Professional Documents
Culture Documents
1703 04851
1703 04851
DIMENSION ON P3
1. Introduction
triple
0 E 2OP3 L(2) 0,
where L is a line bundle over C with Hilbert polynomial PL (k) = 3k such that
L2 6 C ;
(iv) the four components T (s), s = 1, ..., 4, of dimensions 21 + 4s, respectively, whose
general points correspond to sheaves E having singularities along the sets of s distinct
points {q1 , ..., qs } P3 and satisfying the exact triple
s
M
0 E F Oqi 0,
i=1
is a dense open subset of Gr(2, 4) Syms (P3 ) . For (l, Z) X[F ] set
n
r= + 1 s, Qu,r := Ol (r) OZ ,
2
and consider the open dense subset Hom(F, Qu,r )e Hom(F, Qu,r ) consisting of epimor-
phisms : F Qu,r . For any element Hom(F, Qu,r )e we can take the kernel of
:
E = ker .
Xe(n, r, s) = {x = ([F ], u, [x]) | [F ] R(0, 2, n), u X[F ] , [x ] Hom(F, Qu,r )e /Aut(Qu,r )}.
Since R(0, 2, n) are reduced irreducible schemes (see [1]), the sets Xe(n, r, s) have a natural
structure of reduced irreducible schemes. Furthermore, by the above, there is a well-
defined injective morphism
Set
n
X (n, r, s) := f (Xe(n, r, s)) M(3), r+s= + 1, r, s 0.
2
The dimensions of the varieties X (n, r, s) are calculated by the following formula
dim X (n, r, s) = dim Xe(n, r, s) = dim R(0, 2, n) + dim Gr(2, 4) + dim Syms (P3 )+
dim Ext1 (E, E) = dim X (n, r, s) for any point [E] X (n, r, s).
4 A. N. IVANOV AND A. TIKHOMIROV
In Section 3 it will be shown that this equality is satisfied for families X (2, 2, 0), X (4, 3, 0)
and X (4, 2, 1), this giving the proof of Theorem 3. From the construction of the families
it is clear that general points of these irreducible components correspond to sheaves with
singularities along l Sing(F ) Z. Since dim Sing(F ) = 0, these singularities have mixed
dimension.
Throughout this work, the base field k is an algebraically closed field of character-
istic 0. Also, for simplicity, we will not distinguish between a stable rank-2 OP3 -sheaf E
and its isomorphism class [E] as a point in the Gieseker-Maruyama moduli scheme.
Acknowledgements. AST was supported by a subsidy to the HSE from the
Government of the Russian Federation for the implementation of Global Competitive-
ness Program. AST also acknowledges the support from the Max Planck Institute for
Mathematics in Bonn, where this work was finished during the winter of 2017.
Ext1 (IYn (1), OP3 (1)) H 0(Ext1 (IYn (1), OP3 (1))) H 0 (Ext2 (OYn (2), OP3 ))
corresponds to the global section H 0 (Ext2 (OYn (2), OP3 )) such that ()0 = {xi }ni=1 are
n distinct points of the component of Yn which is not a line, and
n
M
(2) Ext1 (F, OP3 ) Oxi .
i=1
0 F (1) Hom(F, F ) G 0,
Proof. Applying the functor Hom(, F ) to (1) we obtain the exact sequence
(4) 0 Hom(IYn (1), F ) Hom(F, F ) F (1)
Ext1 (IYn (1), F ) Ext1 (F, F ).
(5) 0 L1 L0 F 0
of the sheaf F . Applying the functor Hom(IYn (1), ) to (5) and considering an easily
verifiable equality Ext2 (IYn (1), L1 ) = 0, we obtain an exact triple
(6) Ext1 (IYn (1), L1 ) Ext1 (IYn (1), L0 ) Ext1 (IYn (1), F ) 0.
An easy computation shows that the sheaves Ext1 (IYn (1), L0 ) and Ext1 (IYn (1), L1 ) are
locally free OYn -sheaves. Hence, Ext1 (IYn (1), F ) is OYn -sheaf and, moreover, it is generi-
cally a rank 2 locally free OYn -sheaf. Furthermore, the morphism in (4) factors through
the morphism of restriction OYn : F (1) F (1) OYn . Since F (1) OYn is generically
a locally free OYn -sheaf of rank 2, this yields = j (OYn ), where j is a generically in-
jective morphism of OYn -sheaves. Consequently, since ker(OYn ) = F IYn (1), it follows
that the sheaf G := ker satisfies the exact triple
On the other hand, applying the functor Hom(OYn (1), ) to (5) yields an exact triple
... Ext1 (OYn (1), L0 ) Ext1 (OYn (1), F ) Ext2 (OYn (1), L1 ) ...
Since F is locally free outside the points {xi }ni=1 , the sheaf Ext1 (OYn (1), F ) either equals
to zero or is an artinian sheaf. On the other hand, Ext1 (OYn (1), L0 ) = 0 and the sheaf
Ext2 (OYn (1), L1 ) is locally free OYn -sheaf, since L0 and L1 are locally free OP3 -sheaves.
6 A. N. IVANOV AND A. TIKHOMIROV
Hence, the sheaf Ext1 (OYn (1), F ) is a subsheaf of Ext2 (OYn (1), L1 ), so it cannot be a
non-zero artinian sheaf. Consequently, Ext1 (OYn (1), F ) = 0 and (8) follows from (10).
The statement of Lemma now follows from (7) and (8).
h2 (Hom(F, F )) = 0.
Proof. Applying the functor OYn (1) to (1) we obtain the exact sequence
1 0
(11) T or1 (IYn (2), OYn ) OYn F (1) OYn IYn (2) OYn 0.
It is easy to see the sheaves F (1) OYn and IYn (2) OYn are rank-2 OYn -sheaves and 0
is an epimorphism. So ker 0 is artinian sheaf and the morphism 1 is non-zero on each
component of the curve Yn . Next, standard computation gives the following
O O 2 (1) for n = 2,
m C
T or1 (IYn (2), OYn ) T or2 (OYn (2), OYn ) det NYn /P3 (2)
OC 3 (4pt) for n = 4.
L
n
So we obtain coker 1 = ker 0 Oxi due to Sing(F ) = {xi }ni=1 that is the following
i=1
L
n
triple is exact 0 Oxi F (1) OYn IYn (2) OYn 0. This triple as a
i=1
triple of the OYn -sheaves on the smooth curve Yn is splitted. Thus we have the following
isomorphism
n
M
(12) F (1) OYn Oxi IYn (2) OYn .
i=1
Next, we have
2O (1) O 2 O 2 (1) for n = 2,
m C C
(13) IYn (2) OYn NYn /P3 (2)
OC 3 (pt) OC 3 (pt) for n = 4.
It is easy to see that T or1 (F, OYn (1)) = 0 because hdF = 1, so the sheaf F IYn is torsion
free and the following triple is exact
Theorem 3. The closures X (2, 2, 0), X (4, 3, 0), X (4, 2, 1) are irreducible components of
the moduli scheme M(3), of dimensions 22, 24, 26, respectively.
Proof. Consider an arbitrary sheaf [E] X (n, r, s) which according to the construction
satisfies the following exact triple
(17) 0 E F Ol (r) OZ 0,
(19) Ext1 (OP3 , F ) = Ext2 (F, F ) = 0, dim Ext1 (F, OP3 ) = dim Ext1 (F, F ) = 0.
1
(21) : Ext1 (F, E) Ext1 (F, F ).
(24) Ext1 (F, F ) Ext1 (F |U , F |U ) Ext1 (E|U , F |U ).
E|U F |U , U := P3 \ (l Z),
that follows from (17). The inclusion of the artinian sheaf Ext1 (F, F ) into the sheaf
Ext1 (E, F )
Besides, Ext2 (Ol (r) OZ , OU ) (det Nl/P3 )(r) Ol (2 r). Applying the functor
Hom(, OU ) to the exact triple (17) restricted on U and taking into account that the
sheaves Ext1 (F |U , OU ) and Ext2 (F |U , OU ) are vanished because the sheaf F |U is locally
free we have Ext1 (E|U , OU ) Ol (2 r). So Ext1 (E|U , OU ) F |U 2Ol (2 r) due
to (18). The isomorphism Ext1 (E, F )|U 2Ol (2 r) follows from that and (27). Hence,
from (26) follows an isomorphism
Next, in view of (22) applying the functor Hom(Ol (r) OZ , ) to the triple(17) yields
an isomorphism
From here in view of (20) and (21) it follows that the following triple is exact
Next, applying the functor Hom(, E) to (17) and taking into account (25) and (29) we
have the exact triple
(33) 0
=
0 0 // Ol OZ //
0 // Hom(F, E) // Hom(F, F ) // 2Ol (r) 2OZ // 0
0 // Hom(E, E) // Hom(E, F ) // coker // 0
=
// Ol OZ // 0 0
0
It is easy to see that coker |l is a locally free Ol -sheaf (see [6, proof of Lemma 5.1]), so
the right vertical triple in this diagram yields
Applying the functor Hom(E, ) to (17) and taking into account (35) we obtain the exact
sequence
On the other hand, applying the functor Hom(, Ol (r) OZ ) to (17) and taking into
account (23) and standard isomorphisms
(38) Ext1 (E, Ol (r) OZ ) Ext2 (Ol (r) OZ , Ol (r) OZ ) Ol (2) 3OZ .
It is easy to see that im|l is locally free Ol -sheaf in (37), so im is isomorphic to Ol (2r)
OZ . Hence, the following triple is exact
since, in view of (19), Ext2 (F, E) Ext2 (F, F ) 0 and hd(Ol ) = 2. On the other
hand, from (28) it follows that Ext1 (E, F )|Z = 0. So ker |Z = 0 and, in view of (40),
(38), there is an isomorphism
Next, the triple (35) in view of the diagram (33) can be written by the following way
Since E and F are stable, they are simple, i. e. h0 (Hom(E, E)) = h0 (Hom(E, E)) = 1.
Therefore from Lemma 2 and (44) it follows that
(46) h0 (Ext1 (F, F )) + h1 (Hom(F, F )) = dim Ext1 (F, F ) = dim R(0, 2, n) = 13.
we obtain
18 + 4s + 2r for r 2,
dim Ext1 (E, E) =
21 + 4s for r = 0 or 1.
Thus we have the following table of dimensions from which the statement of Theorem
follows.
The family Dimension of the family dim Ext1 (E, E)
X (2, 2, 0) 22 22
X (2, 1, 1) 24 25
X (2, 0, 2) 26 29
X (4, 3, 0) 24 24
X (4, 2, 1) 26 26
X (4, 1, 2) 28 29
X (4, 0, 3) 30 33
For the rest families X (2, 1, 1), X (2, 0, 2), X (4, 1, 2), X (4, 0, 3) which are not in-
cluded into the statement of Theorem 3 we formulate the following conjecture.
Conjecture. The following inclusions are true:
References
[1] M.-C. Chang, Stable rank 2 reflexive sheaves on P3 with small c2 and applications.
Trans. Amer. Math. Soc. 284 (1984), 5789.
[2] G. Ellingsrud, S. A. Stromme, Stable rank 2 vector bundles on P3 with c1 = 0 and
c2 = 3. Math. Ann. 255 (1981), 123135.
[3] R. Hartshorne, Stable vector bundles of rank 2 on P3 . Math. Ann. 254 (1978), 229
280.
[4] R. Hartshorne, Stable Reflexive Sheaves, Math. Ann. 254 (1980), 121176.
[5] D. Huybrechts, M. Lehn, The Geometry of Moduli Spaces of Sheaves, 2nd ed., Cam-
bridge Math. Lib., Cambridge University Press, Cambridge, 2010.
[6] M. Jardim, D. Markushevich, A. S. Tikhomirov, New divisors in the boundary of the
instanton moduli space. Preprint arXiv:1501.00736 [math.AG].
[7] M. Jardim, D. Markushevich, A. S. Tikhomirov, Two infinite series of moduli spaces
of rank 2 sheaves on P3 . Annali di Mat. Pura ed Appl. First online (2017), pp. 1-36.
doi:10.1007/s10231-016-0630-3.