arbitrary characteristic
Georg Hein
June 2, 1999
Abstract
LetX be a smooth variety de ned over an algebraically closed eld of arbitrary characteristic andOX(H) be a very ample line bundle onX. We show that for a semistableX-bundleEof rank two, there exists an integer mdepending only on (E):Hdim(X);2andHdim(X)such that the restric- tion of E to a general divisor in jmHjis again semistable. As corollaries we obtain boundedness results, and weak versions of Bogomolov's theorem and Kodaira's vanishing theorem for surfaces in arbitrary characteristic.
Introduction
Let (
X
OX(1) = OX(H
)) be a smooth polarized variety dened over an al- gebraic closed eld of arbitrary characteristic. We assume OX(1) to be very ample. Additionally, letE
be a -semistable vector bundle of rank two onX
. We want to show that there exists an integerm
only depending on the characteristic numbersH
dim(X) and (c
1(E
)2;4c
2(E
)):H
dim(X);2 such that the restriction ofE
to a general element of jmH
j is semistable. Such e ective bounds have been known only for the case that the characteristic is zero. In this case the restriction theorem of Flenner (see 3]) gives e ective bounds onm
for semistable bundles of arbitrary rank. On the other hand there are results of Mehta and Ramanathan which say that the restriction ofE
to a divisor inj
mH
j is semistable (or stable, forE
a stable vector bundle) ifm
0 (cf. 8], and 9]). A detailled overview on restriction theorems is given inx7 of the book 5] of Huybrechts and Lehn.First we discuss the case of rank two bundles on a surface
X
. Theorem 2.1 shows that for a semistableX
-vector bundleE
of rank two there exists an integerm
such that the restriction ofE
to a general curve in the linear systemj
mH
j is semistable.Using this result we provide a boundedness result for semistable rank two bun- dles (proposition 2.4).
For surfaces dened over C a semistable bundle
E
cannot have positive dis- criminant (E
) (Bogomolov's theorem, cf. 1]). In positive characteristic this does not hold. No more than the Kodaira vanishing holds for positive char- acteristic (see 12]). It is remarkable that semistable bundles which contradict Bogomolov's theorem behave well with respect to restrictions. Applying our1
restriction result, we obtain a weak form of Bogomolov's theorem (corollary 2.7, cf. also 7] for vector bundles of arbitrary rank), and a weak form of Ko- daira's vanishing theorem in arbitrary characteristic (corollary 2.8). The reader familiar with the vector bundle techniques presented in Lazarsfeld's lectures, 6] will deduce Reider type theorems for surfaces in arbitrary characteristic.
If the
X
-bundleE
is semistable but not stable, then it is easy to see that the restriction ofE
to a general curve in jH
j is semistable but not stable.Conversely, we may ask whether stable bundles do restrict to stable objects.
Theorem 2.9 gives an armative answer to this question. The proof follows an idea of Bogomolov (see 2] and 5] theorem 7.3.5) using the weak Bogomolov inequality deduced before.
Finally, we present with theorem 3.2 the higher dimensional version of theorem 2.1. It turns out that its proof is easier than the proof in the surface case. The main reason for this simplication is the fact that two general hyperplanes in a linear system intersect in a irreducible subscheme.
All these results should generalize to vector bundles of arbitrary rank. To prove the corresponding results it seems necessary to consider the complete Harder- Narasimhan ltration.
Acknowledgement:
The author would like to thank D. Huybrechts for many helpful remarks.1 Preliminaries
Let
X H
be a polarized projective variety. We will identify line bundles onX
and their corresponding Cartier divisor classes. Moreover, to any class in the Chow group CHdim(X)(X
) of codimension dim(X
) cycles is assigned via evaluation on the fundamental classX
] ofX
its characteristic number. This allows us to interpretc
i(E
):H
dim(X);i as integers.For a coherent
X
-sheafE
, we writeE
(n
) instead ofE
OX(H
)n. The Hilbert polynomialE :n
7!(E
(n
)) can be written in the following form E(n
) =a
0(E
)n
+ dimX
dimX
+
a
1(E
)n
+ dimX
;1 dimX
;1+
::: :
IfH
is suciently general in the linear systemjH
j (i.e., TorO1X(E
OH) = 0), we havea
i(E
) =a
i(E
jH), for all integersi <
dimX
.We dene the
H
-slopeH(E
) ofE
to be the quotienta
1(E
)=a
0(E
). A coherent sheafE
is called Mumford semistable (resp. stable) with respect toH
, ifE
is torsion free, and for all proper subsheavesF
E
the inequalityH(F
)H(E
) (resp. H(F
)<
H(E
)) holds true. This kind of stability is also named slope stability, weak stability, or -stability. For brevity we simply write stability because we only use this stability concept. We will frequently use the following facts on stable and semistable coherent sheaves:1. If
E
andF
are semistable with H(E
)>
H(F
), then the group Hom(E F
) vanishes.3 2. For a stable bundle
E
on a variety dened over an algebraically closed eld the endomorphism group End(E
) consists of the scalar multiples of the identity.3. If a rank two vector bundle
E
is not semistable, then there exists a unique maximal subsheafE
1E
of rank one which is the maximal destabilizing subsheaf. Or equivalently, there exists a unique destabilizing quotientE
!Q
. The ag 0E
1E
is the Harder-Narasimhan ltration ofE
. See, for example, the article 13] of Shatz.An important invariant of a vector bundle is its discriminant. Let
E
be a vector bundle of rankr
with Chern rootsfigi=0 ::: r. As the name discriminant suggests we dene the discriminant (E
) of the vector bundleE
by (E
) =Pi<j(
i;j)2. Obviously (E
) can be expressed in terms of the Chern classes ofE
, namely (E
) = (r
;1)c
1(E
)2 ;2rc
2(E
). (Unfortunately, there are di erent denitions of (E
) in literature, di ering by a sign or a constant.) In particular, we have (E
) =c
1(E
)2;4c
2(E
), for a rank two vector bundle.A rank two vector bundle
E
on a surfaceX
is named Bogomolov unstable if there exists an injectionA
!E
of coherent sheaves whereA
is anX
-line bundle, the cokernel ofis torsion free, and the inequalities (2A
;c
1(E
))2>
0, and (2A
;c
1(E
)):H >
0 are satised for a polarizationH
ofX
.For a rational number
q
, letdq
e be the least integer not smaller thanq
,bq
c the largest integer smaller or equal toq
, andq
]+ the maximum ofq
and 0.2 Rank two bundles on surfaces
2.1 The semistable restriction theorem
Theorem 2.1
LetX
be a smooth surface over an algebraically closed eld with a very ample line bundle OX(1) =OX(H
). For anX
-vector bundleE
of rank two which is semistable with respect toOX(1) the following holds:1. If (
E
) 0, then the restriction ofE
to a general curve of the linear system jH
jis semistable2. For (
E
)<
0 and any integerl
withl
log2q
;(E)
H2 + 1the restric- tion of
E
to a general curve in j2lH
j is semistable.Proof:
We divide the proof in several steps. First we outline its strategy:We dene the objects which are needed. In particular, we dene the non negative integer
A
(m
) which measures the instability of the restriction ofE
to a general curve of the linear systemjmH
j(step 1-3)We next (step 4-7) compute an upper bound for
A
(1). This bound de- pends only on the Chern number (E
) andH
2After that, we give an upper bound for
A
(2m
) in terms ofA
(m
) (step 8-12)Finally, we combine both estimates to conclude the theorem (step 13).
Step 1:
Letm
be a positive integer. For the linear systemjmH
jwe denote byC
jmHj the universal curve over j
mH
j. We have the morphismsj
mH
joo p CjmHj q //X :
The spacej
mH
jis isomorphic to Ph0(m);1 whereh
0(m
) denotes the dimension ofH
0(X
OX(m
)). Since jmH
j is base point free,q
is a Ph0(n);2-bundle. We denote byg
m the genus of a smooth curve of jmH
j. A curveC
X
rationally equivalent tomH
corresponds to a geometric point in jmH
jwhich we denote byC
].Step 2:
For all integersa
with 2a < c
1(E
):
(mH
) we consider the Quot scheme Quotm a := QuotPq E=a CjmHj=jmHjof
p
-at quotients ofq
E
with Hilbert polynomialP
a(k
) = (mH
2)k
+a
+1;g
mwith respect to the very ample line bundle
q
OX(1) see 4].For
a < g
;1;h
1(E
);h
2(E
(;mH
)) the scheme Quotm a is the empty scheme.To see this, we remark that for any curve
C
] 2jmH
j we have the inequalityh
1(E
jC)h
1(E
) +h
2(E
(;mH
)). Hence, any quotient of the restrictionE
jChas at least Euler characteristic ;(
h
1(E
) +h
2(E
(;mH
))). From that bound, using the Riemann-Roch theorem for curves, we obtain the above bound for the degree of quotients ofE
jC.Thus, we are considering only a nite number of Quot schemes.
Step 3:
Since the schemes Quotm a are projective over jmH
j, their images dominate jmH
j if and only if at least one Quotm a is surjective over jmH
j. If they do not cover jmH
j, then the restriction ofq
E
to the general ber ofp
is semistable. In this case we dene the numberA
(m
) to be zero. Otherwise we deneA
(m
) byA
(m
) := maxc
1(E
):
(mH
);2a
2a < c
1(E
):
(mH
) andQuotm a !j
mH
j is surjective.:
By denitionA
(m
) measures how far the restriction ofE
to the general curve ofjmH
jis from being semistable. The projectivity of the Quot schemes implies that ifA
(m
)>
0, then the restriction ofE
to any curveC
] 2 jmH
j has a quotientQ
with Hilbert polynomial (Q
(k
)) = (mH
2)k
+ 12(c
1(E
):
(mH
);A
(m
)) + 1;g
m:
We will apply this specialization property (see also 13]) in the sequel to re- ducible curves
C
] 2 j2mH
j withC
=C
0C
00 whereC
0 andC
00 are smooth curves in jmH
j, to boundA
(2m
) in terms ofA
(m
).Step 4:
1 From now on we assume thatA
(m
) is positive. We setb
(m
) :=2(
c
1(E
):
(mH
);A
(m
)). By denition ofA
(m
) the subsetY
:=a<b(m)im(Quotm a !j
mH
j)2.1 The semistable restriction theorem 5 is a proper closed subset ofj
mH
j. For all points of the open subsetU
m0 =jmH
jnY
the restriction ofE
the corresponding hyperplane has a minimal destabilizing quotient of degreeb
(m
). The minimality of the destabilizing quotient implies its uniqueness. Therefore the restriction Quotm b(m) jmHjU
m0 of the Quot scheme Quotm b(m) toU
m0 gives a bijection of geometric points of QuotUm0 :=Quotm b(m)jmHj
U
m0 andU
m0 . Thus,p
Um0 : QuotUm0 !U
m0 is an isomorphism or completely inseparable. IfE
jC ////F
] is a geometric point of QuotUm0 , then we have Hom(ker()F
) = 0. Therefore the relative tangent bundle ofp
Um0 vanishes. Eventually, we conclude thatp
Um0 is an isomorphism.Step 5:
IfC
] is a smooth curve inU
m0 , then the minimal quotient of degreeb
(m
) has to be a quotient line bundle. Therefore, by considering the open subsetU
m ofU
m0 parametrizing smooth curves, we obtain the following situation:U
m oo p CUm q //X
and a destabilizing quotient line bundle
L
ofq
E
which isp
-at. Furthermore the degree ofL
on all bers ofp
isb
(m
). The surjectionq
E
!L
denes the following diagram:CUm p
//
q
##FFFFFFFFF P(
E
)U
mX
For a curve
C
X
which is parametrized byU
m we call (p
;1C
]) its canonicalm
-lifting.Step 6:
Now we take two smooth curvesH
1 andH
2 inX
which meet transver- sally and which are contained inU
1jH
j. The pencil spanned by these curves denes a rational map P1 _ _ _ //U
1 // Quot1b(1)U
1. Since Quot1b(1) is projective, we obtain a morphismP1 !Quot1b(1). This corresponds to a at family of degreeb
(1) quotients for all restrictions ofE
to curves of the pencil.To be precise we have the following situation: P1 p
X
~ !qX
where ~X
denotes the blow up ofX
in the points ofH
1\H
2, and a destabilizingp
-at quotientq
E
!Q
which for allp
-bers is of degreeb
(1). Over P1 \U
1 the quotientQ
is a line bundle (see step 5).Q
is at and its restriction to most bers is torsion free. Hence,Q
itself is torsion free of projective dimension at most one. The kernelK
ofq
E
!Q
is a line bundle on ~X
which is isomorphic toq
L
+PHi=12a
iE
i where thefE
igi=1 ::: H2 are the exceptional bers of the blow upq
. Restricting the exact sequence 0 !K
!q
E
!Q
! 0 toE
i we see that the integera
i is at least zero.Step 7:
It results thatQ
is of the form O(q
c
1(E
);q
L
;Pa
iE
i)JZwhereJZ denotes the ideal sheaf of a closed subscheme
Z
of ~X
of nite length.Semistability of
E
implies:H:D
0 (1)where
D
= (2L
;c
1(E
)). Chern class computation gives:c
2(E
) =c
1(K
)c
1(Q
) + length(Z
)=
L:c
1(E
);L
2+Pa
2i+ length(Z
) It followsX
a
2ic
2(E
) +L
2;L:c
1(E
) =c
2(E
);c
1(E
)2 4 +D
24 (2)
The discrepancy to semistability is the number
A
(1)A
(1) = 2c
1(K
):
H
;XE
i;c
1(E
):H
=D:H
+ 2Xa
i (3) Now we use the inequalityH2
X
i=1
a
ivu ut
H
2 HX2
i=1
a
2i and inequalities (2) and (3) to deduce:A
(1)D:H
+ 2qH
2(c
2;c1(E)24 ) +
H
2D42
D:H
+p;H
2(E
) +H
2D
2By the Hodge index theorem
H
2D
2(D:H
)2. Thus, we eventually obtain:A
(1)D:H
+p(D:H
)2;H
2(E
):
The basic properties of the function
x
7!x
+px
2;H
2(E
) together with (1) give a bound forA
(1): If (E
)0, thenA
(1) = 0. For (E
)<
0 we have the upper boundA
(1)<
p;H
2(E
).Step 8:
Take a reducible curveC
=C
0C
00 whereC
0C
002jmH
jare smooth curves which intersect transversally. The singular divisor ofC
consisting ofm
2H
2 nodes we denote byD
. LetE
jC !Q
be a torsion free quotient ofE
with Hilbert polynomial Q(k
) =(Q
OX(k
)) = 2m
H
2k
+b
+ 1;g
2mwhere
g
2m denotes the arithmetic genus ofC
. Torsion free means:Q
does not contain a subsheaf of dimension zero. By the Mayer-Vietoris exact sequence0!OC !OC0 OC0 0!OD !0
we see that
g
2m = 2g
m+m
2H
2;1. Furthermore, we obtain from this exact sequence the following diagram with exact rows and surjective columns0 //
E
jC//
E
jC0E
jC0 0//
E
jD//0
TorO1C(
Q
OD) //Q
//Q
jC0Q
jC0 0 //Q
jD //02.1 The semistable restriction theorem 7 Since TorO1C(
Q
OD) is concentrated inD
, andQ
was assumed to be torsion free, the image of TorO1C(Q
OD) inQ
is zero. Therefore, the equality (Q
) + length(Q
jD) = (Q
jC0) +(Q
jC00) (4) holds true. There are three cases for the ranks ofQ
jC0 andQ
jC0 0. The pair (rk(Q
jC0) rk(Q
jC0 0)) has to be (1 1), (2 0), or (0 2).Step 9:
We next show that if the ranks ofQ
jC0 andQ
jC0 0do not coincide, then the quotientQ
is not destabilizing. Assume thatQ
jC0 has rank two andQ
jC00is torsion. It results that
Q
jC0 is isomorphic toE
jC0, andQ
jC00 is isomorphic toE
jD. Hence,Q
is isomorphic toE
jC0. Therefore we nd (Q
(k
)) =(E
(k
));(E
(k
;m
)) = (2mH
2)k
+(E
);(E
(;m
)):
Analogously we compute the Euler characteristic ofE
jC to be (E
jC(k
)) = (4m
H
2)k
+(E
);(E
(;2m
)):
In order to prove that
Q
is not destabilizing we must show that the inequality (E
);(E
(;m
))2
mH
2>
(E
);(E
(;2m
)) 4mH
2 holds. This inequality is equivalent to2(
(E
);(E
(;m
)))>
(E
);(E
(;2m
)):
The last inequality holds because the function
k
7! (E
(k
)) is strictly convex by the Riemann-Roch theorem for surfaces.Step 10: (General intersection lemma)
LetC
00 be an irreducible curve inX
with a lifting ~C
00 to P(E
). For a general curveC
0] inU
m its canonicalm
-lifting ~C
0 in P(E
) intersects ~C
00 in zero orC
0:C
00 points.Proof:
For brevity, we writeU
instead ofU
m. We consider the following situation:CUX
C
00 //
666666666666666666
C
00
CUP(E)
C
~00o oo 'oo oo oo oo 77 //
CU //
IIIII $$IIIII
X C
~00 ~ //P(E
)uuuuuu ::uuu
U
Since
C
00 !X
and CU !U
are projective morphisms, so is . The same way, we see that the composition morphism'
is projective. For a geometric pointC
0]2U
the ber ofoverC
0] is the intersectionC
0\C
00. Thus,is of relative dimension zero. Analogously we identify the ber of'
with the intersection of the liftings. By constructionCU is an open subset in aPn-bundle overX
. We conclude the irreducibility ofCUXC
00. Thus we see, that if'
is a dominant morphism, then the canonical lifting ~C
0 of a general curveC
0 intersects ~C
00 inC
0:C
00 points. If the morphism'
is not dominant, then the canonical lifting of a general curveC
0 is disjoint from ~C
00. 2(of step 10)Step 11:
LetC
00] be a point inU
m, and ~C
00 be its canonical lifting to P(E
).Let us assume that the lifting ~
C
0 of a general curveC
0]2U
m intersects ~C
00 inC
0:C
00 points. If we consider the pencil spanned byC
00 andC
0, then we obtain (see step 6) a familyQ
over this pencil with alla
i equal to zero. Indeed, if onea
i is positive, then them
-lifting of general curve contained in the pencil spanned byC
0 andC
00 does not intersect ~C
00in the pointP
i. This would imply (see (1) and (3) of step 7) thatA
(m
) = 0.By the above lemma we can assume that
C
0]C
0 0]2U
m are two smooth curves whose canonical liftings are disjoint. We consider now forC
=C
0C
00 a minimal quotientQ
ofE
jC having rank one onC
0 andC
00. We call a pointP
2D
=C
0\C
00 a point of discord if the dimension ofQ
k
(P
) is two. LetM
be the number of points of discord. It is obvious that the maximal torsion subsheaf ofQ
jC0 is concentrated in the points of discord and has lengthM
. The quotient ofQ
C0 modulo its torsion is denoted byQ
0, and analogous we have theC
00-line bundleQ
00. We obtain from (4) that (Q
) = (Q
0) +(Q
00) +M
;m
2H
2 (5)Step 12:
The inequalityA
(2m
)2A
(m
);2m
2H
2]+ holds.Proof:
We consider the unique destabilizing quotientL
0 ofE
jC0. The kernel ofE
jC0 !L
we denote byF
0. We now consider the composition 0 :F
0 !E
jC0 !Q
0.Case 1:
The morphism0 (or0 0) is not trivial.If
0 is not trivial, then it follows that deg(Q
0)>
deg(F
0) and (Q
0(k
)) (F
0(k
)) =(F
0 0(k
)). We obtain from (5) that (Q
(k
)) (F
0 0(k
)) +(Q
0 0(k
));m
2H
2(
F
0 0(k
)) +(L
0 0(k
));m
2H
2=
(E
jC0 0(k
));m
2H
2= 12(
(E
jC(k
))):
Thus, in this case the quotient
Q
is not destabilizing.Case 2:
If0 is trivial, then we obtainF
0 ker(E
C0 !Q
0). HoweverF
0 is the rank one subbundle ofE
C0 of maximal degree. Hence we haveQ
0 =L
0. Since the canonical liftings ofC
0 andC
00do not intersect, we must haveM
=m
2H
2. The equality (5) consequently yields: (Q
(k
)) = 2(L
0(k
))=
(E
j0C(k
));A
(m
)= 12
(E
jC(k
)) + (m
2H
2;A
(m
)) This gives the asserted inequality forA
(2m
).To complete the proof of step 12 we just remark that a similar computation shows that
A
(m
) = 0 impliesA
(2m
) = 0. 2(of step 12)Step 13:
Using induction onl
, we obtain from the inequality of step 12A
(2l)2lA
(1);2l(2l;1)H
2]+:
2.2 A boundedness result 9 Combining this with the upper bound for
A
(1) computed in step 7 the theoremfollows. 2
Remark 1:
The theorem still holds true for a semistable coherentX
-sheafE
of rank two. Indeed, consider the embeddingE
!E
__ ofE
into its double dual. In this way, we obtain the semistable vector bundleE
__ which obviously satises (E
__)(E
). Hence, the theorem applies.Remark 2:
We can extend the theorem to projective surfacesX
with isolated singularities. Reviewing the proof, we see that it is enough to have smooth curves in the linear systems jmH
j. By the same argument, we see that the theorem holds true if we requireOX(1) to be base point free.Remark 3:
If we know the ideal fL:H
gL2Pic(X) Zof intersections withH
, then we can sharpen the inequality of step 7. To illustrate this, let us assume that the Picard group ofX
is generated byOX(1). Furthermore, suppose thatE
is a semistableX
-vector bundle. We have det(E
) =nH
. If (E
)<
0, then we can improve the bound forA
(1) of step 7 byA
(1) ;2H
2+p4(H
2)2;H
2(E
) forn
even;
H
2+p(H
2)2;H
2(E
) forn
odd.2.2 A boundedness result
Let
X
,OX(1) be as before. Furthermore, letE
be a semistableX
-vector bundle of rank two. We next give a boundM
depending only on the characteristic numbersc
2(E
) andc
1(E
):H
suchE
(M
) becomes globally generated. We use Mumford's concept ofm
-regularity:A coherent sheaf
E
on a polarized varietyX
with very ample line bundleOX(1) is calledm
-regular, ifh
i(E
(m
;i
)) = 0, for alli >
0.The following lemma (cf. x14 in 11]) resumes properties of
m
-regular sheaves.Lemma 2.2
LetX
be a projective variety with a very ample line bundleOX(1) = OX(
H
), andE
be a coherentX
-sheaf. IfE
ism
-regular, then it is globally generated, andE
(m
+k
)-regular for allk
0.Let
D
2 jH
j be a divisor such that the sequence 0!E
(;1) !E
!E
jD ! 0 is exact. IfE
jD ism
-regular, thenE
is(m
+h
1(E
(m
;1)))-regular. 2 This lemma outlines our strategy. We rst show that for a suitable curveC
2 jH
j and an integerm
1 the restrictionE
C =E
jC is (m
1 + 1)-regular. In order to obtain the boundedness result, we then compute an upper bound forh
1(E
(m
1)).Lemma 2.3
LetE
C be a rank two vector bundle on a smooth curveC
of genusg
de ned over an algebraically closed eld. We de ne the numberA
to be zero ifE
C is semistable. Otherwise we setA
= maxfdeg(E
C);2deg(Q
)jE
C !Q
is surjective, and rk(Q
) = 1g:
(1) IfL
is aC
-line bundle with deg(L
)>
A;deg(E2 C)+ 2g
;2,then
H
1(C E
CL
) = 0(2) For any
C
-line bundleL
the inequalityh
0(E
CL
)2h1 + deg(L
);deg(E2C);Ai+ holds true.Proof:
(1) Ifh
1(E
CL
)>
0, then there exists, by Serre duality, a non trivial homomorphism'
:E
!!
CL
;1. (Here!
C denotes the dualizing sheaf ofC
.) Thus, the image of'
is a rank one quotient of degree at most 2g
;2;deg(L
).By the very denition of the number
A
we obtain deg(L
) A;deg(E2 C)+2g
;2.(2) Analogously, we see that for deg(
L
)<
deg(E2C);Athere are no global sections ofE
CL
. Thus, the assertion holds for all line bundlesL
of degree less than deg(E2C);A. LetP
2C
be a geometric point ofC
. Then from the exact sequence 0 !E
CL
(;P
) !E
CL
!E
CjP ! 0 we obtainh
0(E
CL
) 2 +h
0(E
CL
(;P
)) which proves the second statement. 2 We now take a smooth curveC
of genusg
in the linear system jH
j such that for the restrictionE
C the numberA
of the above lemma is at mostp;
H
2(E
)]+. We have seen in step 7 of the proof of theorem 2.1 that this is possible. The adjunction formula gives 2g
;2 =H:
(H
+K
X). Obviously, the degree of theC
-line bundleOX(mH
)jC ismH
2. Thus, settingm
1 :=$ 1
H
2p;
H
2(E
)]+;c
1(E
):H
2 +
H:
(H
+K
X)!%
+ 1 we obtain by lemma 2.3 that
E
C is (m
1+ 1)-regular.The semistability of
E
implies thath
0(E
(m
2;1)) = 0, form
2 := l;H:c2H12(E)m. Applying the inequalityh
0(E
(m
))< h
0(E
(m
;1))+h
0(E
C(m
)) obtained from the long exact cohomology sequence yieldsh
0(E
(m
1))m
3 := 2 Xm1m=m2
"
1 +
mH
2;c
1(E
):H
;p;H
2(E
)]+2
#
+
:
Sinceh
2(E
(m
1)) = 0 we deduce thath
1(E
(m
1))m
3 ;(E
(m
1)). Settingm
4:=m
1+m
3;(E
(m
1)), we obtain by lemma 2.2:Proposition 2.4
LetX
be a smooth projective surface over an algebraically closed eld, and OX(1) a very ample line bundle onX
. Furthermore, letE
be a rank twoX
-bundle which is semistable with respect to OX(1). Then form
m
4 we have thatE
(m
) is globally generated. The numberm
4 de ned above depends only on the characteristic numbers ofE
. 2 It follows that any semistable sheafE
of rank two with givenc
1(E
):H
,c
1(E
):K
X, andc
2(E
) is a quotient of OX(;m
4) E(m4). Considering the Quot scheme QuotOEX(;m4) E(m4)=X together with its natural SL E(m4)-action we ob- tain (see 10]) the coarse moduli space of semistable coherent sheaves onX
with Hilbert polynomialE. This proves the next corollary.Corollary 2.5
There exists a projective coarse moduli space for semistable co- herent sheaves of rank two with xed characteristic numbers on a smooth pro- jective surface.2.3 Further applications 11
2.3 Further applications
Proposition 2.6
LetX
be a smooth projective surface over an algebraically closed eld with a very ample line bundleOX(1) =OX(H
). IfE
is aX
-bundle, which is stable with respect toOX(1) and of rank 2, then the inequality(
E
)8>
<
>:
1;4
(OX) ifK
X:H <
02;4
(OX) ifK
X:H
= 0h6;4
(OX) + 4lKHX2:HmK
X:H
i+ ifK
X:H >
0:
holds.Proof:
We compute, using the Riemann-Roch theorem for surfaces that (E
E
_) = (E
) + 4(OX):
The stability of
E
implies thatH
0(E
E
_) = Hom(E E
) is of dimension one. Now we want to boundh
2 :=h
2(E
E
_). By Serre duality,h
2 equals the dimension of Hom(E E
(K
X)) whereK
X denotes the canonical class onX
. Thus, we obtain forK
X:H
0h
2(E
E
_) 0 ifK
X:H <
0 1 ifK
X:H
= 0:
If
K
X:H >
0, we setm
=dKHX2:Heand consider a smooth curveC
in the linear system jmH
j. If (E
) 0, then by theorem 2.1, we may assume that the restrictionE
jC is semistable. Thus, Hom(E
jCE
jC) is at most of dimension 4.By induction we see that for a
C
-line bundleL
of degreed
we can bound the dimension of Hom(E
jCE
jCL
), by 4 + 4d
.By denition of
m
, we have (K
X ;mH
):H
0. Thus, we can bound the dimension of Hom(E E
(K
X ;C
)) by one. From the exact sequence0!Hom(
E E
(K
X ;C
))!Hom(E E
(K
X))!Hom(E E
(K
X)jC) we obtain the estimateh
2(E
E
_)5 + 4K
X:H H
2K
X:H :
Applying the obvious inequality
(E
E
_)h
0(E
E
_) +h
2(E
E
_) weobtain the estimation of the proposition. 2
Corollary 2.7 (Weak Bogomolov inequality)
LetX H
be a very ample polarized smooth surface over an algebraically closed eld. LetE
be a rank 2 vector bundle onX
satisfying(
E
)>
8>
<
>:
1;4
(OX)]+ ifK
X:H <
02;4
(OX)]+ ifK
X:H
= 0h6;4
(OX) + 4lKHX2:HmK
X:H
i+ ifK
X:H >
0:
ThenE
is Bogomolov unstable.Proof:
By proposition 2.6E
cannot be stable with respect to the given polar- izationH
. Thus, we have a short exact sequence0!
A
!E
!JZ(c
1(E
);A
)!0where
Z
X
is a closed subscheme of codimension 2. SinceA
is destabilizing we have (c
1(E
);2A
):H
0. Using the exact sequence to computec
2(E
) yields(
c
1(E
);2A
)2 = (E
) + 4length(Z
)>
0:
Thus, the Hodge index theorem implies (
c
1(E
);2A
):H <
0. Consequently,E
is not semistable. Now the half c^one in the Neron-Severi group NS(X
) dened by positive self intersection and negative intersection with an ample classH
does not depend on
H
. 2Remark:
G. Megyesi proves that for a vector bundleE
of arbitrary rank on a smooth surface dened over a eld of characteristicp >
0 with (E
)>
0 the pullback (F
n)E
ofE
by a large powern
of the absolute FrobeniusF
is Bogomolov unstable (see 7]). Corollary 2.7 gives an e ective bound forn
, for a vector bundleE
of rank two.Corollary 2.8 (Weak Kodaira vanishing)
LetX
be a smooth projective surface de ned over an algebraically closed eld with very ample line bundleOX(
H
). LetL
be a nefX
-line bundle such thatL
2>
8>
<
>:
1;4
(OX)]+ ifK
X:H <
02;4
(OX)]+ ifK
X:H
= 0h6;4
(OX) + 4lKHX2:HmK
X:H
i+ ifK
X:H >
0:
Then the rst cohomology groupH
1(X L
;1) vanishes.Proof:
We take an extensionE
of OX byL
;1. SinceH
1(X L
;1) = Ext1(OXL
;1), we have to show that the short exact sequence0!
L
;1 !E
!OX !0splits. We compute
c
1(E
) =;L
,c
2(E
) = 0, and (E
) =L
2. Consequently, by corollary 2.7,E
has a destabilizing subsheafA
of rank one with (2A
+L
)2>
0, and (2A
+L
):H >
0, for all ample classesH
. Since nef bundles are limits of ample classes, we obtain(2
A
+L
):L
0:
(1)By the same reason, the Hodge index theorem applies
A
2 (A:L
)2L
2:
(2)The subsheaf