COMPLEX GEOMETRY Course notes
COMPLEX GEOMETRY Course notes
COMPLEX GEOMETRY Course notes
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
MARCO A. PÉREZ B.<br />
Université du Québec à Montréal.<br />
Département de Mathématiques.<br />
<strong>COMPLEX</strong> <strong>GEOMETRY</strong><br />
<strong>Course</strong> <strong>notes</strong><br />
December, 2011.
These <strong>notes</strong> are based on a course given by Steven Lu in Fall 2011 at UQÀM. All errors<br />
are responsibility of the author.<br />
On the cover: a picture of the Riemann Sphere<br />
(taken from: http://en.wikipedia.org/wiki/Riemann sphere).<br />
i
TABLE OF CONTENTS<br />
1 <strong>COMPLEX</strong> ANALYSIS 1<br />
1.1 Complex Analysis in one variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1<br />
1.2 Analyticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4<br />
1.3 Complex Analysis in several variables . . . . . . . . . . . . . . . . . . . . . . . . . . 6<br />
2 RIEMANN SURFACES 9<br />
2.1 Complex manifolds, Lie groups and Riemann surfaces . . . . . . . . . . . . . . . . 9<br />
2.2 Holomorphic maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11<br />
2.3 Meromorphic functions and differentials . . . . . . . . . . . . . . . . . . . . . . . . . 12<br />
2.4 Weierstrass P -function on C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15<br />
2.5 Dimension on Riemann surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16<br />
2.6 Covering spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18<br />
2.7 The Riemann surface of an algebraic function . . . . . . . . . . . . . . . . . . . . . 20<br />
2.8 Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21<br />
2.9 Topology of Riemann surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24<br />
2.10 Product structures on ⊕ i Hi dR<br />
(Z) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27<br />
2.11 Questions about (compact) Riemann surfaces . . . . . . . . . . . . . . . . . . . . . 31<br />
2.12 Harmonic differentials and Hodge decompositions . . . . . . . . . . . . . . . . . . . 32<br />
2.13 Analysis on the Hilberts space of differentials . . . . . . . . . . . . . . . . . . . . . 34<br />
2.14 Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36<br />
2.15 Proof of Weyl’s Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37<br />
2.16 Riemann Extension Theorem and Dirichlet Principle . . . . . . . . . . . . . . . . 39<br />
2.17 Projective model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40<br />
2.18 Arithmetic nature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41<br />
iii
3 <strong>COMPLEX</strong> MANIFOLDS 43<br />
3.1 Complex manifolds and forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43<br />
3.2 Kähler manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46<br />
3.3 Metrics and connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48<br />
3.4 Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49<br />
3.5 The Fubini Study metric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50<br />
4 SHEAF COHOMOLOGY 53<br />
4.1 Sheaves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53<br />
4.2 Cohomology of sheaves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58<br />
4.3 Coherent sheaves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59<br />
4.4 Derived functors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60<br />
5 HARMONIC FORMS 63<br />
5.1 Harmonic forms on compact manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . 63<br />
5.2 Some applications of the Main Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 66<br />
5.3 Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67<br />
5.4 Heat equation approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68<br />
5.5 Index Theorem (Heat Equation approach) . . . . . . . . . . . . . . . . . . . . . . . 71<br />
BIBLIOGRAPHY 73<br />
iv
Chapter 1<br />
<strong>COMPLEX</strong> ANALYSIS<br />
1.1 Complex Analysis in one variable<br />
Let U ⊆ C = R 2 be an open subset of the complex plane. We shall denote an element z ∈ C by z = x + iy,<br />
where i = √ −1. An function f : U −→ C is holomorphic on U if it is complex differentiable at all points<br />
of U, i.e.,<br />
f ′ (z 0 ) = df<br />
dz (z 0) = lim z→z0<br />
f(z)−f(z 0)<br />
z−z 0<br />
exists for every z 0 ∈ U. We shall denote this by f ∈ O(U). If S ⊆ C is any subset, we shall say that f is<br />
holomorphic on S (f ∈ O(S)) if f is holomorphic on a open neighbourhood of S.<br />
If the function f is R-differentiable on U then ∂f ∂f<br />
∂x<br />
dx +<br />
∂y makes sense and df(x, y) ∈ Hom R(T z=x+iy U, R 2 ).<br />
Recall that<br />
dz = dx + idy and dz = dx − idy<br />
Using these expressions, we can write the differential df as<br />
df = 1 2<br />
( ) ( )<br />
∂f<br />
∂x − i ∂f<br />
∂y<br />
dz + 1 ∂f<br />
2 ∂x + i ∂f<br />
∂y<br />
dz = ∂f ∂f<br />
∂z<br />
dz +<br />
∂z dz<br />
Notice the following relations<br />
( )<br />
∂f<br />
∂z<br />
= ∂f<br />
∂z<br />
and<br />
( )<br />
∂f<br />
∂z<br />
= ∂f<br />
∂z<br />
Recall that<br />
df =<br />
f is complex differentiable<br />
⇐⇒<br />
f is R-differentiable and ∂f<br />
∂z<br />
= 0 (Cauchy-Riemann condition)<br />
⇐⇒<br />
∂u ∂v<br />
∂z<br />
= −i<br />
∂z<br />
( )<br />
⇐⇒<br />
ux u y<br />
is a rotation matrix up to a real scalar multiple (u<br />
v x v x = v y and u y = −v x ).<br />
y<br />
1
We shall denote V ⊂⊂ U if V ⊆ U is compact (V is precompact in U) and ∂V is rectifiable, i.e., ∂V is<br />
piecewise smooth.<br />
Theorem 1.1.1 (Cauchy). f ∈ O(U) if and only if ∫ f = 0, for every V ⊂⊂ U simple connected.<br />
∂V<br />
Theorem 1.1.2 (Cauchy’s Integral Formula). z 0 ∈ V ⊂⊂ U if and only if<br />
f(z 0 ) = 1 ∫<br />
2πi ∂V<br />
f(z)<br />
z − z 0<br />
dz.<br />
If V = D ɛ (z 0 ) is a disk centered at z 0 of radius ɛ, then we shall denote the previous integral by<br />
f(z 0 ) = Avg ∂V (f) := 1<br />
2π<br />
∫ 2π<br />
0<br />
f(z 0 + ɛe iθ )dθ.<br />
Corollary 1.1.1 (Liouville Theorem). Every holomorphic function on C is constant.<br />
Proof: Let V ɛ = B ɛ (z 0 ), z 0 ∈ C. We show that f ′ (z 0 ) = 0. Using the Cauchy’s Integral formula, we<br />
have<br />
f ′ (z 0 ) = 1 ∫<br />
f(z)<br />
2πi (z − z 0 ) 2<br />
Notice that |f| is bounded on ∂V ɛ . Then |f(z)| ≤ M on ∂V ɛ for some M > 0. So we have<br />
|f ′ (z 0 )| = 1<br />
∫<br />
∣ f(z) ∣∣∣<br />
2π ∣<br />
∂V ɛ<br />
(z − z 0 ) 2 dz ≤ 1 ∫<br />
|f(z)|<br />
2πɛ 2 ∂V ɛ<br />
|z − z 0 | 2 dz<br />
= 1 ∫<br />
2πɛ 2 |f(z)|dz ≤<br />
M ∫<br />
∂V ɛ<br />
2πɛ 2 dz<br />
∂V ɛ<br />
= M<br />
2πɛ 2 · 2πɛ = M ɛ<br />
It follows |f ′ (z 0 )| −→ 0 as ɛ −→ ∞. Hence f ′ (z 0 ) = 0 for every z 0 ∈ C and f is constant in C.<br />
∂V ɛ<br />
Corollary 1.1.2 (Riemann Extension Theorem). If f ∈ O(U − z 0 ), bounded near z 0 and continuous at z 0 ,<br />
then f ∈ O(U).<br />
Proof: If f ∈ O(U − z 0 ) then f ∈ O(U − B ɛ (z 0 )), for some ɛ > 0. Then the result follows since the<br />
Cauchy’s Integral Formula still holds in this case.<br />
2
Theorem 1.1.3 (Local Structure of f ∈ O(U)). If f ∈ O(U) is non-constant at z 0 ∈ U. Let<br />
m = min{n > 0 / f (n) (z 0 ) ≠ 0}.<br />
Then there exists a bi-holomorphic function ϕ : V −→ W from a neighbourhood of V of z 0 to a neighbourhood<br />
W of 0 ∈ C with ϕ(z 0 ) = 0 such that<br />
f(z) − f(z 0 ) = ϕ(z) m , for every z ∈ V.<br />
Proof: Note that f(z) − f(z 0 ) = (z − z 0 ) m g(z) with g(z 0 ) ≠ 0 and g ∈ O(U). Since the quotient<br />
f(z)−f(z 0)<br />
z−z 0<br />
is bounded on U − z 0 and continuous at z 0 , we have by the Riemann Extension Theorem that<br />
(z − z 0 ) m−1 g(z) = f(z)−f(z0)<br />
z−z 0<br />
is holomorphic on U. Proceeding this way, we have that g(z) ∈ O(U).<br />
We study several cases: If n = 1 then f ′ (z 0 ) ≠ 0 and by the Inverse Function Theorem we can choose<br />
ϕ(z) = f(z) − f(z 0 ). Now assume n ≠ 1. Since g(z 0 ) ≠ 0 then g(z) ≠ 0 on a neighbourhood of z 0 . So<br />
we can write g = h m on a neighbourhood V . We have<br />
f(z) − f(z 0 ) = [h(z)(z − z 0 )] m<br />
with ϕ ′ (z 0 ) = h(z 0 ) ≠ 0. Hence, up to a local change of coordinates, f is locally of the form z ↦→ z m for<br />
some m. Such a number m is called the ramification degree of f at z 0 .<br />
Corollary 1.1.3 (Open Mapping Theorem). If f ∈ O(U) is non-constant and U is connected, then f is an<br />
open mapping.<br />
Corollary 1.1.4. If f ∈ O(U) and |f| has a local maximum at z 0 ∈ U, where U is an open connected set,<br />
then f is constant on U.<br />
Proof: Suppose f is not constant. Then by the Open Mapping Theorem, we have that B ɛ (f(z 0 )) ⊆ f(U)<br />
for some ɛ > 0. In this neighbourhood there are some points of modulus greater that 0. Hence f(z 0 ) is<br />
not a local maximum.<br />
3
1.2 Analyticity<br />
A function f : U −→ C is said to be real analytic on U if for every z 0 = (x 0 , y 0 ) ∈ U there exists a<br />
neighbourhood V of z 0 such that<br />
f(z) = ∑ ∞<br />
α,β=0 a α,β(x − x 0 ) α (y − y 0 ) β for every z ∈ V.<br />
Similarly, f is said to be complex analytic on U if for every z 0 ∈ U there exists a neighbourhood V of z 0<br />
such that<br />
f(z) = ∑ ∞<br />
n=0 a n(z − z 0 ) n for every z ∈ V.<br />
In both cases the equality means normal convergence in U, i.e., uniform convergence on compacts in U.<br />
Theorem 1.2.1. f ∈ O(U) if and only if f is complex analytic on U.<br />
Proof: We know that<br />
On the other hand,<br />
f(z) = 1 ∫<br />
f(w)<br />
2πi ∂D r(z 0) w − z dw.<br />
(<br />
1<br />
w − z = 1 1<br />
w − z 1 − z−z0<br />
w−w 0<br />
)<br />
1<br />
=<br />
w − z 0<br />
∞ ∑<br />
n=0<br />
( z − z0<br />
w − z 0<br />
) n<br />
.<br />
It follows that<br />
where<br />
and |w − z 0 | = r on ∂D r (z 0 ).<br />
f(z) = 1 ∫<br />
2πi ∂D r(z 0)<br />
f(z)<br />
∞∑ (z − z 0 ) n<br />
∞<br />
(w − z 0 ) n+1 dw = ∑<br />
a n (z − z 0 ) n<br />
n=0<br />
n=0<br />
a n = 1 ∫<br />
f(w)<br />
2πi ∂D r(z 0) (w − z 0 ) n+1 dw = 1 n! f (n) (z 0 )<br />
Theorem 1.2.2. If f ∈ O(U) is non-constant, where U is connected, then f −1 (0) is discrete in U.<br />
Proof: Suppose f −1 (0) is not discrete. Let γ be an isolated point in f −1 (0) and consider the Taylor<br />
expansion of f about γ,<br />
∞∑ f (n)<br />
f(z) = (z − γ) n<br />
n!<br />
n=0<br />
for every z ∈ D r (γ), where r is the radius of convergence of the series. Since γ is not isolated in f −1 (0),<br />
there exists z ∈ f −1 (0) ∩ D r (γ). We have<br />
0 =<br />
∞∑<br />
n=0<br />
f (n)<br />
(z − γ) n<br />
n!<br />
4
and so f (n) (γ) = 0 for every n ≥ 0. It follows that f is constant on an neighbourhood of γ, getting a<br />
contradiction.<br />
Corollary 1.2.1 (Analytic Continuation or Identity Theorem). If f = g on a non-discrete subset of U and<br />
f, g ∈ O(U), then f ≡ g on U.<br />
Lemma 1.2.1 (Schwartz). If f ∈ O(D) and f ≤ M on ∂D, |f(z)| ≤ M|z| on ∂D r for every r ∈ (−ɛ, 1), then<br />
f(z) = Nz; where |N| < M.<br />
Another version: If |f(z)| ≤ M on D and f(0) = 0, then f(z) = Nz with |N| < M. Here, D<br />
de<strong>notes</strong> the Poncaré disk, i.e., the disk {z ∈ C : |z| 2 < 1} endowed with the metric<br />
|z − w| 2<br />
δ(z, w) = 2<br />
(1 − |z| 2 )(1 − |w| 2 ) .<br />
5
1.3 Complex Analysis in several variables<br />
Let U ⊆ C n be an open subset.<br />
Definition 1.3.1. A function f : U −→ C in C R ′ (U) (R-differentiable on U) is called holomorphic, denoted<br />
f ∈ O(U), if for every u ∈ U, the differential df u ∈ Hom(T u U, C) is a C-linear map.<br />
We prove as before the following result:<br />
Theorem 1.3.1. The following conditions are equivalent:<br />
(1) f ∈ O(U).<br />
(2) For every z 0 ∈ U, f has the form<br />
f(z 0 + δ) = ∑ I<br />
a I δ I (normal convergence)<br />
where I = (i 1 , . . . , i n ) and z I = z i1<br />
1 · · · zin n .<br />
(3) If D = {(ζ 1 , . . . , ζ n ) / |ζ i − a i | < α i ∈ R >0 } is a poly-disk in U, then for every z = (z 1 , . . . , z n ) ∈ Int(D)<br />
( ) n ∫ 1<br />
dζ 1<br />
∧ · · · ∧<br />
dζ n<br />
2πi<br />
ζ 1 − z 1 ζ n − z n<br />
where δD = {|ζ i − a i |} = α i ⊂ ∂D.<br />
|ζ i−a i|=α i<br />
f(ζ)<br />
Note that if f ′ ∈ C R ′ (U) then<br />
df = ∑ ∂f<br />
∂x i<br />
dx i + ∑ ∂f<br />
∂y i<br />
dy i = 1 ∑ ( )<br />
∂f<br />
2 ∂x i<br />
− i ∂f<br />
∂y i<br />
dz i + 1 ∑ ( )<br />
∂f<br />
2 ∂x i<br />
+ i ∂f<br />
∂y i<br />
dz i<br />
Theorem 1.3.2. Let f ∈ C n+1 such that f(0) = 0. Write C n+1 = {(w, z 1 , . . . , z n ) = (w, z)}. If f ≢ 0 on<br />
the w − axis (z = 0) then on some neighbourhood V of 0, we have<br />
where g is never zero on V .<br />
f = (w, z)(w d + a 1 (z)w d−1 + · · · + a d (z))<br />
Denote p(w, z) = w d + a 1 (z)w d−1 + · · · + a d (z). Hence locally we have zero(f) = zero(p). It follows that<br />
the roots of p are single valued holomorphic functions w = b i (z) away from the discriminant locus of f<br />
(∆ f (z) = 0, where ∆ f (z) is a polynomial in the a i ’s). Hence {f = 0} is a étale cover that covers the hyperplane<br />
{w = 0} = {(z 1 , . . . , z n )}. So by induction we see that:<br />
Fact 1.3.1. The zero of a holomorphic function is the disjoint union of submanifolds of lower dimension.<br />
Definition 1.3.2. An analytic set is the set of common zeros of finitely many analytic functions.<br />
6
Theorem 1.3.3 (Riemann Extension Theorem).<br />
• Part I: If f is a holomorphic function and bounded outside an analytic subset of codimension 2 or<br />
higher, then f extends over the subset as a holomorphic function.<br />
• Part II (also known as the Hartogs Theorem):<br />
(1) Let U = ∆(r) = {(z 1 , z 2 ) / |z 1 | < r and |z 2 | < r} and V = ∆(r ′ ), such that r ′ < r and V ⊂⊂ U,<br />
then every f ∈ O(U − V ) extends to a holomorphic function on U.<br />
(2) If S ⊆ U ⊆ C n has complex codimension greater or equal than 2, where S is an analytic subset<br />
and f ∈ O(U − S), then f extends to a holomorphic function on U.<br />
Proof: We only proof the second part.<br />
(1) Take a slice z 1 = const. Then U − V = {r ′ < |z 2 | < r} on this slice. Set<br />
F (z 1 , z 2 ) = 1 ∫<br />
f(z 1 , w 2 )<br />
dw 2<br />
2πi |w 2|=r w 2 − z 2<br />
Hence F : U −→ C is holomorphic in z 1 since<br />
∂f<br />
= 0 =⇒ ∂F<br />
∂z 1 ∂z 1<br />
and clearly also in z 2 (Cauchy’s Integral Formula). Moreover, F = f on U − V by the Cauchy’s<br />
Integral Formula.<br />
(2) The a 2-dimensional slice and apply (1).<br />
Theorem 1.3.4 (Open Mapping Theorem). If f ∈ O(U) then f is open.<br />
Theorem 1.3.5 (Maximum Principle). |f| has no local maximum unless it is locally constant there.<br />
Theorem 1.3.6 (Analytic continuation). f = 0 in an open subset of U and f ∈ O(U), where U is connected<br />
(or arcwise connected), then f ≡ 0 on U.<br />
Proof: For every path α, the set I = {c / f ◦ α(t) = 0 ∀t < c} is open and closed.<br />
7
Chapter 2<br />
RIEMANN SURFACES<br />
2.1 Complex manifolds, Lie groups and Riemann surfaces<br />
Definition 2.1.1. A complex manifold M is a topological manifold whose coordinate charts are open<br />
subsets of C n , such that the transition maps are holomorphic. The number n is called the dimension of M.<br />
A Riemann surface is a complex manifold of dimension 1. A complex Lie group is a group that is a<br />
complex manifold such that the product and inversion maps are holomorphic.<br />
Remark 2.1.1. Manifolds are always connected unless otherwise specified.<br />
Example 2.1.1. The following sets are complex manifolds:<br />
(1) CP n = {[z 0 : · · · : z n ]} = C n+1 − {0}/ ∼, where<br />
(z 0 , . . . , z n ) ∼ (z ′ 0, . . . , z ′ n) ⇐⇒ z = tz ′ , for some t ∈ C ∗ .<br />
Let U 0 = {[z 0 : · · · : z n ] / z 0 ≠ 0}. Let ϕ 0 : U 0 −→ C n be the map given by<br />
(<br />
z1<br />
[z 0 : · · · : z n ] ↦→ , . . . , z )<br />
n<br />
z 0 z 0<br />
which we shall call the 0-th affine chart. Note that there exist n + 1 affine charts that cover CP n .<br />
(2) The compact complex torus C n /Γ, where Γ ∼ = Z n .<br />
(3)<br />
PGL(1, 0) = Aut(CP 1 ) = set of Moebius transformations<br />
{ }<br />
az0 + bz 1<br />
=<br />
/ ad − bc ≠ 0 /{±1}<br />
cz 0 + dz 1<br />
The following sets are Riemann surfaces:<br />
(4) CP 1 .<br />
9
(5) C and C ∗ = C − {0}.<br />
(6) The half-plane model H = {z / Im(z) > 0}.<br />
(7) The Poincaré disk model ∆ = D = {z / |z| < 1} and ∆ ∗ = ∆ − {0}. The models H and D are related<br />
by the map<br />
z ↦→ z − a<br />
z − a<br />
The exponential function exp : C −→ C ∗ is a universal covering.<br />
Definition 2.1.2. The manifold C/Γ (and more generally its nontrivial holomorphic images) is called an<br />
elliptic curve. The manifold CP 1 is called a rational curve.<br />
Note that CP 1 has positive curvature, C has zero curvature (in other words, C is said to be flat), and H and<br />
D have negative curvature. Note that Γ is a lattice {n + α / α ∈ H}. The parameter space of elliptic<br />
curves is the quotient H/SL(2, Z). Note that C/Γ has genus 1.<br />
Example 2.1.2.<br />
(1) Let f(z 0 , . . . , z n ) be a homogeneous polynomial. Then<br />
C = V (f) = {[z 0 : z 1 : z 2 ] / f(z 0 , z 1 , z 2 ) = 0} ⊆ CP<br />
is called an algebraic plane curve over C. This cuve C is smooth (or non-singular) if it is a<br />
submanifold (only need to check df(p) ≠ 0 for every p ∈ C to have C smooth).<br />
(2) x d + y d + z d gives the Fermat curve of degree d in CP 2 . It is smooth since df ≠ (0, 0, 0) on C,<br />
df = (x d−1 dx, y n−1 dy, z n−1 dz).<br />
10
2.2 Holomorphic maps<br />
Definition 2.2.1. Let f : X −→ Y be a continuous map of topological spaces, the inverse image f −1 (p) is<br />
called the fibre of f at p ∈ Y . The map f is called discrete if all its fibres are discrete in X.<br />
Theorem 2.2.1 (Identity Theorem). If f 1 , f 2 : S 1 −→ S 2 are mappings of Riemann surfaces such that they<br />
coincide on a non-discrete subset of S 1 , then f 1 ≡ f 2 .<br />
Theorem 2.2.2. Any non-constant mapping of Riemann surfaces is discrete.<br />
The Open Mapping Theorem implies the following result:<br />
Theorem 2.2.3. Let f : S 1 −→ S 2 be a non-constant map of Riemann surfaces. Assume that S 1 is<br />
compact. Then f is surjective and S 2 is compact. Furthermore, if f is proper (i.e., f −1 (C) is compact<br />
for every compact set C) and discrete with finite fibres (i.e., a finite map) then f is called a branched<br />
covering map (at a branch f looks like z ↦→ z d , where d is called the branched degree).<br />
Let p ∈ S be a ramification point. At such a point we call the multiplicity (or the ramification degree)<br />
of f<br />
mult p (f) = d.<br />
The degree of f is defined by<br />
deg(f) = ∑ p∈F mult p(f)<br />
for every fibre F . The ramification index of f at p is<br />
r p (f) = mult p (f) − 1.<br />
A map is said to be unramified if r p (f) = 0 for every p ∈ S.<br />
11
2.3 Meromorphic functions and differentials<br />
Definition 2.3.1. A meromorphic function on a Riemann surface Z is a holomorphic function on an<br />
open subset U ⊆ Z where Z − U is a discrete set consisting of at most poles of the function.<br />
Recall that a pole p ∈ Z − U is defined by one of the following equivalent conditions:<br />
(a) lim z→p f(z) = ∞.<br />
(b) f can be written locally as a Laurent series<br />
∞∑<br />
f(z) = a i z i<br />
−∞<br />
with a i = 0 for every i < n ∈ Z.<br />
(c) f = g/h, where g, h ∈ O(p), g(p) ≠ 0 and h(p) = 0.<br />
The set of such functions is denoted M(Z). We have<br />
f ∈ M(Z) ⇐⇒ f : Z hol<br />
−→ CP 1<br />
and that<br />
poles of f = f −1 (∞)<br />
Example 2.3.1.<br />
(1) A non-constant polynomial defines a meromorphic function from CP 1 with pole order at ∞ equal to<br />
deg(f) ≥ 1.<br />
(2) A rational function p(z)/g(z) defines a meromorphic function with pole order at ∞ equal to deg(p) −<br />
deg(g). If this difference is negative then f has a zero at ∞.<br />
Fact 2.3.1. M(Z) is a field.<br />
A finite map of Riemann surfaces f : Z 1 −→ Z 2 corresponds to a finite field extension<br />
f ∗ : M(Z 2 ) ↩→ M(Z 1 ).<br />
Definition 2.3.2. A meromorphic differential on a Riemann surface Z is a holomorphic differential ω<br />
on an open U ⊆ Z whose complement Z − U is discrete and consist of poles of ω. Locally, ω = fdz even at<br />
a pole. The pole order of ω is defined by that of f (locally) and its residue at p is the same as that of fdz<br />
(p = 0), denoted Res p (ω).<br />
Theorem 2.3.1 (Residue). V ⊂⊂ Z with rectifiable boundary ∂V and ω differentiable on Z. Then<br />
∫<br />
ω = ∑ Res p (ω)<br />
p∈V<br />
∂V<br />
12
hol<br />
We shall denote the space of meromorphic differentials by M ′ (Z).<br />
Theorem 2.3.2. ω ∈ M ′ (Z), ∑ p∈Z Res p(w) = 0.<br />
Corollary 2.3.1 (Sum rule or product rule, Reciprocity Theorem). If Z is compact and f ∈ M(Z), then<br />
on Z counting multiplicity.<br />
#zero(f) = #poles(f)<br />
Corollary 2.3.2. If f ∈ M(CP 1 ) then f is rational. Hence M(CP 1 ) = C(Z).<br />
Corollary 2.3.3. M ′ (CP ′ ) = C(Z)dz.<br />
Let S be a Riemann surface and U ⊂⊂ S a relative compact open subset of S with good ∂U. Let ω be a<br />
meromorphic differential (ω ∈ M ′ (S)). Then we have<br />
∫<br />
ω = ∑ Res p ω<br />
∂U<br />
p∈U<br />
If ∂U = ∅ (so U = S is compact) then<br />
∑<br />
Res p ω = 0<br />
p∈S<br />
We have that if S is compact then S = R(S), the set of rational functions (later, we are going to study this<br />
in detail).<br />
Example 2.3.2. M(P 1 ) = R(P 1 ), where P 1 = CP 1 . Note that CP 1 is compact and CP 1 = C ∪ C, where<br />
there is a map between charts<br />
z ∈ C w= 1 z<br />
↦→ w ∈ C<br />
Since ∑ p∈S Res pω = 0, we have ω = df f , and Res pω = ord p f.<br />
Recall that f ∈ M(P 1 ) if and only if f : P 1 −→ P 1 is holomorphic.<br />
Example 2.3.3. Let C be a quadratic (conic) curve in CP 2 = P 2 . Let p ∈ C. The space of lines passing<br />
through p is P 2 and so gives a meromorphic map P 2 −→ P 1 . We have the following diagram:<br />
P 2 P 1<br />
P 2 − {p}<br />
hol<br />
C − {p}<br />
13
This shows that C ∼ = hol P 1 . For example, we have the Fermat curve z 2 0 + z 2 1 = z 2 2. If P 1 = {[u : v]}, then<br />
define a map z 0 = (u − v) 2 , z 1 = 2uv and z 2 = (u + v) 2 .<br />
Note that π 1 (S) = 0 where S = P 1 , C, D. Recall that if π 1 (S) = Z then S is not compact.<br />
Example 2.3.4. S 1 = C ∗ , S 2 = D − 1 2D. These two examples are not biholomorphic Riemann surfaces.<br />
Neither D ∗ is biholomorphic to C ∗ . If so, then a biholomorphic function D ∗ −→ C ∗ produces an extension<br />
D −→ S, getting a contradiction.<br />
The Riemann surfaces H and D are biholomorphic via the map<br />
ω ↦→ w − a<br />
w − a<br />
14
2.4 Weierstrass P -function on C<br />
Consider the torus C/Γ, where<br />
Γ = 〈1, τ〉 τ∈C<br />
= Z ⊕ Zτ<br />
is a lattice in C. Define the Weirstrass P -function by the formula:<br />
Hence P ∈ M(C) and satisfies:<br />
(a) P(z) = P(−z),<br />
(b) P(z + λ) = P(z) for every λ ∈ Γ,<br />
(c) there exists no other poles.<br />
P(z) = p(z, Γ) = 1<br />
z<br />
+ ∑ 2 λ∈Γ−{0}<br />
[<br />
]<br />
1<br />
(z−λ)<br />
− 1 2 λ 2<br />
Hence P descends to a meromorphic function on C = C/Γ with a pole at 0 (degree = 2), i.e., we have a<br />
holomorphic function<br />
C −→ f<br />
CP 1<br />
Locally, this map looks like z ↦→ z n about 0. In this case n = 2. This map has exactly degree 2 since there<br />
are no other poles. Moreover, f is branched at 0.<br />
0<br />
∞<br />
CP 1<br />
Now P ′ (z) is also periodic (period Γ) with triple poles on Γ and no other poles. The map C −→ P 2 given by<br />
defines a holomorphic function C − {0} −→ P 2 ,<br />
and it extends over 0.<br />
z ∈ C ↦→ [P(z) : P ′ (z) : 1]<br />
[P(z) : P ′ (z) : 1] =<br />
15<br />
[ ]<br />
P(z)<br />
P ′ (z) : 1 : 1<br />
P ′ (z)
2.5 Dimension on Riemann surfaces<br />
Definition 2.5.1. A divisor D on a Riemann surface S is a formal Z-linear combination of points in S<br />
D = ∑ a i P i<br />
where a i ≠ 0 for every i and {P i } is a discrete subset of S. A divisor D is called effective if a i ≥ 0 for every i.<br />
If supp(D) = {P i / a i ≠ 0} is finite, then<br />
deg(D) := ∑ a i<br />
Definition 2.5.2. Let f ∈ M(S) − {0} = M ∗ (S). The divisor of f is defined by<br />
(f) := (f) 0 − (f) ∞<br />
where<br />
(f) 0 = ∑ (ord P f)P and (f) ∞ = ∑ P ∈f −1 (∞) (mult P f)P<br />
Note that mult P f = −ord P f, so we can rewrite the previos expresion as<br />
(f) = ∑ (ord P f)P<br />
Lemma 2.5.1. If S is a compact Riemann surface, then deg(f) = 0 for every f ∈ M ∗ (S), where deg is a<br />
map Div(S) −→ Z.<br />
Definition 2.5.3. A divisor is called principal if it lies in the image of deg( ) : M ∗ −→ Z.<br />
Definition 2.5.4. Two divisors D 1 and D 2 are said to be linearly equivalent, denoted D 1 ∼ D 2 , if D 1 −D 2<br />
is principal.<br />
Example 2.5.1. D 1 ∼ D 2 on P 1 if and only if deg(D 1 ) = deg(D 2 ).<br />
Example 2.5.2. What condition we need if we want D 1 ∼ D 2 on C = C/Γ. Let p, q ∈ C be two distinct<br />
points in C and suppose that D = p − q = (f) for some f ∈ M ∗ . Then f is a map C −→ P 1 with deg(f) = 1.<br />
We have (f) 0 = p and f is bijective. Then f is a biholomorphic map, getting a contradiction.<br />
Given ω ∈ M ′ (S) ∗ . Recall that this means ω = fdz for a local coordinate z at p and f ∈ M(p).<br />
Definition 2.5.5. ord p ω = ord p f. The divisor of the form<br />
(ω) := ∑ p∈S (ord pω)P<br />
is called a canonical divisor and is denoted K S or simply K. A divisor is canonical if D ∼ (w).<br />
16
Example 2.5.3.<br />
• K C/Γ = (dz) = 0 · P .<br />
• K P 1<br />
= (dz) = −2 · ∞, z = 1 dω<br />
ω<br />
, dz = −<br />
ω<br />
. 2<br />
Theorem 2.5.1 (Riemann - Hurewicz). If f : S 1 −→ S 2 is a finite map, then K S1<br />
is the ramification divisor<br />
R = ∑ r p (f)P,<br />
p∈S 1<br />
∼ K + S 2 + R where R<br />
where each r p (f) ≥ 0.<br />
Proof: Locally, f looks like z ↦→ z n , and dz n = nz n−1 dz.<br />
Corollary 2.5.1 (Riemann - Hurewicz formula).<br />
2 · g(S 1 ) − 2 = degK S1 = (degf) · degK S2 + degR,<br />
where K S1 = (ω), ω ∈ Γ(T ∨ S 1 ) (global sections of the tangent bundle) ←→ ω ∨ ∈ Γ(T S 1 ).<br />
17
2.6 Covering spaces<br />
Recall that an étalé space (or étale map) over X is a continuous map p : ˜X −→ X such that p is a local<br />
homeomorphism: that is, for every x ∈ ˜X, there is an open set U in ˜X containing x such that the image<br />
p(U) is open in X and the restriction of p to U is a homeomorphism p| U : U −→ p(U). A connected covering<br />
space p : ˜X −→ X is a universal cover if ˜X is simply connected. The name universal cover comes from<br />
the following important property: if the map p : ˜X −→ X is a universal cover of the space X and the map<br />
p ′ : X ′ −→ X is any cover of the space X where the covering space X ′ is connected, then there exists a<br />
covering map f : X ′ −→ ˜X such that p ◦ f = p ′ . Any manifold X has a universal cover ˜X with étale covering<br />
map f : ˜X −→ X and π1 (X) acts on ˜X discretely and freely with quotient f. Moreover, there exists a<br />
bijection<br />
étale<br />
π 1 (S) ⊇ H ↦→ { ˜X/H −→ X}<br />
between the set of subgroups of π 1 (X) up to conjugation and the set of étale coverings from a connected<br />
manifold.<br />
G/H for H normal ⇐⇒ Galois (regular) covering.<br />
Recall that a covering map p : ˜X −→ X is said to be Galois if for every x ∈ X and ˜x ∈ p −1 (x), the subgroup<br />
p ∗ π 1 ( ˜X, ˜x) is normal in π 1 (X, x).<br />
If X is a Riemann surface, then the complex charts on X lifts to any covering space.<br />
Lemma 2.6.1. Let f : Z 1 −→ Z be a finite étale covering corresponding to H ↩→ π 1 (Z). Then there exists<br />
a finite regular covering h : Z 2 −→ Z and a finite étale covering g : Z 2 −→ Z 1 such that f ◦ g = h.<br />
Z 1 Z 2<br />
étale covering<br />
Z<br />
∃regular<br />
Proof: H has a finite number of conjugates in π 1 (Z), such a number equals [π 1 (Z) : N(H)] and these<br />
intersection is then of finite index.<br />
Corollary 2.6.1. For every n, there exists a unique étale n-sheeted covering Z −→ D ∗ and it is isomorphic<br />
to D ∗ −→ D ∗ (z ↦→ z n ).<br />
Example 2.6.1. If f : Z 1 −→ Z 2 is finite between Riemann surfaces, then there is a finite étale covering<br />
where ∆ = f(suppR f ) is the branching locus.<br />
Z 1 − f −1 (∆) −→ Z 2 − ∆<br />
18
Conversely, the previous corollary implies:<br />
Theorem 2.6.1. Let ∆ ⊆ Z 2 be a discrete subset. A finite étale covering U −→ Z 2 − ∆, where U is an open<br />
subset, has a unique continuation to a finite map<br />
where Z 1 is a Riemann surface.<br />
U ⊆ Z 1 −→ Z 2<br />
19
2.7 The Riemann surface of an algebraic function<br />
Let Z 2 be a Riemann surface.<br />
Proposition 2.7.1. Let<br />
P (T ) = T n + c 1 T n−1 + · · · + c n<br />
in M(Z 2 )[T ] be an irreducible polynomial. Then there exists a map of Riemann surfaces f : Z 1 −→ Z 2 of<br />
degree n and a meromorphic function F ∈ M(Z 1 ) that satisfies:<br />
F n + f ∗ (c 1 )F n−1 + · · · + f ∗ (c n ) = 0.<br />
(∗)<br />
Proof: Let ∆ ⊆ Z 2 be the discrete set containing the poles of the c i ’s and the points p where<br />
P p (T ) := T n + c 1 (p)T n−1 + · · · + c n (p)<br />
has multiple roots. Then U = {(p, z) ∈ (Z 2 − ∆) × C / P p (z) = 0} is a Riemann surface, and Z 2 − ∆ is<br />
a finite étale cover. We claim that it is connected, i.e.,<br />
Claim: Given f : Z 1 −→ Z 2 , then every F ∈ M(Z 1 ) is algebraic over M(Z 2 ) and satisfies an equality<br />
of the form (∗) but with degree less or equal than the degf.<br />
Corollary 2.7.1. If Z 1 −→ Z 2 is finite, then<br />
is a finite field extension.<br />
f ∗ : M(Z 2 ) −→ M(Z 1 )<br />
Example 2.7.1. M(P 1 ) = C(Z), the field of rational functions in variable z.<br />
Theorem 2.7.1. As soon as there exists a meromorphic function f on Z 1 (←− compact =⇒ f is finite),<br />
M(Z 1 ) is finite algebraic over C(z) of extension deg = degf.<br />
Z 1<br />
f<br />
−→ P 1<br />
Corollary 2.7.2.<br />
(1) If Z 1<br />
f<br />
−→ Z 2 is finite then M(Z 2 ) f ∗<br />
↩→ M(Z 1 ) is a finite field extension of degree (f).<br />
(2) Conversely, let M(Z 2 ) ϕ<br />
↩→ M(Z 1 ) = K be a finite field extension of degree d. Then there exists a finite<br />
map Z 1 −→ Z 2 of degree d whose field extension is isomorphic to ϕ.<br />
(3) A field K of transcendental degree = 1 over C is isomorphic to M(Z) of some compact Riemann<br />
surface. Such a Z is called a model of K.<br />
20
2.8 Review<br />
Note that the ratio of two meromorphic 1-forms is a meromorphic function:<br />
( )<br />
ω 1 , ω 2 ∈ M ′ ω1<br />
(Z) =⇒ = (f) ∈ Div P (Z), where f ∈ M(Z).<br />
ω 2<br />
Here, Div P de<strong>notes</strong> the set of principal divisors. Note that ω ∈ M ′ (Z) if and only if ω ∈ Γ m (T ∨ Z), where<br />
T ∨ Z de<strong>notes</strong> the cotangent bundle of Z and Γ m (T ∨ Z) is the set of holomorphic sections of T ∨ Z. Also,<br />
T ⊗ T ∨ = O (trivial line bundle) and so<br />
( ) 1<br />
1 = · ω ∈ Γ m (O) = O,<br />
ω<br />
where 1 ω ∈ T Z and ω ∈ T ∨ Z.<br />
If ω 1 = fω 2 then (ω 1 ) = (f) + (ω 2 ). So we get the formula<br />
deg( ) = deg(f) + deg( )<br />
On the other hand, deg(f) = 0. To show this, we know that f(z) = z n locally, where n = ord(f). Then<br />
df<br />
f<br />
= n dz<br />
z and ∑ p∈Z Res p df f<br />
= 0, where Z is compact. Hence we get the following result:<br />
Theorem 2.8.1. deg(ω) has the same value for any ω ∈ M ′ (Z), assuming that Z is a compact and<br />
connected Riemann surface.<br />
Let g be the topological genus of Z. We have the following relations:<br />
deg(ω) = 2g − 2 and deg ( 1<br />
ω<br />
)<br />
= 2 − 2g = χ(Z)<br />
where 1 ω<br />
∈ Γ(T Z).<br />
genus<br />
= # of holes<br />
Recall the following results:<br />
21
Theorem 2.8.2. Let P (T ) = T n + c 1 T n−1 + · · · + c n ∈ M(Z)[T ] be an irreducible polynomial in T over Z.<br />
Then there exists a finite map of Riemann surfaces f : Z ′ −→ Z of degree n, unique up to isomorphisms,<br />
and a meromorphic function F on Z ′ satisfying<br />
F n + (f ∗ c 1 )F n−1 + · · · + (f ∗ c n ) = 0<br />
Corollary 2.8.1.<br />
(1) Z 1 −→ Z 2 finite =⇒ f ∗ : M(Z 2 ) ϕ<br />
↩→ M(Z 1 ) is a finite field extension of degree n.<br />
(2) Conversely, any finite field extension of degree n gives rise to a finite map Z 1 −→ Z 2 whose associated<br />
field extension is isomorphic to ϕ.<br />
(3) A field extension K of transcendence degree 1 (i.e., K is a finite field extension over C(Z)) is isomorphic<br />
to M(Z) for some Riemann surface Z (finite over CP 1 ).<br />
Z is called a smooth model of K.<br />
Theorem 2.8.3. Let C ⊆ CP 2 be an irreducible plane curve, i.e., C = V (F ) where F is an irreducible<br />
homogeneous polynomial in (z 0 : z 1 : z 2 ). Then there exists a compact Riemann surface Z and a generically<br />
injective map f : Z −→ CP 2 whose image is C.<br />
Here, generically injective means birational (←→ isomorphic on an open set).<br />
Proof: If F is irreducible then C[x 0 , x 1 , x 2 ]/(F, x 2 − 1) is an integral domain and has a quotient field<br />
K. The mapping f(P ) = (x 0 (P ) : x 1 (P ) : 1) extends to by continuity to a desingularization of C.<br />
Theorem 2.8.4 (Hurwitz). If f : Z 1 −→ Z 2 is a finite map, then K Z1 ∼ f ∗ K Z2 + R (f ∗ = pullback) where<br />
R is the ramification divisor ∑ p∈Z 1<br />
r p (f) · p, r p (f) = ord p (f) − 1, where f : Z 1 −→ Z 2 is locally at p of the<br />
form f(z) = z ordp(f) and<br />
deg(f) =<br />
∑<br />
ord p (f),<br />
p∈ fibre<br />
where this sum is independent of the choice of the fibre.<br />
Corollary 2.8.2 (Riemann-Hurwitz).<br />
deg(K Z1 ) = (deg(f)) · (deg(K Z2 )) + deg(R)<br />
Corollary 2.8.3. g(Z 1 ) ≥ g(Z 2 ), where g is for genus.<br />
22
From this it follows that there exists no map from a rational curve to an elliptic curve.<br />
The following equality is the algebraic geometry definition of topological genus:<br />
g(Z) = dim(Γ(K 2 )) = dim C (Hol ′ (Z))<br />
where Hol ′ (Z) is the set of holomorphic differentials on Z.<br />
23
2.9 Topology of Riemann surfaces<br />
One known fact about Riemann surfaces is that any Riemann surface is orientable. Recall that a manifold<br />
is orientable if its transition functions have positive Jacobian . Let Z be a Riemann surface, consider two<br />
intersecting charts U α and U β , and let z ∈ U α ∩ U β . Then we have<br />
df<br />
dz (z) = f ′ (z)<br />
as an R-matrix.<br />
We write f = u + iv, then df = α + iβ.<br />
β<br />
α ◦ β −1<br />
holomorphic<br />
α<br />
C<br />
By the Cauchy-Riemann equations, we have<br />
( α β<br />
df =<br />
−β α<br />
)<br />
and so<br />
Jac(α ◦ β −1 ) = det(D(α ◦ β −1 )) = |df| 2 = det(df) = α 2 + β 2 > 0.<br />
It is also known that every Riemann surface is obtained by attaching handles to CP 1 = S 2 .<br />
# handles = # holes = : topological genus = g<br />
∞<br />
0<br />
Theorem 2.9.1. Any Riemann surface is triangularizable.<br />
24
This fact is easy to prove for Compact Riemann surfaces, and shows that any such is obtained by attaching<br />
handles to S 2 = CP 1 = C ∪ {∞}. We denote Z g for a Riemann surface Z of genus g. It is known that the<br />
first homotopy group of Z g is given by<br />
π 1 (Z g ) = Z a1 ∗ Z b1 ∗ · · · ∗ Z ag ∗ Z bg / 〈 a 1 b 1 a −1<br />
1 b−1 1 · · · a g b g a −1<br />
g<br />
Notice that every hole gives two generators.<br />
b −1<br />
g<br />
〉<br />
b 1<br />
b g<br />
a g<br />
a 1<br />
a 1<br />
b 1<br />
a −1<br />
1<br />
b −1<br />
1<br />
b −1<br />
g<br />
Let A i denote the space of C ∞ -complex valued i-forms on a Riemann surface, then<br />
A = A 1,0 ⊕ A 0,1 ,<br />
ω ′ = fdx + gdy = hdz + hdz,<br />
where hdz ∈ A 1,0 and hdz ∈ A 0,1 . We also have differential operators making the following diagram commutes<br />
d<br />
A 1<br />
d<br />
C ∞ − functions = A 0 ∂<br />
A 0,1 ∂<br />
A 2<br />
∂<br />
A 1,0<br />
∂<br />
where d = ∂ + ∂.<br />
Definition 2.9.1. An i-form ω ∈ A i is called closed if dω = 0. It is called exact if ω = dα.<br />
Since d ◦ d = 0, we have Im(d) ⊆ Ker(d), and so we define the de Rham cohomology groups of Z and the<br />
quotient<br />
HdR i (Z) = {closed i-forms}/{exact i-forms}<br />
25
Note that HdR 0 (Z) = C if Z is connected. If Z is also compact, by the Poincaré duality Theorem we have an<br />
isomorphism HdR 2 (Z) −→ H0 dR (Z) given by ∫<br />
[ω] ↦→<br />
Also, HdR n (Z) = 0 for every n ≥ 3. For<br />
cohomology is given by<br />
Z<br />
ω<br />
any compact and connected Riemann surface Z g, the middle<br />
H 1 dR (Z g) = π 1 (Z g )/ 〈commutator subgroup〉 = Z a1 ⊕ Z b1 ⊕ · · · ⊕ Z ag ⊕ Z bg<br />
Theorem 2.9.2. Let ω be a holomorphic differential (←→ 1-form). Then ω is d-closed (hence [ω] ∈ H 1 dR (Z)).<br />
Proof: Locally, ω = f(z)dz where f ∈ O. Then<br />
dω = (∂ + ∂)ω = ∂ω + ∂ω =<br />
( ) ( )<br />
∂f<br />
∂f<br />
∂z dz ∧ dz +<br />
∂z dz ∧ dz<br />
= ∂f ∧ dz + ∂f ∧ d = 0 + 0dz ∧ dz, since f is holomorphic.<br />
26
2.10 Product structures on ⊕ i Hi dR (Z)<br />
By taking wedge products of forms, we have a surjective map HdR 1 (Z) × H1 dR<br />
(Z) −→ C given by<br />
∫<br />
([ω 1 ], [ω 2 ]) ↦→ ω 1 ∧ ω 2<br />
In this situation we have H 1 dR (Z) ∼ = (H 1 dR (Z))∨ (finite dimesnion) and so the previous map is a perfect pairing.<br />
Z<br />
We also have the cap product ∩ : H 1 (Z) × H 1 (Z) −→ Z which gives rive to a perfect pairing.<br />
By the Poincaré duality Theorem we have HdR 1 (Z) ∼ = H 1 (Z) ∨ ⊗ C. Also<br />
HdR(Z) 1 = (H 1 (Z) ∗ = H 1 (Z)) ⊗ C.<br />
Recall that the Euler characteristic of Z is given by<br />
χ(Z) := ∑ i (−1)i dimHdR i (Z) = ∑ i (−1)i dim Z H i (Z) = 2 − 2g<br />
Recall H 0 dR (Z) = H2 dR (Z) = C and H1 dR (Z) = C2g .<br />
Definition 2.10.1. Let Ω be the space of differentials (= holomorphic 1-forms) on a compact Riemann<br />
surface Z. The geometric genus of Z is defined by<br />
P g (Z) = dim C (Ω)<br />
Clearly, Ω ↩→ HdR 1 (Z) if Z is compact. We have that<br />
fdz = dg dz = dg = 0<br />
dz<br />
implies that g ∈ O. Since Z is compact, we get that g = const. Similarly, Ω ↩→ H 1 dR (Z). It follows 2P g ≥ 2g.<br />
It is difficult to determine when they are equal. The fact that P g ≥ 2 implies that there exist meromorphic<br />
functions. If P g = g, then a loop C on a compact Riemann surface Z is homotopic to 0 if and only if ∫ C ω = 0,<br />
for every ω ∈ Ω. The map<br />
is called the period map.<br />
ω ↦→ ∫ C ω<br />
Definition 2.10.2. Let S be a Riemann surface, define<br />
H 1 (S; Z) = π 1 (S)/[π 1 (S), π 1 (S)]<br />
where [π 1 (S), π 1 (S)] de<strong>notes</strong> the commutator subgroup of π 1 (S).<br />
Note that if S is compact then H 1 (−) is a free Z-module generated by a 1 , b 1 , . . . , a n , b n . We can consider<br />
each a i and b i as maps S 1 −→ S or curves with parameter t ∈ [0, 1] in S, oriented counterclockwise.<br />
27
1<br />
b g<br />
a g<br />
a 1<br />
Since S is oriented we have that there exists an intersection pairing in H 1 (any two loops, or elements in H 1 ,<br />
are homotopic to loops which are transversal). Note that a and b are the same element in H 1 if and only<br />
if a + (−b) = a − b = ∂U, for some open set U ⊆ S, i.e., a is homologous to b. Here −b de<strong>notes</strong> reverse<br />
orientation. So if a is homotopic to b then they are homologous.<br />
Consider the following picture:<br />
chart<br />
−→ b<br />
y<br />
C<br />
p<br />
−→ a<br />
x<br />
We have −→ a ∧ p<br />
−→ b = kx ∧p y, and denote<br />
and<br />
(a, b) p = sign(k)<br />
(a, b) = ∑ p∈a∩b (a, b) p<br />
Since the previous sum does not depend on the choice of the prepresentative, we have a well defined map<br />
( , ) : H 1 × H 1 −→ Z<br />
If S is a Riemann surface of genus g, we have in terms of a basis that<br />
(a i , b j ) = (b i , b j ) = 0, for every i, j,<br />
(a i , b j ) = −(b j , a i ) = δ ij .<br />
28
Then we have that each (a i , b j ) is a skew-linear pairing which is unimodular (det = 1).<br />
a i<br />
b i<br />
<br />
a i 0 1<br />
b i −1 0<br />
<br />
det<br />
1<br />
Hence H 1 (S; Z) = Hom(H 1 (S; Z), Z) =: H 1 (S; Z).<br />
Let ω ∈ M ′ (S) be a closed meromorphic differential. Recall that ω is exact if and only if ω = df for some<br />
meromorphic function f ∈ M(S). Can we choose a function f ∈ M(S) such that f = ∫ z<br />
ω? As an exercise,<br />
p<br />
think if it is possible that Res p (ω) = 0 for every p ∈ S. Consider the period homomorphism<br />
Π ω : H 1 (S; Z) −→ C<br />
∫<br />
[γ] ↦→ ω, γ is a loop.<br />
γ<br />
This map is well defined. For if γ = γ ′ in H 1 then γ − γ ′ = ∂u, and by the Stokes Theorem we have<br />
∫<br />
γ ω = ∫ γ ′ ω. Also, Π ω is a C-linear map.<br />
By the de Rham Theorem, we have an isomorphism HdR 1 (Z) ∼ = H 1 (S; Z) ⊗ C if Z is compact.<br />
isomorphism is given by<br />
[ω] ∈ HdR(Z) 1 ↦→ Π ω<br />
Such an<br />
Corollary 2.10.1. dim C HdR 1 (S) = 2g.<br />
Proof: We have<br />
2 − 2g = χ(S)<br />
= ∑ (−1) i h i dR(S), where h i dR (S) = dim(H1 dR (S)),<br />
= h 0 dR(S) − h 1 dR(S) + h 2 dR(S)<br />
= 1 − dim C HdR(S) 1 + 1.<br />
Theorem 2.10.1. The de Rham ismomorphism carries wedge product of forms defined by<br />
isomorphically to the (cup) product on H 1 (S).<br />
HdR 1 × HdR 1 −→ C<br />
∫<br />
([ω 1 ], [ω 2 ]) ↦→ ω 1 ∧ ω 2<br />
S<br />
29
Recall that Ω de<strong>notes</strong> the space of holomorphic 1-forms on S. Since<br />
H 1 dR<br />
∼ =<br />
−→ H 1 (S; Z) ⊗ C = Hom C (H 1 , C)<br />
if S is compact, we have an inclusion Ω ↩→ HdR 1 if S is compact. We have<br />
ω exact =⇒ ω = df =⇒ f holomorphic =⇒ f constant =⇒ ω = 0<br />
Definition 2.10.3. P g = dim(Ω) (geometric genus).<br />
It is known that P g ≥ g. When the equality holds, it is because of the Hodge Decomposition Theorem.<br />
30
2.11 Questions about (compact) Riemann surfaces<br />
Recall that if f : Z 1 −→ Z 2 is a finite map of degree n of Riemann surfaces, then any meromorphic function<br />
on Z 1 satisfies a polynomial of degree n over the field of Z 2 . Hence<br />
is a field extension of degree ≠ n.<br />
f ∗ M(Z 2 ) ↩→ f ∗ (Z 1 )<br />
Fact 2.11.1. The equality deg = n implies that all finite extensions of M(Z 2 ) are in natural bijective correspondence<br />
with finite maps up to isomorphisms, and each extension is Galois if and only if so is the finite map.<br />
Example 2.11.1. If Z 2 = P 1 , M(P 1 ) = C(Z) then there is a bijective correspondence between fields of<br />
transcendence degree = 1 and finite covering of P 1 .<br />
The equality deg = n implies that Aut(Z) = C-Aut of M(Z).<br />
Question: We know that 2P g ≥ 2g. When are they equal?<br />
If yes, then a loop C is homologous to 0 if and only if ∫ C ω = 0 for every ω ∈ Ω. Also, H1 dR<br />
= Ω ⊕ Ω. The<br />
values ∫ a i<br />
ω and ∫ b i<br />
ω are called periods.<br />
31
2.12 Harmonic differentials and Hodge decompositions<br />
Recall that A 1 = C ∞ -valued 1-forms.<br />
Definition 2.12.1. Notice that locally every C ∞ -valued 1-form can be written as ω = fdz + gdz, where<br />
fdz ∈ A 1,0 and gdz ∈ A 0,1 . There exists a C-linear map ∗ : A 1 −→ A 1 called the star operator, locally<br />
defined by<br />
∗(pdx + qdy) = −qdx + pdy<br />
Or equivalently,<br />
∗(fdz + gdz) = i(−fdz + gdz).<br />
Every ω is uniquely a sum of ω 1,0 ∈ A 1,0 and ω 0,1 ∈ A 0,1 , and<br />
∗(ω) = i(−ω 1,0 + ω 0,1 ).<br />
Definition 2.12.2. ω ∈ A 1 is harmonic if dω = 0 = d(∗ω). An 1-form ω is called coclosed if d(∗ω) = 0.<br />
In other words, ω is harmonic if it is closed and coclosed.<br />
Locally, ω = fdz + gdz is closed if g z = f z , and is coclosed if g z = −f z . Then if ω is harmonic we have<br />
g z = f z = 0, i.e, ω = fdz + gdz ∈ Ω ⊕ Ω, where fdz is a holomorphic 1-form and gdz is an anti-holomorphic<br />
1-form.<br />
Proposition 2.12.1. Let H 1 be the space of harmonic differentials on Z. Then H 1 = Ω ⊕ Ω.<br />
The Hodge Theorem states that H 1 = H 1 dR .<br />
Example 2.12.1. Ω is nonempty for C/(Z ⊕ Z).<br />
We have a positive answer to all questions we established.<br />
Theorem 2.12.1 (Hodge Decomposition). H 1 dR (Z) = H1 = Ω ⊕ Ω, where the first equality is known as the<br />
Hodge Theorem.<br />
Theorem 2.12.2 (Riemann Existence Theorem). Let Z be a local coordinate around p ∈ Z, and n ≥ 1.<br />
Then there exists an exact harmonic differential ω on Z − {p} such that<br />
( ) 1<br />
ω − d<br />
z n = ω + n dz<br />
zn+1 is harmonic on a neighbourhood U of p, and ω ∈ B ′ S−U , i.e., ∫ S−U ω ∧ ω < ∞.<br />
32
Corollary 2.12.1.<br />
(1) There exist meromorphic differentials on Z with any preassigned finite set of poles p and any principal<br />
parts<br />
∞∑<br />
ω p = a i z i dz, when n ≥ 2.<br />
i=−n<br />
(2) There exist a meromorphic function with any prescribed value at a finite set of points.<br />
(3) f ∗ M 2 ↩→ M 1 has degree = deg(f) for a finite map f : Z 1 −→ Z 2 .<br />
33
2.13 Analysis on the Hilberts space of differentials<br />
There exists a Hermitian inner product for 1-forms ω 1 and ω 2 (at least one which is compactly supported)<br />
(ω 1 , ω 2 ) = ∫ Z ω 1 ∧ ∗ω 2 < ∞ .<br />
Locally, ω i = p i dx + g i dy, with i = 1, 2. Then<br />
ω 1 ∧ ∗ω 2 = (p 1 p 2 + g 1 g 2 )dx ∧ dy.<br />
Hence we can define<br />
∫<br />
||ω|| 2 L = ω ∧ ∗ω < ∞.<br />
2<br />
Definition 2.13.1. Let B ′ be the space of bounded 1-forms ω such that ||ω|| 2 < ∞.<br />
Z<br />
With respect to the L 2 -norm, B 1 is a Hilbert space.<br />
Let E be the closure in B 1 of dA 0 C , where A0 C is the space of C∞ -functions with compact support.<br />
Theorem 2.13.1 (Orthogonal decomposition). Let ω ∈ B 1 . Then there exists a unique orthogonal decomposition<br />
ω = ω h + df + ∗dg<br />
where ω h is bounded harmonic and f, g ∈ A 0 , and df, dg ∈ E.<br />
Proof: The essential point is that the space H is orthogonal to both E and ∗E, and E ⊥ ∗E.<br />
ψ, ϕ ∈ A 0 then<br />
∫<br />
∫ ∫<br />
〈 〉<br />
dϕ, ∗dψ = − dϕ ∧ dψ = ψddϕ + d(ψdϕ)<br />
Z<br />
Z<br />
Z<br />
= 0 + 0.<br />
If<br />
Similarly, saying that ω is closed means that it is orthogonal to ∗E, and coclosed means that it is<br />
orthogonal to E. For example,<br />
∫<br />
∫<br />
0 = 〈d ∗ ω, ϕ〉 = d ∗ ω ∧ ϕ = dϕ ∧ ∗ω = 〈dϕ, ω〉<br />
= 〈ω, dϕ〉 .<br />
Hence<br />
H ⊕ ⊥ E ⊕ ⊥ ∗E ↩→ B 1 .<br />
To show the equality, we go to the L 2 -completion of B ′ first.<br />
Theorem 2.13.2 (Regularity). H = (Ě ⊕ ∗Ě)⊥ in ˇB 1 where ( ∨ ) means L 2 -completion.<br />
34
Corollary 2.13.1 (Hodge decomposition). HdR 1 (Z) = Ω ⊕ Ω.<br />
Proof: Saying that a form is closed means that it is orthogonal to ∗E. Hence the previous theorem<br />
implies that a closed 1-form ω is uniquely ω h + df, i.e., every element of HdR 1 has a unique harmonic<br />
representative.<br />
35
2.14 Review<br />
Theorem 2.14.1 (Orthogonal decomposition). For every ω ∈ A 1 = space of differntial 1-forms, there exists<br />
a unique decomposition<br />
ω = ω h + df + ∗dg<br />
where ω h is a bounded harmonic, f, g ∈ A 0 = space of smooth functions.<br />
We denote A = the space of C ∞ -functions.<br />
Proof: Let H be the space of harmonic differentials. Then H is orthogonal to both E = dA 0 and<br />
∗E = ∗dA 0 . This fact follows easily using the L 2 -inner product and the equality ∗∗ = (−1) k , for<br />
Riemann surfaces one has (−1) k = 1. We have<br />
H⊥E⊥ ∗ E ⊆ A 1 .<br />
Taking completion, we have<br />
Ȟ⊥○Ě ⊕ ∗Ě = A1<br />
where Ȟ⊥○Ě is the orthogonal (Ȟ ⊥ Ě) direct sum of Ȟ and Ě. By the Weyl’s Lemma, we have H = Ȟ.<br />
Lemma 2.14.1. Any distribution (1-form) T (1-current) with ∆T = 0 is the distribution of some differential<br />
function f, i.e., T = T f where<br />
∫<br />
T f [h] = hf<br />
and h is compactly supported in U ⊂⊂ Z, where Z is a Riemann surface.<br />
U<br />
Corollary 2.14.1 (Hodge decomposition). HdR 1 (Z) = Ω ⊕ Ω = H.<br />
36
2.15 Proof of Weyl’s Lemma<br />
Claim 2.15.1. If f ∈ A 0 (C) then there exists ψ ∈ A 0 (C) such that ∆ψ = f.<br />
Definition 2.15.1. A 0 c(U) = {f ∈ C ∞ (U) / f has compact support in U}.<br />
Note that A 0 c(U) is a topological space of uniform convergence (not complete).<br />
A distribution on U is a continuous linear functional T : A 0 c(U) −→ C.<br />
Example 2.15.1. Let h ∈ A 0 c. Define<br />
∫<br />
T h [f] =<br />
Then T is a distribution. Using integration by parts, for h ∈ A 0 c and f ∈ A 0 c, we have<br />
∫<br />
∫<br />
hD α f = (−1) |α| fD α h,<br />
where α = (α 1 , α 2 ), D α = ∂(α 1 +α 2 )<br />
∂x α 1 ∂y α 2<br />
U<br />
hf<br />
and |α| = α 1 + α 2 . In other words, we have<br />
T D α h[f] = (−1) |α| T h [D α f]<br />
Definition 2.15.2. (D α T )[f] = (−1) |α| T [D α f], for every f ∈ A 0 c.<br />
Z<br />
The definition for D α T h is the same as T D α h. Let g ∈ A 0 (U × I) where I is an interval, supp(g) ⊆ K × I<br />
and K ⊂⊂ U.<br />
Let f ɛ (z) = g(z,t+ɛ)−g(z,t)<br />
ɛ<br />
. Then f ɛ −→ ∂g<br />
∂t<br />
So:<br />
Claim 2.15.2.<br />
[ ]<br />
d<br />
dt T z[g(z, t)] = T ∂g(z,t)<br />
z ∂t<br />
.<br />
as ɛ −→ 0. By continuity of T , we get<br />
]<br />
[<br />
d<br />
dt T z(g(z, t)) ←− ɛ→0<br />
T [f ɛ ] −→ ɛ→0 ∂g(z, t)<br />
T z<br />
∂t<br />
Similarly, if U and V are open in C and K ⊂⊂ U, L ⊂⊂ V , g ∈ A 0 (U × V ) are such that supp(g) ⊆ K × L,<br />
then<br />
[∫<br />
] ∫<br />
T z g(z, ɛ)d(Volɛ) = T z (g(z, ɛ))dɛ<br />
V<br />
V<br />
Let ρ ∈ A 0 (C) satisfy supp(ρ) ⊆ D, ρ(z) = ρ 0 (|z|) for every z, and ∫ C ρ = 1. Set ρ ɛ(z) = 1 ɛ<br />
ρ ( )<br />
z<br />
2 ɛ . Then<br />
f ∈ A 0 (U) implies<br />
∫<br />
(ρ ɛ ∗ f)(z) = ρ ɛ (z − ɛ)f(ɛ)dVol(ɛ) ∈ A 0 (U (ɛ) )<br />
where U (ɛ) = {z ∈ U / d(z, ∂U) > ɛ}.<br />
U<br />
37
Claim 2.15.3. For every α ∈ N 2 , we have<br />
D α (ρ ɛ ∗ ρ) = ρ ɛ ∗ (D α f)<br />
This result follows similarly applying a change of variables z ↦→ z + ɛ.<br />
Claim 2.15.4. If z ∈ U (ɛ) and f is harmonic on D(z, ɛ), then<br />
(ρ ɛ ∗ f)(z) = f(z).<br />
Proof:<br />
∫<br />
(ρ ɛ ∗ f)(z) = ρ ɛ (ɛ)f(z + ɛ)dVol(ɛ) =<br />
∫ 2π ∫ ɛ<br />
ρ ɛ (r)f(z + rɛ iθ )rdrdθ = 2πf(z)<br />
D(0,z)<br />
0 0<br />
0<br />
∫ ɛ<br />
ρ ɛ (r)rdr = f(z).<br />
Proof of Weyl’s Lemma: Since ɛ −→ ρ ɛ (ɛ − z) has compact support in U, for every z ∈ U (ɛ) ,<br />
h(z) := T ɛ [ρ ɛ (ɛ − z)]<br />
is defined and belongs to A 0 (U (ɛ) ) by Claim 6.2. By Claim 6.4, it suffices to show that for every<br />
f ∈ A 0 c(U (ɛ) ) we have<br />
∫<br />
T [f] = hf,<br />
U (ɛ)<br />
where h is harmonic since<br />
∆h = T ɛ [∆ z ρ ɛ (ɛ − z)] = ∆T ɛ [ρ(ɛ − z)] = 0.<br />
We want to show that T = T h . We have<br />
Hence it suffices to show<br />
∫<br />
T [ρ ɛ ∗ f] = hf.<br />
U (ɛ)<br />
T [f] = T [ρ ɛ ∗ f].<br />
By Claim 6.1, there exists ψ ∈ A 0 such that ∆ψ = f, where ψ is harmonic on V = C − supp(f). Hence<br />
ψ = ρ ɛ ∗ ψ on V ɛ by Claim 6.4. Therefore, O = ψ − ρ ɛ ∗ ρ ∈ A 0 c(U) satisfies<br />
since ∆T = 0, and T (∆ψ) = 0. Hence<br />
∆O = ∆ψ − ρ ɛ ∗ ∆ψ = f − ρ ∗ f<br />
T [f] = T [ρ ∗ f + ∆O] = T [ρ ɛ ∗ f].<br />
38
2.16 Riemann Extension Theorem and Dirichlet Principle<br />
Theorem 2.16.1 (Riemann Extension Theorem). Let Z be a Riemann surface and p ∈ Z. For n ≥ 1 there<br />
exists a harmonic differntial ω on Z − {p} such that:<br />
(1) ω − d ( 1<br />
z n )<br />
is harmonic on a small neighbourhood of p.<br />
(2) ω ∈ B ′ Z−U , i.e., ω is bounded and smooth outside U, with ||ω||2 L 2 < ∞.<br />
Outline of the proof: Let ρ(z) be a differentiable function on Z such that ρ ≡ 0 outside U and ρ ≡ 1<br />
on a neighbourhood of p. The form ( ) ρ(z)<br />
ψ = d<br />
z n ∈ M(p).<br />
We take p = 0. Now ψ − i ∗ (ψ) is smooth and has compact support on Z (≡ 0 on a neighbourhood of 0<br />
and outside U), where ∗(ψ) = i(udz − vdz) if ψ = udz + vdz. Hence by the Decomposition Theorem we<br />
have<br />
ψ − i ∗ (ψ) = ω h + df + ∗(dg)<br />
and<br />
ω = ψ − df = ω h + i ∗ (ψ) + ∗(dg)<br />
It follows that ω is harmonic ((d ∗ d)ω = 0) since ψ and df are exact.<br />
The uniqueness of ω can be guaranteed by adding<br />
(3) (ω, dh) = 0 for every dh ∈ A 1 such that ||dh|| < ∞ and dh ≡ 0 on a neighbourhood of p.<br />
This condition is the same as the following: Since dh ≡ 0 on a neighbourhood N of p, we have<br />
||ω + dh|| 2 Z−N = ||ω|| 2 + (dh, dh) + (ω, dh) + (ω, dh)<br />
Z−N<br />
= ||ω|| 2 Z−N + ||dh|| 2 Z−N<br />
≥ ||ω|| 2 Z−N .<br />
The Dirichlet Principle states that harmonics ω minimizing || || 2 is given by the class of all differentials<br />
Z−N<br />
ω + dh such that dω = 0 on N. So, uniqueness is an easy consequence of this.<br />
39
2.17 Projective model<br />
Theorem 2.17.1. Any compact Riemann surface can be embedded in a projective space.<br />
Proof: f = (1, f 1 , f 2 , . . . , f n ) : Z −→ CP n if f(P ) = f(Q) for P ≠ Q. Then just adding a meromorphic<br />
function that separates P and Q. This process must terminate by compactness.<br />
Theorem 2.17.2 (Chow). The image of f is a projective algebraic variety, i.e., it is V (P 1 , . . . , P L ), for<br />
some polynomials P 1 , . . . , P l on P n .<br />
( )<br />
( )<br />
Proof: Case n = 2: Consider f : Z −→ CP 2 and [z 0 , z 1 , z 2 ] ∈ C 3 . Then a = f ∗ z 0<br />
z 2<br />
and b = f ∗ z 1<br />
z 2<br />
are algebraic ( dependent ) over C, i.e., there exists a polynomial F (Z 1 , Z 2 ) of degree d such that F (a, b) = 0.<br />
Then x d x 0<br />
2F<br />
x 2<br />
, x1<br />
x 2<br />
defines f(Z).<br />
This result is also known as the 1-st GAGA Principle (after Géométrie Algébrique et Géométrie Analytique).<br />
40
2.18 Arithmetic nature<br />
Recall that an (algebraic) number field F is a finite degree (and hence algebraic) field extension of Q.<br />
An arithmetic Riemann surface is a Riemann surface contained in P n defined over a number field (i.e.,<br />
(P 1 , . . . , P n ) = 0 defines a Riemann surface and P 1 , . . . , P n have coefficients in the same number field).<br />
Theorem 2.18.1 (Belgi ’79). A Riemann surface Z is arithmetic if and only if there exists a holomorphic<br />
map Z −→ CP 1 with 3 ramification points.<br />
Theorem 2.18.2 (Mardell and Faltings). A Riemann surface Z defined over a number field has a finite<br />
number of rational points, i.e., if g(Z) ≥ 2.<br />
Classification: π 1 (Z) = 0 if and only if the Riemann Mapping Theorem holds, Z ∼ = CP 1 , C or D.<br />
Corollary 2.18.1. A Riemann surface Z is rational ( ∼ = CP 1 ) if and only if g(Z) = 0.<br />
41
Chapter 3<br />
<strong>COMPLEX</strong> MANIFOLDS<br />
3.1 Complex manifolds and forms<br />
Recall that for a smooth R-manifold M, there is an ideal I(x) for each x ∈ M, given by<br />
I(x) = {f ∈ C ∞ (M) / f(x) = 0} ↩→ C ∞ (M)<br />
The cotangent plane at x ∈ M can be defined as the quotient<br />
T ∨ x (M) := I(x)/I(x) 2<br />
and the tangent plane at x ∈ M is simply the dual space of the cotangent plane T ∨ x (X), i.e.,<br />
T x (M) := (T ∨ x (M)) ∨<br />
Definition 3.1.1. An almost complex structure on an R-differentiable manifold X of dim R = 2n is an<br />
epimorphism J of T X such that J 2 = −1. Or equivalently, it is the structure of a complex vector bundle on<br />
T X.<br />
A complex structure on X induces an almost complex structure on X by setting J = i = √ −1. We obtain a<br />
map J : T X,R −→ T X,R with √ −1 acting on the domain and J acting in the codomain. We have<br />
∂<br />
∂z = 1 ( ∂<br />
− i ∂ ) ( ) ∂ ∂<br />
↦→ ,<br />
2 ∂x i ∂y i ∂x i ∂y i<br />
Locally, J is defined by<br />
( ) ( ∂ ∂ ∂<br />
, ↦→ , − ∂ )<br />
∂x i ∂y i ∂y i ∂x i<br />
corresponding to<br />
the eigen-value i, an an eigen-value T 0,1<br />
X<br />
corresponding to the eigen-value −i, for the operator J. Note that<br />
is naturally isomorphic to T X,R by taking the real part, and this isomorphism identifies i with J. Hence<br />
is generated by vectors of the form u − iJu, with u ∈ T X,R.<br />
Let (X, J) be an almost complex manifold. Then T X,R ⊗ C contains an eigen-bundle T 1,0<br />
X<br />
T 1,0<br />
X<br />
T 1,0<br />
X<br />
Theorem 3.1.1. A complex manifold has a complex structure J on T X,R and its associated subbundle<br />
⊆ T X,R ⊗ C is naturally the same as T X (by taking the real part).<br />
T 1,0<br />
X<br />
43
Similarly, if Ω X,R := T ∨ X,R , then<br />
with the identification<br />
Ω X,R ⊗ C = Ω X ⊕ Ω X<br />
Ω X ←→ Ω 1,0<br />
X ←→ dz i,<br />
Ω X ←→ Ω 0,1<br />
X ←→ dz i.<br />
Definition 3.1.2. An almost complex structure is called integrable if it comes from a complex structure.<br />
The only spheres with an almost complex structure are S 2 and S 6 . The 6-sphere S 6 has an almost complex<br />
structure via the octonions, by taking the multiplication structure in the multiplicative structure in the<br />
sphere in the purely imaginary part of octonions.<br />
Question: Does it exist complex structures on S 2 ?<br />
Theorem 3.1.2 (Newlande - Niremberg). J is integrable if and only if [T 1,0<br />
X , T 1,0<br />
X ] ⊆ T 1,0<br />
X .<br />
Proposition 3.1.1 (Poincaré Lemma). Let α be a closed differential from a differentiable manifold with<br />
deg(α) > 0. Then locally α = dβ, for some form β.<br />
Proposition 3.1.2 (∂-Poincaré Lemma). Let α be a ∂-closed differential from a differentiable manifold<br />
with (p, q) = deg(α) and q > 0. Then locally α = ∂β, for some form β.<br />
Proof: First we show that we can reduce the problem to the case p = 0 and q = 1. In general, we know<br />
α = ∑ α I,J dz I ∧ dz J<br />
∂α = ∑ dα I,J ∧ dz I ∧ dz J = 0<br />
Then α I = ∑ α I,J<br />
∂z J is ∂-closed and of type (0, q), and so α I = ∂β if and only if α = (−1) p ∂( ∑ ∂z I ∧β I ).<br />
Henceforth assume that α is of type (0, q), i.e., α = ∑ α J dz J . Apply the induction on the largest<br />
integer k such that k ∈ J and α J ≠ 0. Necessarily k ≥ q and k = q implies α = fdz 1 ∧ · · · ∧ dz q . In the<br />
latter case ∂α = 0 if and only if f is holomorphic in the variables z i with i > q. We may know apply<br />
the following result:<br />
Proposition: There exists a differentiable function g holomorphic in the variables z i , with i > q, such<br />
that ∂g<br />
∂z q = f and hence α = (−1)q−1 ∂(gdz 1 ∧ · · · ∧ dz q−1 ).<br />
Now assume the ∂-Poincaré Lemma proved for k − 1 > q. Write α = α 1 + α 2 dz k , α 2 = ∑ α 2,J dz J where<br />
|J| = q − 1 and J ⊆ {1, . . . , k − 1}. So ∂ = 0 implies α 2,J is holomorphic in variables z l , with l > k.<br />
Hence by the previous proposition, we have α 2,J = ∂β 2,J<br />
∂z k<br />
, where β 2,J is holomorphic in z l , l > k. Then<br />
∂β = ∂(β 2,J dz J ) = (−1) q−1 α 2 ∧ dz k + α ′ 1<br />
44
where α ′ 1 involves only the coordinate z l for l < k. Thus,<br />
α = α ′′<br />
1 + ∂β<br />
where α ′′<br />
1 for l < k. Since β is holomorphic in the z l for l > k, we have ∂α ′′<br />
1 = 0, q = deg(α ′′<br />
1) < k. We<br />
conclude by induction.<br />
Proof of the previous proposition: We restrict to the case p = 0 and q = 1. Let α = fdz. As<br />
statement is local, we may assume that supp(f) is compact. Define<br />
g = 1 ∫<br />
2πi C<br />
f(z)<br />
dζ ∧ dζ := lim<br />
ζ − z ɛ→0<br />
∫<br />
1<br />
2πi C\D(z,ɛ)<br />
f(z)<br />
dz ∧ dz.<br />
ɛ − z<br />
This limit exists since f is bounded and 1<br />
ζ−z<br />
in integrable on D. We want to show that ∂g = α = fdz.<br />
∫<br />
Now g = g ɛ where g ɛ = 1<br />
f(γ+z)<br />
2πi C\D(0,ɛ) γ<br />
dγ ∧ dγ. Then<br />
( ∫ )<br />
∂g ɛ (z) = 1<br />
dγ ∧ dγ<br />
∂ z f(γ + z) dz<br />
2πi C\D(0,ɛ)<br />
γ<br />
implies<br />
∂<br />
∂z g(z) = ∂ zg(z) = 1<br />
2πi<br />
∫<br />
1<br />
= lim<br />
ɛ→0 2πi<br />
∫<br />
C<br />
C\D(z,ɛ)<br />
∂<br />
∧ dγ<br />
f(γ + z)dγ<br />
∂z γ<br />
∂<br />
∂ζ<br />
f(ζ)dζ<br />
∧ dζ<br />
ζ − z .<br />
( )<br />
On C\D(z, ɛ), we have ∂f dζ∧dζ<br />
(z)<br />
∂ζ ζ−z<br />
= −d ɛ f(ζ) dζ<br />
ζ−z<br />
. By the Stokes Theorem, we get<br />
Hence ∂g = fdz.<br />
∫<br />
1 ∂f dζ ∧ dζ<br />
2πi C\D(z,ɛ) ∂ζ ζ − z = 1 ∫<br />
f(ζ)<br />
−→ f(z) as ɛ −→ 0.<br />
2πi ∂D(z,ɛ) ζ − z<br />
45
3.2 Kähler manifolds<br />
Let V be a complex vector space with J = √ −1 and W = Hom R (V, R). Recall V C := V ⊗ C = V 1,0 ⊕ V 0,1 .<br />
Hence also W C = W 1,0 ⊕ W 0,1 ⊇ W .<br />
Definition 3.2.1. Let W 1,1 = W 1,0 ⊗ W 0,1 ⊆ ∆ 2 W C ⊇ ∆ 2 W ,<br />
W 1,1 = {(1, 1)-forms} = {sesqui-linear forms on V }.<br />
Let W 1,1<br />
R<br />
= W 1,1 ∩ ∆ 2 W = {real (1, 1)-forms} = {real 2-forms of type (1, 1)} = {alternating forms}. A<br />
(1, 1)-form h ∈ W 1,1 is called Hermitian if h(u, v) = h(v, u) for every u, v ∈ V . Let W 1,1<br />
H<br />
be the space of<br />
such forms.<br />
Fact 3.2.1. There exists a bijective correspondence between Hermitian forms and real alternating forms of<br />
type (1, 1) via<br />
W 1,1<br />
1,1<br />
H<br />
∋ h ←→ Im(h) ∈ WR<br />
Proof: Since h(u, v) = h(v, u), we have that Im(h) is alternating on V , i.e., Im(h) ∈ ∆ 2 W . Conversely,<br />
let ω ∈ W 1,1<br />
R<br />
and set<br />
g(u, v) = ω(u, Jv) = −ω(Ju, v) and<br />
h(u, v) = g(u, v) − iω(u, v).<br />
Then g(u, v) = g(v, u) and thus h(u, v) = h(v, u), i.e., h is Hermitian.<br />
Locally, ω = ∑ i<br />
2 a ijdz i ∧ dz j = −Im(h) ∈ Ω 1,1<br />
X<br />
∩ Ω2 X,R , where (a ij) is hermitian.<br />
Definition 3.2.2. ω ∈ W 1,1<br />
R<br />
is positive if the correspondence h is positive definite.<br />
Definition 3.2.3. A positive real (1, 1)-form on an almost complex manifold (X, J) is a C ∞ associated of<br />
a positive real (1, 1)-form on each tangent space T X,x , x ∈ X.<br />
Definition 3.2.4. A Hermitian metric on a complex vector bundle E over a smooth manifold M is an<br />
element h ∈ Γ(E ⊗ E) ∗ . A Hermitian manifold is a complex manifold with a Hermitian metric on its<br />
holomorphic tangent space. Likewise, an almost Hermitian manifold is an almost complex manifold with<br />
a Hermitian metric on its holomorphic tangent space.<br />
Corollary 3.2.1. There exists a bijective correspondence between real (1, 1)-forms ω on a complex manifold<br />
M and Hermitian metrics on M.<br />
Definition 3.2.5. Let h be a Hermitian metric. We shall say that h is Kähler if ω = Im(h) is closed.<br />
46
Corollary 3.2.2. If a symplectic structure ω on a complex manifold is positive of type (1, 1) (i.e., it vanishes<br />
on Ω 2,0 and hence also on Ω 0,2 and its associated h is positive definite), then it is −Im(h) for a Kähler metric.<br />
Corollary 3.2.3. A Hermitian metric can always be written as h = g + iω where g is a Riemannian metric<br />
invariant under J (g(u, v) = g(Ju, Ju)) and ω is a positive (1, 1)-form, g(u, v) = ω(u, Jv).<br />
Definition 3.2.6. A pair (X, ω) formed by a complex manifold X and a positive (1, 1)-form ω is called a<br />
Kähler manifold.<br />
Lemma 3.2.1. dVol h = ωn<br />
n!<br />
for (X, h), h = h ω = g(u, v), where ω n = ω ∧ · · · ∧ ω of type (n, n).<br />
Proof: Let {e i } be an orthonormal basis of T X,x with respect to h. Then {e i , Je i } is a real orthonormal<br />
basis for TX,x R ωn<br />
with respect to g with positive orientation. It suffices to check<br />
n!<br />
= dx 1 ∧dy 1 ∧· · ·∧dx n ∧dy n<br />
where<br />
∑<br />
{dx 1 , dy 1 , . . . , dx n , dy n } is the dual basis to (e i , Je i ). Let dz j = dx j + dy j . Then we have ω x =<br />
j dz j ∧ dz j and<br />
i<br />
2<br />
ω n x<br />
n!<br />
( ) n i ∏<br />
= dz j ∧ dz j at x,<br />
2<br />
i<br />
2 dz j ∧ dz j = dx i ∧ dy j .<br />
j<br />
Corollary 3.2.4. If X (n) is a compact Kähler manifold then [ω k ] ∈ HdR 2k (X) is nonzero for every k < n.<br />
Proof: ω k = dγ implies ω n = d(ω n−k ∧ γ). The last implies 0 ≠ ∫ X ωn = 0, getting a contradiction.<br />
Corollary 3.2.5. Let X (k) be a compact Kähler submanifold M. Then [x] ∈ H 2k (X) is nonzero.<br />
Proof: Clearly h M | T X = h X and i ∗ ω (M,h) = ω (X,hX ), where i : X ↩→ M is the inclusion. If i(X) = ∂Γ,<br />
then by the Stokes Theorem<br />
∫ ∫<br />
Volume X = i ∗ ωM h = dωM h = 0 since dωM h ≡ 0.<br />
X<br />
Γ<br />
47
3.3 Metrics and connections<br />
Let E −→ X be a C ∞ -vector bundle on X, and let A i (E) be the vector space of C ∞ E-valued forms on X.<br />
Definition 3.3.1. A real (complex) connection on E is a real (resp. complex) linear map<br />
∇ : A 0 (E) −→ A 1 (E)<br />
satisfying the Leibniz rule:<br />
∇(fσ) = df ⊗ σ + f∇σ.<br />
For a vector field ψ and σ ∈ A 0 (E) we write<br />
∇ ψ σ = (∇σ)(ψ) ∈ A 0 (E).<br />
In the case where E is a holomorphic vector bundle, we have the operation<br />
∂ E : A 0 (E) −→ A 0,1 (E)<br />
which defines holomorphic sections of E via Ker(∂ E ). It satisfies the ∂-Leibniz rule instead:<br />
∂(fσ) = ∂f ⊗ σ + f∂σ<br />
but it is not a complex connection.<br />
Proposition 3.3.1 (For a Riemannian manifold). If (M, g) is a R-manifold then there exists a unique<br />
connection ∇ on T M called the Levi-Civita connection satisfying:<br />
(1) d(g(ψ 1 , ψ 2 )) = g(ψ 1 , ∇ψ 2 ) + g(∇ψ 1 , ψ 2 ), i.e., g is ∇-invariant.<br />
(2) ∇ ψ1 ψ 2 − ∇ ψ2 ψ 1 = [ψ 1 , ψ 2 ], i.e., g is torsion free of ∇.<br />
Theorem 3.3.1 (and definition). Let E −→ X be a holomorphic vector bundle with a Hermitian metric.<br />
There exists a unique complex connection ∇ on E, called the Chern connection satisfying:<br />
(1) d(h(σ, τ)) = h(∇σ, τ) + h(σ, ∇τ), i.e., ∇ is invariant under (or compatible with) h.<br />
(2) Let ∇ 0,1 be its composition with A 1 (E) −→ A 0,1 (E). Then ∇ 0,1 = ∂ E .<br />
Theorem 3.3.2. The following statements for a complex Hermitian manifold (X, h) are equivalent:<br />
(1) h is Kähler.<br />
(2) J is flat for the Levi-Civita connection.<br />
(3) Chern connection = Levi-Civita connection.<br />
48
3.4 Review<br />
A complex structure on a real manifold M of dimension 2n is an endomorphism J of T M such that J 2 = −1.<br />
If M is complex, normally we take J = √ −1. A real 2-form h is Hermitian if h(u, v) = h(v, u). It is know<br />
that a form h is Hermitian if and only if h is a positive (1, 1)-form.<br />
Theorem 3.4.1. There exists a bijective correspondence between real alternating forms of type (1, 1) and<br />
Hermitian metrics. Such a correspondence is given by<br />
where Im(h) is a symplectic 2-form.<br />
W 1,1<br />
H<br />
∼<br />
−→ W 1,1<br />
R<br />
h ↦→ Im(h)<br />
Theorem 3.4.2. The following conditions are equivalent for a complex Hermitian manifold (X, h):<br />
(i) h is a Kähler metric, i.e., dw h = 0.<br />
(ii) J is flat for the Levi-Civita connection of h.<br />
(iii) The Chern connection of h on T 1,0<br />
M equals the Levi-Civita connection on T R M .<br />
Proof:<br />
• (iii) =⇒ (ii): It is clear because the Chern connection is C-linear by definition.<br />
• (ii) =⇒ (i): Condition (ii) means that the Levi-Civita connection commutes with J. Then<br />
dω(ϕ 1 , ϕ 2 ) = ω(∇ϕ 1 , ϕ 2 ) + ω(ϕ 1 , ∇ϕ 2 ).<br />
Let C ∞ (M) ∋ ϕ[ω(ϕ 1 , ϕ 2 )] = ω(∇ ϕ ϕ 1 , ϕ 2 ) + ω(ϕ 1 , ∇ ϕ ϕ 2 ). Since<br />
dω(ϕ, ϕ 1 , ϕ 2 ) = ϕω(ϕ 1 , ϕ 2 ) − ϕ 1 ω(ϕ, ϕ 2 ) + ϕ 2 ω(ϕ 1 , ϕ) − ω([ϕ, ϕ 1 ], ϕ 2 )<br />
the result follows from [ϕ i , ϕ j ] = ∇ ϕ1 ϕ j − ∇ ϕ2 ϕ 1 .<br />
• (i) =⇒ (iii): The Chern connction equals the Levi-Civita connection for the flat metric ∑ i dz i ∧dz i .<br />
The result follows from the following proposition.<br />
Proposition 3.4.1. If (X, h) is a Kähler manifold and if x ∈ X, then there exists a holomorphic coordinate<br />
(z 1 , . . . , z n ) centred at x such that<br />
( ) ∂ ∂<br />
h ij = h , = Im + O( ∑ |z i | 2 ).<br />
∂z i ∂z j<br />
The converse is also true.<br />
49
3.5 The Fubini Study metric<br />
Let L = {(l, v) ∈ CP n × C n / v ∈ l}. Consider the diagram<br />
i π<br />
L CP n 2<br />
× C n+1 C n+1<br />
π 1 ◦ i<br />
π 1<br />
CP n<br />
The composite map π 2 ◦ i is called the blow up at 0. We have that L is a holomorphic line bundle over CP n<br />
and is denoted by O(−1).<br />
Definition 3.5.1. O CP n(h) := L −k where L −k := (L ∨ ) ⊗k for k > 0, is a holomorphic line bundle over CP 1 .<br />
The standard metric ∑ |z i | 2 on CP n+1 restricts to a Hermitian metric on L. Its curvature (Ricci or Chern<br />
form) is given by<br />
ω = σ ∗ 2<br />
2π ∂∂log|z i| i =<br />
i<br />
2π ∂∂log|σ|2<br />
for any choice of a holomorphic section σ of L over CP n . Therefore, σ ′ = σf, for f ∈ O and so<br />
where log|f| 2 = logf + logf and it is ∂∂-closed.<br />
log(σ ′ ) 2 = log|σ| 2 + log|f| 2 ,<br />
Lemma 3.5.1. ω is a positive (1, 1)-form.<br />
Proof: We prove only the case n = 1. We have<br />
So<br />
ω =<br />
∂log(1 + |z| 2 ) = ∂(1 + |z|2 )<br />
1 + |z| 2 = zdz<br />
1 + |z| 2 .<br />
i [(1 + |z| 2 )dz ∧ dz − zdz ∧ zdz]<br />
2π (1 + |z| 2 ) 2 = i dz ∧ dz<br />
2π (1 + |z| 2 ) 2<br />
and the conclusion follows from the transitivity of SU(n + 1) on T CP n .<br />
Definition 3.5.2. ω is called the Fubini study metric in CP n and is denoted ω F S . It depends on the<br />
choice of coordinates on C n+1 .<br />
50
Similarly for a holomorphic vector bundle E −→ X with a Hermit metric h, one has the line bundle<br />
i<br />
L π −1 ˜π<br />
(E) E<br />
P(E)<br />
π<br />
X<br />
Π<br />
where P(E) = (E\{zero sections})/C ∗ and L is denoted by L = O P(E) (−1). The composition π ◦ i is called<br />
the blow up of E at its zero section. Here π −1 (E) is the fibre product or pullback of π and Π. Let<br />
F = (PE) x∈X be the fibre of π at x and f : F ↩→ PE the inclusion. Then f ∗ c 1 (| | 2 h<br />
) is a positive (1, 1)-form,<br />
where c 1 (| | 2 h ) = i<br />
2π ∂∂log| |2 h . Hence c 1(| | 2 h<br />
) is a (1, 1)-form on P(E) that is positive in the vertical direction<br />
of π of X, where X is Kähler with Kähler form ω X . Hence P(E) is also Kähler.<br />
Definition 3.5.3. O Eϕ (h) := L −k where L −k := (L ∨ ) ⊗k .<br />
Note that given a vector bundle E −→ ϕ<br />
X, then 1 ϕ = ϕ Eϕ (1) = L ∨ = L −1 is a holomorphic line bundle over<br />
P(E).<br />
Consider the compactification E = P(E ⊕ O) ⊇ E, E −→ X and E ⊕ O −→ ϕ<br />
X are vector bundles over X,<br />
and E is open in P(E ⊕ O). We see that the blow up of E (or E) along its zero section lies in P(ϕ −1 (E)) and<br />
hence it is Kähler.<br />
ϕ<br />
51
Chapter 4<br />
SHEAF COHOMOLOGY<br />
4.1 Sheaves<br />
Definition 4.1.1. Let X be a topological space and A an abelian category. A presheaf F on X is a<br />
collection of objects F(U) of objects in A, for each open subset U ⊆ X, and a collection of morphisms<br />
ρ UV : F(U) −→ F(V )<br />
σ ↦→ σ| V = ρ UV (σ)<br />
for each inclusion of open subsets V ↩→ U such that<br />
ρ UV = ρ V W ◦ ρ UV .<br />
The last equality is known as compatibility.<br />
A presheaf F is called a sheaf if it is saturated, i.e., if it satisfies the following condition: Let s i ∈ F(U i )<br />
be a collection of sections such that<br />
s i | Uij = s j | Uij ,<br />
where U ij = U i ∩ U j , then there exists a unique section s ∈ F(∪U i ) such that s| Ui = s i .<br />
Definition 4.1.2. A morphism of (pre)sheaves is a map ϕ : F −→ G which associates to each open<br />
subset U ⊆ X a morphism<br />
such that for every V ⊆ U open<br />
ϕ U : F(U) −→ G(U)<br />
ρ UV ◦ ϕ U = ϕ V ◦ ρ UV . (compatibility)<br />
Example 4.1.1. Sheaves of sections of vector bundles (C ∞ , C h , C ω for real analytic, O, etc).<br />
53
Lemma 4.1.1. If F is a presheaf over X then there exists a unique sheaf F sh along with a morphism<br />
F −→ F sh that factors thorough all morphisms F −→ G to a sheaf G.<br />
F<br />
G<br />
ϕ ∃!<br />
F sh<br />
If X is a topological space, its structure sheaf CX 0 associated to each open subset U is given by the space<br />
of continuous functions on U. If X is an algebraic variety, its structure sheaf O X is given by the space of<br />
regular functions on (Zariski) open subsets. A complex algebraic manifold is also a complex manifold, and<br />
we write O alg<br />
X<br />
and Ohol X to distinguish the structures:<br />
O alg<br />
X<br />
O hol<br />
X<br />
−→ Zariski topology,<br />
−→ ordinary topology.<br />
Definition 4.1.3. Let A be the sheaf of rings over X. A sheaf F is called an A-module if for every open<br />
set U, F(U) is a module over A(U), compatible with the restriction maps.<br />
Remark 4.1.1. All notions from module theory carry over: Hom, Ker, Im, CoKer, direct sums, tensor<br />
products, exact sequences and homology groups, etc.<br />
Definition 4.1.4. The A-module A ⊕n = A is said to be free of rank n ∈ N.<br />
isomorphic to A ⊕n is called locally free of rank n.<br />
A sheaf that is locally<br />
Remark 4.1.2. There exists a bijection between vector bundles of rank n and locally free sheaves of rank<br />
n.<br />
If A is a sheaf of fields, the rank one sheaves (invertible sheaves with respect to A) form a group under tensor<br />
product, called the Picard group Pic OX (X).<br />
Definition 4.1.5. Let (f, f # ) : (X, A) −→ (Y, B) be a morphism of ringed spaces, i.e., f : X −→ Y is a<br />
continuous function, and<br />
f # : B(V ) −→ A(f −1 (V ))<br />
is a morphism for every open subset V compatible with restrictions. The pullback sheaf f ∗ G of a B-module<br />
G is defined as follows: set f (∗) G(U) = lim G(V ) and then set f ∗ G be the sheaf associated to the<br />
−→f(U)↩→V<br />
presheaf f (∗) G ⊗ f (∗) B A.<br />
Example 4.1.2. The sheaf F of holomorphic sections of a holomorphic vector bundles F over X,<br />
i X : {x} ↩→ X<br />
(1) i (∗)<br />
x F =: F x (the stalk of F at x and its elements are germs of sections of F at x).<br />
54
(2) i ∗ xF = F x has finite dimension over C.<br />
Definition 4.1.6. A sheaf of modules M on an algebraic variety (X, O X ) is said to be (quasi)-coherent<br />
if it is locally isomorphic to the cokernel of a morphism of free sheaves (of finite rank).<br />
Easy fact: f ∗ preserves (locally) free sheaves, rank and invertibility. In particular, f ∗ gives a homomorphism<br />
of Picard groups.<br />
Definition 4.1.7. A short sequence of sheaves<br />
0 −→ F −→ G −→ H −→ 0<br />
is said to be exact if and only if it is exact on the level of stalks.<br />
Let (f, f ∗ ) : (X, A) −→ (Y, B) be a morphism of ringed spaces, i.e., f : X −→ Y is a continuous map, A and<br />
B are shaves of rings. For a sheaf of A-modules F on X, the direct image sheaf f ∗ F of F on X is the sheaf<br />
of B-modules on Y given by V ↦→ F(f −1 (V )). Recall that for a sheaf of B-modules G on Y , its pullback is<br />
defined as follows: Set f (∗) G(U) = lim G(V ), and then set (f ∗ G) to be the sheaf associated with the<br />
−→f(U)⊆V<br />
presheaf f (∗) G ⊗ f (∗) B A.<br />
Example 4.1.3. Let F be the sheaf of holomorphic sections of a vector bundle F and i : {x} −→ (X, O).<br />
(1) i (∗) F =: F x is called the stalk of i, and it equals the set of germs of sections of F .<br />
(2) i ∗ F = F x , the fibre of F at x, is a finite dimensional vector space.<br />
Definition 4.1.8. Recall that a sheaf of modules on an algebraic variety (X, O X ) is said to be quasicoherent<br />
(resp. coherent) if it is locally isomorphic to the cokernel of a morphism of free shaves (resp. of<br />
finite rank). By a free sheaf we mean a sheaf O ⊕n , where n is a cardinal number.<br />
X<br />
Fact 4.1.1. f ∗ preserves local freeness invertibility, in particular f ∗ gives a homomorphism of Picard groups,<br />
where<br />
Pic(X) = group of invertible sheaves ∼ = holomorphic line bundles<br />
A short sequence of shaves 0 −→ F −→ G −→ H −→ 0 is said to be exact if it is exact at the level of stalks.<br />
Remark 4.1.3. Ker, CoKer, ⊗ and f ∗ preserve the property of being (quasi-)coherent. However, f ∗ does not.<br />
Example 4.1.4.<br />
(1) f : C −→ {0}. Then f ∗ O alg<br />
C<br />
(2) i : C ∗ −→ C. Then i ∗ O alg<br />
C ∗<br />
= C[z] which is not finite dimensional.<br />
= C[z, z−1 ] is not of finite type over C[z].<br />
55
Construction: Let X be an affine variety over K, and M a module over K[x]. Then<br />
X = zeroes of {f 1 , . . . , f l } over C N .<br />
We have<br />
K[X] := C[z 1 , . . . , z N ]/ 〈f 1 , . . . , f l 〉<br />
Then U ↦→ M ⊗ K[X] O X (U) is an O X -module and this correspondence preserves tensor product, exactness,<br />
etc. Call this O X -module ˜M. In part, ˜M is quasi-coherent (and coherent if M is of finite type) and any<br />
quasi-coherent O X -module is of this form.<br />
Example 4.1.5 (Important). Ideal sheaf I Y<br />
Y ↩→ X (i.e., Y is an algebraic subvariety).<br />
↩→ O X corresponding to the subsheaf of O X vanishing on<br />
We normally assume that the ground field K is algebraically closed. Then the Nullstellensatz tells us that<br />
V (I) := sup(O X /I) ⊆ X is nonempty if and only if I ≠ O X . Hence the ringed space (V (I), O X /I) is<br />
identified with the subscheme of X corresponding to I. A ringed space locally isomorphic to subschemes<br />
of affine spaces is called al algebraic scheme.<br />
Definition 4.1.9 (Associated fibre spaces). Let A be a quasi-coherent sheaf of a O X -algebra of finite type<br />
(i.e., locally generated by finitely many sections as O X -algebras). We define a scheme S = Specm X A and<br />
a morphism π : S −→ (X, O X ) as follows: Let X be affine. Then set<br />
Specm X A := SpecmA(X) := {maximal ideals in A(X)}.<br />
Recall that if R = A(X) and M is a maximal ideal, then R/M = K if K is algebraically closed.<br />
Let f : S −→ K be a regular function. Then {f = 0} c = a basis of open sets, form the Zariski topology.<br />
Let π be the dual to the K-algebra homomorphism K[X] −→ A(X). If D(f) = {x ∈ X / f(x) ≠ 0}<br />
for f ∈ O X (X) then by the quasi-coherence of A, we have A(D(f)) = A(X) ⊗ K[X] K[D(f)]. So that<br />
SpecmA(D(f)) = π −1 (D(f)).<br />
Special case: Given a coherent sheaf of O X -modules F, let A = Sym OX<br />
F. Then the associated fibre space<br />
is called the vector fibre space, denoted by π : V(F) −→ X. Here, Sym means the symmetric product<br />
⊕ k≥0 (Sym k F). Note that<br />
V(F) x = (F x /M x F x ) ∨<br />
Example 4.1.6 (for algebraic geometry).<br />
(I) Definition of normal and tangent bundles (cones):<br />
– The model for tangent vector space at a point is given by Specm(K[ɛ]/ɛ 2 ).<br />
– The Zariski tangent space: Let x ∈ X be a point of al algebraic variety. Then T x X :=<br />
(M x /M 2 x) ∨ .<br />
– The normal bundle of Y ↩→ X (algebraic subvariety): Let I be the ideal sheaf of Y , then<br />
I/I 2 = I ⊗ OX O Y . The normal vector bundle (or normal bundle) of Y in X is I/I . It is denoted<br />
by N Y |X ( −→ π<br />
Y ).<br />
56
– The normal cone of Y in X is defined by<br />
C Y |X := Specm Y<br />
(<br />
⊕k≥0 I k /I k+1) N Y |X<br />
Y<br />
An easiest definition of the tangent bundle to an algebraic variety X is that it is the normal cone to<br />
the diagonal X ↩→ ∆<br />
X ×X. These are functorial objects since f : X −→ Y , f ×f : X ×X −→ Y ×Y ,<br />
so T f : T X −→ T Y . And it coincides with d x f : T x X −→ T f(x) Y , for every x ∈ X.<br />
– We say that x ∈ X is a smooth point if C x (X) = T x X, and X is smooth (non-singular) if all<br />
points are.<br />
(II) The cotangent sheaf to X is defined as the conormal sheaf to the diagonal in X ×X. Its local sections<br />
are local forms on T X and such a form d gives a map<br />
M x /M 2 x<br />
where Ω ′ X = { differential on O X}.<br />
∼<br />
−→ Ω ′ X(x) := Ω ′ X,x/(Ω ′ X,x ⊗ M x )<br />
(III) Blowing up a subscheme: Let I ↩→ O X be an ideal sheaf defining a subscheme Y ↩→ X, A =<br />
⊕ k>0 I k . Then σ : ˜X = proj(A) −→ X is called the blow up of X along Y , where proj(A) =<br />
Specm(homogeneous decomposition of A). By functoriality, σ −1 (Y ) is the projection of the algebra<br />
A ⊗ OX O Y = ⊕ k≥0 I k /I k+1 , i.e., σ −1 (Y ) −→ Y is the projectivization of the normal cone C Y |N , i.e.,<br />
0 = ⊗O X .<br />
Definition 4.1.10. A sheaf is torsion free is ⊗O X = 0, i.e., it is supported on a subvariety.<br />
Fact 4.1.2.<br />
• Any torsion free O X -module F admits a resolution, i.e., a birational morphism σ : Y −→ X such that<br />
σ ∗ F is locally free.<br />
• (Hironaka) Any variety X (any rational map X −→ Y ) admits a resolution of singularities by repeatedly<br />
blow ups along smooth centres (i.e., smooth subvariety)<br />
˜X<br />
σ<br />
X<br />
Y<br />
57
4.2 Cohomology of sheaves<br />
Let X be a topological space. We consider sheaves F i together with morphisms d i : F i −→ F i+1 such that<br />
d i+1 ◦d i = 0 for every i. Such set of sheaves and morphisms (F i , d i ) is called a complex of sheaves over X.<br />
It is also called a resolution of a sheaf F if there exists an inclusion ι : F −→ F 0 such that i(F) = Ker(d 0 )<br />
and Ker(d i+1 ) = Im(d i ), for every i.<br />
(1) Čech resolution: Let F be a sheaf over X, {U i } i∈N a covering of X. For every finite subset I ⊆ N<br />
set U I = ∩ i∈I U i , j I : U I ↩→ X and<br />
Define F k := ⊕ |I|=k+i F I and d : F k −→ F k+1 by<br />
F I = (j I ) ∗ (F| UI ) (extension by zero outside U I )<br />
(dσ) j0...j k+1<br />
= ∑ i<br />
(−1) i σ j0 . . . ĵ i . . . j k+1 | U∩UI<br />
where j 0 ≤ j 1 ≤ · · · ≤ j k+1 , σ = (σ I ), σ I ∈ F I (U) and |I| = k + 1. Lastly, we define ι : F −→ F 0 by<br />
ι(σ) i = σ| U∩Ui<br />
for σ ∈ F(U).<br />
Proposition 4.2.1. This is a resolution.<br />
(2) de Rham resolution: Let A k be the sheaf of C ∞ (R or C-valued) differential forms of degree k (on<br />
a real or complex manifold). The d-Poincaré Lemma says that the complex (A k , d) is a resolution of<br />
Ker(d 0 ) = R (resp. C), constant sheaves over X.<br />
(3) Dolbeault resolution: Let E be a holomorphic vector bundle over a complex manifold X and E its<br />
sheaf of holomorphic sections (i.e., E = O X (E)). Let A 0,q be the sheaf of C ∞ -sections of Ω 0,q<br />
X ⊗ E.<br />
Generalizing the d-Poincaré Lemma, we get that (A 0,q (E), ∂) is a resolution of Ker(∂ 0 ) = E = O X (E)<br />
(a coherent O X -module).<br />
58
4.3 Coherent sheaves<br />
Let O X be the space of functions on X. The “sheaf version”of this space means that to every open subset<br />
U ⊆ X it is associated an space of forms on U.<br />
Example 4.3.1. To an algebraic variety X we associate the space of rational 1, 1-forms on X without poles<br />
on U. To any complex manifold X we associate the space of holomorphic functions on U.<br />
A coherent sheaf is free if it is of the form O ⊕n<br />
X<br />
. An ideal sheaf I is just a subsheaf of O X which is an ideal<br />
of O X (U) (a ring) for every U. Normally assume I ≠ O X . We have a short exact sequence<br />
0 −→ I −→ O X −→ O Z(I) −→ 0,<br />
where Z(I) is a scheme equal to Z(I) = spec(O X /I) X, which is nonempty. Note that O Z(I) and I are<br />
examples of coherent sheaves over X, and O Z(I) is called the torsion part. Any coherent sheaf is locally a<br />
finite direct sum of these factors. If there are no factors of the form O Z(I) , i.e., not supported on a proper<br />
res<br />
subvariety, then it is called torsion free. Any torsion free sheaf admits a resolution f : ˜X −→ X such<br />
that f ∗ of the sheaf is locally free (i.e., a vector bundle). This is because given an ideal sheaf I defining a<br />
subscheme Y ⊆ X, A = ⊕ k>0 I k is an algebra over O X . Then σ : X = proj(A) −→ X (an isomorphism<br />
outside Y ), called the blowup of X along Y , is a birational map to X that replaces Y by a subscheme of<br />
codimension 1, i.e., σ −1 (Y ) is locally given by one equation and σ −1 (Y ) −→ Y is the projectivization of the<br />
normal cone C Y |X .<br />
Given a collection of sheaves F i over X with morphism d i : F i −→ F i+1 such that d i+1 ◦ d i = 0 for every<br />
i. It is a resolution of a sheaf F if there exists an inclusion i : F −→ F 0 such that j(F) = Ker(d 0 ) and<br />
Ker(d i+1 ) = Im(d i ) for every i.<br />
(1) There exists a sheaf X, {U i } i∈N covering of X. For every finite set I ⊆ N, set U I = ∩ i∈I U i , j I : U I ↩→ X,<br />
and F I = (j I ) ∗ (F UI ) extended by zero outside U I . Set F k = ⊕ |I|=k+1 F I and d : F k −→ F k+1 by<br />
(dσ) j0···j k+1<br />
= ∑ i<br />
(−1) i (σ j0···ĵ i···j k+1<br />
)| U∩UI ,<br />
where σ = (σ I ), σ I ∈ F k I (U), and j : F −→ F 0 is given by j(σ) i = σ| Ui∩U for σ ∈ F(U).<br />
Proposition 4.3.1. This is a resolution, where F i | Ui<br />
has trivial cohomology.<br />
(2) de Rham resolution: Let A k be the sheaf of C ∞ (R or C)-valued k-forms. The Poincaré Lemma<br />
says that the complex (A k , d) is a resolution of Ker(d 0 ) = R or C, constant sheaves.<br />
(3) Let E be a holomorphic vector bundle and E its sheaf of holomorphic sections. Let A 0,q (E) be the<br />
sheaf of C ∞ -sections Ω 0,q<br />
X ⊗ E. The ∂-Poincaré Lemma implies that (A0,q (E), ∂) is a resolution of<br />
Ker(∂ 0 ) = E.<br />
59
4.4 Derived functors<br />
Given a functor F : C −→ C ′ between abelian categories such that C has enough injectives.<br />
Definition 4.4.1. For every object M ∈ C, there exists an object R i F (M) for every i ≥ 0 in C ′ , unique up<br />
to isomorphisms, satisfying the following conditions:<br />
(1) R 0 F (M) = F (M),<br />
(2) Every short exact sequence 0 −→ A −→ B −→ C −→ 0 gives rise to a long exact sequence in C ′<br />
0 −→ F (A) −→ F (B) −→ F (C) −→ R 1 F (A) −→ · · ·<br />
· · · −→ R i F (A) −→ R i F (B) −→ R i F (C) −→ R i+1 F (A) −→ · · ·<br />
Remark 4.4.1.<br />
(1) Let M 0 , i : A −→ M 0 be an acyclic resolution of A (i.e., R j+1 F (M k ) = 0 for every j, k ≥ 0). Then<br />
R j F (A) = H j (F (M 0 )) := CoKer(d i−1 F(M i−1 ) −→ F(M i ))<br />
(2) For a sheaf F over X there exists an acyclic resolution (called the Godement resolution for F).<br />
Definition 4.4.2. A fine sheaf F os a sheaf of A-modules where the sheaf of algebra A admits a partition<br />
of unity: for every open covering {U i }, there exists f i ∈ F(U i ) such that ∑ i f i = 1 (this sum is locally finite).<br />
Proposition 4.4.1. H i (X, F) = 0 for every i ≥ 0 for such F.<br />
Corollary 4.4.1.<br />
(1) Let X be a real (complex) manifold, then<br />
H ∗ (X; R) = Ker(d)<br />
Im(d) = H∗ dR(X; R),<br />
where the same equality holds if we replace R by C.<br />
(2) Given a holomorphic vector bundle E −→ X and E its sheaf of holomorphic sections. Then<br />
H ∗ (X, E) = Ker(∂ : A0,q (E) −→ A 0,q+1 (E))<br />
Im(∂ : A 0,q−1 (E) −→ A 0,q (E))<br />
Corollary 4.4.2 (Grothendieck Vanishing Theorem). Given a holomorphic vector bundle E −→ X, we<br />
have H q (X, E) = 0 for q > n = dim C X.<br />
Ker(d) = F −→ F 0 d<br />
−→ F 1 −→ · · · −→ F k −→ · · ·<br />
Consider a Čech resolution<br />
where the map F −→ F 0 is given by σ ↦→ σ| U∩Ui and F k = ⊕ |I|=k+1 (j I ) ∗ (F| UI ).<br />
60
Theorem 4.4.1. If H q>0 (U I , F) = 0 for every I ⊆ N then<br />
H q (X, F) = Ȟq (U, F) := H q (Γ(F) = F(X), d X ),<br />
e.g., if X is a finite dimensional manifold C ∞ , then there exists a good cover (U I contractible for every I)<br />
and so<br />
Ȟ q (U, Z) = H q cell<br />
(X, Z)<br />
where U is in an open cover and H q cell<br />
(X, Z) de<strong>notes</strong> the cellular cohomology, which is computing via nerves<br />
of the open covering.<br />
61
Chapter 5<br />
HARMONIC FORMS<br />
5.1 Harmonic forms on compact manifolds<br />
Let (X, g) be a Riemannian manifold, where X compact is a blanket assumption. Then we have a metric<br />
( , ) on ∆ ∗ TX,x ∨ . We assume X is oriented. Let α, β ∈ Ak : C ∞ (A k TX ∨ ). Then<br />
∫<br />
〈α, β〉 =<br />
X<br />
〈α, β〉 x<br />
dVol(x)<br />
gives an L 2 -metric on A k . We also have a pointwise isomorphism p : ∆ n−k Tx<br />
∨ −→ Hom(∆ k Tx ∨ , ∆ n Tx ∨ ) given<br />
by v ↦→ v ∧ −, where ∆ n Tx<br />
∨ = RdVol(x), and an isomorphism m : ∆ k Tx<br />
∨ ∼<br />
−→ Hom(∆ k T ∨ , R) given by<br />
e ↦→ 〈e, 〉 ∆ k T ∨.<br />
∼<br />
Definition 5.1.1. The Hodge Star Operator is given by<br />
and the associated global isomorphism by<br />
∗ = p −1 ◦ m : ∆ k T ∨ x<br />
∗ : ∆ k T ∨<br />
∼<br />
−→ ∆ n−k T ∨ x<br />
∼<br />
−→ ∆ n−k T ∨<br />
Ω k (X) −→ Ω n−k (X)<br />
We extend ∗ to complex-valued forms by extending 〈 , 〉 to Hermit metrics on ∆ k C (T ∨ ⊗ C) = (∆ k R T ∨ ) ⊗ C.<br />
We get (α, β) x dVol(x) = α x ∧ β x and so<br />
∫<br />
〈α, β〉 = α ∧ ∗β<br />
is the L 2 -metric on A k C = Ak ⊗ C. In the case X is complex, A k C = ⊕ p+q=k ⊕ A p,q , A p,q = C ∞ (Ω p,q<br />
X ).<br />
X<br />
Fact 5.1.1. The Stokes Theorem implies that (α, d ∗ β) = (dα, β) where d ∗ := (−1) k ∗ −1 d∗.<br />
63
In part, if n is even then d ∗ = − ∗ d∗. Similarly,<br />
Fact 5.1.2. ∂ ∗ = − ∗ ∂∗ and ∂ ∗ = − ∗ ∂∗ are formal adjoint of ∂ and ∂ with respect to the L 2 metric in A k C .<br />
Proof: (∂α, β) = ∫ X ∂α ∧ ∗β = − ∫ X (−1)|α| α ∧ ∂ ∗ β = − ∫ X (−1)|α| α ∧ ∗ ∗ −1 ∂ ∗ β = (α, ∂ ∗ β).<br />
More generally, if (E, h) is a Hermitian vector bundle then there exists a C-anti linear isomorphism of vector<br />
bundles given by h : Ω 0,q<br />
X<br />
⊗ E = Ω0,q (E) −→ (Ω 0,q ⊗ E) ∨ ∼ = Ω n,n−q ⊗ E ∨ , where ∆ 2n<br />
X = Ωn,n = RdVol(x). So<br />
it gives am antilinear isomorphism<br />
∗ E : Ω 0,q (E)<br />
∼<br />
−→ Ω n,n−q (E ∨ ) = K X ⊗ Ω 0,n−q (E ∨ )<br />
called the Hodge Star, where K X = Ω n,0 = ∆ n C T X ∨<br />
bundle of a complex manifold X.<br />
(holomorphic line bundle) is called the canonical<br />
Fact 5.1.3. ∂ ∗ X = (−1) q ∗ −1<br />
E ◦∂ K X ⊗E ∨ : A0,q (E) −→ A 0,q−1 (E) is the formal adjoint of ∂ E .<br />
Fact 5.1.4. (d ∗ ) 2 = (∂ ∗ E) 2 = (∂ ∗ ) 2 = 0.<br />
Definition 5.1.2. Let (X, g) be a Riemannian manifold,<br />
∆ := dd ∗ + d ∗ d = (d + d ∗ ) 2<br />
Definition 5.1.3. Let (X, h) be a Hermitian manifold,<br />
∆ ∂ := ∂∂ ∗ + ∂ ∗ ∂ = (∂ + ∂ ∗ ) 2 ,<br />
∆ ∂<br />
:= ∂∂ ∗ + ∂ ∗ ∂ = (∂ + ∂ ∗ ) 2 .<br />
If further E −→ X is a holomorphic vector bundle with a Hermit metric, we write ∆ E for ∆ ∂E<br />
= (∂ E +∂ ∗ E) 2 .<br />
From construction,<br />
〈α, ∆ d α〉 = ||dα|| 2 + ||d ∗ α|| 2<br />
and analogously for the Hermitian case.<br />
Corollary 5.1.1. Ker(∆ d ) = Ker(d) ∩ Ker(d ∗ ).<br />
Definition 5.1.4. An element of Ker(∆ d ) is called harmonic, i.e., it is killed by d and d ∗ .<br />
64
Theorem 5.1.1 (Main Theorem of the <strong>Course</strong>). Let<br />
{ (<br />
⊕k Ω<br />
(F, φ) = ( k X , ∆ d)<br />
)<br />
⊕q Ω 0,q (E), ∆ ∂E<br />
for the Riemannian case,<br />
for the Hermitian case.<br />
where φ : F −→ F is an automorphism, and F = ⊕ k Ω k X or F = ⊕ qΩ 0,q (E).<br />
65
5.2 Some applications of the Main Theorem<br />
Theorem 5.2.1 (Riemannian case). Let H k be the space of harmonic k-forms. Then the map<br />
H k −→ HdR k (X, R (or C)) given by α ↦→ [α], which is well defined since dα = 0, is an isomorphism.<br />
Proof: By the Main Theorem, β ∈ A k can be written as α + ∆γ = α + dd ∗ γ + d ∗ dγ, where α +<br />
dd ∗ α is d-closed. Since β is closed we have d ∗ dγ is closed and hence it belongs to (Im(d ∗ )) ⊥ . So<br />
0 = (dγ, dd ∗ dγ) = ||d ∗ dγ|| 2 . Similarly, we get the analogous theorem for ∆ ∂E<br />
where we can identify<br />
H q (X, O(E) = E) with H 0,q (X, E) via the Dolbeaut isomorphism.<br />
∂<br />
Theorem 5.2.2. Given a holomorphic vector bundle E −→ X, let H 0,q (E) be the space of (∂ E )-holomorphic<br />
forms of type (0, q) with values in E. Then the map H 0,q (E) −→ H q (X, E) given by α ↦→ [α] is an isomorphism.<br />
Corollary 5.2.1. If X be a compact manifold then H q (X, R) is finite dimensional. If X is a compact<br />
complex manifold and E −→ X is a holomorphic vector bundle then H q (X, E) is finite dimensional.<br />
66
5.3 Review<br />
Theorem 5.3.1. Let X be a compact manifold:<br />
{ (⊕<br />
k<br />
(F, P ) =<br />
Ωk X , ∆ )<br />
( d<br />
⊕ )<br />
q Ω0,q (E), ∆ ∂E<br />
if X is Riemannian,<br />
E −→ X is a holomorphic vector bundle<br />
where P is an elliptic or Laplacian-like operator. This implies that<br />
where Ker(P ) is finite dimensional.<br />
C ∞ (F ) = Ker(P )⊥○P (C ∞ (F ))<br />
Corollary 5.3.1.<br />
(1) There is an isomorphism H k ∼<br />
−→ H k dR (X, Ror C) given by α ↦→ [α], where Hk is the finite dimensional<br />
space of harmonic forms.<br />
(2) There is an isomorphism H 0,q ∼<br />
−→ H q (X, E), where H 0,q is the finite dimensional space of (0, q)-forms.<br />
Both isomorphisms depend on the chosen metric.<br />
Note that<br />
HdR k (X, R) ∼= 1<br />
Hk (X, R)<br />
∼= 2 ∼ =<br />
3<br />
Ȟ(H, R)<br />
where ∼ = 1 is given by the de Rham isomorphism, ∼ = 2 and ∼ = 3 are given by the Poincaré Lemma. Recall the<br />
Dolbeaut isomorphism:<br />
(X, E) ∼ = H 0,q (X, ∂ ∂<br />
Ωp E) ∼ = (∗) H p,q (X, Ω p E) =: H p,q (X, E) (finite dimensional),<br />
Ω p E ∼ = O(∆ p TX ∗ ⊗ E) (space of holomorphic p-forms with values in E).<br />
H p,q<br />
where (∗) comes from the isomorphism Ω p E = Ker∂ 0 (A p,0 )(E) −→ A p,1 (E)). Also,<br />
H p,q (X) = H p,q (X, C) = H q (X, Ω p X<br />
) (sheaf of holomorphic functions).<br />
It is known that ∆ = ∂∆ ∂<br />
in every Kähler manifold. Then the following theorem follows:<br />
Theorem 5.3.2. Let X be a Kähler manifold. Then<br />
H k (X, C) =<br />
⊕<br />
H p,q (X).<br />
p+q=k<br />
In other words, if [α] ∈ H k (X, C) with α harmonic, then α = ∑ p+q=k αp,q , where α p,q is the (harmonic)<br />
component of type (p, q).<br />
67
5.4 Heat equation approach<br />
Given an initial distribution of heat f(x) = F (x, 0) (t = 0) on a Riemannian manifold (X, g), then the heat<br />
F (x, t) at time t is governed by (∂ t + ∆ X )F = 0.<br />
Example 5.4.1. F is easily obtained for every t for S 1 , and in general for the torus as follows:<br />
F (0, t) = ∑ a n (t)e inθ .<br />
We have<br />
∂ t + ∆ θ = 0 =⇒ 0 = ∑ (<br />
a<br />
′<br />
n (t) + n 2 a n (t) ) e inθ<br />
=⇒ a n (t) = a n e −n2t , where a n = a n (0),<br />
=⇒ F (θ, t) = ∑ e −n2t a n e inθ .<br />
n≥0<br />
It follows F (θ, t) −→ a 0 = ∫ S 1 f(θ)dθ = Av S 1(f), where the integral is the initial distribution.<br />
Example 5.4.2. Let X = R. Doing the same exercise as above but using Fourier transforms, we get<br />
F (x, t) = √ 1 ∫<br />
∫<br />
e − (x−y)2<br />
4t f(y)dy = e R (x, y, t)f(y)dy.<br />
4πt<br />
R<br />
The function e R (x, y, t) = 1 √<br />
4πt<br />
e − (x−y)2<br />
4t is called the heat kernel. Similarly, e R n(x, y, t) = 1 √<br />
4πt<br />
e − ||x−y||2<br />
4t .<br />
On S 1 , we have e(x, y, t) = ∑ n e−n2t e in(x−y) = ∑ n e−n2t e inx e iny , where e −n2t are the eigenvalues of ∆, and<br />
e inx and e iny are the eigenfunctions. In general, the existence of e X (x, y, t) is difficult to obtain analytically<br />
but trivial on physical grounds.<br />
R<br />
Remark 5.4.1. F (x, t) is smooth for every t > 0, i.e., immediate smoothing by heat flow.<br />
In general, given a form α on (X, g), wish to solve<br />
{<br />
(∂t + ∆)A(t) = 0<br />
(∗)<br />
A(0) = α<br />
where α(t) is a form on X parametrized by t. Uniqueness of A(t) follows from:<br />
Lemma 5.4.1. ||A(t)|| is decreasing (non-strict) for a solution of (∗).<br />
Proof: ∂ t ||A(t)|| 2 = 2 〈∂ t A, A〉 = −2 〈∆A, A〉 = −2 〈 ||dA|| 2 + ||d ∗ A|| 2〉 ≤ 0.<br />
68
Theorem 5.4.1. Let (X, g) be a compact Riemannian manifold. Then there exists K p (x, y, t) ∈ A p x(X),<br />
depending only on (X, g) and p, called the heat kernel of X on p-forms such that<br />
∫<br />
A(t) = K p (·, y, t)α(y)dVol(y)<br />
solves (∗), for every α ∈ A p (X).<br />
X<br />
Let T t (x) = ∫ K(·, y, t)α(y)dy.<br />
X<br />
Theorem 5.4.2. T t satisfies:<br />
(1) T t1+t 2<br />
= T t1 T t2 .<br />
(2) T t is formally self-adjoint.<br />
(3) T t α tends to a C ∞ harmonic form H(α) as t −→ ∞.<br />
(4) G(α) = ∫ ∞<br />
0 (T tα − Hα)dt is well defined and yields the Green operator G, i.e.<br />
G(α) ⊥ (harmonic forms) and α = H(α) + ∆G(α).<br />
Proof:<br />
(1) Holds because A(t 1 + t) solves the heat equation with initial condition A(t 1 ).<br />
(2) ∂ t 〈T t η, T τ ɛ〉 = 〈∂ t T t η, T τ ɛ〉 = − 〈∆T t η, T τ ɛ〉 = − 〈T t η, ∆T τ ɛ〉 = 〈T t η, ∂ t T τ ɛ〉 = ∂ t 〈T t η, T τ ɛ〉. This<br />
implies that 〈 , 〉 is a function of t + τ, so denote 〈 , 〉 by g(t + τ). Therefore,<br />
(3) (1) + (2) =⇒ ∀h > 0,<br />
〈T t η, ɛ〉 = g(t + 0) = g(0 + t) = 〈η, T τ ɛ〉 .<br />
||T t+2h α − T t α|| 2 = ||T t+2h α|| 2 + ||T t α|| 2 − 2 〈T t+2h α, T t α〉<br />
= ||T t+2h α|| 2 − ||T t α|| 2 − 2(||T t+h α|| 2 − ||T t+2h α||||T t α||)<br />
and ||T α α|| 2 converges, and therefore it is decreasing. Hence ||T t+2h α − T t α|| 2 −→ 0. It follows<br />
T t α −→ H(α), for some H(α) ∈ A p (X) L2 , called the harmonic projection. Fix τ > 0, then<br />
T t α = T τ T t−τ α −→ H(α) := T τ H(α) as t −→ ∞. Hence H(α) is C ∞ since T τ is given by a C ∞<br />
kernel. Hence H = lim t→∞ T t is also formally self-adjoint.<br />
(4) ||T t α − Hα|| can be shown to decay rapidly enough so that G(α) = ∫ ∞<br />
0 (T tα − H(α))dt is well<br />
defined. We verify that G is formally the Green operator:<br />
∆G(α) =<br />
and for β harmonic we have<br />
〈Gα, β〉 =<br />
∫ ∞<br />
0<br />
∫ ∞<br />
〈(T t − H)α, β〉 dt =<br />
0<br />
∆T t αdt =<br />
∫ ∞<br />
0<br />
∫ ∞<br />
0<br />
−∂ t T t αdt = α − H(α),<br />
〈α, (T t − H)β〉 dt = 0, since (T t − H)β = 0.<br />
69
Corollary 5.4.1. There exists an orthogonal direct sum decomposition<br />
where Im(∆) = Im(d) + Im(d ∗ ).<br />
A p (X) = H p (X) ⊕ d(A p−1 (X)) = H p (X) ⊕ d ∗ (A p−1 (X))<br />
Corollary 5.4.2. There is an isomorphism H p (X) ∼ = H p dR<br />
(X) given by α ↦→ [α].<br />
Then O ∈ A p (X) =⇒ O = α + dd ∗ γ + d ∗ dγ, where α ∈ H(X). Also,<br />
||d ∗ dγ|| 2 = 〈d ∗ dγ, α〉 = 〈dγ, dα〉 , where dα = 0.<br />
70
5.5 Index Theorem (Heat Equation approach)<br />
Let ∆ p : A p (X) −→ A p (X), λ ∈ R ≥0 . Let E p λ denote the λ-eigenspace for ∆p (finite dimensonal). The<br />
square root √ ∆ = δ is called the Dirac operator.<br />
Lemma 5.5.1. The sequence<br />
is exact for λ > 0.<br />
0 −→ E 0 λ<br />
d<br />
−→ Eλ 1 d<br />
−→ · · · −→ Eλ n −→ 0<br />
Proof: ω ∈ E p λ =⇒ ∆p+1 dω = d∆ p ω = λdω =⇒ dω ∈ E p+1<br />
λ<br />
.<br />
Now ω ∈ E p λ and dω = 0 =⇒ λω = ∆p ω = d ∗ d + dd ∗ ω =⇒ ω = d ( 1<br />
λ d∗ ω ) . Then ∆d ∗ ω = d ∗ ∆ω = λd ∗ ω.<br />
Corollary 5.5.1. ∑ p (−1)p dim(E p λ ) = 0.<br />
Corollary 5.5.2. Let {λ p i } be the spectrum of ∆p , with terms repeated n times if multi = n. Then<br />
where ∑ ′<br />
i<br />
is over i where λ(p) i = 0.<br />
∑<br />
(−1) p tre −t∆p = ∑<br />
p<br />
p<br />
(−1) p e −tλ(p) i<br />
= ∑ p<br />
(−1) p ′∑<br />
i<br />
e −λ(p) i (t)<br />
Note that ∑ ′<br />
i e−λ(p) i<br />
(t) = dim(Ker(∆ p )). Hence<br />
X (X) = ∑ p<br />
= ∑ p<br />
(−1) p dim(Ker(∆ p )) = ∑ (−1) p tre −t∆(p) , where e −t∆(p) = T t ,<br />
p<br />
(−1) ∑ ∫<br />
p e (p) (x, x, t)dVol(x).<br />
i X<br />
Proposition 5.5.1. e(x, x, t) ∼ (4πt) −n/2 ∑ ∞<br />
k=0 u k(x, t)t k , where u k (x, t) is explicitly given in terms of<br />
components of curvatures.<br />
Hence, as t −→ 0, we have<br />
(<br />
X ∼ 1 n/2 ∑<br />
∞ ∫<br />
4πt<br />
k=0<br />
X<br />
)<br />
∞∑<br />
(−1) p tru p k (x, x)dVol(x) t k .<br />
p=0<br />
71
This implies<br />
4π n/2 ∫<br />
∞∑<br />
{<br />
(−1)tru p 0<br />
k (x, x)dVol(x) = X (X)<br />
X p=0<br />
if k ≠ n/2,<br />
k = n/2.<br />
Theorem 5.5.1 (Gauss-Bonet). Let n = dim(X) be even. Then<br />
∫<br />
X (X) = ω,<br />
where ω is given in a local frame by<br />
∑<br />
ω = c n (signσ)(signτ)R σ(1)σ(2)τ(1)τ(2) · · · R σ(n−1)σ(n)τ(n−1)τ(n)<br />
and c n =<br />
(−1)n/2<br />
(8π) n/2 ( . n 2 )!<br />
σ,τ<br />
X<br />
For n = 2, ω = 1<br />
8π (R 1212 −R 1221 −R 2112 −R 2121 )dA = − 1<br />
2π RdA 1212 = K 2π<br />
dA where K is the Gaussian curvature.<br />
72
BIBLIOGRAPHY<br />
[1] Griffiths Phillip; Harris Joseph. Principles of Algebraic Geometry. Pure & Applied Mathematics. John<br />
Wiley and Sons, Inc. New York (1978).<br />
[2] Voisin, Claire. Hodge Theory I.<br />
[3] Voisin, Claire. Hodge Theory II.<br />
[4] Arapura, Danu. Algebraic Geometry over the Complex Numbers.<br />
[5] Rosemberg, Steven. The Laplacian on a Riemannian Manifold.<br />
[6] Yu, Yan-Lin. The Index Theorem and the Heat Equation Method.<br />
73