |
of nonzero maps between indecomposable T-modules starting with an injective module and ending with a projective module. Because no refinement of the path can be sectional, we have a contradiction to T being quasi-tilted. So, MR∈addLR.
Corollary 4 [1]. Suppose k is a field, and T=R[N]=(R0Nk) is the one-point extension of R by N. If T is quasi-tilted, then
(1) R is quasi-tilted.
(2) N∈addLR.
Next, we will investigate what conditions need to be added to ensure that T=(R0NS) is quasi-tilted.
To investigate when T=(R0NS) is quasi-tilted, we first investigate when gl.dimT≤2. By [1], we have:
Lemma 5. Let R and S be finite dimensional algebras over a field k and N a S-R bimodule finite dimensional over k. Let T=(R0NS). Then, we can get
gl.dimT=max{gl.dimR,gl.dimS,pdS⊗kRopN+1}. |
From this, we get that gl.dimT≤2 if, and only if, gl.dimR≤2 and gl.dimS≤2 and pdS⊗kRopN≤1. From Lemma 5, we have
Theorem 6. Let T=(R0NS),(U,W,φ) be a right T-module. Then,
pdT(U,W,φ)≤sup{pdRU,pdT(0,W,0)}. |
If WS is a projective module, then
pdT(U,W,φ)≤sup{pdRU,pdRN+1}. |
Proof: Consider the following exact sequence:
0⟶(U,0,0)(1U,0)⟶(U,W,φ)(0,1W)⟶(0,W,0)⟶0 |
in mod T. Then, it follows from a well-known result that
pdT(U,W,φ)≤sup{pdT(U,0,0),pdT(0,W,0)}. |
Since pdT(U,0,0)=pdRU, then
pdT(U,W,φ)≤sup{pdRU,pdT(0,W,0)}. |
If WS is a projective module, then we can get the following exact sequence:
0⟶(W⊗N,0,0)(1,0)⟶(W⊗N,W,1)(0,1)⟶(0,W,0)⟶0 |
in mod T. Because (W⊗N,W,1) is projective, we can get that
pdT(0,W,0)=pdT(W⊗N,0,0)+1=pdRW⊗N+1≤pdRN+1. |
In the equation above, suppose WS is not projective, and
π:QS⟶WS⟶0 |
is a projective cover of WS with the first syzygy ZS, then we have the projective cover
(0,π):(Q⊗N,QS,1)⟶(0,WS,0)⟶0 |
with the first syzygy (Q⊗N,ZS,η⊗1), where η:ZS↪QS is the inclusion map. Hence,
pdT(0,W,0)=pdT(Q⊗N,ZS,η⊗1)+1. |
Corollary 7 [2]. Suppose k is a field and T=R[N]=(R0Nk) is the one-point extension of R by N. If (U,ks,φ) is a right T-module, then
pdT(U,ks,φ)≤sup{pdRU,pdRN+1}. |
Corollary 8. Let T=(R0NS). If S is semi-simple and NR is projective, then pdT(U,W,φ)≤sup{pdRU,1} for any T-module (U,W,φ).
Corollary 9 [13]. Let T=(R0NS). If R is hereditary, NR is projective, and S is semi-simple, then T is hereditary.
From Theorem 6, we also have
Corollary 10. Let T=(R0NS). If gl.dimR≤2 and S is semi-simple and pdRN≤1, then gl.dimT≤2. Proof: By Theorem 6, for any T-module (U,W,φ), we can get
pdT(U,W,φ)≤sup{pdRU,pdRN+1}≤2. |
Theorem 11. Let T=(R0NS) with gl.dimR≤2. For any T-module (U,W,φ), if kerφ is not projective, then pdT(U,W,φ)≥2.
Proof: Let ε2:H→W be a projective cover, and δ1:P→U/ϕ(W⊗N) is a projective cover. Set θ=φ∘(ε2⊗1N):H⊗N→U. Let η:U→U/φ(W⊗N) be a natural homomorphism, ε1:(H⊗N)⊕P→U such that ε1|H⊗N=θ,ε1|P=γ1, where γ1:P→U such that δ1=η∘γ1. We have a projective cover
((H⊗N)⊕P,H,j)ε=(ε1,ε2)⟶(U,W,φ). |
According to the above discussion, we can get the following commutative diagram:
kerε2⊗Nσ⊗1N⟶H⊗Nε2⊗1N⟶W⊗N⟶0ψ↓(1,0)↓φ↓0⟶K⟶(H⊗N)⊕Pε1⟶U⟶0. |
According to the snake lemma, it is easy to get the two exact sequences
0⟶kerf⟶cokerψ⟶P⟶cokerφ⟶0 |
and
0⟶kerφ⟶cokerψ⟶Ω1(cokerφ)⟶0. |
We know that pdT(U,W,φ)≤1 if, and only if, (K,kerε2,φ) is projective. Since gl.dimR≤2, we have pdRcokerφ≤2. Hence, kerφ is projective. Therefore, if kerφ is not projective, then pdT(U,W,φ)≥2.
From the above proof, if R is a hereditary ring, then we can get that pdT(U,W,φ)≤1 if, and only if, kerφ is projective.
A sufficient condition for T to be quasi-tilted is given by the following theorem.
Theorem 12. Let T=(R0NS). If R is quasi-tilted, S is semi-simple and NR is projective, then T is quasi-tilted.
Proof: According to Corollary 10, gl.dimT≤2. Let (U,W,φ) be an indecomposable T-module, then W is a semi-simple S-module and U is an indecomposable R-module. Because R is quasi-tilted, then pdRU≤1 or idRU≤1.
If pdRU≤1, then, by Theorem 6, pdT(U,W,φ)≤1.
Assume that idRU≤1 and 0→Ui→I0j→I1→0 is the minimal injective resolution of U, then
0→HomR(N,U)i∗→HomR(N,I0)j∗→HomR(N,I1)→0 |
is exact since NR is projective.
Since HomR(Y⊗SN,U)σ≅HomS(W,HomR(N,U)), for φ:W⊗N⟶U, let φ′:W⟶HomR(N,U) be the map corresponding to φ, that is, σ(φ)=φ′. We can then view (U,W,φ) as the triple (U,W,φ′) where φ′:W⟶HomR(N,U) is a S-module homomorphism. Hence, (U,W,φ′) is an indecomposable T-module. If φ′ is not a monomorphism, then φ′=0 and (U,W,φ′)=(0,W,0), hence, idT(U,W,φ′)<1. Thus, we can assume that φ′ is a monomorphism. Set φ″=i∗∘φ′, then φ″ is a monomorphism. We know that
(U,W,φ′)(i,φ″)⟶(I0,HomR(N,I0),1) |
is an injective envelope of T-module (U,W,φ′).
Furthermore, we have the following commutative diagram:
0→Wφ″→HomR(N,I0)→HomR(N,I0)/φ″(W)→0φ′↓||g↓0→HomR(N,U)→HomR(N,I0)→HomR(N,I1)→0. |
By the Five Lemma in homological algebra, g is epic.
We know that Coker(i,φ″)=(I1,HomR(N,I0)/φ″(W),g), and the injective envelope of Coker(i,φ″) is (I1,HomR(N,I1),1)⊕(0,kerg,0).
We now show that HomR(N,I0)/φ″(W)≅HomR(N,I1)⊕kerg.
Since j∗ and g are epics, we have
HomR(N,I1)≅HomR(N,I0)/kerj∗, |
HomR(N,I1)≅HomR(N,I0)/φ″(W)/kerg. |
Therefore,
HomR(N,I0)/φ″(Y)≅HomR(N,I0)/kerj∗+kerg≅HomR(N,I1)⊕kerg. |
So id(U,W,φ′)≤1 and T is quasi-tilted.
From Theorems 3 and 12, we have:
Corollary 13 [2]. Let R be an algebra with gl.dimR≤2,N a projective R-module, and let T=R[N] be the one-point extension of R by N. Then, T is quasi-tilted if, and only if, R is quasi-tilted.
The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.
We would like to thank the referees for the very helpful comments and suggestions which improved the original version of this paper. This work is supported by the Natural Science Foundation for Yang Scholars of Anhui Province (No.1508085QA04). The authors are grateful to professor Xianneng Du for careful guidance and assistance in this research.
The authors declare there is no conflicts of interest.
[1] | D. Happel, I. Reiten, S. O. Smalϕ, Tilting in abelian categories and quasi-tilted algebras, Am. Math. Soc., 120 (1996), 575. |
[2] |
F. Huard, One-point extensions of quasi-tilted algebras by projectives, Commun. Algebra, 29 (2001), 3055–3060. https://doi.org/10.1081/AGB-5005 doi: 10.1081/AGB-5005
![]() |
[3] |
F. U. Coelho, M. I. R. Martins, J. A. de la Pena, Quasitilted extensions of algebras Ⅰ, Proc. Am. Math. Soc., 129 (2000), 1289–1297. https://doi.org/10.1090/S0002-9939-00-05667-7 doi: 10.1090/S0002-9939-00-05667-7
![]() |
[4] |
F. U. Coelho, M. I. R. Martins, J. A. de la Pena, Quasitilted extensions of algebras Ⅱ, J. Algebra, 227 (2000), 582–594. https://doi.org/10.1006/jabr.1999.8199 doi: 10.1006/jabr.1999.8199
![]() |
[5] | S. Zito, A new proof concerning quasi-tilted algebras, Proc. Am. Math. Soc., 147 (2019), 2757–2760. |
[6] |
I. Assem, R. Schiffler, K. Serhiyenko, Cluster-tilted and quasi-tilted algebras, J. Pure Appl. Algebra, 221 (2016), 2266–2288. https://doi.org/10.1016/j.jpaa.2016.12.008 doi: 10.1016/j.jpaa.2016.12.008
![]() |
[7] |
H. Gao, Z. Huang, Silting modules over triangular matrix rings, Taiwanese J. Math., 24 (2020), 1417–1437. https://doi.org/10.11650/tjm/200204 doi: 10.11650/tjm/200204
![]() |
[8] |
B. L. Xiong, P. Zhang, Gorenstein-projective modules over triangular matrix Artin algebras, J. Algebra Appl., 11 (2012), 125006. https://doi.org/10.1142/S0219498812500661 doi: 10.1142/S0219498812500661
![]() |
[9] |
P. Zhang, Monomorphism categories, cotilting theory and Gorenstein projective modules, J. Algebra, 339 (2011), 181–202. https://doi.org/10.1016/j.jalgebra.2011.05.018 doi: 10.1016/j.jalgebra.2011.05.018
![]() |
[10] |
E. E. Enochs, M. Cortes-Izurdiaga, B. Torrecillas, Gorenstein conditions over triangular matrix rings, J. Pure Appl. Algebra, 218 (2014), 1544–1554. https://doi.org/10.1016/j.jpaa.2013.12.006 doi: 10.1016/j.jpaa.2013.12.006
![]() |
[11] | H. Eshraghi, R. Hafezi, S. Salarian, Z. W. Li, Gorenstein projective modules over triangular matrix rings, Algebra Colloq., 23 (2016), 97–104. |
[12] |
L. X. Mao, Duality pairs and FP-injective modules over formal triangular matrix rings, Commun. Algebra, 48 (2020), 5296–5310. https://doi.org/10.1080/00927872.2020.1786837 doi: 10.1080/00927872.2020.1786837
![]() |
[13] | K. R. Goodearl, Ring Theory, Non Singular Rings and Modules, Marcel Dekker, New York and Basel, 1976. |