亚洲免费av电影一区二区三区,日韩爱爱视频,51精品视频一区二区三区,91视频爱爱,日韩欧美在线播放视频,中文字幕少妇AV,亚洲电影中文字幕,久久久久亚洲av成人网址,久久综合视频网站,国产在线不卡免费播放

        ?

        TheˊEtale Homology and the Cycle Maps in Adic Coefficients*

        2013-08-16 08:57:14LITING

        LI TING

        (Mathematical College,Sichuan University,Chengdu,610064)

        TheˊEtale Homology and the Cycle Maps in Adic Coefficients*

        LI TING

        (Mathematical College,Sichuan University,Chengdu,610064)

        Communicated by Du Xian-kun

        In this article,we def i ne the ?-adic homology for a morphism of schemes satisfying certain f i niteness conditions.This homology has these functors similar to the Chow groups:proper push-forward,f l at pull-back,base change,cap-product, etc.In particular,on singular varieties,this kind of ?-adic homology behaves much better than the classical ?-adic cohomology.As an application,we give a much easier approach to construct the cycle maps for arbitrary algebraic schemes over f i elds.And we prove that these cycle maps kill the algebraic equivalences and commute with the Chern action of locally free sheaves.

        ?-adic cohomology,cycle map,derived category

        1 Introduction

        Theˊetale cohomology,especially the ?-adic cohomology,is one of the most important tools of modern algebraic and arithmetic geometry,which allows us to construct a“good”cohomology theory for varieties over f i elds of arbitrary characteristic.More specif i cally,people use the ?-adic cohomology H*(Xˊet,Z?)to substitute for singular cohomology on varieties of arbitrary characteristic.On a nonsingular varieties,the cohomology H*(Xˊet,Z?)has very good properties and produces rich results.But on singular varieties or more generally on arbitrary schemes,the cohomology H*(Xˊet,Z?)behaves not so good,and many important constructions and results are not valid.So on singular varieties,theˊetale homology is more suitable than theˊetale cohomology.

        In this paper,we generalize theˊetale homology def i ned in[1]in the following three facets.First,we def i ne theˊetale homology in adic coefficients,which we call the ?-adic homology.Secondly,our theory of ?-adic homology is def i ned over schemes separated and offi nite type over base schemes satisfying certain fi niteness conditions,not just the algebraic schemes over separably closed fi elds as in[1].In particular,algebraic schemes over fi elds which are not necessarily separably closed,are considered by us.Since our theory is based on the adic formalism created by Ekedahl[2],the ?-adic homology over base schemes of certain fi niteness conditions shares almost the same good functorial properties,with that over separably closed base fi elds.Thirdly,the ?-adic homology groups H*(X,F)de fi ned by us take value in arbitrary bounded complex F,not just Z?,Q?or Z/nZ as in[1].And almost all functors and properties are preserved when extending to complexes.Basing on this homology,we also extend the cycle maps de fi ned in[1],from separably closed base fi elds,to arbitrary base fi elds of fi nite ?-adic cohomological dimension.

        In Section 2,we brie fl y reiterate the category Dc(Xˊet,R·)together with the Grothendieck's six operations in[2].

        In Section 3,we recite the properties of the functor Rf!and use the language of[2]to rewrite the trace morphisms introduced in[3,4].

        In Section 4,we de fi ne the ?-adic homology groups Hn(X/Y,N)and Hn(X/Y,N)for a morphism X→Y of schemes satisfying certain fi niteness conditions.These homology groups behave similarly in many facets to the bivariant Chow groups A-n(X→Y)de fi ned in Ch. 17 of[5].We de fi ne two maps:the push-forward maps f*and the pull-back maps f*,for the?-adic homology groups,which correspond to these maps on Chow groups CH*(X)de fi ned in§1.4 and§1.7 of[5];and most important,we prove that these two maps commute(see Theorem 4.1),which is essential to construct various cycle maps basing on ?-adic homology. Moreover,we de fi ne the base change maps on the ?-adic homology.

        In Section 5,we apply the ?-adic homology in Section 4 to de fi ne the cycle map

        for arbitrary algebraic scheme X over a fi eld of fi nite cohomological dimension at ?.We prove that the cycle map clX,?commutes with the push-forward map f*and the pull-back map f*.And we prove that the cycle maps kill the algebraic equivalence of algebraic cycles.

        In Section 6,we prove that the cycle map clX,?commutes with the Chow action ci(E)∩· by locally free sheaves.

        The following notations and conventions would be used.

        Let N be the set of natural numbers,Z the domain of integers,and Q the fi eld of the rational numbers.Let Z?and Q?be the ?-adic completions of Z and Q respectively.

        A morphism f:X→Y of schemes is said to fl at(resp.smooth)of relative dimension n if f is fl at(resp.smooth)and all fi bers of f are n-equidimensional.

        A morphism f:X→Y of Noetherian schemes is said to be compacti fi able if it factors as f=?j where j:X→is an open immersion,and:→Y is a proper morphism. By Theorem 4.1 of[6],f is compacti fi able if and only if it is separated and of fi nite type.

        An algebraic scheme over a fi eld k is a scheme separated,of fi nite type over k.A variety over k is an integral algebraic scheme over k.

        If A is a Noetherian ring,we use D(A)to denote the derived category of A-modules,and de fi ne Dfg(A)to be the full subcategory of D(A)consisting of complexes cohomologicallyfi nitely generated.

        If X is a scheme and R is a ring,we use RXto denote the constant sheaf on Xˊetassociated to R.

        If F·is a complex of sheaves on Xˊet,we let F·[r]be the shift of F·to left by r,and F·(r):=F·?μ?nrthe Tate twist,and we write F·〈r〉:=F·(r)[2r].

        Let A and B be categories.If F:A→B and G:B→A is a pair of adjoint functors, then we write F?G for the adjunction.We also say that F is left adjoint to G,and G is right adjoint to F.

        The notation:=means being de fi ned as;means isomorphism;and the notation?in commutative diagrams means Cartesian square.

        2 The ?-adic Sheaves

        In this section,we brie fl y reiterate the theory of Ekedahl[2]about the category Dc(Xˊet,R·) together with the Grothendieck's six operations(see[7,8]).

        Fix a prime number ?,and let R be the integral closure of Z?in a fi nite extension fi eld of Q?.

        Let X be a Noetherian scheme.We denote by S(,R·)the abelian category of inverse systems

        such that each Fnis a sheaf of Rn-modules on Xˊet.Set

        and let Dc(,R·)be the full subcategory of D(,R·)consisting of complexes cohomologically AR-adic and constructible.Let Dc(,R·)be the quotient of Dc(,R·)by inverting AR-quasi-isomorphisms.

        If f:X→Y is a morphism of Noetherian schemes,then we have a triangulated functor

        As to other f i ve operations,we must add some restrictions on the underlying schemes. We consider the following condition(?)related to a scheme X:

        (?)X is Noetherian,quasi-excellent,of f i nite Krull dimension;? is invertible on X and cd?(X)<∞.

        From the Gabber's f i nitenes theorem forˊetale cohomology in[9],we know the following facts:

        (1)If X satisf i es(?),then any scheme of f i nite type over X satisf i es(?).

        (2)Let R be a quasi-excellent,Henselian local ring with residue f i eld k such that cd?(k)<∞.Then SpecR satisf i es(?).

        (3)If ?/=2,then the affine scheme SpecZ[1/?]satisf i es(?)(see X,6.1 of[3]).

        (4)If f:X→Y is a compactif i able morphism of schemes satisfying(?),then both Rf*and Rf!are of f i nite cohomological amplitude.

        In particular,if X is a scheme satisfying(?),then Xˊetsatisf i es the condition A)in[2], and thus we have two bi-triangulated functors

        And if f:X → Y is a compactif i able morphism of schemes satisfying(?),then there are triangulated functors

        For each scheme X satisfying(?),each object F in Dc(Xˊet,R·),and each n∈N,we def i ne

        Note that this de fi nition is compatible with the continuousˊetale cohomology(Xˊet,F) de fi ned in[10].

        When we consider the schemes of fi nite type over a separably closed fi eld,the following Theorem is essential.

        Theorem 2.1 The right derived functors of(Mn)→and the left derived functors of M →(M?RRn)de fi ne a natural equivalence of categories between Dc(R·)and Dfg(R).

        Proof. See Proposition 2.2.8 of[7].

        Now we fi x a separably closed fi eld k.Note that

        Let X be an algebraic scheme over k,and p:X→Speck the structural morphism.Put

        Then for each q∈Z,we have

        And we def i ne

        Theorem 2.2(The Ku¨nneth Formula) Let X and Y be two algebraic schemes over k, Z:=X×kY,f:Z→X and g:Z→Y the projections.Then for each F ∈(Xˊet,R·) and G∈(Yˊet,R·),there are two natural isomorphisms in(R):

        Moreover,there are two exact sequences of R-modules

        For convenience to study the cycle map,we introduce the following notation.Let X be a scheme satisfying(?).For F ∈Dc(Xˊet,R·)and n∈Z,we write

        3 The Functor Rf!and the Trace Morphisms from SGA 4&

        Proposition 3.1 Let f:X→Y be a compacti fi able morphism of schemes satisfying(?) such that all fi bers of f are of dimensions≤d.Then for each a∈Z,Rf!sendsto

        Proof. See XVIII,3.1.7 of[3].

        Lemma 3.1 Let f:X→Y be a compacti fi able morphism of schemes satisfying(?).Then for every pair of objects F and G in(Yˊet,R·),there is a natural morphism

        Proof.First we have a composite morphism

        where φ is induced by the projection formula for Rf!,and ψ is induced by the adjunction.

        Since Rf!is left adjoint to Rf!,the above morphism induces the required morphism

        Proposition 3.2 Let f:X → Y and g:Y → Z be two compacti fi able morphisms of schemes satisfying(?).For every pair of objects F and G in(Zˊet,R·),there is a natural morphism

        in Dc(Xˊet,R·)which is functorial in F and G.

        Proof.We have

        Proposition 3.3 Let

        be a Cartesian square of schemes satisfying(?).Assume that f is compacti fi able.

        (1)For each object F in Dc(Xˊet,R·),there is a natural morphism in Dc(,R·)

        (2)For each object G in Dc(Yˊet,R·),there is a natural morphism in Dc(,R·)

        (3)Assume that Y is an algebraic scheme over a fi eld k,and there exists a k-scheme T such that Y′=Y×kT.Then the morphisms in(1)and(2)are both isomorphisms.

        (4)For each object F in Dc(Xˊet,R·),there is a natural isomorphism in Dc(,R·)

        (5)For each object G in Dc(,R·),there is a natural isomorphism in Dc(Xˊet,R·)

        Proof. (1)is induced by the classical base change morphisms.

        (2)is from[3],XVIII,3.1.14.2.

        (3)is by Th.Finitude,1.9 of[4].

        (4)is by XVII,5.2.6 of[3].

        (5)is by XVIII,3.1.12.3 of[3].

        Now we review the trace morphisms.

        De fi nition 3.1 A morphism f:X →Y of schemes is said to be fl at at dimension d if there exists a nonempty open subset U of X satisfying the following conditions:

        (1)f:U→Y is fl at;

        (2)for each point y∈Y,Uyis either empty or d-dimensional;

        (3)every fi ber of XU→Y is of dimension<d.

        By XVIII,2.9 of[3],for every compacti fi able morphism f:X→Y of schemes satisfying (?)which is fl at at dimension d,and for every object G in Dc(Yˊet,R·),we have a trace morphism:

        Since Rf!is right adjoint to Rf!,the morphism Trfinduces a canonical morphism in Dc(Xˊet,R·):

        Moreover,we have a commutative diagram

        By XVIII,3.2.5 of[3],we have

        Proposition 3.4 Let f:X → Y be a compactif i able smooth morphism of relative dimension d of schemes satisfying(?).Then for any object G in Dc(Yˊet,R·),the canonical morphism

        is an isomorphism in Dc(Xˊet,R·).

        The following Propositions 3.5–3.7 are deduced from XVIII,2.9 of[3].

        Proposition 3.5 Let

        be a Cartesian square of schemes satisfying(?).Assume that f is compactif i able and f l at at dimension d.Then f′is also f l at at dimension d,and for each object G in Dc(Yˊet,R·)we have

        (1)the composite morphism

        is equal to Trf′,where the isomorphism φ is def i ned in Proposition 3.3(4);

        (2)the composite morphism

        is equal to tf′,where the last morphism is def i ned in Proposition 3.3(2).

        Proposition 3.6 Let f:X → Y and g:Y → Z be two compactif i able morphisms of schemes satisfying(?)which are f l at at dimension d and e respectively,and H be an object in Dc(Zˊet,R·).Then we have

        (1)the composite morphism

        is equal to Trg?f;

        (2)the composite morphism

        is equal to tg?f.

        Proposition 3.7 Let f:X→Y be a f i nite morphism of schemes satisfying(?)such that f*Xis a locally freeY-module of degree d.Then for each object F in Dc(Yˊet,R·),the composite morphism

        is equal to the multiplication by n.

        The following proposition shows that the trace morphism is essentially determinated by the generic points.Let A be a Noetherian ring(in particular,A=Rn).

        Proposition 3.8 Let X be an n-dimensional algebraic scheme over k,X1,X2,···,Xrbe all irreducible components of dimension n of X,and F be an A-module.For each i,let Yi/=? be an open subset of X containedand regard Yias a reduced subscheme of X.For each i,let xibe the generic point of Xiand put ai:=length(X,xi).Then there is a canonical isomorphism ω of A-modules which makes a commutative diagram

        4 ?-adic Homology for Morphisms of Algebraic Schemes

        Let f:X→Y be a compactif i able morphism of schemes satisfying(?).For each object N in Dc(Yˊet,R·)and for each n∈Z,we def i ne the n-th ?-adic homology associated to f to be

        which is an R-module.

        For convenience to def i ne pull-backs along f l at morphisms and cycle maps,we also def i ne

        We set

        We also use Hn(X/Y,N)(resp.Hn(X/Y,N))to denote(resp.if no confusion arises.

        If X is an algebraic schemes over a separably closed f i eld k and N is an object in Dfg(R), we write

        By Proposition 3.4 we have

        Lemma 4.1 Let f:X→Y be a compacti fi able morphism of schemes satisfying(?)which is fl at at dimension d.Then for each object N in Dc(Yˊet,R·)and for n∈Z,the morphism tfinduces a canonical homomorphism of R-modules:

        Moreover,if f is smooth of relative dimension d,then the above homomorphism is an isomorphism.

        Proposition 4.1 Let f:X→S be a compacti fi able morphism of schemes satisfying(?), Y be a closed subscheme of X and U:=XY.Then we have a long exact sequence

        Proof.Put M:=Rf!N.Then the proposition follows from the distinguished triangle

        where i:Y→X and j:U→X are the inclusions.

        Proposition 4.2(Mayer-Vietoris Sequence) Let f:X→S be a compacti fi able morphism of schemes satisfying(?),X1and X2be two closed subschemes of X such that X=X1∪X2(as sets).Then we have a long exact sequence

        Proof.Put M:=Rf!N.Then the proposition follows from the distinguished triangle

        where i:X1×XX2→X,i1:X1→X,i2:X2→X are the inclusions.

        Proposition 4.3(Vanishing) Let f:X → Y be a compacti fi able morphism of schemes satisfying(?)and such that all fi bers of f are of dimensions≤d,and N be an object inThen Hn(X/Y,N)=0 whenever n>2d-a.

        Proposition 4.4 Let f:X→S be a compacti fi able morphism of schemes satisfying(?), Y be a closed subscheme of X such that dimYs≤d for all s∈S,and X′:=XY,N be an object in(Sˊet,R·).Then for each integer n>2d+1-a,there is a canonical isomorphism of R-modules

        Proof. Apply Propositions 4.1 and 4.3.

        Let f:X→Y be a compacti fi able morphism of schemes satisfying(?).For each object G in Dc(Yˊet,R·),we de fi ne

        to be the canonical morphisms induced by the adjunctions f*?Rf*and Rf!?Rf!respectively.

        The following map is a kind of variant of the Gysin homomorphism.

        De fi nition 4.1(Push-forward) Let p:X→S and q:Y→S be two compacti fi able morphisms of schemes satisfying(?),and f:X→Y a proper S-morphism.For every object N in Dc(Sˊet,R·)and for every n∈Z,we de fi ne a homomorphism of R-modules

        as follows.For each α∈Hn(X/S,N),f*(α)is de fi ned to be the composition

        Proof.This comes from the following simple lemma.

        Lemma 4.2 Let f:X→Y and g:Y→Z be two compacti fi able morphisms of schemes satisfying(?),and H be an object in Dc(Zˊet,R·).Then we have

        (1)the following composition is equal to δg?f:

        (2)the following composition is equal to θg?f:

        De fi nition 4.2(Pull-back) Let p:X→S and q:Y→S be two compacti fi able morphisms of schemes satisfying(?),and f:X→Y an S-morphism which is fl at at dimension d.For every object N in Dc(Sˊet,R·)and for every n∈Z,we de fi ne a homomorphism of R-modules

        as follows.For each β∈Hn(Y/S,N),f*(β)is de fi ned to be the composition

        Proof. This follows from Proposition 3.6(2).

        Theorem 4.1 Let S be a scheme satisfying(?),and r:Y→ S a compacti fi able morphism.Let

        be a Cartesian square of schemes such that f is proper and q is compacti fi able and fl at at dimension d,N be an object in Dc(Sˊet,R·)and n∈Z.Then we have

        Proof.Put M:=Rr!N.Let α∈Hn(X/S,N).Then q*?f*(α)is equal to the composition

        After applying Proposition 3.5(2)to tp,we obtain that the morphism?p*(α)is equal to the composition

        Consider the following diagram

        where?means commutative square.The commutativity of(a)and(b)are by the following simple Lemma 4.3.So the whole diagram is commutative.Note that the composition along the direction

        *in the above diagram is equal to q?f*(α),and the composition alongis equal to?p*(α).Thus

        Lemma 4.3 Let

        be a Cartesian square of schemes satisfying(?)with all morphisms compacti fi able.Then we have

        (1)For each object G in Dc(Yˊet,R·),the diagram

        is commutative in Dc(,R·),where φ is de fi ned in Proposition 3.3(5)and ψ is de fi ned in Proposition 3.3(2).

        (2)Assume that f is proper.Then for each object G in Dc(Yˊet,R·),the diagram

        is commutative in Dc(,R·),where α is de fi ned in Proposition 3.3(5)and β is induced by the composition

        De fi nition 4.3(Base Change) Let

        be a Cartesian square of schemes satisfying(?)with f compacti fi able.For every object N in Dc(Sˊet,R·)and for every n∈Z,we de fi ne a homomorphism of R-modules

        as follows.For each α∈Hn(X/S,N),u*(α)is de fi ned to be the composition

        where φ is de fi ned in Proposition 3.3(2).

        We have the following three obvious propositions about the base change homomorphisms. Proposition 4.7 Let k?K be two separably closed fi elds,f:X→S be a morphism of algebraic schemes over k,and u:SK→ S be the projection.Then for each object N in Dc(Sˊet,R·)and for each n∈Z,the homomorphism

        is an isomorphism.

        Proof. It follows from Proposition 3.3(3)and Theorem 2.1.

        Proposition 4.8 Let

        be a commutative diagram of schemes satisfying(?)with both squares Cartesian,and all three vertical arrows being compacti fi able.Then for all N ∈Dc(Sˊet,R·)and n∈Z,we have

        Proposition 4.9 Let

        be a commutative diagram of schemes satisfying(?)with both squares Cartesian,and all level arrows being compacti fi able.Let N be an object in Dc(Sˊet,R·)and n∈Z.Then we have

        (1) If f is proper,then

        (2)If f is fl at at dimension d,then

        De fi nition 4.4(Galois action) Let k0be a fi eld,k the separably closed fi eld of k0,G:= Gal(k/k0),X be an algebraic scheme over k,Y0be an algebraic scheme over k0,Y := Y0?k0k,N0be an object in Dc(Y0,ˊet,R·)and N the pull-back of N0on Y.Then there is an action of G on Hn(X/Y,N)de fi ned by

        In particular,if N∈Dfg(R)and n∈Z,then there is a Galois action of G on Hn(X,N).

        The following theorem is used to prove that cycle maps eliminate algebraic equivalent classes.

        Theorem 4.2 Let f:X→Y be a morphism of algebraic schemes over a separably closed fi eld k,Z be a nonsingular variety over k,N be an object in Dc(Yˊet,R·),

        For each z∈Z(k),put

        Proof. By Proposition 4.7 we may assume that k is algebraically closed.Since every two rational points of Z can be jointed by a series of nonsingular curves,we may further assume that Z is a complete nonsingular curve.First we have a commutative diagram with both squares Cartesian:

        By Proposition 3.3,we have

        Since Z is a complete nonsingular curve over k,we have

        where g is the genus of Z,andR are all free R-modules.Now we apply Theorem 2.2 to obtain an isomorphism:

        Let β∈Hn(X/Y,N)be the image of α induced by above isomorphism.Then

        5 The Cycle Maps for Chow Groups

        In this section,we construct the cycle maps for arbitrary algebraic schemes over k,where k is a f i eld such that

        Let f:X→Y be a compactif i able morphism of schemes satisfying(?)which is f l at at dimension d.We def i ne

        Proposition 5.1 Let X→S and Y→S be two compactif i able morphisms of schemes satisfying(?),and f:X→Y a morphism of S-schemes.Assume that Y→S and f:X→Y are f l at at dimension n and d respectively.Then we have

        Proof. This follows from Proposition 3.6(2).

        Proposition 5.2 Let p:X→S and q:Y→S be two morphisms of schemes satisfying (?),both of which are compactif i able and f l at at dimension d,and f:X → Y be a f i nite S-morphism such that f*Xis a locally freeY-module of degree n.Then we have

        Proof. By the de fi nition of f*and Proposition 3.6(2),the element f*c?(X/S)is equal to the composite morphism

        By Diagram(3.1)and Proposition 3.7,we have a commutative diagram as follows:

        Thus we get the proof.

        Let X→S be a compacti fi able morphism of schemes satisfying(?),and i:Y→X a closed immersion.Assume that the morphism Y→S is fl at at dimension d.Then we de fi ne

        Let X be an algebraic scheme over k.Then for each n∈Z,there is a canonical homomorphism of groups

        Proposition 5.3 Let f:X→Y be a proper morphism of algebraic schemes over k.Then for every n∈N,we have a commutative diagram

        Proof.Let X′be an n-dimensional subvariety of X,Y′:=f(X′),i:X′→X and j:Y′→Y the inclusion,and g :X′→Y′the induced morphism.By Proposition 4.5 we have

        Since f*[X′]=deg(X′/Y′)[Y′](see the section 1.4 of[5]),we have only to prove that

        If dimY′<n,then

        And by Proposition 4.3,

        If dimY′=n,we apply the result in Example 3.7 of[11].Since the morphism g is generically fi nite and Y′is an integral scheme,there exists an nonempty subscheme V of Y′such that g:g-1(V)→V is a fi nite morphism and g*X′|Vis a locally freeV-module. Now the proposition follows from Propositions 4.4 and 5.2.

        Proposition 5.4 Let X be an algebraic scheme over k,and Y be a n-equidimensional closed subscheme of X.Then we have

        Proof. This is easily deduced from Proposition 3.8.

        Proposition 5.5 Let f:X→Y be a fl at morphism of relative dimension d of algebraic schemes over k.Then for every n∈N,we have a commutative diagram

        Proof.Let α∈Zn(Y).We may assume that Y is a variety of dimension n and α=[Y]. Then we have only to apply Proposition 5.1.

        Lemma 5.1 Let X be a nonsingular variety of dimension n over k,and D be an e ff ective divisor on X.Then

        Proof. See(3.26)of[10].

        Theorem 5.1 Let X be an algebraic schemeover k.Then for each n∈N,

        Proof. After applying Proposition 1.6 of[5]together with Propositions 5.3 and 5.5,we have only to prove that

        This follows from Lemma 5.1.

        In the following,we de fi ne the degree map for the homology of degree zero.

        De fi nition 5.1 For any proper algebraic scheme X over k,we de fi ne degree map deg?to be the homomorphism

        where p:X→Speck is the structural morphism.

        Lemma 5.2 Let X be an n-dimensional proper algebraic scheme over k.

        (1)We have a commutative diagram

        (2)We have a commutative diagram

        Proof.(1)can be proved by the commutative diagram(3.1).

        (2)can be proved by Proposition 5.3.

        Proposition 5.6 Assume that k is separably closed and let X be a nonsingular complete variety over k.Then deg?:is an isomorphism.

        Proof.Put dimX=n.By Proposition 3.4,we have only to prove that

        is an isomorphism.This is true by VI,11.1(a)of[12].

        The following theorem shows that the cycle map clX,?annihilate algebraic equivalence of cycles.

        Theorem 5.2 Assume that k is separably closed and let X be an algebraic scheme over k.Then for each n∈N,the cycles in CHn(X()which are algebraically equivalent to zero (in the sense of 10.3 of[5])are contained in kerclX,?).

        Proof. By Proposition 4.7 we may assume that k is algebraically closed.Let c1,c2∈CHn(X)be such that c1~ac2,T be a nonsingular curve over k,t1,t2∈T(k),and c∈CHn+1(X×kT)be such that cti=cifor i=1,2.Obviously,we may assume that c=[Y], where Y is an(n+1)-dimensional subvariety of X×kT such that for all t∈T(k),Y is not contained in

        Obviously,the induced morphism Y→T is dominant and fl at.Put

        By Propositions 4.9 and 3.5(2),we have

        So we have only to apply Proposition 4.2.

        6 Cap-products and Compatibility with Chern classes

        First we de fi ne the cap-products for the ?-adic homology.

        De fi nition 6.1(Cap-product) Let f:X→Y and g:Y→Z be compacti fi able morphisms of schemes satisfying(?),M and N two objects in(Zˊet,R·).For every m,n∈Z,there is a cap-product

        where φ is de fi ned in Proposition 3.2.

        Similarly,we may de fi ne the cap-product for H*as follows:

        In particular,if X→S is a compacti fi able morphisms of schemes satisfying(?),and N is an object in(Sˊet,R·),then for every m,n∈Z,there are cap-products

        The following Proposition can be directly calculated.

        Proposition 6.1(Projection Formula) Let f:X →Y and g:Y→ S be morphisms of schemes satisfying(?)with f proper and g compacti fi able,and N an object in Dc(Yˊet,R·). Then we have

        (1)For every α∈Hr(Y,R)and β∈Hn(X/S,N),

        (2)For every α∈Hr(Y,R)and β∈Hn(X/S,N),

        It may be further showed that the cap-product de fi ned in De fi nition 6.1 has many similar properties to bivariant intersection theory de fi ned in Ch.17 of[5],i.e.,has associativity and is compatible with the Pull-back functor f*,the push-out functor f*and the base change functor u*.Since we need not them here,we leave it to the readers.

        Next,we review the cycle maps for locally free sheaves.First by(3.26)a)of[10],we have a homomorphism of groups

        for every scheme X satisfying(?).The following two propositions depict the cycle maps for locally free sheaves.

        Proposition 6.2 Let X be a scheme satisfying(?),E a locally freeX-module of constant rank r+1,P:=P(E),and p:P → X be the projection.Then there is a canonical isomorphism of Z?-algebras

        Proof. See(6.13)of[10]or VII,2.2.6 of[13].

        As a direct application of the above proposition,we have

        Proposition 6.3 Let X be a scheme satisfying(?),E a loca(lly free)-module of constant rank m,P:=P(E∨),and p:P→X be the projection,ξ:=(1).Then for each r∈N there exists a unique element(E)∈Hr(X,Z?)such that

        Now we de fi ne the trace morphisms for regular immersions of codimension 1.Let X be a scheme satisfying(?)and i:D→ X a regular closed immersion of codimension 1.By (3.26)of[10]and its proof,i:D→X determinates an element

        Similar to 2.3.1 of the paper(cycle)of[4],we have

        Proposition 6.4 Let S be a scheme satisfying(?),f:X → S and g:Y→ S be two compacti fi able morphisms which are fl at at dimension n and n-1 respectively,and i:Y→X be a regular closed immersion of codimension 1 such that f?i=g.Then we have

        (1)The composite morphism

        is equal to Trg;

        (2)The composite morphism

        is equal to tg.

        Finally,we prove that the cycle maps are compatible with Chern classes.According to Ch.3 of[5],if X is an algebraic scheme over k and E is a locally free-module,then there is an operation of Chern classes on each Chow group

        Theorem 6.1 Let X be an algebraic scheme over k,E a locally free-module,and α∈CHr(X).Then we have

        Proof.By the the projection formulas(Proposition 6.1 and Theorem 3.2(c)of[5]),we obtain that if f:X′→X is a proper morphism and α′∈CHr(X′)such that f*(α′)=α and the pair(f*E,α′)satis fi es(6.2),then the pair(E,α)also satis fi es(6.2).Thus by the splitting construction(see§3.2 of[5]),we may assume that E=L is an invertible-module and have only to prove that

        Moreover we may assume that X is a variety of dimension r and α=[X].After replacing X with its normalization,we may assume that X is normal.Then we may assume that L=(Y)where Y→X is a regular closed immersion of codimension 1.Then we have only to apply Proposition 6.4 to end our proof.

        References

        [1]Laumon G.Homologieˊetale.In:Sˊeminaire de Gˊeomˊetrie Analytique(Eˊcole Norm.Sup.,Paris, 1974–75).Astˊerisque,No.36–37.Paris:Soc.Math.France,1976:163–188.

        [2]Ekedahl Y.On the Adic Formalism.In:The Grothendieck Festschrift.vol.II.Progr.Math., 87(1990).Boston,MA:Birkh¨auser Boston,1990:197–218.

        [3]Grothendieck A.Thˊeorie des Topos et CohomologieEˊtale des Schˊemas,SGA 4:Lecture Notes in Math.269,270,305.Berlin:Springer-Verlag,1972-73.

        [4]Deligne P.Cohomologieˊetale,SGA:Lecture Notes in Math.,569.Berlin:Springer-Verlag, 1977.

        [5]Fulton W.Intersection Theory.In:Ergebnisse der Mathematik und ihrer Grenzgebiete(3). vol.2.Berlin:Springer-Verlag,1984.

        [6]Conrad B.Deligne's notes on Nagata compacti fi cations.J.Ramanujan Math.Soc.,2007,22: 205–257.

        [7]Behrend K A.Derived l-adic categories for algebraic stacks.Mem.Amer.Math.Soc.,2003, 163:viii+93.

        [8]Laszlo Y,Olsson M.ˊThe six operations for sheaves on Artin stacks II:Adic coefficients.Publ. Math.Inst.HautesEtudes Sci.,2008,107:169–210.

        [9]Gabber O.Finiteness Theorems forE`tale Cohomology of Excellent Schemes.In:Conference in Honor of P.Deligne on the Occasion of His 61st Birthday.Princeton:IAS,October 2005.

        [10]Jannsen U.Continuousˊetale cohomology.Math.Ann.,1988,280:207–245.

        [11]Hartshorne R.Algebraic Geometry.New York:Springer-Verlag,1977.

        [12]Milne J S.Eˊtale Cohomology:Princeton Mathematical Series 33.Princeton,N.J.:Princeton Univ.Press,1980.

        [13]Grothendieck A.Cohomologie l-adique et Fonctions L,SGA 5:Lecture Notes in Math.,589. Berlin:Springer-Verlag,1977.

        tion:14F20,14C25

        A

        1674-5647(2013)01-0068-20

        *Received date:Nov.25,2010.

        The NSF(10626036)of China.

        欧美黑人粗暴多交高潮水最多| 久久国产精品一区av瑜伽| 四虎影在永久在线观看 | 亚洲国产成人AⅤ片在线观看| 99精品又硬又爽又粗少妇毛片| 亚洲av毛片在线免费观看| а天堂中文在线官网| 99精品国产在热久久国产乱| 色偷偷亚洲av男人的天堂| 九九久久精品国产免费av| 国产精品无码aⅴ嫩草| 国产男女猛烈无遮挡免费视频| 99热久久只有这里是精品| 伊人青青草综合在线视频免费播放| 久久香蕉国产线看观看精品yw| 国产AV无码专区亚洲AⅤ| 少妇一级aa一区二区三区片| 99在线视频这里只有精品伊人| 国产午夜精品一区二区| 亚洲欧美国产日韩字幕| 超短裙老师在线观看一区| 99久久免费视频色老| 曝光无码有码视频专区| 99久久国产亚洲综合精品| 国产成人av三级三级三级在线| 国产成人综合日韩精品无码| 亚洲av无码av吞精久久| 国产对白刺激在线观看| 六月婷婷亚洲性色av蜜桃| 丁字裤少妇露黑毛| 亚洲精品有码在线观看| 日本一区二区午夜视频| 老妇高潮潮喷到猛进猛出| 一本之道高清无码视频| 欧美精品一本久久男人的天堂| 久久免费亚洲免费视频| 色爱无码av综合区| 无码精品色午夜| 国产午夜福利在线观看中文字幕| 亚洲午夜成人精品无码色欲| 蜜臀av免费一区二区三区|