关于相对平坦性与相对凝聚性的注记
- 格式:pdf
- 大小:265.61 KB
- 文档页数:9
一相对平坦且面积较大的面作为基准面〔即底面〕,然后用游标卡尺逐一测量该集料颗粒的厚度〔即底面到颗粒的最高点t 〕,长度l ,将l/t ≥3的颗粒挑出,判定为针片状或片状颗粒的总质量m1〔6〕Qe 〔%〕=(m1 / m0 )× 100 阻碍因素:由于沥青路面对粗集料针片状颗粒要求的更为严格,不能用规准仪替代游标卡尺法判定。
采纳规准仪时,第一要通过标准筛将粗集料进行分级,不同粒级的颗粒要对应孔宽和间距来判定,不可错位。
采纳游标卡尺时,第一要确定好颗粒基准面,然后再测量其厚度和长度等相应尺寸。
8 2.0mm 或2.5mm 筛,可假定水泥含水量为0% 2.混合料组成的运算预备。
3.预备5种试样,每种2个样品,水泥计量分别为相差2%每份均重300g 分别放在搪瓷杯内,水泥土混合料的最正确含水量应等于工地预期达到的最正确含水量。
混合料中所加的水应与工地所用的水相同。
取一个盛有试样的搪瓷杯,在杯内加600mL10%氯化按溶剂,用不锈钢搅拌棒充分搅拌3min 〔每分钟搅110-120次〕。
如水泥土混合料中的土是细粒土,那么也能够用1000 mL 具塞三角瓶代替搪瓷杯,手握三角瓶〔瓶口向上〕用力振荡3min(每分钟120次±5次〕,以代替搅拌棒搅拌,放置沉淀4min[如4min 后得到的是混浊悬浮液,那么应增加放置沉淀时刻,直到显现澄清悬浮液为止,并记录所需的时刻,以后所有该种水泥〔或石灰〕土混合料的试验,均应以同一时刻为准],然后将上部清液转移到300mL烧杯内,搅匀,加盖表面皿待测。
4.用移液管吸取上层〔液面下1-2cm 〕悬浮液10.0mL 放人200mL 的三角瓶内,用量筒量取500mL1.8%氢氧化钠〔内含三乙醇胺〕倒人三角瓶中,现在溶液出值为12.5-13.0〔可用pH12-pH14周密试纸检验〕,然后加入钙红指示剂〔体积约为黄豆大小〕,摇匀,溶剂呈玫瑰红色。
用EDTA 二钠标准液滴定到纯蓝色为终点,记录EDTA 二钠的耗量〔以mL 计,读至0.1mL 〕。
高速纺丝对聚酯质量和切片干燥的要求1聚酯质量的要求聚酯纺丝有熔体直接纺丝和切片纺丝两种,不同的聚酯熔体和切片对纺丝和成品的质量影响极大。
纺丝情况及POY的质量不但与聚酯的相对分子质量及其分布、熔体的流变特性和切片的热容量等有关,而且与切片的凝聚粒子含量、聚合时加入催化剂的沉淀物、灰分和其他机械杂质的含量以及所加TiO2的特性等有关。
纺丝工艺不同,会使纺丝情况不同,对原料的要求也不同。
高速纺丝对聚酯质量有如下要求:(1)聚酯中机械杂质及凝聚粒子的含量愈低愈好,熔体特性粘度的波动值最好小于0.01,其中心值在0.63~0.68之间,以偏高为佳。
较高的特性粘度有利于制得良好的POY,但过高会造成纺丝困难和毛丝增多。
(2)聚酯的相对分子质量分布窄,分布指数α小(α<>α大者,纺丝成形不良,产生飘丝和并丝,疵点多,无油丝粘度波动大,纤维的强伸度波动大。
聚酯的相对分子质量高,在纺程上可承受大的拉力,对纺丝有利。
但聚酯相对分子质量太高时,大分子链太长,难以展开和伸直,使分子取向所需要的力也大,并可能导致不完全取向。
聚酯相对分子质量太低时,则大分子链短,从喷丝孔挤出和拉伸取向时,在张力作用下易产生断裂。
故其平均相对分子质量应适中。
聚酯的相对分子质量在很大程度上决定着纤维的性能,同时对纺丝工艺条件有很大的影响。
最佳的相对分子质量范围应选择在对纺丝工艺条件和产品质量最不敏感的区域。
(3)聚酯熔体的过滤性能好。
描述和判定聚酯熔体的过滤性能,可用在过滤面积S(m2)上通过一定时间G(min)的熔体的平均压力降△P来表示,其值A称为过滤系数,表示如下:如果A值小,表示过滤性能好。
过滤性能好的切片,预过滤器的出口起始压力有一个比较稳定的阶段,然后慢慢下降;过滤性能差的,没有稳定阶段,其压力迅速下降,甚至呈直线下降。
(4)切片中的粉屑含量少。
若切片中粉屑含量较多,纺丝时喷丝板粘板严重,新使用的喷丝板8~12h后就会出现粘板现象,从而造成纺丝成形恶化,甚至产生注头或块状疵点,纺丝组件使用寿命缩短,侧吹风窗上会积满粉尘,影响冷却吹风的风速及分布的均匀性,从而使POY质量低劣。
第一章原子结构与键合此章主要掌握概念1.金属键(1)典型金属原子结构的特点是其最外层电子数很少,极易挣脱原子核的束缚而成为自由电子,并在整个晶体内运动,及弥漫于金属正离子组成的晶格之中而形成电子云.这种由金属中的自由电子与金属正离子相互作用所构成的键合为金属键。
(2)绝大多数金属均以金属键方式结合,它的基本特点是电子的共有化,而且无方向性,无饱和性。
(3)金属一般具有良好的到点和导热,以及良好的延展性的原因:自由电子的存在。
2。
离子键大多数盐类、碱类和金属氧化物主要以离子键的方式结合,靠静电引力结合在一起。
3。
共价键共价键是由两个或多个电负性相差不大的原子通过共用电子对而形成的化学键。
共价键又分为非极性键和极性键两种。
有方向性和饱和性。
4.范德瓦耳斯力是借助这种微弱的、瞬时的电偶极矩的感应作用,将原来具有稳定的原子结构的原子或分子结合为一体的键合,没有方向性和饱和性.5.氢键氢键是一种极性分子键,存在于HF、H2O、NF3等分子间,它的键能介于化学键与范德瓦耳斯力之间。
第二章固体结构重点:晶面指数和晶向指数、配位数以及致密度等一些概念、合金相结构的几种类型、间隙固溶体和间隙化合物和间隙相的异同点.主要是简答题按照原子或分子排列的特征可将固态物质分为两大类:晶体和非晶体.1.晶体结构的基本特征是,原子或分子或离子在三维空间呈周期性重复排列,即存在长程有序。
(各向异性)2.空间点阵:阵点在空间呈周期性规则排列,并具有完全相同的周围环境,这种由它们在三维空间规则排列的阵列称为空间点阵。
(概念题)3.晶格:为了便于描述空间点阵的图形,可用许多平行的直线将所有阵点连接起来,构成了一个三维几何架子。
(概念题)4。
晶胞:可在点阵中取出一个具有代表性的基本单元,作为点阵的组成单元。
(概念题)5.选取晶胞的原则:a。
选取的平行六面体应反映出点阵的最高对称性。
b。
平行六面体内的棱和角相等的数目应最多.c。
当平行六面体的棱边夹角存在直角时,直角数目应最多。
凝聚态导论知识点总结一、凝聚态物质概述凝聚态物质是指在常温常压下存在的物质状态,包括固体和液体。
固体具有固定的形状和体积,分子之间的距离较小,分子排列有序;液体具有固定的体积但没有固定的形状,分子之间的距离较大,分子排列无序。
二、固体的结构和性质1. 晶体结构:晶体是由具有三维周期性排列的原子、离子或分子构成的固体。
晶体的结构可以通过晶体学来描述,常见的晶体结构有立方晶系、四方晶系、六方晶系等。
2. 晶格常数:晶体的晶格常数是描述晶体结构的重要参数,它指的是晶胞的尺寸,常用a、b、c表示。
3. 晶体缺陷:晶体中存在各种缺陷,如点缺陷、线缺陷和面缺陷等。
这些缺陷会对晶体的性质和行为产生重要影响。
4. 固体的力学性质:固体的力学性质包括弹性性质、塑性性质和断裂性质等。
这些性质与固体的内部结构和原子间相互作用密切相关。
三、液体的结构和性质1. 流动性:液体具有流动性,这是由于分子之间的相互作用较弱,分子可以相对自由地移动。
2. 表面张力:液体的表面上的分子受到内部分子的吸引力,使得液体表面呈现紧致的状态,形成表面张力。
3. 黏度:液体的黏度是指液体流动时所表现出的阻力大小,与液体的粘性有关。
4. 液晶:液晶是介于液体和固体之间的一种物质状态,具有介于有序固体和无序液体之间的特性。
四、相变和相图1. 相变:物质在一定条件下可以从一种状态转变为另一种状态,这种转变称为相变。
常见的相变有固态到液态的熔化、液态到气态的汽化等。
2. 相图:相图是描述物质在不同条件下各相存在的范围和相变的温度、压力关系的图形表示。
常见的相图有水的三相图和铁的铁-铁碳相图等。
五、凝聚态物理中的重要概念和现象1. 电子能带理论:电子能带理论描述了固体中电子的能级分布情况,解释了固体的导电性和绝缘性等现象。
2. 超导现象:超导是指某些物质在低温下具有零电阻和完全排斥外磁场的特性。
超导现象在科学和工程领域有重要应用。
3. 磁性现象:磁性是物质在外磁场下表现出的各种现象,包括顺磁性、抗磁性和铁磁性等。
- 1 -A Note on Relative Flatness and Coherence 1Xiaoxiang Zhang Jianlong ChenDepartment of Mathematics Southeast University ,Nanjing 210096E-mail: z990303@AbstractLet R be a ring and M a fixed right R -module. A new characterization of M -flatness is given by certain linear equations. For a left R -module F such that the canonical map M ⊗R F →Hom R (M *, F ) is injective, where M * = Hom R (M , R ), the M -flatness of F is characterized via certain matrix subgroups. An example is given to show that R need not be M -coherent even if every left R -module is M -flat. Moreover, some properties of M -coherent rings are discussed. Keywords: M -flat module, matrix subgroup ,M -coherent ring1 IntroductionsRecently, Dauns [2] introduced the notion of coherence of a ring R relative to a right R -module M .Let R be a ring, M a fixed right R -module and σ[M ] the full subcategory of the category of right R -modules subgenerated by M (see [6, p. 118]). Recall from [2] that a left R -module F is σ[M ]-flat if for any exact sequence 0 → X → Y in σ[M ], the sequence 0 → X ⊗R F → Y ⊗R F is exact. Following Wisbauer [6], F is called M -flat if the sequence 0 → K ⊗R F → M ⊗R F is exact for every submodule 0 ≤ M . Dauns [2, Proposition 1.6] proved that R F is M -flat if and only if it is σ[M ]-flat.Following Dauns [2], a right R -module N is M -coherent if for any 0 ≤ A < B ≤ N such that B /A ⊆ mR for some m ∈ M , if B /A is finitely generated, then B /A is finitely presented. R is defined to be M -coherent if the right R -module R R is M -coherent.The purpose of this note is to investigate M -flat modules and M -coherent rings from some new aspects.Throughout R is an associative ring with identity and all modules are unitary. For a positive integer n , R n (resp. R n ) denotes the direct sum of n copies of R R (resp. R R ) whose elements are written as “row (resp. column) vectors”. Similarly, M n stands for the direct sum of n copies of M R . For each m = (m 1, m 2, …, m n ) ∈ M n , the right annihilator of m in R n is symbolized by)(m n R r , that is,)(m n R r = {(r 1, r 2, …, r n )T∈ R n | (m 1, m 2, …, m n ) (r 1, r 2, …, r n )T=∑=ni i i r m 1= 0}.Note that )(m n R r is nothing more than m ⊥ in [2] for all m ∈ M . Given an R -module P , the dual module and the character module are denoted by P * and P + respectively, i.e., P * = Hom R (P , R ) and P += Hom Z (P , Q /Z ), where Q (resp. Z ) is the ring of rational numbers (resp. integers).1This work is supported in part by the NSFC of China (No. 10571026) and the Foundation of Graduate Creative Program of JiangSu (xm04-10) and the Teaching and Research Award Program for Outstanding Young Teachers in Higher Education Institutions of MOE, P.R.China.- 2 -2 Main resultsLet us start with the following result.Theorem 1. Let M be a fixed right R -module . The following are equivalent for a left R -module F .(1) F is M -flat .(2) For any a 1, …, a n ∈ R , and x 1, …, x n ∈ F such thatF m x a R ni ii )(1r ∈∑=for some m ∈ M ,there exist y 1, …, y k ∈ F and n ×k matrix B = (b ij )n ×k over R such that x i = ∑=kj j ijy b 1 for each 1 ≤ i ≤n and m∑=ni iji b a 1= 0 for each 1 ≤ j ≤ k .Proof . (1) ⇒ (2). For any a 1, …, a n ∈ R , and x 1, …, x n ∈ F , ifF m x a R ni ii )(1r ∈∑=, we haver R (m ) ⊆ L = a 1R + … + a n R + r R (m ) ≤ R Rwith L/r R (m ) finitely generated and.)(/)(0)(1F m LF F m F m x a R R R ni ii r r r ∈+=+∑=It follows by [2, Theorem 1.8 (d)] that.))(/())((1F m L x m a R i R n i i ⊗∈⊗+∑=r rLet a = (a 1 + r R (m ), …, a n + r R (m )) and let {e 1, …, e n } be the canonical basis of the free right R -module R n . There is an exact sequence0)(/(0→⎯→⎯⎯→⎯→m L R a R n R n r r ηαwhere α is the inclusion map and η is given by η (e i ) = a i + r R (m ) (i = 1, …, n ). Tensoring by F yieldsthe following exact sequence0)](/[(id id →⊗⎯⎯→⎯⊗⎯⎯→⎯⊗⊗⊗F m L F R F a R n R FF n r r ηαwhere id F : F → F is the identity map. Since,)](/[0))(())(id (11F m L x m a x e R ni i R i n i i i F ⊗∈=⊗+=⊗⊗∑∑==r r η∑=⊗ni i ix e 1∈ Ker(η⊗id F ) = Im(α⊗id F ). So we haveb 1 = ⎥⎥⎥⎦⎤⎢⎢⎢⎣⎡111n b b M , …, b k = ⎥⎥⎥⎦⎤⎢⎢⎢⎣⎡nk k b b M 1∈(a n R r and y 1, …, y k ∈ F such that∑∑==⊗=⊗ni i i kj j j x e y b 11∈ R n ⊗F . But∑∑∑∑∑=====⊗=⊗=⊗kj ni j ij i kj ni j ij i kj j j y b e y b e y b 11111)(])[(.)]([)()(111111∑∑∑∑∑∑======⊗=⊗=⊗=n i kj j ij i n i kj j ij i k j n i j ij i y b e y b e y b eHence x i =∑=⊗kj j j y b 1 for each i . Finally, it is easy to see that m ∑=ni ij i b a 1= 0 for each 1 ≤ j ≤ k .- 3 -(2) ⇒ (1). Suppose that L /r R (m ) is finitely generated with m ∈ M andr R (m ) ⊆ L = a 1R + … + a n R + r R (m ) ≤ R R .Then each element in (L /r R (m ))⊗F is of the form∑=⊗+n i i R i x m a 1))((rwith x i ∈ F (i = 1, …, n ). If∑=⊗+n i i R i x m a 1))((r is contained in the kernel of the natural mapµ: (L /r R (m ))⊗F → LF /r R (m )Fi.e.,F m x a R ni ii )(1r ∈∑=, we have y 1, …, y k ∈ F and n ×k matrix B = (b ij )n ×k over R such that x i =∑=kj j ij y b 1 for each 1 ≤ i ≤ n and m ∑=ni ij i b a 1= 0 for each 1 ≤ j ≤ k by (2). Consequently,∑=⊗+ni i R i x m a 1))((r =)())((11∑∑==⊗+ni kj j ij R i y b m a r=)))(((11∑∑==⊗+n i k j j R ij i y m b a r =]))([(11∑∑==⊗+k j ni j R ij i y m b a r =∑=⊗+k j j R y m 1))(0(r = 0 ∈ (L /r R (m ))⊗F .This shows µ is monic and hence F is M -flat by [2, Theorem 1.8 (d)].Note that a left R -module F is M -flat if and only if 0 → K ⊗R F → M ⊗R F is exact for every finitely generated submodule K of M (see [6, 12.15(1)]. Let K =∑=n i i R m 1and f : R n → M be the compositionof the canonical epimorphism π : R n → K and the inclusion map α: K → M . Applying Hom R (−, R ) to f ,we obtain f *: M * → R n with imageIm f * = {(g(m 1), g (m 2), …, g (m n )) | g ∈ M * = Hom R (M , R ) },which is a matrix subgroup of R n in sense of Zimmermann [7]. We refer the reader to [7] for the general case. Following Zimmermann [7], we write m = (m 1, …, m n ) ∈ M n and H M ,m (R ) = {(g(m 1), g (m 2), …, g (m n )) | g ∈ M *}.Then we have the following commutative diagram with exact bottom row)*,(Hom )),((Hom 0)(,id 1F M F R H FM F R m R m M R R FR ni i →→↓↓⊗⎯⎯→⎯⊗⊗=∑ψϕαwhere ψ : M ⊗R F → Hom R (M *, F ) is given byψ (m ⊗y ): g a g (m )yfor all m ∈ M , y ∈ F and g ∈ M *, and ϕ: (∑=n i i R m 1) ⊗R F → Hom R (H M ,m (R ), F ) is defined such that,)())(,),(),((:)(1211∑∑==⊗ni i i n ni i i y m g m g m g m g y m a L ϕfor all y i ∈ F (i = 1, …, n ) and g ∈ M *. If ϕ is injective then so is α⊗id F and the converse holds in caseψ is injective. So we have the followingProposition 2. Let M be a right R -module and F a left R -module such that the canonical map ψ : M ⊗R F- 4 -→ Hom R (M *, F ) is injective , then the following are equivalent for a left R -module F .(1) F is M -flat .(2) The canonical map ϕ : (∑=n i i R m 1) ⊗R F → Hom R (H M ,m (R ), F ) is injective for all positiveintegers n and all m = (m 1, …, m n ) ∈ M n .Remark 3. It should be pointed out that the hypothesis “M ⊗R F → Hom R (M *, F ) is injective” is essential for (1) ⇒ (2) in Proposition 2. To see this we need only take a finitely generated right R -module M =∑=n i i R m 1which is not torsionless. Then F = R R is M -flat but it does not satisfy (2) incase m = (m 1, …, m n ) ∈ M n . In fact, there exist 0 ≠ m = ∑=n i i i r m 1∈ M and g ∈ M * such that g (m ) = 0since M is not torsionless. If follows that )(1∑=⊗ni i ir m ϕ= 0 but ∑=⊗ni i i r m 1 1)(1⊗∑=ni i i r m = m ⊗ 1 ≠ 0 in M ⊗R .Observing the following commutative diagram with exact rows)),((Hom ),(Hom )),(/(Hom 00)()(,,1F R H F R F R H R F R m F R F m m M R n R m M n R R ni i R n R R n →→→↓≅↓↓→⊗⎯→⎯⊗⎯→⎯⊗∑=ϕσ rwhere σ is the canonical map such that σ(a ⊗y ): b + H M ,m (R )a bay , for all a ∈)(m n R r , y ∈ F and b∈ R n , we have the following proposition.Proposition 4. Let M be a right R -module and F a left R -module . Then the following are equivalent for every m = (m 1, …, m n ) ∈ M n .(1) The canonical map ϕ : (∑=n i i R m 1) ⊗R F → Hom R (H M ,m (R ), F ) is injective .(2) The canonical map σ :(m n R r ⊗R F → Hom R (R n /H M ,m (R ), F ) is surjective .(3) Every f ∈ Hom R (R n /H M ,m (R ), F ) factors through a finitely generated free R -module , i.e., there exists R R k (1 ≤ k ∈ N ), g ∈ Hom R (R n /H M ,m (R ), R k ) and h ∈ Hom R (R k , F ) such that f = hg .(4) R n /H M ,m (R ) is projective with respect to every short exact sequence of the form0 → R L → R P → R F → 0 (∗)i.e., (∗) is R n /H M ,m (R )-pure in sense of Rothmaler [5].Proof . (1) ⇔ (2) follows from the commutative diagram mentioned above.(2) ⇒ (3). For each f ∈ Hom R (R n /H M ,m (R ), F ), by (2), there exist b j = [b 1j , …, b nj ]T ∈)(m n R r (1 ≤ j ≤ k ) and y 1, …, y k ∈ F such that σ (kj 1=Σb j ⊗y j ) = f . Now, define g ∈ Hom R (R n /H M ,m (R ), R k ) via g (a + H M ,m (R )) = aB for all a ∈ R n , where B = (b 1, …, b k ), and define h ∈ Hom R (R k , F ) via h (e j ) = y j for all 1 ≤ j ≤ k , where {e 1, …, e k } is the canonical basis of R k . It is easy to verify that h and g are as desired.- 5 -(3) ⇒ (2). Suppose (3), then for each f ∈ Hom R (R n /H M ,m (R ), F ), we have f = hg for some h ∈Hom R (R k , F ) and g ∈ Hom R (R n /H M ,m (R ), R k ). Let B =⎥⎥⎥⎦⎤⎢⎢⎢⎣⎡++))(())((,,1R H g R H g m M n m M εεM , where {ε1, …, εn } is thecanonical basis of R n , and let y j = h (e j ), where {e 1, …, e k } is the canonical basis of R k . It follows that σ (kj 1=Σb j ⊗y j ) = f , where b j is the j -th column of B (1 ≤ j ≤ k ). Therefore σ is surjective.(3) ⇒(4) follows by the following commutative diagram(4) ⇒ (3). Applying (4) to the exact sequence 0 → R L → R R (I )→ R F → 0 where I is a suitableindex set, we have the following commutative diagramSince R n /H M ,m (R ) is finitely generated, the image of f is contained in a finitely generated free submodule R k of R R (I ). Then (3) follows easily.Corollary 5. Let M be a right R -module and F a left R module such that the canonical map ψ : M ⊗R F → Hom R (M *, F ) is injective , then the following are equivalent .(1) F is M -flat .(2) The canonical map σ : )(m n R r ⊗R F → Hom R (R n /H M ,m (R ), F ) is surjective for all positive integers n and all m = (m 1, …, m n ) ∈ M n .(3) For all positive integers n and all m = (m 1, …, m n ) ∈ M n , every R -homomorphism from R n /H M ,m (R ) to F factors through a finitely generated free R -module .(4) Every exact sequence of the form 0 → R L → R P → R F → 0 is R n /H M ,m (R )-pure for all positive integers n and all m = (m 1, …, m n ) ∈ M n .Next, we consider M -flatness for factor modules of M -flat modules.Proposition 6. Suppose that M is a fixed right R -module and L is a submodule of an M -flat left R -module F . If the canonical map (M /K )⊗L →(M /K )⊗F is injective for each (finitely generated ) K R < M R then F /L is M -flat . The converse holds if the canonical map M ⊗L → M ⊗F is injective .Proof . Let us denote M /K and F /L by M and F respectively. Consider the following exact commutative diagram for each (finitely generated) K R < M R- 6 -FM F M d d F K F K F M L M F M F M d d F M L M F M F M d R d R d R d R ⊗⎯→⎯⊗↓↓⊗⎯→⎯⊗↓↓⊗⎯→⎯⊗→⎯→⎯↓↓↓↓⊗⎯→⎯⊗→⎯→⎯ 621151114387),(Tor ),(Tor ),(Tor ),(Torwhere d 2 is injective since F is M -flat by hypotheses and hence d 1 is surjective. If d 4 is injective then d 3 is surjective and so is d 5. It follows that d 6 is injective. This shows that F is M -flat.Conversely, if d 6 and d 8 are injective then d 5 and d 7 are surjective. Consequently, d 3 is surjective and hence d 4 is injective.Remark 8. Note that the canonical map M ⊗L → M ⊗F in the above proposition need not be injective in general even if every left R -module is M -flat. Recall that a ring R is right FS (see [4]) if the socle of R R is flat as a right R -module. Now, let us take a ring R which is not right FS (e.g., the ring R in Example 11 below) and M = Soc(R R ), the socle of R R . Obviously, every left R -module is M -flat since M is semisimple. But M is not flat hence M ⊗L → M ⊗F need not be injective in general.Corollary 9. Let L be a pure submodule of an M -flat left R -module F then both L and F /L are M -flat .Proof . Let L be a pure submodule of an M -flat left R -module F . Then for every submodule K of M , we have the following commutative diagramFM L M g h FK L K f⊗⎯→⎯⊗↓↓⊗⎯→⎯⊗ where f and g are monic and hence so is h . Therefore L is M -flat. Finally, F /L is M -flat by Proposition 6.Given a ring R and a fixed right R -module M . Let M -F (respectively F ) be the class of all M -flat (respectively flat) left R -modules. Then we haveProposition 10. (1) R is von Neumann regular if and only if M -F = F for every right R -module M .(2) R is right noetherian if and only if R is M -coherent for every (cyclic ) right R -module M .Proof . (1) is obvious.(2). If R is right noetherian then R is right coherent and r R (m ) is finitely generated for all m ∈ M . Hence R is M -coherent by [2, Observations 2.5 (i)]. Conversely, for each right ideal I of R , R is R /I -coherent and hence I = r R (1+I ) is finitely generated. Therefore R is right noetherian.It is well known that R is right coherent if and only if every direct product of flat left R -modules is- 7 -flat. The following example shows that the analogous statement for M -flatness and M-coherence fails. It also shows that the hypothesis “r R (m ) is finitely generated for all m ∈ M ” is essential for [2, Theorem 2.6] and answers the question in [2, p. 308].Example 11. There is a ring R with a module M R such that every left R -module is M -flat but R is not M -coherent .Proof . Note that every left R -module is M -flat if and only if every submodule M is pure in M , i.e., M is regular in Mod-R in sense of Wisbauer [6]. Let Z 2 = Z /2Z and A =∞=⊕1i Z 2 be the direct sum of countably infinite copies of Z 2. Then,R = Z 2 ∝ A = {⎥⎦⎤⎢⎣⎡a a 0α| a ∈ Z 2, α ∈ A },the trivial extension of Z 2 by A is a commutative ring with Soc(R ) = 0 ∝ A . Now, let M = Soc(R ). It follows that every left R -module is M -flat since M is semisimple and hence regular in Mod-R . But theright annihilator ofm = ⎥⎦⎤⎢⎣⎡00),0,0,1(0Lis not finitely generated. Therefore R is not M-coherent by [2, Consequences 2.3 (1)].Note that M = Soc(R ) is the unique maximal (right) ideal of the ring R in Example 11. Moreover, the left annihilator of M in R is M itself. By [4, Theorem 2.4], R is not (right) FS as we claimed in Remark 8.Let L be a left R -module and C a class of left R -modules. Recall from [3] that an R -homomorphism ϕ : L → C with C ∈ C is called a C –preenvelope of L if every ψ ∈ Hom R (L , C') with C' ∈ C factors through ϕ. The class C is said to be preenveloping if every left R -module has a C -preenvelope.Theorem 12. Let M be a right R -module . Then(1) M -F is closed under direct products if and only if M -F is preenveloping .(2) R Γ is M -flat for any set Γ if and only if every projective right R -module P has an M -flat dual module P * if and only if every direct product of flat left R -modules is M -flat .Proof . (1). By Corollary 9, M -F is closed under pure submodules. Thus (1) follows by a slight modification of the proof of [3, Proposition 6.5.1].(2). For every projective right R -module P , we have P ⊕ Q = R (I ) for some Q R and indexed set I . It follows that P * ⊕ Q * = (R (I ))* = R I . If R I is M -flat then so is P *. Conversely, suppose every projective right R -module P has an M -flat dual module P *. Then for every index set I , the free right R -module R (I ) has an M -flat dual (R (I ))* = R I .To complete the proof, it remains only to show that every direct product of flat left R -modules is M -flat provided R Γ is M -flat for any set Γ.For any set {F γ | γ ∈ Γ} of flat left R -modules, we have pure exact sequences- 8 -)(00)(Γ∈→→→→γγγγ F RK Iand00)(→Π→Π→Π→Γ∈Γ∈Γ∈γγγγγγF RK Iwhere)(γγI RΓ∈Π is a pure submodule of γγI RΓ∈Π. But γγI RΓ∈Π is M -flat by hypothesis.Therefore)(γγI R Γ∈Π is M -flat by Corollary 9. Since )(0γγγγI RK Γ∈Γ∈Π→Π→is pure,γγF Γ∈Π is M -flat by Corollary 9 again.Remark 13. So far it is unknown to us whether M -F is closed under products whenever R Γ is M -flat for any set Γ. To find a ring R with a right R -module M such that R Γ is M -flat for any set Γ but M -F is not closed under products, we have to consider those rings which are neither right noetherian nor von Neumann regular. Moreover the module M should not be regular in Mod-R .But by [2, Theorem 2.6], we haveCorollary 14. The following are equivalent for a fixed right R -module M such that r R (m ) is finitely generated for all m ∈ M .(1) R is M -coherent .(2) M -F is preenveloping .(3) P * ∈ M -F for every projective right R -module P .Recall from [6] that a right R module U is M -injective if every diagram of right R -modules with exact rowUMK ↓→→0can be extended commutatively by a morphism M → U . It is proved by Dauns [2, Theorem 1.7] that a left R -module F is M -flat if and only if the character module F + = Hom Z (F , Q /Z ) is M -injective. The following result is motivated by [1, Theorem 1].Proposition 15. Consider the following conditions for a fixed right R -module M :(1) A right R -module N is M -injective if and only if N + is M -flat .(2) A right R -module N is M -injective if and only if N ++ is M -injective . (3) A left R -module F is M -flat if and only if F ++ is M -flat . (4) M -F is closed under direct products . We have (1) ⇔ (2) ⇒ (3) ⇒ (4).Proof . (1) ⇔ (2) ⇒ (3) follows from [2, Theorem 1.7].(3) ⇒ (4). We adopt the method in [1, Theorem 1]. For any index set I and F i ∈ M -F (i ∈ I ), we have- 9 -)()()(0→⊕⎯→⎯Π⎯→⎯⊕↓Π⎯→⎯Π⎯→⎯++∈++∈≅++∈≅++∈∈i I i g i I i i I i i I i fi I i F F F F Fwhere ⊕i ∈I F i is an M -flat left R -module. Then (⊕i ∈I F i )++ is M -flat by (3). Since ⊕i ∈I +i F is a pure submodule of ∏i ∈I +i F by [1, Lemma 1(1)], it follows that g is split epic. Consequently, (⊕i ∈I +i F )+ is M -flat and so is ∏i ∈I ++i F . By [1, Lemma 1(2)], f is pure monic. This guarantees that ∏i ∈I F i is an M -flat left R -module.Remark 16. Note that (3) of Proposition 15 does not implies (1) even if R is M -coherent. For instance, the endomorphism ring R of a countably infinite dimensional vector space V over a field is von Neumann regular. Let M = R R then (3) of Proposition 15 holds since every left R -module F is (M -)flat. By [1, Theorem 2], (1) of Proposition 15 does not hold since R is not right noetherian.References[1] T. J. Cheatham & D. R. Strone (1981) Flat and projective character modules. Proc . Amer . Math .Soc ., 81: 175-177[2] J. Dauns (2006) Relative flatness and coherence. Comm . Algebra , 34: 303-312[3] E. E. Enochs & O. M. G. Jenda (2000) Relative Homological Algebra . de Gruyter Expositions inMath. 30. de Gruyter, Berlin[4] Z. K. Liu (1995) Rings with flat left socle. Comm . Algebra , 23: 1645-1656[5] P. Rothmaler (1994) Mittag-Leffler modules and positive atomicity . Manuscript, Kiel[6] R. Wisbauer (1991) Foundations of Module and Ring Theory . Philadelphia: Gordon and BreachSience Publishers[7] W. Zimmermann (1997) Modules with chain conditions for finite matrix subgroups. J . Algebra ,190: 68-87关于相对平坦性与相对凝聚性的注记张小向 陈建龙东南大学数学系,(210096)摘 要:本文利用某些特定的方程组给出了M -平坦性的一个新的刻画,其中M 为环R 上的一个固定的模。