MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE
|
|
- Easter Lee
- 6 years ago
- Views:
Transcription
1 Dedicated to the memory of Andrei Nikolaevich Tyurin MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE ALEXANDER S. TIKHOMIROV 1. Introduction By a mathematical n-instanton vector bundle (shortly, a n-instanton) on 3-dimensional projective space P 3 we understand a rank-2 algebraic vector bundle E on P 3 with Chern classes (1) c 1 (E) =, c 2 (E) = n, n 1, satisfying the vanishing conditions (2) h (E) = h 1 (E( 2)) =. Denote by I n the set of isomorphism classes of n-instantons. This space is nonempty for any n 1 - see, e.g., [BT], [NT]. The condition h (E) = for a n-instanton E implies that E is stable in the sense of Gieseker-Maruyama. Hence I n is a subset of the moduli scheme M P 3(2;, 2, ) of semistable rank-2 torsion-free sheaves on P 3 with Chern classes c 1 =, c 2 = n, c 3 =. The condition h 1 (E( 2)) = for [E] I n (called the instanton condition) by the semicontinuity implies that I n is a Zariski open subset of M P 3(2;, 2, ), i.e. I n is a quasiprojective scheme. It is called the moduli scheme of mathematical n-instantons. In this paper we study the problem of the irreducibility of the scheme I n. This problem has an affirmative solution for small values of n, up to n = 5. Namely, the cases n = 1, 3, 3, 4 and 5 were settled in papers [B1], [H], [ES], [B3] and [CTT], respectively. The aim of this paper is to prove the following result. Theorem 1.1. For each n = 2m + 1, m, the moduli scheme I n of mathematical n- instantons is reduced and irreducible of dimension 8n 3. A guide to the paper is as follows. In section 3 we remind a well-known relation between mathematical n-instantons and nets of quadrics in arithmetic n-dimensional vector space k n. The nets of quadrics are considered as vectors of the space S n = S 2 (k n ) 2 V, where V = H (O P 3(1)), and those nets which correspond to n-instantons (we call them n-instanton nets) satisfy the so-called Barth s conditions - see definition (13). Thus the description of the moduli space I n of n-instantons reduces to that of the locally closed subset MI n S n of n-instanton nets of quadrics which is crucial for our study. In section 4 we prove one result of general position for the set of (2m + 1)-instanton nets of quadrics MI 2m+1, m 1. Essentially, this result means that the natural map MI 2m+1 S m+1 induced by a generic embedding k m+1 k 2m+1 is dominating - see Remark 8.1. Section 5 is a study of some linear algebra related to a direct sum decomposition ξ : k m+1 k m k 2m+1 giving the above embedding k m+1 k 2m+1. Using the result of section 4 we obtain here the relation (61) which is a key instrument for our further considerations. Also, the decomposition ξ enables us to relate (2m + 1)-instantons E to rank-(2m + 2) symplectic vector bundles E 2m+2 on P 3 satisfying the vanishing conditions h (E 2m+2 ) = h 2 (E 2m+2 ( 2)) =. In section 6 we introduce a new scheme X m as a locally closed subset of the vector space S m+1 Hom(k m, (k m+1 ) 2 V which is defined by linear algebraic data somewhat similar to Barth s conditions. We prove that X m as a reduced scheme is isomorphic to a certain dense
2 2 TIKHOMIROV open subset MI 2m+1 (ξ) of MI 2m+1 determined by the choice of the direct sum decomposition ξ above. This reduces the problem of the irreducibility of I 2m+1 to that of X m. The last ingredient in the proof of Theorem 1.1 is a scheme Z m introduced in section 7 as a closed subscheme of the vector space S m Hom(k m, (k m ) ) 2 V defined by explicit equations. We relate the scheme Z m to the so-called t Hooft instantons. Using the properties of t Hooft instantons (see subsection 5.2) we show that the scheme Z m is reduced and irreducible. In the last section 8 we finish the proof of Theorem 1.1. The proof is based on a study of certain scheme X m containing X m and fibred over the vector space Hom(k, k m+1 ) 2 V. We show that the zero fibre of this projection is scheme-theoretically isomorphic to a direct product of Z m and a certain vector space. This together with the irreducibility of Z m and some other results stated earlier leads to the irreducibility of X m. Acknowledgement. The author acknowledges the support and hospitality of the Max Planck Institute for Mathematics in Bonn where this paper was started during the authors stay there in Winter Notation and conventions Our notations are mostly standard. The base field k is assumed to be algebraically closed of characteristic. We identify vector bundles with locally free sheaves. If F is a sheaf of O X - modules on an algebraic variety or scheme X, then nf denotes a direct sum of n copies of the sheaf F, H i (F) denotes the i th cohomology group of F, h i (F) := dim H i (F), and F denotes the dual to F sheaf, i.e. the sheaf F := Hom OX (F, O X ). If Z is a subscheme of X, by I Z,X we denote the ideal sheaf corresponding to a subscheme Z. If X = P r and t is an integer, then by F(t) we denote the sheaf F O P r(t). [F] will denote the isomorphism class of a sheaf F. For any morphism of O X -sheaves f : F F and any k-vector space U (respectively, for any homomorphism f : U U of k-vector spaces) we will denote, for short, by the same letter f the induced morphism of sheaves id f : U F U F (respectively, the induced morphism f id : U F U F). Everywhere in the paper V will denote a fixed vector space of dimension 4 over k and we set P 3 := P (V ). Also verywhere below we will reserve the letters u and v for denoting the two morphisms in the Euler exact sequence O P 3( 1) u V v O P 3 T P 3( 1). For any k-vector spaces U and W and any vector φ Hom(U, W 2 V ) Hom(U V, W V ) understood as a homomorphism φ : U V W V or, equivalently, as a homomorphism φ : U W 2 V, we will denote by φ the composition U O φ P 3 W 2 V ɛ O P 3 W Ω P 3(2), where ɛ is the induced morphism in the exact triple 2 Ω P 3(2) 2 v 2 V ɛ O P 3 Ω P 3(2) obtained by passing to the second wedge power in the dual Euler exact sequence. Also, shortening the notation, we will omit sometimes the subscript P 3 in the notation of sheaves on P 3, e.g., write O, Ω etc., instead of O P 3, Ω P 3 etc., respectively. Everywhere in the paper for m 1 we denote by S m the vector space S 2 (k m ) 2 V. Following W.Barth [B2], [B3] and A.Tyurin [T1], [T2] we call this space the space of nets of quadrics in the space k m. 3. Some generalities on instantons. Set MI n In this section we recall some well known facts about mathematical instanton bundles - see, e.g., [CTT].
3 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE3 For a given n-instanton E, the conditions (1), (2), Riemann-Roch and Serre duality imply (3) h 1 (E( 1)) = h 2 (E( 3)) = n, h 1 (E Ω 1 P 3) = h2 (E Ω 2 P3) = 2n + 2, h 1 (E) = h 2 (E( 4)) = 2n 2. Furthermore, the condition c 1 (E) = yields an isomorphism 2 E O P 3, hence a symplectic isomorphism j : E E. This symplectic structure j on E is unique up to a scalar, since E as a stable bundle is a simple bundle, i.e. Hom(E, E) = kid. Consider a triple (E, f, j) where E is an n-instanton, f is an isomorphism k n H 2 (E( 3)) and j : E E is a symplectic structure on E. We call two such triples (E, f, j) and (E f, j ) equivalent if there is an isomorphism g : E E such that g f = λf with λ {1, 1} and j = g j g, where g : H 2 (E( 3)) H 2 (E ( 3)) is the induced isomorphism. We denote by [E, f, j] the equivalence class of a triple (E, f, j). From this definition one easily deduces that the set F [E] of all equivalence classes [E, f, j] with given [E] is a homogeneous space of the group GL(k n )/{±id}. Each class [E, f, j] defines a point (4) A n = A n ([E, f, j]) S 2 (k n ) 2 V in the following way. Consider the exact sequences (5) Ω 1 P 3 i 1 V O P 3( 1) O P 3, Ω 2 P 3 2 V O P 3( 2) Ω 1 P, 3 4 V O P 3( 4) 3 V O P 3( 3) i 2 Ω 2 P, 3 induced by the Koszul complex of V O P 3( 1) ev O P 3. Twisting these sequences by E and passing to cohomoligy in view of (2) gives the diagram with exact rows (6) H 2 (E( 4)) 4 V H 2 (E( 3)) 3 V i 2 H 2 (E Ω 2 P ) 3 H 1 (E)) A H 1 (E( 1)) V H 1 (E Ω P 3) where A := i 1 1 i 2. The Euler exact sequence (5) yields the canonical isomorphism ω P 3 4 V O P 3( 4), and fixing an isomorphism τ : k 4 V induces the isomorphisms τ : V 3 V and ˆτ : ω P 3 O P 3( 4). Now the point A = A n in (4) is defined as the composition i 1 =, (7) A : k n V τ k n 3 V f H 2 (E( 3)) 3 V A H 1 (E( 1)) V j j H 1 (E ( 1)) V SD H 2 (E(1) ω P 3) V ˆτ H 2 (E( 3)) V f (k n ) V, where SD is the Serre duality isomorphism. One checks that A n is a skew symmetric map depending only on the class [E, f, j] and not depending on the choice of τ, and that this point A n 2 ((k n ) V ) lies in the direct summand S n = S 2 (k n ) 2 V of the canonical decomposition (8) 2 ((k n ) V ) = S 2 (k n ) 2 V 2 (k n ) S 2 V. Here S n is the space of nets of quadrics in k n. Following [B3], [T1] and [T2] we call A the n-instanton net of quadrics corresponding to the data [E, f, j].
4 4 TIKHOMIROV Denote W A := k n V/ ker A. Using the above chain of isomorphisms we can rewrite the diagram (6) as (9) ker A k n V c A W A ker A A c A = q A (k n ) V. W A Here dim W A = 2n + 2 and q A : W A WA is the induced skew-symmetric isomorphism. An important property of A = A n ([E, f, j]) is that the induced morphism of sheaves (1) a A : W A O P 3 c A (k n ) V O P 3 ev (k n ) O P 3(1) is an epimorphism such that the composition k n O P 3( 1) a A q W A O P 3 A W A O P 3 (k n ) O P 3(1) is zero, and E = ker(a A q A)/ Im a A. Thus A defines a monad a A (11) M A : k n O P 3( 1) a A W A O P 3 with the cohomology sheaf E, (12) E = E(A) := ker(a A q A )/ Im a A. a A q A (k n ) O P 3(1) Note that passing to cohomology in the monad M A twisted by O P 3( 3) and using (12) yields the isomorphism f : k n H 2 (E( 3)). Furthermore, the simplecticity of the form q A in the monad M A implies that there is a canonical isomorphism of M A with its dual which induces the symplectic isomorphism j : E E. Thus, the data [E, f, j] are recovered from the net A. This leads to the following description of the moduli space I n. Consider the set of n-instanton nets of quadrics (i) rk(a : k n V (k n ) V ) = 2n + 2, (ii) the morphism a A : W A O P 3 (kn ) O P 3(1) (13) MI n := A S n defined by A in (1) is surjective, (iii) h (E 2 (A)) =, where E 2 (A) := ker(a A q A)/ Im a A and q A : W A WA is a symplectic isomorphism defined by A in (9) The conditions (i)-(iii) here are called Barth s coditions. These conditions show that MI n is naturally supplied with a structure of a locally closed subscheme of the vector space S n. Moreover, the above description shows that there is defined a morphism π n : MI n I n : A [E(A)], and it is known that this morphism is a principal GL(k n )/{±id}-bundle in the étale topology - cf. [CTT]. Here by construction the fibre πn 1 ([E]) over an arbitrary point [E] I n coincides with the homogeneous space F [E] of the group GL(k n )/{±id} described above. Hence the irreducibility of (I n ) red is equivalent to the irreducibility of the scheme (MI n ) red. The definition (13) yields the following. Theorem 3.1. For each n 1, the space of n-instanton nets of quadrics MI n is a locally closed subscheme of the vector space S n given locally at any point A n MI n by ( ) 2n 2 (14) = 2n 2 5n equations obtained as the rank condition (i) in (13). Note that from (14) it follows that (15) dim [A] MI n dim S n (2n 2 5n + 3) = n 2 + 8n 3
5 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE5 at any point A n MI n. On the other hand, by deformation theory for any n-instanton E we have dim [E] I n 8n 3. This agrees with (15), since MI n I n is a principal GL(k n )/{±id}- bundle in the étale topology. Let S n = {[E] I n there exists a line l P 3 of maximal jump for E, i.e. such a line l that h (E( n) l ) }. It is known [S] that S n is a closed subset of I n of dimension 6n + 2. Thus, since dim [E] I n 8n 3 at any [E] I n, it follows that (16) I n := I n S n is an open subset of I n and (I n) red is dense open in (I n ) red ; respectively, (17) MI n := π 1 n (I n) is an open subset of MI n and we have a dense open embedding (18) (MI n) red dense open (MI n ) red. For technical reasons we will below restrict ourselves to MI n instead of MI n. 4. A result of general position for (2m + 1)-instanton nets Definition 4.1. Let U and U be two vector spaces of dimensions respectively m and n, where m n. Consider the projective space P (U U ). We say that a point x P (U U ) has rank r (and denote this as rk(x) = r), if (i) there exist unique subspaces U r (x) U and U r(x) U of dimensions dim U k (x) = dim U k (x) = r such that x P (U r(x) U r(x)), and (ii) there do not exist subspaces Ũ U and Ũ U of dimension dim Ũ = dim Ũ < r such that x P (Ũ Ũ ). It is well known that each point x P (U U ) has a uniquely defined rank 1 rk(x) n. Fix a positive integer m 3 and a (2m + 1)-instanton vector bundle E such that [E] I 2m+1 and denote H 2m+1 = H 2 (E( 3)) and H 4m = H 2 (E( 4)). The Euler Exact sequence induces the exact triple E Ω P 3 V E( 1) E which gives a natural multiplication map in the first cohomology: (19) H 2m+1 V mult H 4m H 2 (E Ω P 3). Passing to cohomology of the exact triple E Ω 2 P 3 2 V E( 2) E Ω P 3 and using standard equalities = h 2 (E( 2)), h 3 (E Ω 2 P 3 ) = h (E Ω P 3) h (E( 1) V ) = for the instanton bundle E, we obtain: H 2 (E Ω P 3) =. Hence (19) gives the exact triple (2) W 4m+4 H 2m+1 V mult H 4m where (21) W 4m+4 := H 1 (E Ω P 3). We now prove the following main result of this section. Theorem 4.2. Let m 3 and let E be a (2m + 1)-instanton, [E] I 2m+1. Consider the spaces H 2m+1 = H 2 (E( 3)) and W 4m+4 = H 1 (E Ω P 3) together with the injection W 4m+4 H 2m+1 V defined in (2). Then for a generic m-dimensional subspace V m of H 2m+1 one has W 4m+4 V m V = {}.
6 6 TIKHOMIROV Доказательство. According to Definition 4.1 in which we put U = H 2m+1, U = V, each point x P (H 2m+1 V ) has rank 1 rk(x) dim V = 4. Thus (22) P (W 4m+4) = 4 r=1 Z r, where Z r := {x P (W 4m+4) rk(x) = r}, 1 r 4, are locally closed subsets of P (W 4m+4). Consider the Grassmannian and its locally closed subsets G := G(m, H 2m+1) (23) Σ r = {V m G V m U r (x) for some point x Z r }, 1 r 4. The condition that Z r P (V m V ) means that there exists a point x P (U r ) Z r for some r-dimensional subspace U r V m. This together with (22) implies that {V m G P (V m V ) P (W 4m+4) } = 4 r=1 Σ r. Thus, to prove the Theorem, it is enough to show that (24) dim Σ r < dim G, 1 r 4. We are starting now the proof of (24) for r = 4, 3, 2, 1. (i) r = 4. Set Γ 4 := {(x, U) P (W4m+4) G(4, H2m+1) rk(x) = 4 and U = U 4 (x)} and let P (W4m+4) p 4 q 4 Γ 4 G(4, H 2m+1 ) be the projections. By construction, p 4 (Γ 4 )) = Z 4 and the morphism p 4 : Γ 4 Z 4 is an isomorphism. Hence dim q 4 (Γ 4 ) dim Γ 4 = dim Z 4 dim P (W 4m+4) = 4m + 3. By construction we have the graph of incidence Π 4 = {(U, V m ) q 4 (Γ 4 ) Σ 4 U V m } with surjective projections q 4 (Γ 4 ) pr 1 pr 2 Π 4 Σ4 and a fibre over an arbitrary point U q 4 (Γ 4 ). Hence pr 1 1 (U) = G(m 4, H 2m+1/U) dim Σ 4 dim Π 4 = dim q 4 (Γ 4 )+dim G(m 4, H 2m+1/U) 4m+3+(m 4)(m+1) = m(m+1) 1 = = dim G 1 < dim G, i.e. (24) is true for r = 4. (ii) r = 3. Consider a morphism f 3 : Z 3 P (V ) = P 3 : x V 3 (x), where the pair of spaces (U 3 (x), V 3 (x)), U 3 (x) H2m+1 and V 3 (x) V, is determined uniquely by the point x via the condition x P (U 3 (x) V 3 (x)), since rk(x) = 3 (see Definition 4.1). Now for a given subspace V 3 V set (25) Σ 3 (V 3 ) = {V m G V m U 3 (x) for some point x f 1 3 (V 3 )}. Comparing this with (23) for r = 3 yields (26) Σ 3 = V 3 V Σ 3(V 3 ). Hence, (27) dim Σ 3 dim Σ 3 (V 3 ) + 3.
7 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE7 We are going to obtain an estimate for the dimension of Σ 3 (V 3 ) for an arbitrary 3-dimensional subspace V 3 in V. This subspace defines a commutative diagram (28) F Ω P 3 V 3 O P 3( 1) V O P 3( 1) I z O P 3 I x ( 1) O P 3( 1) k z, where z = P (ker : V V3 ) is a point in P 3 and the sheaf F has an O P 3-resolution O P 3( 2) 3O P 3( 1) F. Twisting this resolution by the vector bundle E and passing to cohomology we obtain the equalities H 1 (F E) H 2 (E( 3)) = H 2m+1, H 2 (F E) =. Respectively, passing to cohomology in diagram (28) twisted by E and using the above equalities and evident relations H (E k z ) k 2, H 1 (E k z ) = implies the diagram (29) k 2 H 2m+1 W 4m+4 H2m+1 V λ 3 H2m+1 V k 2 H 1 (E I z ) H 4m mult H 1 (E I z ( 1)) H 2m+1.
8 8 TIKHOMIROV In this diagram the composition ɛ := mult λ is surjective. Hence, setting W 2m+3 (V 3 ) := ker ɛ, where dim W 2m+3 (V 3 ) = 2m + 3, we obtain a commutative diagram (3) W 2m+3 (V 3 ) j W 4m+4 H 2m+1 H2m+1 V λ 3 H2m+1 V H 2m+1 ɛ H 4m H 4m mult Set. Z 3 (V 3 ) := {x P (W 2m+3 (V 3 )) rk(x) = 3}. The inclusion j in diagram (3) yields the bijection (31) Z 3 (V 3 ) f 1 3 (V 3 ). Consider the graph of incidence Γ 3 (V 3 ) := {(x, U) Z 3 (V 3 ) G(3, H2m+1) U = U 3 (x)} with projections Z 3 (V 3 ) p 3 Γ 3 (V 3 ) q 3 G(3, H2m+1). By construction, p 3 (Γ 3 (V 3 )) = Z 3 (V 3 ) and the morphism p 4 : Γ 3 (V 3 ) Z 3 (V 3 ) is an isomorphism. Hence (32) dim q 3 (Γ 3 (V 3 )) dim Γ 3 (V 3 ) = dim Z 3 (V 3 ) dim P (W 2m+3 (V 3 )) = 2m + 2. Consider the graph of incidence Π 3 (V 3 ) = {(U, V m ) q 3 (Γ 3 (V 3 )) Σ 3 (V 3 ) U V m } with projections q 3 (Γ 3 (V 3 )) pr 1 Π 3 (V 3 ) pr 2 Σ 3 (V 3 ) and a fibre pr 1 1 (U) = G(m 3, H 2m+1/U) over an arbitrary point U q 3 (Γ 3 (V 3 )). The projection Π 3 (V 3 ) pr 2 Σ 3 (V 3 ) is surjective in view of (31). Hence, using (32), we obtain dim Σ 3 (V 3 ) dim Π 3 (V 3 ) = dim q 3 (Γ 3 (V 3 ))+dim G(m 3, H 2m+1/U) 2m+2+(m 3)(m+1) = = m 2 1. This together with (27) and the assumption m 3 yields dim Σ 3 m = dim G + 2 m < dim G, i.e. (24) holds for r = 3. Before proceeding to the case r = 2 we need to make a small digression on jumping lines of E. Introduce some more notation. For a given line l P 3 we have E l O P 1(d) O P 1( d) for a welldefined nonnegative integer d called the jump of E l and is denoted d E (l); respectively, the line l is called a jumping line of jump d of E. Set G 2,4 := G(2, V ) and J k (E) := {l G 2,4 d E (l) k}, J k (E) := J k(e) J k+1 (E), k. From the semicontinuity of E l, l G 2,4, it follows that J k (E) (resp., J k (E)) is a closed (resp., locally closed) subset of G 2,4, k. Moreover, by Theorem of Grauert-Mülich, J (E) is a dense open subset of G 2,4. Next, since E I 2m+1, it follows that J 2m+1 (E) =, so that J 2m 1 (E) = J 2m 1(E) J 2m(E). We will use below the following lemma. Lemma 4.3. (1) dim J 2m 1 (E) 1. (ii) dim Jk (E) 3 for 1 k 2m 2.
9 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE9 Proof of Lemma. (1) Suppose the contrary, i.e. dim J 2m (E) 2. Take any irreducible surface S J 2m (E) and let D be the degree of S with respect to the sheaf O G2,4 (1). Fix an integer r 5 and take any irreducible curve C belonging to the linear series OG2,4 (r) S. Then the degree deg C w.r.t. O G2,4 (1) equals to Dr, hence deg C 5. Hence by [C, Lemma 6] there exist two distinct lines, say, l 1, l 2 C, which intersect in P 3. Let the plane P 2 be the span of l 1 and l 2 in P 3. Now the exact triple E( 2) P 2 E P 2 E l1 l 2 implies (33) H (E P 2) H (E l1 l 2 ) H 1 (E( 2) P 2). Next, as [E] I 2m+1, we have h (E( 1)) = h 1 (E( 2)) =, hence the exact triple E( 2) E( 1) E( 1) P 2 implies (34) H (E( 1) P 2) =. Now assume h (E P 2) >. Then a section s H s (E P 2) defines an injection O P 2 E P 2. This injection and (34) show that the zero-set Z of section s is -dimensional and the injection s s extends to a triple O P 2 E P 2 I Z,P 2. Whence (35) h (E P 2) 1. Furthermore, equality together with Riemann-Roch and Serre duality for the vector bundle E( 1) P 2 shows that h 1 (E( 2) P 2) = 2m + 1. Whence in view of (33) and (34) we obtain (36) h (E l1 l 2 ) 2m + 2. On the other hand, let x := l 1 l 2. Since by construction l 1, l 2 J 2m 1 (E), it follows that either E li O P 2(2m 1) O P 2(1 2m), or E li O P 2(2m) O P 2( 2m), hence h (E I x,li ) 2m 1, i = 1, 2. This clearly implies h (E l1 l 2 ) h (E I x,l1 l 2 ) h (E I x,l1 )+h (E I x,l2 ) = 4m 2. Comparing this with (36) we obtain the inequality 2m + 2 4m 2, i.e. m 2. This contradicts to the assumption m 3. Hence, the assertion (1) follows. (2) This is an immediate corollary of Theorem of Grauert-Mülich. Lemma is proved. (iii) r = 2. Our notation and argument is completely parallel to that in the case r = 3. Consider a morphism f 2 : Z 2 G 2,4 : x V 2 (x), where the pair of spaces (U 2 (x), V 2 (x)), U 2 (x) H2m+1 and V 2 (x) V, is determined uniquely by the point x via the condition x P (U 2 (x) V 2 (x)), since rk(x) = 2 (see Definition 4.1). According to the above remarks on jumping lines of E we may assume that l Jk (E) for some k 2m, i.e. respectively, h (E l) = 2, h 1 (E l) =, if l J (E), h (E l) = k + 1, h 1 (E l) = k 1, if l J k(e), 1 k 2m. Now for 1 k 2m and a given subspace V 2 J k set (37) Σ 2,k (V 2 ) = {V m G V m U 2 (x) for some point x f 1 2 (V 2 )}. Then similarly to (26) we have Hence, in view of Lemma 4.3 Σ 2 = 2m Σ 2,k (V 2 ). k= V 2 Jk (38) dim Σ 2 max V 2 J k (dim Σ 2,k (V 2 ) + dim Jk). k 2m
10 1 TIKHOMIROV We are going to obtain an estimate for the dimension of Σ 2,k (V 2 ) for an arbitrary 2-dimensional subspace V 2 in Jk, k 2m. This subspace defines a commutative diagram (39) O P 3( 2) s Ω P 3 F V 2 O P 3( 1) V O P 3( 1) V 2 O P 3( 1) I l O P 3 O l, where l = P (ker V V 2 ) is a line in P 3, V 2 := V /V 2, and F := coker s. Passing to cohomology in diagram (39) twisted by E, we obtain the diagram (4) W 4m+4 H (E l) H 1 (E F ) H 2m+1 V 2 H 2m+1 V H 2m+1 V 2 H (E l) H 1 (E I l ) H 4m mult H 1 (E l) Assume for definiteness that 1 k 2m. (The case k = is treated in a similar way.) In this case diagram (4) leads to a diagram. (41) W k+1 (V 2 ) j W 4m+4 H 4m k+3 H2m+1 V 2 H2m+1 V V 4m k+1 H 4m mult H 2m+1 V 2 W k 1. where we set W k+1 (V 2 ) := H (E l), W k 1 := H 1 (E l), V 4m k+1 := H 2m+1 V 2 /W k+1 (V 2 ).
11 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE11 Set Z 2,k (V 2 ) := {x P (W k+1 (V 2 )) rk(x) = 2}. The inclusion j in diagram (41) yields the bijection (42) Z 2,k (V 2 ) f 1 2 (V 2 ). Consider the graph of incidence Γ 2,k (V 2 ) := {(x, U) Z 2,k (V 2 ) G(2, H2m+1) U = U 2 (x)} with projections Z 2,k (V 2 ) p 2 Γ 2,k (V 2 ) q 2 G(2, H2m+1). By construction, p 2 (Γ 2,k (V 2 )) = Z 2,k (V 2 ) and the morphism p 4 : Γ 2,k (V 2 ) Z 2,k (V 2 ) is an isomorphism. Hence (43) dim q 2 (Γ 2,k (V 2 )) dim Γ 2,k (V 2 ) = dim Z 2,k (V 2 ) dim P (W k+1 (V 2 )) = k. Consider the graph of incidence Π 2,k (V 2 ) = {(U, V m ) q 2 (Γ 2,k (V 2 )) Σ 2,k (V 2 ) U V m } with projections q 2 (Γ 2,k (V 2 )) pr 1 Π 2,k (V 2 ) pr 2 Σ 2,k (V 2 ) and a fibre pr 1 1 (U) = G(m 2, H 2m+1/U) over an arbitrary point U q 2 (Γ 2,k (V 2 )). The projection Π 2,k (V 2 ) pr 2 Σ 2,k (V 2 ) is surjective in view of (42). Hence using (43) we obtain dim Σ 2,k (V 2 ) dim Π 2,k (V 2 ) = dim q 2 (Γ 2,k (V 2 ))+dim G(m 2, H 2m+1/U) k+(m 2)(m+1) = = m 2 m 2 + k = dim G (2m k + 2), 1 k 2m. In a similar way we obtain for k = dim Σ 2, (V 2 ) 1 + (m 2)(m + 1) = m 2 m 1 = dim G (2m + 1). The last two inequalities together with (38), Lemma 4.3 and the assumption m 3 yield dim Σ 2 < dim G, i.e. (24) is true for r = 2. (ii) r = 1. Consider a morphism f 1 : Z 1 P (V ) = (P 3 ) : x V 1 (x), where the pair of spaces (U 1 (x), V 1 (x)), U 1 (x) H2m+1 and V 1 (x) V, is determined uniquely by the point x via the condition x P (U 1 (x) V 1 (x)), since rk(x) = 1 (see Definition 4.1). Now for a given subspace V 1 (P 3 ) set Then similar to (26) we have Σ 1 (V 1 ) := {V m G V m U 1 (x) for some point x f 1 1 (V 1 )}. (44) Σ 1 = V 1 (P 3 ) Σ 1(V 1 ). Hence, (45) dim Σ 1 dim Σ 1 (V 1 ) + 3.
12 12 TIKHOMIROV We are going to obtain an estimate for the dimension of Σ 1 (V 1 ) for an arbitrary 1-dimensional subspace V 1 in V. This subspace defines a commutative diagram (46) Ω P 3 Ω P 3 V 1 O P 3( 1) V O P 3( 1) V 3 O P 3( 1) O P 3( 1) O P 3 O P 2. Note that to the point V 1 (P 3 ) there clearly corresponds a projective plane P (V 1 ) in P 3. Set B(E) := {V 1 (P 3 ) h (E P (V1 )) }. It is known that, for m 1, dim B(E) 2. (see [B1]). Moreover, in view of (35) h (E P (V1 )) = 1, V 1 B(E). Passing to cohomology in diagram (46) twisted by E and using the equality h (E) = for [E] I 2m+1 we obtain the diagram (47) W 4m+4 H (E P (V1 )) W 4m+4 H2m+1 V λ 1 H2m+1 V H2m+1 V 3 H (E P (V1 )) H 2m+1 H 4m mult H 1 (E P (V1 )).
13 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE13 Let V 1 B(E). Setting ɛ := mult λ and W 1 (V 1 ) := ker ɛ = H (E P (V1 )), where dim W 1 (V 1 ) = 1, we obtain from (47) a commutative diagram (48) W 1 (V 1 ) j W 4m+4 W 4m+4/W 1 (V 1 ) H2m+1 λ V 1 H2m+1 V ɛ H 2m+1/W 1 (V 1 ) H 4m mult H 2m+1 V 3 H 1 (E P 2 (V 1 )). Set Z 1 (V 1 ) := if V 1 B(E), resp., Z 1 (V 1 ) := j(w 1 (V 1 )) if V 1 B(E). The diagrams (47) and (48) yield the bijection (49) Z 1 (V 1 ) f 1 1 (V 1 ), V 1 (P 3 ). The rest argument is completely the same as in cases r = 3 and r = 2 above. Consider the graph of incidence Γ 1 (V 1 ) := {(x, U) Z 1 (V 1 ) P (H2m+1) U = U 1 (x)} with projections Z 1 (V 1 ) p 1 Γ 1 (V 1 ) q 1 P (H2m+1). By construction, p 1 (Γ 1 (V 1 )) = Z 1 (V 1 ) and the morphism p 4 : Γ 1 (V 1 ) Z 1 (V 1 ) is an isomorphism. Hence (5) dim q 1 (Γ 1 (V 1 )) dim Γ 1 (V 1 ) = dim Z 1 (V 1 ). Consider the graph of incidence Π 1 (V 1 ) = {(U, V m ) q 1 (Γ 1 (V 1 )) Σ 1 (V 1 ) U V m } with projections q 1 (Γ 1 (V 1 )) pr 1 Π 1 (V 1 ) pr 2 Σ 1 (V 1 ) and a fibre pr 1 1 (U) = G(m 1, H 2m+1/U) over an arbitrary point U q 1 (Γ 1 (V 1 )). The projection Π 1 (V 1 ) pr 2 Σ 1 (V 1 ) is surjective in view of (49). Hence in view of (5) we have dim Σ 1 (V 1 ) dim Π 1 (V 1 ) = dim q 1 (Γ 1 (V 1 )) + dim G(m 1, H 2m+1/U) + (m 1)(m + 1) = = m 2 1. This together with (45) and the assumption m 3 yields dim Σ m = dim G + 2 m < dim G, i.e. (24) holds for r = 1. Theorem is proved. 5. Decomposition k 2m+1 k m+1 k m and related constructions 5.1. Decomposition k 2m+1 k m+1 k m. Fix an isomorphism (51) ξ : k m+1 k m k 2m+1 and let (52) k m+1 i m+1 k m+1 k m im k m
14 14 TIKHOMIROV be the injections of direct summands. For a given (2m + 1)-instanton vector bundle E, [E] I 2m+1, fix an isomorphism f : k 2m+1 H 2 (E( 3)) = H 2m+1 and a symplectic structure j : E E. The data [E, f, j] define a net of quadrics A MI 2m+1 (see section 3), and the exact triple (2) is naturally identified with the dual to the triple ker A k 2m+1 V W A and fits in diagram (9) for n = 2m + 1 (53) ker A k 2m+1 V c A W A Consider the composition ker A (54) j ξ,a : k m+1 V i m+1 k m+1 V k m V A c A = q A (k 2m+1 ) V. W A ξ k 2m+1 V c A W A. Under these notations Theorem 4.2 can be reformulated in the following way: (*) Assume m 3 and let A be an arbitrary (2m + 1)-net from MI 2m+1. Then for a generic isomorphism ξ : k 2m+1 k m+1 k m one has (55) ker A ξ i m+1 (k m+1 V ) = {}. Equivalently, j ξ,a : k m+1 V W A is an isomorphism. Consider the direct sum decomposition corresponding to the isomorphism (51) (56) ξ : Sm+1 (k m ) (k m+1 ) 2 V S m S 2m+1 and let (57) ξ 1 : S 2m+1 S m+1, ξ 2 : S 2m+1 (k m ) (k m+1 ) 2 V, ξ 3 : S 2m+1 S m be projections onto summands. By definition, ξ 1 (A) considered as a skew-symmetric homomorphism k m+1 V (k m+1 ) V coincides with the composition (58) ξ 1 (A) : k m+1 V j ξ,a q W A A W A j ξ,a (k m+1 ) V. This means that assertion (*) can be reformulated as: (**) Assume m 3 and let A be an arbitrary (2m + 1)-net from MI 2m+1. Then for a generic isomorphism ξ in (51) the skew-symmetric homomorphism ξ 1 (A) : k m+1 V (k m+1 ) V is invertible. For A and ξ from (**) we have the commutative diagram (59) k m+1 V i m+1 k m+1 V k m V ξ j ξ,a k 2m+1 V c A W A ξ 1 (A) q A ξ(a) A (k m+1 ) V i m+1 (k m+1 ) V (k m ) V j ξ,a (k 2m+1 ) V c WA A, ξ
15 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE15 ). As j ξ,a in this diagram is invertible, the compo- ( ξ1 (A) ξ where ξ(a) is the matrix 2 (A) ξ 2 (A) ξ 3 (A) sition g ξ,a = j 1 ξ,a c A ξ i m is well-defined, and we obtain a commutative diagram (6) k m V (k m ) V g ξ ξ 2 (A) g ξ ξ 2 (A) k m+1 ξ 1 (A) V (k m+1 ) V. In particular, ξ 3 (A) (61) ξ 3 (A) = ξ 2 (A) ξ 1 (A) 1 ξ 2 (A). For m 1 let Isom 2m+1 be the set of all isomorphisms ξ in (51). Consider the open subset MI 2m+1 of MI 2m+1 defined in (17) and set (62) MI 2m+1 (ξ) := {A MI 2m+1 the skew symmetric homomorphism ξ 1 (A) in (58) is invertible}, ξ Isom 2m+1. The relation (61) together with (**) implies the following corollary of Theorem 4.2. Theorem 5.1. Fom m 3 the following statements hold. (i) The sets MI 2m+1 (ξ), ξ Isom 2m+1, are dense open subsets of the set MI 2m+1 constituting its open cover. (ii) For any ξ Isom 2m+1 and any A MI 2m+1 (ξ) the relation (61) is true. We will need below the following lemma. Lemma 5.2. Let ξ and A MI 2m+1 (ξ) be as in Theorem 5.1 and set (63) B := ξ 1 (A), C := ξ 2 (A). Then the following statements hold. (i) Consider a subbundle morphism (64) α ξ,a := j 1 ξ a A ξ : (k m+1 k m ) O P 3( 1) k m+1 V O P 3. Then there exists an epimorphism (65) λ ξ,a : coker(b α ξ,a ) (k m+1 ) O P 3(1). making commutative the diagram (66) (k m+1 ) V O P 3 coker(b α ξ,a ) can u λ ξ,a (k m+1 ) O P 3(1), where can is a canonical surjection.
16 16 TIKHOMIROV (ii) Consider the commutative diagram (67) k m O P 3( 1) (k m+1 k m ) O P 3( 1) B αξ,a (k m+1 ) V O P 3 can coker(b α ξ,a ) i m+1 k m+1 O P 3( 1) B u (k m+1 ) V O P 3 ɛ ξ,a v B 1 k m+1 T P 3( 1) τ ξ,a k m O P 3( 1), where τ ξ,a and ɛ ξ,a are the induced morphisms. Then the morphism τ ξ,a is a subbundle morphism fitting in a commutative diagram (68) (k m+1 ) V O P 3 v B 1 k m+1 T P 3( 1) C u k m O P 3( 1) Доказательство. (i) Consider the commutative diagram (69) k 2m+1 O( 1) ξ a A W A O j ξ,a q A τ ξ,a k m O P 3( 1). WA O a A (k 2m+1 ) O(1) (k m+1 k m ) O( 1) α ξ,a (k m+1 ) V O B (k m+1 ) V O α ξ,a (k m+1 k m ) O(1) i i m+1 u m+1 u k m+1 O( 1) (k m+1 ) O(1) j ξ,a ξ Here the upper triple is the monad (11) for n = 2m + 1. Whence the statement (i) follows. (ii) Standard diagram chasing using (63) and diagrams (59) and (67) Remarks on t Hooft instantons. Consider the set I th 2m+1 := {[E] I 2m+1 h (E(1)) }, of t Hooft instanton bundles and the corresponding set of t Hooft instanton nets MI th 2m+1 := πn 1 (I th 2m+1). We collect some well-known facts about I th 2m+1 in the following lemma - see [BT], [NT], [T2, Prop. 2.2]. Lemma 5.3. Let m 1. Then the following statements hold. (i) I2m+1 th is an irreducible (1m + 9)-dimensional subvariety of I 2m+1. Respectively, MI2m+1 th is an irreducible (4m m + 1)-dimensional subvariety of I 2m+1. (ii) I2m+1 th := I2m+1 th I 2m+1 is a smooth dense open subset of I2m+1 th and (7) h (E(1)) = 1, [E] I th 2m+1.
17 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE17 (iii) MI th 2m+1 is a smooth dense open subset of the set (71) T H 2m+1 := {A S 2m+1 A = 2m+2 i=1 h 2 w, where h (k 2m+1 ), w 2 V, w w = }. We are going to extend the statement of Theorem 5.1 to the cases m = 1 and 2. To this end, for m = 1, 2 and ξ Isom 2m+1 consider the sets MI 2m+1 (ξ) defined in (62) and set (72) MI 2m+1 := ξ Isom 2m+1 MI 2m+1 (ξ), m = 1, 2. For m 1 let ξ Isom 2m+1 be the standard isomorphism k m+1 k m k m+1 : ((a 1,..., a m+1 ), (a m+2,..., a 2m+1 )) (a 1,..., a 2m+1 ). Let {h 1 = (1,,..., ),..., h 2m+1 (,...,, 1) be the standard basis in (k 2m+1 ) and let e 1,..., e 4 be some fixed basis in V. Consider the nets A (m) T H 2m+1, m = 1, 2, defined as follows (73) A (1) = h 2 1 (e 1 e 2 + e 3 e 4 ) + h 2 2 (e 1 e 3 + e 4 e 2 ), A (2) = h 2 1 (e 1 e 2 + e 3 e 4 ) + h 2 2 (e 1 e 3 + e 4 e 2 ) + h 2 3 (e 1 e 4 + e 2 e 3 ). It is an exercise to show that the homomorphisms ξ 1(A (m) ) : k m+1 V (k m+1 ) V, m = 1, 2, are invertible. On the other hand, for a given ξ Isom 2m+1 the condition that a homomorphism ξ 1 (A) : k m+1 V (k m+1 ) V is invertible is an open condition on the net A T H 2m+1. Hence, since the sets MI 2m+1, m = 1, 2, are irreducible, Lemma 5.3 yields the following corollary. Corollary 5.4. Let 1 m 2. (i) For m = 1, 2 the set MI 2m+1 is a dense open subset of MI 2m+1 and of MI 2m+1, and the statement of Theorem 5.1 extends to the cases m = 1 and 2, with MI 2m+1 being substituted by MI 2m+1. (ii) Let m 1. The set { MI MI2m+1 th th := 2m+1, m 3, MI 2m+1 MI2m+1, th m = 1, 2, is a dense open subset of MI th 2m+1 and of MI th 2m+1 covered by dense open subsets (74) MI th 2m+1(ξ) := MI th 2m+1 MI 2m+1 (ξ), ξ Isom 2m+1. Note that (18), Theorem 5.1 and Corollary 5.4 yield Corollary 5.5. Let m 1. Then for any ξ Isom 2m+1 the scheme (MI 2m+1 (ξ)) red is dense open in (MI 2m+1 ) red. In particular, (75) dim MI 2m+1 (ξ) = dim MI 2m Invertible nets of quadrics from S m+1 and symplectic rank-(2m + 2) bundles. Introduce more notations. Set (76) N m+1 := {B S m+1 B : k m+1 V (k m+1 ) V is an invertible homomorphism}. The set N m+1 is a dense open subset of the vector space S m+1, and it is easy to see that for any B N m+1 the following conditions are satisfied. (1) The morphism B : k m+1 O P 3( 1) (k m+1 ) Ω P 3(1) induced by the homomorphism B : k m+1 V (k m+1 ) V is a subbundle morphism, i.e. (77) E 2m+2 (B) := coker( B)
18 18 TIKHOMIROV is a vector bundle of rank 2m + 2 на P 3. This follows from the diagram (78) k m+1 O P 3( 1) B (k m+1 ) Ω P 3(1) e E 2m+2 (B) u k m+1 V O P 3 B v (k m+1 ) V O P 3 E 2m+2 (B) k m+1 T P 3( 1) v B u (k m+1 ) O P 3(1) (2) The homomorphism B : k m+1 (k m+1 ) 2 V induced by B : k m+1 V (k m+1 ) V is injective. This follows from the commutative diagram extending the upper horizontal triple in (78) (79) (k m+1 ) T P 3( 2) (k m+1 ) T P 3( 2) k m+1 O P 3 B (k m+1 ) 2 V O P 3 can H (E 2m+2 (B)(1)) O P 3 k m+1 O P 3 w B (k m+1 ) Ω P 3(2) e ev E 2m+2 (B)(1), where w is the morphism induced by the morphism v from the Euler exact sequence in (78). From this diagram we obtain the isomorphism (8) coker( B) H (E 2m+2 (B)(1)). (3) Diagram (78) and the Five-Lemma yield an isomorphism (81) θ : E 2m+2 (B) E 2m+2 (B) which is in fact symplectic, θ = θ,
19 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE19 since the homomorphism B : k m V (k m ) V is skew-symmetric. The isomorphism θ together with the upper triple from (78) and its dual fits in the commutative diagram (82) k m+1 O P 3( 1) B (k m+1 ) Ω P 3(1) e E 2m+2 (B) k m+1 O P 3( 1) B u (k m+1 ) V O P 3 v v B 1 e θ k m+1 T P 3( 1) u (k m+1 ) O P 3(1) B (k m+1 ) O P 3(1) Note that this diagram immediately implies that. (83) h (E 2m+2 (B)) = h i (E 2m+2 (B)( 2)) =, i. Let ξ and A MI 2m+1 (ξ) be as in Theorem 5.1 for m 3, respectively, in Corollary 5.4 for m = 1, 2. Then the homomorphism B : k m+1 V (k m+1 ) V defined in (63) by definition lies in N m+1. Hence by Lemma 5.2 diagrams (66) and (66) hold. These diagrams together with (82) imply B τ ξ,a =, so that there exists a morphism (84) ρ ξ,a : k m O( 1) E 2m+2 (B) such that τ ξ,a = e θ ρ ξ,a. Since τ ξ,a is a subbundle morphism, ρ ξ,a is also a subbundle morphism. Moreover, diagrams (68) and (82) yield the commutative diagram (85) (k m+1 ) Ω P 3(1) E 2m+2 (B) C ρ ξ,a v k m O( 1) e θ C τ ξ,a (k m+1 ) V v B O 1 k m+1 T P 3( 1). Diagrams (82) and (85) yield the commutative diagram e (86) k m O( 1) (k C ρ ξ,a m+1 ) V O v E 2m+2 (B) Ω P 3(1) D C θ C e e θ e E 2m+2 (B) e T P 3( 1) ρ ξ,a v C (k m ) O(1) k m+1 V O, C B 1
20 2 TIKHOMIROV where D C := C B 1 C = u (C B 1 C) u is the zero map. In fact, by (61) and (63) we have D C = p 2 (ξ 3 (A)), where p 2 : 2 ((k n ) V ) 2 (k n ) S 2 V is the projection onto the second direct summand of the decomposition (8). Since by (57) ξ 3 (A) lies in the first direct summand of (8) it follows that D C =. We thus obtain the monad (87) k m O( 1) ρ ξ,a E 2m+2 (B) θ ρ ξ,a (k m ) O(1) with the cohomology sheaf (88) E 2 (ξ, A) := ker(θ ρ ξ,a)/ Im ρ ξ,a which is a vector bundle since ρ ξ,a is a subbundle morphism. Furthermore, by (83) it follows from the monad (87) that E 2 (ξ, A) is a (2m + 1)-instanton, (89) [E 2 (ξ, A)] I 2m+1. Lemma 5.6. E 2 (ξ, A) E(A), where the sheaf E(A) is defined in (12). Доказательство. Diagram chasing using (59), (6), (67)-(69), (78)-(79) and (82). 6. Scheme X m. An isomorphism between X m and an open subset of the space MI 2m Space X m. Consider the vector space S m+1, respectively, its dual space S m+1 and set (9) (S m+1) := {B S m+1 D : (k m+1 ) V k m+1 V is an invertible homomorphism}, (91) Σ m+1 := Hom(k m, (k m+1 ) 2 V ) According to our convention on notations we will understand an arbitrary point C Σ m+1 either as a homomorphism C : k m V (k m+1 ) V, or as a homomorphism C : k m (k m+1 ) 2 V, or as an induced morphism C : k m O( 1) (k m+1 ) Ω(1). Note also that the set (S m+1) is a dense open subset of the vector space S m+1. Consider the set (92) (i) (C D C : k m V (k m ) V ) S m, (ii) the map (k m+1 k m ) O (D 1,C) u (k m+1 ) V O(1) is a subbundle morphism, X m := (D, C) (S m+1) Σ m+1 (iii) the composition Ĉ : km C (k m+1 ) 2 V can (k m+1 ) 2 V / Im( D 1 ) H (E 2m+2 (D 1 )(1)) yields a subbundle morphism k m O P 3( 1) ρ D,C E 2m+2 (D 1 ), i.e. ρ D,C is surjective and E 2(D, C) := Ker( t ρ D,C )/ Im(ρ D,C ) is locally free.
21 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE21 By definition X m is a locally closed subset of (S m+1) Σ m+1. Hence it is naturally supplied with the structure of a reduced scheme. Note that in the condition (iii) of the definition of X m we set t ρ D,C := θ ρ D,C, where θ : E 2m+2 (D 1 ) E 2m+2(D 1 ) is a natural symplectic structure on E 2m+2 (D 1 ) defined in (81). Theorem 6.1. Let m 1 and let ξ be as in Theorem 5.1 and Corollary 5.4. (i) There is an isomorphism of reduced schemes (93) f m : (MI 2m+1 (ξ)) red X m : A (ξ 1 (A) 1, ξ 2 (A)). (ii) The inverse isomorphism is given by the formula (94) g m : X m (MI 2m+1 (ξ)) red : (D, C) ξ(d 1, C, C D C). 1 Доказательство. (i) We first show that the image of the map f m : (MI 2m+1 (ξ)) red (S m+1) Σ in m,m+1 lies in X m, i.e. satisfies the conditions (i)-(iii) in the definition of X m. Indeed, the condition (i) is automatically satisfied, since (57) and (61) give C D C = ξ 2 (A) ξ 1 (A) 1 ξ 2 (A) = ξ 3 (A) S 2 (k m ) 2 V. Next, the morphism ρ D,C defined in (iii) above coincides by its definition with the morphism ρ ξ,a defined in (84). In fact, the upper triangle of the diagram (85) twisted by O(1) and the lower part of the diagram (79) in which we put (95) B = D 1 (note that D is invertible) fit in the diagram (96) k m+1 O D 1 (k m+1 ) 2 V O C k m+1 O D 1 (k m+1 ) Ω(2) w C can k m O H (E 2m+2 (D 1 )(1)) O Ĉ ρ ξ,a e E 2m+2 (D 1 )(1), where the composition Ĉ = can C is defined in the condition (iii) of the definition of X m. Whence (97) ρ D,C = ρ ξ,a. Since ρ ξ,a is a subbundle morphism, the condition (iii) is satisfied and, moreover, Ĉ is a subbundle morphism as well. Thus, the lower part of the diagram (96) shows that the morphism ( D 1, C) : (k m+1 k m ) O (k m+1 ) Ω(2) is a subbundle morphism. Hence its composition with the subbundle morphism v : (k m+1 ) Ω(2) (k m+1 ) V O(1) is a subbundle morphism as well. By definition, this composition coincides with (D 1, C) u. Hence the condition (ii) in the definition of X m is satisfied. This shows that f m ((MI 2m+1 (ξ)) red ) lies in X m. Last, the equality g m f m = id follows directly from (57) and (61). (ii) We first prove that the image of the map (98) g m : X m S 2m+1 : (D, C) (D 1, C, C D C) 2 1 Here we use the decomposition (56) fixed by the choice of ξ. 2 We identify here the triple (D 1, C, C D C) with a point in S 2 (k 2m+1 ) 2 V via the decomposition (56). ev
22 22 TIKHOMIROV lies in (MI 2m+1 (ξ)) red. In fact, the subbundle morphism A := (D 1, C) u : (k m+1 k m ) O (k m+1 ) V O(1) and its dual extend to the right and left exact sequence (99) (k m+1 k m ) O( 1) A (k m+1 ) V O A D (k m+1 k m ) O(1). ( ) D Furthermore, by definition A D A = u 1 C A u, where A is the matrix C C. D C Since the condition (i) is satisfied, under the direct sum decomposition (56) this matrix A can be treated an element of S 2m+1. Hence u A u =, i.e. (99) is a monad. Show that its cohomology bundle E(D, C) := ker(a D)/ Im A is an (2m + 1)-instanton, this giving the desired inclusion g(x m ) (MI 2m+1 (ξ)) red. For this, consider the diagram (67) in which we substitute B α ξ,a by A, respectively, B by D 1, denote G := coker A, and change the notation for τ ξ,a and ɛ ξ,a, respectively, to τ D,C and ɛ D,C (1) k m O P 3( 1) (k m+1 k m ) O P 3( 1) i m+1 k m+1 O P 3( 1) A D 1 u (k m+1 ) V O P 3 (k m+1 ) V O P 3 can G ɛ D,C v D k m+1 T P 3( 1) τ D,C k m O P 3( 1). In these notations the diagram (82) becomes the display of the antiselfdual monad (11) k m+1 O( 1) D 1 u (k m+1 ) V O u (k m+1 ) O(1) with the symplectic cohomology sheaf E 2m+2 (D 1 ): (12) E 2m+2 (D 1 ) = ker(u )/ Im(D 1 u). Moreover, as in (84) and (85) we obtain a subbundle morphism (13) ρ D,C : k m O( 1) E 2m+2 (D 1 ) such that (14) τ D,C = e θ ρ D,C, where θ : E 2m+2 (D 1 ) E 2m+2 (D 1 ) is a symplectic structure on E 2m+2 (D 1 ). Besides, as in (83) we have (15) h (E 2m+2 (D 1 )) = h i (E 2m+2 (D 1 )( 2)) =, i. Furthermore, as before, the antiselfdual monads (99) and (11) imply the (antiselfdual) monad (87) (16) k m O( 1) ρ D,C E 2m+2 (D 1 ) θ ρ D,C (k m ) O(1) with the cohomology sheaf E(D, C), (17) E(D, C) = ker(θ ρ D,C)/ Im(ρ D,C ). Now (15) and (16) yield h (E(D, C)) = h i (E(D, C)( 2)) =, i, i.e. E(D, C) is an (2m + 1)-instanton. Thus Im g m I 2m+1 (ξ). The fact that f m g m = id follows directly from (93) and (94).
23 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE Scheme Z m. Set 7. Variety Z m (18) Λ m := 2 (k m ) S 2 V, Φ m := Hom(k m, (k m ) ) 2 V, and consider the set (19) Z m := (D, φ) S m Φ m Θ m (D, φ) := φ D φ : k m V (k m ) V satisfies the condition. Θ m (D, φ) S m (Here, as in (9), we understand a point D S m as a homomorphism (k m ) V k m V.) Consider the standard decomposition with the induced projections 2 ((k m ) V ) = S m Λ m pr 1 S m 2 ((k m ) V ) pr 2 Λ m. We have a morphism h m : S m Φ m Λ m : (A m, φ m ) pr 2 (Θ(A m, φ m )). By the definition Z m we have (11) Z m = h 1 m (). Convention: If Z m is nonempty, we supply Z m with a scheme structure of a scheme-theoretic fibre h 1 m () of the morphism h m. Assume that (111) Z m. Then from the definition of Z m we obtain the estimate for the dimension of Z m at each point z Z m (112) dim z Z m = dim h 1 m () dim(s m Φ m ) dim 2 (k m ) S 2 V = = 3m(m + 1) + 6m 2 5m(m 1) = 4m(m + 2). Consider the open dense subset Φ m := {φ Φ m φ : k m (k m ) 2 V ) is injective} of Φ m and set (113) Z m := { (D, φ) Z m (S m) Φ m The set Z m is by definition an open subset in Z m. Assume Z m. Pick a point z = (D, φ) Z m and set Im( φ) Im( (D 1 ) : k m (k m ) 2 V ) = {} } W 5m := (k m ) 2 V / Im( (D 1 )), dim W 5m = 5m. Let i(z) be the composition in the diagram (114) k m φ i(z) k m (D 1 ) (k m ) 2 V can W 5m The lower horizontal triple in (114) yields the diagram (115) k m O P 3 k m O P 3 (D 1 ) (k m ) 2 V O P 3 ev D 1 (k m ) Ω P 3(2) can W 5m O P 3 ev can E 2m (D 1 )(1),
24 24 TIKHOMIROV where E 2m (D 1 ) is a symplectic bundle (see (81)). From this diagram we deduce the equalities (116) h i (E 2m (D 1 )( 2)) =, i, and the isomorphism (117) h (ev) : W 5m H (E 2m (D 1 )), i, Moreover, the diagrams (114) and (115) define the composition (118) i z : k m O P 3( 1) i(z) W 5m O P 3( 1) ev E 2m (D 1 ). Note that from the definition of the set Z m it follows that (119) t i z i z =, where t i z := i z θ and θ : E 2m ((D 1 )) E 2m ((D 1 )) is the symplectic structure on E 2m ((D 1 )) mentioned above, i.e. we have an antiselfdual complex (12) k m O P 3( 1) iz E 2m (D 1 ) t i z (k m ) O P 3(1). (Warning: this complex is not right exact.) Twisting the sequence (118) by O P 3(1) and passing to sections, we obtain in view of Furthermore, the standard embedding (121) j : k m 1 k m : (a 1,..., a m 1 ) (a 1,..., a m 1, ) and the morphism i z from (118) define the composition (122) j z : k m 1 O P 3( 1) j k m O P 3( 1) iz E 2m (D 1 ) 7.2. Varieties Z m and N th 2m 1. Assume, as above, that Z m and set (123) Z m = {z = (D, φ) Z m j z : k m 1 O P 3( 1) E 2m (D 1 ) is a subbundle morphism}. By definition, Z m is an open subset of Z m, hence also of Z m. If Z m, then for any point z = (D, φ) Z m we obtain from (119) that t j z j z =, where t j z := j z θ. Thus j z defines a monad (124) k m 1 O P 3( 1) j z E 2m (D 1 ) t j z (k m 1 ) O P 3(1), and in view of (116) the cohomology sheaf of this monad is an instanton bundle (125) E 2 (z) := Ker( t j z )/ Im(j z ), [E 2 (z)] I(2m 1). Consider the subvariety I th 2m 1 I 2m 1 of t Hooft instanton bundles I th 2m 1 := {[E] I 2m 1 h (E(1)) }. Lemma 7.1. Assume Z m. Then for any z = (D, φ) Z m the bundle E 2 (z) is a t Hooft instanton bundle, i.e. [E 2 (z)] I th 2m 1. Proof. Consider the complexes (12) and (124) and set H m 1 := k m 1 O P 3( 1), H m := k m O P 3( 1), K m+1 := coker j z, K m := coker i z.
25 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE25 The complexes (12) and (124) are antiselfdual, hence they extend to a commutative diagram (126) H m 1 E 2m (D 1 ) j i H m z E 2m (D 1 ) t j z O P 3( 1) O P 3(1) t i z j z Hm 1 j Hm O P 3(1), K m γ H m E 2 (z) τ K m+1 β δ Hm 1 j s α O P 3( 1) in which α, β, γ, δ and τ are the induced morphisms. In this diagram we have β α = and j γ β = δ. Hence δ α =. This implies that α factors through the morphism τ, i.e. there exists an injection s : O P 3( 1) E 2 (z) such that α = τ s. This injection s is a nonzero section s H (E 2 (z)(1)). Hence E 2 (z) is a t Hooft bundle. We will show that Zm is an irreducible variety of dimension 4m(m+2), hence it is nonempty. For this, fix an isomorphism (127) ξ : k m k m 1 k 2m 1 and consider the variety MI th 2m 1(ξ) defined in (74). Take an arbitrary point A MI th 2m 1(ξ). The point A defines a point B = ξ 1 (A) and a monad k m 1 O P 3( 1) ρ ξ,a E 2m (B) t ρ ξ,b (k m 1 ) O P 3(1) with the cohomology bundle [E 2 (A)] = π 2m 1 (A) (see subsection 5.3). The display of this monad twisted by O P 3(1) is (128) E 2 (A)(1) k m 1 ρ O ξ,a P 3 E 2m (B)(1) K m+1 (A)(1) t ρ ξ,a (k m 1 ) O P 3(2), where K m+1 (A) := coker ρ ξ,a. Note that from (7) and the definition of MI2m 1(ξ) th it follows that h (E 2 (A)(1)) = 1. Hence, passing to sections in the diagram (128) we obtain a well defined epimorphism (129) b(ξ, A) : H (E 2m (B)(1)) h (ɛ) H (K m+1 (A)(1)) can H (K m+1 (A)(1))/H (E 2 (A)(1)) k 4m. ɛ
26 26 TIKHOMIROV On the other hand, similar to (115) and (117) we obtain the exact triple (13) k m B 1 (k m ) 2 V c(a) H (E 2m (B)(1)). Denote by c(a) the epimorphism (k m ) 2 V H (E 2m (B)(1)) in this triple and set (131) V 2m (ξ, A) := c(a) 1 (ker b(ξ, A)) k 2m, V 2m(ξ, A) := {v V 2m (ξ, A) Span(Im (ξ 1 (A) 1 ), Im (ξ 2 (A)), kv) = V 2m (ξ, A)}, (132) V 2m (ξ) := {(A, v) A MI th 2m 1(ξ), v V 2m (ξ, A)}. Here the projection V 2m (ξ) MI2m 1(ξ) th : (A, v) A is a k 2m -bundle over MI2m 1(ξ), th hence by Lemma 5.3 and Corollary 5.4 V 2m (ξ) is irreducible of dimension (133) dim V 2m (ξ) = dim MI th 2m 1(ξ) + 2m = 4m(m + 2). Besides, V 2m(ξ, A) is a dense open subset of V 2m (ξ, A) for each A MI th 2m 1(ξ), (134) V 2m(ξ, A) dense open V 2m (ξ, A) k 2m. Next, set Π m := Hom(k m, (k m ) 2 V) and (135) N(ξ, A) := (φ : km V (i) Span(Im (ξ 1 (A) 1 ), Im φ) = V 2m (ξ, A), (k m ) V ) Π m (ii) φ j = ξ 2 (A), (iii) φ (ξ 1 (A) 1 ) φ S m, (136) N th 2m 1(ξ) := {(A, φ) A MI th 2m 1(ξ), φ N(ξ, A)}. Consider the standard decomposition k m = k m 1 k, so that the injection j in (121) is an embedding of the left direct summand of this decomposition. Then each monomorphism ( φ : k m (k m ) 2 V ) N(ξ, A) in view of the conditions (i)-(iii) of (135) is uniquely determined by its restriction onto the right direct summand k of the standard decomposition, satisfying the conditions and φ k : k V 2m (ξ, A) (k m ) 2 V : 1 v Span(Im (ξ 1 (A) 1 ), Im φ) = Span(Im (ξ 1 (A) 1 ), Im (ξ 2 (A)), kv) = V 2m (ξ, A). (ξ 2 (A) + φ k V ) (ξ 1 (A) 1 ) (ξ 2 (A) + φ k V ) S m. These conditions and the definition of V2m(ξ, A) mean that N(ξ, ) is a closed subset of V2m(ξ, A), hence by (134) it is a locally closed subset of V 2m (ξ, A). As a result, we have (137) N th 2m 1(ξ) In particular, locally closed V 2m (ξ). (138) dim N th 2m 1(ξ) dim V 2m (ξ) = 4m(m + 2). Now consider the map (139) h m : N th 2m 1(ξ) Z m : (A, φ) (D := ξ 1 (A) 1, φ). This map is well defined. In fact, take any point (A, φ) N2m 1(ξ). th Since A MI2m 1(ξ), th we have D (S m), so that the vector bundle E 2m (D 1 ) is well-defined. Next, since φ j = ξ 2 (A) (see condition (ii) in (135)), it follows from Theorem 6.1 that the morphism j z : k m 1 O( 1) E 2m (D 1 )
27 MODULI OF MATHEMATICAL INSTANTON VECTOR BUNDLES WITH ODD c 2 ON PROJECTIVE SPACE27 for z = (D, φ) coincides with the subbundle morphism ρ ξ,a satisfying diagram (96). Note that in view of (97) we can rewrite this also as (14) j z = ρ D,C, C = φ j. The diagram (96), in turn, implies that the condition Im( D) Im( φ) = {} is satisfied. This together with the injectivity of j z and the condition (iii) in (135) precisely means that z Zm. As a result, it follows that Zm and, respectively, Z m is nonempty. Moreover, since Zm is supplied with the structure of a reduced scheme and N2m 1(ξ) th is smooth (hence reduced) it follows that the map h m given by formula (139) is a morphism of reduced schemes. Next, consider the set where Z m(ξ) := {z Z m z = (D, φ) satisfies the condition ( )} (D 1, φ j) u : (k m k m 1 ) O( 1) (k m ) V O is a subbundle morphism. ( ) Since the condition (*) is open and Z m(ξ) contains a subset h m (N th 2m 1(ξ)), it follows that Z m(ξ) is a nonempty open subset of Z m. Consider the map (141) λ m : Z m(ξ) S 2m 1 : z = (D, φ) A := ξ(d 1, φ j, (φ j) D (φ j)). Since (φ D φ) S m by the definition of Z m, it follows that (142) (φ j) D (φ j) S m 1, i.e. the map λ m in (141) is well-defined. Moreover, since Z m(ξ) is a reduced scheme, the map λ m is a morphism of reduced schemes. Theorem 7.2. Let m 1 and ξ be a fixed isomorphism (127). Then Z m(ξ) is a smooth irreducible variety of dimension 4m(m + 2) and there is an isomorphism of smooth varieties (143) ν m : Z m(ξ) N th 2m 1(ξ) : (D, φ) (A, φ), where A is given by (141). Proof. Consider the set X m 1 defined in (92) and the morphism of reduced schemes (144) η m : Z m(ξ) X m 1 : z = (D, φ) (D, φ j). This morphism is well-defined since (142), (*) and (14) are precisely the conditions (i), (ii) and (iii) of the definition of X m 1. Next, comparing (94), (141) and (144) we obtain that λ m = g m 1 η m for m 1. Whence Im λ m MI 2m 1 (ξ). Moreover, for any point z = (D, φ) the diagram (126) defines a section s E 2 (A)(1) for A = λ m (z), so that [E 2 (A)] I2m 1, th i.e. A MI2m 1(ξ). th Hence (A, φ) N2m 1(ξ), th and the morphism ν m in (143) is well-defined. Comparing now (139) and (143), we obtain that h m = νm 1, i.e. ν m is an isomorphism of reduced schemes. Next, since by definition Zm(ξ) is an open subset of Z m, it follows from (112) that dim Zm(ξ) 4m(m + 2). This together with (138) and the isomophism ν m shows that dim Z m(ξ) = dim N th 2m 1(ξ) = dim V 2m (ξ) = 4m(m + 2). Whence by (137) and the irreducibility and smoothness of V 2m (ξ) we obtain that Zm(ξ) N2m 1(ξ) th is a dense open subset of V 2m (ξ), so that Zm(ξ) is smooth and irreducible of dimension 4m(m + 2).
h M (T ). The natural isomorphism η : M h M determines an element U = η 1
MODULI PROBLEMS AND GEOMETRIC INVARIANT THEORY 7 2.3. Fine moduli spaces. The ideal situation is when there is a scheme that represents our given moduli functor. Definition 2.15. Let M : Sch Set be a moduli
More informationSPACES OF RATIONAL CURVES ON COMPLETE INTERSECTIONS
SPACES OF RATIONAL CURVES ON COMPLETE INTERSECTIONS ROYA BEHESHTI AND N. MOHAN KUMAR Abstract. We prove that the space of smooth rational curves of degree e on a general complete intersection of multidegree
More informationAFFINE PUSHFORWARD AND SMOOTH PULLBACK FOR PERVERSE SHEAVES
AFFINE PUSHFORWARD AND SMOOTH PULLBACK FOR PERVERSE SHEAVES YEHAO ZHOU Conventions In this lecture note, a variety means a separated algebraic variety over complex numbers, and sheaves are C-linear. 1.
More informationExercises of the Algebraic Geometry course held by Prof. Ugo Bruzzo. Alex Massarenti
Exercises of the Algebraic Geometry course held by Prof. Ugo Bruzzo Alex Massarenti SISSA, VIA BONOMEA 265, 34136 TRIESTE, ITALY E-mail address: alex.massarenti@sissa.it These notes collect a series of
More informationMATH 8253 ALGEBRAIC GEOMETRY WEEK 12
MATH 8253 ALGEBRAIC GEOMETRY WEEK 2 CİHAN BAHRAN 3.2.. Let Y be a Noetherian scheme. Show that any Y -scheme X of finite type is Noetherian. Moreover, if Y is of finite dimension, then so is X. Write f
More informationWe can choose generators of this k-algebra: s i H 0 (X, L r i. H 0 (X, L mr )
MODULI PROBLEMS AND GEOMETRIC INVARIANT THEORY 43 5.3. Linearisations. An abstract projective scheme X does not come with a pre-specified embedding in a projective space. However, an ample line bundle
More information1. Algebraic vector bundles. Affine Varieties
0. Brief overview Cycles and bundles are intrinsic invariants of algebraic varieties Close connections going back to Grothendieck Work with quasi-projective varieties over a field k Affine Varieties 1.
More informationARITHMETICALLY COHEN-MACAULAY BUNDLES ON HYPERSURFACES
ARITHMETICALLY COHEN-MACAULAY BUNDLES ON HYPERSURFACES N. MOHAN KUMAR, A. P. RAO, AND G. V. RAVINDRA Abstract. We prove that any rank two arithmetically Cohen- Macaulay vector bundle on a general hypersurface
More informationHARTSHORNE EXERCISES
HARTSHORNE EXERCISES J. WARNER Hartshorne, Exercise I.5.6. Blowing Up Curve Singularities (a) Let Y be the cusp x 3 = y 2 + x 4 + y 4 or the node xy = x 6 + y 6. Show that the curve Ỹ obtained by blowing
More informationSPACES OF RATIONAL CURVES IN COMPLETE INTERSECTIONS
SPACES OF RATIONAL CURVES IN COMPLETE INTERSECTIONS ROYA BEHESHTI AND N. MOHAN KUMAR Abstract. We prove that the space of smooth rational curves of degree e in a general complete intersection of multidegree
More informationTheta divisors and the Frobenius morphism
Theta divisors and the Frobenius morphism David A. Madore Abstract We introduce theta divisors for vector bundles and relate them to the ordinariness of curves in characteristic p > 0. We prove, following
More informationAlgebraic Geometry Spring 2009
MIT OpenCourseWare http://ocw.mit.edu 18.726 Algebraic Geometry Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.726: Algebraic Geometry
More informationwhere m is the maximal ideal of O X,p. Note that m/m 2 is a vector space. Suppose that we are given a morphism
8. Smoothness and the Zariski tangent space We want to give an algebraic notion of the tangent space. In differential geometry, tangent vectors are equivalence classes of maps of intervals in R into the
More informationSegre classes of tautological bundles on Hilbert schemes of surfaces
Segre classes of tautological bundles on Hilbert schemes of surfaces Claire Voisin Abstract We first give an alternative proof, based on a simple geometric argument, of a result of Marian, Oprea and Pandharipande
More informationPolarized K3 surfaces of genus 18 and 20
Polarized K3 surfaces of genus 18 and 20 Dedicated to Professor Hisasi Morikawa on his 60th Birthday Shigeru MUKAI A surface, i.e., 2-dimensional compact complex manifold, S is of type K3 if its canonical
More informationAlgebraic Geometry Spring 2009
MIT OpenCourseWare http://ocw.mit.edu 18.726 Algebraic Geometry Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.726: Algebraic Geometry
More informationAPPENDIX 3: AN OVERVIEW OF CHOW GROUPS
APPENDIX 3: AN OVERVIEW OF CHOW GROUPS We review in this appendix some basic definitions and results that we need about Chow groups. For details and proofs we refer to [Ful98]. In particular, we discuss
More informationModuli spaces of reflexive sheaves of rank 2
Moduli spaces of reflexive sheaves of rank 2 Jan O. Kleppe Abstract Let F be a coherent rank 2 sheaf on a scheme Y P n of dimension at least two and let X Y be the zero set of a section H 0 (F). In this
More information10. Smooth Varieties. 82 Andreas Gathmann
82 Andreas Gathmann 10. Smooth Varieties Let a be a point on a variety X. In the last chapter we have introduced the tangent cone C a X as a way to study X locally around a (see Construction 9.20). It
More information8 Perverse Sheaves. 8.1 Theory of perverse sheaves
8 Perverse Sheaves In this chapter we will give a self-contained account of the theory of perverse sheaves and intersection cohomology groups assuming the basic notions concerning constructible sheaves
More informationALGEBRAIC GROUPS. Disclaimer: There are millions of errors in these notes!
ALGEBRAIC GROUPS Disclaimer: There are millions of errors in these notes! 1. Some algebraic geometry The subject of algebraic groups depends on the interaction between algebraic geometry and group theory.
More informationVector bundles in Algebraic Geometry Enrique Arrondo. 1. The notion of vector bundle
Vector bundles in Algebraic Geometry Enrique Arrondo Notes(* prepared for the First Summer School on Complex Geometry (Villarrica, Chile 7-9 December 2010 1 The notion of vector bundle In affine geometry,
More informationChapter 2 Linear Transformations
Chapter 2 Linear Transformations Linear Transformations Loosely speaking, a linear transformation is a function from one vector space to another that preserves the vector space operations. Let us be more
More informationLECTURE 7: STABLE RATIONALITY AND DECOMPOSITION OF THE DIAGONAL
LECTURE 7: STABLE RATIONALITY AND DECOMPOSITION OF THE DIAGONAL In this lecture we discuss a criterion for non-stable-rationality based on the decomposition of the diagonal in the Chow group. This criterion
More informationMATH 233B, FLATNESS AND SMOOTHNESS.
MATH 233B, FLATNESS AND SMOOTHNESS. The discussion of smooth morphisms is one place were Hartshorne doesn t do a very good job. Here s a summary of this week s material. I ll also insert some (optional)
More informationStable bundles with small c 2 over 2-dimensional complex tori
Stable bundles with small c 2 over 2-dimensional complex tori Matei Toma Universität Osnabrück, Fachbereich Mathematik/Informatik, 49069 Osnabrück, Germany and Institute of Mathematics of the Romanian
More informationEVEN AND ODD INSTANTON BUNDLES ON FANO THREEFOLDS OF PICARD NUMBER ONE
EVEN AND ODD INSTANTON BUNDLES ON FANO THREEFOLDS OF PICARD NUMBER ONE DANIELE FAENZI Abstract. We consider an analogue of the notion of instanton bundle on the projective 3-space, consisting of a class
More informationStable maps and Quot schemes
Stable maps and Quot schemes Mihnea Popa and Mike Roth Contents 1. Introduction........................................ 1 2. Basic Setup........................................ 4 3. Dimension Estimates
More informationVector Bundles on Algebraic Varieties
Vector Bundles on Algebraic Varieties Aaron Pribadi December 14, 2010 Informally, a vector bundle associates a vector space with each point of another space. Vector bundles may be constructed over general
More informationSTEENROD OPERATIONS IN ALGEBRAIC GEOMETRY
STEENROD OPERATIONS IN ALGEBRAIC GEOMETRY ALEXANDER MERKURJEV 1. Introduction Let p be a prime integer. For a pair of topological spaces A X we write H i (X, A; Z/pZ) for the i-th singular cohomology group
More informationSome Remarks on Prill s Problem
AFFINE ALGEBRAIC GEOMETRY pp. 287 292 Some Remarks on Prill s Problem Abstract. N. Mohan Kumar If f : X Y is a non-constant map of smooth curves over C and if there is a degree two map π : X C where C
More informationON A THEOREM OF CAMPANA AND PĂUN
ON A THEOREM OF CAMPANA AND PĂUN CHRISTIAN SCHNELL Abstract. Let X be a smooth projective variety over the complex numbers, and X a reduced divisor with normal crossings. We present a slightly simplified
More informationYuriy Drozd. Intriduction to Algebraic Geometry. Kaiserslautern 1998/99
Yuriy Drozd Intriduction to Algebraic Geometry Kaiserslautern 1998/99 CHAPTER 1 Affine Varieties 1.1. Ideals and varieties. Hilbert s Basis Theorem Let K be an algebraically closed field. We denote by
More informationSERRE FINITENESS AND SERRE VANISHING FOR NON-COMMUTATIVE P 1 -BUNDLES ADAM NYMAN
SERRE FINITENESS AND SERRE VANISHING FOR NON-COMMUTATIVE P 1 -BUNDLES ADAM NYMAN Abstract. Suppose X is a smooth projective scheme of finite type over a field K, E is a locally free O X -bimodule of rank
More informationFormal power series rings, inverse limits, and I-adic completions of rings
Formal power series rings, inverse limits, and I-adic completions of rings Formal semigroup rings and formal power series rings We next want to explore the notion of a (formal) power series ring in finitely
More informationInjective Modules and Matlis Duality
Appendix A Injective Modules and Matlis Duality Notes on 24 Hours of Local Cohomology William D. Taylor We take R to be a commutative ring, and will discuss the theory of injective R-modules. The following
More informationMotivic integration on Artin n-stacks
Motivic integration on Artin n-stacks Chetan Balwe Nov 13,2009 1 / 48 Prestacks (This treatment of stacks is due to B. Toën and G. Vezzosi.) Let S be a fixed base scheme. Let (Aff /S) be the category of
More informationMath 248B. Applications of base change for coherent cohomology
Math 248B. Applications of base change for coherent cohomology 1. Motivation Recall the following fundamental general theorem, the so-called cohomology and base change theorem: Theorem 1.1 (Grothendieck).
More informationCOHEN-MACAULAY RINGS SELECTED EXERCISES. 1. Problem 1.1.9
COHEN-MACAULAY RINGS SELECTED EXERCISES KELLER VANDEBOGERT 1. Problem 1.1.9 Proceed by induction, and suppose x R is a U and N-regular element for the base case. Suppose now that xm = 0 for some m M. We
More informationChern classes à la Grothendieck
Chern classes à la Grothendieck Theo Raedschelders October 16, 2014 Abstract In this note we introduce Chern classes based on Grothendieck s 1958 paper [4]. His approach is completely formal and he deduces
More informationALGEBRAIC GEOMETRY I, FALL 2016.
ALGEBRAIC GEOMETRY I, FALL 2016. DIVISORS. 1. Weil and Cartier divisors Let X be an algebraic variety. Define a Weil divisor on X as a formal (finite) linear combination of irreducible subvarieties of
More informationCoherent sheaves on elliptic curves.
Coherent sheaves on elliptic curves. Aleksei Pakharev April 5, 2017 Abstract We describe the abelian category of coherent sheaves on an elliptic curve, and construct an action of a central extension of
More information4. Images of Varieties Given a morphism f : X Y of quasi-projective varieties, a basic question might be to ask what is the image of a closed subset
4. Images of Varieties Given a morphism f : X Y of quasi-projective varieties, a basic question might be to ask what is the image of a closed subset Z X. Replacing X by Z we might as well assume that Z
More informationAn Atlas For Bun r (X)
An Atlas For Bun r (X) As told by Dennis Gaitsgory to Nir Avni October 28, 2009 1 Bun r (X) Is Not Of Finite Type The goal of this lecture is to find a smooth atlas locally of finite type for the stack
More informationthe complete linear series of D. Notice that D = PH 0 (X; O X (D)). Given any subvectorspace V H 0 (X; O X (D)) there is a rational map given by V : X
2. Preliminaries 2.1. Divisors and line bundles. Let X be an irreducible complex variety of dimension n. The group of k-cycles on X is Z k (X) = fz linear combinations of subvarieties of dimension kg:
More informationARITHMETICALLY COHEN-MACAULAY BUNDLES ON THREE DIMENSIONAL HYPERSURFACES
ARITHMETICALLY COHEN-MACAULAY BUNDLES ON THREE DIMENSIONAL HYPERSURFACES N. MOHAN KUMAR, A. P. RAO, AND G. V. RAVINDRA Abstract. We prove that any rank two arithmetically Cohen- Macaulay vector bundle
More informationLines on Projective Hypersurfaces
Lines on Projective Hypersurfaces Roya Beheshti Abstract We study the Hilbert scheme of lines on hypersurfaces in the projective space. The main result is that for a smooth Fano hypersurface of degree
More informationProjective Schemes with Degenerate General Hyperplane Section II
Beiträge zur Algebra und Geometrie Contributions to Algebra and Geometry Volume 44 (2003), No. 1, 111-126. Projective Schemes with Degenerate General Hyperplane Section II E. Ballico N. Chiarli S. Greco
More informationarxiv:alg-geom/ v1 21 Mar 1996
AN INTERSECTION NUMBER FOR THE PUNCTUAL HILBERT SCHEME OF A SURFACE arxiv:alg-geom/960305v 2 Mar 996 GEIR ELLINGSRUD AND STEIN ARILD STRØMME. Introduction Let S be a smooth projective surface over an algebraically
More informationMATH 326: RINGS AND MODULES STEFAN GILLE
MATH 326: RINGS AND MODULES STEFAN GILLE 1 2 STEFAN GILLE 1. Rings We recall first the definition of a group. 1.1. Definition. Let G be a non empty set. The set G is called a group if there is a map called
More informationSpecial cubic fourfolds
Special cubic fourfolds 1 Hodge diamonds Let X be a cubic fourfold, h H 2 (X, Z) be the (Poincaré dual to the) hyperplane class. We have h 4 = deg(x) = 3. By the Lefschetz hyperplane theorem, one knows
More informationON THE NUMBER AND BOUNDEDNESS OF LOG MINIMAL MODELS OF GENERAL TYPE
ON THE NUMBER AND BOUNDEDNESS OF LOG MINIMAL MODELS OF GENERAL TYPE DILETTA MARTINELLI, STEFAN SCHREIEDER, AND LUCA TASIN Abstract. We show that the number of marked minimal models of an n-dimensional
More informationREPRESENTATION THEORY, LECTURE 0. BASICS
REPRESENTATION THEORY, LECTURE 0. BASICS IVAN LOSEV Introduction The aim of this lecture is to recall some standard basic things about the representation theory of finite dimensional algebras and finite
More informationON THE INTERSECTION THEORY OF QUOT SCHEMES AND MODULI OF BUNDLES WITH SECTIONS
ON THE INTERSECTION THEORY OF QUOT SCHEMES AND MODULI OF BUNDLES WITH SECTIONS ALINA MARIAN Abstract. We consider a class of tautological top intersection products on the moduli space of stable pairs consisting
More informationSome remarks on symmetric correspondences
Some remarks on symmetric correspondences H. Lange Mathematisches Institut Universitat Erlangen-Nurnberg Bismarckstr. 1 1 2 D-91054 Erlangen (Germany) E. Sernesi Dipartimento di Matematica Università Roma
More informationHonors Algebra 4, MATH 371 Winter 2010 Assignment 4 Due Wednesday, February 17 at 08:35
Honors Algebra 4, MATH 371 Winter 2010 Assignment 4 Due Wednesday, February 17 at 08:35 1. Let R be a commutative ring with 1 0. (a) Prove that the nilradical of R is equal to the intersection of the prime
More informationIntroduction to Chiral Algebras
Introduction to Chiral Algebras Nick Rozenblyum Our goal will be to prove the fact that the algebra End(V ac) is commutative. The proof itself will be very easy - a version of the Eckmann Hilton argument
More informationCOMPLEX VARIETIES AND THE ANALYTIC TOPOLOGY
COMPLEX VARIETIES AND THE ANALYTIC TOPOLOGY BRIAN OSSERMAN Classical algebraic geometers studied algebraic varieties over the complex numbers. In this setting, they didn t have to worry about the Zariski
More informationarxiv:math/ v1 [math.ag] 9 Sep 2002
arxiv:math/0209094v1 [math.ag] 9 Sep 2002 ELLIPTIC CURVES AND RANK-2 VECTOR BUNDLES ON THE PRIME FANO THREEFOLD OF GENUS 7 A. ILIEV AND D. MARKUSHEVICH Abstract. According to Mukai, any prime Fano threefold
More informationCourse 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra
Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra D. R. Wilkins Contents 3 Topics in Commutative Algebra 2 3.1 Rings and Fields......................... 2 3.2 Ideals...............................
More informationNote that the first map is in fact the zero map, as can be checked locally. It follows that we get an isomorphism
11. The Serre construction Suppose we are given a globally generated rank two vector bundle E on P n. Then the general global section σ of E vanishes in codimension two on a smooth subvariety Y. If E is
More informationAPPENDIX 1: REVIEW OF SINGULAR COHOMOLOGY
APPENDIX 1: REVIEW OF SINGULAR COHOMOLOGY In this appendix we begin with a brief review of some basic facts about singular homology and cohomology. For details and proofs, we refer to [Mun84]. We then
More informationPacific Journal of Mathematics
Pacific Journal of Mathematics STABLE REFLEXIVE SHEAVES ON SMOOTH PROJECTIVE 3-FOLDS PETER VERMEIRE Volume 219 No. 2 April 2005 PACIFIC JOURNAL OF MATHEMATICS Vol. 219, No. 2, 2005 STABLE REFLEXIVE SHEAVES
More informationmult V f, where the sum ranges over prime divisor V X. We say that two divisors D 1 and D 2 are linearly equivalent, denoted by sending
2. The canonical divisor In this section we will introduce one of the most important invariants in the birational classification of varieties. Definition 2.1. Let X be a normal quasi-projective variety
More informationarxiv:math/ v1 [math.ag] 17 Oct 2006
Remark on a conjecture of Mukai Arnaud BEAUVILLE Introduction arxiv:math/0610516v1 [math.ag] 17 Oct 2006 The conjecture mentioned in the title appears actually as a question in [M] (Problem 4.11): Conjecture.
More informationNodal symplectic spheres in CP 2 with positive self intersection
Nodal symplectic spheres in CP 2 with positive self intersection Jean-François BARRAUD barraud@picard.ups-tlse.fr Abstract 1 : Let ω be the canonical Kähler structure on CP 2 We prove that any ω-symplectic
More informationNotes on p-divisible Groups
Notes on p-divisible Groups March 24, 2006 This is a note for the talk in STAGE in MIT. The content is basically following the paper [T]. 1 Preliminaries and Notations Notation 1.1. Let R be a complete
More information3. Lecture 3. Y Z[1/p]Hom (Sch/k) (Y, X).
3. Lecture 3 3.1. Freely generate qfh-sheaves. We recall that if F is a homotopy invariant presheaf with transfers in the sense of the last lecture, then we have a well defined pairing F(X) H 0 (X/S) F(S)
More information0.1 Spec of a monoid
These notes were prepared to accompany the first lecture in a seminar on logarithmic geometry. As we shall see in later lectures, logarithmic geometry offers a natural approach to study semistable schemes.
More informationHungry, Hungry Homology
September 27, 2017 Motiving Problem: Algebra Problem (Preliminary Version) Given two groups A, C, does there exist a group E so that A E and E /A = C? If such an group exists, we call E an extension of
More informationHOMOMORPHISMS OF VECTOR BUNDLES ON CURVES AND PARABOLIC VECTOR BUNDLES ON A SYMMETRIC PRODUCT
PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 140, Number 9, September 2012, Pages 3017 3024 S 0002-9939(2012)11227-4 Article electronically published on January 24, 2012 HOMOMORPHISMS OF VECTOR
More informationLINKED ALTERNATING FORMS AND LINKED SYMPLECTIC GRASSMANNIANS
LINKED ALTERNATING FORMS AND LINKED SYMPLECTIC GRASSMANNIANS BRIAN OSSERMAN AND MONTSERRAT TEIXIDOR I BIGAS Abstract. Motivated by applications to higher-rank Brill-Noether theory and the Bertram-Feinberg-Mukai
More informationSynopsis of material from EGA Chapter II, 4. Proposition (4.1.6). The canonical homomorphism ( ) is surjective [(3.2.4)].
Synopsis of material from EGA Chapter II, 4 4.1. Definition of projective bundles. 4. Projective bundles. Ample sheaves Definition (4.1.1). Let S(E) be the symmetric algebra of a quasi-coherent O Y -module.
More information1 Notations and Statement of the Main Results
An introduction to algebraic fundamental groups 1 Notations and Statement of the Main Results Throughout the talk, all schemes are locally Noetherian. All maps are of locally finite type. There two main
More informationUnder these assumptions show that if D X is a proper irreducible reduced curve with C K X = 0 then C is a smooth rational curve.
Example sheet 3. Positivity in Algebraic Geometry, L18 Instructions: This is the third and last example sheet. More exercises may be added during this week. The final example class will be held on Thursday
More information9. Birational Maps and Blowing Up
72 Andreas Gathmann 9. Birational Maps and Blowing Up In the course of this class we have already seen many examples of varieties that are almost the same in the sense that they contain isomorphic dense
More informationIntroduction to Arithmetic Geometry Fall 2013 Lecture #17 11/05/2013
18.782 Introduction to Arithmetic Geometry Fall 2013 Lecture #17 11/05/2013 Throughout this lecture k denotes an algebraically closed field. 17.1 Tangent spaces and hypersurfaces For any polynomial f k[x
More informationElliptic curves, Néron models, and duality
Elliptic curves, Néron models, and duality Jean Gillibert Durham, Pure Maths Colloquium 26th February 2007 1 Elliptic curves and Weierstrass equations Let K be a field Definition: An elliptic curve over
More informationSynopsis of material from EGA Chapter II, 5
Synopsis of material from EGA Chapter II, 5 5. Quasi-affine, quasi-projective, proper and projective morphisms 5.1. Quasi-affine morphisms. Definition (5.1.1). A scheme is quasi-affine if it is isomorphic
More informationPERVERSE SHEAVES. Contents
PERVERSE SHEAVES SIDDHARTH VENKATESH Abstract. These are notes for a talk given in the MIT Graduate Seminar on D-modules and Perverse Sheaves in Fall 2015. In this talk, I define perverse sheaves on a
More informationarxiv:math/ v2 [math.ag] 3 Mar 1999
arxiv:math/9809140v2 [math.ag] 3 Mar 1999 THE ABEL JACOBI MAP OF A MODULI COMPONENT OF VECTOR BUNDLES ON THE CUBIC THREEFOLD D. MARKUSHEVICH AND A. S. TIKHOMIROV Abstract. The Abel Jacobi map of the family
More informationAlgebraic Geometry Spring 2009
MIT OpenCourseWare http://ocw.mit.edu 18.726 Algebraic Geometry Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.726: Algebraic Geometry
More information5 Quiver Representations
5 Quiver Representations 5. Problems Problem 5.. Field embeddings. Recall that k(y,..., y m ) denotes the field of rational functions of y,..., y m over a field k. Let f : k[x,..., x n ] k(y,..., y m )
More informationBetti numbers of abelian covers
Betti numbers of abelian covers Alex Suciu Northeastern University Geometry and Topology Seminar University of Wisconsin May 6, 2011 Alex Suciu (Northeastern University) Betti numbers of abelian covers
More informationclass # MATH 7711, AUTUMN 2017 M-W-F 3:00 p.m., BE 128 A DAY-BY-DAY LIST OF TOPICS
class # 34477 MATH 7711, AUTUMN 2017 M-W-F 3:00 p.m., BE 128 A DAY-BY-DAY LIST OF TOPICS [DG] stands for Differential Geometry at https://people.math.osu.edu/derdzinski.1/courses/851-852-notes.pdf [DFT]
More informationON THE NUMBER AND BOUNDEDNESS OF LOG MINIMAL MODELS OF GENERAL TYPE
ON THE NUMBER AND BOUNDEDNESS OF LOG MINIMAL MODELS OF GENERAL TYPE DILETTA MARTINELLI, STEFAN SCHREIEDER, AND LUCA TASIN Abstract. We show that the number of marked minimal models of an n-dimensional
More informationTopics in Algebraic Geometry
Topics in Algebraic Geometry Nikitas Nikandros, 3928675, Utrecht University n.nikandros@students.uu.nl March 2, 2016 1 Introduction and motivation In this talk i will give an incomplete and at sometimes
More informationINJECTIVE MODULES: PREPARATORY MATERIAL FOR THE SNOWBIRD SUMMER SCHOOL ON COMMUTATIVE ALGEBRA
INJECTIVE MODULES: PREPARATORY MATERIAL FOR THE SNOWBIRD SUMMER SCHOOL ON COMMUTATIVE ALGEBRA These notes are intended to give the reader an idea what injective modules are, where they show up, and, to
More information3.1. Derivations. Let A be a commutative k-algebra. Let M be a left A-module. A derivation of A in M is a linear map D : A M such that
ALGEBRAIC GROUPS 33 3. Lie algebras Now we introduce the Lie algebra of an algebraic group. First, we need to do some more algebraic geometry to understand the tangent space to an algebraic variety at
More informationTHE HODGE DECOMPOSITION
THE HODGE DECOMPOSITION KELLER VANDEBOGERT 1. The Musical Isomorphisms and induced metrics Given a smooth Riemannian manifold (X, g), T X will denote the tangent bundle; T X the cotangent bundle. The additional
More informationarxiv: v1 [math.ag] 14 Mar 2019
ASYMPTOTIC CONSTRUCTIONS AND INVARIANTS OF GRADED LINEAR SERIES ariv:1903.05967v1 [math.ag] 14 Mar 2019 CHIH-WEI CHANG AND SHIN-YAO JOW Abstract. Let be a complete variety of dimension n over an algebraically
More informationMath 797W Homework 4
Math 797W Homework 4 Paul Hacking December 5, 2016 We work over an algebraically closed field k. (1) Let F be a sheaf of abelian groups on a topological space X, and p X a point. Recall the definition
More informationHere is another way to understand what a scheme is 1.GivenaschemeX, and a commutative ring R, the set of R-valued points
Chapter 7 Schemes III 7.1 Functor of points Here is another way to understand what a scheme is 1.GivenaschemeX, and a commutative ring R, the set of R-valued points X(R) =Hom Schemes (Spec R, X) This is
More informationA NEW PROOF OF SERRE S HOMOLOGICAL CHARACTERIZATION OF REGULAR LOCAL RINGS
A NEW PROOF OF SERRE S HOMOLOGICAL CHARACTERIZATION OF REGULAR LOCAL RINGS RAVI JAGADEESAN AND AARON LANDESMAN Abstract. We give a new proof of Serre s result that a Noetherian local ring is regular if
More informationCHAPTER 0 PRELIMINARY MATERIAL. Paul Vojta. University of California, Berkeley. 18 February 1998
CHAPTER 0 PRELIMINARY MATERIAL Paul Vojta University of California, Berkeley 18 February 1998 This chapter gives some preliminary material on number theory and algebraic geometry. Section 1 gives basic
More informationThe Moduli Space of Rank 2 Vector Bundles on Projective Curves
The Moduli Space of Rank Vector Bundles on Projective urves T. Nakamura, A. Simonetti 1 Introduction Let be a smooth projective curve over and let L be a line bundle on. onsider the moduli functor VB (,
More informationPROBLEMS, MATH 214A. Affine and quasi-affine varieties
PROBLEMS, MATH 214A k is an algebraically closed field Basic notions Affine and quasi-affine varieties 1. Let X A 2 be defined by x 2 + y 2 = 1 and x = 1. Find the ideal I(X). 2. Prove that the subset
More informationAlgebraic Geometry Spring 2009
MIT OpenCourseWare http://ocw.mit.edu 18.726 Algebraic Geometry Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.726: Algebraic Geometry
More informationNORMALIZATION OF THE KRICHEVER DATA. Motohico Mulase
NORMALIZATION OF THE KRICHEVER DATA Motohico Mulase Institute of Theoretical Dynamics University of California Davis, CA 95616, U. S. A. and Max-Planck-Institut für Mathematik Gottfried-Claren-Strasse
More informationAn Outline of Homology Theory
An Outline of Homology Theory Stephen A. Mitchell June 1997, revised October 2001 Note: These notes contain few examples and even fewer proofs. They are intended only as an outline, to be supplemented
More information