Θ αβ = i. Θ αβjk dz j dz k.

Σχετικά έγγραφα
6.1. Dirac Equation. Hamiltonian. Dirac Eq.

Ordinal Arithmetic: Addition, Multiplication, Exponentiation and Limit

Every set of first-order formulas is equivalent to an independent set

2 Composition. Invertible Mappings

Lecture 2: Dirac notation and a review of linear algebra Read Sakurai chapter 1, Baym chatper 3

Example Sheet 3 Solutions

Exercise 1.1. Verify that if we apply GS to the coordinate basis Gauss form ds 2 = E(u, v)du 2 + 2F (u, v)dudv + G(u, v)dv 2

Chapter 6: Systems of Linear Differential. be continuous functions on the interval

EE512: Error Control Coding

Parametrized Surfaces

forms This gives Remark 1. How to remember the above formulas: Substituting these into the equation we obtain with

C.S. 430 Assignment 6, Sample Solutions

Nowhere-zero flows Let be a digraph, Abelian group. A Γ-circulation in is a mapping : such that, where, and : tail in X, head in

ST5224: Advanced Statistical Theory II

Math 446 Homework 3 Solutions. (1). (i): Reverse triangle inequality for metrics: Let (X, d) be a metric space and let x, y, z X.

Homework 3 Solutions

Uniform Convergence of Fourier Series Michael Taylor

Phys460.nb Solution for the t-dependent Schrodinger s equation How did we find the solution? (not required)

3.4 SUM AND DIFFERENCE FORMULAS. NOTE: cos(α+β) cos α + cos β cos(α-β) cos α -cos β

Section 8.3 Trigonometric Equations

Finite Field Problems: Solutions

Space-Time Symmetries

Problem Set 3: Solutions

Reminders: linear functions

MATH423 String Theory Solutions 4. = 0 τ = f(s). (1) dτ ds = dxµ dτ f (s) (2) dτ 2 [f (s)] 2 + dxµ. dτ f (s) (3)

Math221: HW# 1 solutions

Lecture 13 - Root Space Decomposition II

Statistical Inference I Locally most powerful tests

A Note on Intuitionistic Fuzzy. Equivalence Relation

Areas and Lengths in Polar Coordinates

4.6 Autoregressive Moving Average Model ARMA(1,1)

Fourier Series. MATH 211, Calculus II. J. Robert Buchanan. Spring Department of Mathematics

derivation of the Laplacian from rectangular to spherical coordinates

Lecture 15 - Root System Axiomatics

Other Test Constructions: Likelihood Ratio & Bayes Tests

Congruence Classes of Invertible Matrices of Order 3 over F 2

Second Order Partial Differential Equations

ANSWERSHEET (TOPIC = DIFFERENTIAL CALCULUS) COLLECTION #2. h 0 h h 0 h h 0 ( ) g k = g 0 + g 1 + g g 2009 =?

Srednicki Chapter 55

Areas and Lengths in Polar Coordinates

= λ 1 1 e. = λ 1 =12. has the properties e 1. e 3,V(Y

Partial Differential Equations in Biology The boundary element method. March 26, 2013

Concrete Mathematics Exercises from 30 September 2016

The Simply Typed Lambda Calculus

Απόκριση σε Μοναδιαία Ωστική Δύναμη (Unit Impulse) Απόκριση σε Δυνάμεις Αυθαίρετα Μεταβαλλόμενες με το Χρόνο. Απόστολος Σ.

Practice Exam 2. Conceptual Questions. 1. State a Basic identity and then verify it. (a) Identity: Solution: One identity is csc(θ) = 1

Chapter 6: Systems of Linear Differential. be continuous functions on the interval

w o = R 1 p. (1) R = p =. = 1

Solutions to Exercise Sheet 5

Fractional Colorings and Zykov Products of graphs

2. Let H 1 and H 2 be Hilbert spaces and let T : H 1 H 2 be a bounded linear operator. Prove that [T (H 1 )] = N (T ). (6p)

Section 9.2 Polar Equations and Graphs

HOMEWORK 4 = G. In order to plot the stress versus the stretch we define a normalized stretch:

Approximation of distance between locations on earth given by latitude and longitude

Section 7.6 Double and Half Angle Formulas

On a four-dimensional hyperbolic manifold with finite volume

Homework 8 Model Solution Section

Arithmetical applications of lagrangian interpolation. Tanguy Rivoal. Institut Fourier CNRS and Université de Grenoble 1

CRASH COURSE IN PRECALCULUS

Second Order RLC Filters

SCITECH Volume 13, Issue 2 RESEARCH ORGANISATION Published online: March 29, 2018

6.3 Forecasting ARMA processes

Symmetric Stress-Energy Tensor

If we restrict the domain of y = sin x to [ π, π ], the restrict function. y = sin x, π 2 x π 2

Matrices and Determinants

The challenges of non-stable predicates

12. Radon-Nikodym Theorem

If we restrict the domain of y = sin x to [ π 2, π 2

( ) 2 and compare to M.

b. Use the parametrization from (a) to compute the area of S a as S a ds. Be sure to substitute for ds!

ORDINAL ARITHMETIC JULIAN J. SCHLÖDER

CHAPTER 25 SOLVING EQUATIONS BY ITERATIVE METHODS

Coefficient Inequalities for a New Subclass of K-uniformly Convex Functions

Tridiagonal matrices. Gérard MEURANT. October, 2008

Generating Set of the Complete Semigroups of Binary Relations

PARTIAL NOTES for 6.1 Trigonometric Identities

Econ 2110: Fall 2008 Suggested Solutions to Problem Set 8 questions or comments to Dan Fetter 1

5. Choice under Uncertainty

SCHOOL OF MATHEMATICAL SCIENCES G11LMA Linear Mathematics Examination Solutions

Numerical Analysis FMN011

Pg The perimeter is P = 3x The area of a triangle is. where b is the base, h is the height. In our case b = x, then the area is

1 String with massive end-points

Jesse Maassen and Mark Lundstrom Purdue University November 25, 2013

ECE Spring Prof. David R. Jackson ECE Dept. Notes 2

SOLUTIONS TO MATH38181 EXTREME VALUES AND FINANCIAL RISK EXAM

On geodesic mappings of Riemannian spaces with cyclic Ricci tensor

Lecture 2. Soundness and completeness of propositional logic

SOME PROPERTIES OF FUZZY REAL NUMBERS

Inverse trigonometric functions & General Solution of Trigonometric Equations

The Pohozaev identity for the fractional Laplacian

2. THEORY OF EQUATIONS. PREVIOUS EAMCET Bits.

Higher Derivative Gravity Theories

F19MC2 Solutions 9 Complex Analysis

9.09. # 1. Area inside the oval limaçon r = cos θ. To graph, start with θ = 0 so r = 6. Compute dr

Problem Set 9 Solutions. θ + 1. θ 2 + cotθ ( ) sinθ e iφ is an eigenfunction of the ˆ L 2 operator. / θ 2. φ 2. sin 2 θ φ 2. ( ) = e iφ. = e iφ cosθ.

Exercises 10. Find a fundamental matrix of the given system of equations. Also find the fundamental matrix Φ(t) satisfying Φ(0) = I. 1.

= {{D α, D α }, D α }. = [D α, 4iσ µ α α D α µ ] = 4iσ µ α α [Dα, D α ] µ.

Lecture 34 Bootstrap confidence intervals

Lecture 10 - Representation Theory III: Theory of Weights

DESIGN OF MACHINERY SOLUTION MANUAL h in h 4 0.

Transcript:

Kobayashi-Lübke inequalities for Chern classes of Hermite-Einstein vector bundles and Guggenheimer-Yau-Bogomolov-Miyaoka inequalities for Chern classes of Kähler-Einstein manifolds Let (X, ω) be a compact Kähler manifold, n = dimx, and let E be a holomorphic vector bundle over X, r = rank E. We suppose that E is equipped with a hermitian metric, h and denote by D E,h the Chern connection on (E, h). The Chern curvature form is Θ h (E) = D E,h. In a (local) orthonormal frame (e ) 1 r of E, we write Θ h (E) = (Θ β ) 1,β r where the Θ β are complex valued (1, 1)-forms satisfying the hermitian condition Θ β = Θ β. We denote Θ β = i Θ βjk dz j dz k. 1,β r, 1 j,k n The hermitian symmetry condition can then be read Θ βjk = Θ βkj. If at some point x 0 X the coordinates (z j ) are chosen so that (dz j (x 0 )) is an orthonormal basis of TX,x 0, we define ( ) Tr ω Θ h (E) = Θ βjj C (X, hom(e, E)). j Definition. The hermitian vector bundle (E, h) is said to be Hermite- Einstein with respect to the Kähler metric ω if there is a constant λ > 0 such that Tr ω Θ h (E) = λid E. Recall that the Chern forms c k (E) h are defined by the formula det ( I + tθ h (E) ) = det(δ β + tθ β ) = 1 + t c 1 (E) h +... + t r c r (E) h. This gives in particular the identities c 1 (E) h = Θ, c (E) h = <β Θ Θ ββ Θ β Θ β = 1 Θ Θ ββ Θ β Θ β.,β The trace Tr ω Θ h (E) can be computed by the formula Θ h (E) ωn 1 (n 1)! = Tr ωθ h (E) ωn n!. 1

By taking the trace with respect to the indices in E and taking the Hermite- Einstein equation into account, we find c 1 (E) h ωn 1 (n 1)! = λr ωn n!. This implies that the number λ in the definition of Hermite-Einstein metrics is a purely numerical invariant, namely λ = n c 1 (E) ω n 1/ ω n. r X X Kobayashi-Lübke inequality. If E admits a Hermite-Einstein metric h with respect to ω, then [(r 1)c1 (E) h r c (E) h ω n 0 at every point of X. Moreover, the equality holds if and only if Θ h (E) = 1 r c 1(E) h Id E. Observe that the equality holds pointwise already if we have the numerical equality [ (r 1)c1 (E) r c (E) ω n = 0. X If we introduce the (formal) vector bundle Ẽ = E (det E) 1/r (Ẽ normalized vector bundle such that det Ẽ = O), then c1(ẽ) h = 0 and is the Θ h (Ẽ) = ( Θ h (E) 1 r c 1(E) h Id E ) Id (det E) 1/r. By the formula for the chern classes of E L, the Kobayashi-Lübke inequality can be rewritten as c (Ẽ) h ω n 0, with equality if and only if Ẽ is unitary flat. In that case, we say that E is projectively flat. Proof. By the above, (r 1)c 1 (E) h r c (E) h =,β Θ Θ ββ + r Θ β Θ β. Taking the wedge product with ω n /(n )! means taking the trace, i.e. the sum of coefficients of the terms i dz j dz j i dz k dz k for all j < k. For this, we have to look at products of the type (i dz j dz j )(i dz k dz k ) or (i dz j dz k )(i dz k dz j ). This yields [ (r 1)c 1 (E) h r c (E) h ω n = (n )! (Θ jj Θ ββkk Θ jk Θ ββkj ) + r(θ βjj Θ βkk Θ βjk Θ βkj ).

The initial factor comes from the fact that the final sum is taken over all unordered indices j, k (terms with j = k cancel). The Hermite-Einstein condition yields j Θ βjj = λ δ β, so we get λ δ δ ββ + rλ δ β δ β Θ jj Θ ββkk + r Θ βjj Θ βkk =,β = r λ + r λ = 0. Hence, using the hermitian symmetry of Θ βjk, we find [ (r 1)c 1 (E) h r c (E) h ω n = = r = r (n )! Θ jk Θ ββjk r Θ βjk β,j,k β,j,k Θ βjk + ( Θ jk Θ ββjk r Θ jk ) j,k,β Θ βjk 1 Θjk Θ ββjk 0. This proves the expected inequality. Moreover, the equality holds if and only if we have Θ βjk = 0 for β, Θ jk = γ jk for all. where γ = i j,k γ jkdz j dz k is a (1, 1)-form (take e.g., γ jk = Θ 11jk ). Hence Θ h (E) = γ Id E. By taking the trace with respect to E in this last equality, we get c 1 (E) h = r γ. Therefore the equality occurs if and only if Θ h (E) = 1 r c 1(E) h Id E. Corollary 1. Let (E, h) be a Hermite-Einstein vector bundle with c 1 (E) = 0 and c (E) = 0. Then E is unitary flat for some hermitian metric h = h e ϕ. Proof. By the assumption c 1 (E) = 0, we can write c 1 (E) h = i π ψ for some global function ψ on X. The equality case of the Kobayashi-Lübke inequality yields Θ h e ψ/r(e) = Θ h (E) 1 i r π ψ Id E = 0. Corollary. Let X be a compact Kähler manifold with c 1 (X) = c (X) = 0. Then X is a finite unramified quotient of a torus. Proof. By the Aubin-Calabi-Yau theorem, X admits a Ricci-flat Kähler metric ω. Since Ricci(ω) = Tr ω Θ ω (T X ), we see that (T X, ω) is a Hermite-Einstein vector bundle, and c 1 (T X ) ω = Ricci(ω) = 0. By the Kobayashi-Lübke inequality, we conclude that (T X, ω) is unitary flat, given by a unitary representation 3

π 1 (X) U(n). Let X be the universal covering of X and ω the induced metric. Then (T X, ω) is a trivial vector bundle equipped with a flat metric. Let (ξ 1,...,ξ n ) be an orthonormal parallel frame of T X. Since ξ j = 0, we conclude that dξ j = 0, and it is easy to infer from this that [ξ j, ξ k = 0. The flow of each vector field λj ξ j is defined for all times (this follows from the fact the length of a trajectory is proportional to the time, and ω is complete). Hence we get an action of C n on X, and it follows easily that ( X, ω) (C n, can). Now, π 1 (X) acts by isometries on this C n. The classification of subgroups of affine transformations acting freely (and with compact quotient) shows that π 1 (X) must be a semi-direct product of a finite group of isometries by a group of translations associated to a lattice Λ C n. Hence there is an exact sequence 0 Λ π 1 (X) G 0 where G is a finite group of isometries. It follows that there is a finite unramified covering map C n /Λ X/π 1 (X) X of X by a torus. We now discuss the special case of the tangent bundle T X in case (X, ω) is a compact Kähler-Einstein manifold. The Kähler-Einstein condition means that Ricci(ω) = λω for some real constant λ, i.e., Tr ω Θ ω (T X ) = λid TX. In particular, (T X, ω) is a Hermite-Einstein vector bundle. Here, however, the coefficients (Θ βjk ) 1 n of the curvature tensor Θ ω (T X ) satisfy the additional symmetry relations ( ) Θ βjk = Θ jβk = Θ kjβ = Θ jkβ. These relations follow easily from the identity Θ βjk = ω β / z j z k in normal coordinates, when we apply the Kähler condition ω β / z j = ω jβ / z. It follows that the Chern forms satisfy a slightly stronger inequality than the general inequality valid for Hermite-Einstein bundles. In fact (T X, ω) satisfies a similar inequality where the rank r = n is replaced by n + 1. Guggenheimer-Yau inequality. Let (T X, ω) be a compact n-dimensional Kähler-Einstein manifold, with constant λ IR. If λ = 0, then c (T X ) ω ω n 0. If λ 0, we have the inequality [ nc1 (X) (n + )c (X) (λc 1 (X)) n 0, and the equality also holds pointwise if we replace the Chern classes by the Chern forms c k (T X ) ω. The equality occurs in the following cases: (i) If λ = 0, then (X, ω) is a finite unramified quotient of a torus. (ii) If λ > 0, then (X, ω) (IP n, Fubini Study). (iii) If λ < 0, then (X, ω) (IB n /Γ, Poincaré metric), i.e. X is a compact unramified quotient of the ball in C n. Corollary (Bogomolov-Miyaoka-Yau). Let X be a surface of general type with K X ample. Then there is an inequality c 1 (X) 3 c (X), and the equality occurs if and only if X is a quotient of the ball IB. 4

Miyaoka has shown that the inequality holds in fact as soon as X is a surface with general type. Proof. As in the proof of the Kobayashi-Lübke inequality, we find n c 1 (T X ) ω (n + )c (T X ) ω =,β Θ Θ ββ + (n + 1)Θ β Θ β. Taking the wedge product with ω n /(n )!, we get [ n c 1 (T X ) ω (n + ) c (T X ) ω ω n = (n )! (Θ jj Θ ββkk Θ jk Θ ββkj ) + (n + 1)(Θ βjj Θ βkk Θ βjk Θ βkj ). If we had the factor r = n instead of (n + 1) in the right hand side, the terms in jj and kk would cancel (as they did before). Hence we find [ n c 1 (T X ) ω (n + ) c (T X ) ω ω n = = (n )! Θ jk Θ ββkj + Θ βjj Θ βkk (n + 1)Θ βjk Θ βkj Θ jk Θ ββkj (n + 1)Θ βjk Θ βkj. Here, the symmetry relation ( ) was used in order to obtain the equality of the summation of the first two terms. Using also the hermitian symmetry relation, our sum Σ can be rewritten as Σ = Θ jk Θ ββjk (n + 1) Θ βjk + n Θ jk,j,k = Θ jk Θ ββjk (n + 1) Θ βjk, pairwise [ (n + 1) 8 Θ jk + 4 Θ j + 4 Θ jj + Θ j <j [ + n Θ jk + Θ j + Θ jj + Θ j <j = Θ jk Θ ββjk (n + 1) Θ βjk β,j,k (4n + 8) + (n 1) Θ. Θ jk 4 j, pairwise Θ j 4 <j Θ jj All terms are negative except the last one. We try to absorb this term in the 5

summations involving the coefficients Θ jj. This gives Σ = Θ jk Θ ββjk (n + 1) Θ βjk β,j k (4n + 8) β,j, pairwise Θ jk 4 Θ j j Θ jj Θ ββjj 4 Θ jj + (n 1) Θ. <j The last line is equal to Θ jj Θ ββjj Θ ββ Θ ββββ β j β 4 Θ ββ + (n 1) Θ <β = Θ jj Θ ββjj β j (n 1) Θ 8 <β Θ ββ + β Θ ββ Θ ββββ + Θ ββ Θ ββββ = β j Therefore we find Σ = Θ jj Θ ββjj β β,j k (4n + 8) β j Θ Θ ββ. Θ jk Θ ββjk (n + 1) Θ jk 4 j Θ jj Θ ββjj β, pairwise Θ j Θ Θ ββ. Θ βjk This proves the expected inequality Σ 0. Moreover, we have Σ = 0 if and only if there is a scalar µ such that Θ ββ = Θ ββ = Θ ββ = µ for β, Θ = µ, and all other coefficients Θ βjk are zero. By taking the trace j Θ jj, we get λ = (n+1)µ. We thus obtain that the hermitian form associated to the curvature tensor is Θ ω (T X )(ξ η), ξ η ω = λ ξ η β ξ n + 1 η β + ξ η β ξ β η ( ),β = λ ( ξ η + ξ, η ) n + 1 for all ξ, η T X. When λ = 0, the curvature tensor vanishes identically and we have already seen that X is a finite unramified quotient of a torus. Assume from 6

now on that λ 0. The formula ( ) shows that the curvature tensor is constant and coincides with the curvature tensor of IP n (case λ > 0), or of the ball IB n (case λ < 0), relatively to the canonical metrics on these spaces. By a well-known result from the theory of hermitian symmetric spaces, it follows that (X, ω) is locally isometric to IP n (resp. IB n ). Since the universal covering X is a complete and locally symmetric hermitian manifold, we conclude that X IP n, resp. X IB n. In the case λ > 0, X is a Fano manifold, thus X is simply connected and X = X. The proof is complete. To compute the curvature of IP n and IB n, we use the fact that the canonical metric is ω = i π log(1 + z ) on IP n, resp. ω = i π log(1 z ) on IB n, with respect to the non homogeneous coordinates on IP n. We thus get ω = i ( dz dz dz, ) z, resp. ω = i ( dz dz dz, ) z +. π 1 + z (1 + z ) π 1 z (1 z ) By computing the derivatives Θ βjk = ω β / z j z k at z = 0, we easily see that the curvature is given by ( ) with λ = ±(n + 1). The equality also holds at any other point by the homogeneity of IP n and IB n. Observe that the riemannian exponential map exp : T X,0 X at the origin of C n IP n or IB n is unitary invariant. It follows that the holomorphic part h of the Taylor expansion of exp at 0 is unitary invariant. This invariance forces h to coincide with the identity map in the standard coordinates of IP n and IB n. From this observation, it is not difficult to justify intuitively the local isometry statement used above. In fact let X be a hermitian manifold whose curvature tensor is given by ( ), λ 0. Then ω is proportional to c 1 (T X ) ω and so ω is a Kähler-Einstein metric. Since the Kähler-Einstein equation is (nonlinear) elliptic with real analytic coefficients in terms of any real analytic Kähler form, it follows that ω is real analytic, and so is the exponential map. Fix a point x 0 X and let h : C n T X,x0 X be the holomorphic part of the Taylor expansion of exp at the origin. Then h must provide the holomorphic coordinates we are looking for, i.e. h ω must coincide with the metric of IP n (resp. IB n ) in a neighborhood of the origin. See for example, F. Tricerri et L. Vanhecke, Variétés riemanniennes dont le tenseur de courbure est celui d un espace symétrique riemannien irréductible, C. R. Acad. Sc. Paris, 30 (1986), 33-35. 7