Class Notes for MATH 366.

by S. W. Drury

Copyright c 20022007, by S. W. Drury.

Contents

1 Complex Number Basics 1.1 Polar Representation . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Complex Multiplication as a Real Linear Mapping . . . . . . . . 2 Analytic Functions of a Complex Variable 2.1 Analytic Functions . . . . . . . . . . . 2.2 The Complex Exponential . . . . . . . 2.3 Manipulation of Power Series . . . . . 2.4 The Complex Logarithm . . . . . . . . 2.5 Complex Derivatives of Power Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 3 5 6 8 9 10 18 19 22 23 24 28 30 34 38 42 49 51 53 54 56 58 65 68

3 A Rapid Review of Multivariable Calculus 3.1 The Little o of the Norm Class . . . . . . . . . . . . . . . 3.2 The Differential . . . . . . . . . . . . . . . . . . . . . . . 3.3 Derivatives, Differentials, Directional and Partial Derivatives 3.4 Complex Derivatives and the CauchyRiemann Equations . 3.5 Exact Equations and Conjugate Harmonics . . . . . . . . .

4 Complex Integration 4.1 Fundamental Theorem of Calculus for Holomorphic Functions . 4.2 The Winding Number . . . . . . . . . . . . . . . . . . . . . . . 4.3 Cauchy Integral Formula . . . . . . . . . . . . . . . . . . . . . 5 Holomorphic Functions Beyond Cauchys Theorem 5.1 Zeros of Holomorphic Functions . . . . . . . . . 5.2 Bounded Entire Functions . . . . . . . . . . . . . 5.3 The Riemann Sphere and Mobius Transformations 5.4 Preservation of Angle . . . . . . . . . . . . . . . 5.5 Moreras Theorem . . . . . . . . . . . . . . . . . i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 The Maximum Principle 6.1 Harmonic Functions Again . . . . . . . . . . 6.2 Digression Poisson Integrals in Rn . . . . . 6.3 Maximum Principles for Harmonic Functions 6.4 The PhragmenLindelof Method . . . . . . .

. . . .

. . . .

. . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

70 70 71 75 78 81 85 89 95 102 104 107 107 108 110 114 116 118 120 123

7 Isolated Singularities and Residues 7.1 Laurent Expansions . . . . . . . . . . . . . . . . . 7.2 Residues and the Residue Theorem . . . . . . . . . 7.3 Method of Residues for Evaluating Denite Integrals 7.4 Singularities in Several Complex Variables . . . . . 8 Variation of the Argument and Rouch s Theorem e 8.1 Meromorphic Functions . . . . . . . . . . . . 8.2 Variation of the Argument . . . . . . . . . . . 8.3 Rouch s Theorem . . . . . . . . . . . . . . . e 8.4 Hurwitzs Theorem . . . . . . . . . . . . . . 9 Conformal Mapping 9.1 Some Standard Conformal Maps 9.2 Montels Theorem . . . . . . . . 9.3 The Riemann Mapping Theorem 9.4 Conformal maps between Annuli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10 Odds and Ends 127 10.1 The Schwarz Reection Principle . . . . . . . . . . . . . . . . . 127 10.2 The Gamma Function . . . . . . . . . . . . . . . . . . . . . . . 128

ii

1
Complex Number Basics

Complex numbers take the form z = a + ib where a and b are real. The symbol i is precisely that a symbol. The set of all complex numbers is denoted C. We call a the real part of z and use the notation a = z and b the imaginary part of z with the notation b = z. If z = b = 0, then we say that z is real and identify it to the real number a. Thus we view the real line R as a subset of C. We dene addition and multiplication according to the following laws (a1 + ib1 ) + (a2 + ib2 ) = (a1 + a2 ) + i(b1 + b2 ), (a1 + ib1 ) (a2 + ib2 ) = (a1 b1 a2 b2 ) + i(a1 b2 + a2 b1 ). It can be shown that addition is commutative and associative. The multiplication is also commutative and associative and the standard distributive laws hold. A particular case of multiplication is multiplication by a real number t (a2 + ib2 ) = (ta2 + itb2 ) from putting a1 = t, b1 = 0 in the multiplication law and now it comes that t1 (a1 + ib1 ) + t2 (a2 + ib2 ) = (t1 a1 + t2 a2 ) + i(t1 b1 + t2 b2 ) and it follows that we can view C with addition and real multiplication as a twodimensional real vector space. Hence the expression the complex plane . So, in this analogy, we are indentifying (a + ib) C (a, b) R2 Following the analogy, complex addition is just vector addition in the plane and multiplication by a real number is scalar multiplication in the plane. 1

The multiplication law is based on the idea that i2 = 1. Sometimes we hear it said that i is the square root of 1. The key point being that there is no real square root of 1. A more sophisticated approach to complex numbers would dene C = R[i]/I the quotient of the ring of polynomials in the indeterminate i with real coefcients by the ideal consisting of all such polynomials that are multiples of i2 + 1. However, as far as we are concerned this is overkill1 ! But in any case, the complex numbers form a commutative ring with identity, the identity being 1 + i0 which we will simply denoted by 1. In fact C is a eld. This means that if a + ib = 0 (translate logically to (a, b) = (0, 0), or equivalently not both a = 0 and b = 0), we have (a + ib) (x + iy) = (x + iy) (a + ib) = 1 where x = a(a2 + b2 )1 and y = b(a2 + b2 )1 . These are both well dened since a2 + b2 > 0 since not both a = 0 and b = 0. Thus every non-zero element of C possesses a reciprocal (multiplicative inverse) in C. Some additional denitions are as follows. The complex conjugate of a complex number z = a + ib is a ib which actually means a + i(b) always assuming that a and b are real. It is denoted by z. Visually, the mapping z z takes a point in the complex plane to its reection in the x-axis. It is a real linear mapping i.e. we have t1 z 1 + t 2 z 2 = t 1 z 1 + t 2 z 2 for t1 and t2 real and furthermore a ring homomorphism i.e. z1 z2 = z1 z2 . Also it is involutary, namely z = z. For a and b real, we dene |a + ib| = a2 + b2 , the Euclidean norm of (a, b) in R2 . This called the modulus or absolute value of the complex number a + ib. It is both a norm, in particular sublinear over R |t1 z1 + t2 z2 | |t1 ||z1 | + |t2 ||z2 |, t1 , t2 R, z1 , z2 C

and satises |z| = 0 z = 0 as well as being multiplicative |z1 z2 | = |z1 ||z2 |. There are many other important identities and inequalities which we list below.
We could also say that C is the splitting eld of the polynomial i2 + 1 over R, but this would be beyond overkill!
1

zz = zz = |z|2 z= z+z 2

1 z = 2 z |z| zz z= 2i |z 1 | = |z|1 |z w|2 = |z|2 2 zw + |w|2 |z w| |z| + |w| | z| |z|

|z| = |z| |z + w|2 = |z|2 + 2 zw + |w|2 |z1 z3 | |z1 z2 | + |z2 z3 | | z| |z|

1.1

Polar Representation

The polar representation of complex numbers corresponds with the representation of points in the plane in polar coordinates. We write z = r cos() + ir sin() and it is clear that r = x2 + y 2 = |z|. The quantity is only dened if z = 0 and then only up to an integer multiple of 2, that is if is a possible value, then so is + 2n where n Z. The quantity is called the argument of z. It is a consequence of the addition laws for cos and sin that (r1 cos(1 ) + ir1 sin(1 ))(r2 cos(2 ) + ir2 sin(2 )) = r1 r2 cos(1 ) cos(2 ) sin(1 ) sin(2 ) + ir1 r2 sin(1 ) cos(2 ) + cos(1 ) sin(2 ) = r1 r2 cos(1 + 2 ) + ir1 r2 sin(1 + 2 ) It follows that to take the product of two complex numbers, the modulus of the product is the product of the moduli, but the argument of the product is the sum of the arguments. 3

To cut down on writing we sometimes see cis() = cos() + i sin() and then we can write a complex number in polar representation as z = r cis(). In fact, we will see later that cis() = exp(i) where the exponential function of a complex argument is dened by its power series. We have cis(1 + 2 ) = cis(1 ) cis(2 ) and also cis(n) = cis()n for R/2Z and n Z. Other useful relations are cis() = cis(), | cis()| = 1.

These identities make it possible to solve easily equations of the type z n = with given and unknown z. Let z = r cis() and = cis() be the corresponding polar representations. Then we get r n = and n = (mod 2). + 2k 1 Thus r = n and = where k is an integer. But the latter sequence is n periodic with period n taking account of the equivalence of arguments that differ by an integer multiple of 2. Thus the solutions are given by = + 2k , n k = 0, 1, 2, . . . , n 1

In particular the nth roots of unity the solutions of the equation z n = 1 are given by z = cis(2k/n) for k = 0, 1, 2, . . . , n 1. Since polynomials play a key role in complex analysis, its worth observing that a polynomial of degree n can have no more then n roots. Let p(z) = z n + p1 z n1 + p2 z n2 + + pn1 z + pn and suppose that is a root, i.e. p() = 0. Then
n

p(z) = p(z) p() =

k=0

pnk (z k k ).

But, for each nonnegative integer k, (z k k ) is divisible by z since z k k = (z )(z k1 + z k2 + + k2 z + k1) 4

and it follows that p(z) is divisible by z . The quotient of this division is a polynomial of degree n 1 and a simple induction argument nishes the proof. We will see later in the course that every monic2 polynomial of degree n with complex coefcients can be factored as
n

p(z) =
k=1

(z k )

for some 1 , . . . , n C. Finally, let us not lose sight of the fact that C is a normed vector space over R and hence a metric space with distance function d(z1 , z2 ) = |z1 z2 |. As a metric space, it inherits all the baggage that goes with metric spaces open subsets, closed subsets, convergent sequences and, where mappings are dened, continuity. In particular, since C is nite dimensional over R we see that C is complete as a metric space. Furthermore, according to the HeineBorel Theorem, every closed bounded subset of C is compact. Actually, in this course we will not use the open covering version of compactness. We will work with sequential compactness. 1.2 Complex Multiplication as a Real Linear Mapping

Consider multiplication by the complex number a + ib. We have (u + iv) = (a + ib) (x + iy) = (ax by) + i(bx + ay) or u = ax by and v = bx + ay which can be viewed as a real linear mapping from R2 to itself u a b x = v b a y We therefore obtain the following observation. A real linear mapping from R 2 to itself can be realised as a complex multiplication if and only if its matrix a11 a21 satises a11 = a22 and a21 = a12 .
2

a12 a22

A polynomial is monic iff its leading coefcient is unity.

2
Analytic Functions of a Complex Variable

Let us consider a function dened by a power series expansion with complex coefcients f (z) = a0 + a1 z + a2 z 2 + a3 z 3 + The following theorem tells us where this series converges. T HEOREM 1 There is a number [0, ] such that the series (2.1) converges if |z| < and does not converge if |z| > . There are two extreme cases. In the case = 0, the series converges only if z = 0. In this case, the series is for all intents and purposes useless. The other extreme case is when = and then the series converges for all complex z. We give two proofs, the rst a simple-minded one and the second uses the root test and actually produces a formula for . The number is called the radius of convergence . First proof. The series always converges for z = 0, so let us dene = sup{|z|; Series (2.1) converges}, with the understanding that = in case the set is unbounded above. Then, by denition, |z| > implies that (2.1) does not converge. It remains to show that |z| < implies that (2.1) converges. In this case, there exists with |z| < || such that a0 + a1 + a2 2 + a3 3 + converges, for otherwise |z| would be an upper bound for the set over which the sup was taken. Therefore a n n 0 as 6 (2.1)

n . So, there is a constant C such that |an n | C for all n Z+ . But, now an z n converges absolutely by comparison with a geometric series n=0 z |an z | |an | n=0 n=1
n n n

C
n=0

<

Note that the proof actually shows that (2.1) converges absolutely for |z| < . Furthermore the proof shows that if 0 < r < , then the series converges uniformly on the set |z| r. Another way of saying this is that the series converges uniformly on the (sequentially) compact subsets of {z; z C, |z| < }. A further consequence is that the function dened by a power series is continuous in {z; z C, |z| < }. Second proof. and does not converge if lim supn |an z n | n > 1, because the proof of the root test shows that the terms do not tend to zero. This gives the formula = lim inf |an | n
n 1

By the root test, series (2.1) converges if lim sup n |an z n | n < 1
1

eventually, or if inf nN |an | n tends properly to as N . So, in the complex case, we get a disk of convergence with convergence holding on the open disk and failing on the interior of its complement. In general, not much can be said about convergence on the boundary of the disk |z| = . Power series allow us to dene many of the elementary functions in the complex setting.

which has to be interpreted by taking |an | n = if an = 0 and = if an = 0

1

exp(z) =
n=0

zn n! ()n z 2n (2n)! z 2n+1 (2n + 1)!

cos(z) =
n=0

sin(z) =
n=0

()n

cosh(z) =
n=0

z 2n (2n)! z 2n+1 (2n + 1)!

sinh(z) =
n=0

It is easy to check that all the power series listed above have an innite radius of convergence and therefore dene functions in the entire complex plane. 2.1 Analytic Functions

We make the following denition. D EFINITION Let be a nonempty open subset of C. Let f : C. Then f is analytic if and only if for each point , there is a power series expansion

n=0

an ()(z )n

(2.2)

with strictly positive radius () such that the sum of the series 2.2 equals f (z) in the disk |z | < ().
These functions are the main subject matter of this course. We will see later that the hypotheses are too strong. It will in fact sufce for f to possess a complex derivative at every point z of in order for f to be analytic in . There is also a concept of analyticity on the real line that is sometimes used. D EFINITION Let U be a nonempty open subset of R. Let f : R or C. Then f is real analytic if and only if for each point U , there is a power series expansion

n=0

an ()(x )n

(2.3)

with strictly positive radius () such that the sum of the series 2.3 equals f (x) in the interval () < x < + ().
Of course, if f is real-valued, then the coefcients an () will be real.

Since the series 2.3 will converge in the disk {x; x C, |x | < ()}, we may let = {z; z C, |z | < ()}
U

and suspect that f will have an extension f to . This is in fact the case, but a little beyond our scope at this point in the course. 2.2 The Complex Exponential

L EMMA 2

For z1 , z2 C we have
exp(z1 + z2 ) = exp(z1 ) exp(z2 )
N

Proof. Let expN (z) =

n=0

zn , then we get n!
N N n n z1 1 z2 2 . n ! n2 ! =0 1

expN (z1 ) expN (z2 ) =

n1 =0 n2

On the other hand

N

expN (z1 + z2 ) =
n=0

(z1 + z2 )n = n!

n=0

1 n n nn C n1 z 1 1 z 2 1 = n!

n1 ,n2 0 n1 +n2 N

n n z1 1 z2 2 n1 ! n2 !

It follows that expN (z1 ) expN (z2 ) expN (z1 + z2 ) = and at this point we put in the absolute values | expN (z1 ) expN (z2 ) expN (z1 + z2 )| |z1 |n1 |z2 |n2 . n1 ! n2 !
n n z1 1 z2 2 n1 ! n2 !

0n1 ,n2 N n1 +n2 >N

0n1 ,n2 N n1 +n2 >N

1 Now if n1 + n2 > N then either n1 > 1 N or n2 > 2 N (or both). Hence we have 2

0n1 ,n2 N n1 +n2 >N

|z1 |n1 |z2 |n2 n1 ! n2 !

1 n n1 > N 2 2

|z1 |n1 |z2 |n2 + n1 ! n2 ! =0

1 n n2 > N 1 2

|z1 |n1 |z2 |n2 n1 ! n2 ! =0 |z2 |n2 n2 !

exp(|z2 |)

1 n1 > N 2

|z1 |n1 + exp(|z1 |) n1 !

1 n2 > N 2

The expression
1 n> N 2

|z|n is a tail sum of the (convergent) series for exp(|z|) and n!

hence tends to zero as N tends to innity. It follows that exp N (z1 ) expN (z2 ) expN (z1 + z2 ) 0 as N . But expN (zj ) converges to exp(zj ) for j = 1, 2 and expN (z1 + z2 ) converges to exp(z1 + z2 ). it follows that exp(z1 + z2 ) = exp(z1 ) exp(z2 ) as required. It is easy to check from the power series denitions that if y is real, then

exp(iy) =
n=0

(iy)n = cos(y) + i sin(y) n!

because i2k = (1)k and i2k+1 = (1)k i. So, if x is also real, we have exp(x + iy) = exp(x) exp(iy) = ex cos(y) + iex sin(y) allowing us to understand for an arbitrary complex number z, what the real and imaginary parts of exp(z) are in terms of the real and imaginary parts of z. 2.3 Manipulation of Power Series

Power series can be manipulated in obvious ways. The standard theorems concerning linear combinations, products, quotients and compositions hold good in the complex case. The proofs are essentially the same as those given in MATH 255. For the sake of the record, we repeat those proofs here, but we will eventually establish the results by other means. In this section we will assume that f (z) = a0 + a1 z + a2 z 2 + a3 z 3 + 10 (2.4)

with radius r and for |z| < r and that g(z) = b0 + b1 z + b2 z 2 + b3 z 3 + with radius s and for |z| < s. We already know from general principles the following result. P ROPOSITION 3 The series (an + bn )z n has radius at least min(r, s) n=0 and it converges to f (z) + g(z) for |z| < min(r, s). It is easy to nd examples where the radius of larger than min(r, s).
n=0 (an

+ bn )z n is strictly

T HEOREM 4 The formal product series cn z n has radius of convergence at n=0 least min(r, s) and it converges to f (z)g(z) for |z| < min(r, s). Explicit formul for cn are given by
n n

cn =
p=0

ap bnp =
q=0

anq bq ,

and furthermore

n=0

|cn |t

p=0

|ap |t

p q=0

|bq |tq

(2.5)

for 0 t < min(r, s).

Proof. Let 0 |z| t < min(r, s) and > 0. We denote by
N

fN (z) =
p=0 N

ap z p

gN (z) =
q=0 N

bq z q

hN (z) =
n=0

cn z n

11

then, a tricky calculation shows that fN (z)gN (z) hN (z) = This gives |fN (z)gN (z) hN (z)|
0p,qN p+q>N 0p,qN

ap bq z p+q

ap bq z p+q =
0p,q p+qN 0p,qN p+q>N

ap bq z p+q .

|ap ||bq |tp+q

and, since p + q > N implies p > N/2 or q > N/2

0pN N/2<qN N

|ap ||bq |tp+q +

N

0qN N/2<pN q

|ap ||bq |tp+q

N N

<

p=0

|ap |t |ap |t

p q= p q=
N 2 N 2

|bq |t + |bq |t +
q

q=0

|bq |t |bq |t

q p= q p=
N 2 N 2

|ap |tp |ap |tp

p=0

q=0

if N is large enough. But, on the other hand, we also have if N is large enough. Therefore, for N large enough |f (z)g(z) fN (z)gN (z)| < |f (z)g(z) hN (z)| < 2 .

Since is an arbitrary positive number this shows that the partial sums (h N ) of the formal product series converge to the product of the sums of the given series. This holds for all z with |z| < t, but since t may approach min(r, s), it holds for all z with |z| < min(r, s). So the radius of convergence of the formal product series is at least min(r, s). To show (2.5) we remark that
N N N

n=0

|cn |tn

0p,q p+qN

|ap ||bq |tp+q

0p,qN

|ap ||bq |tp+q =

p=0

|ap |tp

q=0

|bq |tq

and let N tend to . 12

E XAMPLE The radius of convergence of a product series can exceed the minimum of the individual radii. To see this, take 1+z = 1 + 2z + 2z 2 + 2z 3 + 2z 4 + 1z

and

1z = 1 2z + 2z 2 2z 3 + 2z 4 + 1+z both of which have radius 1. But, not surprisingly, the product series is 1 = 1 + 0z + 0z 2 + 0z 3 + 0z 4 + which has innite radius.

2 cK,n z n of
n=0

C OROLLARY 5

Let K Z+ . The formal K -fold product series

K

an z n has radius at least r and converges to f (z)

n=0

. Furthermore

n=0

|cK,n|t

n=0

|an |t

for 0 t < r.
E XAMPLE Again, the product series may have larger radius of convergence than the original. Consider (1 + 2z) 2 = 1 + z
1

1 2 13 3 135 4 z + z z + 2! 3! 4!

which has radius 1 . However, the square of this series is just 1 + 2z which has 2 innite radius. 2 Now we come to compositions. Usually, the formal composition does not make sense. The formal K-fold product series has constant term C K,0 = aK and 0 so the constant term of g f would be bK aK which is an innite sum. So, 0 K=0 in general, composition of power series is not a formal operation. However, if we suppose that a0 = 0, then it does become a formal operation. In this special case, 13

the series { an z n } starts with the term in z K (or later if a1 = 0). Hence, in n=0 computing the coefcient of z n in g f we need only consider K = 0, 1, . . . , n. So each coefcient of g f is in fact a nite sum. In practice, it is convenient to break up the discussion of general compositions of power series into two separate operations. One of these is the special type of composition with a0 = 0 which is a formal operation and the other is the recentering of power series which is not a formal operation. We will deal with recentering later. T HEOREM 6 Suppose that a0 = 0. Then the formally composed series of g f has strictly positive radius and converges to g(f (z)) for z in some neighbourhood of 0. In most situations, one can say nothing about the radius of convergence, except that it is strictly positive. However, if s is innite, then the formally composed series has radius of convergence at least r. Proof. Let (t) =
n=1

|an |tn .

The series has radius r > 0 and so is continuous at 0. Hence, there is a number > 0 such that 0t< = |(t)| < s. Now we have for |z| < , |f (z)| = |
n=1

an z n | (|z|) < s, so that

K

g f (z) =

bK f (z)
K=0

Note that if s = , then we may take = r if r is nite or to be any positive number (as large as we please) if r = . Now, using the Corollary 5 we get

g f (z) =

bK
K=0 n=K

cK,n z n .

(2.6)

The inner sum could be taken from n = 0 to innity, but c K,n = 0 for 0 n < K. What we would like to do is to interchange the order of summation in (2.6). This would yield
n

g f (z) =

bK cK,n z n .
n=0 K=0

14

and indeed, n bK cK,n is the coefcient of z n in the formal powers series for K=0 gf . To justify this interchange, we must apply the theorem dealing with changing the order of summation (Fubinis Theorem). We need to show

K=0

|bK |

n=K

|cK,n |tn < .

(2.7)

But, according to Corollary 5,

n=K

|cK,n |tn (t)

and (2.7) holds since (t) < s. The radius of convergence of the formally composed series is then at least . In the special case s = , (2.7) holds for = r if r is nite, or for every nite > 0 if r is innite. The radius of convergence of the series is therefore at least r. C OROLLARY 7

Suppose that a0 = 0. Then

1 = d0 + d1 z + d 2 z 2 + d3 z 3 + f (z)

with strictly positive radius. In fact, the coefcients d0 , d1 , . . . can be obtained by successively solving the recurrence relations
1 = a 0 d0 0 = a 1 d0 + a 0 d1 0 = a 2 d0 + a 1 d1 + a 0 d2 0 = a 3 d0 + a 2 d1 + a 1 d2 + a 0 d3

et cetera.
Proof. We can assume without loss of generality that a0 = 1. Now, let us dene h(z) = an z n and g(w) = (1 + w)1 . Then, applying Theorem 6, we see n=1 that 1 1 = = (g h)(z) f (z) 1 + h(z) 15

has a power series expansion with strictly positive radius. Once we know this, 1 then both f and have expansions with strictly positive radius and f f (z) 1 =1 f (z)

so the Product Theorem 2.5 allows us to conclude that the coefcients are in fact obtained by formal multiplication, leading to the recurrence relations cited above. Finally we deal with recentering power series. This is not a formal power series operation. We will start with a power series centered at 0, namely a0 + a 1 z + a 2 z 2 + Let us suppose that it has radius r > 0. Now let || < r. Then we wish to expand the same gadget about z = b0 + b1 (z ) + b2 (z )2 + b3 (z )3 + Substituting w = z and comparing the coefcient of w n in

(2.8)

ak (w + )k =
k=0 n=0

bn w n

(2.9)

we nd the formula

bn =
k=n

Cn ak kn .

(2.10)

So the coefcients of the recentered series are innite sums (as opposed to nite sums) and this is why the recentering operation is not an operation on formal power series. T HEOREM 8 Under the hypotheses given above, the series (2.10) dening b n converges for all n Z+ . The radius of convergence of (2.8) is at least r ||. Finally, the identity (2.9) holds provided that |w| < r ||.

16

Proof. Let || < < r and = ||. Then

k k=n k

n=0

Cn |ak |||

kn

=
k=0

|ak |

k n=0

Cn n ||kn

=
k=0

|ak |( + ||)k <

since the order of summation can be interchanged for series of positive terms and since +|| = < r. In particular it follows that for each xed n, the inner series
k k=n

Cn |ak |||kn converges and hence the series (2.10) converges absolutely for

each n Z+ . The same argument now shows that |bn | n n

n=0 k=n k

n=0

Cn |ak |||kn <

and so (2.8) converges absolutely whenever |w| < r ||. So the radius of convergence of the recentered series is at least r ||. Finally, we use Fubinis theorem to show that (2.9) holds. Effectively, since
k

k=0

|ak |

k n=0

Cn |w|n||kn <

we have

bn w =
n=0 n=0

n k=n k

Cn ak kn

=
k=0

ak
n=0

Cn w n kn

=
k=0

ak (w + )k

by interchanging the order of summation.

17

2.4

The Complex Logarithm

We would like the logarithm to be the inverse function of the exponential. The exponential function never takes the value zero. This is a consequence of exp(z) exp(z) = exp(z z) = exp(0) = 1 so let z C \ {0}. Let z = r cis(). Then the equation exp(w) = z can be solved. If w = u + iv with u and v real, we get eu = r and cis(v) = cis(). Hence u = ln(r) and = v + 2ni where n Z. So, unfortunately there are innitely many solutions. There are three ways of proceeding. The politicians solution to this situation is to do nothing and to assert that the complex logarithm is a multivalued function. This solution is largely unworkable. The engineers solution is based on the observation that if you decide to choose a particular solution, say the one which would have log(1) = 0, then you will run into trouble because as the point z starting at 1 makes an anticlockwise tour of the origin and comes back to 1 and if the logarithm is continuous along this path, then on returning to the origin, the value taken would be 2i. The problem arises from making a circuit of the origin and the engineer sees that the solution is to prevent the making of circuits around the origin. To do this he/she makes a cut from the origin out to innity. The position of the cut is somewhat arbitrary, but usually it is located in the most unobtrusive location, namely along the negative real axis N =] , 0]. Restricted to the set C \ N , the angle may be dened in the range < < and the complex logarithm so obtained (and called the principal branch of the logarithm ) is given by log(r cis()) = ln(r) + i where indeed, < < . The nal solution to this situation (the mathematicians solution) is to assert that it is necessary to dene the complex logarithm on a different space (the universal covering space of C \ {0}) which in this context can be given the structure of a Riemann surface. We do not explore this solution at the moment, but possibly will do so later if time allows. Since we are dealing with power series, we should consider the power series

f (z) =
n=1

(1)(n1)

zn n

which is easily seen to have radius 1. Note that if z is real and 1 < z < 1 then f (z) = ln(1 + z). This series vanishes at z = 0 and so the Composition 18

Theorem for power series applies and tells us that the function z exp(f (z)) has a power series expansion with radius at least 1. If z is real and 1 < z < 1 then exp(f (z)) = 1 + z. Since the coefcients of a real power series are uniquely determined by the function represented by the sum, we see that the relationship exp(f (z)) = 1 + z must continue to hold for all z complex with |z| < 1. So f (z) is a logarithm of 1 + z, but which one? Note that if |z| < 1, then (z) > 0 and it follows that the argument of 1 + z is in the range < < . We establish that 2 2 this is in fact the imaginary part of f (z) by continuity. Explicitly, we observe that y h(x, y) = (f (x + iy)) arctan 1+x takes values in 2Z and is continuous in the open unit disk {(x, y); x 2 + y 2 < 1} in the plane. At (x, y) = (0, 0) we have that h(0, 0) = 0 since f (0) = 0 and arctan(0) = 0. The function h is a continuous function of a connected space (the unit disk) into a discrete space (2Z) and hence is constant. If you have difculty with this you can also apply the Intermediate Value Theorem to obtain the same result. To do this, you take an arbitrary point (x, y) of the unit disk and dene H(t) = h(tx, ty) and continuous function of [0, 1] into 2Z. If H(0) = H(1), then the Intermediate Value Theorem gives a contradiction. 2.5 Complex Derivatives of Power Series

L EMMA 9

We have the identity

n2

(z + h) z nhz Sketch Proof.

n2

n1

=h

2 =0

(n 1)(z + h) z n 2

We have by the summation formula for a geometric series

n2 n 1

(z + h) (tz)
=0

= (tz)

n1 =0

z+h tz

tz(z + h)n1 (tz)n (1 t)z + h

19

Now, differentiating both sides partially with respect to t and then setting t = 1 we get
n2

z
=0

(n 1)(z + h) z n 2 =

hz(z + h)n1 nhz n + z 2 (z + h)n1 z n+1 h2

Multiplying by h2 z 1 and simplifying, now gives the result.

T HEOREM 10

Let the power series f (z) =

n=0

an z n have radius > 0 and

dene a function f in |z| < . Then f has a complex derivative f (z) at every point of |z| < and f (z) = radius .
nan z n1 and the derived power series also has
n=0

Proof. It is clear from the formula for the radius of convergence that the two series have the same radius of convergence . We only need to establish that f (z) = g(z) in |z| < where

g(z) =
n=0

nan z n1 .

If |z| < , then we can nd r such that |z| < r < and we insist that |h| < r|z|, so that |z + h| |z| + |h| < r also. We get from the lemma that f (z + h) f (z) hg(z) h2 =
n=2 n2 n=2

|an | |an |r

=0

(n 1)|z + h| |z|n 2
n2

n2 =0

n=2

(n 1)

1 n(n 1)|an |r n2 = C < 2

Hence we nd

f (z + h) f (z) g(z) C|h| h

and the result follows. 20

C OROLLARY 11

Let the power series f (z) =

n=0

an z n have radius > 0 and

dene a function f in |z| < . Then f has a complex derivatives f (k) (z) of all n! + (k) an z nk and orders k Z at every point of |z| < and f (z) = (n k)! n=k each of these power series also has radius .

C OROLLARY 12

Let the power series f (z) =

n=0

an z n have radius > 0. Then

the coefcients an are uniquely determined by f from the formula

an = 1 (n) f (0) for n Z+ . n!

21

3
A Rapid Review of Multivariable Calculus

In a single variable, differential calculus is seen as the study of limits of quotients of the type f (v) f (v0 ) . v v0 This approach works when the domain of the function f is one-dimensional. D EFINITION Let g : ]a, b[ V where V is a nite-dimensional normed real vector space. Let t ]a, b[. Then the quotient f (s) = (s t)1 (g(s) g(t)) V

is dened for s in ]a, b[ \ {t}. It is not dened at s = t. If

lim f (s)
st

exists, then we say that g is differentiable at t and the value of the limit is denoted g (t) and called the derivative of g at t. It is an element of V .
In several variables this approach no longer works. We need to view the derivative at v0 as a linear map dfv0 such that we have f (v) = f (v0 ) + dfv0 (v v0 ) + error term. Here, the quantity f (v) has been written as the sum of three terms. The term f (v0 ) is the constant term. It does not depend on v. The second term df v0 (v v0 ) is a linear function dfv0 of v v0 . Finally the third term is the error term. The linear map dfv0 is called the differential of f at v0 . The differential is also called 22

the Fr chet derivative . Sometimes we collect together the rst and second terms e as an afne function of v. A function is afne if and only if it is a constant function plus a linear function. This then is the key idea of differential calculus. We attempt to approximate a given function f at a given point v0 by an afne function within an admissible error. Which functions are admissible errors for this purpose? We answer this question in the next section. There are two settings that we can use to describe the theory. We start out using abstract real normed vector spaces. However as soon as one is faced with real problems in nitely many dimensions one is going to introduce coordinates i.e. one selects bases in the vector spaces and works with the coordinate vectors. This leads to the second concrete setting which interprets differentials by Jacobian matrices. 3.1 The Little o of the Norm Class

Let V and W be nite-dimensional real normed vector spaces. D EFINITION Let V be an open set and let v0 . Then a function : W is in the class E,v0 called little o of the norm at v0 iff for all > 0 there exists > 0 such that (v) v v0

It is clear from the denition that if E,v0 then (v0 ) = 0. If we replace the norms on V and W by equivalent norms then it is clear that the class of functions E,v0 does not change. Since all norms on a nite dimensional real vector space are equivalent, we see that the class E ,v0 is completely independent of the norms on V and W . In other words, the class E ,v0 is an invariant of the linear space structure of V and W . The following Lemma is very important for the denition of the differential. It tells us that we can distinguish between a linear function of v v 0 and an admissible error function. L EMMA 13 given by

for all v with v v0 < .

Let V be an open set and let v0 . Let : W be

(v) = (v v0 ) v

where : V W is a linear mapping. Suppose that E,v0 . Then (v) = 0 for all v .
23

Proof. Let u V . Then for all > 0 we have (v0 + tu) tu

for all values of t such that |t| is small enough. Using the specic form of we obtain (tu) tu . Using the linearity and the denition of the norm, this leads to |t| (u) |t| u . Choosing now t small and non-zero, we nd that (u) u .

Since this is true for all > 0 we have (u) = 0. But this holds for all u V and the result follows. The next Proposition is routine and will be used heavily in these notes.

Let V be an open set and let v0 . Then E,v0 is a P ROPOSITION 14 vector space under pointwise addition and scalar multiplication.
We leave the proof to the reader. 3.2 The Differential

In this section, U , V and W are nite-dimensional real normed vector spaces. D EFINITION Let V be an open set and let v0 . Then a function f : W is differentiable at v0 with differential dfv0 (a linear map from V to W ) iff there exists a function : W in the class E,v0 such that f (v) = f (v0 ) + dfv0 (v v0 ) + (v) v . (3.1)

In this situation, the quantity dfv0 is called the differential of f at v0 . It is an immediate consequence of Lemma 13 that if the derivative df v0 exists then it is unique. 24

E XAMPLE If f is a linear mapping from V to W , then it is everywhere differentiable and its derivative is given by dfv0 (v) = f (v). The error term is zero. E XAMPLE If is a bilinear mapping : Ra Rb Rk , then we have = (x0 , y0 ) + (x0 , y y0 ) + (x x0 , y0 ) + (x x0 , y y0 ). (3.2) 2

(x, y) = (x0 + (x x0 ), y0 + (y y0 ))

The rst term in (3.2) is the constant term, the second and third terms are linear. The last term is little o of the norm since (x x0 , y y0 ) Here
op op

x x0

y y0 . 2

stands for the bilinear operator norm.

We use the notation U (v, t) for a t > 0 and v a vector in a nite-dimensional vector space V to designate the open ball centred at v of radius t. In symbols U (v, t) = {w V ; w v < t}

Let V be an open set and let f : W be a function P ROPOSITION 15 differentiable at v0 . Then f is Lipschitz at v0 in the sense that there exists > 0 and 0 < C < such that
f (v) f (v0 ) C v v0

whenever v U (v0 , ). In particular, f is continuous at v0 .

Proof. Using the notation of (3.1), we have dfv0 (v v0 ) dfv0 and for = 1, there exists > 0 such that (v) v v0 . for v U (v0 , ). Combining (3.3) and (3.4) with (3.1) we nd f (v) f (v0 ) ( dfv0 for v U (v0 , ) as required. Proposition 15 has a partial converse. 25
op op

v v0

(3.3)

(3.4)

+ 1) v v0

P ROPOSITION 16 Let V be an open set and let f : W be a function differentiable at v0 . Suppose that there exists > 0 and 0 < C < such that f (v) f (v0 ) C v v0

whenever v U (v0 , ). Then dfv0

Proof. We write

op

C.

f (v) = f (v0 ) + dfv0 (v v0 ) + (v v0 ) where E,v0 . Let > 0. The, there exists 1 with 0 < 1 < such that v , v v0
V

< 1 = (v v0 )
V W

v v0 .

and consequently, for v with v v0 dfv0 (v v0 )

< 1 we nd

(C + ) v v0

Since vv0 is free to roam in a ball centered at 0V , it follows that dfv0 op C + . Finally, since is an arbitrary positive number, we have the desired conclusion. The following technical Lemma will be needed for the Chain Rule. L EMMA 17 Let V be an open set, an open subset of W , and let f : be a function Lipschitz at v0 . Let : U be in E,f (v0 ) . Then the composed function f is in E,v0 . Proof. There exists 1 > 0 and 0 < C < such that f (v) f (v0 ) C v v0 (3.5) whenever v U (v0 , 1 ). Let > 0. Dene 1 = C 1 > 0. Then since is little o of the norm, there exists 2 > 0 such that we have (w)
1

w f (v0 )

provided w and w f (v0 ) < 2 . Now dene = min(1 , C 1 2 ) > 0. Then, using (3.5), v and v v0 < together imply that f (v) f (v0 ) < 2 and hence also (f (v))
1

f (v) f (v0 ) C

v v0 .

Since = C 1 , this completes the proof. 26

T HEOREM 18 (C HAIN R ULE ) Let V be an open set, an open subset of W , let f : be a function differentiable at v0 and let g : U be differentiable at f (v0 ). Then the composed function g f is differentiable at v 0 and d(g f )v0 = dgf (v0 ) dfv0 . Proof. We use the differentiability hypotheses to write f (v) = f (v0 ) + dfv0 (v v0 ) + (v) and g(w) = g(f (v0 )) + dgf (v0 ) (w f (v0 )) + (w) w (3.7) where : W is in the class E,v0 and : U is in the class E,f (v0 ) . Combining (3.6) and (3.7) yields g(f (v)) = g(f (v0 )) + dgf (v0 ) (dfv0 (v v0 ) + (v)) + (f (v)) Using the linearity of dgf (v0 ) we can rewrite this in the form g f (v) = g f (v0 ) + (dgf (v0 ) dfv0 )(v v0 ) + dgf (v0 ) ((v)) + (f (v)),(3.8) for all v . The rst term on the right of (3.8) is constant and the second term is linear because it is the composition of two linear functions. Since E ,v0 is a vector space, it sufces to show that the third and fourth terms on the right of (3.8) are in E,v0 . For dgf (v0 ) ((v)) this is a consequence of the continuity of dgf (v0 ) , and for (f (v)) it is a consequence of Lemma 17. There is no product rule as such in the multivariable calculus, because it is not clear which product one should take. E XAMPLE For the most general case of the product rule, is a bilinear mapping : Ra Rb Rk . Let now be open in V and let x0 . Let f and g be mappings from into Ra and Rb respectively differentiable at x0 . Then let h(x) = (f (x), g(x)) x . Applying the chain rule and using the derivative of found earlier, we nd that h is differentiable at x0 and the derivative is given by dhx0 v = (f (x0 ), dgx0 v) + (dfx0 v, g(x0 )). 2 27 v . v (3.6)

3.3

Derivatives, Differentials, Directional and Partial Derivatives

We have already seen how to dene the derivative of a vector valued function on page 22. How does this denition square with the concept of differential given in the last chapter? Let V be a general normed vector space, g : ]a, b[ V and t a point of ]a, b[. Then, it follows directly from the denitions of derivative and differential that the existence of one of f (t) and dft implies the existence of the other, and dft (1) = f (t). This formula reconciles the fundamental difference between f (t) and dft , namely that f (t) is a vector and dft is a linear transformation. In effect, the existence of the limit f (t) = lim(s t)1 (f (s) f (t))
st

as an element of V , is the same as showing that the quantity f (s) (f (t) + (s t)f (t)) is little o of s t. Thus, dft (s t) = (s t)f (t) or equivalently dft (1) = f (t). For a one-dimensional domain, the concepts of derivative and differential are closely related. We can attempt to understand the case in which the domain is multidimensional by restricting the function to lines. Let us suppose that be an open subset of a normed vector space U and that u0 , u1 U . We can then dene a function g : R U by g(t) = u0 + tu1 . The function g parametrizes a line through u0 . We think of u1 as the direction vector , but this term is a misnomer because the magnitude of u1 will play a role. For |t| small enough, g(t) . Hence, if f : V is a differentiable function, the composition f g will be differentiable in some neighbourhood of 0 and (f g) (0) = d(f g)0 (1) = dfu0 dg0 (1) = dfu0 g (0) = dfu0 (u1 ). (3.9)

since both g and f g are dened on a one-dimensional space. Equation (3.9) allows us to understand what dfu0 (u1 ) means, but unfortunately it cannot be used to dene the differential. D EFINITION The directional derivative Du1 f (u0 ) of the function f at the point u0 in the direction u1 is dened as the value of (f g) (0) if this exists. In symbols Du1 f (u0 ) = lim s1 (f (u0 + su1 ) f (u0 )).
s0

(3.10)

28

Clearly, in case f is differentiable, we can combine (3.9) and (3.10) to obtain dfu0 (u1 ) = Du1 f (u0 ). E XAMPLE Consider the function f : R2 R dened by f (x, y) =
x2 y x2 +y 2

(3.11)

if (x, y) = (0, 0), if (x, y) = (0, 0).

It is easy to check that f is linear on every line passing through the origin (0, 0). Hence the directional derivative D(,) f (0, 0) exists for every direction vector (, ) R2 . In fact, it comes as no surprise that D(,) f (0, 0) =
2 2 + 2

if (, ) = (0, 0), if (, ) = (0, 0).

and this is not a linear function of (, ) and therefore cannot possibly be equal to df(0,0) (, ) which would necessarily have to be linear in (, ). It follows from 2 (3.9) that df(0,0) cannot exist. Let be an open subset of Rm . Faced with a mapping f : Rk , we will typically write this mapping as f (x1 , x2 , . . . , xm ) = (f1 (x1 , x2 , . . . , xm ), f2 (x1 , x2 , . . . , xm ), . . . , fk (x1 , x2 , . . . , xm )) where f1 through fk denote the corresponding coordinate functions. Then, the existence of all the partials fi /xj as i runs over 1 to k is equivalent to the existence of the directional derivative Dej f . In case that f is differentiable at x, the k m Jacobian matrix fi (x) xj ij is precisely the matrix representing the linear transformation df x with respect to the usual bases in Rm and Rk . Symbolically we have f1 (x) f1 (x) f1 (x + ) f1 (x) 1 x1 xm f2 (x) f2 (x) f2 (x + ) f2 (x) x 2 xm = . + 1 . . . + error term, . . .. . . . . . . . . . . fk fk m fk (x + ) fk (x) (x) xm (x) x1 where the error term is little o of {
1 m 2 2 j=1 j } .

29

fi for (1 i xj k, 1 j m) implies the existence and continuity of the differential df x for all x . L EMMA 19

The existence and continuity of all the partials

Sketch proof. The case k = 1, m = 2 is entirely typical and captures the idea of the proof. We have f1 (x1 + 1 , x2 + 2 ) f1 (x1 , x2 ) = f1 (x1 + 1 , x2 + 2 ) f1 (x1 + 1 , x2 ) + f1 (x1 + 1 , x2 ) f1 (x1 , x2 ) f1 f1 (x1 + 1 , x2 + t2 2 )2 + (x1 + t1 1 , x2 )1 x2 x1 by the Mean Value Theorem and where 0 t1 , t2 1, = = f1 f1 (x1 , x2 )2 + (x1 , x2 )1 + o( (1 , 2 ) ) x2 x1

since for example the difference f1 f1 (x1 + 1 , x2 + t2 2 )2 (x1 , x2 )2 x2 x2 is o(|2 |) since f1 f1 (x1 + 1 , x2 + t2 2 ) (x1 , x2 ) x2 x2 f1 tends to zero as (1 , 2 ) (0, 0) by the continuity of . x2 Complex Derivatives and the CauchyRiemann Equations

3.4

Let C be open and let f : C. Using the standard identication of C with R2 , we can write f = u + iv, where u and v are real-valued functions on and we write the variable z as x + iy with x and y real. In this way, we may equate f (z) = u(x, y) + iv(x, y). The equations u v u v = and = x y y x 30 (3.12)

are known as the CauchyRiemann equations. The CauchyRiemann equations are a necessary and sufcient condition that the Jacobian matrix
u x v x u y v y

is the matrix representing a complex multiplication. T HEOREM 20

(i) If the function f possesses a complex derivative at a point, then the Fr chet e derivative exists at that point and is a complex multiplication. e (ii) Conversely, if f has a Fr chet derivative derivative at a point which is a complex multiplication, then f possesses a complex derivative at the point. (iii) If the function f possesses a complex derivative in , then the partials u u v v exist in and satisfy (3.12). , , , x y x y (iv) If the partials
u u v v , , , exist and are continuous in and satisfy x y x y (3.12), then f has a complex derivative in and the derivative is continuous in .

Proof. (i) Since f possesses a complex derivative, say at , then for h complex, we have f ( + h) f () f () h as h 0. Multiplying by h we see that f ( + h) = f () + f ()h + (h) (3.14) (3.13)

where |(h)| is o(|h|). So the Fr chet derivative exists at . Furthermore e the linear mapping h f ()h is a complex multiplication.

31

(ii) Conversely, if f possesses a Fr chet derivative A at , then e f ( + h) = f () + A(h) + (h) where |(h)| is o(|h|). But A is also a complex multiplication, then A is multiplication by a complex number which we will designate f (z). We now have (3.14) and can divide by h to obtain (3.13). (iii) Follows immediately from (i). (iv) Since the partials are all continuous in , it follows from Lemma 19 that f has a Fr chet derivative in and the derivative is continuous in . But, e now we apply (ii) to see that in fact f has a complex derivative at each point of

Next, we observe that if u and v satisfy the CauchyRiemann equations and have more regularity, then they are harmonic. before we can establish this, we need the symmetry of the second derivative. f f 2 f 2f , , and all exist and are continuous in an x y xy yx 2f 2f open subset of the plane, then = in . xy yx L EMMA 21

If f ,

Proof. Let h and k be non-zero real numbers. Then consider f (x + k, y + h) f (x + k, y) f (x, y + h) + f (x, y) = k( f f (, y + h) (, y)) x x (3.15)

from applying the Mean Value Theorem to the function g given by g(t) = f (t, y + h) f (t, y). The point lies between x and x + k. Applying the Mean Value Theorem a second time gives f (x + k, y + h) f (x + k, y) f (x, y + h) + f (x, y) = kh( 32 2f (, )). yx

where lies between y and y + h. An exactly similar argument yields f (x + k, y + h) f (x + k, y) f (x, y + h) + f (x, y) = kh( 2f (1 , 1 )), xy

where 1 lies between x and x + k and 1 lies between y and y + h. For kh = 0 we now get 2f 2f (1 , 1 ) = (, ). xy yx Using the continuity of both second partials at (x, y), it sufces to let k and h tend to zero to conclude that 2f 2f (x, y) = (x, y). xy yx

L EMMA 22 If u and v be C 2 (i.e. possess continuous partial derivatives of all homogenous orders 0, 1 and 2) and if (3.12) holds, then 2u 2u 2v 2v + = 0 and 2 + 2 = 0. x2 y 2 x y Proof. We have u u 2u 2u + 2 = + 2 x y x x y y = v + x y y = 0. v x

A similar argument shows that

2v 2v + = 0. x2 y 2

33

3.5

Exact Equations and Conjugate Harmonics

In the theory of ordinary differential equations (ODE) we meet the concept of an exact equation. This is an equation of the form A(x, y)dx + B(x, y)dy = 0 which satises B F A = . We wish to nd a function F (x, y) such that =A y x x F A B and = B. The exactness condition = is necessary since it ensures y y x 2F 2F = . that xy yx

P ROPOSITION 23 Let A and B be given as above and C 1 in the rectangle x1 < x < x2 , y1 < y < y2 . Then a suitable C 2 primitive F can be found in the same rectangle. Proof. Let us assume without loss of generality that the origin lies in the rectangle. Then we dene
x y

F (x, y) =
t=0

A(t, 0)dt +
s=0

B(x, s)ds

which is effectively a line integral along the path from (0, 0) to (x, 0) along the x-axis and then from (x, 0) to (x, y) parallel to the y-axis. We get F = B(x, y) y directly from the Fundamental Theorem of Calculus. We also get F = A(x, 0) + x
y s=0

B (x, s)ds x A (x, s)ds y

from the Fundamental Theorem of Calculus and by differentiation under

y

= A(x, 0) +
s=0

by the exactness condition = A(x, 0) + (A(x, y) A(x, 0)) = A(x, y) 34

from the Fundamental Theorem of Calculus again. Note that a little extra work shows that F is continuous and hence, since A and B are supposed C 1 , it follows that F is C 2 . C OROLLARY 24 Let u be a C 2 harmonic function dened in an open rectangle with sides parallel to the coordinate axes. Then there is a C 2 harmonic function v dened in the same rectangle satisfying (3.12). Proof. We need to solve u v u v =A= and =B= . x y y x The exactness condition is just the fact that u is harmonic. By the previous result, there is a C 2 solution v. The fact that v is harmonic follows from equating the mixed partials of u. D EFINITION Let be an open subset of C. Then f : C is holomorphic in if it has a complex derivative f (z) at every point z of and the map z f (z) is continuous on . If u is the real part of a holomorphic function and is also C 2 , then the material above shows how to construct the imaginary part v. Such a v is called the conjugate harmonic of u. E XAMPLE Let u = x4 6x2 y 2 + y 4 , then for v we need to solve v v = 12x2 y 4y 3 and = 4x3 12xy 2 x y and we can eyeball that v = 4x3 y 4xy 3 +C where C is a constant of integration. The corresponding holomorphic function is u+iv = (x 4 6x2 y 2 +y 4 )+i(4x3 y 4xy 3 + C) = (x + iy)4 + iC. 2
1 E XAMPLE Let u = 2 ln(x2 + y 2 ). Then u is a nice function away from the origin and somewhat tedious calculations show that u is harmonic away from the origin. We wish to solve v y x v = 2 = 2 and 2 x x +y y x + y2

35

and again, we can nd v = arctan(x1 y) + C in x > 0. A solution can be found locally anywhere in C \ {0}, but not globally in C \ {0}. We get u + iv = log(x + iy) + iC. 2 There is a more sophisticated way of looking at the CauchyRiemann equations that uses the differential operators 1 = z 2 i x y and 1 = z 2 +i x y

There are two ways of understanding these denitions. The rst is to point out that with dz = dx + idy and dz = dx idy, we get f f f f dz + dz = dx + dy = df z z x y so that z and z mimic a coordinate system on C. and x y evaluated at a point (x, y) of the plane form a basis of the tangent space M (x,y) to the plane at that point. The dual space M(x,y) is called the cotangent space and the dual basis consists of dx and dy. Both of these spaces have complexications1 . , form a basis in the complexication of the tangent space and The vectors z z the vectors dz and dz form the corresponding dual basis in the complexication of the cotangent space. The key point is that This is not what is actually happening however. The vectors f 1 = z 2 (u + iv) (u + iv) +i x y = 1 2 u v x y 1 + i 2 v u + x y

so that the equation

1

f = 0 is equivalent to the CauchyRiemann equations. z

For V a real vector space, we may construct its complexication V iV which becomes a complex vector space when a complex scalar multiplication is dened by (x + iy)(u iv) = (xu yv) i(xv + yu). The dimension of V iV as a complex vector space is the same as the dimension of V as a real vector space. Those familiar with tensor products of vector spaces can think of the complexication as C V , the tensor product of real vector spaces, with the complex multiplication dened by extending (z v) = (z) v by linearity.

36

The function f (z) = |z|2 has a complex derivative only at the origin. f = z. Clearly f is innitely differentiable and we check from f (z) = zz that z Thus, f possesses a complex derivative if and only if z = 0. 2 E XAMPLE It is also worth pointing out that 2f 1 2f = = zz zz 4 2f 2f + 2 x2 y = 1 f. 4

A C 2 holomorphic function is necessarily harmonic! Short proof: 2f 2f f + 2 =4 = 4 0 = 0. x2 y z z z We nish this section by showing how Cauchys Theorem can be obtained from Greens Theorem. Greens Theorem can be stated as T HEOREM 25 (G REEN S T HEOREM ) Let be a bounded connected open subset of R2 such that consists of a nite number of piecewise smooth closed curves. Let P, Q be C 1 functions dened on an open subset containing cl(). Then Q P P dx + Qdy = dxdy x y

Applying this yields (u + iv)(dx + idy) =

u v + x y

dxdy + i

u v x y

dxdy

so that if f = u + iv is holomorphic in an open subset containing cl(), then we have f dz = 0.

37

4
Complex Integration

We want to dene complex integrals along curves. In this course we will work with curves that have a piecewise C 1 parametrization. A more general theory, which we do not attempt, deals with rectiable curves. Before we can approach this subject, we need to dene the standard Riemann integral of a complex-valued function. D EFINITION

Let f : [a, b] C be continuous. Then we dene

b b b

f (t)dt =
a a

f (t)dt + i
a

f (t)dt.

It is easy to see that this denition is complex linear we check in turn that it is additive, stable under real scalar multiplication and stable under scalar multiplication by i. We have L EMMA 26
b a b

f (t)dt

|f (t)|dt

(4.1)

Proof. Write r=

b a b

f (t)dt = r where r 0 and || = 1. Then

b b b

f (t)dt =
a a

f (t)dt =
a

f (t)dt =
a

(f (t))dt

38

and so
b a b b b

f (t)dt = r

| (f (t))|dt

|f (t)|dt =

|f (t)|dt

using the inequality (4.1) for real-valued functions. We will say that a function t z(t) from [a, b] to C is a C 1 path if it is oneto-one and continuously differentiable. The underlying set of the path is the direct image z([a, b]). If we have two distinct points z1 and z2 of the underlying set then they arise uniquely from t1 , t2 such that zj = z(tj ) for j = 1, 2 and we can determine whether t1 < t2 or t2 < t1 , i.e. we have a concept of order along the path. Let f : C be continuous. Then we dene the integral along the parametrized path by
b

f (z)dz =
a

f (z(t))

dz (t)dt dt

(4.2)

and initially the integral appears to depend upon the parametrization of the path. The integral on the right is dened because the integrand is continuous (compositions and product of continuous functions are continuous). A Riemann sum for the integral on the right of (4.2) is
n

S=
k=1

f (z(k ))

dz (k ) (tk tk1 ) dt

where a = t0 < t1 < < tn = b and k [tk1 , tk ] for k = 1, 2, . . . , n. There is a related Riemann sum, namely
n

S =
k=1

f (z(k ))(z(tk ) z(tk1 )).

By the Fundamental Theorem of Calculus (take real and imaginary parts)

tk

z(tk ) z(tk1 ) = and so z(tk ) z(tk1 )

tk1

dz (t)dt dt

dz (k ) (tk tk1 ) = dt 39

tk tk1

dz dz (t) (k ) dt dt dt

Leading to z(tk ) z(tk1 ) dz (k ) (tk tk1 ) |tk tk1 | dz (|tk tk1 |) (4.3) dt dt

if the partition has step less than > 0, then this leads to |S S | sup |f (z)||b a| dz ()
z
dt

Letting 0, we have that the Riemann sum S dz f (z(t)) (t)dt. hence also S dt a Since the Riemann sum S can be written in the form
n b

f (z(t))
a

dz (t)dt and dt

S =
k=1

f (k )(zk zk1 )

for node points zk and tag points k related appropriately to the order on the curve, we see that the path integral could have been dened in terms of such Riemann sums and that our denition is independent of the C 1 parametrization used. The denition is easily extended to curves which are obtained by joining a nite number of segments each of which has a C 1 parametrization end to end. As we have dened path integrals (with a one-to-one parametrization), the use of piecewise parametrizations is strictly necessary for path integrals around loops. We also note that the denition (4.2) of the integral along a parametrized path is also perfectly valid for paths which are not one-to-one and indeed it is easy to see from the real change of variables formula that if s t(s) is a C 1 map of [c, d] onto [a, b] with t(c) = a and t(d) = b, then
d

f (z(t(s)))
c

dz(t(s)) ds = ds

f (z(t))
a

dz (t)dt. dt

In some instances (such as in the construction of homotopic parametrized curves), the parametrization may not be one-to-one, but this is of no consequence. The purpose of insisting that parametrization be one-to-one above is merely for the purpose of establishing an intrinsic denition for an ordered curve. 40

E XAMPLE

Let m Z. We have

(z )m dz =

0 if m = 1 2i if m = 1

where is an anticlockwise circle centred at . To see this we set z = + re i where the constant r is the radius of the circle and is the parameter for the curve. Technically, since this is a closed curve, we should integrate over two pieces [0, ] and [, 2], but in practice integrating over [0, 2] will also yield the correct result. Since dz = irei d, we get
2

(z ) dz =

r m eim irei d
0 2 m+1 0

= ir

cos((m + 1)) + i sin((m + 1)) d 2

which yields the stated result. L EMMA 27

Let f : C be continuous. Then we have the estimate

b

f (z)dz sup |f (z)|

z b

dz (t) dt dt

(4.4)

and the interpretation of

length(). Proof. Starting with
a

dz (t) dt is the pathlength of and will be denoted dt

f (z)dz =
a

f (z(t))

dz (t)dt dt dz (t) dt dt

we clearly get
b

f (z)dz

|f (z(t))|

from (4.1). The inequality (4.4) follows. 41

dz dz (t) dt would be (k ) (tk tk1 ). By dt dt a k=1 (4.3), this quantity differs from the Riemann sum n |z(tk ) z(tk1 )| by at k=1 most |b a| dz (), where is the step of the partition. We can dene the pathdt length of as sup n |zk zk1 | where zk are points on written in order k=1 along the curve, the sup being taken over all such nite sets of points. Since, from the extended triangle inequality, the quantity under the sup increases as the b dz (t) dt may be partition gets ner and |b a| dz () 0, the quantity dt dt a interpreted as the pathlength of . A Riemann sum for 4.1 Fundamental Theorem of Calculus for Holomorphic Functions

T HEOREM 28

Let be an open subset of C, let F be holomorphic in . Then

F (z)dz = F () F ()

where is any C 1 path contained in starting at and ending at .

Proof. Choosing a parametrization of the path , we see that we must show
b

F (z(t))
a

dz (t) dt = F () F (). dt

But the chain rule gives

dz d F (z(t)) = F (z(t)) (t). The interpretation of this dt dt in terms of differentials is that dt (F z) is dz(t) F dt z, dt z is the multiplication dz operator from R to C dened by (t) and dz(t) F is the multiplication operator dt from C to C dened by F (z(t)). Thus, it remains to show that
b a

d F (z(t)) dt = F () F (). dt

and now it sufces to take real and imaginary parts and apply the real variable Fundamental Theorem of Calculus. As in the real variable situation, there is a second version of the Fundamental Theorem. This is more subtle however. 42

D EFINITION An open subset of C is star shaped at if and only if for every point z of , then line segment joining to z lies entirely in . T HEOREM 29 Let be an open subset of C which is star shaped at . Let f be holomorphic in with complex derivative f . Let
z

F (z) =

f (w)dw

the integral being taken along the line segment from to z . Then F is holomorphic in and F (z) = f (z) for all z .
Proof. We may parametrize the line segment by t + t(z ) as t runs from 0 to 1. So
1

F (z) =
0

f ( + t(z ))(z )dt

We differentiate under the integral sign (possible since the integrand is a C 1 function of z) to get F = z and F = z =
0 1 1 0 1 1 0

f ( + t(z ))(z ) dt = z f ( + t(z ))(z ) dt z

0 dt = 0
0

f ( + t(z )) + f ( + t(z ))t(z )dt tf ( + t(z )) dt t

1 0

=
0

= tf ( + t(z ))

= f (z)

F F and are both continuous in . So x y the Fr chet derivative of F exists in . Thus F is holomorphic and F (z) = f (z) e in . It will follow from these identities that

43

C OROLLARY 30 Let be an open subset of C which is star shaped at . Let f be holomorphic in with complex derivative f . Let
z

F (z) =

f (w)dw

the integral being taken along any C 1 path from to z lying in . Then F is holomorphic in and F (z) = f (z) for all z .
Proof. First dene F (z) by the integral along the line segment, as in Theorem 29. Then since F is holomorphic, F (z) = f (z), we can apply Theorem 28. Taking account of F () = 0, we get the desired conclusion. This now gives us a version of Cauchys Theorem C OROLLARY 31 Let be an open subset of C which is star shaped at . Let f be holomorphic in . Let be a piecewise C 1 loop passing thru . Then f (z)dz = 0.

Proof. Choose a point 1 on the loop different from . Let 1 be the path from to 1 going forward along the loop and 2 be the path from to 1 going backwards along the loop. According to the previous corollary f (z)dz =
1 2

f (z)dz

and so f (z)dz =
1

f (z)dz

f (z)dz = 0.
2

We can generalize this result to loops that are nicely homotopic. T HEOREM 32 Let be an open subset of C. Let (s, t) z(s, t) be a mapping of the rectangle [0, 1] [a, b] into continuous on the given rectangle. Let z k for k = 1, . . . , n be the restriction of z to strips [0, 1] [tk1 , tk ]. Suppose that zk zk 2 zk 2 zk zk , , , , are all continuous on the strip [0, 1] [tk1 , tk ] on t s st ts 44

which it is dened. Suppose that z(s, a) = z(s, b) for all s [0, 1]. Let (s) be the piecewise continuous loop dened by t z(s, t). Then independent of s and consequently
f (z)dz =
(1) (0) (s)

f (z)dz is

f (z)dz.

Proof. We consider s f (z)dz = s (s)

b b

f (z(s, t))
a

z (s, t)dt t

=
a b

f (z(s, t)) t

z 2z z (s, t) (s, t) + f (z(s, t)) (s, t) dt s t st z (s, t) dt s

=
a

f (z(s, t))

= f (z(s, b)) since z(s, b) = z(s, a) and

z z (s, b) f (z(s, a)) (s, a) = 0 s s

z z (s, b) = (s, a). A detailled proof requires splits s ting up the range of integration [a, b] into its constituent intervals [t k1 , tk ] and applications of differentiation under and the Fundamental Theorem of Calculus on each constituent interval. In case that t z(0, t) is constant, we get f (z)dz = 0.
(1)

Why is this an extension of Cauchys Theorem for star shaped regions? Well, given a region star shaped at and a piecewise C 1 loop t z(t) in , it sufces to consider (s, t) (1 s) + sz(t). The star shaped condition implies that (1 s) + sz(t) for all s with 0 s 1 and all t and we observe that for s = 0, (1 s) + sz(t) = and for s = 1, (1 s) + sz(t) = z(t). It is awkward to deal with C 1 homotopies. Usually homotopies are required only to be continuous. We settle this with the following technical theorem. 45

T HEOREM 33 Let (s, t) z(s, t) be a continuous map from [0, 1] [a, b] to the open subset of C such that z(s, a) = z(s, b) for all s [0, 1]. Suppose that t z(0, t) and t z(1, t) are piecewise C 1 loops 0 and 1 . Let f be holomorphic in . Then f (z)dz =
(1) (0)

f (z)dz.

Proof. Without loss of generality, we can always assume that a = 0 and b = 1. The ambient space of the t variable is then the interval [0, 1] with the points 0 and 1 identied, in other words, the circle. We realize the circle in the form R/Z and treat it as a quotient group. The group operations + and that appear in the integration formul below are to be taken in this group, i.e. modulo 1. Since the image of z is compact and C \ is closed it follows that there is some wiggle room say r > 0 between the two sets. Formally r = inf distC\ (z(s, t)) > 0
s,t

because the minimum is attained. Since z is uniformly continuous, there exists 1 > > 0 such that |s1 s2 | < , |t1 t2 | < = |z(s1 , t1 ) z((s2 , t2 )| < r Next, for 0 < u < 1 , let u be a nonnegative function on R/Z zero outside a 2 interval of halength u about 0, with integral 1 and continuously differentiable1 . We can easily build u from parabolic segments as follows 3 2 for u x 1 u, 2u (x + u) 2 1 1 3 2 u 2u x for 1 u x 2 u, 2 u (x) = 2u3 (x u)2 for 1 u x u, 2 0 otherwise Now construct a new homotopy by2
1 1

w(s, t) =
0
1

s(1s) (x)z(s, t x)dx =

s(1s) (t x)z(s, x)dx

The family (u )u>0 is an example of a summability kernel on the circle group as u 0+. My MATH 355 notes give more detail on this topic. 2 This is an example of a convolution integral. We can write the formula as w s = s(1s) zs with denoting the convolution product. My MATH 355 notes give more detail on this topic.

46

4.0

3.0

2.0

1.0

-0.5

-0.4

-0.3

-0.2

-0.1

0.1

0.2

0.3

0.4

0.5

Figure 4.1: The function u for u = 0.25. for 0 < s < 1, a change of variables showing that the two expressions are equal and by w(0, t) = z(0, t) and w(1, t) = z(1, t) for the cases s = 0, 1. Differentiating under the integral sign (in t) in the second expression, we see that t w(t, s) is C 1 for each s [0, 1]. We leave it to the reader to check that (s, t) w(s, t) is continuous on ]0, 1[[0, 1]. A reader familiar with more advanced material would express this by pointing out that s zs is continuous from ]0, 1[ to C(R/Z), that s s(1s) is continuous from ]0, 1[ to L1 (R/Z) and that convolution is continuous as a map L1 (R/Z) C(R/Z) C(R/Z). A key point is that
1

w(s, t) z(s, t) = so that

s(1s) (x)(z(s, t x) z(s, t))dx

1

|w(s, t) z(s, t)|

s(1s) (x)|z(s, t x) z(s, t)|dx sup |z(s, t x) z(s, t)| < r

|x|<s(1s)

showing that w(s, t) . It remains to check that (s, t) w(s, t) is continuous on s = 0 and s = 1. Since the two cases are similar, we restrict attention to 47

s = 0. The continuity is equivalent to showing that w s converges to w0 uniformly on [0, 1] as s 0+. This follows since ws w0 = s(1s) leading to ws w0

zs z0 = s(1s)

(zs z0 ) + (s(1s)

z0 z0 )

z s z0

s(1s)

+ z0 (s(1 s)) 0
s0

zs z0

+ s(1s)

z0 z0

since z0 is uniformly continuous and uniform continuity of z implies that z s z0 as s 0+ uniformly. The new homotopy w has now been shown to satisfy all the conditions of the original homotopy z and the map t w(s, t) is C 1 for all s with 0 < s < 1. We will now assume that the original homotopy z has this additional property and maintain the denitions of r and for this new z. Then we can nish the proof quickly by showing that two C 1 loops which are sufciently close are homotopic by a highly regular homotopy. We choose a nite sequence 0 = s 0 < s1 < < sn = 1 with sk sk1 < for k = 1, 2, . . . , n. It is enough to show that f (z)dz =
(sk ) (sk1 )

f (z)dz.

for k = 1, 2, . . . , n. Let us x k. Note that |z(sk , t) z(sk1 , t)| < r and it follows that the piecewise linear homotopy w(s, t) = sz(s k , t) + (1 s)z(sk1 , t) (dened for 0 s 1 and 0 t 1) remains in . But this homotopy satises all the regularity conditions of Theorem 32 and the result follows. We are now almost ready to establish another version of Cauchys Theorem. D EFINITION An open subset of C is contractible if there is a point in and a continuous map : [0, 1] such that (1, z) = z for all z and(0, z) = for all z . T HEOREM 34 (C AUCHY S T HEOREM FOR C ONTRACTIBLE O PEN S ETS ) Let be a contractible open subset of C. Let F : C be holomorphic. Let be a C 1 loop in . Then F (z)dz = 0.

48

4.2

The Winding Number

The following concept is of great importance in the sequel. D EFINITION Let be a piecewise C 1 loop in C. Let C\, then the winding number wind () is dened by wind () = 1 2i dz . z

P ROPOSITION 35

In the above situation wind () is always an integer.

Proof. Suppose is C 1 and x a C 1 parametrization t z(t) of say over a t b and dene

t

g(t) =
s=a

z(s)

z (s)ds.

We get by applying the fundamental theorem of calculus to calculate g (t) d dt z(t) exp(g(t)) = z (t) exp(g(t)) z(t) g (t) exp(g(t)) = 0, So, z(t) exp(g(t)) is independent of t. Hence z(a) exp(g(a)) = z(b) exp(g(b)) and since z(a) = z(b) = 0, we nd exp(g(b) g(a)) = 1 It follows that g(b) g(a) 2iZ and the result follows. The case of a piecewise C 1 loop is only slightly more complicated. Its also necessary to understand the winding number from an intuitive point dz d dr of view. Putting z(t) = r(t)ei(t) , we get = ei + irei and dt dt dt (z(t) )1 49 dz dt = d dt

1 2 0

Figure 4.2: A loop and the winding number for points not on the loop. so that

1 wind () = 2

b a

d dt dt

viewing as a multivalued function. The interpretation is that 2 wind () is the angle that z(t) winds about (in the anticlockwise sense) as t runs from a to b. We clearly have wind (1 ) wind (2 ) = = | wind (1 ) wind (2 ) | 1 2i 1 2i 1 1 z 1 z 2 dz

|1 2 | length() 2 dist (1 ) dist (2 )

1 2 dz (z 1 )(z 2 )

so that wind () is continuous on C \ . A continuous function that takes only integer values is locally constant. For those who understand the concept of connected component, this can be expressed by saying that wind () is constant on each connected component of C \ . 50

4.3

Cauchy Integral Formula

Let be open in C. Let 0 and let = dist(0 , C \ ) > 0. Let 0 < r < . Then the closed disk D = {z; |z 0 | r} is contained in . Let be the circle |z 0 | = r traversed anticlockwise. T HEOREM 36 | 0 | < r . Proof. The function g(z) =

Let f be holomorphic in . Then f () =

1 2i

f (z) dz for z

f (z) f () is holomorphic in \{}. Let 0 < s < z r |z| and we imagine that s is very small. It is easy to see that is homotopic to the circular loop of radius s about in the region \ {}. Therefore
2

g(z)dz =
=0

f ( + sei ) f () i ise d sei

and we get
2

g(z)dz

=0

|f ( + sei ) f ()| d

2f |D (s) But f is uniformly continuous on D, so letting s 0 gives now get 1 2i as required. f (z) 1 dz = z 2i f () dz = f () z g(z)dz = 0. We

This result has far reaching consequences for 1 1 = = z (z 0 ) ( 0 )

n=0

(z 0 )n1 ( 0 )n

We note that if |z 0 | = r, | 0 | u < r, then f (z)(z 0 )n1 ( 0 )n sup |f (z)|r 1

zD

u r

51

It follows that

f () =
n=0

an ( 0 )n

(4.5)

with uniform and absolute convergence on | 0 | u where an =

n

1 2i

and |an | supzD |f (z)|r . Given any xed u with |u| < , we may always choose r with u < r < and hence (4.5) must have radius at least . Uniqueness considerations for the coefcients of power series guarantee that the a n derived from two different values of r would have to be the same. Effectively then, Theorem 36 has the following corollaries.

f (z) dz (z 0 )n+1

Let be open in C and let f be holomorphic in . Let 0 C OROLLARY 37 and let = dist(0 , C \ ) > 0. Then there exist complex numbers (an ) such n=0 that the power series
n=0

an ( 0 )n

has radius at least and converges to f () in | 0 | < .

In particular, the concepts analytic and holomorphic are equivalent. C OROLLARY 38 Let be open in C. Let 0 and let = dist(0 , C\) > 0. Let 0 < r < . Let be the circle |z0 | = r traversed anticlockwise and n Z+ . Then f (z) n! f (n) () = dz. 2i (z )n+1 Proof. It sufces to differentiate n times with respect to the result of Theorem 36 under the integral sign. In the special case = 0 , it can also be deduced from the formula for the coefcients of a power series in terms of the derivatives of the function it represents at the central point. C OROLLARY 39 (C AUCHY S E STIMATE ) Let be open in C. Let 0 and let = distC\ (0 ) > 0. Let 0 < r < . Let n Z+ . Then f (n) (0 ) n! sup |f (z)|. r n |z0 |=r 52

5
Holomorphic Functions Beyond Cauchys Theorem

In the last chapter, we saw that as soon as a function has a continuous complex derivative in an open set, then locally about every point in the open set, it has a power series expansion about that point with strictly positive radius of convergence. The corresponding result in the real setting is far from being true. In fact there are innitely differentiable functions that do not have a power series expansion at a point. The most well known example is (x) = exp(x2 ) if x = 0, 0 if x = 0.

is innitely differentiable on the whole of R and has (n) (0) = 0 for all n Z+1 . If had a power series expansion about 0 it would have to be 0 + 0x + 0x2 + and it is clear that is not identically zero in any neighbourhood of 0. This example does not work in C since exp(z 2 ) tends properly to innity as z tends to zero along the imaginary axis. The corresponding function is not even continuous at 0. So, holomorphic functions are very special and have many unexpected properties that we try to investigate in this chapter.
1

To prove this show by induction that (n) (x) =

Z+ where pn is a polynomial.

pn (x1 ) exp(x2 ) 0

if x = 0, for all n if x = 0.

53

5.1

Zeros of Holomorphic Functions

The situation with regard to zeros is summarized by the following theorem. T HEOREM 40 Let f be holomorphic in an open subset of C. Let . Then there is a number m Z+ {} called the order of as a zero of f such that m = 0 if f does not have a zero at , i.e. f () = 0. m = if f vanishes identically in a neighbourhood of . (z )m f (z) 1 (m) m N if g(z) = f () m! tion in and g() = 0.

if z = , if z = . denes a holomorphic func-

Proof. The assertion of the Theorem is that one of the cases listed must necessarily hold. By Corollary 37 we have

f (z) =
n=0

an (z )n

valid in |z | < r for some r > 0. If an = 0 for all n Z+ then we are clearly in case m = . So, if we are not in this case, there exists n such that a n = 0. Now the nonempty subset {n; an = 0} of Z+ has a least element by the Well Ordering Principle. Let this least element be m. If m = 0, then f () = a 0 = 0 and f does not have a zero at . So we are left with the case m N and then f () = a 0 = 0 so that f does have a zero at . We consider the function dened by an (z )nm if |z | < r, g(z) = n=m (z )m f (z) if z \ {}.

We note that

n=m

an (z )nm agrees with (z )m f (z) in 0 < |z | < r

since a0 , . . . , am1 all vanish. The function z

n=m

an (z )nm is holomor-

phic in |z | < r because it is given by a convergent power series expansion. 54

The function z (z )m f (z) is holomorphic in \ {} since it is the product of two holomorphic functions in that open set (i.e. by the CauchyRiemann equations). Note that g() = am = 0. C OROLLARY 41 Let be a path connected open subset of C, f holomorphic in and f () = 0 for a point . Then either f vanishes identically in or is an isolated zero of . Proof. We apply the previous result. If f (z) = (z )m g(z) with g holomorphic in , m N and g() = 0, then the continuity of g at implies that g(z) = 0 in some neighbourhood V of . It is then evident that if f (z) = 0 and z V , then z = , i.e. the zero is isolated. Therefore, if the zero is not isolated, then f vanishes identically in a neighbourhood of . Suppose that 1 with f (1 ) = 0. Since is path connected, there is a continuous path : [0, 1]

with (0) = and (1) = 1 . Unfortunately we wish to avoid situations in which has an interval of constancy containing 0. Consider rst = inf{t; 0 t 1, (t) = }.

The point t = 1 is an element of the set, so the inf is well-dened. By continuity of , we see that 0 < 1. Now replace by 1 (t) = ((1 t) + t). In this way, we can assume that there a points t > 0 arbitrarily close to 0 such that (t) = . Now consider = inf{t; 0 t 1, f ((t)) = 0}. Since f ((1)) = 0, this is the inmum of a nonempty set. By continuity of at t = 0 and since f vanishes in a neighbourhood of we have > 0. For t < , f ((t)) = 0 and therefore by continuity f (( )) = 0. By the previous part of the proof, either f vanishes in a neighbourhood of ( ) or ( ) is an isolated zero of f. Suppose rst that ( ) is an isolated zero of f . Then, there exists > 0 such that 0 < |z ( )| < = f (z) = 0. But, by denition of , f ((t)) = 0 for 0 t and it follows that 0 t = either |(t) ( )| = 0 or |(t) ( )| . 55

But the Intermediate Value Theorem yields that |(t) ( )| = 0 for all t [0, ] 1 for else, the value 2 would have to be taken by t |(t)( )| on this interval. But this means that has [0, ] as an interval of constancy, a situation that we were careful to eliminate earlier. So we are left with the case that f vanishes in a neighbourhood V of ( ). But then f ((t)) = 0 for t in a neighbourhood of . This contradicts the denition of unless = 1. Let us understand this in detail. By continuity of there exists > 0 such that |t | < implies that (t) V and hence that f ((t)) = 0. Since we know already that f ((t)) = 0 for 0 t , this yields f ((t)) = 0 for 0 t < + provided always that t 1. This gives a contradiction to the denition of unless = 1. But then we have f (( )) = 0 as before which contradicts f ( 1 ) = 0. We need to understand this result in context. If is a path connected bounded open subset of C and f holomorphic in then f can have innitely many zeros in , but necessarily the zeros accumulate only at the boundary of . 5.2 Bounded Entire Functions

D EFINITION An entire function is a function that is holomorphic in the entire complex plane. T HEOREM 42 (L IOUVILLE S T HEOREM ) sarily constant.

A bounded entire function is neces-

Proof. Suppose that |f (z)| M for all z. Then applying Corollary 39 (Cauchys Estimate) we get M |f (z)| r for all r > 0. It sufces to take r sufciently large to see that f (z) = 0 for all z C. Since f can be recovered from its derivative by means of f (z) = f (0) +

f (z)dz

where is a C 1 path from 0 to z (for example a line segment), it follows that f is constant. 56

C OROLLARY 43 (F UNDAMENTAL T HEOREM OF A LGEBRA ) Let p be a monic polynomial of degree m 1, then there exist 1 , . . . , m such that
m

p(z) =
k=1

(z k )

(5.1)

both in the sense of functions (i.e. that and in the sense that (5.1) holds for all z C) and in the sense of polynomials, (i.e. in the ring C[z]).
Proof. The proof is by induction on m. For m = 1 the result is evident. Suppose that the result has been proved for monic polynomials of degree m 1 and let p be a monic polynomial of degree m. We claim that p has a zero in C. If not, then consider 1 f (z) = p(z) in C. This function is the composition of two holomorphic functions z p(z) and w w 1 , the latter being dened on C \ {0}, where p takes its values. So
m

f is entire. But f is also bounded. to see this, we observe that if p(z) = with pm = 1 we get
m1 k=0

pk z k

|p(z)| |z| |z|m

1 |z|m 2

k=0

|pk ||z|k
m1

k=0

|pk ||z|km

for |z| large,

1 since the quantity in brackets exceeds 2 if |z| is large enough. Since f is continuous, it is necessarily bounded on any bounded set and hence is bounded on the whole of C. So f is constant and f (0)p(z) = 1 for all z C which contradicts the fact that p has degree 1. Thus, p vanishes somewhere, say at m . Then the fact (easily veried) that k z m divides z k m for every k Z+ yields that z m divides p(z) = p(z) p(m ). So we may write p(z) = (z m )q(z). We see that q is a monic polynomial of degree m 1 and the induction hypothesis nishes the proof. The

57

fact that polynomial functions and polynomials are identical concept on C is an easy consequence of the standard formula for the vandermonde determinant. We leave the details to the reader. Similar to Liouvilles Theorem we have the following proposition. P ROPOSITION 44 Let f be entire and satisfy |f (z)| C(1 + |z|) where C and are absolute constants, m Z+ and < m + 1. Then f is a polynomial of degree at most m. Proof. Again, Cauchys Estimate (Corollary 39) gives |f (m+1) (z)| (m + 1)! sup |f ()| r m+1 |z|=r (m + 1)! (1 + |z| + r)m+ r m+1

Letting r , we get f (m+1) (z) = 0 for all z C. Integrating up m + 1 times now shows that f is a polynomial of degree at most m. 5.3 The Riemann Sphere and Mobius Transformations

There are various ways of thinking about the Riemann Sphere which will be denoted C . As a set, it is the abstract union of C with {}, the singleton consisting of the point at innity. From the metric or topological space point of view it is the one point compactication of C. This means that we can put a metric on C for which a sequence (zn ) converges to innity if and only if (zn ) diverges properly to . If you want an explicit metric that embodies this, you could take d(z1 , z2 ) = max z1 |z2 | z2 |z1 | , 2 2 1 + |z1 | 1 + |z2 | 1 + |z1 | 1 + |z2 | |z| z , 1 2 1 + |z| 1 + |z|

with the specic interpretation d(z, ) = d(, z) = max

58

The Riemann Sphere is compact. to see this take a sequence in C . If the sequence is bounded, then it has a convergent subsequence by the Bolzano Weierstrass Theorem. If it is unbounded, then it has a subsequence properly divergent to 2 . A more sophisticated way of thinking of C is as the complex projective space CP1 , that is the space of one dimensional complex linear subspaces of C 2 . For most of those subspaces, we can nd a basis vector in the form (1, z), but the subspace {0} C is the sole exception to this. If we identify z with linspan{(1, z)}, then the point at innity is identied to the line {0}C = linspan{(0, 1)}. There is actually nothing special about the one-dimensional subspace {0} C, to an impartial observer, it looks just like any one-dimensional subspace and consequently, there is nothing special about the point at innity in C . Yet another way of viewing C is as a manifold. This is a complicated concept, but one worth exploring a little bit. We start in the real setting. It is clear that one can think for example of the sphere S 2 in R3 and consider differentiable functions dened just on S 2 (and not on R3 \ S 2 ). Here, the sphere and its differentiable structure come from the fact that S 2 is embedded in the space R3 . We would like to nd a way of describing the differentiable structure of S 2 without reference to this embedding. This is the concept of a manifold. Nicely embedded surfaces are manifolds, but the manifold structure should not reference this embedding. So, to dene a manifold, we rst decide upon a dimension d for the manifold (in the case of S 2 this will be 2) and we insist that the manifold M is a metric space. Sometimes additional conditions are imposed already at this point. The manifold M has an atlas which is a collection of charts : U V as runs over an index set I. The subsets U are open in M and they cover M , i.e. U = M . The sets V are open subsets of Rd and the mappings are
I

continuous bijections with continuous inverses. The analogy with the charts of a real-world atlas is based on the idea that we do not try to understand the world in its entirety. We understand it a bit at a time by examining individual maps (charts) of specic regions. However, its very important that where two charts overlap, they carry the same information. So there is a compatibility condition which concerns the overlap mapping 1 : (U U ) (U U )
We just showed that C is sequentially compact. A well-known theorem tells that every sequentially compact metric space is in fact compact.
2

59

This mapping is certainly continuous, but we may impose additional regularity. If we are trying to dene a C k manifold, we will insist that each overlap map is C k . Observe that the overlap mappings are dened from one open subset of R d to another and so questions of differentiability have a perfectly good meaning. Once this has been done, we can dene C k functions on M . Given say f : M R, f will be C k on M if and only if each of the mappings f 1 : V R is C k . Here we are using the C k regularity only as an example. The avour of the overlap mappings essentially determines the avour of the manifold. More generally it is possible to dene C k functions from one C k manifold to another. This is just a very naive introduction to the subject. In practice there are all sorts of problems that arise3 . So, to come back to the Riemann Sphere, it is a complex analytic manifold. We can work with just two charts. U1 = V1 = {z; z C}, 1 is the identity map. U2 = {z C; z = 0} {}, V2 = V1 , 2 (z) = z 1 0 if z C \ {0}, if z = .

Both overlap mappings 2 1 and 1 1 are the map z z 1 on the 1 2 punctured plane z C; z = 0 which is holomorphic. Hence with this atlas, C is a holomorphic manifold (complex analytic manifold). We have the right to dene holomorphic functions on C . At rst sight, it seems that we have gone to a lot of trouble here for nothing. What are the holomorphic functions from C to C? Well they are continuous and C is compact, so they are necessarily bounded and they are clearly holomorphic on the subset C of C , so according to Liouvilles Theorem, they are constant on C and hence by continuity on C . However, there might be non-trivial holomorphic mappings from C to itself. The M bius transformations on C are the transformations induced on C o by invertible linear transformations of C2 , viewing C as projective space in C2 . Let T =
One for example that occurs almost immediately is that completely different atlases can dene equivalent structures (two real-world atlases from real-world publishers are supposed to carry the same information, but the selection of charts will be different), so one needs a concept of equivalent atlases.
3

60

with det(T ) = = 0. Then z 1 = z + z +

so that the corresponding transformation of C is z z + z +

it being understood that the point at innity maps to 1 and 1 maps to the point at innity. In case = 0 the point at innity is preserved and the mapping is actually afne (constant plus linear) on C. Note that if t C \ 0, then the matrix tT yields the same Mobius transformation as T . It is easy to check that Mobius transformations are holomorphic on C . To check analyticity at , we must check that + w w 1 + = w 1 + w + w is analytic at w = 0 and this is OK unless = 0 when we need to check that w + w + w
1

w + w = + w + w

is analytic at w = 0. This is good since if = 0 then the determinant condition is violated. L EMMA 45 Every Mobius Transformation can be expressed as a composition of translations, dilations and inversions. Proof. Throughout the proof, denotes a generic non-zero complex number. We have to build up every Mobius Transformation as a composition of Mobius Transformations of one of the following forms: z z + , z z, z z 1 .

Now every Mobius Transformation is given by a non-singular matrix and every non-singular matrix is a product of elementary matrices (remember gaussian reduction). We need therefore, only work with Mobius Transformations that are generated by elementary matrices. 61

1. 2. 3. 4. 5.

0 0 1 1 0 0 0 1 1 0 1 0 1 1 0 1

, , , , =

z z. z 1 z. z z 1 . z z + . 0 1 1 0 1 0 1 0 1 . 1 0

Cases 1 thru 4 are of the desired type and in case 5 we show how to build this case out of the previous four cases. P ROPOSITION 46 C .

The group of Mobius Transformations is triply transitive on

Proof. Mobius Transformations clearly form a group since invertible 2 2 matri ces do. Let a, b, c be distinct in C , then consider (z) = c(b a)z + a(c b) (b a)z + (c b)

Similarly, if a , b , c are distinct in C , there exists a Mobius Transformation taking 0, 1 and to a , b and c . It follows that 1 takes a, b and c to a , b and c respectively. 62

then is a genuine Mobius Transformation and takes 0, 1 and to a, b and c respectively. The denition of needs modication in certain special cases: cz + c b if a = , z cz a (z) = if b = , z1 (b a)z + a if c = . 1

Actually, the Mobius Transformation that does this is uniquely determined. To see this, suppose that 1 and 2 both take a, b and c to a , b and c respectively. Then = 1 2 xes a, b and c. Letting 1 (z) = z + z +

we see that the equation (z) = z has three distinct roots and it follows that the quadratic equation z 2 + ( )z = 0 must vanish identically. So = = 0 and = . (Of course, = 0, so that 2 = 0 and = 0. We nd (z) = z for all z C , or equivalently 1 = 2 . We next dene the cross ratio of four distinct complex numbers z 1 , z2 , z3 and z4 to be (z1 z2 )(z3 z4 ) . [z1 , z2 , z3 , z4 ] = (z1 z4 )(z3 z2 ) The order of the numbers is important. The denition can be extended to the case of zj C for j = 1, 2, 3, 4 provided {z1 , z2 , z3 , z4 } has at least 3 elements but also takes values in C . L EMMA 47 We have [(z1 ), (z2 ), (z3 ), (z4 )] = [z1 , z2 , z3 , z4 ] for a Mobius Transformation. Proof. It sufces to check this in the case that is a translation, dilation or inversion. In each of these cases, the verication is routine. It is worth observing that [z(1) , z(2) , z(3) , z(4) ] = [z1 , z2 , z3 , z4 ] if is a double transposition in S4 . For example, we have [z2 , z1 , z4 , z3 ] = (z2 z1 )(z4 z3 ) (z1 z2 )(z3 z4 ) = = [z1 , z2 , z3 , z4 ] (z2 z3 )(z4 z1 ) (z1 z4 )(z3 z2 )

Let = [z1 , z2 , z3 , z4 ]. Now, it is well known that the set of double transpositions together with the identity permutation form a normal subgroup H of S 4 which is clearly transitive in its action on the 4-tuple (z1 , z2 , z3 , z4 ), so each coset of H will contain a permutation that xes 1 say. There are six such permutations which accounts for all the elements in the coset space (quotient). Now there is a Mobius transform f with f (z2 ) = 0, f (z3 ) = 1 and f (z4 ) = and since Mobius transforms preserve cross-ratio, we have [z1 , z2 , z3 , z4 ] = = [, 0, 1, ]. Thus f (z1 ) = . It is therefore sufcient to compute [, 0, 1, ] = , [, 1, , 0] = ( 1)/, [, , 0, 1] = 1/(1 ), 63

[, 0, , 1] = /( 1), [, , 1, 0] = 1/, [, 1, 0, ] = 1 . in order to determine all possible values of [z(1) , z(2) , z(3) , z(4) ] as runs over S4 . By a circle in C , we mean either a circle in C or a straight line in C together with the point at innity. (i.e. we interpret a straight line as a circle that passes thru ). L EMMA 48

A Mobius transform maps circles to circles.

Proof. It is easy to see that any circle can be represented by the equation p|z|2 2 (az) + q = 0 where a C, p, q R and pq < |a|2 . It is also easy to see that substituting z = w + , z = w or z = w 1 into such an equation yields an equation of the same type. Since these transformations generate all Mobius transforms, the result follows. C OROLLARY 49 If z1 , z2 , z3 and z4 are distinct points of C , then [z1 , z2 , z3 , z4 ] is real if and only if there is a circle to which all four points belong. Proof. We use the fact that the group of Mobius transformations is triply tran sitive, preserves the cross-ration and maps circles to circles to show that without loss of generality we may take z2 = 0, z3 = 1 and z4 = . In this way, the problem reduces to showing that (for z C \ {0, 1}, z = [z, 0, 1, ] is real if and only if z lies on the circle thru 0, 1 and , i.e. the real axis. Of course, we usually think of a circle as having an inside and an outside. So which is which. If we take z2 , z3 and z4 in order and orient the circle S thru z2 , z3 and z4 by traversing in that order, then it can be shown that the set {z; z C , [z, z2 , z3 , z4 ] > 0} is the connected component of C \ S on the left and {z; z C , [z, z2 , z3 , z4 ] < 0} is the connected component of C \ S on the right.

64

5.4

Preservation of Angle

At points where the derivative fails to vanish, holomorphic functions are orientation preserving. In fact, we can actually say much more. P ROPOSITION 50 Let f be holomorphic in a neighbourhood of a point and suppose that f () = 0. Then the Jacobian matrix of f is a positive scalar multiple of a rotation matrix. Proof. This is an immediate consequence of the CauchyRiemann equations. We can write the Jacobian as p q J= q p u v v u =p= and = q = . In case x y x y p and q are not both zero, we can write where f = u + iv, u and v are real, p q for r = q p =r cos() sin() sin() cos()

p2 + q 2 > 0 and suitable real .

Note that the Jacobian determinant is r 2 = p2 + q 2 > 0, so that locally near , f preserves orientations. We can understand Proposition 50 in the following way. Suppose that t (t) and t (t) are two smooth curves thru with (0) = (0) = . Then t f ((t)) and t f ((t)) are also smooth curves thru f () with f (0) = f (0) = f (). We have from the chain rule (f ) (0) = J( (0)) (f ) (0) = J( (0))

The angle between the curves and at is just the angle between the vectors (0) and (0) and since J is a positive multiple of a rotation matrix, this is the same as the angle between (f ) (0) and (f ) (0) which is just the angle between the image curves f and f at f (). Thus, where a holomorphic transformation has non-zero derivative, it preserves oriented angle between curves. Even at a point where the derivative vanishes, we can still say something about angles between curves. 65

Figure 5.1: Argument modulus plot. P ROPOSITION 51 Let be an open subset of C, and f a holomorphic function on with a zero of order m at . Suppose 1 m < . Then there is a neighbourhood U of and a holomorphic function g dened on U such that m f (z) = g(z) for z U . Necessarily, g has a single zero at (i.e. g () = 0). Proof. Obviously, if m = 1 we can take U = and g = f . By Theorem 40 we can write f (z) = (z )m h(z) with h holomorphic in a neighbourhood of and h() = 0. Now consider an open disk centred at h() and of radius |h()|. Then we can dene a holomorphic mth root in . In fact, with a little trouble we could even write down a power series expansion for centred at h() and convergent in . Now set g(z) = (z )(h(z)) dened in h 1 (). Then g is holomorphic since it is the product of holomorphic functions ( h being the composition of holomorphic functions). Also we have (g(z)) m = (z )m ((h(z)))m = (z )m h(z) = f (z) as required. What has this to do with angles? Well if f is holomorphic and at a point we have f () = 0, . . . , f (m1) () = 0 but f (m) () = 0, then z f (z) f () has a zero of order m at . Therefore, locally about we can write f (z) = 66

f ()+(g(z))m where g is holomorphic near and g () = 0. So, at , g preserves angles and it follows that at the function f multiplies angles by m. To be explicit about this, if we take two curves and emanating from then the angle between the curves f and f will be m times the angle between the curves and . It has to be stressed that this occurs only at the point . As soon as you move slightly away from you will necessarily have a non-zero derivative and the usual preservation of angles holds. Figure 5.1 shows an argumentmodulus plot for a function f . To be more explicit, the function shows in the z-plane, curves on which the argument and modulus of w = f (z) are constant. The argument is constant on the red and black curves and the modulus is constant on the blue curves. The increment in the argument is /6 (i.e. 30 degrees) and the increments in ln(r) are also /6 (for the curves |w| = r). Here are some questions that the reader might like to consider. There are three zeros of f in the gure. Where are they? One of the zeros is a double zero. Which one? The other two are single zeros. How can you be sure? In most places, the blue curves meet the red and black curves at right angles. Why is this? The curviquadrilaterals of which most of the gure is composed are approximately square in shape. Why is this? How can you be sure that it is not a mixture of zeros and poles that is being displayed. What changes in the gure would you expect if the zero with the largest y coordinate were replaced with a pole? Where approximately are the zeros of f ? Two of the regions depicted have eight sides. How do you account for this and what conclusions can you draw? If there were a region with twelve sides, what conclusion would you draw? In argumentmodulus plots in general, is it possible for a curve of constancy of the modulus to be a smooth gure eight?

67

T HEOREM 52 (O PEN M APPING T HEOREM ) Let be an open path connected subset of C. Let f : C be non-constant and holomorphic. Then for every U open in , f (U ) is open in C. Proof. It is enough to show that every point of possesses an open neighbourhood V such that f (V ) is open. If f () = 0, the we may apply the Inverse Function Theorem (from MATH 354) to deduce the result. If not, then has nite multiplicity as a zero of f (since if not then f would vanish identically on by Corollary 41 and f would be constant on . So, according to the previous result, we have f (z) = f () + g(z) for z in a neighbourhood of and g () = 0. It therefore sufces to show that if V is open in C, then W = {z m ; z V } is also open in C. If w W \ {0}, then w = z m with z V and z = 0 and the result follows as above since mz m1 = 0. If 0 W , then 0 V and there exists > 0 such that U (0, ) V whence U (0, m ) W . 5.5 Moreras Theorem
m

Moreras Theorem is a converse to Cauchys Theorem. Its strength lies in the fact that the only regularity imposed on the function is continuity. T HEOREM 53 (M ORERA S T HEOREM )

Let f be a continuous function dened

f (z)dz = 0.
T

on an open subset of C. Suppose that for every triangle T we have Then f is holomorphic in .

Proof. Being holomorphic is a local property, so it sufces to prove the result on an open disk whose closure is contained in . Without loss of generality on {z; z C, |z| < 1}. Dene F (z) = f (w)dw where L(a, b) denotes the
L(0,z)

0, z and z + h form a triangle T so the hypothesis F (z + h) F (z) = f (w)dw 68

line segment from a to b. Then consider F (z + h) F (z). Now the three point f (z)dz = 0 can be written f (w)dw.
L(z,z+h) T

f (w)dw =
L(0,z)

L(0,z+h)

Parametrizing L(z, z + h) by t z + th for t running from 0 to 1, we get

1

f (w)dw = h
L(z,z+h) 0

f (z + th)dt.

It now follows that

1

F (z + h) = F (z) + f (z)h + h
0

f (z + th) f (z) dt.

and h
0

f (z + th) f (z) dt |h|f |K (h)

where K = {z; z C, |z| 1}. Since f is continuous on K it is uniformly continuous and it follows that F has a complex derivative f (z). But f is continuous, so F is holomorphic. But the derivative of a holomorphic function is again holomorphic. Moreras Theorem is not a curiosity, it is a powerful tool. Its main application is the following result which might be less easy to prove by other means. C OROLLARY 54 Let be an open subset of C. Let (fn ) be a sequence of holomorphic functions converging uniformly on the compacta of to a function f . Then f is holomorphic. Proof. Let T be a triangle in , then we will show
T

f (z)dz = 0. Since T will

be arbitrary in we can apply Moreras Theorem to obtain the desired conclusion. By Cauchys Theorem we have
T

fn (z)dz = 0. We can prove this either from the

Greens Theorem version of Cauchys Theorem, or using the fact that triangles are contractible. The uniform on compacta convergence yields the convergence of the integrals. f (z)dz fn (z)dz =
T T zT

(f (z) fn (z))dz

sup |f (z) fn (z)| length(T ) 0 as n . This gives the desired equality 69 f (z)dz = 0.
T

6
The Maximum Principle

The maximum principle actually applies to harmonic functions rather than analytic functions, so it makes sense to revisit the subject of harmonic functions. It is also very little trouble to tackle this issue in Rd rather than just in R2 . 6.1 Harmonic Functions Again

We start with some motivational material in two dimensions. Let u be a realvalued harmonic function dened in a neighbourhood of the closed unit disk {z C; |z| 1}. Then we can nd a conjugate harmonic function v dened in a possibly smaller neighbourhood of the closed unit disk. Then, applying the Cauchy Integral formula (Theorem 36) we have 1 (u + iv)(z) = 2 = 1 2
2 t=0 2 t=0

(u + iv)(eit ) it e dt eit z

(u + iv)(eit )
n=0

eint z n

dt

for |z| < 1. Now from Cauchys Theorem we have 1 2

2

(u + iv)(eit ) eint dt =
t=0

1 2i

(u + iv)() n1d = 0,

where the path for the integral on the right is the unit circle oriented anticlockwise. We can therefore write 1 (u + iv)(z) = 2
2

(u + iv)(e )
t=0 n=0

it

int n

z +
n=1

eint z n dt

70

again for |z| < 1. Computing the geometric sums in the brackets and combining, we get 2 1 1 |z|2 dt. (u + iv)(z) = (u + iv)(eit ) 2 t=0 |z eit |2 But the function 1 |z|2 is positive and taking real parts we get |z eit |2 1 u(z) = 2 1 |z|2 u(e ) dt |z eit |2 t=0
it 2

which is the Poisson Integral formula for a harmonic function. The case z = 0 is particularly important 2 1 u(eit )dt u(0) = 2 t=0 and tells that the average value of a harmonic function on a circle is the value at the centre of the circle. 6.2 Digression Poisson Integrals in Rn

This whole theory works out almost as well in Rd . So imagine that a real-valued function f is continuous on the unit sphere S d1 = {y Rd ; |y| = 1} in Rd . We will form the Poisson Integral of f , F (x) = 1 Ad1 1 |x|2 f (y) d(y) |x y|d (6.1)

for |x| < 1 and where is the d 1 dimensional area measure on S d1 and Ad1 denotes the d 1 dimensional area of S d1 . Let now y S d1 be xed and let p(x) = 1 |x|2 and q(x) = |x y|d . We 2 v d 1 v + nd (pq) = p q + 2 p q + q p, using the formula v = r 2 r r for radial functions. This gives (q) = d(d+1)|xy|d2 d(d1)|xy|d2 = 2d|x y|d2 and (pq) = 2d(1 |x|2 )|x y|d2 + 4d|x y|d2 x (x y) 2d|x y|d = 2d|x y|d2 1 |x|2 + 2|x|2 2x y |x|2 + 2x y |y|2 = 2d|x y|d2 1 |y|2 = 0, 71

since |y| = 1. Hence, differentiating twice (i.e. applying the Laplace operator) under the integral sign in (6.1), we get F = 0. The differentiation under the integral sign is valid because so long as x is kept in a ball |x| r with 0 < r < 1 1 |x|2 is innitely differentiable with derivatives and |y| = 1, the function x |x y|d of all orders being bounded and continuous and the integral over the sphere will somehow be built up out of iterated integrals involving d 1 one-dimensional Riemann integrals. It will therefore sufce to apply the regular differentiation under the integral sign theorem, multiple times. Our objective is T HEOREM 55 For f a continuous function on the unit sphere S d1 we dene F (x) by (6.1) if |x| < 1 and F (x) = f (x) if |x| = 1, thereby obtaining a function on the closed unit ball. We claim that F is continuous on the closed unit ball and harmonic on the open unit ball. In the parlance of PDE, F solves the Dirichlet Problem on the unit ball. Proof. In case d = 1 we proceed by direct calculation to verify that F (x) = 1 f (1)(1 + x) + f (1)(1 x) 2

which is easily seen to satisfy the requirements of the theorem. Hence we may always assume that d 2. Now consider the case where f 1, we have G(x) = 1 Ad1 1 |x|2 d(y) |x y|d

is harmonic in the open unit ball and it is clear from the denition that G(x) depends only on |x|. This is ultimately a consequence of the fact that the area measure on S d1 is rotationally invariant. But, we can solve for radial harmonic functions and we obtain G(x) = A + B ln(|x|) A + B|x|d+2 if d = 2, if d 3.

Since G is clearly bounded, we deduce in either case that B = 0 and therefore G is constant. But G(0) = 1, so we nd G(x) = 1 for all x with |x| < 1. 72

To complete the proof of the theorem, we proceed to cut F onto spherical shells. The function fr will be essentially the restriction of F to a sphere centred at the origin of radius r, but parametrized on the unit sphere, namely fr (z) = F (rz) = = where Pr (z, y) = plicitly 1 Ad1 1 Ad1 f (y) 1 r2 d(y) for |z| = 1 |rz y|d

Pr (z, y)f (y)d(y)

1 r2 . We claim that Pr is a summability kernel, ex|rz y|d

Pr (z, y) 0, for all z, y S d1 and 0 r < 1. Pr (z, y)d(y) = 1 for all z S d1 and 0 r < 1. Pr (z, y)d(y) 0 for all > 0.
r1

sup

zS d1

|zy|>

The Summability Kernel Theorem then shows that fr f uniformly on S d1 as r 1. A moments thought shows that this implies the continuity of F on the closed unit ball. The rst and second conditions for a summability kernel have been shown. It remains to verify the third. We claim that
1 |rz y| 2 |z y|.

To see this, it sufces to show 4(r 2 |z|2 2rz y + |y|2 ) |z|2 2z y + |y|2 or equivalently
1 When 0 r 4 the worst case is when z y = 1. It boils down to 4r(r+2) 0 1 and when 4 r < 1, the worst case is when z y = 1 leading to 4(1 r)2 0. The claim is established. Therefore

4r 2 + (2 8r)z y + 2 0.

|zy|>

Pr (z, y)d(y)

1 Ad1
S d1

(1 r 2 )2d d(y) = 2d d (1 r 2 ). d

73

The proof is complete. Let us now back up and establish the result on summability kernels. We will state this in terms of a sequence of kernels n on a compact metric space X and some measure on X (this concept is made precise in MATH 355). We rst assume that n is sufciently regular on X X. Continuity on X X is sufcient, but there are cases where one is interested in kernels that are not continuous. Then come the three dening properties n (x, y) 0 for all x, y X and n N.
X

n (x, y)d(y) = 1 for all x X and n N. n (x, y)d(y) 0

n

For all > 0 we have sup

xX

d(x,y)>

T HEOREM 56 (S UMMABILITY K ERNEL T HEOREM ) Let (n) be a summan=1 bility kernel as above. Let f be a continuous real-valued function on X (complexvalued will also work). Then fn (x) = n (x, y)f (y)d(y) f (x)

uniformly on X as n .
Proof. We have by the second condition fn (x) f (x) = and by the rst |fn (x) f (x)| n (x, y) f (y) f (x) d(y) n (x, y) f (y) f (x) d(y)

Let > 0. Then, since f is uniformly continuous (it is continuous on a compact space), we can nd > 0 such that f () < 1 . So 2 |fn (x) f (x)| 74

d(x,y)

n (x, y) f (y) f (x) d(y) +

xX

d(x,y)>

n (x, y) f (y) f (x) d(y)

n (x, y)f ()d(y) + 2 sup |f (x)|

d(x,y)>

n (x, y)d(y)
d(x,y)>

1 + 2 sup |f (x)| 2 xX

n (x, y)d(y) <

provided that n is large enough by the third condition. 6.3 Maximum Principles for Harmonic Functions

Let be a bounded open P ROPOSITION 57 (F IRST M AXIMUM P RINCIPLE ) subset of Rd . Suppose that f is continuous real-valued on cl() and harmonic in . Then sup f (x) sup f (x)
xcl() x

Proof. Suppose not. Then there is a point z such that f (z) > sup f (x).
x

Now is bounded, so there exists > 0 such that

x

sup |x z|2 < f (z) sup f (x).

x

Now dene g(x) = f (x) + |x z|2 . We have g(z) = f (z) > sup f (x) + sup |x z|2 sup g(x)
x x x

(6.2)

Now, cl() is compact and g is continuous on cl(), so it attains its maximum value at a point y cl(). But because of (6.2), we see that the maximum cannot be taken on . It follows that y cl() \ = . It follows from basic calculus that the Hessian of g at y is negative semi-denite. In particular the trace of the Hessian is 0. But the trace of the Hessian is just the Laplacian, so 0 a contradiction. 75 g(y) = f (y) + 2 d = 2 d > 0

C OROLLARY 58 Let be a bounded open subset of Rd . Suppose that f is continuous complex-valued on cl() and harmonic in . Then
xcl()

sup |f (x)| sup |f (x)|

x

Proof. Suppose not. Then there is a point z such that |f (z)| > sup |f (x)|.
x

Let be the complex sign of f (z). Then let g(x) = f (x). Then g is continuous real-valued on cl() and harmonic in . But g(z) = |f (z)| > sup |f (x)| sup g(x)
x x

contradicting Proposition 57. A further corollary is now given. C OROLLARY 59 Let G be a function continuous on the closed unit ball and harmonic on the open unit ball. Then G(x) = 1 Ad1 G(y) 1 |x|2 d(y). |x y|d (6.3)

for |x| < 1. In particular, putting x = 0, we have

G(0) = 1 Ad1 G(y)d(y),

the mean-value property.

Proof. Let F (x) be the right-hand side of (6.3) for |x| < 1 and F (x) = G(x) for |x| = 1. Then according to Theorem 55, F is continuous on the closed unit ball and harmonic on the open unit ball. We consider H(x) = F (x) G(x). Then by Corollary 58 sup |H(x)| sup |H(x)| = 0.
|x|1 |x|=1

So, H vanishes identically and we have our result. Obviously, this last corollary can be scaled to any ball. 76

P ROPOSITION 60 (S ECOND M AXIMUM P RINCIPLE ) Let be a open connected d subset of R . Suppose that f is real-valued harmonic in and that f attains its maximum in . Then f is constant in . Proof. Let the maximum value be M . Let Z = {x ; f (x) = M }. Then Z is a non-empty subset of which is relatively closed in . It is the inverse image of a singleton by a continuous function and singletons are necessarily closed. We will show that Z is open. Let x Z and choose r > 0 such that U (x, r) . Consider 0 < t < r and according to the mean-value principle scaled to a ball of radius t about x that M = f (x) = 1 Ad1
S d1

f (x + ty)d(y).

(6.4)

Now clearly f (x + ty) M . Suppose that for some z with |z| = 1 we have f (x + tz) < M , then, using the continuity of f , there is an > 0 and an open neighbourhood V of z in S d1 such that f (x + ty) < M for all y in V . But now f (x + ty)d(y) f (x + ty)d(y) +
S d1 \V V

f (x + ty)d(y) d(y)
V

S d1

M =M

S d1 \V

d(y) + (M )
V

S d1

d(y)

d(y) < M Ad1

leading to a contradiction with (6.4). Hence f (z) = M on U (x, r). It follows that Z is open relative to . But is a connected set and it follows that Z = . C OROLLARY 61 Let be a open connected subset of Rd . Suppose that f is complex-valued harmonic in and that |f | attains its maximum in . Then f is constant in . Proof. Let |f | attain its maximum M at z . Let be the complex sign of f (z). Then let g(x) = f (x). Then g is real-valued harmonic in , g(z) = M and g(x) M for all x . So, according to Proposition 60, g is constant in . This says that f (x) = M for all x and it follows that |f (x)| M for all 77

x . But |f (x)| M for all x so |f (x)| = M for all x . Combining this with f (x) = M , we see that f (x) = M for all x . Since in d = 2, holomorphic functions are examples of complex-valued harmonic function, the results above for complex-valued harmonic functions also holds in that context. We do have the following result. P ROPOSITION 62 Let f be holomorphic in a connected open set of C and suppose that |f | attains its minimum value. Then either f is constant, or the minimum value is zero. Proof. Suppose that the minimum value is not zero. Then |f (z)| > 0 for all z . Hence g(z) = (f (z))1 is holomorphic in and attains its maximum value. By Corollary 61, g is constant in and hence, so is f . We can now answer the question posed on page 67 concerning holomorphic functions with constant modulus on a gure eight curve. Certainly the constant modulus cannot be zero, for the zero set of a holomorphic function either consists of isolated points unless the function is locally zero. Clearly the function is not locally constant and by compactness it must attain its minimum modulus in each of the closed loops of the gure eight. So, by Proposition 62, the function must vanish somewhere inside each of the closed loops. So, let us try to construct such a function and see what happens. We take f (z) = 1z 2 with zeros at 1. At z = 0 the function f takes the value 1. Where do we have |f (z)| = 1. We will use polar coordinates. The equation |1z 2 | = 1 becomes (1r 2 cos(2))2 +r 4 (sin(2))2 = 1 which simplies to r 4 2r 2 cos(2) = 0 or, since the origin is on the curve anyway, r 2 = 2 cos(2). This curve is called a lemniscate and has exactly the form of a gure eight. So the answer to the question is yes. 6.4 The PhragmenLindelof Method

The rst result is an extension of the maximum modulus principle to a strip. Proposition 57 does not apply because a strip is unbounded.

Let = {x + iy; x, y real , |y| < }, f be continuous on P ROPOSITION 63 2 the closure of , holomorphic in and satisfy |f (z)| 1 for z . Further, suppose that there are constants a < and b < 1 such that
|f (z)| exp(aeb|
z|

) for all z .

(6.5)

Then |f (z)| 1 for z .

78

The growth condition (6.5) is necessary. Consider for example f (z) = exp(e z ). Roughly speaking the proposition says that the growth has to be really wild before the maximum modulus principle will fail on the strip. Proof. Choose c with b < c < 1. let > 0 be arbitrary, but xed for the moment. Let h (z) = exp( cosh(cz)) an entire function. Then for z = x + iy , with x, y real, cosh(cz) = cosh(cx) cos(cy) cosh(cx) where = cos(c/2) > 0. Thus |f (z)h (z)| exp(aeb|x| cosh(cx)) 0 as |x| since b < c. On |f (z)h (z)| exp( cosh(cx)) 1. We now apply Proposition 57 on the open subset R = {x + iy; x, y real , |x| < R, |y| < } where R is sufciently large. The result is that 2 |f (z)h (z)| 1 for all z R . Letting R increase to innity, we now obtain |f (z)h (z)| 1 for all z , or equivalently |f (z)| e desired result.
cosh(z)

. Letting

decrease to zero now gives the

The next step in this saga is the following theorem which turns out to play a crucial role in functional analysis. To go into the details of why this result is so important is unfortunately beyond the scope of this course. T HEOREM 64 (T HE T HREE L INES T HEOREM ) Let = {x + iy; x, y real , 0 < x < 1}, f be continuous on the closure of , holomorphic in . Suppose that there are positive constants M0 , M1 , a < and b < such that exp(aeb|y| ) for 0 < x < 1, y R, |f (x + iy)| M0 for x = 0, y R, M1 for x = 1, y R. 79

1x x Then |f (x + iy)| M0 M1 for 0 < x < 1, y R.

Proof. We start by adapting the previous result to the strip involved here. We then apply that result to the function
1+z z g(z) = M0 M1 f (z).

The desired conclusion follows.

80

7
Isolated Singularities and Residues

Morally speaking, a singularity is a place where a function has bad behaviour or is not properly dened, with the understanding that the function has good behaviour nearby. D EFINITION Let where is open in C. Then a function (dened on ) which is holomorphic in a punctured neighbourhood V = {z C; 0 < |z | < r} of or radius r > 0 is said to have an isolated singularity at . Here we are assuming that V = {z C; |z | < r} There is a classication of isolated singularities as follows removable singularities (i.e. singularities that arent really there at all), poles and essential singularities (i.e. everything else). Here are the denitions. D EFINITION Let where is open in C. Let f be a function dened on which is holomorphic in V . Then is a removable singularity , if there is a number a C such that the function f dened by f (z) = a f (z)

if z = , if z V .

is holomorphic in V .
is a pole if the function f dened by f (z) = f (z) 81

if z = , if z V .

is holomorphic as a map from V to C .

is an essential singularity if it does not t into either of the two above cases. T HEOREM 65 Let f be a function bounded and holomorphic in V . Then f has a removable singularity at . In fact, boundedness is too strong, one can get away with replacing the boundedness with the condition
z

lim |z ||f (z)| = 0.

Proof. The idea is to dene g in V by g(z) = 0 (z )2 f (z) if z = , if z V .

Then clearly g has a complex derivative g (z) = (z )2 f (z) + 2(z )f (z) in V . Also g(z) g() = (z )f (z) 0 z as z . Hence g () exists and equals zero. We claim that g is holomorphic in V . It remains only to check that g is continuous in V . This is obvious, except at . Let > 0. By hypothesis, there exists > 0 such that 0 < |w | < = |w ||f (w)| <
1 Now choose z with 0 < |z | < 2 . We apply the Cauchy estimate on the disk 1 {w C; |w z| 2 |z |} V to get

|g (z)|

2 |g(w)| sup |z | |wz|= 1 |z|

2

2 |w |2 |f (w)| sup |z | |wz|= 1 |z|

2

3 3

sup
1 |wz|= |z| 2

|w ||f (w)|

82

3 3 since 0 < 1 |z | |w | 2 |z | 4 . it follows that g is continuous at 2 and completes the claim.

But now we may expand g as a power series g(z) =

n=0

an (z )n . Since

g() = g () = 0, it follows that a0 = a1 = 0 and therefore

f (z) =
n=0

an+2 (z )n

for z V . C OROLLARY 66 If f is holomorphic in V and tends properly to at , then f has a pole at . Proof. Consider g(z) = 1 dened on W = {z V ; f (z) = 0}. Then g is f (z) holomorphic on W and tends to zero as z . Hence g extends to a holomorphic function g on W = {} {z V ; f (z) = 0} which is a neighbourhood of . It follows that f has a pole at .

Since g has a zero at and g is not identically zero in a neighbourhood of , it follows that g has a zero of order m at where m is an integer m 1. This number is called the order of the pole . In case m = 1, we say that is a simple pole or single pole . If m = 2, we say that is a double pole . If f has a pole of order m at , then z (z )m f (z) has a removable singularity at . Also, if f has a pole of order m at we can express f (z) in terms of an expansion

f (z) =
n=m

an (z )n

valid in 0 < |z | < for some > 0. Such an expansion is called a Laurent expansion . It will turn out that if f has an essential singularity at , then it has a Laurent expansion of the form

f (z) =
n=

an (z )n

but this expansion does not terminate as we progress backwards through the negative powers. 83

sin(z) . Then ofcially, f is undened at z = 0, but is z holomorphic on C \ {0}. It is easy to see that 0 is a removable singularity and indeed, we have 1 f (z) = (1)k z 2k (2k + 1)! k=0 E XAMPLE Let f (z) = with innite radius. E XAMPLE Let f (z) = 2 z . Then ofcially, f is dened and holoexp(z) 1 z morphic on {z C; exp(z) 1 z = 0}. At every point = 0 where exp() 1 = 0, there is a simple pole at . This is because z exp(z) 1 z has a simple zero at . We can tell that the zero is simple, because if not then both exp(z) 1 z and its derivative exp(z) 1 both vanish at z = and this can only happen if = 0. At = 0, we see that z exp(z) 1 z has a double zero and z z has a simple zero. Hence f also has a simple pole at z = 0. 2 1 , a holomorphic function on C \ {0}. It is z fairly clear that f has an essential singularity at z = 0. To prove this rigorously, let y be real with |y| large and solve the equation z + z 1 = iy, or z 2 iyz + 1 = 0. 1 The sum of the two roots is iy so at least one of them has absolute value 2 |y|. The product of the two roots is 1, so the other root has absolute value 2|y| 1 . As iy tends to , the smaller root tends to zero and the value of f at the smaller root is cos(y) + i sin(y) which does not converge to anything, nor does it properly diverge to . 2 E XAMPLE Let f (z) = exp z +

We can also discuss the concept of singularities at . We are thinking of the Riemann Sphere and using the chart z z 1 which acts as the chart mapping near . So, f has a removable singularity (respectively pole) at if the function z f (z 1 ) has a removable singularity (respectively pole) at 0. We see that such a function has an expansion
m

f (z) =
n=

an z n

with m an integer m 0 in the case of a removable singularity and m 1 in the case of a pole (in which case, m is the order of the pole). A function holomorphic on C except at nitely many points (if the singularities of f are isolated, this will force them to be nite in number by compactness) 84

where there are poles is necessarily a rational function (i.e. quotient of polynomials). To see this, let f be such a function. Now construct a polynomial q whose zeros match the poles of f in C. If a pole of f has order m, then we insist that the corresponding zero of q has order m. Then p = qf has no poles in C (they have become removable singularities). At innity, it has a pole of order m where m is the order of the pole of f at plus the degree of q. It now follows that there is a constant C such that |p(z)| C(1 + |z|)m . Now apply Proposition 44 to show that p is a polynomial. 7.1 Laurent Expansions

Laurent expansions are more general than the expansions introduced earlier. They apply to holomorphic functions dened in annuli. For convenience, we will take our annuli centred at 0. T HEOREM 67 Let 0 r < R and let be the annulus dened by r < |z| < R. If f is holomorphic in , then we may nd a n C for n Z such that

f (z) =
n=

an z n

(7.1)

where convergence is uniform on the compacta of . Furthermore, we have

an = 1 2n
2

f (ei )ein d
=0

(7.2)

for all with r < < R.

Proof. Let z0 be xed. Choose r1 and R1 such that r < r1 < |z0 | < R1 < R. We start with a small circular contour of radius s oriented anticlockwise and centred at z0 and lying in . By the Cauchy Integral Formula, we have f (z) = 1 2i f () d z

for all z such that |z z0 | < s. Now consider the following contour, 1 . Let z = ei . We start at R1 ei make an anticlockwise loop around |z| = R1 , 85

9.0

6.0

3.0

-9.0

-6.0

-3.0

0.0 0.0

3.0 z

6.0

9.0

-3.0

-6.0

-9.0

Figure 7.1: Contour used in the proof of Theorem 67. returning to R1 ei , then along the straight line path to r1 ei then a clockwise loop around |z| = r1 , returning to r1 ei and nally back along the straight line path to R1 ei . It is fairly clear that and 1 are homotopic in \ {z; |z z0 | 1 1 s}. So again for |z z0 | < 2 s, we get 2 f (z) = = 1 2i 1 2 1 2
1

2 0

f () d z

since the integrals over the straight line portions of 1 cancel

2 0

R1 ei f (R1 ei ) d R1 ei z f (R1 ei ) d + 1 1 R1 ei z
2 0

2 0

r1 ei f (r1 ei ) d r1 ei z

= =

z 1 r1 ei f (r1 ei ) d 1 z 1 r1 ei

an z n
n= 2

where an =

1 n 2R1

f (R1 ei )ein d
=0

86

2 1 f (r1 ei )ein d an = n 2r1 =0 for n < 0. If M is an upper bound for |f | on r1 |z| R1 , then we have n n |an | R1 M for n 0 and |an | r1 M for n < 0 and the series converges uniformly on the compacta of r1 < |z| < R1 . So, the series converges to a holomorphic function g on the annulus r1 < |z| < R1 by Corollary 54. Also f 1 and g agree on |z z0 | < 2 s and therefore, they agree on r1 < |z| < R1 . So (7.1) holds for r1 < |z| < R1 . Since we can choose r1 as close to r (but with r1 > r) as we please and R1 as close to R (but with R1 < R) as we please, this gives that (7.1) holds on r < |z| < R and also that the series converges uniformly on the compacta of r < |z| < R. We can also see that for all n Z,

for n 0 and

an =

1 2i

f (z) dz z n+1

with the integral taken round |z| = anticlockwise is independent of so long as r < < R. This is because the circular paths corresponding to different are homotopic and the integrand z n1 f (z) is holomorphic in r < |z| < R. The formula (7.2) shows that the coefcient of the Laurent expansion are related to the Fourier coefcient of the function restricted to a circle. The uniqueness assertion is straightforward. Suppose that the series (7.1) converges uniformly on compacta to zero. Then we obtain by uniform convergence of integrals that for any with r < < R
2 N

0 = lim as required. E XAMPLE f (z) =

an n ein
0 n=N

k eik d = ak

Here is a simple example. Let 1 1 1 1 (z 1)1 (z + 2)2 (z + 2)1 2 = 9 3 9 (z 1) (z + 2) 1 1 1 (1 z)1 (1 + z/2)2 (1 + z/2)1 9 12 18 87

In the region |z| < 1 we write f (z) =

= giving

1 9

n=0

zn

1 12

n=0

(1)n (n + 1)2n z n

1 18

(1)n 2n z n
n=0

1 1 1 an = (1)n (n + 1)2n (1)n 2n 9 12 18 for n 0 and an = 0 for n < 0. For the region 1 < |z| < 2, the series expansion for (1 z)1 is no good, we need to use f (z) = 1 1 z 1 z 1 9
1 n 1

1 1 (1 + z/2)2 (1 + z/2)1 12 18 (1) (n + 1)2

n n n

1 = 9 giving

1 z 12 n=

n=0

1 z 18

(1)n 2n z n
n=0

an = for n 0 and an = f (z) = =

1 9

1 1 (1)n (n + 1)2n (1)n 2n 12 18 for n < 0. Finally, for 2 < |z|, then we write
1

1 1 z 1 z 1 9 1 9

1 2 z 1 + 2z 1 3

1 1 z 1 + 2z 1 9

n=1

z n

1 1 2 (1)n (n + 1)2n z n z 1 (1)n 2n z n z 3 9 n=0 n=0

giving an = 0 for n 0 and an = 1 1 1 + (1)n (n + 1)2n + (1)n 2n 9 12 18

for n < 0. This last expansion is a little bit misleading since the rst two terms vanish. Numerically, it looks like z 3 3z 4 + 9z 5 23z 6 + 57z 7 135z 8 + 313 z 9 + 2

88

E XAMPLE Find the Laurent expansion of cosec(z) in 0 < |z| < . since sin has a simple zero at z = 0, cosec has a simple pole. We can use the identity cosec(z) = cot z cot(z) 2

and the expansion for cot (from one of the assignments) to get

cosec(z) = z 1 +
k=1

(1)k

(4k 2)B2k 2k1 z (2k)! 2

7.2

Residues and the Residue Theorem

First, lets state a more powerful version of the Cauchy Integral Theorem. We could have used this theorem in the proof of the existence of Laurent expansions.

Let be a bounded connected open subset of C with piecewise T HEOREM 68 smooth boundary. Let denote the oriented boundary of . Let and f continuous on cl() holomorphic in . Then
f (z)dz = 2if () z

Proof. The idea is to cut a small closed disk D out of . Let U = \ D. Then U = D. We apply the Greens Theorem version of Cauchys Theorem to get f (z)dz =0 U z since z f (z) is holomorphic in U . This gives z

f (z)dz = z

f (z)dz = 2if () z

from Theorem 36. 89

D EFINITION Let be an isolated singularity of a function f holomorphic in a punctured neighbourhood of . Then f has a Laurent expansion about . The coefcient of (z )1 in this Laurent expansion is called the residue of f at and will be denoted Res(f, ). z . Then f has a simple pole at 0. This makes it 1z easy to calculate the residue as limz0 zf (z) = 2. 2 E XAMPLE Let f (z) = ez E XAMPLE The case of poles of higher order is trickier. For example let g(z) = z . Then we will expand the denominator far enough 1 ez 1 z 2 z 2 z g(z) = 1 3 1 4 z + 24 z + 6

1 = 6z 2 (1 + z + )1 4 1 = 6z 2 (1 z + ) 4 3 = 6z 2 z 1 + 2 3 and the residue is seen to be 2 .

E XAMPLE Even in case of an essential singularity, the concept of residue is still valid. For example, if h(z) = exp(z + z 1 ) we can calculate the residue as an integral Res(h, 0) = 1 2
2

exp(ei + ei )d =
=0 k=0

1 = I1 (2) k! (k + 1)! 2

where I1 is one of the Bessel function family.

1

ez a E XAMPLE Find the residue of at z = 0. Clearly, is analytic in a 1z 1z neighbourhood of z = 0. This means that the desired residue is also the residue 1 ez a of at z = 0 and we choose a = e to kill the singularity at z = 1. Then, 1z with w = z 1 , we have ez e ew e =w 1z w1
1

(7.3)

90

ew e is entire after we have resolved the removable singularity at w1 e0 e = e 1, this is also the coefcient of w in w = 1. The value at w = 0 is 01 (7.3) and hence the desired residue. 2 and w ez at z = 0. We need the coefcient of w (1 z)3 ew . So with f (w) = ew , we have in the Laurent expansion of w 3 (w 1)3
1

E XAMPLE

Find the residue of

f (1) + f (1)(w 1) + 1 f (1)(w 1)2 = 1 e(w 2 + 1). 2 2 So, since the singularity of f (w) (f (1) + f (1)(w 1) + 1 f (1)(w 1)2 ) 2 is (w 1)3 removable at w = 1 and the resulting function is entire, the coefcient of w in w
3 f (w)

(f (1) + f (1)(w 1) + 1 f (1)(w 1)2 ) 2 (w 1)3 z2 + 1 . But z 2 (1 z)3

1 is zero. Hence the desired residue is also the residue of 2 e

z2 + 1 = z 2 + 3z 1 2(z 1)3 + 2(z 1)2 3(z 1)1 2 (1 z)3 z 3e . 2 2 Before tackling the Residue Theorem, we need the version that applies to a single singularity and a small circle centred at the singularity. and the desired residue is P ROPOSITION 69 Let f be continuous on the punctured disk 0 < |z | r and holomorphic in 0 < |z | < r. Then f (z)dz = 2i Res(f, ),

where denotes the circle |z | = r traversed anticlockwise.

91

Proof. If we had taken (s) the circle |z | = s traversed anticlockwise with 0 < s < r, then f (z)dz = 2i Res(f, )
(s)

would follow from the Laurent Expansion Theorem. Its enough to show that f (z)dz f (z)dz

(s)

as s r and this is any easy consequence of the fact that f is uniformly continuous on {z C; 1 r |z| r}. 2 T HEOREM 70 Let be a bounded connected open subset of C with piecewise smooth boundary. Let denote the oriented boundary of . Let f be continuous on cl() holomorphic in \ F , where F is a nite set of singularities. Then f (z)dz = 2i
F

Res(f, )

Since the set of singularities is nite, it follows that each singularity is isolated. Proof. The proof is similar to that of Theorem 68. For each F make a small closed disk D , centred at and such that all the disks are disjoint and contained in . Let U be the set obtained from excising these disks from . Then f is holomorphic on U and this yields f (z)dz =
F D

f (z)dz = 2i
F

Res(f, )

from proposition 69. To get to the more advanced versions of the Residue Theorem, we will have to look again at the winding number. P ROPOSITION 71 Let be a piecewise C 1 loop in C. Then for in the unbounded connected component of C \ , wind () = 0.

92

Proof. Since dz 1 2i z it is easy to see that | wind () | 0 as . But wind () takes integer values and is constant on the connected components of C \ . Hence the result. wind () = We now come to the winding number version of the Cauchy Integral Formula. T HEOREM 72 Let be an open subset of C. Let 1 , . . . , m be piecewise C 1 loops in such that
m

W ()

windk () = 0
k=1

C \ .
m k=1

Let f be holomorphic in . Then for z \

m k

k (7.4)

k=1

f (w) dw = 2i wz

windk (z) f (z)

k=1

Proof. Let f (w) f (z) wz (z, w) = f (z)

m

if z = w, if z = w.

then, it is easy to see that is continuous on (use power series expansions to establish continuity at points on the diagonal) and holomorphic in each variable separately. Let U = { C; / k , W () = 0}
k=1

an open subset of C containing the set { C; || > R} for R sufciently large and which by hypothesis satises U = C. Let m (z, w)dw if z , k=1 k g(z) = m f (w) dw if z U . wz k=1 k 93

We will need to check that this is well-dened. In case U we have

m k=1 k

f (w) (z, w) dw = wz

k=1

f (z) dw wz
m

= 2if (z)
k=1

windk () = 0

It is easy to see that g is everywhere holomorphic (i.e. entire) and also that g(z) 0 as z since we can use the denition for U for this purpose. Since g is a bounded entire function, Liouvilles Theorem (Theorem 42) asserts that g is constant and therefore zero. But then, repeating the above argument assuming that z \ m k , we get (7.4) as required. k=1 C OROLLARY 73 (W INDING N UMBER V ERSION OF C AUCHY S T HEOREM ) Let 1 be an open subset of C. Let 1 , . . . , m be piecewise C loops in such that
m

k=1

windk () = 0 C \ .
m k=1

Let h be holomorphic in . Then for z \

m k=1 k

k (7.5)

h(w)dw = 0

Proof. Choose z . Apply Theorem 72 with f (w) = (w z)h(w). Finally, we now come to C OROLLARY 74 (W INDING N UMBER V ERSION OF THE R ESIDUE T HEOREM ) Let be an open subset of C. Let 1 , . . . , m be piecewise C 1 loops in such that m
k=1

windk () = 0 C \ .

Let f be holomorphic in \ F , where F is a nite set of singularities not meeting m k=1 k . Then
m

f (w)dw = 2i
k=1 k zF

W (z) Res(f, z)

(7.6)

94

where W (z) has its usual meaning W (z) = of times that the contours wind about z .

m k=1

windk (z) as the total number

Proof. For each z F , we strip out a small disk Dz . These disks are chosen so small that they avoid the contours k , are contained in and avoid each other. We would like to replace by 1 = \ zF Dz . We have that f is holomorphic on 1 , but we have messed up the winding number condition. To x this, for each z F we introduce an additional loops Lz winding around each of the disk Dz W (z) times. The loop Lz is located in 1 so very close to Dz that it does not interfere with anything else. Applying Corollary 7.5 we now get
m

f (w)dw +
k=1 k zD Lz

f (w)dw = 0

Proposition 69 now gives f (w)dw = 2i Res(f, z) windLz (z) = 2i Res(f, z)W (z)

Lz

and (7.6) follows. 7.3 Method of Residues for Evaluating Denite Integrals

This section is all examples. sin x dx. We proceed by integrating a funcx 0 tion related to the given one around a contour. Ler r > 0 be very small and eiz and integrate around the R > 0 be very large. In this case, we choose f (z) = z contour comprising four sections E XAMPLE We wish to evaluate 1. Along the real axis from r to R. 2. Around the semicircle Rei from = 0 to = . 3. Along the real axis from R to r. 4. Around the semicircle rei from = to = 0. 95

Figure 7.2: Contour for

0

sin x dx. x

This is a C 1 loop and f has no singularities inside the loop. The only singularity of f is at z = 0 and the contour has winding number zero about z = 0. So, we obtain
R

0=
r R

eix dx + x eix dx + i x
R

eiRe iRei d + i Re

r R

eix dx + x
0

eire irei d i re
i

=
r

eiRe d
0

R r

eix dx i x

eire d eire d
0
i

= 2i
r

sin x dx + i x

eiR cos() eR sin() d i

irei

Obviously, we are going to let r 0+ and R . As r 0+, we have 1 uniformly in (since rei 0 uniformly in ) and so
0

eir cos() er sin() d

r0+

We also have
0

eiR cos() eR sin() d 0

R

96

but this is less obvious. To justify this we proceed by

0

eiR cos() eR sin() d

eR sin() d
0 1 R 2 1 R 2

d +
0 1

1 R 2

e
1 2)

R sin()

d +

1 R 2

2R 2 + eR sin(R Finally, this gives 2i

0

sin x dx = i x

and we conclude
0

sin x dx = . x 2 2

The next example is computationally more intensive. d . here we will put z = ei and 32 18 cos + 5 sin 2 0 integrate over the unit circle anticlockwise call it . We get dz = ie i d and it follows that d = iz 1 dz. So, the integral becomes E XAMPLE Find i = i
2

zdz 5 9z 3 + 32z 2 9z + 2 i 1 2 6 The roots of the denominator are at 6 + 12 i + 5 31 5 i 31, 5 + 12 i 5 5 5 2 2 1 31 + 5 i 31, 5 1 i, and 2 i. The rst and third roots listed are in the unit 5 5 circle and the second and fourth are outside it. Only the residues coming from the rst and third roots will contribute. So the integral is
5 4 2 iz

dz 5 z(32 9z 9z 1 2 iz 2 + 5 iz 2 ) 2

(2i)(i)

+ g () g ()

(217 + 13 31) 3270

5 where and stand for the rst and third roots and g(z) = 2 iz 4 9z 3 +32z 2 5 9z + 2 i. 2

97

The next example is useful in the theory of Fourier integrals and can be used to establish the Plancherel Theorem and Inversion formula for Fourier Integrals. E XAMPLE We start with the formula

e 2 x dx =

which is proved in the MATH 255 notes. Unfortunately, there does not seem to be any way to establish this with contour integrals. What we would like to evaluate is 1 1 2 1 2 x2 iux 2 e e 2 (x+iu) ei 2 u dx dx = e

for u real. We will show that

1

e 2 (x+iu) dx =
2

2 and it follows that

1
2

e 2 x eiux dx =
1
2

2ei 2 u

So here, we take f (z) = e 2 z and integrate around the rectangle from R to

-R + iu R + iu

-R

Figure 7.3: Rectangular contour for Fourier Integral example. R along the real axis, from R to R + iu along a vertical line, then from R + iu to R + iu horizontally and nally from R + iu to R vertically. Cauchys Theorem yields
R R

e 2 x dx +
0

e 2 (R+it) idt

R R

e 2 (x+iu) dx

u 0

e 2 (R+it) idt = 0

98

We estimate the second and fourth integrals by

u 0

e 2 (R+it) dt =
0

e 2 R e

1 iRt 2 t2 e dt

|u|e 2 u e 2 R

Both these integrals tend to zero as R while keeping u constant. Thus, letting R , we get
1 2 e 2 (x+iu) dx

e 2 x dx

as required. The application of this (with most of the details beyond the scope of this course) is to the Fourier integral of a function f dened on the line, continuous and with compact support say. f (u) =

eiux f (x)dx

is the Fourier transform of f dened for u real. Then for t > 0 1 2

|f (u)|2 e 2 t

2 u2

du = =

1 2

eiux eiuy e 2 t eiux eiuy e 2 t t1 e 2 t

1 1

2 u2

f (x)f (y)dxdydu duf (x)f (y)dxdy

1 2 1 = 2

2 u2

2 (xy)2

f (x)f (y)dxdy

Letting t 0+ we get Plancherels Theorem 1 2

|f (u)|2 du =

|f (x)|2 dx

1 2 2 1 using among other things, the fact that kt (x, y) = t1 e 2 t (xy) is a sum2 mability kernel on the line as t 0+. 2

99

xa dx for 1 < a < 1. The integral has to be (1 + x)2 0 treated as an improper integral at the upper limit and in case 1 < a < 0 also as an improper integral at the lower limit. The integrand looks like x a for small x, so the condition a > 1 is needed for the integral to converge near x = 0. The integrand looks like xa2 for large x, so the condition a < 1 is needed for convergence at innity. Now the standard way of approaching this is to integrate the function E XAMPLE Consider f (z) = za (1 + z)2

around the famous keyhole contour. Doing this depends on selecting a nonstandard branch of the power z a , namely one that has a cut along the positive real axis. The conceptual difculties that arise can be avoided by making a substitution in the original integral before considering how to embed the problem in the complex domain. Our rst step is to put x = et and then the desired integral becomes eat dt. t t e + 2 + e eaz and we integrate over the rectangular contour We now take f (z) = z e + 2 + ez from R to R along the real axis, from R to R+2i along a vertical line, then from R + 2i to R + 2i horizontally and nally from R + 2i to R vertically. In some obvious sense, this is equivalent to integrating the original function around the keyhole contour. The function f has its poles, where (1 + cosh(z)) = 2 cosh2 z has its zeros 2 and the only zero of 1 + cosh(z) inside our new contour is at i. To nd the residue, substitute z = i + w. We nd f (z) = ei(1+a) e(1+a)w (ew 1)2
1 = ei(1+a) e(1+a)w (w + 2 w 2 + )2

1 = ei(1+a) w 2 (1 + (1 + a)w + )(1 + 2 w + )2

and the residue is seen to be aei(1+a) . Thus, we get

R

= ei(1+a) w 2 (1 + aw + )

2iae

i(1+a)

=
R

eat dt + et + 2 + et 100

2 0

ea(R+is) ids eR+is + 2 + eRis

ea(t+2i) dt et + 2 + et

2 0

ea(R+is) ids eR+is + 2 + eRis

As R , the second and fourth integrals tend to zero giving

2iaei(1+a) =

eat dt et + 2 + et

ea(t+2i) dt et + 2 + et

which simplies to

2iaei(1+a) 2ia a eat dt = = ia = t + 2 + et 2ia ia e 1e e e sin(a) 2

xa dx for 1 < a < 1. The integral has to be 1 + x2 0 treated as an improper integral at the upper limit and in case 1 < a < 0 also as an improper integral at the lower limit. The integrand looks like x a for small x, so the condition a > 1 is needed for the integral to converge near x = 0. The integrand looks like xa2 for large x, so the condition a < 1 is needed for convergence at innity. Our rst step is to put x = et and then the desired integral becomes E XAMPLE Consider

eat dt. et + et

We now take f (z) =

eaz and we integrate over the rectangular contour from ez + ez R to R along the real axis, from R to R + i along a vertical line, then from R + i to R + i horizontally and nally from R + i to R vertically. The function f has its poles, where cosh has its zeros and the only zero of cosh eia/2 1 = 2 ieia/2 . inside our new contour is at i/2. The residue is i/2 e ei/2 Thus, we get
R

eia/2 =
R

eat dt + et + et
R a(t+i)

ea(R+is) ids eR+is + eRis

0

e dt et et 101

ea(R+is) ids eR+is + eRis

As R , the second and fourth integrals tend to zero giving

eia/2 = 2i 1 ieia/2 = 2 which simplies to

eat dt + et + et

ea(t+i) dt et + et

a eia/2 eat dt = = sec t + et ia e 1+e 2 2 2

7.4

Singularities in Several Complex Variables

The situation as regards isolated singularities in several variables is bizarre. First of all, we need to agree on what a holomorphic function is. We give two denitions, one minimal and the other maximal and in fact, they agree. We will leave this as an exercise. D EFINITION

Let Cd . Let f : C. Then

f is holomorphic if it is continuous and holomorphic in each variable separately. f is analytic if for each , there is a neighbourhood U of in and an C for n Z+d such that the series
nZ+d

an (z )n

converges unconditionally uniformly on the compacta of U to f (z). In the second denition, n is a multiindex and we interpret
d

(z ) =

j=1

(zj j )nj .

The result that we wish to prove here is the following.

102

P ROPOSITION 75 Let = {(z1 , z2 ) C2 ; 1 < |z1 |2 + |z2 |2 < 9}. Let f be holomorphic in . Then f extends to a holomorphic function in {(z 1 , z2 ) C2 ; |z1 |2 + |z2 |2 < 9}. In other words, the whole of the inner ball is a removable singularity! Proof. For |z1 |, |z2 | < 2, let us dene g(z1 , z2 ) = 1 2
2 0

f (2ei , z2 ) i 2e d 2ei z1

It is easy to see that g is continuous and holomorphic in each variable separately (use differentiations under the integral sign) on {(z 1 , z2 ) C2 ; |z1 | < 2, |z2 | < 2}. Now suppose that 1 < |z2 | < 2, so that z1 f (z1 , z2 ) is holomorphic on {z1 ; |z1 | < 9 |z2 |2 }. Then the Cauchy Integral Formula shows that f (z1 , z2 ) = g(z1 , z2 ). Now free z2 and x z1 such that |z1 | < 2. Then f (z1 , z2 ) = g(z1 , z2 ) for 1 |z1 |2 < |z2 | < 2 in case |z1 | 1 and for |z2 | < 2 in case |z1 | > 1 since it is already known to hold for 1 < |z2 | < 2. We have just shown that f and g agree where both are dened. It follows that we can fabricate a glued function h holomorphic on the union of the domains of denition. But this is just {(z1 , z2 ) C2 ; |z1 |2 + |z2 |2 < 9} as required. In the Proposition above, we have chosen the domain between balls or radius 1 and 3, but the same result is true for the domain between balls of radius r and R so long as r < R.

103

8
Variation of the Argument and Rouch s Theorem e

We will start with the path lifting lemma L EMMA 76 Let f : [0, 1] T = {z C; |z| = 1} be continuous. Then f possesses a continuous lift, i.e. a continuous mapping f : [0, 1] R such if (t) that f (t) = e for all t [0, 1]. Furthermore, if f (0) is specied (satisfying if (0) e = f (0)), then f is uniquely determined. Proof. It is clear that if f maps into a closed interval on T, then the result is obvious. This is because for every such interval K, there is an interval I in the line such that the map s eis is a bijection from I onto K continuous in both directions. We now handle the general case. Since f is continuous [0, 1] it is on uniformly continuous. So there exists > 0 such that f () < 2. So every closed interval of [0, 1] of length less than gets mapped onto an interval in T. We can therefore break up
n

[0, 1] =
k=1

k1 k , n n

where n < 1 and then for each k there will exist a continuous lift g k dened on k1 k , n such that eigk = f on this interval. Now n eigk (k/n) = f (k/n) = eigk+1 (k/n) and it follows that gk+1 (k/n) = gk (k/n) + 2mk , for some integer mk . We then adjust the lifts by setting new continuous lifts
k1

hk (t) = gk (t) 2

mj
j=1

k1 k , n n

104

which are coherent in the sense hk+1 (k/n) = hk (k/n). The gluing lemma allows the hk to be glued into a single continuous function f which satises the required conditions. The uniqueness assertion amounts to showing that if g : [0, 1] R is continuous eig = 1 and g(0) = 0, then g is identically zero. Since g takes values in 2Z and g is continuous, it must be constant. This is of course the start of some interesting topology. We say that a path connected metric space X is simply connected if every continuous loop in X is homotopic to a constant map. We will leave the following theorem as an exercise.

Let X be a path connected, locally path connected, simply conT HEOREM 77 nected metric space. Let f : X T be continuous. Then f possesses a continuous lift f : X R.
Before we can prove this result, we will need several lemmas. L EMMA 78 Let X be a simply connected metric space. Let , be two paths in X (i.e. continuous maps from [0, 1] to X ) such that (0) = (0) and (1) = (1). Then and are homotopic via a homotopy that respects the endpoints. Explicitly, there is a continuous map H : [0, 1] [0, 1] X such that H(0, t) = (t), H(1, t) = (t), H(t, 0) = (0) and H(t, 1) = (1) for all t [0, 1]. Proof. We make a loop by adjoining with the reversal of , explicitly (t) = (2t) (2 2t)
1 if 0 t 2 , 1 if 2 t 1.

1 Note that ( 2 ) is well-dened and that (0) = (1). Let K be the homotopy linking to a constant loop with value . We have K(0, t) = (t) and K(1, t) = for all t [0, 1] and also we have K(s, 0) = K(s, 1) for all s [0, 1] since this is a homotopy of loops. We now cut this mapping apart and reassemble it.

L(s, t) =

1 K(2s, 2 t) 1 K(2 2s, 1 2 t) 1 2

1 if 0 s 2 , 1 if 2 s 1.

The denitions agree for s = 1 0 s 2,

since K(1, t) = for all t [0, 1]. Also for

L(s, 0) = K(2s, 0) = K(2s, 1) = L(1 s, 0),

1 L(s, 1) = K(2s, 2 ) = L(1 s, 1).

105

One can check that H is continuous by using the glueing lemma. We note also that the choice of root r satisfying r(1r) = t in the rst case is irrelevant because if r is a root, then the other root is 1 r and we know L(r, 0) = L(1 r, 0). Similarly for the second case. We have H(s, 0) = L(0, 0) = K(0, 0) = (0) for all s [0, 1] and similarly H(s, 1) = (1) for all s [0, 1]. On the other hand H(0, t) = L(0, t) = K(0, 1 t) = (t) and H(1, t) = L(1, t) = K(0, 1 1 t) = 2 2 (t) for all t [0, 1]. L EMMA 79 Let and be homotopic paths in T (via a homotopy that respects the endpoints). We are assuming that (0) = (0) and (1) = (1). Let and be lifts to R such that (0) = (0). Then (1) = (1). Sketch proof. If and are uniformly close, say |(t) (t)| < 2 for all t [0, 1], then |1 (t)1 (t)| < 2 or ((t)1 (t)) > 0. It is then easy to 1 lift this path to a path in ] 2 , 1 [ in R. The lifted path therefore takes the value 2 zero at t = 1. The uniqueness assertion of Lemma 76 shows that (1) = (1). In general, and are not uniformly close. Let H : [0, 1] [0, 1] T be the homotopy between them satisfying H(0, t) = (t), H(1, t) = (t), H(t, 0) = (0) and H(t, 1) = (1) for all t [0, 1]. Then H is uniformly continuous and k for k = 1, 2, . . . , n, the paths t H( k1 , t) and t H( n , t) are uniformly n close. The result follows. Sketch proof of Theorem 77. Let f : X T. Fix a point x0 X. Let 0 = f (x0 ) T and nd t0 R such that 0 = eit0 . Now for every x1 X, nd a continuous path from x0 to x1 , possible since X is path connected. Then f is a path in T which therefore has a unique lift f such that f (0) = t0 . Then we dene f (x1 ) = f (1). Now if is another continuous path from x0 to x1 , then by Lemma 78, and are homotopic via a homotopy that respects the endpoints. The same is then 106

Unfortunately L does not satisfy the necessary requirements, because for s L(s, 0) and s L(s, 1) are not constant. We x this by waisting L. That is, we dene L(r, 0) for 0 t s(1 s) and where r(1 r) = t, L(r, 1) for 0 1 t s(1 s) and where r(1 r) = 1 t, H(s, t) = L(s, r) for s(1 s) t 1 s(1 s) and where (1 r)s(1 s) + r(1 s(1 s)) = t.

true for f and f . It then follows from Lemma 79 that f (1) = f (1). This shows that f (x1 ) is independent of the path chosen. It remains to show that f is continuous. Let > > 0, then the open neigh bourhood ]f (x1 ) , f (x1 ) + [ of f (x1 ) lives above a interval neighbourhood V of f (x1 ) in T. Since X is locally path connected and f is continuous, we can nd a path connected neighbourhood U of x1 in X such that f (U ) V . Let x2 U . Let be a continuous path from x1 to x2 lying entirely in U . Then we make a continuous path from x0 to x2 by adjoining and to produce a new path = . Then f = (f ) (f ). Then f = (f ) (f ) and we can compute f directly by lifting from V to ]f (x1 ) , f (x1 ) + [. Thus This shows that f is continuous. 8.1 Meromorphic Functions f (x2 ) = f (1) = f (1) ]f (x1 ) , f (x1 ) + [

Here is another denition. D EFINITION Let be a connected open subset of C and f : C . Then f is meromorphic if it is not identically innite and is holomorphic as a map from to C in the sense of complex manifolds. Thus f is allowed to have isolated singularities (they cannot accumulate in since then 1/f would have a non-isolated zero and f would be identically innite). Each isolated singularity is necessarily a pole corresponding to the zeros of 1/f . Both the zeros and poles of f can accumulate on the boundary of . 8.2 Variation of the Argument

T HEOREM 80 (VARIATION OF THE A RGUMENT ) Let be a connected open subset of C and f : C be meromorphic. Let be a closed piecewise C 1 loop in not passing through any zero or pole of f . Then 1 2i f (z) dz = f (z) wind () m()

where the sum on the right is over the zeros and poles of f . If is a zero, then m() is the order of as a zero of f . If is a pole, then m() is the order of as
107

a pole of f . Since wind () = 0 for all in the unbounded component of C \ , it follows that all but nitely many terms in the sum vanish.
f (z) is holomorphic except at the zeros and poles of f (z) f . If is such a zero or pole, we write f (z) = (z )m g(z) where g() = 0 and where g is holomorphic in a neighbourhood of . The we get locally near Proof. The function z f (z) m g (z) = + f (z) z g(z) g (z) is holomorphic in this neighbourhood. The result now follows g(z) from the Residue Theorem. and z This result is called the variation of the argument because of the way that 1 f (z) dz can be interpreted. Let f () be the C 1 loop in f () dened by 2i f (z) t f (z(t)) where is the loop dened by t z(t). Then a change of variables gives f (z) 1 1 1 dz = dw = windf () (0) . 2i f (z) 2i f () w C OROLLARY 81

Under the same hypotheses as Theorem 80, we have

wind () m() = windf () (0),

where the sum on the left is over the zeros and poles of f and is interpreted in the same way.

8.3

Rouch s Theorem e

It is now possible to establish Rouch s Theorem, which is a seat of the pants way e of counting zeros.

108

T HEOREM 82 Let be a connected open subset of C and f, g : C be holomorphic. Let be a closed piecewise C 1 loop in not passing through any zero of f . Further, suppose that |g(z)| < |f (z)| for all z . Then wind f () (0) = windh() (0) where h = f + g and consequently wind () mf () =
f 1 ({0}) h1 ({0})

wind () mh ().

Proof. We have h(z) = f (z) + g(z) = f (z) 1 + h(z) f (z) = 1+ g(z) f (z) >0 (1 + w) > 0. Now parametrize the sgn(h(z(t)) = ei(t) g(z) f (z)

where we have used the fact |w| < 1 = loop by t z(t). Then let sgn(f (z(t)) = ei(t) and we nd and

(ei(t) ei(t) ) > 0 and it follows from {s R; eis > 0} =

mZ

](2m 1 ), (2m + 1 )[, 2 2

the fact that is continuous and the Intermediate Value Theorem that there is 1 an integer m such that (2m 2 ) < (t) (t) < (2m + 1 ) for all t. It follows 2 that 2 < ((1) (1)) ((0) (0)) < 2. But the quantity in the middle is an integer multiple of 2 since f (z(0)) = f (z(1)) and h(z(0)) = h(z(1)) and so it follows that windf () (0) = windh() (0) as required. E XAMPLE Let h(z) = 10 + 7z 2 + 2z 3 . Then on |z| = 1 we take f = 10, g = 7z 2 + 2z 3 , we get |f | = 10, |g| 9 < 10. So h has no zeros in |z| < 1. Then, on |z| = 2 we take f = 7z 2 and g = 10+2z 3 , we get |f | = 28 and |g| 26 < 28. So h has exactly two zeros in |z| < 2. Then, on |z| = 4 we take f = 2z 3 and g = 10 + 7z 2 , we get |f | = 128 and |g| 122 < 128. So h has exactly three zeros in |z| < 4. 109

Using other facts one can say considerably more. The single root in 2 < |z| < 4 must be real since otherwise there would be at least two roots with the same radius. Clearly this root is negative and you can locate it easily with a search as lying between -3.84 and -3.83. The product of all roots is positive, so if the remaining roots are real, then one is positive and one is negative. This is impossible, since h is increasing on the positive axis and h(0) = 10. So the it is a complex conjugate pair of roots a + ib that lie in 1 < |z| < 2. Since the sum and product of all three roots are known, it is easy to nd approximate locations. 2 E XAMPLE Consider the power series

f (z) =
n=0

3n z n

which clearly has innite radius and denes an entire function. Consider the 2 2 circle |z| = 9k . On this circle, |3k z k | = 3k . On the other hand

3
n=k

n2 n

3
n=k

(nk)2 k 2

3 2

k2

3m < 3k
m=1

It follows that for k = 0, 1, 2, . . . f has exactly k zeros in |z| < 9k . Further analysis shows that all the zeros are real and negative. 2 E XAMPLE Suppose that a and b are real constants such that |b| < 1 and |a b | < 1. We will show that h(z) = sin(z) (a + bz) has only one zero in 2 < z < , namely the one on the real axis. The idea is to apply Rouch s e 2 2 Theorem on a rectangular contour with sides corresponding to x = 2 and y = Y where Y is a large number. We take f (z) = sin(z) and g(z) = a + bz. We have |f (x + iy)|2 = (sin x)2 + (sinh y)2 and |g(x + iy)|2 = (a + bx)2 + b2 y 2 for x and y real. On a side of the form z = + iy with Y y Y , 2 it is easy to see by the hypotheses 1 + (sinh y)2 > (a b )2 + b2 y 2 . On 2 a side of the form z = x iY with x , it is easy to see that 2 2 (sin x)2 + (sinh Y )2 > (a + bx)2 + b2 Y 2 since for Y sufciently large, we will 2 have (sinh Y )2 b2 Y 2 > sup (a + bx)2 (sin x)2 ).
x 2 2

8.4

Hurwitzs Theorem

Hurwitzs Theorem can be obtained as a corollary of Rouch s Theorem. e 110

T HEOREM 83 Let be open in C and let (fk ) be a sequence of holomork=1 phic functions converging uniformly on compacta to a (necessarily holomorphic) function f . Let and suppose that f has a zero of nite order m at . Let > 0. Then there exists > 0 and K N such that < . {z; |z | < } For k K , fk has exactly m zeros in {z; |z | < }. Proof. We write f (z) = (z )m g(z) where g is holomorphic near and g() = 0. Now choose > 0 smaller than and such that |z | < = |g(z)| > 1 |g()|. Such a exists by the continuity of g at . Now choose K such that 2 k K = sup |f (z) fk (z)| < 1 m |g()| < inf |f (z)| 2
|z|= |z|=

Applying Rouch s Theorem now yields that f and f k have the same number of e zeros in {z; |z | < }. Hence the result. C OROLLARY 84 Let be connected open in C and let (fk ) be a sequence of k=1 one-to-one holomorphic functions converging uniformly on compacta to a (necessarily holomorphic) function f . The either f is constant or one-to-one. Proof. Suppose that f is not one-to-one. Then There exist z 1 = z2 in such that f (z1 ) = f (z2 ). Subtracting f (z1 ) from both f and fk we can assume without loss of generality that f (z1 ) = f (z2 ) = 0. Now if either z1 or z2 is a zero of innite order, then f is identically zero. We can therefore assume that they are zeros of nite order ( 1). Choose the of Hurwitzs Theorem to be half the distance from z1 to z2 . Then we obtain two disjoint disks D1 and D2 centred at z1 and z2 respectively and such that for k sufciently large, fk has at least one zero in both D1 and D2 . But this implies that fk is not one-to-one contrary to hypothesis. There is of course an inverse function theorem for suitably differentiable functions dened on subsets of Rd . The following version is particular to holomorphic functions and the proof uses the ideas of this section.

111

T HEOREM 85 (I NVERSE F UNCTION T HEOREM FOR H OLOMORPHIC F UNCTIONS ) Let be open in C, and f : C be holomorphic. Suppose that f () = 0. Then there is a neighbourhood U of , a neighbourhood V of f () such that f maps U onto V bijectively. Furthermore the inverse function g : V U C is holomorphic. Proof. By subtracting f () from f we can assume without loss of generality that f () = 0, and in view of f () = 0, is a single zero of f . As in the proof of Hurwitzs Theorem, we can nd > 0 such that = inf |z|= |f (z)| > 0. Let V = {w; |w| < 1 }. Then V is certainly open in C. For w V , the 2 function z f (z) w has a single zero in {z; |z | < } again as in the proof of Hurwitzs Theorem. Let this single zero be designated g(w). Then U = {z; |z | < } f 1 (V ) is open in and f is one-to-one on U . Now x w V and consider 1 2i

zf (z) dz f (z) w

(8.1)

where is the circle |z | = traversed anticlockwise. The integrand has a single simple pole at z = g(w) and we compute the residue by the usual method for simple poles Res g(w)f (g(w)) zf (z) , g(w) = = g(w) f (z) w f (g(w))

Therefore (8.1) evaluates to g(w). To show that g is holomorphic, it sufces to differentiate under the integral sign. We obtain g =0 w g (w) = 1 2i

zf (z) dz (f (z) w)2

and we note that g (w) is clearly a continuous function of w for w V since f (z) 1 for z and w V are always separated by at least 2 . C OROLLARY 86 Let be open in C and f be a holomorphic function on which is one-to-one. Then f cannot vanish on . 112

Proof. We suppose the contrary, namely that there exists such that f () = 0. Then by Proposition 51, we can write f (z) = f () + (h(z)) m locally near where m is an integer m 2, h is holomorphic near , h() = 0 and h () = 0. According to Theorem 85, h is bijective in a neighbourhood of . Fix = e 2i/m . Hence, for r > 0 small, there exist z1 , z2 such that h(z1 ) = r and h(z2 ) = r. Clearly z1 = z2 . However, f (z1 ) = f (z2 ) contradicting the fact that f is one-toone.

113

9
Conformal Mapping

We start with the Schwarz Lemma , a simple result with far reaching consequences. L EMMA 87 Let f be a holomorphic mapping of the open unit disk to itself. Suppose that f (0) = 0. Then

(i) |f (z)| |z| for |z| < 1. Furthermore if equality |f ()| = || holds for a single with 0 < || < 1, then there exists C with || = 1 such that f (z) = z for all |z| < 1. (ii) |f (0)| 1. Furthermore if equality |f (0)| = 1 holds, then there exists C with || = 1 such that f (z) = z for all |z| < 1.
Proof. We can always write f (z) = zg(z) where g is holomorphic. Let 0 < r < |f (z)| 1. Then on |z| = r we have |g(z)| < < r 1 . Therefore by the maximum |z| principle |g(z)| r 1 for |z| r. Let r 1, then |g(z)| 1 for |z| < 1, and the result of (i) follows. If |f ()| = || for such that 0 < || < 1, then |g()| = 1. By Corollary 61, g is constant (it attains its maximum modulus at an interior point). Call the constant . Then using |f ()| = || again we see that || = 1. We have now f (z) = z as required. The assertion (ii) follows in the same way from the Cauchy Estimate for f (0) based on the values of f on |z| = r and by letting r 1. Clearly f (0) = g(0), so the case of equality is also handled similarly. 114

C OROLLARY 88 Let f be a bijective holomorphic mapping of the open unit disk onto itself, with f (0) = 0, then there exists C with || = 1 such that f (z) = z for all |z| < 1. Proof. Let h be the inverse mapping. By Corollary 86, f does not vanish on the open unit disk. Hence, by Theorem 85, h is holomorphic and also a bijective mapping of the open unit disk onto itself. Applying the Schwarz Lemma 87, we get |f (z)| |z| and |h(w)| |w| for all z and w in the open unit disk. Putting f (z) w = f (z), we nd |f (z)| = |z| for |z| < 1. Now let g(z) = , then g is z holomorphic with constant modulus. We differentiate |g(z)| 2 = g(z)g(z) with respect to z to get 0 = g(z) g g (z) + g(z) (z) = g(z)g (z) z z

If both g and g are not identically zero on the open unit disk, then both have a most countably many zeros and we have a contradiction. So at least one of these functions is identically zero and either way, g is constant. it follows that there exists C with || = 1 such that f (z) = z for all |z| < 1. C OROLLARY 89 Let f be a bijective holomorphic mapping of the open unit disk onto itself, then there exists C with || = 1 and a C with |a| < 1 such that za f (z) = for all |z| < 1. 1 az

Proof. Let a be the point of the open unit disk that gets mapped to 0. Then dene a+w g(w) = f ((w)) where (w) = 1 + aw We remark that is a Mobius transformation preserving the open unit disk, so that g is a bijective holomorphic mapping of the open unit disk onto itself with g(0) = f (a) = 0. By Corollary 88, there exists C with || = 1 such that g(w) = w. Put za w= 1 az 115

and we discover that (w) = z (this is the inverse Mobius transformation) and hence za . f (z) = g(w) = w = 1 az This result tells that the group of bijective holomorphic mappings of the open unit disk onto itself (under composition) consists of the Mobius transformations that have the same property. Each Mobius transformation is associated with an invertible linear transformation of C2 via w1 where (z, 1) = (w1 , w2 ) (z) = w2 and we think of as acting on the one-dimensional subspaces of C 2 . The fact that preserves the open unit disk corresponds to the fact that preserves the lines that live in |w1 | < |w2 |, or what amounts to the same thing, the sign of the quadratic form |w2 |2 |w1 |2 . So, in our case, the matrix U of will be U= a a 1 0 1 .

If we let the matrix of the quadratic form |w2 |2 |w1 |2 be J= 1 0 a 1 .

We will have the matrix relation U JU = (1 |a|2 )J or in longhand a 9.1 a 1 1 0 0 1 a =

(1 |a|2 ) 0 0 (1 |a|2 )

Some Standard Conformal Maps

The really obvious ones are translations, rotations and scalings. Lets look at some of the less obvious ones. Usually, we are trying to map onto a disk or a region that we already know how to map conformally onto a disk. E XAMPLE The halfspace. We can map the right halfspace {z C; z > 0} onto the open unit disk by the Mobius transformation f (z) = z1 z+1 2 116

E XAMPLE We can map a sector with opening angle 2 (with < ) say {z C; < arg(z) < } onto the right halfspace with f (z) = z 2 . This is the principal branch that is intended here, i.e. f (rei ) = r 2 ei 2

The same transformation can also be used to a sector {z C; < arg(z) < , |z| < 1} to the intersection of the unit disk and the right halfspace {z C; z > 0, |z| < 1} 2 E XAMPLE The intersection of the unit disk and the right halfspace {z C; z > 0, |z| < 1} can be mapped conformally onto the rst quadrant {z C; z > 0, z > 0} by z+i f (z) = zi Note that i gets mapped to 0 and i gets mapped to the point at innity. The rst quadrant is a sector so that can then be mapped conformally onto a halfspace. More generally, the eye-shaped region between two circles can be mapped to a sector by z f (z) = z where and are the points where the circles intersect. This is a Mobius transfor mation, so that the angle of the resulting sector is the same as the angle between the two given circles. 2 E XAMPLE A strip can be mapped to a sector by the exponential mapping. For example, the strip {z C; a < z < b} can be mapped to a sector by f (z) = e icz with c real and suitably chosen. The region given by {z C; 0 < z < 1, z > 0} can be mapped to a sector 2 by f (z) = ei 2 z Finally lets look at examples with slits. The general strategy is: rst move the slit to where you want it (the ends should be either at 0 or at or both) and second open the slit by taking a square root. Let 1 < a < 1. Consider the unit disc with the segment of the real za axis from 1 to a removed. First use the Mobius transformation f (z) = 1 az which preserves the unit disk and will reduce to the case a = 0. Then the principal square root will map onto {z C; z > 0, |z| < 1}. 2 E XAMPLE 117

r > 0, < < .

The nal example is C \ [0, 1]. Map rst with the Mobius transfor z which maps onto C \ [, 0]. Then take the principal mation f (z) = 1z square root to map onto the right halfspace. 2 E XAMPLE 9.2 Montels Theorem

The objective of this section is to apply the AscoliArzela Theorem in the context of spaces of holomorphic functions. There is an advanced version of the Ascoli Arzela Theorem that applies more directly in the present context, but we will make do with the version established in MATH 354. T HEOREM 90 (A SCOLI A RZELA T HEOREM ) Let K be a compact metric space. We denote by C(K) the space of bounded complex-valued continuous functions on K . Let F C(K). Then the following are equivalent statements. F has compact closure in C(K). F is bounded in C(K) and F is equicontinuous. We start with the following Lemma. L EMMA 91 Let K, L be a compact subsets of C, > 0 such that K + D() L C where is open in C. We have denoted D() the closed disk centred at 0 of radius . Let f be holomorphic in , then |f (z1 ) f (z2 )| 4 1 |z1 z2 | sup |f (z)|
zL

for z1 , z2 K.

Proof. In case that |z1 z2 | 1 , we simply use the estimate 2 |f (z1 ) f (z2 )| |f (z1 )| + |f (z2 )| 2 sup |f (z)|.
zL

So, we can assume that |z1 z2 | 1 . In that case, L contains the union of 2 the two disks centred at z1 and z2 respectively of radius and therefore L also 118

contains the disk centred at 1 (z1 + z2 ) of radius 3 . We apply the Cauchy Integral 2 4 formula on the boundary of that disk to get f (z1 ) f (z2 ) = = |f (z1 ) f (z2 )| 1 2i 1 2i 1 1 z z1 z z1 f (z)dz

1 |z1 z2 ||f (z)| ds 2 |z z1 ||z z2 | 3 1 1 (2 )|z1 z2 | sup |f (z)| 1 2 2 4 ( 2 ) zL

(z1 z2 )f (z) dz (z z1 )(z z2 )

since for z we have |z zj | 1 for j = 1, 2. Hence the result. 2 P ROPOSITION 92 Let be an open subset of C, F a family of holomorphic functions on . Suppose that for every compact subset of , the family F |K = {f |K ; f F }

is a bounded subset of C(K), then for every compact subset K of , F | K is equicontinuous in C(K).
Proof. For every n N we dene Kn = {z; distC\ (z) 1 , |z| n}. n

Then Kn is a closed bounded (and hence compact) subset of C contained in . On the other hand, if K is a compact subset of C contained in , then dist C\ attains its minimum value and z |z| attains its maximum value on K. It follows that there exists n N such that K Kn . If as a shorthand we denote Fn = F |Kn , then our hypothesis is that Fn is bounded in C(Kn ) for every n N and the desired conclusion is that Fn is equicontinuous in C(Kn ) for every n N. 1 However, Kn + D( n(n+1) ) Kn+1 and it follows immediately from Lemma 91 that the boundedness of Fn+1 in C(Kn+1 ) implies the equicontinuity of Fn in C(Kn ).

119

T HEOREM 93 (M ONTEL S T HEOREM ) Let be an open subset of C, F a family of holomorphic functions on . Suppose that for every compact subset of , the family F |K is bounded in C(K). Then every sequence (fn ) in F posn=1 sesses a subsequence that converges uniformly on the compacta of to a function holomorphic in . Montels Theorem is the HeineBorel Theorem of complex function theory. If we let H() be the space of holomorphic functions on with the convergence of uniform on compacta convergence (which can in fact be realised as a metric space), then every closed bounded subset of H() is compact. Proof. With the same notations as in Proposition 92, we use the AscoliArzela Theorem to extract a subsequence (fm(1,n) ) of (fn ) with (fm(1,n) |K1 ) n=1 n=1 n=1 converging uniformly on K1 . Then from that subsequence, a further subsequence (fm(2,n) ) of functions such that (fm(2,n) |K2 ) converges uniformly on K2 . n=1 n=1 Then from that subsequence, a further subsequence (fm(3,n) ) of functions such n=1 that (fm(3,n) |K3 ) converges uniformly on K3 . . . and so forth. Finally we see n=1 that the diagonal subsequence (fm(n,n) ) converges to a limit f uniformly on n=1 each Kn and hence on all the compacta of . The reason for this is that the sequence (fm(n,n) ) is in fact a subsequence of the sequence (fm(k,n) ) which n=k n=k is known to converge uniformly on Kk . Finally, let us note that the limiting function f is holomorphic in by Moreras Theorem. 9.3 The Riemann Mapping Theorem

T HEOREM 94 (T HE R IEMANN M APPING T HEOREM ) Let be an open connected proper subset of C which is simply connected. Let . Then there is a unique bijective holomorphic map f : {z C; |z| < 1} such that f () = 0 and f () > 0. The theorem stated under the additional assumption that the boundary of is piecewise smooth was established by Bernhard Riemann in 1851. The rst proof of the theorem in the generality given above is due to Constantin Carath odory in e 1912. The uniqueness assertion is easy. It follows directly from Corollary 88. It is clear that the whole of C is not conformally equivalent to the open unit disk, since the latter has non-constant bounded holomorphic functions and the former does not. 120

L EMMA 95 Let be a connected simply connected open subset of C and h : C \ {0} be holomorphic. Then h has a holomorphic square root. Proof. We have that is path connected, locally path connected and simply connected. The map h(z) z k(z) = |h(z)| is a continuous map from to T. Therefore, by Theorem 77, there is a lift k(z) and therefore a logarithm z (z) = ln(|h(z)|) + ik(z). Clearly h(z) = exp( (z)) and since the exponential map is locally invertible, we see that is holomorphic. A desired square root is given by exp( 1 (z)). 2 P ROPOSITION 96 It sufces to prove the Riemann mapping Theorem in the case where is a connected simply connected bounded nonempty open subset of C. Proof. Let a C \ . Then by the previous lemma, there exists a holomorphic mapping : C such that ((z))2 = z a. Then z1 , z2 , (z1 ) = (z2 ) implies that z1 = z2 so that is one-to-one. Now is non-constant and therefore () is open in C by the Open Mapping Theorem52. There is a small disk V centred at () contained in (). We will show that V is disjoint from (). Indeed, if z V and z = (z1 ) with z1 , then there also exists z2 such that z = (z2 ). But then z1 = z2 , z = 0, (z1 ) = 0 and z1 = a. This is a contradiction since z1 and a . Hence there is a whole disk disjoint / from (). It sufces to compose with a Mobius transformation to map conformally onto a bounded open subset of C. Since the resulting conformal transformation is continuous and has a continuous inverse, it is connected and simply connected (as well as being nonempty). P ROPOSITION 97 Let be a bounded nonempty open subset of C and let . Let F be the set of holomorphic functions from to the open unit disk such that f () = 0, f () > 0 and f is one-to-one. The F is nonempty. Let a = sup f () > 0.
f F

Then the supremum is attained.

121

Proof. Taking f (z) = (z ) for > 0 but sufciently small shows that F is nonempty. Let (fn ) be a sequence of functions in F for which the sup is approached. Then by Montels Theorem (fn ) possesses a subsequence converging uniformly on compacta to a holomorphic function f . We rename the subsequence to (fn ). Clearly f () = a since uniform on compacta convergence will also imply uniform on compacta convergence of the derivative. Also f () = 0 and f takes values in the closed unit disk. Now suppose that z1 , z2 and f (z1 ) = f (z2 ), we will produce a contradiction. Let D be a closed disk centred at z 2 with z1 D / and D . Then z fn (z) fn (z1 ) does not vanish on D. By Hurwitzs Theorem, since fn fn (z1 ) converges uniformly on compacta to f f (z1 ), either f is identically f (z1 ) on D or f f (z1 ) never vanishes on D. But z2 D and so it must be that f is identically f (z1 ) on D and hence also on . But then f () = 0 a contradiction. Hence f is one-to-one on . But now, from the Open Mapping Theorem, f () is open in C and hence f takes values in the open unit disk. So f F as required. Proof of the Riemann Mapping Theorem. We can assume that is bounded and with the notations of Proposition 97 let f be a function in F for which the supremum is attained. It sufces to show that f maps onto the open unit disk. Let C, || < 1 and suppose that f does not take the value . Then applying Lemma 95 again, there is a holomorphic function h in such that (h(z))2 = f (z) . 1 f (z) (9.1)

Clearly h takes values in the open unit disk and is one-to-one on . However, h does not necessarily map to 0. We therefore dene g(z) = |h ()| h () h(z) h() (9.2)

1 h()h(z)

and clearly g is one-to-one holomorphic from to the open unit disk and g() = 0. Differentiating (9.1) and setting z = yields 2h()h () = (1 ||2 )f () and processing (9.2) similarly gives g () = |h ()| h () (1 |h()|2 )h () (1 |h()|2 )2 122 = |h ()| . 1 ||

Therefore g () = 2 ||(1 ||) 1 ||2 f () = 1 + || 2 || f () > f (),

since || < 1. Hence g F and we have a contradiction with the denition of f .

9.4

Conformal maps between Annuli

In this section we let 0 < rj < 1 for j = 1, 2 and let Aj = {z C; rj < |z| < 1} be the corresponding annulus. We ask what conformal maps are possible from A 1 onto A2 . T HEOREM 98 Let 1 be a conformal map from A1 onto A2 , then r1 = r2 and there exists C with || = 1 such that either 1 (z) = z or 1 (z) = r1 z 1 . Sketch proof 1 . Let 2 be the inverse map to 1 . Let Sj = {z C; ln(rj ) < z < 0} be open strips in the complex plane for j = 1, 2. The exponential map takes Sj onto Aj but is not one-to-one. By Theorem 77 and the fact that strips are simply connected, we can construct a continuous map 1 : S1 S2 such that 1 exp = exp 1 . Since locally 1 = log 1 exp where log is some branch of the logarithm, 1 is holomorphic. Let s1 S1 . Let s2 = 1 (s1 ) S2 . Similarly, we may construct a holomorphic mapping 2 : S2 S1 such that 2 exp = exp 2 and additionally we may arrange that 2 (s2 ) = s1 . Clearly 2 1 (z) = z for all z near s1 and hence for all z S1 . Similarly, 1 2 (z) = z for all z S2 . Hence 1 is a conformal map from S1 onto S2 . We know how to classify these, but before we get into that aspect of the proof, let us consider 1 (z + 2i) 1 (z) which takes values in 2iZ. Since a continuous mapping from a connected space to a discrete one is necessarliy constant, there is an integer m1 such that 1 (z + 2i) = 1 (z) + 2m1 i for all z S1 . Similarly, there exists m2 Z such that 2 (z + 2i) = 2 (z) + 2m2 i for all z S2 . We now have (with some details omitted) z + 2i = 2 1 (z + 2i) = 2 (1 (z) + 2m1 i) = 2 1 (z) + 2m1 m2 i = z + 2m1 m2 i
1

With thanks to Paul Koosis for suggesting this approach.

123

forcing m1 m2 = 1 and therefore m1 = m2 = 1. Let H = {z C; z > 0}, the upper halfspace. The conformal mapping j : Sj H is given by j (z) = exp iz ln(rj ) .

Dening 1 = 2 1 1 we have a conformal map 1 of H onto H. We 1 2 2 > 1, so that j (z + 2i) = tj j (z) and it follows dene tj = exp ln(rj ) that 1 (t1 w) = tm1 1 (w) for all w H. Now by Corollary 89 and the standard 2 conformal equivalence between disk and halfspace, we see that 1 has the form 1 (w) = aw + b cw + d

with a, b, c, d real and ad bc > 0. We now obtain the identity aw + b at1 w + b = 1 (t1 w) = tm1 1 (w) = tm1 . 2 2 ct1 w + d cw + d We multiply out and equate the coefcients of 1, w and w 2 to get ac = tm1 ac, 2 adt1 + bc = tm1 (ad + bct1 ) and bd = tm1 bd. So, either a = 0 or c = 0 and either 2 2 b = 0 or d = 0. Two of the resulting four cases violate ad bc > 0. The remaining two are b = c = 0, t1 = tm1 and since t1 , t2 > 1, m1 = 1, t1 = t2 , 1 (w) = w 2 with > 0. a = d = 0, t1 = tm1 and since t1 , t2 > 1, m1 = 1, t1 = t2 , 1 (w) = 2 w 1 with > 0. Tracing this information back and using the fact that 1 preserves S1 (now known to equal S2 ) shows that 1 assumes one or other of the forms 1 (z) = z + i or 1 (z) = ln(r1 ) z + i,

with R and nally tracing back to 1 establishes the result. In the same vein, we can show the following theorem.

124

T HEOREM 99 Let , u be continuous functions from T to ]0, [ such that < u everywhere. Let be the open subset {rei ; (ei ) < r < u(ei )}. Then there exists s ]0, 1[ such that is conformally equivalent to the annulus {z C; s < |z| < 1}. Sketch proof. The proof follows a very similar line to the proof of Theorem 98. Let U = {z; exp(z) } and V = {exp(iaz); z U } where a > 0 is chosen sufciently small that a(ln(sup u) ln(inf )) < which forces V to lie in a halfspace. Now clearly, U is contractible and so is V , so by the Riemann Mapping Theorem, V is conformally equivalent to the upper halfspace H. There is some freedom for the choice of this mapping which we will exploit later. Let : V H be such a conformal equivalence. Now z U z + 2i U , so z V e2a z V . Since the only conformal transformations of H onto H are Mobius transformations, we have az + b (e2a 1 (z)) = cz + d for all z H and where a, b, c, d are real and ad bc > 0. Equivalently (e2a z) = a(z) + b c(z) + d (9.3)

for all z V . We now exploit the freedom in the choice of the mapping . This allows us to replace the matrix A= with the matrix a c b d

a b S c d where S is a matrix with real entries and positive determinant. Therefore, using the Jordan canonical form for real matrices, we can are arrange that the matrix A has one or other of the special forms S 1 0 0 , , R, > 0

1 , R \ {0} 0 , , R, = 0 125

Thus (9.3) can always be rewritten in one or other of the following forms (i) (e2a z) = (z) with > 0. (ii) (e2a z) = (z) + with R \ {0}. (iii) (e2a z) = (z) with R \ {0}. (z) + 1

We can eliminate case (iii) immediately. Let z V be the point such that (z) = i. Then i (e2a z) = = i, i + 1 and since is one-to-one, e2a z = z resulting in z = 0 which is not correct since 0 V . / Next we consider case (ii). We dene (z) = 2i1 (exp(iaz)) and it can be shown that is a conformal map of U onto the right or left halfspace (according as > 0 or < 0) and it satises (z + 2i) = (z) + 2i. It therefore respects the exponential mapping and factors down onto a conformal map of onto either {z C; 0 < |z| < 1} or {z C; |z| > 1}. In either case we nd that is conformally equivalent to a punctured disk and it is easily seen that this is not the case. In case (i), we rst observe that = 1 for else is not one-to-one. Let us assume that < 1, then we choose b > 0 such that b1 = 1 ln() and dene 2 by the relation (exp(iaz)) = exp(ib(z)) The is a conformal map of U onto the strip b1 < z < 0. Furthermore (z + 2i) = (z) + 2i. Again respects the exponential map and factors down onto a conformal map of onto the annulus {z C; s < |z| < 1} where ln(s) = b1 (i.e. s = ). The case > 1 is similar, but it is necessary to ip the boundaries of the annulus.

126

10
Odds and Ends

10.1

The Schwarz Reection Principle

The Schwarz Reection Principle addresses the question of analytic continuation. Thats the situation in which a holomorphic function in one domain is extended to a holomorphic function in a larger domain. We have seen instances of this already in these notes and it is a very common theme in complex analysis. A systematic treatment however is beyond the scope of this course. Here we present just one theorem in this vein. T HEOREM 100 Let be an open subset of C with the property that z z . Let f be dened on {z ; z 0} and be continuous on that set, as well as being holomorphic in {z ; z > 0}. Suppose that z , z real implies that f (z) is real. Then f extends to a holomorphic function f : C. Proof. We dene f (z) = f (z) f (z) if z , z 0, if z , z 0.

The two denitions agree by hypothesis. We see that f is a continuous map from to C by the Glueing Lemma, the subsets {z ; z 0} and {z ; z 0} being closed subsets in the relative topology of . It remains to show that f is holomorphic. We do this using Moreras Theorem. We need to show that if T is a solid triangle contained in , then
T

f (z)dz = 0.

127

This is obvious (by Cauchys Theorem) if T is contained in either of the sets {z ; z > 0} and {z ; z < 0}, but is not immediately obvious in the case that the triangle straddles the real axis, or in fact even touches it. In this situation, the idea is to write the triangular path T as the sum of two closed polygonal paths P1 and P2 with P1 in the upper half space z 0 and P2 in the lower half space z 0. It will be enough to show that f (z)dz = 0 for j = 1, 2.
Pj

To see this for j = 1, we observe by Cauchys Theorem that

Pj +i

f (z)dz = 0

for > 0 since now the translated path Pj + i lives in the strict upper half space z > 0 where f is known to be analytic. But f (z)dz =
Pj +i Pj

f (z + i)dz

f (z)dz
Pj

and f is uniformly continuous on such a subset. It follows that and the case j = 2 is exactly similar. 10.2 The Gamma Function

as 0, since the values of z under consideration form a closed bounded subset

P1

f (z)dz = 0

The Gamma Function is dened by

(z) =
0

et tz

dt = t

et tz1 dt
0

for z > 0. The integral always converges at . The condition z > 0 is imposed to make it converge at t = 0. In the right halfplane is a uniform on
n

compacta limit of
n1

et tz1 dt as n , so is holomorphic in the right

halfplane. It is easy to prove by integration by parts and by induction that for n N and x>0 n n nx n! t tx1 dt = 1 n x(x + 1)(x + 2) (x + n) 0 We claim that e
t

2t2 1 n

128

t 1 n

et

(10.1)

for n N and 0 t n. The right-hand inequality boils down to 1 x e x for 0 x 1. The left-hand one is more subtle. Taking logs, we need gn (t) = n ln 1 for 0 t < t n + t ln 1 2t2 n 0

n , the inequality being obvious otherwise. Clearly g n (0) = 0, so it 2 will sufce to show that gn (t) 0. We have 4t t(3n 4t + 2t2 ) t + = 0 n t n 2t2 (n t)(n 2t2 ) in the desired range since 3n 4t > 3n 2 2 n > 2 2(n n) 0. From(10.1) it follows that gn (t) =
n

(x)

t n

tx1 dt

2t2 t x1 e t dt + n

et tx1 dt 0
n

for x > 0. It follows that (x) = lim again for x > 0. We set about dening nx n! , n x(x + 1)(x + 2) (x + n)

1 for all complex z by means of (z)

1 = zez (z) n=1 where is Eulers constant. For |w| |log(1 + w) w| and consequently log(1 +
1 2

1+

z z e n n

(10.2)

one has

1 1 (1)k1 w k |w|2 k 4 k=2 z z |z|2 ) 2 n n 4n 129

for n 2|z|. Thus, the product (10.2) converges uniformly on compact of C to an entire function. We nd for x > 0
N

xe

x n=1

1+

with the consequence that the two denitions of are both holomorphic and agree on the positive real axis. Therefore they agree everywhere on z > 0. 1 We note that has zeros only at 0, 1, 2, . . . and these are therefore the (z) only poles of . This is an example of analytic continuation. A function dened intitially only in a halfspace turns out to have a holomorphic extension to a much larger region. Next consider (z + 1) nz (n + 1)! = lim n z(z + 1)(z + 2) (z + n + 1) z = lim
n

n n+1

nz+1 n! z(z + 1)(z + 2) (z + n + 1)

= (z)

So that (z + 1) = z(z) (except where has its poles). From (10.2) we also have z2 1 = z 2 1 2 (z)(z) n n=1 or equivalently 1 z2 =z 1 2 (z)(1 z) n n=1

It turns out that we can relate the innite product on the right of (10.3) to the sin function. We will need the following lemma L EMMA 101

Let an,k C and suppose that limk an,k = an for all n N.

n=1

Suppose further that |an,k | Mn for all n and k and that

130

Mn < . Then

x x e n n

x(x + 1)(x + 2) (x + N ) x(+ln(N ) e Nx N!

N 1 n=1 n

(10.3)

(i) lim

an,k =
n=1 n=1

an .

(ii) lim

1 + an,k =
n=1 n=1

1 + an .

Proof. This is Analysis 2 stuff, well almost. Let > 0 Then since
k N

lim

n=1

an,k

n=1

an
n>N

n=1

|an,k an | + 2
1 4

Mn
n>N

we rst choose N so large that

Mn <

and then K so large that

|an,k an | <

1 2N

holds for n = 1, 2, . . . , N and k > K. We obtain

k

lim

n=1

an,k

an <
n=1

for k > K. This proves assertion (i). For assertion (ii) we rst choose N so large that Mn < enough to show that
k

1 2

for n > N . It will be

lim

1 + an,k =
n=N +1 n=N +1

1 + an .
1 2

Equivalently, we can assume without loss of generality that M n < now, using the principal branch of the logarithm we have | log(1 + an,k )| 2|an,k | 2Mn

for all n. But

1 using | log(1 + z)| 2|z| for |z| < 2 , and the result follows from (i) since exp is continuous.

131

P ROPOSITION 102

We have

z
n=1

z2 n2

= sin(z)

(10.4)

for all z C.
Proof. Let m = 2 be an even integer. Consider the polynomial Pm (z) = 1 2i 1+ iz m
m

iz m

and note that it has degree m1, since the term in z m will cancel when the brackiz ets are expanded, but the term in z m1 will not. Since lim m log(1 + = i m m (the left hand side is the derivative of z log(1 + iz) at z = 0), it follows that
m

lim mPm (z) = sin(z).

(As an exercise, show that this is a uniform on compacta limit). On the other hand, the roots of Pm satisfy m + iz m iz and hence have the form
m

= 1,

m + iz 1 = e2ikm m iz for k = 0, 1, 2, . . . , m 1. Solving for z gives z= m(e2ikm 1) m = tan 1 i(e2ikm 1)

1

k m

We now realise that k = does not yield a root and that we have our full complement of m 1 roots. We can therefore write
1

Pm (z) = Cz
k=1

2z2 m2 (tan(k/m))2

132

for suitable C. From the coefcient of z obtained from the binomial expansion, we see that C = . Hence
1

Pm (z) = z
k=1

2z2 m2 (tan(k/m))2

We now apply Lemma 101 with 2z2 2 2 an,k = 4k (tan(n/2k)) 0 an = z2 n2 Mn = to show that (10.4) holds. C OROLLARY 103 |z|2 n2

if n < k, if n k.

We have
1 sin(z) = (z)(1 z)

for all z C.

One can also point out that since ( 1 ) = 2

1 ( 2 )

t 2 et dt > 0, it must be that

1 2.

Putting t =

1 2 s 2

in the integral, this provides conrmation that

0

e 2 s ds =

. 2

133

Index
absolute value, 2 afne, 23 analytic, 8 argument, 3 AscoliArzela Theorem, 118 Cauchys Estimate, 52 Cauchys Theorem, 48, 94 Chain Rule, 26 complex conjugate, 2 complex plane, 1 conjugate harmonic, 35 cross ratio, 63 derivative, 22 differentiable, 22 differentiable at v0 , 24 differential, 22 direction vector, 28 directional derivative, 28 Dirichlet Problem, 72 double pole, 83 entire function, 56 essential singularity, 82 First Maximum Principle, 75 Fr chet derivative, 23 e Fundamental Theorem of Algebra, 56 Greens Theorem, 37 Inverse Function Theorem, 111 isolated singularity, 81 Jacobian matrix, 29 Laurent expansion, 83 Liouvilles Theorem, 56 Lipschitz at v0 , 25 Little o of the norm, 23 meromorphic, 107 modulus, 2 Montels Theorem, 119 Moreras Theorem, 68 Mobius transformations, 60 Open Mapping Theorem, 67 order of a zero, 54 order of the pole, 83 pole, 81 principal branch of the logarithm, 18 radius of convergence, 6 real analytic, 8 removable singularity, 81 residue, 90 Residue Theorem, 94 Riemann Mapping Theorem, 120 Riemann sphere, 64 Schwarz Lemma, 114 134

Second Maximum Principle, 76 simple pole, 83 simply connected, 105 single pole, 83 star shaped, 43 Summability Kernel Theorem, 74 Three Lines Theorem, 79 Variation of the Argument, 107

135