An Introduction to Complex Variables and Analytic Functions

INCOMPLETE MANUSCRIPT UNDER DEVELOPMENT

Bengt Fornberg and Cécile Piret

March 14, 2017

2
Contents

1 Complex Numbers 1
1.1 How to think about a complex number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Complex numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Complex number system: Algebraic introduction . . . . . . . . . . . . . . . . . . . . . 2
1.2.2 Complex number system: Graphical representation . . . . . . . . . . . . . . . . . . . . 3
1.2.3 Examples of polar rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Stereographic projection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.1 Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Functions of a Complex Variable 9

2.1 Derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 Cauchy-Riemann equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.2 Test whether or not a function is analytic . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Some elementary functions generalized to complex argument by means of their Taylor expansion 13
2.2.1 Relations between exponential and trigonometric functions . . . . . . . . . . . . . . . 14
2.2.2 Trigonometric functions represented in terms of the exponential function . . . . . . . . 14
2.2.3 Logarithm function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2.4 Inverse hyperbolic and and trigonometric functions . . . . . . . . . . . . . . . . . . . . 20
2.3 Additional observations on Taylor expansions of analytic functions . . . . . . . . . . . . . . . 23
2.3.1 Get the expansion from an already known expansion . . . . . . . . . . . . . . . . . . . 24
2.3.2 Method of undetermined coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.4 Singularities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4.1 Removable singularities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4.2 Poles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4.3 Essential singularities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4.4 Branch points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4.5 Clusters of poles or other singularities . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.4.6 Natural boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.5 Multi-valued functions - Branch cuts and Riemann sheets . . . . . . . . . . . . . . . . . . . . 29
2.5.1 Branch cuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.5.2 Riemann sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.6 Sequences of analytic functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.7 Definitions of functions by integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.8 Select proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.9 Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.9.1 MATLAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.9.2 Mathematica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.10 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

3
4 CONTENTS

3 Analytic continuation 43
3.1 Introductory examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.2 Some methods for analytic continuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.2.1 Circle-Chain method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.2.2 Schwarz reflection principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.2.3 Use of a functional equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2.4 Partitioning of an integration interval . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2.5 Replace Taylor coefficients by integrals or sums . . . . . . . . . . . . . . . . . . . . . . 49
3.2.6 Subtraction of a similar series or integral . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2.7 Borel summation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2.8 Padé approximations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2.9 Ramanujan’s formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.3 Select proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

4 Complex Integration 57
4.1 Integration when a primitive function F (z) is available . . . . . . . . . . . . . . . . . . . . . . 57
4.1.1 Cauchy’s and Morera’s theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.2 Contour integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.2.1 Cauchy’s argument principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.2.2 Cauchy’s integral formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.2.3 Liouville’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.2.4 The fundamental theorem of algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2.5 Mean value theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2.6 Revisiting Max/Min theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2.7 Revisiting Taylor expansions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.2.7.1 Weierstrass M-Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.2.7.2 Radius of Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.3 Laurent series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.3.1 Summary of Taylor versus Laurent series around a point z0 . . . . . . . . . . . . . . . 68
4.3.1.1 Taylor series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.3.1.2 Laurent series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3.2 Revisiting singularities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.4 Residue calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.4.1 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.4.2 Special contours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.4.3 Principal value integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.4.4 Rouche’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.5 Infinite sums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.6 Analytic continuation with use of contour integration . . . . . . . . . . . . . . . . . . . . . . . 94
4.6.1 General change of contour for the defining integral . . . . . . . . . . . . . . . . . . . . 95
4.6.2 Modification of the integration path to a Hankel-type contour . . . . . . . . . . . . . . 95
4.6.3 Modification of the integration path to a Pochhammer-type contour . . . . . . . . . . 96
4.7 Weierstrass products and Mittag-Leffler expansions . . . . . . . . . . . . . . . . . . . . . . . . 97
4.7.1 Weierstrass products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.7.2 Mittag-Leffler expansions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.8 Select proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.9 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
CONTENTS 5

5 Gamma, zeta, and related functions 105

5.1 The Γ- function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.1.1 Product representations of the Gamma function . . . . . . . . . . . . . . . . . . . . . 105
5.1.2 The Beta function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
5.2 The ζ- function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
5.2.1 The critical strip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.2.2 Relation to prime numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.3 Select proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

6 Elliptic functions 113

6.1 Some introductory remarks on simply periodic functions . . . . . . . . . . . . . . . . . . . . . 113
6.2 Some basic properties of doubly periodic functions . . . . . . . . . . . . . . . . . . . . . . . . 113
6.3 The Weierstrass ℘-function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6.3.1 Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6.3.2 Some nonlinear ODEs that are satisfied by the ℘-function . . . . . . . . . . . . . . . . 115
6.3.3 Closed form expressions for certain integrals based on the ℘-function . . . . . . . . . . 117
6.3.4 The modular function λ(τ ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
6.3.5 Picard’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
6.4 The Jacobi elliptic functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.4.1 The Jacobi sn(z,k ) function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.4.2 Some other Jacobi elliptic functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.5 Select proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

7 Conformal Mappings 125

7.1 Relations between conformal mappings and analytic functions . . . . . . . . . . . . . . . . . . 125
7.2 Mappings provided by bilinear functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
7.3 Riemann’s mapping theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.4 Mappings of polygonal regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.5 Re-visiting the Jacobi elliptic function sn(z ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
7.6 Some applications of conformal mappings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.7 Select proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.8 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

8 Transforms 135
8.1 Fourier transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
8.2 Laplace transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
8.3 Hilbert transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
8.4 z -transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
8.5 Select proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
8.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

9 Special functions defined by ODEs 141

9.1 Airy’s equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
9.1.1 Real-valued independent variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
9.1.2 Complex-valued independent variable . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
9.2 Bessel’s equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
9.3 Hypergeometric functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
9.4 Converting linear ODEs to integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
9.4.1 Fourier-Laplace method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
9.4.1.1 Airy’s equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
9.4.1.2 Bessel’s equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
9.4.2 Euler’s method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
9.4.2.1 Legendre’s equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6 CONTENTS

9.5 The Painlevé equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

9.5.1 The PI equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
9.5.2 The PII equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
9.5.3 The PIII equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
9.6 Select proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
9.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

10 Asymptotic Analysis of Integrals 153

10.1 Asymptotic vs. convergent expansions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
10.2 Laplace integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
10.2.1 Watson’s lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
10.2.2 Applications of Watson’s lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
10.2.3 Leading order approximations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
10.3 Steepest descent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
10.3.1 Concept and background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
10.3.2 Application of steepest descent to Laplace integrals . . . . . . . . . . . . . . . . . . . . 163
10.3.2.1 Integrals that become increasingly peaked . . . . . . . . . . . . . . . . . . . . 163
10.3.2.2 Integrals that become increasingly oscillatory . . . . . . . . . . . . . . . . . . 165
10.3.3 Steepest descent analysis of Ai(z) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
10.3.3.1 Ai(z) for z → +∞ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
10.3.3.2 Ai(z) for z → + − ∞ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

A Illustrations of Conformal Mappings 171

A.1 Elementary functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
A.2 Special functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
CONTENTS 7

Preface

The topics of Complex Variables and Analytic Functions are of fundamental importance throughout
applied mathematics and engineering. Their theorems and formulas also simplify many results from Calculus
and for functions of real variables. Not surprisingly, there is already a wide choice available of introductory
text books, raising the question why anyone would want to see yet another one. Our main answer lies in the
evolution that has occurred in other fields. For at least the last half century or so, it has been unthinkable
to use introductory text books √ for Calculus that do not graphically illustrate the basic elementary functions,
such as f (x) = x2 , f (x) = x, f (x) = sin x, f (x) = log x, etc. When we first taught a course on Complex
Variables and Analytic Functions, we became puzzled about why complex variables texts should not also
visually build on the real cases, familiar to all students, and then liberally illustrate how these same functions
extend away from the real axis. While formulas alone for some students might provide a feasible alternative
to visual impressions, we are here aiming the text for students that find the latter to be helpful for gaining
at least their initial intuitive feeling for the subject.
Another issue where text books differ is the order in which topics are covered. One strategy is to make
sure each step follows rigorously from previous steps. While that can have some appeal, it might not at all
be the best order for building up conceptual understanding as effectively as possible. We will instead focus
on ther latter, which might at times require issues to be re-visited once additional tools have fallen in place.
In either case, the end knowledge will be similar, but it is our belief that the latter approach is the one best
suited for an introductory text. A case in point concerns singularities, where we here bring these up early on,
before having developed all the tools of contour integration that are needed for a certain (Laurent-expansion
based) singularity characterization. To arrive at that stage, some earlier insights into singularities are very
helpful.
We have in this book never sacrificed correctness for simplicity. However, there inevitably arises situations
where it becomes unavoidable to choose a compromise path between the extremes of strictest mathematical
rigor (some might call nitpicking) and quite heuristic arguments. Believing that the former is better suited
for finalizing proofs than for gaining insights and for solving problems (and can be acquired when needed
from sources with that as focus), our preference here has generally been somewhat in favor of the latter.
The emphasis that is given here to different topics also differs somewhat from many other texts. For
example, we view analytic continuation as fundamental to a good understanding of the nature of analytic
functions. Instead of just listing one approach for this task (the circle-chain theorem, which incidentally is
quite impractical for computations), we have included numerous approaches. We also discuss and illustrate
multi-valued functions and their Riemann sheets more extensively than what is customary in an introductory
text, while we somewhat de-emphasize for example conformal mappings.
The present material originated from lecture notes first developed by Bengt Fornberg in 2006, and then
used jointly by us in 2008 (at University of Colorado, Boulder). After that, they were mostly put aside
until Cécile Piret revitalized them when resuming teaching the topic in 2015, at Michigan Technological
University.
As noted above, a main purpose of extending from the real axis to the complex plane is to greatly
simplify a vast number of tasks in applied mathematics, engineering, etc. The eminent mathematician
Jaques Hadamard expressed this succinctly (paraphrasing an earlier statement from 1900 by Paul Painlevé):
“The shortest path between two truths in the real domain passes through the complex
domain”
However, complex numbers go far beyond being just a mathematical trick that, once having done its
magic, ought to quickly and gracefully disappear. It is our hope that this book will make its readers not
only appreciate their utility, but also come to regard them as equally ‘natural’ as, say, negative integers.

Acknowledgments: ......
8 CONTENTS
Chapter 1

Complex Numbers

1.1 How to think about a complex number

At first (a few thousand years ago), the only numbers that were used were positive integers: 1,2,3,... Such
numbers were used to count with apples, oranges, days, frying pans, etc. Rational numbers, such as 3/7 or
11/8, are clearly useful in measuring things like distances and weights. These can still be associated with
counting (like pieces of a whole pie) - and they simplify division by now allowing all divisions (apart from
zero). An important point to note is that it is usually best not to think of a rational number as a pair of
two integers; even when represented as a ratio of two integers, the combination is usually best viewed as a
single number (and we use it as such in, say, representing a single position on the real axis); not as some
sort of a 2-component vector.
Negative numbers made originally little sense. However, much algebra - such as subtraction - became
much easier if negative numbers are allowed. Else there would need to be extra rules about when subtraction
is allowed and when it is not allowed. Also, there were many applications for which both the input data
and the final answer are positive, but when nevertheless going via negative numbers makes it easier to reach
the final answer, we have now given up on the idea that numbers need to directly correspond to counting
objects.
The Greeks, over 2000 years ago noted that even the rational numbers were insufficient to measure all
distances, such as the length of the diagonal of a square with side length one. This discovery was deeply
agonizing to them - there was great fear that the gods would avenge this human discovery of their imperfection
in creation, and the discovery was kept as a secret for quite some time.
The historical perspective above may be useful as a background when we now are extending the real
number system to complex numbers. There are no 3 + 5i apples, just as there are no −7 apples, so we
are already used to numbers being generalized past being merely counters for objects. Just like for rational
numbers, a complex number can be represented by means of a pair of numbers (and we graphically use
these two real numbers as coordinates in picturing a complex number’s location in the complex plane).
Nevertheless, it is often best to think about z = 3 + 5i as a single number, just as r = 3/7 is a single number
- and not as any sort of a two-component vector. This course is largely ´ about what happens to our usual
functions, such as y = sin(x) and to our calculus tools, such as d/dx, f (x)dx, etc... when we use z complex
instead of x real. Any other way of thinking of z than as a single number would for such work be very
awkward.
It is natural to wonder if one next should continue generalizing complex numbers to some kind of hyper-
complex ones, with more than two components. It can be shown that such attempts (of which there have
been many), will require very severe sacrifices - rules such as a · b = b · a and/or others that we want to take
for granted, will have to be abandoned, typically causing much greater losses than gains. Complex numbers
in many ways is THE most natural number system possible. Making excursions into the complex plane
often provides the simplest path also for problems where all ingredients as well as the final answers are all
real. Also, analytic functions (meaning functions w = f (z) for which df /dz exists) are in many ways much
simpler than real functions f (x). Just to mention a few aspects which we will encounter early on in this
book:

1
2 CHAPTER 1. COMPLEX NUMBERS

1. If a function f (z) can be differentiated once, it can be differentiated infinitely many times.
2. If an analytic function f (z) is defined uniquely on any interval, no matter how short, it is automatically
defined away from that line segment as well.
3. If the magnitude of an analytic function is everywhere bounded by some constant, the only possibility
is that the function is identically constant.
4. When a sequence of analytic functions converges point-wise to some limit, the convergence is always
uniform, and the limit is itself an analytic function.
5. A polynomial equation of degree n will always have exactly n roots.

The term complex is unfortunate since going to complex numbers very much simplifies much of calculus and
algebra. The complex number system is the most natural system in which to do most mathematics - the
other cases (such as integers, rational numbers, real numbers, etc.) are just restrictive subclasses, with often
more difficult rules.

1.2 Complex numbers

We recall that negative numbers were introduced in order to always make subtraction possible, with rather
immediate practical applications. In contrast, when complex numbers were first conceived (to have some
formal notation for all square roots and for solutions to all quadratic equations, such as x2 + 1 = 0), their
practical utility remained very unclear. A key step in advancing these complex numbers from being mostly
meaningless notational abstractions occurred when Girolamo Cardano (1501-1578) described a formula for
3
solving a general cubic equation. One √ such equation that he came to consider was z − 15z −√4 = 0, 1/3with
the three
√ roots z 1 = 4, z2,3 = −2 ± 3. However, his procedure gave one root as z1 = (2 + −121) +
(2 − −121)1/3 . On observing that, in our modern notation, (2 + i)3 = 2 + 11i and (2 − i)3 = 2 − 11i, he
simplified this to z1 = (2 + i) + (2 − i) = 4. For a purely real valued problem, a temporary excursion into the
world of complex numbers had thus produced a real-valued solution that no previously available systematic
approach had been able to reach. This is likely the first known example of the (much later formulated) quote
in the Introduction: ‘The shortest path between two truths in the real domain passes through the complex
domain’. √
Euler assigned the symbol i to the imaginary units satisfying i = −1. Not until around 1800 did the
complex number plane become recognized as a natural extension of the number line. The real axis becomes
then just a special case.

1.2.1 Complex number system: Algebraic introduction

Let i have the property that i2 = −1. Assume that all the other rules hold
1. z1 + z2 = z2 + z1 (commutativity in addition)
2. z1 z2 = z2 z1 (commutativity in multiplication)
3. (z1 + z2 ) z3 = z1 z3 + z2 z3 (distributivity)
4. (z1 z2 ) z3 = z1 (z2 z3 ) (associativity)
Next, let z1 = a + ib and z2 = c + id, where a, b, c, d are real. Then,
1. z1 ± z2 = (a ± c) + i(b ± d)
2. z1 z2 = (a + ib)(c + id) = (ac − bd) + i(ad + bc)
  
(ac+bd)+i(bc−ad)
3. zz21 = a+ib
c+id = a+ib
c+id
c−id
c−id = c2 +d2 .

The results are in all cases complex numbers.

1.2. COMPLEX NUMBERS 3

Im(z)

z = x + iy
iy

x Re(z)

Figure 1.1: Representation of a number z = x + iy in the complex plane.

Im(z)
Im(z)

z 1 + z2

z1
z1

z1 − z 2

z2 z2

Re(z) Re(z)

Figure 1.2: Addition and subtraction of two complex numbers.

1.2.2 Complex number system: Graphical representation

A complex number z = x + iy is made of a real part x, denoted by Re z, and of an imaginary part, y = Im z.

The complex z-plane was originally devised by Caspar Wessel, a Danish-Norwegian cartographer, in a paper
published in 1797, but which would receive little attention for nearly a century outside of Scandinavia, maybe
due to the fact that it was written in Danish. It can be viewed as a two-dimensional cartesian coordinate
system whose horizontal axis corresponds to the real part of the numbers and whose vertical axis, to the
imaginary part of the numbers. We can thus display z on the complex plane as a point with abscissa x and
ordinate y, as in Figure 1.1.
We can also display operations on the numbers in the complex plane. Addition and subtraction are
illustrated in Figure 1.2. One can add and subtract by forming parallelograms, in the same way that one
would perform addition or subtraction of two vectors. In order to display multiplication and division of two
complex numbers, we need first to represent z1 and z2 in polar form. A complex number has √ a magnitude
and an argument (or angle), cf. Figure ??. The point z1 = a + ib has a magnitude r1 = |z1 | = a2 + b2 and
an argument θ1 . The point z2 has a magnitude r2 and an argument θ2 . Then, z1 = r1 (cos(θ1 ) + i sin(θ1 ))
and z2 = r2 (cos(θ2 ) + i sin(θ2 )).
4 CHAPTER 1. COMPLEX NUMBERS

z1 z2 = r1 (cos(θ1 ) + i sin(θ1 )) · r2 (cos(θ2 ) + i sin(θ2 ))

= r1 r2 ((cos(θ1 ) cos(θ2 ) − sin(θ1 ) sin(θ2 )) + i(cos(θ1 ) sin(θ2 ) + sin(θ1 ) cos(θ2 )))
= r1 r2 (cos(θ1 + θ2 ) + i sin(θ1 + θ2 ))

and similarly
z1 r1
= (cos(θ1 − θ2 ) + i sin(θ1 − θ2 ))
z2 r2

The magnitude of a product is r1 r2 , the product of the magnitudes and the argument of a product is
θ1 + θ2 , the sum of the arguments. Similarly, the magnitude of a quotient is r1 /r2 and the argument of a
quotient is θ1 − θ2 . Expressed in formulas:

|z1 · z2 | = |z1 | · |z2 | and arg(z1 · z2 ) = arg(z1 ) + arg(z2 )

|z1 /z2 | = |z1 |/|z2 | and arg(z1 /z2 ) = arg(z1 ) − arg(z2 )

The complex conjugate of a complex number z = x + iy is z̄ = x + iy = x − iy. A useful formula for

finding the magnitude of a complex number is r2 = |z|2 = z̄z. One can easily check that

• z1 ± z2 = z1 ± z2

• z1 · z2 = z1 · z2

• z1 /z2 = z1 /z2

If p(z) is a polynomial with real-valued coefficients, it follows from these rules that p(z) = p(z). This is
a special case of the Schwartz Reflection Principle described in Section 3.2.2. While we now can take the
square root of a negative number, we still cannot divide by zero. We will next introduce a spherical version
of the complex plane which addresses this.

1.2.3 Examples of polar rules

1. i2 = −1. This makes sense since arg(i) = π/2 and |i| = 1. If we square i, we square the magnitude
and double the argument, see illustration in Figure 1.3. This has better be true since it is the property
that initiated the whole present topic of complex variables)

2. We can graphically find solutions to z k + 1 = 0 by alternatively considering z k = −1. We find all the
points on the unit circle for whom k arg(z) = π + 2nπ, where n is an integer. For example, if k = 2,
arg(z) = π/2 + nπ when n = 0, 1. The k = 3 case is displayed in Figure 1.4 (a).

3. We can similarly find solutions to z k − 1 = 0 by considering that z k = 1. Those k solutions are called
the k th roots of unity. They are equispaced around the unit circle. The third roots of unity are shown
in Figure 1.4 (b).
n
4. The product rule implies de Moivre’s formula (cos θ + i sin θ) = cos nθ + i sin nθ in the special case of
n integer.

Example 1.2.1 Use the polar rules to simplify z = √ 1+i√ .

1+i 3

√
For the numerator, we get√|1 + i| = 2 √ and arg(1√ + i) = arctan 1 = π/4. √ Regarding the√denominator, we
start by considering 1 + i 3, i.e. |1 + i 3| = 1 + 3 = 2, and arg(1 + i 3) = arctan 3 = π/3. Thus,
|z| = 21/2 /21/2 = 1 and arg z = π4 − 12 · π3 = 12
π π
. Hence z = cos 12 π
+ i sin 12 .
1.2. COMPLEX NUMBERS 5

Im(z)

z1 i

π/2
r1 π
z2
θ1
−1 1
r2
θ2
Re(z)

(a) Polar form of complex numbers. (b) Display of i2 = −1.

Figure 1.3: If we square i, we square the magnitude (here equal to one) and double the argument (here π/2),
i.e. the result ends up at -1.

(a) z 3 + 1 = 0. (b) z 3 − 1 = 0.

Figure 1.4: Illustrations of the solutions to two single cubic equations.

6 CHAPTER 1. COMPLEX NUMBERS

1.3 Stereographic projection

Although addition, subtraction, multiplication and division are all well defined in the numbering system
defined so far, there still persists a notable limitation. We cannot divide by zero. In terms of complex
numbers, this is naturally handled if we project the complex number plane to a sphere of radius one, placed
on top of the origin in the plane, in the way shown in Figure 1.5. We clearly get a one-to-one map between
the plane and the sphere. However, the ∞ (in all directions) in the plane is just the North Pole, a point like
any other point on the sphere. There is a trade-off, however, in defining complex numbers as points on the
sphere instead of on the plane. By doing so, we lose the nice immediate illustrations of +, −, ∗, / but instead
gain a nicer way to think about divisions by zero, and a consistency in thinking about branch cuts (see for
example Section ??).

Examples
1
Consider the function f1 (z) = z−(1+i) . It diverges to ∞ at z = 1 + i, a point on the plane and on the sphere,
described as a pole. Next, consider the function f2 (z) = z. This function goes to ∞ in the limit of z → ∞
following any direction in the complex plane. The limits in all of these direction map on the sphere to the
North Pole. Thus, on the sphere, this function has a pole at the North Pole.

This stereographic projection has a number of nice properties, such as

• Locally, angles between intersecting curves are preserved.
• A circle in the complex plane maps to a circle on the sphere.
• A circle on the sphere which contains the North Pole maps to a straight line in the plane (discussed
further in the Exercises in Section 1.4).

1.3.1 Formulas
It is easy to find the correspondence between points z = x + iy in the plane and points (X, Y, Z) on the
sphere:

Plane → Sphere Sphere → Plane

4x
X= x2 +y 2 +4
2X
4y
x= 2−Z
Y = x2 +y 2 +4

2(x2 +y 2 ) 2Y
Z= x2 +y 2 +4 y= 2−Z

1.4 Exercises
Exercise 1.4.1 Find the magnitude and the argument of the following complex numbers
√
(a) 1 − i √ (d) 3 + 4i
(b) 1 + i 3 (e) The roots of z 7 + 128 = 0
 4 √
1+i 1+i 3 π
(c) √
2 (f) 1+i . Compute from this cos 12 .

Exercise 1.4.2 Express each of the following numbers in the form a + bi, where a and b are real

1
 2
(a) 6+2i (c) 1+ 3
1+i
(2+i)(3+2i) √
(b) 1−i (d) The roots of z 2 + 32i z − 6i = 0
1.4. EXERCISES 7

iy

Figure 1.5: The stereographic projection of the lines Re z = 0 and Im z = 0 (in bold) and two arbitrary
circles. (Figure courtesy of Douglas Baldwin)

Exercise 1.4.3 Find the magnitude and the argument of the following complex numbers
√
(a) 1 − i √ (e) 3 + 4i
(b) 1 + i 3 (f) The roots of z 7 + 128 = 0
 4 √
1+i 1+i 3 π
(d) √
2 (g) 1+i . Use your result to compute cos 12 .

Exercise 1.4.4 Express each of the following numbers in the form a + bi, where a and b are real

1 (2+i)(3+2i)
(a) 6+2i (d)
√  1−i 2
3
(b) 1+i 3 (e) 1 + 1+i
√ √
(c) 1+i 3 (f) The roots of z 2 + 32i z − 6i = 0

Exercise 1.4.5 Let z and w be any two complex numbers. Prove the relations

(a) z + z̄ = 2 Re(z) (c) |zw| = |z||w|

(b) |z − w| ≤ |z| + |w| (d) |wz + wz| ≤ 2|wz|

Exercise 1.4.6 Draw the set of points that satisfy

(a) Im(z + 2) = 3 (d) |z − 1| + |z + 1| = 3

(b) (b) |z − i| < 2
(c) |z − i + 2| = |z + 2i − 1| (e) |z − 1| − |z + 1| = 2

Exercise 1.4.7 Show that, for the stereographic projection, a circle in the z-plane corresponds to a circle
on the sphere. Hint: Recall the correspondences between points z = x + iy in the regular complex z-plane
and points (X, Y, Z) on the sphere as given in Section 1.3.1. Furthermore, note that a circle on the sphere
is given by the intersection of it with a plane AX + BY + CZ − D = 0.
8 CHAPTER 1. COMPLEX NUMBERS

Exercise 1.4.8 Given a point in the complex z-plane, show that the following steps produce the value w =
1/z: (i) Form 4z and project this value to the sphere, (ii) Rotate the sphere half a turn around a line through
the sphere center, parallel to the X-axis, and (iii) Project the point back again from the sphere to the z-plane.
Chapter 2

Functions of a Complex Variable

A real function of one variable has the form y = f (x). A complex function has similarly the form w = f (z),
where w and z now are complex quantities. Let z = x + iy and w = u + iv, with x, y, u, v real. Thus,
w = u(x, y) + iv(x, y) where u(x, y) and v(x, y) are two real functions of two real variables. If we want to
fully visualize w = f (z), we can plot u(x, y) and v(x, y) separately. For example, let f (z) = z 2 = (x + iy)2 =
(x2 − y 2 ) + i (2xy), i.e. u(x, y) = x2 − y 2 and v(x, y) = 2xy. Thus, to picture w, we can plot u(x, y)
and v(x, y), and it is often helpful to also display |w(x, y)| and arg(w(x, y)), as shown in the middle row
of subplots in Figure 2.1. The top and bottom rows of subplots show similarly the functions f (z) = z and
f (z) = z 3 , respectively. Along the x-axis (marked by a thick red curve in the left and center subplots and as
a black curve in the right subplots), we recognize as special cases the standard results when z = x is a real
variable.
We will soon come to consider much more complicated functions than pure monomials. However, this
type of graphical display will continue to work very well for showing also such functions over the complex
plane. As an example, Figure 2.2 illustrates similarly the function f (z) = z + z1 . Sometimes, it is helpful to
also include a fourth subplot, here showing what the phase portrait of the function looks like (same as the
third subplot, but viewed from straight above, making the magnitude information invisible unless contour
lines for this are included). It is critically important to know what the functions ‘look like’ to understand
and best utilize them.

2.1 Derivative
In the real-valued case, f 0 (x) = lim∆x→0 f (x+∆x)−f
∆x
(x)
, where we require the limit to be the same from the
right and the left. In the case of a complex function, f 0 (z) = lim∆z→0 f (z+∆z)−f
∆z
(z)
, where the limit should
be the same when 4z → 0 from any direction of the complex plane. When this holds, the function f (z)
is said to be analytic. In particular, the limits in two main directions must be the same: the horizontal
direction (when ∆z is purely real) and the vertical direction (when ∆z is purely imaginary). This suffices
for obtaining the Cauchy-Riemann equations.

2.1.1 Cauchy-Riemann equations

f (z+∆z)−f (z)
We defined f 0 (z) = lim∆z→0 ∆z where f (z) = u(x, y) + iv(x, y). Next:
   
1. Choose ∆z = ∆x real. f 0 (z) = lim∆x→0 u(x+∆x,y)−u(x,y)
∆x + i v(x+∆x,y)−v(x,y)
∆x = ∂u
∂x
∂v
+ i ∂x
   
u(x,y+∆y)−u(x,y) v(x,y+∆y)−v(x,y)
2. Choose ∆z = i∆y purely imaginary. f 0 (z) = lim∆y→0 i∆y +i i∆y =
−i ∂u
∂y + ∂v
∂y

For the two results to be the same, u(x, y) and v(x, y) must therefore related by the Cauchy-Riemann (C-R)
equations:

9
10 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

u(x, y) =Re f (z) v(x, y) =Im f (z) |f (z)| together with arg f (z)

4 4 4
2 2
0 0 2
-2 -2
-4 -4 0
2 2 2

1 1 1

2 2 2
0 0 0
1 1 1

-1 0 -1 0 -1 0
y -1 y -1 y -1
-2 x -2 x -2 x
-2 -2 -2
z

4 4 4
2 2
0 0 2
-2 -2
-4 -4 0
2 2 2

1 1 1

2 2 2
0 0 0
1 1 1

-1 0 -1 0 -1 0
y -1 y -1 y -1
-2 x -2 x -2 x

z2
-2 -2 -2

4 4 4
2 2
0 0 2
-2 -2
-4 -4 0
2 2 2

1 1 1

2 2 2
0 0 0
1 1 1

-1 0 -1 0 -1 0
y -1 y -1 y -1
-2 x -2 x -2 x

z3
-2 -2 -2

Figure 2.1: Re f (z), Im f (z), and |f (z)| together with arg f (z) displayed for the three functions f (z) =
z, z 2 , z 3 . In the first two columns of subplots, some positive contour lines√are shown in blue and some negative
ones in green. In the last column of subplots, the magnitude |f (z)| = u2 + v 2 is displayed vertically, and
the phase angle is displayed according to the ‘color wheel’ at the bottom left of each of these subplots
(showing how colors are associated with phase angles).
2.1. DERIVATIVE 11

10 10

5 5

0 0

-5 -5

-10 -10
2 2

1 2 1 2

0 1 0 1
0 0
-1 -1
-1 -1
y -2 y -2
-2 x -2 x

(a) Re f (z) (b) Im f (z)

(c) |f (z)| (elevation) and arg f (z) (color) (d) As (c), but seen from above (with contours for |f (z)|)

Figure 2.2: Graphical representation of f (z) = z + z1 . The function values along the real axis are marked
by thick red or black lines. The colors in (c) and (d) are related to the function values’ phase angle (arg)
according to the small color circle. The phase portrait (subplot (d)) will in most further cases be omitted
(unless it reveals some features not apparent from the first three subplots). In this present case, it also
includes contour lines for the magnitude.
12 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

(
∂u ∂v
∂x = ∂y
∂v
(2.1)
∂x = − ∂u
∂y

Theorem 2.1.1 If the function f (z) is differentiable, the Cauchy-Riemann (C-R) equations hold.

We have already shown this result just above. For the reverse direction, some minor extra conditions are
needed. Theorem 2.8.1shows that the C-R equations together withux , uy , vx , vy being continuous at a point
suffices for f 0 (z) to exist at that point. If the C-R equations are valid in a full vicinity of a point, the
extra requirement get reduced to u and v themselves being continuous (i.e. we do no longer need to verify
continuity also of the partial derivatives). There are many consequences to the C-R equations:

1. If f (z) is differentiable once, it is differentiable an infinite amount of times. This result will be shown
in Section ??. There is no counterpart of this for real-valued functions.

2. The functions u(x, y) and v(x, y) each satisfy Laplace’s equation

uxx + uyy = (ux )x + (uy )y = (vy )x + (−vx )y = 0

and similarly,
vxx + vyy = 0.
Thus, neither u nor v can have a local minimum or a local maximum. Typically, at a local maximum,
both uxx < 0 and uyy < 0, which are incompatible with uxx + uyy = 0. The general case follows from
Theorem 4.2.8, which tells that the value at any point of a Laplace equation solution is the average of
the values around the periphery of any circle centered at the point.

3. Given u, we can compute v up to a constant, and vice versa. They are called the harmonic conjugates
of each other.

4. The gradient vectors for u(x, y) and v(x, y) satisfy ∇u · ∇v = (ux , uy ) · (vx , vy ) = ux vx + uy vy =
−vy uy + uy vy = 0, i.e. one or both are zero (at a location with f 0 (z) = 0), or they are orthogonal
to each other. The same holds for level curves to the two surfaces u(x, y) and v(x, y), as such are
orthogonal to their gradient vectors - see Figure 2.3.

Definition The function f is analytic at z0 if f (z) is differentiable in some neighborhood of z0 (no matter
how small). f is analytic in a region if it is analytic at all points in that region. The function f is holomorphic
if it is analytic. The terms are synonyms. A function is entire if it is differentiable everywhere in C, the
finite complex plane, excluding infinity.

2.1.2 Test whether or not a function is analytic

If a function has a Taylor expansion, it is analytic at least within its radius of convergence. It is sufficient
to find a Taylor expansion in x and the replace x by z. For example,
d 1
• One can easily find a series expansion for log(1 + x), using the fact that dx log(1 + x) = 1+x and the
1 x2 x3
geometric series identity 1+x = 1 − x + x2 − x3 ± .... We find that log(1 + x) = x − 2 + 3 ± ...

• f (z) = z is analytic. It is its own Taylor series.

• f (z) = z̄ = x − iy does not have a Taylor series in z. Checking with the C-R equations, we see that
∂u ∂v
∂x = 1 and ∂y = −1, so (2.1) does not hold. Thus, this function is nowhere analytic.

• If g(z) is analytic, then f (z) = g(z̄) is analytic. Proof: Let z = x + iy and g(z) = u(x, y) + iv(x, y).
Then, g(z̄) = u(x, −y) − iv(x, −y) = a(x, y) + ib(x, y). So ax = ux , ay = −uy , bx = −vx , by = vy ,
which implies that, since g is analytic, ax = ux = vy = by and that ay = −uy = vx = −bx . Since the
Cauchy-Riemann equations are satisfied, f (z) = g(z̄) is analytic.
2.2. SOME ELEMENTARY FUNCTIONS GENERALIZED TO COMPLEX ARGUMENT BY MEANS OF THEIR TAYLOR

-1
-1 0 1

Figure 2.3: This plot shows for f (z) = z 2 − 3 the level curves of u(x, y) as solid curves and of v(x, y) as
dashed curves. They are orthogonal apart from at at z = 0 as f 0 (0) = 0.

2.2 Some elementary functions generalized to complex argument

by means of their Taylor expansion
P∞
If we are familiar with a function f (x), x real, that has a Taylor series expansion f (x) = k=0 ak xk (where
P∞around x = 0), and this expansion converges for −R < x < R, then we can
we, for simplicity, have expanded
swap x and z and let f (z) = k=0 ak z k be the generalization of the function to a complex argument. This
expansion will later be shown to converge for all z satisfying |z| < R. Typically, the function f (z) remains
completely defined outside this circle as well. How to find its values for |z| ≥ R is the topic of Analytic
continuation, which will be addressed in Chapter 3 .

Example 2.2.1 Extend

x2 x3
y(x) = ex = 1 + x + + + ...
2! 3!
to complex argument.

Since the functional definition is in the form of a Taylor series, we just replace x with z, and thus obtain

z2 z3
w(z) = ez = 1 + z + + + ... 
2! 3!

It is now easy to verify that all the standard relations for the exponential function hold also in the complex
case. For example, when x1 , x2 are real, it holds that ex1 · ex2 = ex1 +x2 . This implies that if we Taylor
expand (in the two variables) the function

f (x1 , x2 ) = ex1 · ex2 − ex1 +x2

x2 x3 x2 x3
= (1 + x1 + 1 + 1 + ...)(1 + x2 + 2 + 2 + ...)
2! 3! 2! 3!
(x1 + x2 )2 (x1 + x2 )3
−(1 + (x1 + x2 ) + + + ...)
2! 3!
= a0,0 + (a1,0 x1 + a1,1 x2 ) + (a2,0 x1 + a2,1 x1 x2 + a2,2 x22 ) + ...
2
14 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

then every Taylor coefficient ai,j must be zero. These Taylor coefficients do not at all depend on what values
we substitute for x1 and x2 , so the result will be zero also if we substitute in complex numbers z1 and z2 .
Consequently, the relation ez1 · ez2 = ez1 +z2 must hold also for complex arguments z1 and z2 .
In the same way, we can conclude that any functional relation that holds when argument(s) are real (and
the function(s) involved are Taylor expandable) will again hold when the argument(s) are complex.

2.2.1 Relations between exponential and trigonometric functions

If z = x + iy, we get from the result above that
ez = ex+iy = ex · eiy
2 3 4 5 2 4 3 5
where furthermore eiy = 1 + iy − y2! − i y3! + y4! + i y5! ... = (1 − y2! + y4! ...) + i(y − y3! + y5! ...) = cos y + i sin y.
The generalization of the exponential function to complex arguments will therefore obey
ez = ex+iy
= ex · eiy
= ex (cos y + i sin y) (2.2)
πi
An interesting special case is Euler’s identity (discovered in 1748): e + 1 = 0, connecting the five
fundamental numbers 0, 1, e, π and i.
Figure 2.4 shows what the exponential function looks like in the complex plane (numerous similar illus-
trations for other standard analytic functions will soon be central to our description of them). This figure
was produced by the MATLAB statements shown in Section 2.9.1. We can recognize the real-valued function
y = ex along the real axis in the upper left subplot, along the real axis, drawn in red. Using Mathematica,
similar plots are obtained by the statements shown in Section 2.9.2.
We saw for the monomials f (z) = z, z 2 , z 3 in Figure 2.1 and now again for the exponential function
that a Taylor expandable function, previously defined only along the real, will have a unique extension to
the complex plane. A strong statement about this will soon be given as Theorem 2.3.2.

2.2.2 Trigonometric functions represented in terms of the exponential function

Since the relation eiy = cos y + i sin y holds for y real, it must according to our discussion above also hold
when y is complex, i.e. we have Euler’s relations
eiz = cos z + i sin z (2.3)
e−iz = cos z − i sin z (2.4)
Adding and subtracting these equations give
eiz + e−iz
cos z = (2.5)
2
and
eiz − e−iz
sin z = (2.6)
2i

Graphically, we can see what this becomes for cos z (Figure 2.5) and for sin z (Figure 2.6). The two
functions differ only by a phase shift (translation) in the real direction (cos z = sin(z + π2 )), and they grow
exponentially fast in the imaginary direction (away from the real axis). These relations motivate how the
functions cosh x and sinh x are defined for real x, and thus extended to complex z as

ez + e−z
cosh z = = cos(iz)
2
ez − e−z 1
sinh z = = sin(iz)
2 i
Many trigonometric identities (at first sight not having anything to do with complex numbers) can be
derived very easily with use of Euler’s formulas:
2.2. SOME ELEMENTARY FUNCTIONS GENERALIZED TO COMPLEX ARGUMENT BY MEANS OF THEIR TAYLOR

60 60

40 40

20 20

0 0

-20 -20

-40 -40

-60 -60
10 10

5 4 5 4

0 2 0 2
0 0
-5 -5
-2 -2
y -10 y -10
-4 x -4 x

(a) Re ez . (b) Im ez .

(c) |ez | and arg ez .

Figure 2.4: Graphical representation of f (z) = ez .

16 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

30 30

20 20

10 10

0 0

-10 -10

-20 -20

-30 -30
4 4

2 10 2 10

0 5 0 5
0 0
-2 -2
-5 -5
y -4 y -4
-10 x -10 x

(a) Re cos z. (b) Im cos z.

(c) | cos z| and arg cos z.

Figure 2.5: Graphical representation of f (z) = cos z.

2.2. SOME ELEMENTARY FUNCTIONS GENERALIZED TO COMPLEX ARGUMENT BY MEANS OF THEIR TAYLOR

30 30

20 20

10 10

0 0

-10 -10

-20 -20

-30 -30
4 4

2 10 2 10

0 5 0 5
0 0
-2 -2
-5 -5
y -4 y -4
-10 x -10 x

(a) Re sin z. (b) Im sin z.

(c) | sin z| and arg sin z.

Figure 2.6: Graphical representation of f (z) = sin z.

18 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

Example 2.2.2 Derive the addition theorems for sin(α + β) and cos(α + β).

We have
ei(α+β) = cos(α + β) + i sin(α + β)
and also

ei(α+β) = eiα eiβ

= (cos α + i sin α)(cos β + i sin β)
= (cos α cos β − sin α sin β) + i(cos α sin β + sin α cos β)

Equating the real and imaginary parts now give the standard addition theorems. 
PN
Example 2.2.3 Evaluate k=−N cos kx.

N
X N
X
eikx = e−N ix 1 + eix + e2ix + ... + e2N ix


cos kx =
k=−N k=−N
| {z }
PN
Imaginary part vanishes since k=−N sin kx = 0; now a finite geometric progression.

1 − e(2N +1)ix e−(N +1/2)ix − e(N +1/2)ix

= e−N ix =
1 − eix e−ix/2 − eix/2
| {z }
Routine way to simplify denominator of this type; here multiply numerator and denominator by e−ix/2

−2i sin N + 12 x sin N + 12 x

= =
−2i sin 12 x sin 12 x
| {z }
This is a purely real-valued result (when x is real) - as to be expected

While the result can also be proven by induction over N , the present derivation is more practical, since it
does not require the answer to be known in advance. 

2.2.3 Logarithm function

From the (real-valued) Taylor expansion for log(1 + x) follows that we can define the extension to complex
z as
1 1 1
log(1 + z) = z − z 2 + z 3 − z 4 ± ... (2.7)
2 3 4
By the principle that functional relations that hold in real cases must carry over to complex variables, we
get log(z1 z2 ) = log(z1 ) + log(z2 ). Writing z = reiθ gives

log z = log r + iθ = log |z| + i arg z, (2.8)

a counterpart of (2.2) in terms of extending an elementary function from the real axis to the complex plane
(and also outside the radius of convergence R = 1 for (2.7)). Displayed over the domain −4 ≤ Re z ≤ 4,
−4 ≤ Im z ≤ 4, the logarithm function becomes as shown in Figure 2.7. It is clear that something we have
not seen before happens along the negative real axis (leading us soon to introduce the concepts of branch
cuts and Riemann surfaces).
In the displays of different functions f (z), we typically show surface plots for Re f (z), Im f (z), and |f (z)|.
For the first two of these, we noted in Section 2.1.1 that they can never have any local (finite) maxima or
minima, since they obey Laplace’s equation. From (2.8) follows a similar result for |f (z)|. This quantity has
a local minimum (with value zero) wherever f (z) is zero, but cannot otherwise have any local finite extreme
points. If it did, so would log |f (z)| = Re log f (z), which is not allowed for the real part of an analuytic
function (this is later expressed in Theorem 4.2.10).
2.2. SOME ELEMENTARY FUNCTIONS GENERALIZED TO COMPLEX ARGUMENT BY MEANS OF THEIR TAYLOR

2
2
1.5
1
1

0.5 0

0
-1
-0.5

-1
-2

-1.5 -3
-2 4
4
2 4
2 4
0 2
0 2
0
0 -2
-2 -2
-2
y y -4 x
-4 -4 x -4

(a) Re log z. (b) Im log z.

(c) | log z| and arg log z. (d) arg log z and contour lines for | log z|.

Figure 2.7: Graphical representation of f (z) = log z.

20 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

1.5
2

1.5 1

1 0.5

0.5 0

0 -0.5
-0.5
-1
-1
-1.5
-1.5
-2
-2 4
4
2 4
2 4
0 2
0 2
0
0 -2
-2 -2
-2
y y -4 x
-4 -4 x -4

(a) Re arctanh z. (b) Im arctanh z.

(c) |arctanh z| and arg arctanh z.

Figure 2.8: Graphical representation of f (z) = arctanh z. Notice the lines of discontinuity, especially visible
on the imaginary part of the function, and not present for the real part.

2.2.4 Inverse hyperbolic and and trigonometric functions

Solving for z in the RHSs of (2.5) and (2.6) (and their hyperbolic counterparts) gives expressions for their
inverse functions, for ex.:  
1 1+z
arctanh z = log (2.9)
2 1−z
arcsin z = −i log(iz + (1 − z 2 )1/2 ) (2.10)
arccos z = −i log(z + (z 2 − 1)1/2 ) (2.11)
 
i 1 − iz
arctan z = log (2.12)
2 1 + iz
Figures 2.8 - 2.10 illustrate the first three of these.
 
1 1+z
Example 2.2.4 Derive equation (2.9) above: arctanh z = 2 log 1−z , and compare it to its counterpart
for z real.
ew −e−w e2w −1
We first note that w = arctanh z corresponds to z = tanh w = ew +e−w = e2w +1 , from which follows
1+z
e2w = 1−z and thus (2.9): w= 21 log 1−z
1+z
.
2.2. SOME ELEMENTARY FUNCTIONS GENERALIZED TO COMPLEX ARGUMENT BY MEANS OF THEIR TAYLOR

3
2

1.5 2

1 1

0.5 0

0 -1
-0.5
-2
-1
-3
-1.5
-4
-2 5
5
10
10
0 5
0 5
0
0
-5 -5
y y -5 x
-5 -10 x -10

(a) Re arcsin z. (b) Im arcsin z.

(c) | arcsin z| and arg arcsin z.

Figure 2.9: Graphical representation of f (z) = arcsin z.

22 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

1
4
0
3.5
-1
3
-2
2.5
-3
2
-4
1.5 -5
5

0 0
0.5
-10
-5
0 0
-5 y y 5 5
-10 -8 -6 -4 -2 10
0 2 4 6 8 10 x
x

(a) Re arccos z. (b) Im arccos z.

then
(c) | arccos z| and arg arccos z.

Figure 2.10: Graphical representation of f (z) = arccos z.

2.3. ADDITIONAL OBSERVATIONS ON TAYLOR EXPANSIONS OF ANALYTIC FUNCTIONS 23

2 2

1.5 1.5

1 1

0.5 0.5

0 0
y

y
-0.5 -0.5

-1 -1

-1.5 -1.5

-2 -2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2
x x

(a) The function y = tanh x. (b) The inverse function y = arctanh x.

Figure 2.11: Graphical representation of the functions y = tanh x and y = arctanh x. The two curves are
identical apart from a reflection in the line y = x (exchanging the roles of x and y).

x −x
For x real, tanh x = eex −e
+e−x asymptotes to ±1 when x → ±∞, and we obtain the inverse function by
reflecting its graph across the line y = x, cf Figures 2.11 (a,b). The curve in the latter figure matches
the central (red) section in Figure 2.8 (a). Nothing else in these Figures 2.8 (a,b) is apparent from the
case of a real argument. Regarding the imaginary part in Figure 2.8 (b), we see two shifted copies of the
logarithm function from Figure 2.7 (b), as to be expected from writing the arctanh formula as arctanh z =
1 1
2 log(1 + z) − 2 log(1 − z). 

2.3 Additional observations on Taylor expansions of analytic func-

tions
We introduced above the standard analytic functions through their Taylor expansions. Before using them
much further, below are some observations of their domains of convergence.

Theorem 2.3.1 The following are four key results on the region in which Taylor expansions converge (as-
suming for notational simplicity expansions around z0 = 0 - else replace z by z − z0 below):
1. Convergence occurs within a circular domain |z| < R with divergence for |z| > R, where R is some
value from 0 to ∞ (including 0 and ∞). Convergence on the boundary |z| = R can vary from case to
case.
2. The radius R is the largest possible, such that f (z) is differentiable at every point with |z| < R.
P∞
3. If the Taylor expansion takes the form f (z) = n=0 an z n , then R = 1/ lim|an |1/n where lim denotes
the limit superior (upper limit) of the sequence.

4. The ratio test: In case that lim n→∞ |a|an+1

n|
| exists, then R is equal to this limit.

The first two statements will be proven later in Section ??, while the last two are known from basic
calculus. Here, lim denotes the limit superior (upper limit). The following is three equivalent ways to define
it. A = lim Sn of a sequence {Sn }∞n=1 if (i) The sequence is bounded from above and has A as its largest
limit point, (ii) A = limN →∞ supn>N Sn , and (iii) A is the smallest number such that for  arbitrarily small,
Sn > A +  only a finite amount of times.
1 1 1
Example 2.3.1 Determine R for f (z) = ez = 1 + 1! z+ 2! z2 + 3! z3 + . . .

Since f 0 (z) = ez , the derivative exists and is finite for all finite z, By part 2 above, the radius of convergence
is R = ∞ (meaning that it converges for all |z| < R = ∞). The result follows immediately also from parts 3
and 4 of the theorem. 
24 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

Example 2.3.2 Determine R for f (z) = log(1 + z) = z − 21 z 2 + 13 z 3 − + . . .

Since f 0 (z) = 1
1+z , the only singularity is located at z = −1, implying that the radius of convergence is
R = 1. Alternatively, we can note that R = 1/ lim|1/n|1/n = 1 or R = lim n→∞ |a|an+1
n|
| = 1. 

1
Example 2.3.3 Determine R for 1+z 2 = 1 − z2 + z4 − z6 + − . . .

The radius of convergence is again R = 1, since the function is differentiable everywhere apart from at
z = ±i. Also R = 1/ lim|an |1/n = 1, but lim n→∞ |a|an+1
n|
| does not exist.

´∞ e−t
Example 2.3.4 Determine R for 0 1+zt = 1 − 1!z + 2!z 2 − 3!z 3 + 4!z 4 − + . . .

This function will be considered in more detail in Example 3.2.12. The radius of convergence is R = 0,
easiest seen from R = lim n→∞ |a|an+1
n|
| = 0, but follows also from the other tests. 
P∞
It is easy to verify that if f (z)P= n=0 an z n is valid in the disk of radius R, then we can differentiate
∞ ∞
(f 0 (z) = n=0 nan z n−1 , f 00 (z) = n=0 n(n − 1)an z n−2 , . . .) or integrate however many times we want and
P
obtain expansions that will still converge in the disk of radius exactly R.

Theorem 2.3.2 If two analytic functions coincide on any curve segment, no matter how small (or any point
set with a limit point), then the two functions are identically the same.

Proof If two analytic functions coincide on any curve segment, then sufficient data is given to determine
all the Taylor coefficients, which then determine the function uniquely. 

This result raises fundamental questions about different possible representations of a single function. As
1
illustrated in Fig 2.12, f (z) = 1−z is defined everywhere except at z = 1, where the function has a pole.
2 3
Next, define g(z) = 1 + z + z + z + .... This function g(z) converges only in the disk |z| < 1 and diverges
for |z| ≥ 1. Do these functions represent the same function everywhere? This will be the topic of analytic
continuation. Although the function itself exists everywhere, its particular representation as a Taylor series
fails outside |z| = 1.
Given, at a point a values for f (a), f 0 (a), f 00 (a), ..., we can write down the Taylor expansion at that
point
(z − a)2 00
f (z) = f (a) + (z − a)f 0 (a) + f (a) + ... (2.13)
2!
This relation is theoretically important, but it is usually impractical to differentiate a non-trivial function
many times. We will next describe two methods (beyond (2.13)) to obtain Taylor expansions of given
functions. To go in the reverse direction, obtain a function from its Taylor expansion, we will need analytic
continuation - the topic of Chapter 3.

2.3.1 Get the expansion from an already known expansion

1 1
Example 2.3.5 Derive Taylor expansions for log(1 + z) and 1+z 2 based on the geometric series 1+z =
1 − z + z2 − z3 + z4 − + . . .

Integrating the geometric series gives log(1 + z) = z − 12 z 2 + 13 z 3 − 41 z 4 + 15 z 5 − + . . . + C. To find C, let z = 0,

1
leading to C = 0. Replacing z by z 2 in the geometric series gives 1+z 2 4 6
2 = 1 − z + z − z + z − +...
8


Example 2.3.6 Determine the Taylor expansion for f (z) = arcsin z.

We first note that f 0 (z) = √ 1

1−z 2
. The binomial theorem generalized to non-integer powers states that
s(s−1) 2 s(s−1)(s−2) 3
s
(1+z) = 1+sz + 2! z + 3! z +... Using this with s = −1/2 and swapping z for −z 2 gives f 0 (z) =
(−1/2)(−3/2) (−1/2)(−3/2)(−5/2)
1+ 21 z 2 + 2! z4 − 3! z 6 +. . . and therefore f (z) = arcsin z = z + 61 z 3 + 40
3 5 5 7
z + 112 z +. . .

2.3. ADDITIONAL OBSERVATIONS ON TAYLOR EXPANSIONS OF ANALYTIC FUNCTIONS 25

Im(z) Im(z)

1 1
Re(z) Re(z)

(a) f (z) is defined for all z except z = 1. (b) g(z) converges only in |z| < 1.

Figure 2.12: Although the function f (z) exists everywhere, its particular representation g(z) as a Taylor
series fails on and outside |z| = 1

2.3.2 Method of undetermined coefficients

This method applies directly when one wants to convert a rational function to a Taylor expansion.

3−4z
Example 2.3.7 Find the Taylor expansion of f (z) = 1+2z .

Let the expansion be

3 − 4z
= a0 + a1 z + a2 z 2 + a3 z 3 + . . .
1 + 2z
Multiplying up the denominator gives

3 − 4z = a0 + (a1 + 2a0 ) z + (a2 + 2a1 ) z 2 + (a3 + 2a2 ) z 3 + . . .

Since this is an identity, the coefficients on the two sides must match. We thus obtain

3 = a0
−4 = a1 + 2a0
0 = a2 + 2a1 (2.14)
0 = a3 + 2a2
···

from which all the coefficients follow recursively: a0 = 3, a1 = −10, an+2 = −2an−1 , n = 0, 1, 2, . . . or
evaluated: f (z) = 3 − 10 z + 20 z 2 − 40 z 3 + 80 z 4 + − . . . 

With the power of modern computers, many results relating to analytic functions are nowadays first discov-
ered numerically, and only later (if at all) verified analytically. This brings up the intriguing issue of the
amount of numerical evidence that is needed for it to become plausible that an observed result will always
remain true. The book The Computer as a Crucible, subtitled An Introduction to Experimental Mathemat-
ics [2] contains a fascinating discussion of this subject. The following example, again concerning Taylor
expanding a rational function, can be found in it. The example is highly exceptional, but it can still serve
as a useful warning to be careful!

Example 2.3.8 Find the leading terms in the Taylor expansion of

8 + 7z − 7z 2 − 7z 3
f (z) = .
1 − 6z − 7z 2 + 5z 3 + 6z 4
 
Does the following recursion a0 = 8, a1 = 55, an+2 = 1 + a2n+1 /an , n = 0, 1, 2, . . . (with the brackets b·c
meaning rounding down to nearest integer) produce the exact result for all the Taylor coefficients of f (z)?
26 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

For a problem involving quite a lot of algebra, computational tools greatly simplify tedious labor (as well as
greatly reduce the risk of algebraic errors). For Mathematica, note in this context it’s command “Series”,
and for MATLAB, the functions “conv” and “deconv”. One finds the leading Taylor coefficients to be
a0 = 8
a1 = 55
a2 = 379
a3 = 2612
a4 = 18002
a5 = 124071
a6 = 855106
a7 = 5893451
a8 = 40618081
a9 = 279942687
···
The singularity of f (z) located nearest to the origin occurs around z = 0.145094, implying that the coeffi-
cients (all integers) will forever grow each step roughly by a factor of 7. It has been found that the proposed
recursion formula for the coefficients indeed is exact all the way up thorough n = 11, 055, but it fails for
n = 11, 056. 

g(z)
Example 2.3.9 Determine the Taylor expansion for f (z) = arcsin√ z
1−z 2
= h(z) based on the known expansions
1 3 3 5 5 7 1 2 1 4 1 6 5 8
g(z) = z + 6 z + 40 z + 112 z + . . . and h(z) = 1 − 2 z − 8 z − 16 z − 128 z − . . .. .

This problem generalizes the rational function case above in that the degree is infinite in both numerator
and denominator. That makes however no difference in how one proceeds. Noting that the ratio will be an
g(z) g(z)
odd function, we let the expansion for h(z) be h(z) = a1 z + a3 z 3 + a5 z 5 + . . . Multiplying up the denominator
and equating coefficients gives again a set of recursive equations that can be solved explicitly:

1 = a1
1
6 = − 21 a1 + a3
3
40 = − 81 a1 − 12 a3 + a5 (2.15)
5 1
112 = − 16 a1 − 18 a3 − 12 a5 + a7
···

arcsin z
and thus f (z) = √
1−z 2
= z + 23 z 3 + 8 5
15 z + 16 7
35 z + ... 

Another common case of using the method of undetermined coefficients arises if the function f (z) satisfies
some simple ODE (ordinary differential equation).

arcsin z
Example 2.3.10 Determine again the Taylor expansion for f (z) = √
1−z 2
based on an ODE that it satisfies

Once we differentiate and simplify, we notice that f (z) satisfies (1 − z 2 )f 0 (z) = 1 + zf (z) with f (0) = 0.
Now assume that f (z) = a0 + a1 z + a2 z 2 + ..., then f 0 (z) = a1 + 2a2 z + 3a3 z 2 + .... After substituting
these expansions into the ODE, we can equate the coefficients and get any number of coefficients recursively,
starting with a0 = 0, obtaining again f (z) = arcsin
√
1−z 2
z
= z + 23 z 3 + 15
8 5
z + 16 7
35 z + . . . 

Especially for nonlinear ODEs of the general form f 0 (z) = F (z, f (z)) where F is analytic in both its
arguments, even the method of undetermined coefficients often becomes cumbersome if one needs many
terms. Assume we have already obtained the beginning of the Taylor expansion f (z) = a0 + a1 z + . . . + an z n
(at first just knowing a0 ), substituting this truncated series into the ODE, then re-expanding and integrating
once gives one more term. This is relatively straightforward to implement computationally, and can be
repeated indefinitely to obtain any number of terms.
2.4. SINGULARITIES 27

3 3

2 2

1 1

0 0

-1 -1

-2 -2

-3 -3
2 2

1 2 1 2

0 1 0 1
0 0
-1 -1
-1 -1
y -2 y -2
-2 x -2 x

1 1
(a) Re 1+z 2 . (b) Im 1+z 2 .

1 1
| 1+z 2 | and arg 1+z 2 .

1 1
Figure 2.13: Graphical representation of f (z) = 1+z 2 .The Runge function f (z) = 1+z 2 has two first order

pole singularities in the complex plane but is differentiable everywhere on the real axis.

2.4 Singularities
Functions f (x), x ∈ R with which we are familiar, typically can be extended to become analytic functions
w = f (z). However, they can become infinite (or non-differentiable) at some points or regions in the complex
plane. These points or regions are called singularities. We have encountered some examples already. For
1
example, consider the function f (z) = 1+z 2 . It is everywhere differentiable on the real axis, but exhibits

singularities at z = ±i in the complex plane (cf. Figure 2.13). We next want to describe the different types
of ways that an analytic function can be singular. There are six types of singularities. The first four refer to
isolated singularities, meaning that the function is analytic in a full vicinity of the singular point.

2.4.1 Removable singularities

This type of ‘singularity’ is a problem of a particular representation of the function, and not of the function
itself. It should therefore not be counted as a type of singularity, but it nevertheless traditionally is - thus
being a misnomer. In contrast to genuine singularities, removable ones do not limit the radius of convergence
for a Taylor series.

Example 2.4.1 The function f (z) = sinz z has a singularity at z = 0, at which point it is undefined. However,
1 3 1 5
sin z = z − 3! z + 5! z − +.... Thus sinz z = 1 − 3!
1 2 1 4
z + 5! z − +... converges for all z, and is thus analytic for
all z. It is an entire function. Wherever we canter a Taylor expansion of this function, it will have R = ∞.
28 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

A more formal characterization of the next two types of singularities has to await our description of
Laurent expansions in Section ?? We can however already now describe their key features, and note how
they differ both in character and when displayed visually.

2.4.2 Poles
These singularities typically arise through division by zero. In the vicinity of such a point, |f (z)| get uniformly
large. One can recognize a pole by its signature ‘chimney look’ in the graph of |f (z)|, together with a simple
phase angle (color) pattern. See for example Figures 2.2 and 2.13.

Example 2.4.2 The function f (z) = z1 has a first order pole singularity at z = 0, f (z) = z12 has a second
order pole singularity at z = 0 and f (z) = sin1 z has forst order pole singularities at z = πk, k = 0, ±1, ±2, ....
in all cases caused by divisions by zero. 

Near a pole singularity, the phase angle coloring shows great simplicity and regularity. Returning to Figure
2.2 with f (z) = z + z1 , we see a first order pole (from the term z1 ) at z = 0 and first order zeros at z = ±i.
In the phase portrait (subplot (d)), the we note that the color pattern near the pole is identical to that
surrounding the zeros, apart from a reversal in the order of the colors. For zeros and poles of higher orders
(such as z 2 , 1/z 2 , etc.), the colors (phase angles) would be traversed the same number of thies as the order.

2.4.3 Essential singularities

The structure of a function near an essential singularity is far more complicated. In fact (as will be shown
later in Theorem ??), the function comes arbitrarily close to every single finite value infinitely many times
in any vicinity of the singular point (no matter how small).

Example 2.4.3 The function f (z) = e1/z has an essential singularity at z = 0.

1 x−iy x
y y
This function f (z) = e1/z = e x+iy = e x2 +y2 = e x2 +y2 (cos( x2 +y 2 ) − i sin( x2 +y 2 )) converges to 0 as z → 0

from the negative real axis, converges to +∞ as z → 0 from the positive real axis and is rapidly oscillatory
as z approaches 0 from any other direction. Due to unbound oscillations of infinite density near the origin,
the funcion cannot there be well represented by our usual mesh plots of its real and imaginary parts. The
magnitude / phase plots in Figure 2.14 show near the singularity a totally different, well structured but still
infninitely ‘dense’ phase angle pattern. 
For all the singularity types above, we return to the same value if we follow the function along a small
path that goes around the singular point.

2.4.4 Branch points

Example 2.4.4 Consider the function f (z) = log z = log reiθ = log r + iθ.

Moving around the origin once in the positive direction increases θ by 2π, as we saw in Figure 2.7 (b). How
to ‘deal with’ such functions will be disduccesd in Section 2.5. Very briefly, a branch cut is a ‘barrier curve’
that one can draw in the complex plane if one wishes to make a multi-valued function appear as it was single
valued. The path of a branch cut is quite arbitrary and will mostly depend on conventions. For example, if
one wants log z to be single valued for θ ∈ (−π, π], one would place the branch cut (barrier not to be crossed)
along the negative real axis. If one instead wants to work within θ ∈ [0, 2π), one would place it along the
positive real axis. While the shape of this artificial barrier (branch cut) is arbitrary, its end points are not.
For the log function, it needs to go along some non-self-intersecting path from z = 0 out to infinity.

Example 2.4.5 If not imposing any branch cut restrictions, then Im(log 1) = iθ, where θ = 0, ±2π, ±4π, ....
and log( 1+i
√ ) = i( π + 2kπ), k ∈ Z. With the standard branch cut, only the k = 0 cases are relevant (i.e. the
2 4
result is then single valued).

Example 2.4.6 In the case of the square root function f (z) = z 1/2 , Figure 2.17 shows two common branch
cut choices (discussed much further next in Section 2.5).
2.5. MULTI-VALUED FUNCTIONS - BRANCH CUTS AND RIEMANN SHEETS 29

(a) |e1/z | and arg e1/z . (b) arg e1/z .

Figure 2.14: Graphical representation of f (z) = e1/z . Illustration of the essential singularity.

Note that we can encounter similar cases for real variables. For example, one needs to restrict the x-values of
y = arcsin x to keep the function from becoming multi-valued. The convention is that we allow −1 ≤ x ≤ 1
and thus only define the y−values in −π/2 ≤ y ≤ π/2, thus ignoring that there are more function values
that are possible.
Back to complex variables: Software such as Mathematica or MATLAB typically impose branch cuts
following common conventions - for ex. along the negative real axis in cases where this is appropriate.
Another much less ‘crude’ approach for achieving uniqueness is to extend the complex plane through Riemann
sheets (or surfaces), making branch cuts unnecessary (see Section 2.5 below).

2.4.5 Clusters of poles or other singularities

An example would be a function with poles located at an infinite set of points with a finite limit point.

Example 2.4.7 The function f (z) = 1/ sin z1 is singular at z = πk 1

, k = 0, ±1, . . . These are all poles, apart
from the cluster point at z = 0. Note that this point is not an essaential singularity, since it is not isolated.

2.4.6 Natural boundary

A natural boundary is a genuine barrier (or boundary) past which a function does no longer exist (i.e. not
just that a certain functional form fails to work, such as a Taylor expansion beyond its radius of convergence).
We will discuss on this later with analytic continuation (including an illustration in Figure 3.1).

2.5 Multi-valued functions - Branch cuts and Riemann sheets

The simplest example of multivalued functions is provided by the square root function f (z) = z 1/2 . When
using only real numbers, the function y = x2 can be inverted graphically by reflection in the line y = x (cf.
Figures 2.15 (a,b)). For x positive,
√ there are two choices for numbers
√ that, when squared, return the value
x, traditionally denoted by ± x, with the understanding that x denotes the positive choice.
With instead a complex argumant z, Figure 2.16 reveals a much richer functional behavior. We recognize
√
again the red curve from Figure 2.15 (b) in Figure 2.16 (a), but nothing else. The two choices ± x are
merely curves on smooth connected surfaces. The function f (z) = z 1/2 is double-valued for all z, apart from
an anomaly (branch point) at z = 0 (and a counterpart branch point at z = ∞).
To analyze f (z) = z 1/2 in more detail, consider z in its polar form representation z = reiθ , where we can
√ √ iθ
add any term 2πik, k integer, to the exponent without changing the result. Thus, z = re 2 +ikπ , where
30 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

2 2

1.5 1.5

1 1

0.5 0.5

0 0
y

y
-0.5 -0.5

-1 -1

-1.5 -1.5

-2 -2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2
x x
√
(a) y = x2 near the origin. (b) y = ± x near the origin.
√
Figure 2.15: The real valed function y = x2 , and its inverse y = ± x, as obtained by reflection in the line
y = x (shown in blue).

√
it makes sense to limit k to k = 0,√1 as further choices become just repetitions. For instance, 1 = ekiπ , for
k = 0, 1, recovering the real case 1 = ±1 ; squaring either +1 or −1 gives the result +1.
There are two main ways to proceed when representing a multivalued function:

1. Impose some ‘barrier’ (branch cut) in the complex plane such that, if we do not allow z to move across
it, the result will be single-valued, and

2. Accept double (or, more generally, multi-valued) surfaces of function values, such as those displayed
here in Figure 2.16 (a,b) and also in several earlier figures.

We will next discuss these two choices in some more detail.

2.5.1 Branch cuts

√ √ iθ
Staying for now with the f (z) = z 1/2 example, we can limit ourselves to z = re 2 +ikπ , θ ∈ (−π, π], and
the single case of k = 0. We now permit z only to vary in the z-plane region shown in Figure 2.17 (a), with
an artificial ‘barrier’ (branch cut) placed along the negative real axis. With this restriction, all we will be
able to ‘see’ of the function is what is at or above height zero in Figure 2.16 (a), and on the surface sheet
that tilts upwards for incresing imaginary part in Figure 2.16 (b). These parts are single valued, and in
many cases suffice for what may be needed. This k = 0 case would correspond to the function’s principal
value or primary Riemann sheet. Instead choosing k = 1 will give us the rest of the function - the parts that
were omitted when choosing k = 0.
Even when considering both cases k = 0 and k = 1, the results will depend on our quite arbitrary choice
of locating the branch cut in the z-plane along the negative real axis. Figure 2.17 (b) illustrates the branch
cut instead placed along (immediately below) the positive real axis. If we define θ ∈ [0, 2π), √ we get a jump
discontinuity at the branch√ cut on the positive real axis. The figure shows that the value of i become the
same in either case, but −i differs. In Figure 2.16 (a), the primary sheet would now be what we encounter
when starting from the top red curve and moving one full turn in the positive (counter-clockwise) direction
until immediately before we reach the bottom red curve. In part (b), we would follow the top surface all the
way round. It will depend on the application what choice of branch cut is the most√convenient to use.
How can we test whether a point is a branch point? For example, let f (z) =√ z − z0 . We can test z0
by centering a small circle at it: let z − z0 = reiθ . The function becomes f (z) = reiθ/2 . We then monitor
the function values at each point around the circle. If the starting value differs from the end value, it means
that the point that is being tested is a branch point, and that we need to introduce a branch cut if we
want to make the function appear single-valued. For example, if we choose θ ∈ [0, 2π), let’s test the point
2.5. MULTI-VALUED FUNCTIONS - BRANCH CUTS AND RIEMANN SHEETS 31

(a) Re z 1/2 . (b) Im z 1/2 .

(c) |z 1/2 | .

Figure 2.16: Visualization of f (z) = z 1/2 .

32 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

√ √
(1 + i)/ 2 (1 + i)/ 2

i 1 i 1
−i −1

√ √
(1 − i)/ 2 (−1 + i)/ 2

√ √
(a) z with θ ∈ (−π, π). (b) z with θ ∈ (0, 2π).

Figure 2.17: Two branch cut choices for f (z) = z 1/2 . The numbers shown give the function values at the
respective locations.

2 2

1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5

-2 -2
1 1
0.5 3 0.5 3
2 2
0 1 0 1
0 0
-0.5 -1 -0.5 -1
-2 -2
-1 -3 -1 -3
y x y x

(a) Re f (z). (b) Im f (z).

Figure 2.18: Real and imaginary parts of f (z) = (z − 1)1/2 (z + 1)1/2 . Each consists of two sheets.

√
z = z0 . Let’s follow a small circle
√ around √it, starting at θ = 0. At the starting point, f (z) = r. After one
revolution, θ = 2π and f (z) = reiπ = − r. Thus z = z0 is a branch point.
A branch cut will need to join two branch points. We know that z = z0 is one, but where is the other
one? Let’s test z = ∞. Considering the stereographic projection, z = ∞ is just one point like any other.
To test it, let’s consider z = 1t around t = 0. Then f (t) = √1t , and we see that t = 0 and thus z = ∞ is a
branch point too. The branch cut will need to link those two points.

Example 2.5.1 Discuss different branch cut options for f (z) = (z − 1)1/2 (z + 1)1/2 .

Figure 2.18 shows the real and imaginary parts of f (z). The function clearly has two solution sheets, and
branch points at z = 1 and z = −1. To explore if infinity also is a branch point, we set z = 1t and obtain
√
(1−t)(1+t)
f (t) = t . Thus, there’s a pole at t = 0, but not a branch point. The cut will thus need to link
z = −1 and z = 1 together. Preferring to place cuts along the real axis, there are two natural choices: (i)
Directly along [−1, +1], and (ii) via infinity, i.e. along [−∞, −1] and [1, ∞]. Figure 2.19 focuses on Im f (z),
and illustrates the corresponding single-valued solutions that arise from the two branch cut strategies. For
each choice of branch cuts, it is clear that the two single-valued functions can be combined to form the multi
valued one shown in Figure 2.18 (b).
We can also verify that the cut presented in Figure 2.20 makes the function single-valued. Let’s follow
the contour around the cut to make sure that we end up with the same function value than the one with
2.5. MULTI-VALUED FUNCTIONS - BRANCH CUTS AND RIEMANN SHEETS 33

2 2

1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5

-2 -2
1 1
0.5 3 0.5 3
2 2
0 1 0 1
0 0
-0.5 -1 -0.5 -1
-2 -2
-1 -3 -1 -3
y x y x

(a) Cut [−1, 1], Primary sheet, . (b) Cut [−1, 1], Secondary sheet.

2 2

1.5 1.5

1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1

-1.5 -1.5

-2 -2
1 1
0.5 3 0.5 3
2 2
0 1 0 1
0 0
-0.5 -1 -0.5 -1
-2 -2
-1 -3 -1 -3
y x y x

(c) Cuts [−∞, −1] and [1, ∞], Primary sheet. (d) Cuts [−∞, −1] and [1, ∞], Secondary sheet.

Figure 2.19: Imaginary parts of f (z) = (z − 1)1/2 (z + 1)1/2 . (a), (b) The two solutions when the branch cut
is placed along [−1, 1], (c), (d) The two solutions when branch cuts are placed along [−∞, −1] and [1, ∞].
34 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

θ2
IV

I θ1 III
−1 1

II
p
Figure 2.20: Branch cut for f (z) = (z − 1)(z + 1).

Step Changes in Θ
I θ1 gains 2π
II No change in the angle
III θ2 gains 2π
IV No change in the angle
Total Θ = θ1 +θ
2
2
has increased by 2π

Table 2.1: Changes in the angle Θ as one follows the path shown in Figure 2.20.

1/2 1/2 θ1 +θ2

which we started. Let z − 1 = R1 eiθ1 and z + 1 = R2 eiθ2 . Thus, (z − 1)1/2 (z + 1)1/2 = R1 R2 ei 2 .
Let Θ = θ1 +θ
2 . We want to make sure that the function’s argument Θ has changed by an integer multiple
2

of 2π after having gone around the contour. Table 2.1 shows that after going around the cut, the function’s
argument has increased by 2π. Thus the value of the function remains unchanged. The function is therefore
single-valued with this cut. 

2.5.2 Riemann sheets

If we don’t want to restrict the domain by a branch cut, we obtain a multi-valued function with multiple
Riemann sheets. Fig 2.16 illustrated this for the function f (z) = z 1/2 . Figure 2.21 shows similar illustrations
for f (z) = z 1/3 - a function with three
√ Riemann sheets. We note along the entire real axis one option that
matches the real valued case of y = 3 x (with imaginary part zero), but we see here all the three options for
all complex values z.
While one can think of Figures 2.21 (a,b) as three solution sheets over a single complex ‘base plane’, an
interesting alternative is to think of the base plane replaced to a triple layered complex plane, with layers
connected as if one just ‘flattened down’ either of the triple surfaces, to get three sheets that lie together
without any vertical separations. From the perspective of this triply connected complex base plane, the
function has become entirely single valued (and with no branch cuts needed).
The function f (z) = (z − 1)1/2 (z + 1)1/2 just considered in Example 2.5.1 displayed also how the full set
of function values can be veiewed either as a Riemann sheets, or decomposed into pieses with two different
branch cut strategies.
Figure 2.22 displays in an equivalent way f (z) = log z, with an infinity of Riemann sheets (which all
coincide for the real part - reminiscent of Figure 2.7, but now with additional sheets displayed). As noted
above, it will depend on the application if it is best to consider the infinity of sheets, or make a branch cut,
for ex. along the negative real axis, at the expense of only having a small part of the function available at a
time.
 
z+1
Example 2.5.2 Illustrate the function f (z) = log z−1 .

Apart from a trivial factor of 12 , we are here revisiting the arctanh-function (2.9), shown in Figure 2.8.
As in Example 2.5.1, both z = 1 and z = −1 are branch points. Like for the function f (z) = log z, the
multivaluedness is visible only in the imaginary part; see Figure 2.23. We see a function with an infinity of
Riemann sheets. However, if we choose a path that takes us around both branch points (and does not cross
2.6. SEQUENCES OF ANALYTIC FUNCTIONS 35

(a) Re z 1/3 . (b) Im z 1/3 .

(c) |z 1/3 | .

Figure 2.21: Visualization of f (z) = z 1/3 .

the real axis between −1 and +1), we return to the same function value. The point z = ∞ is in this case
not a branch point. It is nevertheless entirely possibly to place branch cuts along [−∞, −1] and [1, +∞],
and we will obtain a function that is analytic and single valued within the cut domain. In the figure, that
choice would correspond to using the upper halfplane from one level and the lower half plane from the level
above. 

2.6 Sequences of analytic functions

Consider real functions in 1D. Let fn (x) be continuous on [a, b], and let the sequence of functions fn (x)
converge at every point x, as n → ∞. This does not imply that limn→∞ fn (x) = f (x) is continuous.

Example 2.6.1 Discuss the convergence of the function sequence

0 ≤ x < 1 − n1

0
fn (x) =
n (x − 1) + 1 1 − n1 ≤ x ≤ 1
as n → ∞.

This sequence converges pointwise (a every x-location on [0,1]), to zero for 0 ≤ x < 1, and to one for x = 1
- see Figure 2.24 (a). The limit function (in green) is obviously not continuous. 
36 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

(a) Re log z. (b) Im log z (showing only three Riemann sheets).

Figure 2.22: Visualization of f (z) = log z.

-2

-10
-4
10 -5

5 0
0
5
-5
y
x -10 10
   
z+1 z+1
(a) Re log z−1 . (b) Im log z−1 (showing three Riemann sheets).
 
z+1
Figure 2.23: Real and imaginary parts of f (z) = log z−1 Note the different wiev angle compared to Figure
2.8.
2.7. DEFINITIONS OF FUNCTIONS BY INTEGRALS 37

1.2

4
n = 20
1
3 n=3

0.8 2

1
0.6

y
y

0.4
-1
n=2 4 10
-2
0.2

-3

0
-4

-0.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -6 -4 -2 0 2 4 6
x x

0 ≤ x < 1 − n1

0 Pn k+1 sin(kx)
(a) fn (x) = (b) fn (x) = 2 k=1 (−1) .
n (x − 1) + 1 1 − n1 ≤ x ≤ 1 k

Figure 2.24: The functions fn (x) in Examples 2.6.1 and 2.6.2, respectively, shown for some different n-values.
Both cases feature pointwise convergence of a sequence of continuous functions. However, in both cases, the
convergence is non-uniform, and the limit functions are discontinuous.

Example 2.6.2 Discuss the convergence of the function sequence

n
X sin(kx)
fn (x) = 2 (−1)k+1
k
k=1
as n → ∞.
Each function fn (x) is 2π-periodic, converges pointwise to x for x ∈ (−π, π) and equals zero for x = ±π.
Convergence is this time not uniform, due to the Gibbs’ phenomenon - a narrow ‘overshoot’ next to the
jumps (occurring at x = ±π, ±3π, . . .), reaching there approximately 9% of the jump height - see Figure 2.24
(b). 
It can be useful to consider limits of functions that converge pointwise, and then be able to tell something
about the limit. For real functions, a small step in that direction is the concept of uniform convergence:
Given  > 0, ∃ N () such that |fn (x) − f (x)| < , ∀x ∈ [a, b] whenever n > N ()
The two examples above were not uniformly convergent. With uniform convergence, the fact that the
functions fn (x) are continuous implies that the limiting function f (x) is continuous as well. Still this does
d d
not guarantee that the derivative(s) dx fn (x) will converge toward dx f (x), the derivative of the limiting
function.
Example 2.6.3 The sequence fn = sinnnx converges uniformly to zero. In fact, Given  > 0, |fn (x)−f (x)| <
, ∀x whenever n > N () = 1 . However, dx
d d
fn (x) = cos(nx) does not converge towards dx 0 = 0.
Back to complex variables. Life is much easier for analytic functions.
Theorem 2.6.1 (Vitali’s convergence theorem) If fn (z) is analytic on a domain in C and if limn→∞ fn (x) =
dk dk
f (x) pointwise, then f (z) is also analytic and limn→∞ dx k fn (x) = dxk f (x) for all k = 1, 2, 3, . . .. Further-

more, pointwise convergence only on a point set with a finite limit point suffices.

2.7 Definitions of functions by integrals

Consider the function f (z) defined as follows
ˆ b
f (z) = g(z, t)dt
a
38 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

Let g(z, t) be an analytic function of z and a continuous function in t. We want to know whether f (z) is
df
´b
analytic. If |a|, |b| < ∞, then dz = f 0 (z) = a ∂g∂z (z, t)dt must exist. Thus, f (z) has one derivative and is
´n
therefore analytic. If b = ∞, we can redefine f (z) as the limit of the sequence fn (z) = a g(z, t)dt and then
use the convergence of sequence results (Vitali’s convergence theorem). If the integral exists and is finite,
then f (z) is again analytic. The process is identical if a = −∞.
For example, let’s consider ˆ ∞ −t
e
f (z) = dt (2.16)
0 1 + zt
If z is a real negative number, the integral doesn’t converge as the denominator reaches 0. Another example
is the Γ function ˆ ∞
Γ(z) = e−t tz−1 dt (2.17)
0
Let’s analyze it piece by piece, by observing the potential difficulties of the integral
1. The infinite interval can be a problem. However, since e−t makes the integrand decay very rapidly, the
limit of the sequences of functions with the upper limit of integration b → ∞ exists and is finite. The
sequence is thus always convergent.
2. Divergence of the integrand at the origin if the lower limit of integration a → 0 and z < 1. We want
to show that the integral nevertheless converges if Re z > 0. As t & 0, the Taylor expansion of e−t
becomes extremely accurate, and we can express´ z−1e−t t z−1 = tz−1 + 1!
1 z 1 z+1
t + 2! t + . . . The first term
1 z
is the most singular one at the origin; we get 0 t dt = z t t=0 . Writing t z = ez log t , we see that t z
for t & 0 is well defined only for Re z > 0 (then goes to zero; further terms in the expansion imposes
the weaker restrictions, Re z > −1, > −2, etc.)
By means of (2.17), Γ(z) is thus defined when Re z > 0 and it is analytic because of limit arguments at both
limits of integration (a → 0 and b → ∞). We have now defined Γ(z) in the right half plane.
Integrating by parts reveals the interesting property
ˆ ∞ ˆ ∞ ˆ ∞
−t z−1 −t −t z−1 ∞
Γ(z) = e t dt = − z−1
t d(e ) = −e t |0 + e−t tz−2 (z − 1)dt
0 0 0
= (z − 1)Γ(z − 1) if Re z > 1
Functional equations, such as this one arise often for analytic equations. This one can alternatively be written
Γ(z + 1) = zΓ(z) if Re z > 0. (2.18)
´1
Noticing that Γ(1) = 0 e−t dt = 1, then Γ(2) = 1 · 1, Γ(3) = 2 · 1, Γ(4) = 3 · 2 · 1, etc., we have generalized
the factorial function n!, previously solely defined for integers n.
The natural follow-up question to having generalized Γ(z) to an analytic function in the complex plane
becomes: Is the boundary limiting the Γ function to the right hand side of the complex plane (Re z > 0)
an actual limitation of the function, or just an artifact of the functional representation used in (2.17)? It
will soon turn out that we can analytically continue Γ(z) to the left half plane by several methods (all, by
necessity, giving the same result).

2.8 Select proofs

Theorem 2.8.1 If the C-R equations (2.1) hold and the partial derivatives ux , uy , vx , vy are continuous
functions of x and y, then f (z) is differentiable.
1 dx−i dy
Proof Let f (z) = u(x, y) + iv(x, y) and introduce the infinitesimal quantities dz = dx + i dy ⇒ dz = |dz|2 .
Then
f (z + dz) − f (z) (dx − i dy)
= [{u(x + dx, y + dy) − u(x, y)} + i {v(x + dx, y + dy) − v(x, y)}]
dz |dz|2
(dx − i dy)
= [{ux dx + uy dy} + i {vx dx + vy dy}]
|dz|2
2.9. CODES 39

where we, in the last line, replaced u(x + dx, y + dy) (and likewise for v) with the leading terms of their
2-D Taylor expansions around (x, y). Multiplying out the product, some terms drop out due to the C-R
equations, and we are left with f (z+dz)−f (z) 1

dz = |dz|2 (dx)2 (ux + ivx ) + (dy)2 (vy − iuy ) which, after using
(dx)2 +(dy)2
the C-R equations again, further simplifies to |dz|2 (ux + ivx ) = ux + ivx , thus showing that f 0 (z) exists.


2.9 Codes
2.9.1 MATLAB
The following MATLAB script together with function display function produce the three subplots of Figure
2.4.

% Test code for displaying an analytic function; here exp(z)

clear; close all;
bx = [-4,4,-10,10]; % Domain (box) surrounding origin to be displayed
nx = 41; % Number of nodes along real axis for re & im plots
ny = 41; % Number of nodes along imag axis for re & im plots
bounds = [ -60,60; -60,60; 0,60]; % Lower and upper in the three displays
f = @(z) exp(z); % Function to display
display function(f,bx,nx,ny,bounds); % Create the three subplots

function display function(f,bx,nx,ny,bounds)

% The input parameters described in the script that calls this routine
lw = 4; % Set LineWidth for highlighting values along real axis
x = linspace(bx(1),bx(2),nx); y = linspace(bx(3),bx(4),ny);
[xr,xi] = meshgrid(x,y); z = complex(xr,xi);

figure(1) % Plot the real part using mesh

mesh(xr,xi,real(f(z))); colormap([0 0 0]); hold on;
xlabel(’\itx’); ylabel(’\ity’); title(’Real part’)
xlim(bx(1:2)); ylim(bx(3:4)); zlim(bounds(1,:));
plot3(x,zeros(size(x)),real(f(x)),’r’,’LineWidth’,lw); % Highlight real axis

figure(2) % Plot the imaginary part using mesh

mesh(xr,xi,imag(f(z))); colormap([0 0 0]); hold on;
xlabel(’\itx’); ylabel(’\ity’); title(’Imaginary part’)
xlim(bx(1:2)); ylim(bx(3:4)); zlim(bounds(2,:));
plot3(x,zeros(size(x)),imag(f(x)),’r’,’LineWidth’,lw); % Highlight real axis

figure(3) % Plot the magnitude as a surface

% Increase resolution for smoother display
x = linspace(bx(1),bx(2),nx*4); y = linspace(bx(3),bx(4),ny*4);
[xr,xi] = meshgrid(x,y); z = complex(xr,xi);
p = surf(xr,xi,abs(f(z)),angle(-f(z))); % Display the surface
set (p,’EdgeColor’,’none’); colormap hsv(600); hold on;
xlabel(’\itx’); ylabel(’\ity’); title(’Magnitude, with phase plot’)
xlim(bx(1:2)); ylim(bx(3:4)); zlim(bounds(3,:));
plot3(x,zeros(size(x)),abs(f(x)),’k’,’LineWidth’,lw); % Highlight real axis
axes(’Position’,[0.05 0.05 .17 .17]) % Add color wheel for phase information
[th,r] = meshgrid(linspace(-pi,pi),linspace(0,1));
[X,Y] = pol2cart(th+pi,r);
contourf(X,Y,th,100,’linestyle’,’none’); hold on % Show wheel colors
plot([-1 1],[0,0],’k’); plot([0 0],[-1,1],’k’); % Show Re and Im axes
plot(cos(0:0.01:2*pi),sin(0:0.01:2*pi),’k’); colormap hsv(600);
axis equal; axis off
40 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE

2.9.2 Mathematica
If one wants to quickly get surface plots of a complex function with Mathematica (here without highlighting
the real axis or inserting a color wheel), single line statements suffice for each plot. The first line below
specifies the exponential function f (z) = ez , and the following three lines produces plots of its real and
imaginary parts, and magnitude (with phase angle coloring).

f[z ] = E^z;
Plot3D[Re[f[x+I y]],{x,-4,4},{y,-10,10},ViewPoint->{-1.3,-2.4,2}]
Plot3D[Im[f[x+I y]],{x,-4,4},{y,-10,10},ViewPoint->{-1.3,-2.4,2}]
Plot3D[Abs[f[x+I y]],{x,-4,4},{y,-10,10},ColorFunctionScaling->False,Mesh->False,
ColorFunction->Function[{x,y,z},Hue[(Pi+Arg[f[x+I y]]/(2 Pi))]], PlotPoints->61,
ViewPoint->{-1.3,-2.4,2}]

2.10 Exercises
Exercise 2.10.1 Determine the real and imaginary parts of f (z) = z z . Does this function satisfy the
Cauchy-Riemann equations?

Exercise 2.10.2 Derive the equations (2.9) and (2.10), for arctanh z andarcsin z, respectively.

Exercise 2.10.3 Are the following functions analytic? Check using the Cauchy-Riemann (C-R) equations?
If they are, write down as analytic functions of z.

(a) f (x, y) = y 3 − 3x2 y + i(x3 − 3xy 2 + 2) (c) f (x, y) = ex−iy

(b) f (x, y) = z4z+1 (d) f (x, y) = e1/(z−1)

Exercise 2.10.4 Let f (x, y) = u(x, y) + iv(x, y) be an analytic function, where u(x, y) is given below. Find
v(x, y), the harmonic conjugate of u(x, y), and express the function in terms of z.
y
(a) u(x, y) = x2 +y 2 (b) u(x, y) = cos x cosh y

Exercise 2.10.5 Express the Cauchy-Riemann equations in terms of polar coordinates.

Exercise 2.10.6 Use the MATLAB code given in the manuscript to plot the real part, imaginary part,
magnitude and phase portrait of the following functions:

(a) f (z) = ez (d) f (z) = 1/z

(b) f (z) = log z (e) f (z) = arctan z
(c) f (z) = z 6 (f) f (z) = tan z

Exercise 2.10.7 Derive the formulae

(a) sin−1 z = −i log(iz + (1 − z 2 )1/2 ) (b) d

dz sin−1 z = 1
(1−z 2 )1/2

Exercise 2.10.8 Find the radius of convergence of the following series

P∞ k P∞ k! k
(a) z (c) k=1 kk z
Pk=1
∞ zk 1+z
(b) k=1 (k+1)! (d) The Taylor series of 2−z around z0 = i

Exercise 2.10.9 Find the Taylor series expansions around z = 0 of the following functions, valid in the
given regions
2.10. EXERCISES 41

2
1 ez −1−z 2
(a) 1−z 2 , |z| < 1 (b) cosh z, |z| < ∞ (c) , 0 < |z| < ∞
z3
P∞ Ek k
Exercise 2.10.10 The Euler numbers {Ek }∞ k=0 are defined such that
1
cosh(z) = k=0 k! z . Similarly, the
P∞
Bernoulli numbers {Bk }∞k=0 are defined such that ezz−1 = k=0 Bk!k z k
(a) Find the first three Euler numbers and Bernoulli numbers.
(b) Find the radius of convergence of both series.

Exercise 2.10.11 Discuss the singularities of the following functions

cos z−1 1
(a) f (z) = z4 (b) f (z) = z 3 e1/(z−1) (c) f (z) = ez −1
42 CHAPTER 2. FUNCTIONS OF A COMPLEX VARIABLE
Chapter 3

Analytic continuation

A particular representation of a function may be valid only in a limited region of the complex plane, although
the function itself may exist well past that region. For example, a function defined through its Taylor series
can only be evaluated (by immediate summation) inside this expansion’s circular domain of convergence,
which is limited in size by the distance to the nearest singularity.

3.1 Introductory examples

Example 3.1.1 Consider the following three functions
P∞ k 2 3
1. f1 (z) = k=0 z = 1 + z + z + z + ... From the general formula for the radius of convergence
R = 1/limk→∞ |ak |1/k follows that R = 1.
´∞
2. f2 (z) = 0 e−t(1−z) dt. If the integral had been over a finite interval rather than over [0, ∞], it would
have defined an entire function. With the upper limit ∞, it will diverge if Re z ≥ 1.
1
3. f3 (z) = 1−z . This expression represents an analytic function in the whole complex plane, with the
exception of a single first order pole located z = 1.

The key point in this example is that all the three functions are identically the same. The boundaries for
f1 (z) and f2 (z) are not natural ones, but just artifacts of the particular way we used for representing the
function f3 (z). 

The goal of analytic continuation is to reformulate functional expressions so they can be used in larger
regions. It can occur that continuation is not possible. In the following example, the unit circle forms a
natural boundary, and it is not possible to continue the function past it.
P∞
Example 3.1.2 Consider the function f (z) = k=1 z k! = z + z 2 + z 6 + z 24 + ...

The general formula for the radius of convergence gives again R = 1, and the domain of convergence will
be the same as for the function f1 (z) in Example 3.1.1. Consider a point z = e2πik/m where k and m are
arbitrary integers. Then z n! = e2πin!k/m = 1 for all n ≥ m. This means that f (z) → +∞ whenever z is of
the form z = re2ik/m and r → 1. The function f (z) must therefore have singularities located densely all the
way around the unit circle, making |z| = 1 a natural boundary. It can not be continued past it.
What does a function like this look like graphically? Figures 3.1 (a,b) shows its magnitude and phase
angles when displayed within circles of radii 0.99 and 0.999999, respectively. There is of course no difference
between the two displays throughout the first region, but we see more fine structure entering near the edge
in the second case. Moving even closer to the unit circle, the fine structure becomes infinitely dense. We can
also note that |f (z)| does not increase monotonically as the unit circle is approached - indeed, we can in the
phase angle color patterns see tell-tale signs of increasing numbers of zeros. Near the origin, this function
f (x) becomes of course very close to f (z) = z, as displayed in the top right subplot of Figure 2.1. 

43
44 CHAPTER 3. ANALYTIC CONTINUATION

(a) Display of f (z) over |z| ≤ 1 − 10−2 . (b) Display of f (z) over |z| ≤ 1 − 10−6 .
P∞
Figure 3.1: Graphical representation of f (z) = k=1 z k! over two slightly different circles around the origin.
Illustration of a natural boundary.

Whenever a function is given in a form which only permits it to be evaluated in some limited region of
the complex plane, it becomes of great interest to see if it can be rewritten if some other form that permits
the domain to be extended - preferably to the complete complex plane.

3.2 Some methods for analytic continuation

We next review nine different methods for analytic continuation that do not require the use of complex
integration:

1. Circle-chain method

2. Schwarz reflection

3. Use of a functional equation

4. Partitioning of an integration interval

5. Replace Taylor coefficients by integrals or sums

6. Subtract similar series / integral

7. Borel summation

8. Ramanujan’s formula

9. Padé approximations

The last of these approaches is noteworthy also from the point of view that it can provide quite accurate
continuations from highly incomplete information, such as from a relatively low numbers of leading terms
in truncated Taylor expansions. Some examples of continuation methods that use complex integration are
given in Section ??.
3.2. SOME METHODS FOR ANALYTIC CONTINUATION 45

Figure 3.2: Schematic illustration of a circle-chain continuation.

3.2.1 Circle-Chain method

This approach is seldom practical for actual applications, but it is of interest in that it gives insights into
what analytic continuation amounts to. Suppose that we have a function with point-type singularities (poles,
branch points or essential singularities) in the locations marked by black dots in Figure 3.2. If we are given
initially a Taylor expansion centered at the small circle at the origin of the complex plane, it will converge
only in the domain shown by the solid, extending out to the nearest singularity. Within this circle, the
expansion completely describes the function, i.e. we can choose any other point inside it and create a new
expansion centered at this new location. Its domain of convergence will likely reach outside the original one.
The process can in theory be repeated indefinitely, allowing us to proceed for example as is indicated by the
chain of circles in this figure.
From a practical (as opposed to a theoretical) perspective, this approach is virtually always impractical.
If a complete expansion is known analytically, it is rare that one can find convenient closed forms also for the
subsequent expansions - however one such example is given as Example 3.2.1 and another one as Exercise
3.4.1. If the initial expansion is only known numerically, this continuation process turns out to become
extremely ill conditioned.
This circle-chain method also requires the initial Taylor expansion’s radius of convergence R to be non-
zero. Although the theory for continuation based on an R = 0 initial expansion is unclear, at least three of
the following approaches can nevertheless do this successfully - see Example 3.2.12 and Exercises 3.4.3 (b,c)
below.
An interesting situation arises if one tries to continue around a singularity - following separate paths
above and below it. Will the results then agree when the two circle-chain continuations meet up on the
other side of the singularity? The answer will be ’Yes’ when going around poles and essential singularities,
but ’No’ in the case of a branch point.

zn
P∞
Example 3.2.1 Continue the function f (z) = n=1 n by means of a chain of circles once in the clockwise
direction around the singularity at z = 1.

We can start by noting (but will not use further) the fact that f (z) = − log(1 − z), so the end result after
having step-by-step continued around the singularity should return the same function as we started with,
apart from an increase in its imaginary part by 2πi.
The original expansion is centered at z0 = 0. Let the subsequent expansions be centered at z1 , z2 , ..., zp =
46 CHAPTER 3. ANALYTIC CONTINUATION

z0 where zk = 1 − e−2iπk/p , k = 1, 2, . . . , p (see Figure 3.3). From the Taylor expansion of f (z) follows
∞
X 1
f 0 (z) = z n−1 = (3.1)
n=1
1−z
(m−1)!
and therefore f (m) (z) = (1−z)m The Taylor expansion centered at z1 therefore becomes
∞ ∞  n  
X 1 (n) X 1 z − z1 z − z1
f (z) = f (z1 )(z − z1 )n = f (z1 ) + = f (z1 ) + f (3.2)
n=0
n! n=1
n 1 − z1 1 − z1

Differentiating (3.2), we obtain from its first and third expressions

∞  n−1
1 X z − z1 1
f 0 (z) = = ,
1 − z1 n=1 1 − z1 1−z

confirming that (3.1) again holds when now expanding around z = z1 (and its different domain of conver-
gence). Looking at the first and last expressions in (3.2), a pattern becomes clear. For the next expansion,
around the point z2 , we get similarly
   
z2 − z1 z − z2
f (z) = f (z1 ) + f +f ,
1 − z1 1 − z2
etc. After p steps, we arrive at
       
z2 − z1 z3 − z2 zp − zp−1 z − zp
f (z) = f (z1 ) + f +f + ... + f + f
|{z} 1 − z1 1 − z2 1 − zp−1 1 − zp
new f (z) | {z } | {z }
additive constant Old f (z) since zp = 0

Are we back? In the ‘additive constant’, all the arguments are equal, and since zk = 1 − e−2iπk/p , they all
evaluate to z1 = 1 − e−2iπ/p . The value of the additive constant thus becomes
 z12

−2πi/p f (z1 z1
  
−2πi/p
pf (z1 ) = p 1 − e =p 1−e 1+ + + ...
| {z } z1 | {z }| 2
{z
3
}
z1 →2πi →1

The result must be a fixed number, independent of p. To determine its value, we can let p → ∞, which shows
that it must evaluate to 2πi. Continuation with the circle-chain method all the way around the singularity
at z = 1 thus did not come back to the same expansion as we started with, telling that the singularity was a
branch point. Every time we continue in this direction around the singularity, we gain a constant of 2πi. .

3.2.2 Schwarz reflection principle

Suppose that we are given a function f (z) that is
1. Known to be analytic in some region D (cf. Figure 3.4), and
2. Real on the real axis (or on some part of it)
We showed in Section 2.1.2, last bullet item, that the function g(z) = f (z̄) will then also be analytic.
However, it is not defined on D but on D (the reflection of D in the real axis). Since g(z) agrees with f (z)
along a section of the real axis, the two functions must be the same (Theorem 2.3.2). The function as f (z)
has thus been continued from D to D̄. Another way to formulate the result is:
Theorem 3.2.1 An analytic function f (z) that is real on any segment of the real axis takes complex conju-
gated values at complex conjugated locations in the z-plane.
The result can be generalized significantly. For any analytic function, there will be some curve(s) along
which Im f (z) = 0. If we can find some change of variable z = z(w) such that the segment corresponds to
real values of w, we can apply the reflection principle to f (z(w)).
3.2. SOME METHODS FOR ANALYTIC CONTINUATION 47

Figure 3.3: Illustration of the sequence of Taylor expansion centers in Example 3.2.1.

Figure 3.4: Schematic illustration of the Schwarz reflection principle.

3.2.3 Use of a functional equation

´∞
We will illustrate this approach by a series of examples. The first two examples extend Γ(z) = 0
e−t tz−1 dt,
convergent for Re z > 0, to the whole complex plane.

Example 3.2.2 Continue the Gamma function by utilizing the functional equation

Γ(z + 1) = zΓ(z).

We noted in Section 2.7 that the functional equation follows from integration by parts in the defining integral.
If we write the above equation in the form Γ(z) = Γ(z+1) z , it clearly gives the same result as the integral
formulation when Re z > 0, but it can be used down to Re z > −1. Thus, it must provide the analytic
continuation of Γ(z) through the strip −1 < Re z ≤ 0 (and we can note that Γ(z) will have a pole at z = 0).
We can then use the functional equation again, to continue through the strip −2 < Re z < −1, etc. This
process can be repeated indefinitely, showing that Γ(z) exists uniquely defined throughout the entire complex
plane, and that it has poles at z = 0, −1, −2, ... as its only singularities. 
48 CHAPTER 3. ANALYTIC CONTINUATION

Example 3.2.3 Continue the Gamma function by utilizing the functional equation
π
Γ(z)Γ(1 − z) = . (3.3)
sin πz
The proof for this equation will be given later. This relation gives the value of the Γ-function at the location
z whenever it is available at a location 1 − z. Whenever Re z < 0 (i.e. we are outside where the integral
definition holds), Re(1 − z) ≥ 1, so Γ(1 − z) can be evaluated. Again, Γ(z) has been continued over the
whole complex plane (omitting singularities). 

The next example requires a convergence theorem for Dirichlet series:

P∞
Theorem 3.2.2 The regions of convergence and divergence for a Dirichlet series n=1 annz are half-planes
Re z > α and Re z < α respectively, with α the least value that makes the right half-plane singularity free.
Example 3.2.4 Continue the Riemann zeta function
∞
X 1
ζ(z) = (3.4)
n=1
nz

with use of the functional equation

1
ζ(z) = 2z π z−1 sin( πz)Γ(1 − z)ζ(1 − z) (3.5)
2
The functional equation (which is proved be contour integration in Section ??) will immediately (just as was
the case for the Gamma function in the above example) give the continuation everywhere that Re z ≤ 21 if
values are available for Re ≥ 12 . By Theorem 3.2.2, we can find the convergence boundary by inspecting
P∞ ´∞
the sum (3.4) for z = x real. Comparing the sum n=1 n1x with the integral 1 tdtx = x−1 1
shows that (3.4)
converges for Re z > 1. What remains for us is therefore to somehow continue (3.4) from Re z > 1 down to
Re z = 1/2; then (3.5) completes the task of continuation to the whole complex plane.
One idea for this is to turn the sum in (3.4) into an alternating one. From
1 1 1 1 1 1
ζ(z) = z
+ z + z + z + z + z + ...
1 2 3 4 5 6
follows
2 2 2 2
z
ζ(z) = z + z + z + ...
2 2 4 6
Subtracting this second series from the first one gives
  1 1 1 1 1 1
1 − 21−z ζ(z) = z − z + z − z + z − z + −...
1 2 3 4 5 6
This new sum is again of Dirichlet type, and it therefore suffices to test convergence for z = x real. Since the
sum in this case alternates and the magnitudes of the terms decay for all x > 0, it will, by Theorem 3.2.2,
also convergence for all z with Re z > 0. The continuation of 3.4 is complete. 

As a sideline, we can note that

1 − 21−z 1 − e(1−z) log 2
lim = lim = − log 2
z→1 1 − z z→1 1−z
and  
1 1 1 1 1 1
lim − z + z − z + z − z + −... = log 2
z→1 1z 2 3 4 5 6
implying that limz→1 (z − 1)ζ(z) = 1. One can then deduce from (3.5) that the pole of ζ(z) at z = 1 is
the only singularity of the ζ(z)-function in the complex plane (and Palso that ζ(0) = − 12 ). We can further
∞
note a few other values of the zeta function. The value of ζ(2) = n=1 n12 is known as the Basel problem
P∞ 2
and turns out to be ζ(2) = n=1 n12 = π6 (as will be shown in Section ??). With this value for ζ(2), the
1
functional equation gives ζ(−1) = − 12 . Finding similar closed form expressions for ζ(k) when k = 3, 5, 7, . . .
is an unsolved problem.
3.2. SOME METHODS FOR ANALYTIC CONTINUATION 49

3.2.4 Partitioning of an integration interval

Example 3.2.5 Extend Γ(z) (again) from its integral definition (2.17), but without use of a functional
relation.
´∞
In the integral definition Γ(z) = 0 e−t tz−1 dt, the upper integration limit causes no problem (thanks to the
fast decay of the exponential function). The limitation Re z > 0 comes from what happens at the origin.
Thus, we split the interval in two parts:
ˆ 1 ˆ ∞
Γ(z) = e−t tz−1 dt + e−t tz−1 dt
| 0 {z } |1 {z }
Analytic for Re z > 0; Continue this part Analytic for all z; Entire function.

On a finite interval, the Taylor expansion of the exponential function is particularly fast converging:
ˆ 1 ˆ 1 ∞
! ∞ ˆ 1
z−1 −t z−1
X (−t)n X (−1)n
t e dt = t dt = tz−1+n dt
0 0 n=0
n! n=0
n! 0
| {z }
1
= x+n if Re z > −n

So it might not look like we have achieved anything extra. The first term in the sum has n = 0, so the
restriction would still seem to be Re z > 0. However, let us ignore this and consider the function
∞ ˆ ∞
X (−1)n 1
η(z) = + e−t tz−1 dt (3.6)
n=0
n! z + n 1
| {z } | {z }
Entire function.
Converges for all z. Poles at z = 0, −1, −2, ...

Since η(z) = Γ(z) for Re z > 0, η(z) must be the continuation of Γ(z) to the whole complex plane. 

3.2.5 Replace Taylor coefficients by integrals or sums

Again, we describe the approach by means of an example:

Example 3.2.6 Continue the function

∞
X zn
f (z) = √ .
n=1
n

The Taylor series converges for |z| < 1 and diverges for |z| > 1. We start by noting that a simple change of
√ ´∞ 2 ´∞ 2
variable in the relation 2π = 0 e−x dx gives √1n = √2π 0 e−nx dx. Therefore

∞ ˆ ˆ ∞ ∞
! ˆ
2 X n ∞ −nx2 2 X
n −nx2 2z ∞
dx
f (z) = √ z e dx = √ z e dx = √ .
π n=1 0 π 0 n=1
π 0 ex2 − z

In the last equality, we made use of the fact that the sum inside the integral is a geometric series and therefore
can easily be summed in closed form. The integral form of f (z) is reminiscent of the Stieltjes integral, which
we have come across earlier. It defines f (z) everywhere away from a branch cut along the line z ≥ 1 .

3.2.6 Subtraction of a similar series or integral

zk
P∞
Example 3.2.7 Continue the function f (z) = k=0 1+2−k .

The
P∞ original series diverges for |z| > 1. For large k, the terms become very close to those of the series
k 1
k=0 z = 1−z . This makes it natural to look at the difference between the two expansions. After minor
simplifications:
∞
1 X (z/2)k
f (z) − =− = −f (z/2). (3.7)
1−z 1 + 2−k
k=0
50 CHAPTER 3. ANALYTIC CONTINUATION

So we stumbled on to a functional equation

1
f (z) − = −f (z/2). (3.8)
1−z

Since the original sum works for |z| < 1, (3.8) allows us to compute f (z) for |z| < 2; then repeating again,
for |z| < 4, etc. The function is therefore continued to the whole complex plane. We can note that it will
have poles at z = 1, 2, 4, 8, ...
1 1
It is tempting to write (3.8) as f (z) = 1−z − f (z/2) and apply this repeatedly, to obtain f (z) = 1−z −
1 1
1−z/2 + 1−z/4 − + . . . However, this sum diverges for all z. A better idea is to repeat the original subtraction
idea, first on the sum in (3.7) and then on the sum this gives rise to, etc. After some manipulations, this
gives  
1 1 1/2 1/4 1/8
f (z) = + z − + − + ... , (3.9)
2 1−z 1 − z/2 1 − z/4 1 − z/8
which converges rapidly for all values of z. 

3.2.7 Borel summation

P∞
Assume that we are given a function f (z) defined by a Taylor series f (z) = k=0 ak z k that it is convergent
in |z| < R. Then
∞
X ak z k
φ(z) =
k!
k=0

is an entire function. Next, we form the function

ˆ ∞ ˆ ∞ ∞ ∞ ˆ ∞ ∞
X ak z k tk X e−t tk X
η(z) = e−t φ(zt)dt = e−t dt = ak z k dt = ak z k = f (z)
0 0 k! k!
k=0 k=0 |0 {z } k=0
= 1 for all k

This new function η(z) clearly agrees with f (z) within the domain of convergence |z| < R. But is it a
continuation of f (z)? It can be shown that η(z) will converge in the smallest polygon that can be constructed
from the singularities of f (z) in the way that is indicated in Figure 3.5. In this schematic figure, the solid
dots mark the singularities of f (z) and the circle shows where the Taylor expansion around the origin will
converge. Through each singularity, we draw a line orthogonal to the direction to the origin. The function
η(z) will converge inside the resulting polygon (marked by solid lines in the figure).
P∞
Example 3.2.8 Apply Borel summation to f (z) = k=0 z k = 1 + z + z 2 + z 3 + ....
P∞ zk z
We get immediately φ(z) = k=0 k! = e , and therefore the Borel extension takes the form f (z) =
´ ∞ −t ´ ∞ −t(1−z)
0
e φ(zt)dt = 0 e dt.This has brought us from f1 (z) to f2 (z) to in Example 3.1.1. As we noted
there, the new form converges for Re z < 1, entirely agreeing with the description above of where integrals
obtained through Borel summation will converge. 

3.2.8 Padé approximations

This methodology can be extremely powerful, especially if all that is known of a Taylor expansion is the
numerical values of a finite number of leading coefficients. It is very well suited for numerical computing,
although its convergence theory is not fully understood.
Taylor expansions (and often also asymptotic expansions, cf. Chapter 10) can often be accelerated quite
dramatically (or turned from divergent to convergent) by being rearranged into a ratio of two such expansions.
A Padé approximation
PN
N an z n
PM (z) = Pn=0
M
(3.10)
n
n=0 bn z
3.2. SOME METHODS FOR ANALYTIC CONTINUATION 51

Figure 3.5: Schematic illustration of the region of convergence of a Borel sum.

(normalized by b0 = 1) generalizes the Taylor expansion with equally many degrees of freedom

M
X +N
TM +N (z) = cn z n , (3.11)
n=0

with the two being the same in case M = 0. The Padé coefficients are normally found by starting from a
Taylor expansion:
a0 + a1 z + a2 z 2 + ...
c0 + c1 z + c2 z 2 + ... =
1 + b1 z + b2 z 2 + ...

and then require that both sides match to the highest degree possible at the origin. Multiplying up the
denominator gives the following equivalent set of coefficient relations

a0 = c0 (3.12)
a1 = c1 + c0 b1
a2 = c2 + c1 b1 + c0 b2
a3 = c3 + c2 b1 + c1 b2 + c0 b3
...

With the ci given, each new line introduces two new unknowns, ai and bi . The system would appear to be
severely under-determined. However, if we specify the degree of the numerator to be N , of the denominator
to be M , and of the truncated Taylor expansion to be M + N , there will be just as many equations as
unknowns (ignoring all terms that are O(z M +N +1 )). We can then solve for all the unknown coefficients, as
the following example shows:

Example 3.2.9 Given T5 (z), determine P32 (z).

In this case of M = 3, N = 2, M + N = 5, the system (3.12) becomes ’cut off’ as follows

52 CHAPTER 3. ANALYTIC CONTINUATION

a0 = c0
a1 = c1 + c0 b1
 a2 = c2 + c1 b1 + c0 b2 
 0 = c3 + c2 b1 + c1 b2 + c0 b3 

No more a’s available 0 = c4 + c3 b1 + c2 b2 + c1 b3  No more b’s available

0 = c5 + c4 b1 + c3 b2 + c2 b3
 
past limit O(x2+3+1 )
In the simplest (which is not the most robust) version of the Padé method, we first solve the bottom three
equations for b1 , b2 , b3 , after which the top three explicitly give a1 , a2 , a3 . This same idea carries through
for any values of M and N. .

A key usage of Padé approximations is to extract the information from power series expansions with only
a few known terms. Transformation to Padé form usually accelerates convergence, and often allows good
approximations to be found even well outside a Taylor expansion’s radius of convergence (which might even
be R = 0).

Example 3.2.10 Find the increasing order Padé approximations for f (z) = 1 − z + z 2 − z 3 + −....

The Padé table based on the truncated Taylor sums becomes as shown in Table 3.1. The main diagonal (and
the diagonal below it) usually gives the best results. The example above was trivial (and untypical) in that
every entry with M > 0 happened to recover the exact result. 

Table 3.1: Beginning of Padé table for f (z) = 1 − z + z 2 − z 3 + −...

N - order of numerator
0 1 2 3 ...
0 1 1 − z 1 − z + z 2 1 − z + z 2 − z 3 ...
M - order of 1 1 1
1 1+z 1+z 1+z ...
denominator 1 1
2 1+z 1+z ...
1
3 1+z ...
... ...

Example 3.2.11 Approximate f (2) when we only know the first few terms in the expansion f (z) = 1 − 21 z +
1 2 1 3 1 4 ln(1+z)
3 z − 4 z + 5 z − +... (= z , but convergent only if |z| < 1).

The Padé table 3.2 is laid out like Table 3.1, but it shows only the numerical values for z = 2 and, in
parenthesis, the errors in these compared to 21 log 3 ≈ 0.5493. In spite of z = 2 being well outside the domain
of convergence for the Taylor series for f (z) (as is again visible from the top M = 0 row in the table), the
Padé method allows fast and accurate calculation of the analytically continued result at z = 2.
´ ∞ e−t
Example 3.2.12 Compare Taylor- and Padé approximations for the Stieltjes’ function f (z) = 0 1+zt dt.

The integral defining f (z) is singular for z < 0 (real and negative), but it is well defined for other values of
z in the complex plane. Figure 3.6 (a) shows the result of a direct evaluation of Im(f (z)); the presence of a
branch jump along the negative real axis is obvious. If we Taylor expand f (z) around z = 0 (for example by
repeated integration by parts, or by noting that f (z) satisfies z 2 f 0 (z) + (1 + z)f (z) − 1 = 0, f (0) = 1, and
then equate coefficients), the resulting expansion becomes
∞
X
f (z) ≈ (−z)k k! .
k=0

This diverges for all values of z 6= 0 (its radius of convergence is R = 0). Truncation after the sixth power
gives
T6 (z) ≈ 1 − z + 2z 2 − 6z 3 + 24z 4 − 120z 5 + 720z 6 .
3.2. SOME METHODS FOR ANALYTIC CONTINUATION 53

Table 3.2: Values for f (2) from Padé approximations; in prentheses their difference to 12 log 3.
N - order of numerator
0 1 2 3 4 ...
1 0 1.3333 −0.6667 2.5333
0
(0.4507) (−0.5493) (0.7840) (−1.2160) (1.9840)
0.5714
1
(0.0221)
M - order of
0.5507
denominator 2
(0.0014)
0.5494
3
(0.0001)
0.5493
4
(0.0000)
... ... ...

Predictably, this partial Taylor sum gives nonsense when evaluated numerically, as is shown for the imaginary
part in Figure 3.6 (b). However, when T6 (z) is converted to the Padé approximation

1 + 11z + 26z 2 + 6z 3
P33 (z) = (3.13)
1 + 12z + 36z 2 + 24z 3
we recover a somewhat respectable approximation of the original function (Figure 3.6 (c)); the rational
approximation has even arranged for its pole singularities to appear along the negative real axis in an
attempt to mimic the branch discontinuity there - a bit surprising since the location of a branch cut is not
uniquely determined). Finally, Figure 3.6 (d) compares, along the positive real axis, the true function f (z)
(dashed) against T6 (z) and P33 (z).
In this particular case of the Stieltjes’ function, it can be proven that the Padé approximations will
converge to the true function exponentially fast (as higher degrees are used) everywhere in the complex
plane away from the negative real axis. Somehow, the everywhere divergent Taylor expansion does contain
complete information about the function, and the Padé approach allows it to be recovered.

3.2.9 Ramanujan’s formula

This method is different from the previous ones in that, technically we are not supplying enough information
to determine the analytic function uniquely. Ramanujan noted that, if function values are given for non-
negative integers, then we can construct an analytic function which obeys these by means of the formula
ˆ ∞
π
tz−1 φ(0) − tφ(1) + t2 φ(2) − +... dt


φ(−z) = (3.14)
sin πz 0

It is easy to see that both φ(z) ≡ 0 and φ(z) = sin πz satisfy φ(0) = φ(1) = φ(2) = ... = 0. However, if
we also know that |φ(z)| < CeA|z| with A < π holds when Re z > 0, it can be shown that (3.14) uniquely
determines φ(z). A typical use of (3.14) would be to determine an unknown function φ(z) based on values
for φ(0), φ(1), φ(2), .... In the two examples below, we enter instead known functions, and obtain quite
nontrivial formulas as results.

Example 3.2.13 Substitute φ(k) = 1, k = 0, 1, 2, . . . into Ramanujan’s formula (3.14).

Obviously, φ(z) ≡ 1 will satisfy the given data (and also the growth condition for the right half plane).
Equation (3.14) then gives the relation
ˆ ∞ z−1
t π
dt = .
0 1 + t sin πz
1
Example 3.2.14 Substitute φ(k) = k! , k = 0, 1, 2, . . . into (3.14).
54 CHAPTER 3. ANALYTIC CONTINUATION

´∞ e−t
(a) Imaginary part of f (z) = 0 1+zt dt. (b) Im T6 (z) (note vertical scale)

2.0

1.5

1.0

0.5

0.0 0.5 1.0 1.5 2.0 2.5 3.0

(c) Im P33 (z). (d) Comparison for z real; f (z), T6 (z), P33 (z).

Figure 3.6: Stieltjes’ function and its Taylor and Padé approximations.
3.3. SELECT PROOFS 55

We arrive similarly at ˆ ∞
π
e−t tz−1 dt = .
0 Γ(1 − z) sin πz
Since we already recognize the integral as Γ(z), we have arrived at (3.3) .

An alternative version of Ramanujan’s formula is

ˆ ∞
t2
 
z−1 t
Γ(z)φ(−z) = t φ(0) − φ(1) + φ(2) − +... dt (3.15)
0 1! 2!

3.3 Select proofs

3.4 Exercises
P∞
Exercise 3.4.1 Consider the Taylor expansion f (z) = n=0 (−z)n .
(a) Determine the radius of convergence R for the Taylor expansion of the series above.
1
(b) Using only the Taylor series coefficients for f (z) (i.e. not the fact that it sums to 1+z ), re-expand f (z)
1
around the point z = 2 , in order to obtain the expansion
∞  n+1  n
X 2 1
f (z) = (−1)n z−
n=0
3 2

(c) From the expansion coefficients in part 2), determine the radius of convergence for the new expansion.
Draw in the complex plane where the two expansions converge, and mark the point z = P−1, which happens to
∞ k! 1
be the only singularity of the continued function. Note: You can assume it known that k=m (−1)k m!(k−m)! 2k−m
=
m 2 m+1


(−1) 3
P∞ k
Exercise 3.4.2 Consider the function f (z) = k=0 z (2 ) ,
(a) Determine the radius of convergence of the Taylor expansion for f (z).
(b) Show that f (z) satisfies the functional equation

f (z) = z + f (z 2 ).

(c) By means of the functional equation, show that the function cannot be continued outside the Taylor
series radius of convergence (i.e. its boundary forms a natural boundary for the function itself ). Hint: From
Pm−1 k
the functional equation, first deduce that f (z) = f (z 2m ) + k=0 z (2 ) holds for m = 1, 2, 3, . . . The result
then follows by a similar argument to the one used in Example 3.1.2.

Exercise 3.4.3
(a) Determine the radius of convergence of the Taylor series
∞
X
f (z) = (−1)k k!z k . (3.16)
k=0

(b) Use Borel summation to obtain from (3.16) the formula

ˆ ∞ −t
e
f (z) = dt, (3.17)
0 1 + zt

and determine the range of validity for this integral representation of f (z).
(c) Use instead the method from ´Section 3.2.5 to obtain (3.17) from (3.16). Hint: Note from the definition
∞
of the Gamma function that k! = 0 e−t tk dt.
56 CHAPTER 3. ANALYTIC CONTINUATION

(d) Starting from (3.17), derive (3.16) by repeated integration by parts.

(e) Again, start from (3.17) and arrive at (3.16) by the different method of first showing that (3.17) implies
that f (z) satisfies the ODE
z 2 f 0 (z) + (1 + z)f (z) − 1 = 0 (3.18)
with initial condition f (0) = 1. Then assume that f (z) has a Taylor expansion

f (z) = a0 + a1 z + a2 z 2 + a3 z 3 + ...

and use (3.18) to determine the unknown coefficients.

Exercise 3.4.4
(a) By a change of variable in the integral definition of the Gamma function, derive
ˆ ∞
1 1
= tz−1 e−nt dt.
nz Γ(z) 0

(b) Following one of the ideas in the analytic continuation notes, show that the Riemann zeta-function
∞
X 1
ζ(z) = z
n=1
n

can be rewritten as ˆ ∞
1 tz−1
ζ(z) = dt.
Γ(z) 0 et − 1

(c) Determine for what values of z the integral in part (b) converges.

Exercise 3.4.5 From two sheets of paper and some Scotch tape (or similar), construct a Riemann surface
for the real part of the function f (z) = z 1/2 . Mark on the top of it the real and the imaginary axis.
Chapter 4

Complex Integration

Real-valued case:

In real variable calculus, we have, as in Figure 4.1 (a)

ˆ b X
f (x)dx = lim f (xk )∆xk
a ∆xk →0
k

´b
The key theorem states that, if there is a function F (x) such that F 0 (x) = f (x), then a f (x)dx = F (b)−F (a).
Finding such a function F (x) (which may not exist in terms of elementary functions) is often the only means
available for evaluating integrals analytically.

Complex-valued case:

In the complex plane, we have to consider, as in Figure 4.1 (b)

ˆ zb X
f (z)dz = lim f (zk )∆zk .
za ∆zk →0
k

There are now two major novelties:

(i) The result contains nothing about the path Γ (also known as contour), as long as F (z) is single-
d
valued. We assumed here that there exists a function F (z) such that dz F (z) = f (z).

(ii) By utilizing the path independence, it will turn out that many definite integrals can be evaluated
in closed form, without us needing to obtain any closed form expression for F (z) (the topic of
Residue Calculus in Section 4.4).

4.1 Integration when a primitive function F (z) is available

´ zb
Theorem 4.1.1 If there exists a function F (z) such that F 0 (z) = f (z), then za
f (z) dz = F (zb ) − F (za ).

57
58 CHAPTER 4. COMPLEX INTEGRATION

Im z

zb
y
y = f (x)

Re z
za
a ∆x k b x

∆zk

(a) Integration in real variables. (b) Integration in complex variables.

Figure 4.1: Conceptual difference between integrating in a real and a complex variable.

Proof
ˆ ´ z+∆z ´z
z f (ξ)dξ − f (ξ)dξ
d za za
f (ξ)dξ = lim
dz za ∆z→0 ∆z
´ z+∆z
f (ξ)dξ
z
= lim
∆z→0 ∆z
f (z)∆z
= lim
∆z→0 ∆z
= f (z)
d
= F (z).
dz
Therefore,  ˆ z 
d
F (z) − f (ξ)dξ =0
dz za

which implies ˆ z
F (z) − f (ξ)dξ = C,
za

where C is a constant. Setting z = za , we get C = F (za ), and setting z = zb , we obtain

ˆ zb
f (ξ)dξ = F (zb ) − F (za ). 
za

´ z as in Figure 4.2, where za = zb . We expect that F (zb ) − F (za ) = 0 if F (z)

Next, consider a closed loop
is single-valued. Therefore, zab f (ξ)dξ = 0 if F (z) exists and is single-valued inside or on Γ, (i.e. if F (z) has
no branch point inside Γ).
We made above the assumption that F (z) exists. This can be justified by considering a circle-chain
continuation of f (z) along the integration path. We can then can integrate each of these Taylor expansions
(with one free constant, as to be expected). Recall that the radius of convergence never changes if we
integrate or differentiate a Taylor series. So F (z) exists along a path if f (z) exists along it.
4.1. INTEGRATION WHEN A PRIMITIVE FUNCTION F (Z) IS AVAILABLE 59

Im z

Re z

za = z b

¸
Figure 4.2: Integration along a closed contour Γ is denoted by Γ
f (z)dz.

We have not yet

¸ proven that Taylor series converge to the nearest singularity. In fact, that proof will use
the properties of f (z)dz. So, without invoking the existence of some F (z), we would like to arrive at the
result ˛
f (z)dz = 0,
Γ
where Γ a closed loop containing no singularity. We accomplish this in the following Cauchy’s theorem.

4.1.1 Cauchy’s and Morera’s theorems

¸
Theorem 4.1.2 (Cauchy) Γ
f (z)dz = 0 when f (z) is an analytic function, and there is no singularity
inside the contour Γ.

Proof
˛ ˛
f (z)dz = (u + iv) (dx + idy)
Γ Γ | {z } | {z }
f (z) dz
˛ ˛
= (udx − vdy) + i (udy + vdx)
Γ Γ

Recall from multivariable calculus Green’s theorem: Let g(x, y) and h(x, y) and their first partial derivatives
be continuous in a domain D bounded by a closed loop Γ. Then,
˛ ¨  
∂h ∂g
(gdx + hdy) = − dxdy
Γ D ∂x ∂y
In our case,
˛ ˛ ˛
f (z)dz = (udx − vdy) + i (udy + vdx)
Γ Γ
¨  Γ ¨  
∂v ∂u ∂u ∂v
= − + dxdy + i − dxdy
D ∂x ∂y D ∂x ∂y
| {z } | {z }
=0 by C-R =0 by C-R
= 0 

Morera’s theorem is the reverse to Cauchy’s theorem.

60 CHAPTER 4. COMPLEX INTEGRATION

¸
Theorem 4.1.3 (Morera) If f is continuous and if Γ
f (z)dz = 0 for any contour Γ in the domain, then
f (z) is an analytic function.
A proof is given in Section 4.8. The main themes in this chapter will be the following:
¸
• Evaluating Γ f (z)dz along a certain contour Γ might be problematic. We will learn how to change
paths to make the evaluation of the integral easier.
• We will see what happens when the closed loop Γ actually does contain a singularity.
´b
• If we want to evaluate a real integral a f (x)dx, we will see that we can modify the path and instead
evaluate the integral in the complex plane.
A combination of these techniques will lead to very powerful methods to evaluate integrals in cases where we
cannot find F (z). These methods also lead to accurate estimates integrals in various singular limit situations,
as discussed in Chapter 10 Asymptotic analysis of integrals.

4.2 Contour integration

We start by an often useful estimate
Theorem 4.2.1 The integral along a path Γ can be bounded from above by
ˆ 
 
 f (z)dz  ≤ M · L,
 
Γ

where M is the upper bound of |f (z)| on Γ and where L is the length of Γ.

Proof
ˆ   X 
   
 f (z)dz  =  lim f (zk )∆zk 
  ∆zk →0
Γ
X
≤ lim |f (zk )| |∆zk |
∆zk →0 | {z }
≤M
X
≤ M lim |∆zk |
∆zk →0
| {z }
=L
≤ M ·L 
Example 4.2.1 Evaluate ˆ
z̄dz
Γ
where Γ is the path shown in Figure 4.3.
Note: f (z) is NOT an analytic function, so we know from Morera’s theorem that the result will be be path
dependent. Also, the contour is not a closed loop, so Green’s theorem is not available to us. We thus need
to work out this integral one line segment at a time.
f (z) = z̄
= x − iy
So
ˆ ˆ 1 ˆ 1
z̄dz = xdx + (1 − iy)(idy)
Γ 0 0
1 1
x2 y2
 
= +i y−i
2 0 2 0
= 1+i 
4.2. CONTOUR INTEGRATION 61

Im z

1+i

0 1 Re z

Figure 4.3: Path for the integral in Example 4.2.1.

Γ Γ

(a) Path for the integral in Example 4.2.2 (b) Modified contour, with a red dot marking the pole at the origin.

Figure 4.4: Different integration paths considered in Example 4.2.2. Note in part (b) that we have marked
the path direction for the small circle in the positive direction (causing the minus sign in equation (4.1)).

Example 4.2.2 Evaluate ˆ

z n dz
Γ

where n is an integer and Γ is any closed contour that goes around the origin once in the positive (counter-
clockwise) direction.

We will show two ways to solve the problem:

1. by changing the path and evaluating the integral explicitly around the simplified path

2. by using the primitive form F 0 (z) = f (z).

´
1. Modifying the Contour: If n ≥ 0, then f (z) is analytic inside Γ, which leads to Γ z n dz = 0.
However, if n ≤ 0, then there is a pole at the origin. In that case, one can consider a different contour,
such as the one represented in Fig 4.4 (b). This new contour excludes the origin, and thus the pole. It will
62 CHAPTER 4. COMPLEX INTEGRATION

integrate to zero:
˛ ˆ ˆ ˆ ˆ 
f (z)dz = + − + f (z)dz (4.1)
Γ ← γ →
= 0
´ ´
Since one can choose the paths → and ← arbitrarily close to each other, ←
f (z)dz and →
f (z)dz cancel
out. Therefore ˛ ˛
f (z)dz = f (z)dz ,
γ Γ
where both are followed in the positive direction. The argument shows that we, in general, can move the
contour as we want without changes to the value of the integral, as long we don’t move it across any singular
point. We can thus integrate around the small circle γ centered at the origin with radius R. Let z = Reiθ ,
and therefore dz = iReiθ dθ. Then
˛ ˛
z n dz = z n dz
Γ γ
ˆ 2π
= Rn einθ iReiθ dθ
0
ˆ 2π
= iRn+1 ei(n+1)θ dθ
0
1 h i2π
= iRn+1 ei(n+1)θ
i(n + 1) 0
= 0 if n 6= −1
However, if n = −1, then
˛ ˛
z −1 dz = z −1 dz
Γ γ
ˆ 2π
= R−1 e−iθ iReiθ dθ
0
ˆ 2π
= idθ
0
= 2πi

n+1
2. Finding the primitive function F (z): Let f (z) = z n . If n 6= −1, then F (z) = zn+1 . Therefore, after
having followed the closed contour Γ, F (zend ) = F (zstart ) since F (z) is single valued. Therefore,
˛
z n dz = 0.
Γ

If n = −1, then F (z) = log(z). After having gone through one cycle of the contour, the value of F (z) must

have increased by 2πi. Therefore, ˛

z n dz = 2πi.
Γ
We arrive at the same conclusion as before. 
The above result can alternatively be formulated as
Theorem 4.2.2
˛

 0 if z = z0 is outside Γ
1 dz
= 0 if z = z0 is inside Γ and n 6= 1
2πi Γ (z − z0 )n
1 if z = z0 is inside Γ and n = 1


for n integer.
4.2. CONTOUR INTEGRATION 63

4.2.1 Cauchy’s argument principle

Example 4.2.3 Let p(z) be a polynomial of degree n. Evaluate the following integral
˛ 0
1 p (z)
I= dz .
2πi Γ p(z)

Let p(z) have the following form

p(z) = A(z − a1 )(z − a2 )...(z − an )

Therefore
p0 (z) d
= log (p(z))
p(z) dz
d
= (logA + log(z − a1 ) + log(z − a2 ) + ... + log(z − an ))
dz
1 1 1
= + + ... + (4.2)
z − a1 z − a2 z − an

Therefore, the value of I must equal the number of roots lying inside Γ. We can thus tell the number of
roots inside a contour without inspecting the function inside it. 
This result generalizes in various ways, to Cauchy’s argument principle (next) and to Rouché’s theorem
(Theorem 4.4.4); see also Gauss-Lucas theorem (Exercise 4.9.9).

Theorem 4.2.3 (Cauchy’s argument principle) If f (z) is a meromorphic function (its only singularities are
poles), and if it has N zeros and P poles as only singularities (accounting for their multiplicities) inside the
contour Γ, then ˛ 0
1 f (z)
N −P = dz.
2πi Γ f (z)

Proof We write f (z) as

QN
(z − ak )
f (z) = Qk=1
P
g(z)
k=1 (z − bk )
where g(z) has neither zeros nor poles within Γ. Then

f 0 (z) d
= log (f (z))
f (z) dz
N P
!
d X X
= log(z − ak ) − log(z − bk ) + log g(z)
dz
k=1 k=1
N P  
X 1 X 1 function that is singu-
= − + ,
z − ak z − bk laratity free within Γ
k=1 k=1

and the result follows when integrating around Γ. 

We will re-visit this result in Section 4.4.4 describing Rouche’s theorem and winding numbers.

4.2.2 Cauchy’s integral formula

Theorem 4.2.4 Let f (z) be analytic inside Γ. Then,
˛
1 f (ξ)
f (z) = dξ. (4.3)
2πi Γ ξ−z
64 CHAPTER 4. COMPLEX INTEGRATION

Significance

1. We have mentioned before that f (z) is completely determined by its values along any curve segment.
So the existence of a formula of this kind is not altogther surprising. If we know its values along a
whole closed contour, Cauchy’s integral formula provides its values at points inside the contour in a
remarkably simple way.

2. We will soon deduce from it that if f (z) is analytic, it is not just once but infinitely many times
differentiable.

3. This formula forms the foundation for a much more solid understanding of Taylor expansions, and
their generalization to Laurent expansions .

Proof As in Figure 4.4, we modify our contour Γ to become a circle centered at z with radius δ, and call
this new contour γ. Thus
˛ ˛
1 f (ξ) 1 f (ξ)
dξ = dξ
2πi Γ ξ − z 2πi γ ξ − z
˛
1 (f (ξ) − f (z)) + f (z)
= dξ
2πi γ ξ−z
˛ ˛
1 (f (ξ) − f (z)) 1 dξ
= dξ +f (z)
2πi γ ξ−z 2πi γ ξ − z
| {z } | {z }
=Iδ =1

Consider the first of the integrals Iδ , and let δ become arbitrarily small. Then, the continuity of f implies
that |f (ξ) − f (z)| <  whenever |z − ξ| < δ. Thus (by Theorem 4.2.1)

|Iδ | ≤ · |{z}
2πδ
δ
|{z} L
M
= 2π
→ 0, when δ → 0

Therefore,
˛
1 f (ξ)
dξ = f (z) 
2πi Γ ξ−z

Theorem 4.2.5 If f (z) is analytic (once complex differentiable) in some region, it can then be differentiated
any number of times.

Proof The RHS of (4.3) clearly possesses any number of derivatives, and then, so must the LHS. In fact,we
obtain the following formula for the k th derivative:
˛
(k) k! f (ξ)
f (z) = dξ (4.4)
2πi Γ (ξ − z)k+1

4.2.3 Liouville’s theorem

Theorem 4.2.6 If f (z) is entire (no singularity for finite z) and bounded (|f (z)| ≤ M ) then f (z) ≡ C, for
C constant.

Proof Equation (4.4) becomes for the first derivative (k = 1)

˛
1 f (ξ)
f 0 (z) = dξ
2πi Γ (ξ − z)2
4.2. CONTOUR INTEGRATION 65

and we choose as the contour a circle of radius R centered at the point z. Then,
˛ 
0 1  f (ξ) 
|f (z)| = dξ 
2π  Γ (ξ − z)2 
1 M M
≤ 2πR = .
2π R2 R
Since M is a constant and R can be arbitrarily large, this inequality is a contradiction unless f 0 (z) = 0.
From this follows then that f (z) must be a constant. 

4.2.4 The fundamental theorem of algebra

Theorem 4.2.7 A polynomial p(z) of degree n ≥ 1 has at least one root.

Proof Suppose that p(z) has no root and that n ≥ 1. We wish to construct a contradiction. To this end,
1
we consider f (z) = p(z) . It satisfies the two conditions needed to apply Liouville’s theorem:
1
1. p(z) is an entire function (since p(z) has no root).
1
2. p(z) is bounded since |p(z)| → ∞ as |z| → ∞.

By Liouville’s theorem, p(z) ≡ const. This violates the assumption that p(z) is a polynomial of degree n ≥ 1.
The hypothesis that p(z) has no root must therefore have been false. 

After dividing out the root we now know must have existed, the argument can be repeated. A polynomial
pn (z) of degree n will therefore always have exactly n roots (not necessarily at different locations from each
other).

4.2.5 Mean value theorem

Theorem 4.2.8 The Mean Value Theorem:
ˆ 2π
1
f (z) = f (z + Reiθ )dθ (4.5)
2π 0

(assuming no singularity inside the circle).

Proof Use Cauchy’s integral formula

˛
1 f (ξ)
f (z) = dξ
2πi Γ (ξ − z)

and let the contour be a circle centered at z with radius R. A change of variables yields ξ − z = Reiθ and
dξ = iReiθ dθ. Thus ˆ 2π
1
f (z) = f (z + Reiθ )dθ 
2π 0
This last theorem states that the value of an analytic function at any interior point z is that average of the
function integrated over a circle centered at z. One can separate the real from the imaginary part in the
above equation, so the theorem holds also for Re f (z) and for Im f (z) separately.

4.2.6 Revisiting Max/Min theorems

Theorem 4.2.9 Max/Min theorems. u(x, y) = Re f (z) and v(x, y) = Im f (z) cannot have a local (finite)
maximum or minimum.

Proof Assume that the real part u(x, y) reaches a local maximum at some point z. However the value of u
at that point is the average of the values of u around a circle centered at z. This is a contradiction. 
66 CHAPTER 4. COMPLEX INTEGRATION

Theorem 4.2.10 Can |f (z)| attain a local minimum or maximum? The answer is: No except at at a
z-value for which |f (z)| = 0 and at a singular point.

Proof Assume that |f (z)| attains a local min/max that it is neither 0 nor ∞. Then log(|f (z)|) also attains
a local min/max since log is a monotonic function. However, log(|f (z)|) is the real part of the function
log(f (z)), which is analytic. By the theorem above, this is impossible.

Somewhat related to this theorem is whether anything can be said if it only known that a function f (z)
is analytic (without singularities) in a half-plane, and bounded along its edge.

Theorem 4.2.11 (Phragmén-Lindelöf) Let f (z) be analytic (singularity free) in the right half-plane Re z ≥
β
0 and |f (iy)| < M (constant) along the imaginary axis. Then, if |f (z)| < K e|z| for some constant K and
β < 1, in the half-plane, it will hold everywhere across it that |f (z)| < M .
γ
Proof Define F (z) = f (z) s−ε z where ε > 0 is arbitrarily small, and β < γ < 1. The maximum principle
tells that |F (z)| < M throughout the half-plane, and the result follows by letting ε → 0. 

Ruoghly speaking, the theorem tells that the function must either grow exponentially fast or stay bounded
in the half-plane - nothing in-between can happen. There are many theorems of this general character, for
ex. for functions that are bounded along two intersecting lines z = r eiθ1 and z = r eiθ2 , r ∈ [0, ∞]. The
formulation above corresponds to θ2 − θ1 = π.

4.2.7 Revisiting Taylor expansions

4.2.7.1 Weierstrass M-Test
Theorem 4.2.12 Weierstrass M-Test. P Let |bj (z)| ≤ Mj , an upper
P bound for each analytic function bj (z),
for j = 0, 1, 2, 3, ... in some region R. If j Mj converges, then j bj (z) converges uniformly in R.

4.2.7.2Radius of Convergence
P∞
Theorem 4.2.13 If f (z) = j=0 bj z j converges for z0 6= 0, then it also converges for all |z| < |z0 |

Significance: The domain of convergence for a Taylor series must be a circle. Anything else would become
a contradiction.
P∞
Proof Assume that f (z) = j=0 bj z j converges for z0 6= 0. Then, from some index j ≤ J, |bj z0j | < 1. Then
also form this index,
 j
z
j
|bj z | = |bj z0j |  
z0 
 j
z
<  
z0
= Mj
P
P∞ |M |j < 1 since |z| < |z0 |. Since
with Mj converges, by the Weierstrass M-test, we can conclude that
j=0 bj z converges. 

In the following two theorems, we consider for notational simplicity expansions centered at z0 = 0.In the
summary in Section 4.3.1, z0 denotes an arbitrary point.,
P∞ j f (j) (0)
Theorem 4.2.14 If f (z) is analytic in |z| ≤ R, then its Taylor series f (z) = j=0 bj z with bj = j!
converges at each point |z| < R.

Significance: This theorem implies that a Taylor series always converges in the greatest circle up to the
nearest singularity.
4.3. LAURENT SERIES 67

Proof Let Γ be the circle of radius R, and |z| < R. By Cauchy’s integral formula,
˛
1 f (ξ)
f (z) = dξ
2πi Γ ξ − z
˛ !
1 1 1
= f (ξ) dξ.
2πi Γ ξ 1 − zξ
 
Since |ξ| = R and |z| < R,  zξ  < 1, which implies that
 

∞
!
1 1 X zj
z = j+1
ξ 1− ξ j=0
ξ

is a convergent series. Thus,

∞  ˛ ∞
f (j) (0) j

X 1 f (ξ)dξ j
X
f (z) = z = z
j=0 |
2πi Γ ξ j+1 j=0
j!
{z }
f (j) (0)
=bj = j!

is also convergent. 
The largest number R for which the power series converges inside the disk |z| < R is called the radius of
convergence.

4.3 Laurent series

P∞
Laurent series are the natural extension of a Taylor series. The Taylor series takes the form j=0 aj z j and
P∞ a
it converges when |z| < R. If we define t = z1 , we obtain a series j=0 tjj and it converges for |t| > R1 . If we
P∞
are given a sum j=−∞ cj z j , it would be natural to split it
   
∞
X −1
X X∞
cj z j =  cj z j  + c0 +  cj z j  ,
j=−∞ j=−∞ j=1

for which the region of convergence could be expected to be a circular annulus centered at the origin, with
a singularity on each of its bounding circles, cf. Figure 4.5.

Theorem 4.3.1 If f (z) is analytic in an annulus, then there is a Laurent series corresponding to it, con-
verging everywhere inside the annulus.

Proof Consider the annulus shaded in gray in Figure 4.6, and the integration path C = C1 + C2 , oriented
as shown, and with the radial segments adjacent to each other, i.e. with canceling integrals. Let z be a point
inside the annulus, and choose radii r1 and r2 such that R1 < r1 < |z| and R2 > r2 > |z|. By Cauchy’s
integral theorem,
ˆ
1 f (ξ)
f (z) = dξ
2πi C ξ − z
ˆ ˆ
1 f (ξ) 1 f (ξ)
= dξ + dξ. (4.6)
2πi C1 ξ − z 2πi C2 ξ − z

For the first integral, |z| > |ξ|, and

!
1 1 1 1 ξ ξ2
=− ξ
= − − 2 − 3 − ...
ξ−z z 1− z
z z z
68 CHAPTER 4. COMPLEX INTEGRATION

Im z

z2

z1
I II III IV
z0

z3

Re z

Figure 4.5: Regions of validity for the Taylor (I) and Laurent (II, III and IV) expansions centered at z0 with
singular points z1 , z2 and z3 .

Figure 4.6: {Figure to be created and inserted here, showing the paths used in the derivation of the Laurent
Theorem 4.3.1 }

converges. Likewise, for the second integral, |ξ| > |z|, and
!
1 1 1 1 z z2
= z = + 2 + 3 + ...
ξ−z ξ 1− ξ ξ ξ ξ

converges. Substituting these into (4.6) gives

ˆ ˆ
   
−1 +∞
X
 1 f (ξ)  n X  1 f (ξ)  n
f (z) = − dξ z + dξ z .
n=−∞
2πi ξ n−1 n=0
2πi ξ n−1
C1 C1

Here C1 was followed in the negative (clockwise) direction. If we follow both contours in the positive
direction, no further change occurs if we change both paths to become a single path going around once
counterclockwise within the annulus. We rename it C. 

4.3.1 Summary of Taylor versus Laurent series around a point z0

4.3.1.1 Taylor series

A Taylor series is valid out to the nearest singularity

∞ ˆ
X
n 1 f (ξ) f (n) (z0 )
f (z) = an (z − z0 ) with an = dξ = ,
n=0
2πi (ξ − z)n+1 n!
C

for any path C going around z0 once counterclockwise, inside the region of analyticity.
4.3. LAURENT SERIES 69

4.3.1.2 Laurent series

A Laurent series is valid in an annulus between singularities. Each annulus limited my two singularities is
associated with a separate Laurent expansion.
∞ ˆ
X
n 1 f (ξ)
f (z) = an (z − z0 ) with (again) an = dξ,
n=−∞
2πi (ξ − z)n+1
C

now for −∞ < n < +∞. The contour C loops around z0 once, counterclockwise, within the annulus in
(n)
question. In this Laurent case, there is no longer any counterpart to the closed form expression an = f n!(z0 )
from the Taylor case.

Example 4.3.1 Give the three expansions around the origin of the following function
1
f (z) = .
(z − 1)(z − 2)
Since there are two distinct poles, the three expansions will be
1. a Taylor series converging inside the disk of radius 1.
2. a Laurent series converging in the annulus between the circles of radius 1 and radius 2.
3. a Laurent series converging outside the circle of radius 2.
1 −1 1
First, note that one can easily find that partial fraction expansion of (z−1)(z−2) = (z−1) + (z−2) . Considering
1 1
the latter form f (z) = (1−z) + (z−2) will make the expansion process considerably easier.

1. The first expansion is a Taylor series converging in

Pa∞disk centered at the origin with radius 1. The first
1
function term (1−z) can be directly expanded as k=0 z k , which converges for |z| < 1 as required. The
1
second term, (z−2) 1
, can be rewritten as (z−2) = −1 1
2 (1− z2 ) . When |z| < 2, this term can be expressed
P∞
k
as k=0 z2 . Adding the expansions for both terms gives
∞  
X 1
f (z) = 1− zk
k=0
2(k+1)

2. The second expansion is a Laurent series converging in an annulus centered at the origin, between the
circles of radii 1 and 2. The first term of the function needs to be expanded outside of the unit circle.
1
We thus it as (1−z) = z1 (1−1 1 ) . The expansion that we found above for the second term is still valid in
z
the present annulus. The resulting series reads
∞
X
f (z) = ak z k
k=−∞

where 
−1 k ≤ −1
ak = 1
2k+1
k≥0

3. Finally, the third expansion is a Laurent series converging outside the circle of radius 2. The first term
1
of the function can be expanded as above. The second term can be rewritten as follows: (z−2) = z1 (1−1 2 )
z
to obtain a convergence in the requested region. The resulting series is
∞  k+1
X
k 1
f (z) = (2 − 1) 
z
k=0
70 CHAPTER 4. COMPLEX INTEGRATION

Im z

III
II
I

1 2 Re z

Figure 4.7: Regions of validity for the expansions in Example 4.3.1

1
Example 4.3.2 Expand f (z) = ez− z around z = 0.

This function has essential singularities at z = 0 and at z = ∞. We can thus find a Laurent expansion that
1
will converge between them (thus for all finite z except at z = 0). Direct Taylor expansion f (z) = ez− z =
2 3
1 + z − z1 + 12 z − z1 + 3! 1
z − z1 + ... suggests that, when expanding out the different powers in the

sum, we ought to end up to a Laurent-type expansion.

It is in this case quite tricky to obtain the Laurent coefficients more explicitly. However, one can proceed
1 1 P∞ zr
P∞ (−1)s z−s  P∞ P∞ s r−s
as follows: f (z) = ez− z = ez · e− z = r=0 r! s=0 s! = r=0 s=0 (−1)r!s!z . If we now set
P∞ P∞ s
r = n + s, i.e. r − s = n, the double sum can be re-arranged into n=−∞ cn z n where cn = s=0 s!(−1) (n+s)! .
This last sum turns out to appear appear also in the context of Bessel functions (cf. Section ??), and its
values can through that connection be given in closed form as cn = Jn (2). 

4.3.2 Revisiting singularities

We have now better tools to characterize singularities. The lowest index in a Laurent sum, centered at a
singularity, will give us information about the type of singularity with which we are faced. Let’s revisit the
four first types:

1. Removable Singularities: We saw earlier that a removable singularity is only an artifact of the
function representation. Any trace of this ’singularity’ completely disappears once the function has
been expanded into a power series centered at that point. Since the singularity has vanished, the
resulting expansion around this point is strictly a Taylor series.

2. Poles: A function exhibiting a N th order pole at a point z0 can be expanded into a Laurent series
centered at z0 . It will have the following form
∞
X
cj (z − z0 )j
j=−N

We call c−N the pole’s strength, the coefficient associated with the lowest index (−N ). Another
coefficient of interest is c−1 , which we call the residue.
4.3. LAURENT SERIES 71

3. Essential Singularities: An essential singularity is characterized by a Laurent series with cj 6= 0 for

arbitrarily large negative indices j. The Laurent series will thus take the form
∞
X
cj z j
j=−∞

Once again, we shall draw particular attention to the residue c−1 .

4. Branch Points: Like a Taylor series, a Laurent series can only produce single-valued results. In the
Laurent case, there may however be branch points inside the inner circle, as long as these cancel in the
sense that the result becomes single valued in the annulus of convergence.
Example 4.3.3 Describe the possible Taylor and Laurent expansions around the origin for the function
f (z) = (z 2 − 1)1/2 .
This function was illustrated in Figures 2.18 and 2.19. The branch points are located at z = ±1. If we are
interested in Taylor expanding around z = 0, we orient the branch cuts so they become directed outwards,
for ex. [−∞, −1] and [+1, +∞]. Each of the two solution sheets have now its own Taylor expansion, i.e. (by
the binomial expansion (1 + z)s = 1 + sz + s(s−1) 2
2! z +
s(s−1)(s−2) 3
3! s + . . .)
 

1/2 1 1 1
f (z) = ±i 1 − z 2 = ±i 1 − z 2 − z 4 − z 6 − . . . .
2 8 16
Regarding Laurent expansions for |z| > 1, we need to place the branch cut between +1 and −1 instead inside
the unit circle (which is possible since z = ∞ is not a branch point). There are then again two solution
sheets, each with a separate Laurent expansion:
 
1 1 1
f (z) = ± z − − 3− − . . . .
2z 8z 16z 5
As can be inferred from Figure 2.19, there is no one-to-one correspondence between the Taylor and the
Laurent versions. The Taylor expansion holds within the unit circle, and assumes the branch cut is away
from this region, as in parts (c) and (d) of the figure, whereas the Laurent expansions assume regularity
outside of the unit circle, as depicted in parts (a) and (b) of the figure. 
(z−1)2
Example 4.3.4 Describe singularities of f (z) = z(z+1)3 .
There are two pole singularities (at z = 0 and at z = −1):
 
1 (z−1)2 (z−1)2
• Consider the Laurent series centered at z = 0. f (z) = z (z+1)3 . The remaining factor (z+1)3 is
2
analytic at z = 0, and can be expanded as 1 − 5z + 13z − .... The coefficient c−1 = 1 is both the
strength of the pole z = 0 of order 1, and the residue.
 
(z−1)2 2
1
• Consider next the Laurent series centered at z = −1. f (z) = (z+1)3 z . The function (z−1)
z is
2
analytic around z = −1 and thus has a Taylor series centered at −1, (z−1)
z = −4−(z+1)2 −(z+1)3 −....
Thus the function exhibits a 3rd order pole with a strength of −4. 
z 1/2 −1
Example 4.3.5 Describe the singularities of f (z) = z−1 .

Due to the z 1/2 -term in the numerator, there will be a branch point at z = 0, giving rise to two solution
sheets. At z = 1, there occurs a divide by zero. On one solution sheet, 11/2 = +1 and the singularity
becomes removable, whereas on the other sheet 11/2 = −1 and we get a first order pole with residue -2.
Figure 4.8 displays this function. In part (a), we see the real parts of the two solutions coinciding for x < 0
and the imaginary parts coinciding for x > 0. For the real part, we see the pole at z = 1 affecting only
one of the two solution branches. Part (b) illustrates Im f (z) in the complex plane, showing again how one
solution sheet is unaffected by the pole that is present on the other one. 
For the examples we have shown previously of multi-valued functions, there were great similarities between
the solutions on different sheets. This latest example illustrates that this by no means needs to be the case.
The following is an interesting result for the character of an analytic function near an essential singularity.
72 CHAPTER 4. COMPLEX INTEGRATION

Re f(x)
3

1 3

0
2
-1

-2
1
-3
-3 -2 -1 0 1 2 3
x 0

Im f(x)
1
-1

0.5
-2

-3
4
-0.5
2
3
0 2
1
-1 -2 0
-1
-3 -2 -1 0 1 2 3 -4 -2
y -3
x x

(a) Re f (x) and Im f (x) along x-axis. (b) Im f (z) in the complex plane near the origin
1/2
Figure 4.8: The function f (z) = z z−1−1 , displayed along the real axis, and its imaginary part in the complex
plane (with a part of the pole cut out to better display the pole-free solution sheet).

Theorem 4.3.2 (Picard) Take any arbitrarily small neighborhood of an essential singularity. The func-
tion f (z) then takes every complex value (with at most one exception) infinitely many times within this
neighborhood.

4.4 Residue calculus

Let z1 , z2 , z3 , ..., zn be the singularities of function f (z), inside contour Γ. Assume that these singularities
are either poles or essential singularities.

˛ ˆ ˆ ˆ ˆ
+ + + +... + =0 (4.7)
Γ z1 z2 z3 zn

Changing direction thanks to a change of sign, we obtain

˛ ˆ ˆ ˆ ˆ
= + + +... + (4.8)
Γ z1 z2 z3 zn

We can thus change the contour to just go around each singular point. Recall that

˛

 0 if z = z0 is outside Γ
1 dz
= 0 if z = z0 is inside Γ and n 6= −1
2πi Γ (z − z0 )n
1 if z = z0 is inside Γ and n = −1


for n integer. Thus for each singularity,

∞
X
f (z) = cn (z − zj )n
n=−∞
4.4. RESIDUE CALCULUS 73

z2
z3
z1

z4
z5

Γ
Figure 4.9: Change of contour in residue calculus.

Only the term associated with n = −1, the residue (c−1 ) will contribute to the integral. So
˛ N
X 
f (z)dz = 2πi c−1 

Γ zj
j=1
| {z }
Sum of the residues
 
XN 
= 2πi Res(f, zj )
 
j=1

This result is known as the Cauchy Residue Theorem.

4.4.1 Examples
Example 4.4.1 Evaluate
˛
I= ze1/z dz .
|z|=1

The only singularity inside the unit circle is the essential singularity at z = 0. We thus expand the integrand
at that point to find the function’s residue there:

1 1
ze1/z = z(1 ++ + ...)
z 2!z 2
1
= z+1+ + ...
2!z

Thus, Res(ze1/z , z = 0) = 1/2. Therefore,

˛
1
I= ze1/z dz = 2πi · = πi .
|z|=1 2

Figure 4.10 shows that the integrand is of extreme complexity around the singularity point. It is remarkable
how that did not at all complicate the calculation of the integral going around it - all that mattered was the
value of the residue. 

Shortcuts to calculate the residue We wish to compute the residue Res(f (z), z0 ) without having to
work out the Laurent expansion. In case the singularity is a pole of order m, the following gives three
options::
74 CHAPTER 4. COMPLEX INTEGRATION

2
0
0.2
-2
0.1

0.2 0
0.1
-0.1
0
-0.1 -0.2
-0.2 x
y -0.3
-0.3

(a) Re z e1/z (b) |z e1/z |, with phase angle coloring

Figure 4.10: Illustrations of the integrand in Example 4.4.1, f (z) = z e1/z , near the origin.

N (z)
1. In case of m = 1, write f (z) = D(z) , where N (z0 ) 6= 0 and D(z) has a simple root at z0 . Then,
N (z0 )
Res(f (z), z0 ) = Note that what one decides to include in N (z) and in D(z) is quite arbitrary;
D 0 (z0 ) .
making D(z) as simple as possible will make it easier to take the derivative.
N (z0 )+N 0 (z0 )(z−z0 )+... N (z0 ) 1
Justification: Taylor expanding around z0 gives f (z) = 0+D 0 (z0 )(z−z0 )+... = D 0 (z) · (z−z0 ) + O(1).
φ(z)
2. If m is arbitrary and f (z) can conveniently be written in the form f (z) = (z−z0 )m , then Res(f (z), z0 ) =
(m−1)
φ (z0 )
(m−1)! .

Justification: Expanding φ(z) in a Taylor series centered at z0 gives φ(z) = φ(z0 ) + (z − z0 )φ0 (z0 ) +
(z−z0 )2 00
2! φ (z0 ) + . . . and therefore,

φ(z) φ(z0 ) φ0 (z0 ) φ00 (z0 ) φ(m−1) (z0 )

= + + + ... + + ...
(z − z0 )m (z − z0 )m (z − z0 )m−1 2!(z − z0 )m−2 (m − 1)!(z − z0 )
with the last displayed term above giving the result.
3. If m is arbitrary and the rewrite described above is not convenient, we can instead use Res(f (z), z0 ) =
1 dm−1 m
(m−1)! limz→z0 dz m−1 ((z − z0 ) f (z)), where the limit usually is best handled by l’Hospital’s rule.

Justification: Around the pole

c−m c−m+1 c−1
f (z) = + + ... + + ...
(z − z0 )m (z − z0 )m−1 (z − z0 )1
and therefore

(z − z0 )m f (z) = c−m + c−m+1 (z − z0 )1 + . . . + c−1 (z − z0 )m−1 + . . .

and
dm−1
((z − z0 )m f (z)) = (m − 1)! c−1 + m! c0 (z − z0 ) + . . .
dz m−1
and the result follows.
z
e
Example 4.4.2 Compute Res( z cos z , z = 0).

z z e0
e e
We can let N (z) = cos z and D(z) = z. Then Res( z cos z , z = 0) =
cos 0
1 = 1. Alternatively, we can choose
z 0
z e e
N (z) = e and D(z) = z cos z, which gives Res( z cos z , z = 0) = cos 0−0 = 1. 
4.4. RESIDUE CALCULUS 75

Im z

R
z2 z1

Re z
z4 z3

Figure 4.11: Contour in example 4.4.3.

Example 4.4.3 Evaluate

ˆ ∞
x2
I= dx .
−∞ 1 + x4
This integral must be real valued. Let’s try computing this integral with residue calculus. Consider the
z2 √ , z2 = −1+i √ , z4 = −1−i
function f (z) = 1+z 4. This function has 4 simple poles: z1 = 1+i 2
√ , z3 = 1−i
2 2
√ .
2
Consider the open ´ contour  Γ, a semi-circle of radius R, presented in Figure 4.11. Recall that the magnitude
z2
of the integral  Γ 1+z 4 dz  ≤ M · L, where L is the length of the contour (πR) and where M is the upper
 
bound of the function on the contour. In order to find this bound, recall the following triangle  inequalities:
 z2  R2
|z1 + z2 | ≤ |z1 | + |z2 |, |z1 + z2 | ≥ |z1 | − |z2 | but also |z1 + z2 | ≥ |z2 | − |z1 |. Therefore  1+z 4  ≤ R4 −1 . Thus,
´  ´ z2
z2 πR3
4 dz  ≤ R4 −1 → 0 as R → ∞. Therefore, dz = 0 when R is sufficiently large. Notice that
 
 Γ 1+z Γ 1+z 4
the contour Γ is an open contour. Now consider the closed contour, say Λ, which is made of the contour Γ
followed by the section of the real axis from −R to R. Since 2 poles are located inside of this new contour,
˛ ˆ ˆ R
z2 z2 x2
4
dz = 4
dz + 4
dx
Λ 1+z Γ 1+z −R 1 + x
ˆ ∞
x2
= 0+ 4
dx as R → ∞
−∞ 1 + x
= 2πi{Res(f, z1 ) + Res(f, z2 )}
N (z1 ) 2
Using the formulas above, we find Res(f, z1 ) = D 0 (z ) , where N (z) = z and D(z) = z 4 + 1. Thus,
√ 1
√
Res(f, z1 ) = 4z11 = 2(1 + i)/8. Similarly, Res(f, z2 ) = 4z12 = 2(−1 − i)/8. Thus, 2πi{Res(f, z1 ) +
Res(f, z2 )} = √π2 . Therefore,
ˆ ∞
x2 π
4
dx = √ . 
−∞ 1 + x 2

´∞
Note When the integral to compute is of the type −∞ f (x)dx, the open half-circle contour Γ shown
above
´ will often be used, with R → ∞. We showed that, for our particular integrand in Example 4.4.3,
 f (z)dz  → 0. Note that this result will always be true in the case where both numerator and denominator
Γ
are polynomials, if the degree of the denominator is at least 2 more than the degree of the numerator.
Example 4.4.4 Evaluate
ˆ ∞
cos kx
I= dx
−∞ (x + b)2 + a2
76 CHAPTER 4. COMPLEX INTEGRATION

Im z

ia − b Γ

Re z

−ia − b

Figure 4.12: Contour in example 4.4.4

with k, a > 0 and b ∈ R.

cos kz
Consider the integrand, the function f (z) = (z+b) 2 +a2 . We find the two simple poles to be z1 = ia − b

and z2 = −ia − b. We would love to bound the magnitude of the integral as we did in the previous example.
However, cos kz oscillates with amplitudes that grow exponentially fast once we consider its values far away
ikz −ikz
from the real axis. Indeed, recall that cos kz = e +e 2 . In order to bypass this issue, let’s instead consider
eikz
the function f (z) = (z+b)2 +a2 . In this case,
ˆ ∞ ˆ ∞ ˆ ∞
eikx cos kx sin kx
dx = dx +i dx
−∞ (x + b)2 + a2 −∞
2
(x + b) + a2 2
−∞ (x + b) + a
2
| {z } | {z }
Real part Imaginary part

´∞ eikx
Thus, we can evaluate −∞ (x+b) 2 +a2 dx and retain the real part only as the answer. As for the boundedness
−ky ikx
of the integrand, e ikz
= e e , which implies that |eikz | = e−ky → 0 as y → ∞(had k been negative,
we could similarly have used a contour in the lower half plane). Let x = R cos θ and y = R sin θ be the
coordinates of z = x + iy, a point on the open semi-circle contour Γ of radius R as illustrated in Figure 4.12.
Let’s find an upper bound for the magnitude of the integral. We have
ˆ
eikz eikz
  
   
 (z + b)2 + a2 dz  ≤ 
    (πR)
2 2
(z + b) + a 
Γ
e−Rk sin θ
= (πR)
|(z + b)2 + a2 |
πe−Rk sin θ
≤
R |(cos θ + i sin θ + Rb )2 | − | R
a 2


|
→0 as R → ∞

Let’s now consider the closed contour Λ, which is made of the half circle Γ and which closes by covering the
whole real axis from −R to R, as R → ∞.
˛ ˆ ˆ R
eikz eikz eikx
dz = dz + dx
Λ (z + b)2 + a2 2
Γ (z + b) + a
2 2
−R (x + b) + a
2
ˆ ∞
eikx
= 0+ 2 2
dx as R → ∞
−∞ (x + b) + a
= 2πiRes(f, ia − b)
4.4. RESIDUE CALCULUS 77

Calculating the residue at z = ia − b, we obtain

e−ka (−i cos kb − sin kb)
Res(f, ia − b) =
2a
(here using the assumption that a > 0; for a < 0 the relevant pole would have been the one at z = −ia − b).
Thus, ˆ ∞
eikx πe−ka (cos kb − i sin kb)
2 2
dx =
−∞ (x + b) + a a
Finally, by only retaining the real part, we obtain the result
ˆ ∞
cos kx πe−ka cos kb
I= 2 2
dx = .
−∞ (x + b) + a a
As a concluding observation, we can note from the form of the integral that the answer must be an even
function in both a and k. So we can drop that the requirements that these parameters are positive, giving
the final answer as ˆ ∞
cos kx πe−|ka| cos kb
I= 2 2
dx =
−∞ (x + b) + a |a|
for a, b, k all real (with divergence if a = 0). 
´∞
Integrals of the type −∞ f (x)eikx dx are very common. The semi-circle contour Γ will be used in such
´
cases. The following result gives us an easy way to ensure that Γ f (z)eikz dz → 0 as R → ∞.
Theorem 4.4.1 Jordan’s Lemma. ´ If f (z) → 0 uniformly as R → ∞, Γ is a half-circle as shown in Figures
4.11 and 4.12, and k > 0, then Γ f (z)eikz dz → 0.

Example 4.4.5 Evaluate

ˆ 2π
dθ
I=
0 A + B sin θ
where A > 0 and where |B| < A (to avoid divides by zero on the integration path).
´ 2π
We often encounter the more general form 0 g(sin θ, cos θ)dθ. The following illustrates a couple of general
approaches.

Method 1: Let
1
f (z) =
A + B sin z
1
Considering the contour sketched in Figure 4.13 (a), we find that for the integrand f (z) = A+B sin z
ˆ
f (z)dz → 0
Γ3

since sin z grows´ exponentially

´ fast the further we go up the complex plane. We also notice that f (z) is
2π periodic, so Γ2 f (z)dz + Γ4 f (z)dz = 0. The three extra paths Γ1 , Γ2 , Γ3 will therefore together not
contribute anything to the integral. Next step is to find the singularities inside the contour, illustrated again
together with f (z) in Figure 4.13 (b). The singe pole inside the contour occurs where A + B sin z0 = 0, i.e.
A
at z0 = − arcsin B . The residue becomes
 
N (z0 ) 1  1  1
Res(f, z0 ) = 0 = = = √ .

p
D (z0 ) B cos z z=− arcsin A


B
2
B 1 − (sin z) z=− arcsin A
 i A − B2
2
B

Therefore,
2π
I=√
A2 − B 2
78 CHAPTER 4. COMPLEX INTEGRATION

Im z

Γ3

Γ4 Γ2

0 Γ1 2π Re z

(a) Schematic integration path (b) Magnitude/phase plot with path in case A = 3, B = 1.

Figure 4.13: Contour and function in example 4.4.5.

Method 2: The second ‘routine’ approach for trigonometric integrals over a full period is to let z = eiθ .
Then
 
1 iθ
1 1
cos θ = e + e−iθ = z+
2 2 z
 
1 iθ 1 1
e − e−iθ =


sin θ = z−
2i 2i z
1
and dθ = iz dz. For 0 ≤ θ ≤ 2π (or equivalently −π ≤ θ ≤ π), the path for z becomes once around the unit
circle, in the positive direction. With this variables change,
˛
1 dz
I = 1 1



|z|=1 A + B 2i z − z iz
˛
2 dz
= A
B |z|=1 z 2 + 2i B z−1
A
The denominator factorizes as z 2 + 2i B z − 1 = (z − z1 )(z − z2 ) with
√
−A ± A2 − B 2
z1,2 = i .
B
√
2 2
Since A > |B|, the second root (pole) z2 = i −A− BA −B will fall outside the unit circle, and from z1 ·z2 = −1
follows then that z1 falls inside it. From
 
1 1
Res , z = z1 =
(z − z1 )(z − z2 ) z1 − z2
follows then that the relevant residue evaluates to √B and, with that,
2i A2 −B 2

2 2π
I = B (2πi) 2i √AB2 −B 2 =√ .
A2 − B2
Example 4.4.6 Evaluate
˛
dz
I=
Γ z 2 + a2
around the contour shown in Figure 4.14.
4.4. RESIDUE CALCULUS 79

Im z

ia

Γ Re z
−ia

Figure 4.14: Contour in example 4.4.6.

We find that  
1 −1
Res , −ai =
z + a2
2 2ai
and similarly,
 
1 1
Res 2 2
, ia =
z +a 2ai
Since the contour goes around the pole z = ia twice, we shall count the function’s residue at that point
twice, while the residue at z = −ia only once. We thus find
 
1 1 π
I = 2πi 2 − =
2ia 2ia a

The winding number w(zj ) of the curve Γ around a point is the number of times that Γ winds around zj
(positive means counterclockwise). We have for example w(ia) = 2 and w(−ia) = 1. 

Example 4.4.7
ˆ ∞
dx
I= , a>0
0 x3 + a3

One can again consider using as´ contour a circle of radius R centered at the origin, since far out in any
direction in the complex plane, f (z)dz → 0 as R → ∞. The problem is that the limits of integration go
x
from 0 to ∞ and not from −∞ to ∞. Before describing a more general approach in the next Example 4.4.8,
here is a more specialized idea.
Since z only appears as z 3 , then we that both z = r and z = re2πi/3 give the same result, z 3 = r3 . We
thus choose the contour as in Figure 4.15 and perform both substitutions z = r and z = re2πi/3 . We obtain
ˆ ∞ ˆ ˆ 0
dr dz ei2π/3 dr 1
+ + = 2πi Res( 3 , z1 = aeπi/3 )
0 r + a3
3 z + a3
3
∞
3
r +a 3 z + a3
x
80 CHAPTER 4. COMPLEX INTEGRATION

Im z

z1 = ae i π/3

2π/3
z2 = −a

Γ Re z

z3 = ae5i π/3

Figure 4.15: Contour in example 4.4.7

´ dz 1 πi/3 1
Since z 3 +a3 → 0 as R → ∞, and since Res( z3 +a3 , z1 = ae )= 3a2 e2πi/3
we obtain that
x
ˆ ∞ 
dr 2πi
(1 − e2πi/3 ) = ,
0 r + a3
3 3a2 e2πi/3
which simplifies to ˆ ∞
dr 2π
= √ .
0 +a 3
3 3a2 r3
It was in this case quite lucky that the integrand itself provided a return path from infinity back to the
origin, along which it took the same values as along the real axis. As we will see next, this type of ‘luck’
is quite unnecessary. The particular behavior of Im log z (recall Figure 2.7 (b)) offers a much more general
approach for creating a suitable return path.

4.4.2 Special contours

Example 4.4.8 Evaluate
ˆ ∞
dx
I= , a, b > 0
0 (x + a)(x + b)
We can solve this exercise using partial fractions, but
´ ∞we will ignore that fact for now. We will use an idea
that can be used for much more general integral 0 f (x)dx. We only need the integrand to decay fast
enough for the contribution of its value on a path distant from the origin to be insignificant.
Instead of tackling the problem itself, consider
˛
(log z)dz
J= , a, b > 0
Γ (z + a)(z + b)
along the keyhole contour ´shown in Figure 4.16. ¸
∞
In general, to compute 0 f (x)dx, we will consider Γ (log z)f (z)dz Why is this a good idea? If we follow
the contour, we often the following contributions to the integral
´ ´∞
1. → (log z)f (z)dz = 0 log(z)f (z)dz
4.4. RESIDUE CALCULUS 81

Figure 4.16: Contour in example 4.4.8

´
2. (log z)f (z)dz → 0 as R → ∞
R
´ ´∞ ´∞
3. ← ((log z) + 2πi)f (z)dx = − 0 (log z)f (z)dz − 2πi 0 f (z)dz
´
4.  (log z)f (z)dz → 0 as  → 0

Thus
ˆ ˆ ˆ ˆ  ˆ ∞
+ + + (log z)f (z)dz = −2πi f (x)dx
→ R ←  0
X
= 2πi Res((log z)f (z), zj )
j

Thus,
ˆ ∞ X
f (x)dx = − Res((log z)f (z), zj )
0 j

In our example, we have

 
log(−a) log(−b)
I = − (Res((log z)f (z), −a) + Res((log z)f (z), −b)) = − +
−a + b −b + a
log(b/a)
= .
b−a
In this example, introducing an extra log z allowed the evaluation over [0, +∞] in place of over [−∞, +∞].
If what we wanted to integrate already contained a factor of log x in the numerator, we can still multiply
with an extra log z, see Exercises 4.9.7 and 4.9.8.

Example 4.4.9 Evaluate

ˆ 1
√
1 − x2
I= dx
−1 1 + x2

√ the integrand over the integration interval x ∈ [−1, 1] (assuming the conventional sign
Figure 4.17 shows
choice such that 1 = +1). As usual, displaying the function over the complex plane reveals a much richer
picture. The function has two branch points, at z = ±1 (but none at z = ∞), making it natural to introduce
a branch cut between -1 and +1. The real and imaginary parts of the integrand f (z) then become as shown
82 CHAPTER 4. COMPLEX INTEGRATION

0.8

0.6

f(x)
0.4

0.2

-0.2
-1 -0.5 0 0.5 1
x

√
1−x2
Figure 4.17: The integrand f (x) = 1+x2 displayed over −1 ≤ x ≤ +1.

√
2
in Figure 4.18. We note in particular the (first order) poles at z = ±i, both with residue 2i , as follows from
N (z0 )
the Res(f (z), z0 ) = D 0 (z ) -rule:
0

√  √  √
1 − z 2  1 − z 2  2
+ =− = .
2z  2z  2i
 
z=+i z=−i

We give next three different methods for proceeding.

Method 1: A simple fact that could quite easily escape attention if it wasn’t for Figure 4.18 is that
Re f (z) is identically zero over [−∞, −1] and [+1, +∞] and Im f (z) is anti-symmetric along the real axis,
meaning that there will be no difference in value if we change the integration interval from [−1, +1] to
[−R, +R] for any R > 1. We then close the loop via a large (radius R) half-circle Γ in the upper half-plane.
It will then hold
ˆ √
I + f (z)dz = 2πi · {residue of pole at z = +i} = π 2 . (4.9)
Γ
´
It remains to evaluate Γ
f (z)dz. Since f (z) is singularity-free for |z| > 1, it will there possess a Laurent
expansion:
√
q
1
1− z2 ±iz 1− z2 i

3

f (z) = = 2 1

=± 1− + ... .
1 + z2 z 1+ z2
z 2z 2

Taking the integral over the semi-circle, only the leading term ± zi will contribute in the limit of R → ∞.
We need however to choose the sign so it corresponds to the function branch we are interested in - with
f (0) = +1, as shown in Figure 4.18 (a). As z moves up the imaginary axis, the value should
´ (beyond´ the pole)
be negative and increase towards zero. That requires the negative sign. We then find Γ f (z)dz = Γ (− zi )dz,
´
which, after changing variables {z = Reiθ , dz = Ri eiθ dθ}, evaluates to Γ (− zi )dz = π. From (4.9) now
√
follows I = π( 2 − 1).

Method 2: When one wants to integrate along one side of a branch cut, it is often advantageous to extend
the contour to also run along the other side of the cut, and in the opposite direction. Hence, we consider
´ illustrated´ in Figure 4.19. On the return path, below the cut, both f (z) and´ dz have changed
the contour
sign, so ← f (z)dz = → f (z)dz = I. Since, the integrals around the two circular paths  f (z)dz → 0 as
their radii  → 0, the integral around the closed loop will be 2 I. Had there been a residue associated with
the branch cut, this would have provided a value for I. The alternate strategy in cases such as this is to just
expand the contour into a very large circle, radius R, surrounding the origin. Expanding the contour in this
way will (i) cause the integral to change according to the two poles that now have been included (cf. Figure
4.20), and (ii) allow an exact evaluation for R → ∞.
√
2
From the discussion above, each of the poles have residue 2i , and the integral around the full circle will
4.4. RESIDUE CALCULUS 83

3 3

2 2

1 1

0 0

-1 -1

-2 3 -2 3
2 2
-3 1 -3 1
-3 0 -3 0
-2 -2
-1 -1 -1 -1
0 0
1 -2 1 -2
2 y 2 y
x 3 -3 x 3 -3
√ √
1−z 2 1−z 2
(a) Real part of f (z) = 1+z 2 . (b) Imaginary part of f (z) = 1+z 2 .
√
1−z 2
Figure 4.18: Real and imaginary parts of f (z) = 1+z 2 , using a branch cut along [−1, +1]. The sign is

chosen so the function is positive along the cut and smooth above it.

Im z

Γ
−1 1
Γ Re z

−i

Figure 4.19: First contour in example 4.4.9; the second contour will be to expand it to a large circle radius
R surrounding the origin.

be twice of what we evaluated for the half-circle under ‘Method 1’. Putting this together:
ˆ √
2I + f (z)dz = 2πi {sum of residues} , i.e. 2I + 2π = 2 2π ,
"
and we therefore again obtain
√ 
I=π 2−1 .

√
Method 3: Let’s perform the change of variable x = sin θ. Then, 1 − x2 = cos θ and dx = cos θdθ. Also,
1 + x2 = 1 + sin2 θ. Therefore,
ˆ π/2
cos2 θ
I = 2 dθ
−π/2 1 + sin θ
ˆ
1 π cos2 θ
= dθ
2 −π 1 + sin2 θ
84 CHAPTER 4. COMPLEX INTEGRATION

Im z

Γ
−1 1
Re z

−i

Figure 4.20: Illustration of the ‘first’ and ’second’ contours (as shown in Figure 4.19 and the large circle,
respectively), and why the integrals around these two contours add up to the contributions from the two
poles.

because the integrand is π-periodic.

The standard further change of variables when dealing

with trigonometric functions over a full period is
to set z = eiθ . Then cos θ = 12 z + z1 , sin θ = 2i
1
z − z1 and ieiθ dθ = dz implying dθ = −i


z dz. Therefore,

˛ 1 1 2


z+
 
1 4 z −i
I =
2 dz
2 |z|=1 1+ −1
z − z1 z
4
˛ 2 2
−i (z + 1)
= dz
2 |z|=1 z(4z − (z 2 − 1)2 )
2
˛
i (z 2 + 1)2
= dz (4.10)
2 |z|=1 z(z − z1 )(z − z2 )(z − z3 )(z − z4 )
√
where zj = ±1
√ ± 2, j = 1, 2, 3, 4. √The only roots that are located inside the unit circle are at z = 0,
z = z1 = 1 − 2, and z = z2 = −1 + 2; see Figure 4.21. We find that

Res(f, 0) = 1
−1
Res(f, z1 ) = √
2
−1
Res(f, z2 ) = √
2
and therefore
i √
I = 2πi (− 2 + 1)
2√
= π( 2 − 1) 

Example 4.4.10 Evaluate

ˆ
+∞
eαx
I= dx.
1 + ex
−∞
4.4. RESIDUE CALCULUS 85

Figure 4.21: Magnitude / phase portrait of the integrand in (4.10), with the integration path marked by a
dashed black circle.

In order for the integral to converge at x = −∞ and at x = +∞, we need Re α > 0 and Re α < 1,
eαz
respectively. In the complex z-plane, f (z) = 1+e z has poles at z = ±iπ, ±3iπ, ±5iπ, etc. - see Figures 4.22

and 4.23. This pole pattern, and even more how the phase angle for f (z) seems to be constant along lines
parallel to the x-axis suggests the integration path that is shown, following the real axis from −R to +R,
and returning from R + 2πi to −R + 2πi , with also two vertical sides along which the contributions go to
´ eα(z+2πi) 2πiα
´ eαz
zero when R → ∞. Along the upper return path, ← 1+e (z+2πi) dz = e ← 1+ez
dz = −e2πiα I, implying
that for the full loop, we get
ˆ ˆ  αz 
e
+ = (1 − e2πiα ) I = 2πi Res (f, z = πi) = 2πi = 2πi e(α−1)πi ,
ez z=πi
→ ←

from which follows

π
I= .
sin πα

Figure 4.22: Contour for Example 4.4.10 .

Example 4.4.11 Evaluate

ˆ ∞
xa−1
I= dx, a ∈ R
0 1+x
86 CHAPTER 4. COMPLEX INTEGRATION

Figure 4.23: Magnitude / phase portrait of the integrand in Example 4.4.10 (shown in the case of α = 1/2),
with the same integration path as shown schematically in Figure 4.22 here marked by a dashed rectangle.

In order to have convergence, we must have a > 0 (x = 0) and a < 1 (x = ∞). A notable feature of
α−1
the integrand, which we write as f (z) = z1+z , is its branch point at z = 0. That suggests a couple of
opportunities. First, we have seen before that, when integrating from a branch point, it can be convenient
to locate the branch cut along it, and choose a closed loop that returns on the opposite side of the same
cut (a ‘keyhole’ contour). This will become our Method 1. Another possibility is to start by rewriting
xα−1 = e(α−1) log x and change variable t = log x, after which the integrand becomes free of branch points.
Just possibly, the new integral provide different solution opportunities.
Method 1: Placing the branch cut immediately below the outgoing integration path and creating a closed
loop that returns just below it suggests the contour shown in Figure 4.24.

Im z

x = −1 Re z

Figure 4.24: Better contour in example 4.4.10

Let’s consider all the contributions to the closed loop integral:

a−1
1. Res(f, −1) = (−1)1 = eiπ(α−1)
´ ´ ∞ a−1
2. → f (z)dz = 0 x1+x dx = I
4.4. RESIDUE CALCULUS 87

´ ´∞ ea−1 elog x+i2π

3. ←
f (z)dz = − 0 1+x dx = e2πi (a−1) I
´
4. 
f (z)dz → 0 as  → 0
´
5. R
f (z)dz → 0 as R → ∞
Putting everything together, we obtain
2πi eiπ(α−1)
I =
1 − e2πi (a−1)
π
=
sin πa

Method 2: Alternatively, one can consider some analytic modifications (partial integration, change of
variables, etc...) before applying contour integration. Given that xα−1 = e(α−1) log x , the variable change
t = log x is ‘natural’. Then x = et , dx = et dt, and the integral becomes
ˆ ∞ a−1 ˆ ∞ t(a−1) t
x e e
= t
dt
0 1+x −∞ 1 + e
ˆ ∞
eta
= t
dt
−∞ 1 + e
π
=
sin πa
where the last equality follows from the preceding Example 4.4.10. 

An interesting follow-up result to this example is one of the many ways to prove one of the functional
identities for the Γ(z)-function:
Theorem 4.4.2 The Γ(z)-function satisfies the functional equation
π
Γ(z)Γ(1 − z) = . (4.11)
sin(πz)
Proof
ˆ ∞  ˆ ∞ 
z−1 −t z −s
Γ(z)Γ(1 − z) = t e dt s e ds
0 0
| {z }
Change of variable: t→su
ˆ ∞  ˆ ∞ 
z−1 −su z −s
= (su) e sdu s e ds
0 0

Combining into a double integral, let s outside be the same as s inside so s−z cancels sz . Thus
ˆ ∞ˆ ∞
Γ(z)Γ(1 − z) = sz−1 uz−1 e−su ss−z e−s duds
0 0

Now split up again

ˆ ∞ ˆ ∞  ˆ ∞ z−1
u
uz−1 e−su e−s ds du = du
0 0 0 1 +u
| {z }
1
1+u

π
=
sin πz
Recalling the idea of analytic continuation,
π
f (z) = Γ(z)Γ(1 − z) − ≡ 0 when 0 < Re z < 1,
sin πz
implies that this will hold for all z. 
88 CHAPTER 4. COMPLEX INTEGRATION

´ ´ ´
(a) Γ
f (z)dz = 0 (b) Γ
f (z)dz = 2πiRes(f, z0 ) (c) Γ
f (z)dz = πiRes(f, z0 )

Figure 4.25: The principal value integral can be thought of as the average of integrals with a pole in and
out the contour.

Γ Γ Γ

−ε ε
ε ε
−ε ´−ε ´
(a) Principal value integral contour (b) Γ f (z)dz = 0 (c) Γ
f (z)dz = 2πiRes(f, z0 )

Figure 4.26: Contours considered for deriving the key principal value integral result.

4.4.3 Principal value integrals

A natural follow-up to discussing the effect on an integral when a singularity is located inside vs. outside a
contour is what happens if it is located exactly on a contour (cf. Figure 4.25). For example, does a concept
´1
such as −1 dx x make sense? Will both left and right halves diverge to ±∞ ? Figure 4.26 (a) illustrates the
´ − dx ´1
case of −1 x +  dx x , with ε small, and the contour closed in the upper half-plane. Parts (b) and (c) give
the integral values for two choices of connecting z = −ε to z = +ε. Going around this small inner circle
one full revolution in the positive direction contributes 2πi Res(f, 0) and going around it half a revolution
similarly contributes ±1πi Res(f, 0), with the ± sign according to direction. Starting from either the case in
part (b) or part (c) of the figure, we see that the integral value in case (a) (as ε → 0) should be 1πi Res(f, 0).
We have thus arrived at:
´
Key Result As is illustrated in Figure 4.25 (c) with a singularity on the boundary, Γ f (z)dz = πiRes(f, z0 )
(the average of the integrals in the situations for which the singularity is inside and outside of the contour
Γ). The singularity must be on a curve segment that is locally smooth, and cannot be on a corner.
Example 4.4.12
ˆ ∞ ˆ ∞
sin x 1 sin x
I= dx = dx
0 x 2 −∞ x
´
1 ∞ iz
Following the idea from Example 4.4.4, we compute 2 −∞ ez dz and extract its imaginary part. Closing
the contour by a half-circle in the upper half-plane (cf. Figure 4.27) does not, by Jordan’s lemma, change
the value (as R → ∞). Since the pole at z = 0 is located right on the integration path, we will use half
izz ´ ∞ iz
the residue Res( e z , z = 0) = 1. Thus, −∞ ez dz = 2πi 21 = πi. Separating real and imaginary parts give
´ ∞ cos z ´ ∞ sin z ´∞
−∞ z
dz = 0 and −∞ z dz = π, i.e. I = 0 sinx x dx = 21 π. 
4.4. RESIDUE CALCULUS 89

Im z

Re z

Figure 4.27: Contour in example 4.4.12

4.4.4 Rouche’s theorem

We start by recalling Theorem 4.2.3 (Cauchy’s argument principle):

Lemma 4.4.3 Consider function f (z) and the contour Γ. If the function f (z) has N zeros and P poles
1
¸ f 0 (z)
inside Γ, and there are no other singularity than these poles inside of Γ, then 2πi f (z) dz = N − P .

We proved this earlier. Here is a slightly different argument that also shows it. Let’s modify the contour
Γ to go around each zero and each pole separately in small circles. Say z0 is a zero or order m. Then,
f (z) = (z − z0 )m g(z), where g(z) is non-zero and analytic inside the circle. Then

f 0 (z) m(z − z0 )m−1 g(z) + (z − z0 )m g 0 (z)

=
f (z) (z − z0 )m g(z)
m g 0 (z)
= +
z−z g(z)
| {z 0} | {z }
contributes m2iπ contributes nothing

We can use the same reasoning for poles. The pole of order n contributes −2πin to the integral. 

Figure 4.28: Contour in Lemma 4.4.3, where we identify the poles as × and the zeros as ·
90 CHAPTER 4. COMPLEX INTEGRATION

From this result, we can introduce the concept of winding number

˛ 0
1 f (z)
N −P = dz
2πi f (z)
1
= [log f (z)]end
start
2πi
1
= [ log |f (z)| +i arg(f (z)) ]end
start
2πi | {z } | {z }
Same at start and end changes by 2π each time f (z) goes around the origin

= {winding number of f (z) with respect to Γ}

The color plots we have frequently used to illustrate phase angles of functions also show the concept
of winding number ‘in action’. As an example, Figure 2.2 (d) illustrates the phase angles for the function
f (z) = z + z1 , which has one first order pole (at z = 0) and two simple zeros (at z = ±i). If we follow any
curve that circles close to one of the zeros, the color changes along it correspond to once around the color
wheel, while going around the pole near to it goes through the colors in reversed order. For a completely
arbitrary closed path in the z-plane, we can follow the net circulation it has produced with regard to the
color wheel. For example, going around the origin once in the positive direction far out from it, the net
effect in this case matches once around the color wheel in the the positive directions, telling that N − P for
that path equals to one. The number of times we have gone around the color wheel when following a path
equals to the winding number for the path.
Example 4.4.13 Say f (z) has no poles inside Γ (P =0). If in this case, the winding number is two (cf.
Figure 4.29), then N = 2, i.e. f (z) has two zeros inside Γ.

Γ Im w

w = f (z)
st ar t/end

Im z Re w

Re z

Figure 4.29: Contour in Lemma 4.4.3, where we identify the poles as × and the zeros as ·

A consequence of this observation is the following theorem:

Theorem 4.4.4 Rouché’s theorem. Let f (z) and g(z) be analytic inside Γ (closed). If |f (z)| > |g(z)| on Γ
then f (z) and f (z) + g(z) have the same number of zeros inside Γ.
Proof Let z go around Γ. If |f (z)| > |g(z)| on Γ, then |(f (z) + g(z)) − f (z)| < |f (z)|. Thus f (z) + g(z)
is closer to f (z) than the origin. Think of walking a dog many times around a tree. We want to make the
leash shorter than the distance to the tree. Thus, f (z) + g(z) must be dragged around the origin just the
same number of times. Therefore it must have the same number of zeros in side Γ. 

As a consequence of Rouche’s theorem, we obtain another proof of the Fundamental Theorem of Algebra
Theorem 4.4.5 A polynomial of degree n has exactly n roots.
Proof Consider
Pn (z) z n + (an−1 z n−1 + an−2 z n−2 + ... + a1 z + a0 )
= |{z}
| {z }
f (z) g(z)
4.5. INFINITE SUMS 91

Im w

Re w

Figure 4.30: Illustration of w−plane in Rouché’s Theorem 4.4.4. We see f (z) as a solid line and g(z) as a
dashed line.

Let z go around a very big circle around the origin. f (z) has clearly n roots, and |f (z)| > |g(z)| around the
| {z }
Rn
circle. |g(z)| < n · max|aj |Rn−1 . Then just make R > n · max|aj |. Then, |f (z)| = Rn > n · max|aj |Rn−1 >
|g(z)|. Therefore, Pn (z) has also n roots. 

Example 4.4.14 How many zeros does p(z) = z 4 + 6z + 3 have in the complex plane? Also determine how
many of these satisfy |z| < 2.

The degree of the polynomial is 4. Therefore, there are 4 zeros. Now, let f (z) = z 4 and let g(z) = 6z + 3.
When |z| = 2, |f (z)| = 24 and |g(z)| = |6z + 3| < 6|z| + 3 = 15. Then, since |f (z)| > |g(z)| around the
contour, f (z) and p(z) = f (z) + g(z) have the same number of zeros inside Γ (|z| < 2). The function f (z)
has 4 zeros and thus p(z) has also 4 zeros which satisfy |z| < 2. 

4.5 Infinite sums

One of the applications of residue calculus is to infinite sums.

Example 4.5.1 Evaluate

∞ ∞
X 1 1 X 1 1 1
I= 4 + α4
= 4 + α4
−
n=1
n 2 n=−∞
n 2 α4

Consider the following integral

˛
π cot πz
I= dz
Γ z 4 + α4
P∞ ´
In general, when trying to evaluate n=−∞ f (n), we will consider Γ f (z)π cot πzdz.
The idea is based on the fact that π cot πz = π cos πz
sin πz has a simple pole at each integer value. The
residue of the integrand at each of these poles is Res(f (z)π cotPπz, z = n) = f (n), thus the sum of the
∞
residues associated with the integer values is the original sum n=−∞ f (n). We choose a large square as
contour Γ, which encloses the integers −N, −N + 1, ..., N − 1, N along the real axis, and consider the limit
of N → ∞. The function π cot πz converges to −πi away from the real axis in the positive imaginary side of
the complex plane and to πi away from the real axis in the negative imaginary side of the complex plane;
cf. Figure 4.32. Also we are considering that the vertical sides of our contour cross the real axis halfway
between two consecutive poles. So the function π cot πz is bounded along the contour. Further, we note
92 CHAPTER 4. COMPLEX INTEGRATION

Im z

Re z

Figure 4.31: Contour in example 4.5.1

 
 1  1
that  z4 +α4  ≤ R4 −α4 → 0 as N → ∞. Thus, we can assume that the integral along the contour vanishes:
¸ π cot πz
Γ z 4 +α4
dz = 0, which implies that
N 3
X π cot πz X π cot πz
Res( 4 4
, z = n) + Res( 4 , z = zj ) = 0
z +α j=0
z + α4
n=−N
| P∞
{z }
n=−∞ f (n)

1
where zj are the poles of the function f (z) = z 4 +α4 . Therefore, we obtain the formula
∞
X X
f (n) = − Res(f (z)π cot πz, z = zj )
n=−∞ j

iπ ijπ
where zj are the poles of f (z). In our case, zj = αe 4 + 2 , for j = 0, 1, 2, 3.. Thus Res(f (z)π cot πz, z =
π cot πz
zj ) = 4z3 j . After some simplification, we obtain
j

∞ √ √ !
X 1 1 1 √ sinh πα 2 + sin πα 2
4 4
=− 4
+ 4 2πα3 √ √ .
n=1
n +α 2α cosh πα 2 − cos πα 2

Example 4.5.2 Evaluate

∞
X 1
I(m) = 2m
, m = 1, 2, 3, . . .
k
k=1
´ π cot πz
Following the same idea, we start by noting that z 2m dz will vanish when integrated around the same
contour as shown in Figure 4.31, i.e.

2I(m) + Res (f (z), z = 0) = 0 .

The residue at z = 0 follows most easily from the Taylor expansion

π 2 2 π 4 4 2π 6 6
πz cot πz = 1 − z − z − z − ...
3 45 945
4.5. INFINITE SUMS 93

10 10

5 5

0 0

-5 -5

-10 -10
3 3

2 2

1 1
-3 -3
0 0
-2 -2
-1 -1 -1 -1
0 0
-2 1 -2 1
x 2 x 2
-3 3 y -3 3 y

(a) Re π cot πz. (b) Im π cot πz.

(c) |π cot πz|.

Figure 4.32: Graphical representation of f (z) = π cot πz, as seen from a viewpoint in the third quadrant
(i.e. the positive real axis is directed upwards and to the left).
94 CHAPTER 4. COMPLEX INTEGRATION

i.e.
π2 π4 π6
I(2) = , I(4) = , I(6) = , ...
6 90 945
Note: The P
Taylor coefficients above are closely related to the Bernoulli numbers
∞ P∞Bk , k = B0,2k1, 2, . . .2kdefined
by ezz−1 = k=0 Bk!k z k , since this relation can be rearranged into πz cot πz = k=0 (−1)k (2k)! (2πz) .

Example 4.5.3 Evaluate

∞
X 1
I= 2 − n2
n=−∞
ψ

Consider the following integral

˛
π cot πz
I= dz
Γ ψ2 − z2

We again consider that the integral vanishes along the contour, which is the same as in Figure 4.31. We notice
1 π cot πψ
that function f (z) = ψ2 −z 2 has two simple poles at z = ±ψ. Thus Res(f (z)π cot πz, z = ±ψ) = −2ψ .
Therefore
∞
X 1 π cot πψ
2 − n2
=
n=−∞
ψ ψ

With a slightly different notation, we obtain

∞
X 1 π cot πz
= 
n=−∞
z2 −n 2 z

For alternating series, we will use the function sinππz instead of π cot πz with the same contour as before.
πf (z) n
The residues of sin πz at each integer value n is
now (−1) f (n). In case an infinite ´sum contains a binomial
n 1 (1+z)n
factor, one might be able to use the relation k = {coeff. of z k in (1 + z)n } = 2πi C z k+1
dz where C is a
contour around the origin.

Example 4.5.4 Evaluate

∞  
X 2n 1
S= · .
n=0
n 5n

Substituting the integral above for the sum gives

∞  ∞ ˆ ˆ ∞ ˆ
!
(1 + z)2n dz (1 + z)2n

X 2n 1 1 X 1 X dz 5 dz
· n = · = · = .
n=0
n 5 2πi n=0 C (5z)n z 2πi C n=0
(5z)n z 2πi C 3z − 1 − z 2

It was here necessary that the geometric progression is convergent,

 which fails both when |z| is small and
 (1+z)2  4
when it is large. However, for |z| = 1, it holds that  5z  ≤ 5 , so we choose C as the unit circle. The
√ √
integrand has first order poles at z1 = 12 (3 − 5) ≈ 0.38 and at z2 = 12 (3 + 5) ≈ 2.62. Only the first one is
1 1
√ √
inside the unit circle, giving S = 5 · Res( 3z−1−z 2 , z = 2 (3 − 5)) = 5. 

4.6 Analytic continuation with use of contour integration

The three methods mentioned below apply to analytic functions which are defined by integrals, and for
which the limitations in the domains of validity come from divergence at one or both ends of the integration
interval.
4.6. ANALYTIC CONTINUATION WITH USE OF CONTOUR INTEGRATION 95

4.6.1 General change of contour for the defining integral

´∞ ez t
Example 4.6.1 Continue the function f (z) = −∞ 1+et
dt.

As it stands, this integral defines f (z) only for 0 < Re z < 1, since it otherwise diverges at t = −∞ and at
t = ∞, respectively. However, we showed in Example 4.4.10 that the integral (for such z-values) evaluates to
f (z) = sinππz . Therefore, this amounts to an analytic continuation of the f (z)-function to the full complex
plane. 

In a more typical situation, applying contour integration would produce, instead of a closed form expres-
sion another integral (arising from another side of the extended path), and for which the domain of validity
might be different (and hopefully larger) from what the case was for the original integral.

4.6.2 Modification of the integration path to a Hankel-type contour

´∞
Example 4.6.2 Continue the Gamma function Γ(z) = 0
e−t tz−1 dt.

The definition, as it stands, works only for Re z > 0, since it otherwise diverges at the lower integration limit
t = 0. ´
Let us consider C et tz−1 dt (note the sign in the exponent) over the path marked ”C” in Figure 4.33(entering
from −∞ just below the branch cut along the negative real axis, going around the origin and exiting back
to −∞ just above the cut).
Under the assumption that Re z > 0, the contribution when going around the origin will be vanishingly
small. With t = reiθ :
 z−1 −iπ(z−1)
z−1 r e : on path in towards origin (θ = −π)
t =
rz−1 eiπ(z−1) : on path in towards origin (θ = π)

Thus
ˆ ˆ ∞ ˆ ∞
−iπ(z−1) −r z−1
t z−1
et dt = e e r dr − e iπ(z−1)
e−r rz−1 dr
C 0 0
= −2i sin π(z − 1)Γ(z)
= 2i sin(πz)Γ(z)

from which follows

ˆ
1
Γ(z) = et tz−1 dt.
2i sin πz C

We next change the contour to the one that is marked ”H” (for Hankel) in the figure. This will not change
the value of the integral, but the fact that the contour now keeps some distance away from the origin means
that the integral will converge for all values of z, i.e. Γ(z) has been continued as a meromorphic function to
the complete complex plane. 

Figure 4.33: The contours C and H in Example 4.6.2.

96 CHAPTER 4. COMPLEX INTEGRATION

P∞ 1
Example 4.6.3 Continue the Riemann zeta function ζ(z) = n=1 nz

The definition, as it stands, works only for Re z > 1. In a first stage, we rewrite the sum into an integral,
which features the same domain restriction. However, the integral formulation can then be continued by
changing to a Hankel-type contour.
The change of variable t → nt in the integral definition of the Gamma function
ˆ ∞
Γ(z) = et tz−1 dt
0
´∞
gives 1
nz = 1
Γ(z) 0
tz−1 e−nt dt. Therefore

∞
X 1
ζ(z) = z
n=1
n
∞ ˆ ∞ 
1 X z−1 −nt
= t e dt
Γ(z) n=1 0
ˆ ∞ ∞
!
1 z−1
X
−nt
= t e dt
Γ(z) 0 n=1
ˆ ∞ z−1
1 t
= dt
Γ(z) 0 et − 1

For the same reason as with the original integral definition of the Gamma function, this integral representation
´ z−1
is limited to Re z > 0. The remedy is now exactly the same as in that case. After considering C ett −1 dt
where C is the same path as in Figure 4.33, and using the relation Γ(z)Γ(1 − z) = π/ sin(πz) to further
simplify the final result, we obtain ˆ
Γ(1 − z) tz−1
ζ(z) = t
dt (4.12)
2πi C e −1

We next change the contour from C to H where H is the same Hankel contour as previously. This new
representation for the zeta function is defined and finite for all z 6= 1 (the additional poles of the Gamma
function gets canceled by the fact that the integral vanishes when z = 1, 2, 3, ...
Apart from the branch singularity at t = 0, the integrand also features poles at t = ±2πi, ±4πi, ... It is
assumed that the contour H is not deformed so much that any of these fall inside it. However, in case that
H is so deformed that all these poles become enclosed by H, it is shown in Section 5.3 how this leads to a
proof of the functional equation for the zeta function:
1
ζ(z) = 2z π z−1 sin( πz)Γ(1 − z)ζ(1 − z) 
2

4.6.3 Modification of the integration path to a Pochhammer-type contour

The primary application for this method is to functions defined by an integral over a finite interval, and
where the domain restrictions come from singularities at both ends.
´1
Example 4.6.4 Continue the Beta function B(p, q) = 0
tp−1 (1 − t)q−1 dt.

The Beta function is analytic in both of its variables p and q. The integral definition works only when
Re p > 0 and Re q > 0 (since it otherwise diverges at t = 0 and at t = 1 respectively).
One way to continue the Beta function is to derive the closed form expression B(p, q) = Γ(p)Γ(q)/Γ(p+q),
and then rely on an analytically continued representation of the Gamma function.
For the Pochhammer approach, we note that the integrand g(t) = tp−1 (1 − t)q−1 is analytic in a complex
t−plane, apart from branch point singularities at t = 0 and t = 1 (unless p and/or q are integers). We start
at the origin in a complex t-plane, and move along the path marked 1 → 2 → 3 → 4 in Figure 4.34 a, taking
us back to the origin. The magnitude of g(t) will be single valued along this path, while the argument of
4.7. WEIERSTRASS PRODUCTS AND MITTAG-LEFFLER EXPANSIONS 97

Figure 4.34: The contours C and P in Example 4.6.4. For the contour C, the paths between t = 0 and t = 1
all coincide with the real line segment [0, 1], and the loops at t = 0 and t = 1 are infinitesimally small.

g(t) will change according to the sequence 0 → −2π(q − 1) → −2π(q − 1) − 2π(p − 1) → −2π(p − 1) → 0,
i.e. g(t) is back to its original value when the full trajectory is completed. We note that each of the two
singularities t = 0 and t = 1 has been gone around twice, but for each, the two times have been in opposite
(canceling) directions.
With Re p > 0 and Re q > 0, the contributions around the small near-circular parts will be vanishingly
small. The integral around the full loop will thus add up to
ˆ
= B − e−2π(q−1)i B + e−2π(q−1)i e−2π(p−1)i B − e−2π(p−1)i B
C
= −4e−iπ(p+q) sin pπ sin qπB

i.e.
ˆ
−eiπ(p+q)
B(p, q) = tp−1 (1 − t)q−1 dt.
4 sin pπ sin qπ C

We next change the contour from C to the one marked P (for Pochhammer) in Figure 4.34 b. This will
not change the value of the integral, but the fact that the contour now keeps some distance away from the
singularities implies that the integral will converge for all values of p and q.
We can note that it would not have been helpful to introduce branch cuts from the two branch points
(at t = 0 and t = 1) in order to create a single valued function. No matter how such cuts would have been
placed, the contour would have needed to move across these artificially introduced barriers. With a Riemann
surface approach to multi-valuedness, there is no difficulty in this regard. 

4.7 Weierstrass products and Mittag-Leffler expansions

Weierstrass Products Mittag-Leffler Expansions
Applies to entire functions Meromorphic functions
(poles are only singularities)
Describes a functions through its zeros poles
through its

The general treatment of these expansions is quite involved, so we will limit ourselves to illustrative
examples.
98 CHAPTER 4. COMPLEX INTEGRATION

4.7.1 Weierstrass products

Let f (z) be entire. Its only zeros are simple and at z = a1 , a2 , a3 , ..., an . Let pn (z) = (z − a1 )(z − a2 )(z −
a3 )...(z − an ). Then pfn(z) f (z)
(z) is an entire function. Therefore, h(z) = log pn (z) must also be free of singularities
and thus entire. So f (z) = pn (z)eh(z) . Just knowing the zeros of f (z) is enough to determine pn (z), but not
enough to exactly determine f (z).
We will now consider what happens when f (z) has infinitely many zeros. Let a1 , a2 , ... be the zeros of
f (z). We are this forming  
z
p(z) = Π∞
n=1 1 −
an
and are wondering whether this infinite product is convergent. We are investigating the tail end of the
product Π∞ z
n=M (1 − an ) for M  1 to see whether it converges to a limit, neither 0 nor ±∞.
P∞
Theorem 4.7.1 Π∞
n=1 (1 + un (z)) converges (uniformly) in z if n=1 |un (z)| converges (uniformly).

Example 4.7.1 Find a product expansion for sin πz.

The function sin πz has zeros at 0, ±1, ±2, ... so sinzπz has zeros at ±1, ±2, ... making it natural to form
n z z z  o n z z z o
p(z) = 1− 1− 1− ... 1+ 1+ 1+ ...
| 1 {z2 3 } | 1 {z2 3 }
We get harmonic series by the theorem above P∞ z
| |=|z|
P∞ 1
diverge
k=1 k k=1 k

Note that n z  z/k o

p(z) = Π∞
k=1 1− e
k
and n z  z/k o
p(z) = Π∞k=1 1 − e
k
 2
both converge. To see why, ez/k = 1 + kz + O kz 2 . Then each term becomes
 2  
z2
  3 
 z z z z
1− 1+ +O = 1 − + O
k k k2 k2 k3
P∞  2 P∞
and the resulting sum n=1  kz  = |z|2 k=1 k12 converges.
Using this on our example would just lead to the exponential terms e−z/k and ez/k to cancel out. Instead,
we can combine terms and obtain
z2
 
∞
p(z) = Πk=1 1 − 2
k
We will now expect that

z2
   
sin πz
= Π∞
k=1 1 − e h(z)
z k2
where h(z) is an entire function, which often turns out to have a quite simple form. In this case, it turns
out that eh(z) ≡ π
How can we see this? Taking the log on both sides, we obtain
∞
z2
X  
log(sin πz) = log z + h(z) + log 1 − 2
k
k=1

Differentiating, we get
∞
1 X 1
cot πz = + h0 (z) + 2z
z z 2 − k2
k=1
4.7. WEIERSTRASS PRODUCTS AND MITTAG-LEFFLER EXPANSIONS 99

From before,
∞
X 1
π cot πz = z
z2 − k2
k=−∞
0
Therefore, h (z) = 0 and thus h(z) = h is a constant. Taking the limit on both sides, we get that eh = π.
Therefore
z2
 
sin πz ∞
= Πk=1 1 − 2 (4.13)
πz k
A classical special case was computed by Wallis in 1670, who arrived at it by a different way. If z = 1/2,
then
π 2.2 4.4 6.6
= · · · ... 
2 1.3 3.5 5.7

4.7.2 Mittag-Leffler expansions

This method consists in representing a function in terms of its poles.
Example 4.7.2 Find a closed form expression for
∞
X 1
f (z) = .
(z − k)2
k=−∞

Clearly the series converges and it has period 1 (indeed, f (z + 1) = f (z)). Because it is 1-periodic, it has
the same Laurent expansion at every pole. Check at z = 0:
1 X 1
f (z) = 2
+
z k (1 − z/k)2
2
k6=0
2
z2

1 X 1 z
= + + 1 + + 2 + ...
z2 k2 k z
k6=0
∞
!
1 0 X 1
= 2
+ + 2 + 0z + (...)z 2 + ...
z z k2
k=1

π2
Next consider the function sin2 πz
. It is also 1-periodic. Expanding around the origin, we get the Laurent
expansion
π2 1 0 π2
= + + + 0z + ...
sin2 πz z2 z 3
2 P∞
Therefore, h(z) = sinπ2 πz − k=1 (z−k)
1
2 is an entire function and it is also 1-periodic. It is quite easy to

show that h(z) → 0 up and down the period strip shown in Figure 4.35. Thus by Liouville’s theorem,
h(z) ≡ constant. Since both functions tend to 0 far out in the period strip, this constant must be h = 0.
We obtain this result, which is a quite typical example of a Mittag-Leffler expansion
∞
π2 X 1
2 = . (4.14)
sin πz k=1 (z − k)2
P∞ 1
An interesting side result is the following alternative way to arrive at the sums I(m) = k=1 k2m , m=
1, 2, 3, . . . (cf. Example 4.5.2). Equating the Laurent series of both functions gives the sequence of relations:
∞
X 1 π2
=
k2 6
k=1
∞
X 1 π4
=
k4 90
k=1
···
100 CHAPTER 4. COMPLEX INTEGRATION

Im w

Period Strip
Re w

π2
P∞ 1
Figure 4.35: Period strip in sin2 πz
and in k=1 (z−k)2

Finding corresponding closed form expressions for odd positive integers 3,5,7,... is a longstanding open
problem.

Example 4.7.3 Find a closed form expression for

∞
X 1
f (z) = . (4.15)
z−k
k=−∞

As it stands, the series is only conditionally convergent. The following are two ideas to bypass this issue:
 
P∞ 1 1
P 1 1
1. Add a term which will cancel out. Notice that f (z) = k=1 z−k = z + k6=0 + .
z−k k
| {z }
 
1
size O k2
, so converges

2. Pair up the terms k and −k. We can again rewrite f (z) as

m ∞  
X 1 1 X 1 1
lim = + +
m→∞ z−k z z+k z−k
k=−m k=1
∞
1 X 1
= + 2z
z z 2 − k2
k=1

We next show two ways to evaluate this sum:

Method 1: We notice from the last equation above that this function has the same pole structure as
π cot πz. Furthermore, both functions have period one and are bounded up and down the period strip. By
Liouville’s theorem
P∞ again, we find that they differ only by a constant. Since π cot πz is an odd function and
1
since z1 + 2z k=1 z2 −k 2 is odd also, the only constant by which they could possibly differ is zero. Thus

∞
X 1
= π cot πz .
z−k
k=−∞
P∞ ´
Method 2: The standard approach for evaluating k=−∞ f (k) is to consider f (z)π cot πzdz around the
4.8. SELECT PROOFS 101

large rectangular path in Figure 4.31. As usual (although slightly more subtle to show in this particular
case), the integral vanishes
P∞ as the path is moved increasingly far out. To avoid a confusion of notation, we
1
write (4.15) as f (ξ) = k=−∞ ξ−k and thus consider the integral of π cot πz
ξ−z . Apart from at z integer, there
is one more pole, at z = ξ, with residue −π cot πξ. Since all residues should add up to zero, we arrive again
at f (ξ) = π cot πξ. 
Note that we can similarly find that
∞
X (−1)k π
= .
z−k sin πz
k=−∞

Mittag-Leffler expansions provide a supplementary approach to the one described in Section 4.5 for
evaluating many infinite sums.
Example 4.7.4 Find a closed form expression for
∞
X 1
S= . (4.16)
1 + k2
k=−∞

The present idea is to form the 1-periodic analytic function

∞
X 1
G(z) =
1 + (z + k)2
k=−∞

for which we want the value G(0). Considering G(z) in the complex plane, its only singularities will be first
N 1
order poles at z = −k ± i, k integer, with residues D = = ∓ 2i . Recalling that π cot πz has

0 2(z+k) 
z=−k±i
poles with residues +1 at all integers, it is clear that G(z) has exactly the same pole structure (locations
and residue) as the function
πi
H(z) = [cot(π(z + i)) − cot(π(z − i))] ,
2
cf. Figure 4.36. The difference G(z) − H(z) is therefore an entire, 1-periodic function which furthermore is
easily seen to approach zero when |Im z| → ∞. By Liouville’s theorem, the difference is identically zero, and
we obtain G(0) = H(0) = π coth π. 

4.8 Select proofs

4.9 Exercises
Exercise 4.9.1 Consider the function f (z) = (z + 1)1/2 (z − 1)1/2 . Does this function have no, one, or two
Taylor expansions around the origin? If any exist, give the first three non-zero coefficients.
Hints: (i) Recall the binomial expansion (1 + z)s = 1 + sz + s(s−1) 2
2! z +
s(s−1)(s−2) 3
3! z + ..., and (ii) Compare
with Figures 2.18 and 2.19.
Exercise 4.9.2 Consider the function f (z) = ((z + 1)(z − 1)(z − i))1/3 . It is proposed to make this function
single valued by means of making a branch cut as shown in Figure 4.37 (a). Determine if this branch cut is
sufficient by means of the following two approaches:
(a) Inspect if z = ∞ is a branch point (by making the change of variables z = 1/t).
(b) Follow the function around the path outlined in Figure 4.37 (b) and see if we get back to the same
function value as we started with.
¸
Exercise 4.9.3 Let Γ be the unit square, with corners at ±1 ± i. Evaluate Γ z 2 dz in the following three
ways:
(a) Give the answer directly by referring to some general theorem
(b) Evaluate explicitly the line integrals along the four sides, and add together the results.
(c) Use Green’s theorem to convert the contour integral over to a double integral, and then work out this
double integral. (i.e. do NOT refer to the Cauchy-Riemann equations).
102 CHAPTER 4. COMPLEX INTEGRATION

Figure 4.36: Magnitude / phase plot of the function G(z) = H(z) in Example 4.7.4.

(a) Branch cuts (b) Considered integration path.

Figure 4.37: Branch cuts and integration paths considered in Exercise 4.9.2.
4.9. EXERCISES 103

Figure 4.38: Six function illustrations for Exercise 4.9.5

¸
Exercise 4.9.4 Let Γ again be the unit square, with corners at ±1 ± i. Evaluate Γ z̄ 2 dz in the following
two ways:
(a) Evaluate explicitly the line integrals along the four sides, and add together the results.
(b) Use Green’s theorem to convert the contour integral over to a double integral, and then work out this
double integral.

Exercise 4.9.5 The six functions in Figure 4.38 (labeled a, b, c, d, e, f; with the real axis pointing towards
the right and slightly upwards) illustrate six of the twelve functions f (z) that are listed below. Match each
of the pictures with its corresponding function, and also tell whether the figure shows Re f (z), Im f (z), or
|f (z)| . Give thorough explanations!

• log z • z • ζ(z) • z −3 • (z − 2)i • z −4

• sin(iz) • z2 • i
z3 • z −1 • (z + 1)−3 • zi

Exercise 4.9.6 Prove the fundamental theorem of algebra directly from Theorem 4.2.3. Let f (z) = pn (z)
be a polynomial of degree n and choose as the integration path a circle |z| = R where R is very large.
´∞ dx 1 b

log z
Exercise 4.9.7 Show that 0 (x+a) 2 +b2 = b arctan a , (a ≥ 0, b > 0) by considering f (z) = (z+a)2 +b2 .

´∞ log x dx 1 b

√
Exercise 4.9.8 Show that 0 (x+a)2 +b2 = b arctan a log a2 + b2 , (a ≥ 0, b > 0) by considering f (z) =
(log z)2
(z+a)2 +b2 .

Exercise 4.9.9 One striking feature in Figure 4.39 is that the roots of the derivative of a polynomial seem
to fall within the convex hull of the polynomial’s roots. Prove the Gauss-Lucas theorem, which states that
this will always be the case.
Hints: Note that (i) it suffices to show that, if all roots of p(z) fall in the right half-plane, so do those of
p0 (z), (ii) the result is true for a P
root of p0 (z) that is also a root of p(z),
P and (iii) P
equation (4.2) tells that if
0 1 1 ak
z is a simple root of p (z), then z−ak = 0. Rearrange this into z |z−ak |2 = |z−ak |2 , and deduce the
result.
104 CHAPTER 4. COMPLEX INTEGRATION

1.5
3
1
2.5
1
0.5
2

0
1.5 0.5
-0.5
1

0.5 -1 0

0 -1.5
-0.5
-0.5 -2

-1 -2.5
-1

-2 -1 0 1 2 3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 -1.5 -1 -0.5 0 0.5 1 1.5

Figure 4.39: Three cases of randomly chosen degree 10 polynomials p10 (z) with their roots shown in red and
the 9 roots of p010 (z) shown in green.

P∞ 1
Exercise 4.9.10 Show that k=1 zk2 = −1 where the zk are the solutions to the equation z = ez .
z
−1
Hint: The function eez −z has a pole of residue one at each of these zk -locations. Therefore consider
´ z
e −1
C z 2 (ez −z)
dz around a large rectangle.
Chapter 5

Gamma, zeta, and related functions

5.1 The Γ- function

We introduced the Γ(z) function in Section 2.7 as
ˆ ∞
Γ(z) = e−t tz−1 dt
0

and showed that it satisfies the functional equations Γ(z + 1) = zΓ(z) (equation (2.18)) and (in Theorem
π
4.4.2; equation (4.11)) Γ(z)Γ(1 − z) = sin(πz) . It can be evaluated rapidly for all complex z by combining
either of the functional equations with the asymptotic expansion (10.21), which is highly accurate when Re z
is large. We can see from either (2.18)) or (3.6) that the only singularities of the Gamma function are poles
at the non-positive integers. As seen in Figure 5.1, the rapid growth in magnitude for increasing Re(z) is
associated with oscillations in the real part (and likewise in the imaginary part) as we move away from the
real axis. The function Γ(z) approaches zero very rapidly as we move out in the left half plane (except for
the string of poles along the negative real axis).

5.1.1 Product representations of the Gamma function

There are two important product representations of the Γ(z) function:

Theorem 5.1.1
n! nz
Γ(z) = lim (5.1)
n→∞ z(z + 1) · . . . · (z + n)

Theorem 5.1.2
∞ h
1 Y z  −z/n i
= z eγz 1+ e (5.2)
Γ(z) n=1
n
1 1 1


where γ = limn→∞ 1 + 2 + 3 + ... + n − log n ≈ 0.57722 .

Proofs for both are given in Section 5.3. From these relations, several further results follow. For example
(5.2) shows immediately that 1/Γ(z) is an entire function, and in particular that Γ(z) will never take the
1
value of zero. Forming Γ(z)Γ(−z) leads in a few steps (utilizing equation (4.13)) to the functional relation
π 1
Γ(z)Γ(1 − z) = sin(πz) (cf. equation (4.11)). Similarly, forming (with y real) leads to an exact
Γ(iy)Γ(iy)
formula for the magnitude of the Gamma function up/down the imaginary axis:
 1/2
π
|Γ(iy)| = .
y sinh πy

105
106 CHAPTER 5. GAMMA, ZETA, AND RELATED FUNCTIONS

(a) Re Γ(z). (b) |Γ(z)|.

Figure 5.1: Plots in the complex plane of Γ(z).

5.1.2 The Beta function

The Beta function B(p, q) is defined by
ˆ 1
B(p, q) = tp−1 (1 − t)q−1 dt (5.3)
0

(choosing the principal values for the powers) and is related to the Gamma function through the relation

Γ(p)Γ(q)
B(p, q) = , (5.4)
Γ(p + q)

proved below in Section 5.3. While the definition (5.3) only works for Re p > 0 and Re q > 0 (as the
integral otherwise diverges), we saw in Example 4.6.4 how it could be continued to all p and q by means of
a Pochhammer integral. The explicit formula (5.4) provides also such a continuation. Replacing t by 1 − τ
shows that B(p, q) = B(q, p). Other substitutions that lead to often useful identities are:

ˆπ/2
2
t = sin θ gives B(p, q) = 2 (sin θ)2p−1 (cos θ)2q−1 dθ (5.5)
0

and
ˆ∞
τ tp−1
t= gives B(p, q) = dt. (5.6)
1+τ (1 + t)p+q
0

In Sections ?? and ??, we will come across a generalized version of the Beta function,
´z which has one
additional argument z replacing one as the end point if the integration: B(z; p, q) = 0 tp−1 (1 − t)q−1 dt.

5.2 The ζ- function

P∞
In Section 3.2.3, we continued the Riemann ζ-function, introduced as ζ(z) = n=1 n1z , from Re z > 1 to
P∞ n+1
Re z > 0 (using the re-formulation ζ(z) = (1−211−z ) n=1 (−1) nz ), and finally to the entire complex plane
z z−1 1
by means of the functional equation ζ(z) = 2 π sin( 2 πz)Γ(1 − z)ζ(1 − z). We also found that ζ(z)’s only
singularity is a simple pole at z = 1, with residue 1.
5.2. THE ζ- FUNCTION 107

\sin\theta)ˆ{2p-1}(\cos\theta)ˆ{2q-

1}

Figure 5.2: Plots of ζ(z): Re ζ(z), |ζ(z)| and 1/|ζ(z)| in a strip surrounding the imaginary axis. Part (c)
shows first nontrivial zeros along the line Re z = 12 as narrow spikes.

Regarding values of the zeta function at integer arguments, we deduce from Example 4.5.2 that

(−1)k+1 B2k (2π)2k

ζ(2k) = , k = 1, 2, 3, . . .
2(2k)!

and then, via the functional equation, that

Bk+1
ζ(−k) = (−1)k , k = 0, 1, 2, . . .
k+1
P∞ Bk k
where the B-coefficients stand for the Bernoulli numbers defined by ezz−1 = k=0 k! z , i.e. B0 = 1,
B1 = − 12 , B2 = 16 , B3 = 0, B4 = − 301
, B5 = 0, B6 = 42 1
, . . . In particular, ζ(k) vanishes at all the negative
even integers - forming the trivial zeros of the zeta function.. Missing from this summary of values at integer
locations are the odd positive integers 3, 5, 7, . . . Surprisingly little is known about trese, with the exception
that ζ(3) is known to be irrational.

5.2.1 The critical strip

The critical strip denotes the vertical strip 0 ≤ Re z ≤ +1. Figure 5.2 shows the function in the somewhat
wider strip −5 ≤ Re z ≤ +5 surrounding the imaginary axis, and Figure 5.3 shows it along the real axis,
with the first few of the trivial zeros, located at the negative even integers, barely visible. Figure 5.4 shows
|ζ(z)| along two sections of the critical line Re z = 21 . The nontrivial zeros are seen to be quite irregularly
distributed.
The Riemann Hypothesis is one of the most famous unsolved problems in all of mathematics. It states
that all of the (infinitely many) nontrivial zeros lie exactly on the line Re z = 12 , by now known to be true
for the first 1013 of them. Proving or disproving the Riemann Hypothesis would have vast consequences in
mathematics.
108 CHAPTER 5. GAMMA, ZETA, AND RELATED FUNCTIONS

-10 -5 5 10

-1

Figure 5.3: ζ(x) plotted along the real axis. We notice the pole at x = 1. The trivial zeros at x =
−2, −4, −6, ... are on this scale visible.

3.0
5

2.5
4

2.0
3
1.5

2
1.0

1
0.5

10 20 30 40 110 120 130 140

(a) |ζ( 12 + it)| shown for 0 ≤ t ≤ 40. (b) |ζ( 12 + it)| shown for 100 ≤ t ≤ 140.

Figure 5.4: Plots of |ζ( 12 + it)|.

5.3. SELECT PROOFS 109

5.2.2 Relation to prime numbers

The ζ- function has a close connection to prime numbers. The exact distributions of the primes can be
deduced from the locations of the ζ- function’s nontrivial zeros - which however still (and for the foreseeable
future) contain many unsolved questions. The discovery in 1896 by Hadamard and de la Vallée Poussin
(independently) that none of these zeros can have real part equal to one sufficed to prove the prime number
theorem, limx→∞ x/π(x)
log(x) = 1, where π(x) is the number of primes not exceeding x.
Already the ancient Greeks knew that there are infinitely many prime numbers. As a minor illustration
of the extensive connections between primes and the ζ- function, we will show below this same result based
on the fact that ζ(z) has a pole at z = 1.
Theorem 5.2.1 There are infinitely many primes.
Proof Let {p1 , p2 , p3 , ...} = {2, 3, 5, ...} be the set of all the prime numbers. Then
Theorem 5.2.2 There are infinitely many primes.
   
1 1 1 1 1 1
1++ + ... 1 + + + ... 1 + + + ... ...
pz1 p2z
1 pz2 p2z
2 pz3 p2z
3
1 1 1 1 1
= z
+ z + z + z + z + ...
1 2 3 4 5
= ζ(z)
because every integer can be written exactly in one way as a product of primes. Each sum in the top line
above is an infinite geometric progression that can be summed in closed form, giving the relation
 
1
ζ(z) = 1/Π∞k=1 1 − . (5.7)
pzk
If there were only finite number of primes, this would be a finite and not an infinite product. For z = 1, it
would then evaluate to a finite rational number, contradicting that ζ(1) = ∞. 

5.3 Select proofs

z
n! n
Theorem 5.1.1 stated: Γ(z) = limn→∞ z(z+1)·...·(z+n) .
´ n −t z−1
n
Proof Since Γ(z) = limn→∞ 0 e t dt and e−t = limn→∞ (1 − nt , we can write
ˆ n n ˆ 1
t z−1 z
Γ(z) = lim 1− t dt = lim n (1 − τ )n τ z−1 dτ.
n→∞ 0 n n→∞ 0

For n integer, repeated integration by part gives

ˆ 1 ˆ 1 1 ˆ
n 1

1
(B(p, q) = tp−1 (1 − t)q−1 dt1 − τ )n τ z−1 dτ = τ z (1 − τ )n + (1 − τ )n−1 τ z−1 dτ
0 0 z 0 z 0
n(n − 1) · . . . · (1)
=... = . 
z(z + 1) · . . . · (z + n)
Q∞ 
1 z
e−z/n .


Theorem 5.1.2 stated: Γ(z) = z eγz n=1 1+ n

Proof Equation (5.1) can be written

1 h  z z  z  −z log n i
= lim z 1 + 1+ · ... · 1 + e
Γ(z) n→∞ 1 2 n
h  z −z
  z −z
  z  − z [1+ 12 +...+ n1 −log n]z i
= lim z 1 + e 1 1+ e 2 · ... · 1 + e n ·e
n→∞ 1 2 n
∞ h
γz
Y z  −z/n i
= ze 1+ e . 
n=1
n
110 CHAPTER 5. GAMMA, ZETA, AND RELATED FUNCTIONS

´1
Theorem 5.3.1 The Beta function B(p, q) = 0
tp−1 (1 − t)q−1 dt satisfies

Γ(p)Γ(q)
B(p, q) = .
Γ(p + q)

Proof Starting with the integrals for Γ(p) and Γ(q), we obtain
ˆ ∞ ˆ ∞
Γ(p)Γ(q) = e−t tp−1 dt e−t tq−1 dt
ˆ0 ∞ 0 ˆ ∞ 
−t p−1
= e t dt tq e−tx xq−1 dx
ˆ0 ∞ ˆ ∞ 0
= q−1
xe dx e−t(x+1) tp+q−1 dt
0 0
ˆ ∞ ˆ ∞
xq−1
= p+q
dx e−τ τ p+q−1 dτ
0 (1 + x) 0

and the result follows from using (5.6). 

Theorem 5.3.2 The ζ-function satisfies the functional equation

1
ζ(z) = 2z π z−1 sin( πz)Γ(1 − z)ζ(1 − z) . (5.8)
2
Proof The following outlines
P∞the key steps in one of the proofs provided by Riemann. Example 4.6.3 started
from the definition ζ(z) = n=1 n1z and then showed both
ˆ ∞
1 tz−1
ζ(z) = dt ,
Γ(z) 0 et − 1

for which the contour (the interval [0, +∞]) is indicated in part (a) of Figure 5.5, and
ˆ
Γ(1 − z) tz−1
ζ(z) = t
dt , (5.9)
2πi Γ2 e − 1

(now valid for all z), and where Γ2 is the Hankel contour shown in Figure 5.5 (b). This last integrand has
poles at t = ±2πi, ±4πi, . . ., (red dots in Figure 5.5 (c)) with residues n1−z1
e±πi/2(z−1) . A direct estimate
shows that the integral around the path Γ3 (shown in green in Figure 5.5 (c)) will vanish if N → ∞ (the
path is moved increasingly far out while passing half-way between poles on the imaginary axis). Hence
ζ(z) (based onP (5.9)) will equal Γ(1−z)
2πi · 2πi · {sum of residues}, where the sum of the residues becomes
π ∞ 1
2 cos( 2 (z − 1)) n=1 n1−z = 2 cos( π2 (z − 1))ζ(1 − z). The last equality holds for Re z < 0, but the resulting
identity will then by continuation hold for all z. A concluding variable change z → 1 − z gives the functional
equation in the form (5.8). 

5.4 Exercises
Exercise 5.4.1 Show that the primes pk , k = 1, 2, 3, . . . in the long run are more densely distributed that
the sequence k 1+ε , k = 1, 2, 3, . . ., no matter how small we choose ε > 0.
P∞ P∞
Hint: Show that k=1 p1k diverges, while k=1 k1+ε 1
converges. Equation (5.7) and the theorem below (from
basic calculus) are helpful.
P∞ P∞
Theorem
Q∞ 5.4.1 If the sequence {ak } is real and k=1 a2k converges, then the series k=1 ak and the product
k=1 (1 + ak ) converge or diverge together. Convergence of the product to zero occurs only if it has a zero
factor (i.e. an ak = −1).
5.4. EXERCISES 111

← +(2N-1)πi

↓ ↑

← -N +N
→

Γ1 → Γ2 Γ3

↓ ↑

→ -(2N-1)πi

Figure 5.5: The three integration paths Γ1 , Γ2 , Γ3 considered in the proof of Theorem 5.3.2, shown in parts
(a) - (c), respectively. There is in all cases a branch cut along the negative real axis.
112 CHAPTER 5. GAMMA, ZETA, AND RELATED FUNCTIONS
Chapter 6

Elliptic functions

6.1 Some introductory remarks on simply periodic functions

Consider first a function f (z) with a single period ω, satisfying f (z + ω) = f (z). An obvious example would
be f (z) = sin z, with period 2π, illustrated previously in Figure 2.6 and again here in Figure 6.1 (a). The
values it takes in its period strip, shaded in Figure 6.1 (b), repeats in each equally sized strip that is parallel
to it. The basic strip, as shown, is not uniquely determined, as we could shift it any amount sideways, say
to α ≤ Re z ≤ α + 2π where α is any real number. Using α = −π is also common in this case. For more
complicated periodic functions, with singularities, it can be convenient to choose α so that the edges of the
period strip become singularity free.
If a function is, say, 2π-periodic, it has of course also the periods −2π, ±4π, ±6π, . . . We note that the
ratio between these different choices is purely real valued, and we will in general focus on the smallest one,
choosing here ω = 2π.
When z is a complex variable (and, as such, typically illustrated in a 2-D complex plane), it becomes
natural to ask if it is possible for an analytic function to have two different periods ω1 and ω2 . If the ratio
τ = ω1 /ω2 is real, it is easy to see that there are no more choices possible than the trivial multiples described
above – each period will be an integer multiple of a smallest one. The case when the ratio τ = ω1 /ω2 is
not real leads us to the topic of the present chapter – elliptic functions. It will turn out that these types
of functions have important features, such as solving certain nonlinear ODEs in closed form. They also
provide a closed form expression for the circumference of an ellipse – hence their name. This is not possible
to achieve using only the standard functions of calculus.

6.2 Some basic properties of doubly periodic functions

Figure 6.2 illustrates the concept of a 2-D period box, here shown with the distinct periods ω1 = 1 and ω2 = i
(i.e. with the ratio τ = ω1 /ω2 not real). The first thing we note is that, apart from that we can translate
the period box (convenient in case we wish to avoid singularities on the box sides), we can also define it
genuinely differently, as shown with the dashed parallelogram.

Theorem 6.2.1 A non-constant doubly periodic function must have at least one singularity inside its period
box.

Proof If it does not have any singularity inside the period box, it will be bounded in magnitude there and
therefore, by the periodicities, everywhere across the complex plane. By Liouville’s theorem (Section 4.2.3),
it becomes a constant. 

Definition A doubly periodic meromorphic function (poles as the only singularities) is called an elliptic
function.

Theorem 6.2.2 A (non-constant) elliptic function has equally many poles as it has zeros inside the period
box.

113
114 CHAPTER 6. ELLIPTIC FUNCTIONS

Figure 6.1: (a) Re sin z, (b) A period strip (here 0 ≤ Re z ≤ 2π) for the function f (z) = sin z.

Proof Consider the difference between the number of zeros and number of poles in the period box: N − P =
1
´ f 0 (z) ´
2πi  f (z) dz = 0 (Theorem 4.2.3; here with  denoting integration around the four sides of the period
box). The last equality follows from the fact that f 0 (z) must have the same double periodicity, so the
contributions to the integral from the opposite box sides will cancel. The number N = P is called the order
of the elliptic function.

Corollary 6.2.3 An elliptic function takes every value equally many times.

Theorem 6.2.4 There does not exist any elliptic function of first order.

´ 0
 
residue at the 1
Proof If there was one, then = 2πi 
f (z)dz = 0 (by cancellation between the sides of
simple pole
the period box), i.e. the pole does in fact not exist.

6.3 The Weierstrass ℘-function

6.3.1 Construction
Although we have identified some properties that elliptic functions must possess if they exist, we have not
yet demonstrated that they in fact do exist. The easiest demonstration is to explicitly construct one. Our
first example will be the Weierstrass ℘-function. This function has a double pole at the origin, with zero
residue, and we denote its two periods by ω1 and ω2 (with τ = ω1 /ω2 not real).
First attempt of construction: Let us place a double pole of the type 1/z 2 at the origin and also at
each periodic repetition of this point, i.e. at ω = n1 ω1 + n2 ω2 where n1 and n2 run through all integers
(zero, positive and negative). A simplest attempt would be to consider
X 1
P (z) = . (6.1)
ω=n1 ω1 +n2 ω2
(z − ω)2

However, the convergence properties of this sum are not obvious. It is not absolutely convergent, as we can
6.3. THE WEIERSTRASS ℘-FUNCTION 115

Figure 6.2: Period box(es) for a function with primary periods ω1 = 1 and ω2 = i.

see by comparing with

ˆ∞ ˆ∞ ˆ2π ˆ∞ ˆ∞
 
1 1 1
dxdy =  dr dθ = 2π dr = ∞. (6.2)
r2 r 2 r
∞ ∞ 0 0 0

where we in all the integrals have excluded some region around the origin and only focused on behaviors
towards infinity.
Second attempt of construction: The idea will be to change the function P (z) from (6.1) in concept
as little as possible (so that it still features unit strength double poles at the same locations). So here is
another attempt:
1 X 1 1

℘(z) = 2 + − 2 . (6.3)
z (z − ω)2 ω
ω6=0

With the 1/z -term separated out in front, and then and having inserted −1/ω 2 inside the sum (which no
2

longer includes the case of ω = 0), each term becomes (for z fixed and ω large) of size O(1/ω 3 ) instead of
as before O(1/ω 2 ) . That extra power of ω means that we get one more power of r in the denominators of
(6.2) showing that (6.3) will converge for z fixed. With convergence thus taken care of, it remains to verify
that we did not lose the double periodicity when modifying (6.1) to (6.3).

Theorem 6.3.1 The Weierstrass ℘-function, as defined by (6.3), is doubly periodic.

Proof From (6.3) follows, by differentiation, that ℘0 (z) = −2 ω (z−ω)

1
P
3 , which converges (since its terms

are again of size O(1/ω )). The sum is also doubly periodic, implying that ℘0 (z + ωi ) − ℘0 (z) = 0, i = 1, 2.
3

Integration gives
℘(z + ωi ) − ℘(z) = C. (6.4)
Substituting z = −ωi /2 into (6.4) gives ℘(ωi /2) − ℘(−ωi /2) = C. Since (6.3) shows that ℘(z) is an even
function of z, the constant C must be zero. Equation (6.4) then confirms that indeed is doubly periodic.

Figures 6.3 and 6.4 show the ℘-function in the special case when the periods have been chosen as ω1 = 1
and ω2 = i. The pattern of equispaced double poles extends over the full complex plane.

6.3.2 Some nonlinear ODEs that are satisfied by the ℘-function

Elliptic functions satisfy various different nonlinear ODEs, many of which arise in applications. We continue
to use the ℘-function as illustrative example.
116 CHAPTER 6. ELLIPTIC FUNCTIONS

Figure 6.3: Real and imaginary parts of the Weierstrass ℘-function in the case of periods ω1 = 1 and ω2 = i.

1/k at z = iK 0

Figure 6.4: Magnitude and phase plot for the same case as in Figure 6.3.
6.3. THE WEIERSTRASS ℘-FUNCTION 117

The Laurent expansion of ℘(z) around the origin will take the form

1
℘(z) = + 0 + a2 z 2 + a4 z 4 + . . . (6.5)
z2

where the coefficients depend on the periods: ak = (k + 1) ω6=0 1/ω k+2 , k = 2, 4, 6, . . . Since ℘(z) is even,
P
all coefficients for odd powers vanish. That is also the case for a0 , seen from letting z → 0 in (6.3). From
(6.5) follows
2
℘0 (z) = − 3 + 2a2 z + 4a4 z 3 + . . .
z
and therefore
4 8a2
℘0 (z)2 = − 2 − 16a4 + . . .
z6 z
4 3
We obtain the same leading term z6 if we form 4[℘(z)] , implying that we can cancel it out by subtraction

20a2
℘0 (z)2 − 4℘(z)3 = − − 28a4 + . . .
z2

Adding 20a2 ℘(z) to this will additionally cancel out the first term in the RHS, leading to

℘0 (z)2 − 4℘(z)3 + 20a2 ℘(z) = −28a4 + . . . (6.6)

The LHS of (6.6) is doubly periodic with its only possible singularity location (in the period box) at z = 0.
However, the RHS is singularity free at this point. By Theorem 6.2.1, it must be a constant. All further
terms in the RHS of (6.6) (beyond −28a4 ) must therefore also have vanished.
It is conventional to introduce the new parameters g2 = 20a2 and g3 = −28a4 . We can then write the
ODE satisfied by ℘(z) as
℘0 (z)2 = 4℘(z)3 − g2 ℘(z) − g3 . (6.7)

If g2 and gg are specified, one can compute matching periods ω1 and ω2 from the formula for ak . The same
procedure as just employed will give many additional ODEs, such as

1
℘00 (z) = 6℘(z)2 − g2 , (6.8)
2

℘000 (z) = 12℘(z)℘0 (z),

etc. We will return to (6.8) in Section 9.5.1.

6.3.3 Closed form expressions for certain integrals based on the ℘-function
Equation (6.7) can be written as 1 = ℘0 (z)/ 4℘(z)3 − g2 ℘(z) − g3 . Using d℘
p dz
p dz = 1/ d℘ , we obtain for the
inverse function z(℘) the relation z 0 (℘) = 1/ 4℘3 − g2 ℘ − g3 . At this point, ℘ has become the name of an
independent variable, which we just as well can call, say, x. By means of the inverse function z(x), we have
thus obtained an analytic expression for the indefinite integral
ˆ
dx
p (6.9)
4x3 − g2 x − g3

for arbitrary constants g2 and g3 . By choosing other elliptic functions, (6.9) can be generalized to
ˆ
dx
√ .
arbitrary quartic polynomial in x
118 CHAPTER 6. ELLIPTIC FUNCTIONS

Figure 6.5: Illustration of the mapping w = λ(z). In particular, it maps the edge of the domain Ω shown in
part (a) to the real w-axis. The shaded regions (extending to infinity) are mapped one-to-one to each other.
It holds that λ(0) = 1, λ(1) = ∞, λ(∞) = 0.

6.3.4 The modular function λ(τ )

The RHS in (6.7) is a cubic, which can be factored as

℘0 (z)2 = 4(℘(z) − e1 )(℘(z) − e2 )(℘(z) − e3 ), (6.10)

where the ei , i = 1, 2, 3 are the three zeros of the order 3 elliptic function ℘0 (z). From the basic properties
of the ℘-function, it transpires that e1 = ℘(ω1 /2), e2 = ℘(ω2 /2), e3 = ℘((ω1 + ω2 )/2) will satisfy this (cf.
proof in Section 6.5). These ei -values are always different from each other. If one further notes that each
of them becomes multiplied by 1/t2 if the two ωi are both multiplied by t, one can conclude that the ratio
(e3 − e2 )/(e1 − e2 ) will become a single valued analytic function of the ratio τ = ω1 /ω2 . This function
(e3 −e2 )
λ(τ ) = (e 1 −e2 )
is known as a modular function. It is of great interest in its own right, independently of its
definition based on an elliptic function. We can’t here go through its properties in any detail (for that, see
more advanced books, such as Ahlfors [3] or Copson [1]), but will just mention that w = λ(z) maps the
boundary of the region Ω shown in Figure 6.5 (a) to the real axis, and the inside one-to-one to the upper
half plane (Figure 6.5 (b)). Given this, it is natural to display this function graphically over the region Ω
rather than, as we usually do, over the region surrounding origin in the complex plane – see Figure 6.6.
In Figure 6.7 (a), we show also the region Ω reflected across the imaginary axis, denoting this by Ω.
By Schwartz’ reflection principle, the function λ(z) will in this region take take complex conjugated values
comparet to in Ω, implying that its values will be as shown in 6.7 (b), where we note the branch cuts along
[−∞, 0] and [1, ∞]. We denote the inverse function to w = λ(z) as z = λ−1 (w). It can be shown to have no
other singularities than these branch points at w = 0 and w = 1 .

6.3.5 Picard’s theorem

This theorem provides intriguing insights into the possible behaviors of entire functions (functions that are
singularity free in the finite complex plane). As a background, let us first recall that a polynomial of degree
d takes every finite value d times (when counting the multiplicity of roots). We also know from Liouville’s
Theorem 4.2.6 4.2.6 that, if the magnitude of an entire function is bounded, it is a constant. Weierstrass
theorem ?? tells that an entire function comes arbitrary close to any finite value infinitely many times. Some
entire functions actually take (rather than just come arbitrarily close to) every value infinitely many times
(e.g. sin z) while others avoid one value altogether (e.g. ez is never zero for any finite z).

Theorem 6.3.2 (Picard, 18??) An entire function that avoids two values must be a constant.

Proof (The following is a brief sketch of the original proof by Picard): If the function F (z) avoids the two
(finite) values a and b, we consider f (z) = F (z)−a
b−a and it remains to prove that an entire function that avoids
6.4. THE JACOBI ELLIPTIC FUNCTIONS 119

Figure 6.6: Real and imaginary parts of the modular function λ(z). For part (a) Note that Re λ(z) increases
monotonically as z follows the boundary 1 → ∞ → 0 → 1. For (b) we note that Im λ(z) is zero along the
whole boundary and everywhere positive inside it.

both 0 and 1 must be a constant. The essential idea is to consider the function Λ(z) = λ−1 (f (z)) and then
apply Morera’s Theorem 4.1.3. If we choose any closed path in the z-plane, we know that f (z) along this
path can never encircle either of the points 0 or 1 (as we otherwise could shrink the path to a point at
which this would conflict with the fact that f (z) cannot take these values). This means that the integral
also of Λ(z) around the path will be zero, and thus (by Morera’s theorem) represent an analytic function of
z. However, the values of Λ(z) are all in the upper halfplane (cf. Figure 6.7 (a)). The function ei Λ(z) is then
a bounded entire function, and according to Liouville’s theorem a constant. Then, so are Λ(z), λ(z), f (z),
and F (z).

6.4 The Jacobi elliptic functions

This family of elliptic functions differs from the Weierstrass ℘-functions primarily in that the perid box
contains two first order poles (of opposite residues) instead of a single pole of order two and residue zero.
Much of the notations for these functions come from their connection with conformal mappings - see Section
7.5. For example, the Jacobi sn(z) function arises there in the implicit form

ˆ
sn(z)
du
z= 1/2
, (6.11)
[(1 − u2 )(1 − k 2 u2 )]
0

where k is a parameter in the range 0 ≤ k ≤ 1. However, we start here with a more direct approach to the
Jacobi elliptic functions by defining a periodic box, and then just placing two poles inside it.

6.4.1 The Jacobi sn(z,k ) function

We first create a period box of size 0 ≤ Re z ≤ 4K, 0 ≤ Im z ≤ 2K 0 , where both K and K 0 depend on a
single free parameter k ∈ [0, 1] through the relations
ˆ1
du p
K(k) = 1/2
and K 0 (k) = K(k 0 ), with k 0 = 1 − k2 . (6.12)
[(1 − u2 )(1 − k 2 u2 )]
0
120 CHAPTER 6. ELLIPTIC FUNCTIONS

Figure 6.7: The one-to-one mapping given by λ(z) between a part of the upper half plane (composed of the
regions Ω and Ω) and the entire complex plane. Note in this the branch cuts along [−∞, 0] and [1, ∞].

6
K
5
K'

2
π/2
1

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
k

Figure 6.8: The two functions K(k) and K 0 (k). K(0) = K 0 (1) = π
2. The two curves intersect at k = √1 ,
2

2 √
with K( √12 ) = K 0 ( √12 ) = Γ( 41 ) /(4 π).

(We follow here again notational conventions - K and K 0 are related functions, and k and k 0 are related
parameter values, with the primes here having nothing to do with derivatives). Exercise 6.6.1 shows how
this function K(k) (and thereby also K 0 (k)) can be calculated numerically extremely rapidly. Figure 6.8
shows the two functions. In the period box, we next place a first order pole with residue +1/k at z = iK 0
and one with residue −1/k at z = 2K + iK 0 . Repeating this this period box across the full complex plane,
and summing the contributions from all the resulting infinitely many poles gives, like for our first attempt
of constructing the Weierstrass ℘-function, rise to a divergent double sum. However, this divergence can be
avoided similarly to the ‘second attempt’ (also in Section 6.3.1). The result is known as the Jacobi sn(z, k)
function, illustrated over one period box in Figure 6.9.

If we let the parameter k approach zero, the period box approaches 2π in width, while its height increases
towards infinity. The sn(z, k) function then approaches sin z (which we can see indications of along the real
axis already in Figure 6.9 (a)). In the opposite limit of k approaching one, the period box increases towards
infinity in width while shrinking in height to π, and sn(z, k) then approaches tanh z. The red curves in
Figure 6.10 show these variations in sn(x, k) over one period [−2K, +2K] in the real variable x for some
different values of k.
6.4. THE JACOBI ELLIPTIC FUNCTIONS 121

5 5

0 0

-5 -5

4 4

3 3

2 2
6 6
5 5
1 4 1 4
3 3
y 2 y 2
1 1
0 x 0 x
0 0

(a) Re sn(z, k) for k = 12 . (b) Im sn(z, k) for k = 12 .

(c) |sn(z, k)| for k = 12 .

Figure 6.9: Visualization across one period box of sn(z, k)in the case of k = 21 . For this k-value, K ≈
1.6857503548 and K 0 ≈ 2.1565156475, i.e. the period box is roughly of size [0, 6.7430] × [0, 4.3130]. In parts
(a) and (b), some positive contour lines are highlighted in blue and negative ones in green. As usual, values
along the real axis are shown in red for (a) and (b), and in black for (c).
122 CHAPTER 6. ELLIPTIC FUNCTIONS

k = 0.999
1

-1
-10 -8 -6 -4 -2 0 2 4 6 8 10
k = 0.99
1

-1
-10 -8 -6 -4 -2 0 2 4 6 8 10
k = 0.9
1

-1
-10 -8 -6 -4 -2 0 2 4 6 8 10
k=0
1
sn
0 cn
dn
-1
-10 -8 -6 -4 -2 0 2 4 6 8 10

Figure 6.10: The variations of the sn(x, k), cn(x, k) and dn(s, k)-functions along the real axis over one spatial
period, in the cases of four different k-values.
6.5. SELECT PROOFS 123

6.4.2 Some other Jacobi elliptic functions

Similar to how we from sin z obtain cos z, we can from sn(z, k) create numerous additional elliptic functions,
with the most commonly used ones being cn(z, k) and dn(z, k), defined through the relations

cn2 (z, k) + sn2 (z, k) = 1 and dn2 (z, k) + k 2 sn2 (z, k) = 1.

The green and blue curves in Figure 6.10 show counterparts to the red sn-curves for the functions cn(x, k)
and dn(x, k), respectively. These elliptic functions solve numerous nonlinear ODEs in closed form. For
example:
 2
d2 y 2 2 3 dy
sn(x, k) solves dx 2 + (1 + k )y − 2k y = 0 and dx = (1 − y 2 )(1 − k 2 y 2 )
  2
d2 y 2 2 3 dy
cn(x, k) solves dx 2 + (1 − 2k )y + 2k y = 0 and dx = (1 − y 2 )(1 − k 2 + k 2 y 2 )
  2
d2 y 2 3 dy
dn(x, k) solves dx 2 − (2 − k )y + 2y =0 and dx = (y 2 − 1)(1 − k 2 − y 2 ).

There is also a wealth of closed form integrals and other relations satisfied by the Jacobi elliptic functions.
The NIST handbook [??], Chapter 22, contains a summary. One of the integrals provides a closed form
expression for the circumference of an ellipse.

6.5 Select proofs

Theorem 6.5.1 The three zeros of ℘0 (z) in the period box are z1 = ω1 /2, z2 = ω2 /2, and z3 = (ω1 + ω2 )/2.

Proof Since ℘(z) is even and has a period ω1 , it holds that ℘(z) = ℘(−z) = ℘(ω1 − z), and therefore
℘0 (z) = −℘0 (ω1 − z). For z = ω1 /2,we then get ℘0 (ω1 /2) = −℘0 (ω1 /2), i.e. ℘0 (ω1 /2) = 0. The same
argumant can be used for the other two roots. Since ℘0 (z) is an elliptic function of order 3, it cannot have
any further zeros (in the period box). 

6.6 Exercises
Exercise 6.6.1 The arithmetic-geometric mean (AGM): If one starts with two non-negative numbers a and
b, sets a0 = a, b0 = b, and then iterates
a n + bn p
an+1 = , bn+1 = an bn , n = 0, 1, 2, . . . , (6.13)
2
the two sequences {an } and {bn } approach each other extremely rapidly, converging to a joint limit we denote
AGM (a, b). The convergence rate is quadratic, meaning that the number of digits in common roughly doubles
each iteration. In 1799, C.F. Gauss, by numerical hand calculations, ‘stumbled on to’ and then quickly proved
that
π
K(k) =
2 AGM (1, k 0 )
(and thus K 0 (k) = 2 AGM
π
(1,k) ), in the notation of equation (6.12).
For some k-values in the range k ∈ [0, 1], evaluate the corresponding values for K(k) and K 0 (k). Verify
that your results match what is qouted in the captions of Figures 6.8 and 6.9.

Exercise 6.6.2 Gauss found different analytic ways to link the AGM to integrals. One of his approaches
was based on showing that
 2  2  2
1 1 1·3 1·3·5
=1+ x2 + x4 + x6 + . . .
AGM (1 + x, 1 − x) 2 2·4 2·4·6
ˆπ
1 dθ
= √ .
π 1 − x2 cos2 θ
0
124 CHAPTER 6. ELLIPTIC FUNCTIONS

Another is his observation that (for x, y ≥ 0)

ˆπ/2 ˆπ/2
dθ dϕ
= (6.14)
√
p q
x2 cos2 θ + y 2 sin2 θ x+y 2


cos2 ϕ + xy

2
sin2 ϕ
0 0 2

2x sin ϕ
(by means of the ingenious change of varialble sin θ = (x+y)+(x−y) sin2 ϕ
). Assuming the truth of (6.14), show
that
ˆπ/2
dθ π
p = .
2 2 2
x cos θ + y sin θ2 2 AGM (x, y)
0

Exercise 6.6.3 The Jacobi elliptic function sn(z, k) is characterized by having first order poles of opposite
residue placed at the locations z = 0 + iK 0 and z = 2K + iK 0 within the period box, of horizontal and vertical
sides of lengths 4K and 2K 0 , respectively. Show that immediate double summation over all repetitions of this
period box leads to a not absolutely convergent sum, and then show how a certain re-grouping of the terms
will provide this (convergence even if we take the magnitude of all its terms).
Chapter 7

Conformal Mappings

The previous chapters have contained many illustrations of analytic functions f (z), often in the form of
displaying Re f (z), Im f (z), and / or |f (z)| as surfaces over the complex z-plane. A different type of
illustration is obtained by choosing some region in the z-plane and then showing in a complex w-plane
where the corresponding values of w = f (z) end up. Figure 7.1 (a) shows an arbitrarily chosen region in
the z-plane, and parts (b-f) the matching regions in the w-plane for some different simple choices of f (z).
All these cases illustrate one-to-one mappings for the displayed regions (i.e. over these regions, the inverse
function z = f −1 (w) is also single-valued). Only case (b) (a combination of a translation and a rotation)
would be single valued in both directions for arbitrary regions.
Maybe the most striking feature in these figures is that the grid, shown in green in part (a), will correspond
to a curve set in the w-plane where again all the curves intersect at right angles. This will turn out to be
the case at any point where f (z) is analytic, with f 0 (z) 6= 0. This furthermore is a special case of a more
general property - not only right angles but all angles are locally preserved - and, in reverse, only analytic
functions w = f (z) can have this conformal mapping property.
Figure 7.1 was somewhat unsatisfactory in that neither the initial region in the z-plane, nor the resulting
regions in the w-plane were of any special significance. What makes the topic of conformal mappings
important for certain applications is that one can specify both regions independently, and there will then
exist (a 3-parameter family of) analytic functions w = f (z) that performs that mapping in a one-to-one
manner. Although it is somewhat rare to be able to find an appropriate f (z) in closed form, it can typically
be computed accurately (with algorithms that fall outside the scope of this book).
The rest of this chapter will highlight

• Relation between conformal mappings and analytic functions (Section 7.1)

• Mappings provided by bilinear functions (Section 7.2)

• Mappings of polygonal regions (Section 7.4)

• Riemann’s mapping theorem (Section 7.3)

• Some applications of conformal mappings (Section 7.6)

7.1 Relations between conformal mappings and analytic functions


u = u(x, y)
A pair of functions (u, v, x, y real) form a conformal (angle-preserving) map if any two inter-
v = v(x, y)
secting curve segments in the x, y-plane become two intersecting curve segments in the u, v-plane, such that
the angle between the curves (at the intersection points) has stayed the same. The following theorem shows
that this property is unique to analytic functions.

125
126 CHAPTER 7. CONFORMAL MAPPINGS

2
2.5
7
1.8

6
1.6 2

1.4
5

1.2 1.5
y

y
1
y

3
0.8 1

0.6
2

0.4 0.5
1
0.2

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0.5 1 1.5 2 -4 -3 -2 -1 1 2 3 4
x x x

1+i
(a) Initial domain in complex z-plane. (b) The mapping w = 1 + √
2
z. (c) The mapping w = z 2 .

2.5

7
2

1.5

y
5

0.6
4
y

0.5 1
3
0.4
y

0.3 2
0.5
0.2
1
0.1

0.2 0.4 0.6 0.8 1 1.2 -3 -2 -1 1 2 3 4 5 6 7 -1.5 -1 -0.5 0.5 1 1.5

x x x

1/3 z
(d) The mapping w = z . (e) The mapping w = e . (f) The mapping w = arctan z.

Figure 7.1: (a) An arbitrarily chosen domain in the first quadrant, (b)-(f) The images of this domain (and of
the green lattice) when mapped to a complex w-plane through some different mapping functions w = f (z).
7.2. MAPPINGS PROVIDED BY BILINEAR FUNCTIONS 127

u = u(x, y)
Theorem 7.1.1 The mapping with ux uy , vx , vy continuous is conformal at a location z =
v = v(x, y)
x + iy if and only if f (z) = u + iv is an analytic function, with f 0 (z) 6= 0.

Proof If f (z) is analytic at z0 , its Taylor expansion there starts w = f (z) = f (z0 ) + (z − z0 )f 0 (z0 ) + . . ., i.e.
it is locally around z0 a combination of a translation (given by f (z0 )), a uniform scaling (given by |f 0 (z0 )|),
and a rotation (given by arg f 0 (z)), all of which preserves angles.
The proof for the reverse statement (local angles preserved implies analyticity) is more lengthy, and is
given in Section 7.7 as Theorem 7.7.1. 

The next theorem makes it relatively simple to check that a mapping function is one-to-one.

Theorem 7.1.2 Let D be a finite region in the z-plane, with boundary δD, corresponding to E and δE in
the w = f (z) plane, with neither boundary self-intersecting. Then
(a) w = f (z) provides a one-to-one map from D to E, and
(b) If z moves around δC in the positive direction, so does w around δE.

Proof Let w0 be a point inside E, and consider

ˆ ˆ
1 f 0 (z) dz 1 dw
I= =
2πi f (z) − w0 2πi w − w0
δD δE

where z moves around δD in the positive direction. By Theorem 4.2.3, I = N − P where N and P are the
number of zeros and poles, respectively, of f (z) − w0 . With P = 0, N ≥ 0, and the RHS = ±1 according to
the direction w moves around δE, the theorem follows. 

7.2 Mappings provided by bilinear functions

It turns out that bilinear functions
az+b
w= , (7.1)
cz+d
the ratio between two linear expressions in z, has a number of remarkable properties when it comes to
mapping simple regions:

Theorem 7.2.1 Any straight line or circle in the z-plane becomes mapped by a bilinear function into another
circle or straight line.

Proof If c = 0, the result is obvious, since (7.1) then just amounts to a sequence of (i) scaling/rotation
(forming a z) , (ii) shift (adding b), and (iii) scaling/rotation (dividing by d), none of which can change a
circle/line away from still being a circle/line. For c 6= 0, we rewrite (7.1) as
 
az+b 1 1
w= = (bc − ad) +a ,
cz+d c cz + d

and it is again just a sequence of scaling/rotations and shifts, with the only novelty that it also includes
one inversion. Hence, it remains to show that if z is on a circle/line in the complex plane, so is w = 1/z.
However, this follows from Exercises 1.4.7 and 1.4.8. A circle/line in the complex plane becomes a circle on
the stereographic sphere, and inversion just rotates the sphere, i.e. preserves the circular shape. Then, it
again becomes a circle/sphere when projected back to the complex plane. 

Theorem 7.2.2 Successive bilinear mappings become again a bilinear mapping.

Proof Straightforward algebra shows that w = ac11z+d

z+b1
1
followed by ζ = a2 w+b2
c2 w+d2 can be written as ζ = a3 z+b3
c3 z+d3 ,
where     
a3 b3 a 2 b2 a1 b1
= .  (7.2)
c3 d3 c2 d2 c1 d1
128 CHAPTER 7. CONFORMAL MAPPINGS

z-plane w-plane z-plane w-plane

1 1 1 1

-1 1 -1 1 1 -1 1
-1 -1 -1 -1

Figure 7.2: Mapping by the bilinear function w = 1/z of the inside of the unit circle to its outside, and by
a shifted unit circle to a half-plane.

z-plane w-plane z-plane w-plane

1 1
1 1

-1 1 -1 1 -1 1 -1 -11
-1 -1

Figure 7.3: Two mappings by the bilinear function w = (z − 21 )/(1 − 12 z). Since this function is of the form
w = (z − a)/(1 − a z), the mapping takes the unit circle to itself (red) and inside to inside and outside to
outside. The blue circle becomes again a circle (since the mapping is of bilinear form).

a1 z+b1
Theorem 7.2.3 The inverse of a bilinear mapping w = c1 z+d1 takes also the form of a bilinear map
z = ac22w+d
w+b2
2
.
 
a1 b1
Proof We can assume det 6= 0, since otherwise w = ac11z+d
z+b1
reduces to w = constant (with no
c1 d1 1
   −1
a2 b2 a1 b1
z-dependence). By (7.2), we should choose as , since the combined mapping then
c2 d 2 c1 d 1
becomes the identity, i.e. starting from z, we get back to z. 

Among the family of bilinear mappings, the special one

z−a
w= , (7.3)
1 − az
where a is an arbitrary complex constant, is particularly important due to the following result:

Theorem 7.2.4 The bilinear mapping (7.3) maps z = a → 0 and |z| = 1 to |w| = 1.
  
z−a z−a
Proof The result z = a → 0 is obvious. The second part needs verification: |w|2 = 1−a z 1−a z =
z z−a z−z a+a a
1−a z−a z+a a z z = 1. In the last equality, we utilized twice that z is on the unit circle, i.e. that z z = 1. 

Figures 7.2 and 7.3 illustrate the mappings performed by two different bilinear functions. This can be
contrasted to the mapping in Figure 7.4, which is not bilinear, and hence can map a circle to a different type
of curve - here to an ellipse. The function is this figure is the same as the function shown earlier in Figure
2.2.
7.3. RIEMANN’S MAPPING THEOREM 129

z-plane w-plane z-plane w-plane

1
1 1 1

-1 1 -1-1 1 -1-1 1 -1-1 1

-1

Figure 7.4: Two mapping by the function w = z + 1/z. This function is not bilinear, and we note in
the second part a circular arc becoming an elliptic arc. The mapping is singular at the points z = ±1
(corresponding to w = ±2), and local angles are doubled at these points.

Theorem 7.2.5 There is no other way to conformally (one-to-one) map the inside of the unit circle and
its center to themselves than through w = λz, where λ is a complex constant of magnitude one.

Proof Let w = f (z) be another such mapping, and consider g(z) = f (z)/z. The singularity at z = 0
is removable, and f (z) cannot have any other zero inside |z| = 1 (since the mapping is one-to-one), and
|g(z)| = 1 around |z| = 1, i.e. |g(z)| has to be identically one (cf. Section 2.2.3 and also Theorem 4.2.10). 

The theorem above can quite easily be strengthened to apply also when f (z) is known to be analytic only
for |z| < 1 rather than for |z| ≤ 1 by considering a limit process as |z| % 1; see for ex. Schwarz’ Lemma in
[??,??,??].
Between Theorems 7.2.4 and 7.2.5, we can conclude that a mapping from the unit circle to itself has
precisely three free (real) parameters, of which two can be used to choose the (complex) point a and the third
to specify λ (obeying |λ| = 1). Next section tells how this result can be greatly generalized from the unit
circle to arbitrarily shaped domains (assumed to be simply connected, neither the whole plain, nor regions
with multiple ‘holes’ as seen in the last subplot of Figure 7.3).

7.3 Riemann’s mapping theorem

Theorem 7.3.1 Let D be a simply connected domain. There exists then a 3-parameter family of functions
f (z) that maps the unit circle one-to-one to D (or in reverse, D to the unit circle).

This result was formulated by Riemann as a small part of his 1851 Ph.D. thesis. Weierstrass later found a
flaw in the proof, and the first rigorous one was not given until 1912. While strict proofs remain technically
difficult, we sketch in Section 7.7 a heuristic argument that makes the theorem seem very plausible.
Current rigorous proofs are all of ‘existence type’ and not ‘constructive’ for actually obtaining a mapping
function from the domain shape D. There is however a rich literature on how to achieve this numerically
[??,??,??]. In the particular case when D is a polygon, there exists a systematic approach that we will
describe next. However, even this will require computational methods for all but the simplest cases [??,??].

7.4 Mappings of polygonal regions

The classical method for mapping to the inside of a polygon comes in two versions dependent on the region
we map from; either from the upper half-plane or from the inside of the unit circle (without an intermediate
bilinear mapping going via the upper half-plane). The resulting formulas turn out to be virtually identical
in the two cases. We follow here the first choice - for the second one, see for ex. Ahlfors [3], Chapter 6.
The following example of creating a mapping function from the upper half-plane to the inside of a triangle
generalizes immediately to the case when the target region instead is a general polygon.
130 CHAPTER 7. CONFORMAL MAPPINGS

z-plane w-plane w-plane

πα 2 w2

w0
πα 1
z 1 =-1 z 2 =+1 πα 0 w1

Figure 7.5: Steps in mapping the upper half-plane to the inside of a traingle, as described in Example
7.4.1. In part (c), the corners are denoted w0 , w1 , w2 and the exterior angles are πα0 , πα1 , πα2 with here
α0 = 17/20, α1 = 3/4, α2 = 2/5.

Example 7.4.1 Map conformally the upper half-plane from Figure 7.5 (a) to the inside of the triangle shown
in Figure 7.5 (b).

We simplify first the problem by rotating and translating the target triangle (and scaling it if we so wish), so
one of its sides (here marked blue) falls along the w-axis, cf. Figure 7.5 (c). Let then z0 = −∞ correspond to
w0 = w(−∞), z1 correspond to w1 , and z2 correspond to w2 . Since the mapping has three free parameters,
we can for convenience choose z1 = −1 and z2 = +1.When a point z moves along the z-axis over [−∞, z1 ],
the mapping function w(z) needs to satisfy arg w0 (z) = 0 (so that w(z) follows the first side of the triangle,
shown in blue). Then for z ∈ [z1 , z2 ], it has to hold that arg w0 (z) = πα1 (to match the slope of the second
triangle side, shown in red). Finally for z ∈ [z2 , +∞] arg w0 (z) needs to be increased with an additional πα2
(with side shown in green). In other words, arg w0 (z) needs, along the real z-axis, to be a piecewise constant
function which at z = z1 and z = z2 jumps by the amounts πα1 and πα2 , respectively.
Recalling the behavior of the log z function, cf. Figure 2.7 (b), we recognize in this function the perfect
‘building block’ for creating functions that are piecewise constant along the real axis. For ex.

g(z) = −α1 log(z1 − z) − α2 log(z2 − z)

looks over the upper half-plane as shown in Figure 7.6 in the present case of z1 = −1, α1 = 3/4 and z2 = +1,
α2 = 2/5 (i.e. the jumps are of sizes πα1 and πα2 , respectively, matching the angles in Figure 7.5).
What we needed was a function w0 (z) such that arg w0 (z) (rather than Im w0 (z)) has this step structure.
However, the Euler identity ez = ex+i y = ex (cos y + i sin y) tells that the two functions arg and Im are
closely related. For any complex z, it holds that

arg ez = Im z .

Hence, we form
w0 (z) = eg(z) = (z1 − z)−α1 (z2 − z)−α2 , (7.4)
and arg w0 (z) now matches the requirements imposed by the angles and the straight sides of the triangle.
Integrating (7.4) introduces an integration constant B (corresponding to a translation of the triangle) and
we also introduce a factor A that scales and rotates the triangle, as desired. With these two constants A
and B, can be brought back to its original location, orientation, and size. 
The further generalization from a target triangle to a general target polygon with corners at w0 , w1 , . . . wn
is now immediate:

Theorem 7.4.1 (Schwarz-Christoffel) The function w(z) that maps the upper half-plane to the inside of
a polygon, and [−∞, z1 , z2 , . . . , zn ] (on the real axis) to the polygon with corners [w0 , w1 , w2 , . . . , wn ] and
7.5. RE-VISITING THE JACOBI ELLIPTIC FUNCTION SN(Z) 131

-1
1

0.8

0.6

0.4 3
2
1
0.2 0
y -1
0 -2
-3 x

Figure 7.6: Imaginary part of the function g(z) = −α1 log(−1 − z) − α2 log(+1 − z), constructed to be
piecewise constant along the real axis, with jumps of sizes 43 π and 25 π at the locations z1 = −1 and z2 = 1,
respectively. The colors correspond to those of Figure 7.5.

external angles [πα0 , πα1 , πα2 , . . . , παn ] takes the form

ˆz
w(z) = A (ξ − z1 )−α1 (ξ − z2 )−α2 · . . . · (ξ − zn )−αn dξ + B . (7.5)
−∞

In practical use of these mappings obtained by the Schwarz-Christoffel theorem, one should note

• Three (real-valued) parameters can be chosen arbitrarily. We have above used one to enforce that
z = −∞ maps to the polygon corner w0 . Given that the integral typically is difficult to handle in
closed form, further choosing z1 = −1 and z2 = +1 may simplify the algebra.

• The constants A, B, and z3 , z4 , . . . , zn will then need to be obtained from enforcing that (7.5) obeys
the mapping requirements that z = [−∞, z1 , z2 , . . . , zn ] correspond to w = [w0 , w1 , w2 , . . . , wn ]. The
algebra in this step may become difficult or impossible unless approached numerically.

Example 7.4.2 ...

Example 7.4.3 ...

Example 7.4.4 ...

7.5 Re-visiting the Jacobi elliptic function sn(z )

From our discussion of the Schwarz-Christoffel mapping, it becomes apparent that (6.11) represents the
conformal mapping illustrated in Figure 7.7 - mapping the upper half-plane shown in part (a) to the inside
of the rectangle in part (b). The inverse function z(w) is clearly real-valued whenever w lies on one of the
straight sides shown in part (b). By Schwarz reflection principle (Section 3.2.2), reflecting this rectangle as
indicated by going from the red rectangle in Figure 7.8 to the composite of four rectangles shown here in
four black (dashed line) rectangles will have created a doubly periodic function. As such, we can shift the
period box as shown to arrive at the rectangle marked in green. By this, we have recovered exactly the same
Jacobi elliptic sn(z) function as we introduced previously in Section 6.4. This confirms that (6.11) indeed
represents a doubly periodic function over the claimed period box.
132 CHAPTER 7. CONFORMAL MAPPINGS

z-plane
w-plane

iK'(k)

-K(k) K(k)
-1/k -1 +1 +1/k

Figure 7.7: (a) Upper half-plane with the marked segments color-coded, (b) The resulting map in the w-plane
through the relation (6.11), with matching color coding..

2iK'

iK'

-K K 3K

Figure 7.8: Illustrations of reflecting and shifting the rectangle shown in Figure 7.7 (b) to obtain the period-
box for the Jacobi elliptic function sn(z) .
7.6. SOME APPLICATIONS OF CONFORMAL MAPPINGS 133

7.6 Some applications of conformal mappings

7.7 Select proofs

u = u(x, y)
Theorem 7.7.1 If the mapping with ux uy , vx , vy continuous is conformal in the vicinity of
v = v(x, y)
a location z = x + iy , then f (z) = u + iv is an analytic function, with f 0 (z) 6= 0.

Proof Consider infinitesimal steps dx and dy in the x, y-plane, with corresponding steps du and dv in the
u, v-plane. We have then du = ux dx + uy dy and dv = vx dx + vy dy. Preserving angles locally means that all
sides of any infinitesimal triangle scale with the same factor h(x, y), i.e. du2 + dv 2 = h2 (dx2 + dy 2 ). With
the expressions above for du and dv, this becomes

(u2x + vx2 − h2 )dx2 + (u2y + vy2 − h2 )dy 2 + (ux uy + vx vy )dxdy = 0

For this to hold identically for any combination of dx and dy, all three coefficients need to vanish, i.e.

 u2x + vx2 = h2
u2y + vy2 = h2 .
u x u y + vx vy =0


We can satisfy the top equation with ux = h cos α, vx = h sin α and the second by uy = h cos β, vy = h sin β,
after which the third equation gives α − β = ± π2 , from which follows either {ux = vy , uy = −vx } or
{ux = −vy , uy = vx }. We recognize the first option as the Cauchy-Riemann equations for u + iv = f (x + iy),
implying that f (z) is analytic. The second option similarly corresponds to u + iv = f (x + iy). Although the
latter version also preserves scales locally, and sizes of angles, it reverses directions, and is hence not what
we accept as being conformal. 

7.8 Exercises
134 CHAPTER 7. CONFORMAL MAPPINGS
Chapter 8

Transforms

8.1 Fourier transform

For f (x) periodic on [−π, π], the Fourier series of f (x) can be defined as
∞
X ∞
X
f (x) = a0 + ak cos(kx) + + bk sin(kx)
k=1 k=1

where
ˆ π
1
a0 = f (x)dx
2π −π
ˆ π
1
ak = f (x) cos(kx)dx
2π −π
ˆ π
1
bk = f (x) sin(kx)dx, k = 1, 2, ...
2π −π

Thus f (x) can be seen as a superposition of Fourier modes and the coefficients tell us how much of each
mode is needed to recover the original function. This is another example of how algebra gets nicer if we
replace trigonometric functions with complex exponentials. In this case, the Fourier series takes the form
∞
X
f (x) = ck eikx
−∞

where ˆ π
1
ck = f (x)eikx dx
2π −π

If f is not periodic and if it is defined on (−∞, ∞), then we have a continuum of frequencies.
We defined the Fourier transform as follows:
ˆ ∞
1
f (x) = fˆ(k)eikx dk
2π −∞

where k is not an integer but varies continuously over (−∞, ∞) and where
ˆ ∞
ˆ
f (k) = f (x)e−ikx dx
−∞

Fourier series and transforms are important tools in applied mathematics and analysis. The integrals that
arise are often perfectly suited for contour integral treatment.

135
136 CHAPTER 8. TRANSFORMS

Example 8.1.1 Consider

f (x) = e−|x|
We find
ˆ ∞
fˆ(k) = e−|x| e−ikx dx
−∞
ˆ 0 ˆ ∞
= ex e−ikx dx + e−x e−ikx dx
−∞ 0
2
=
k2 + 1
Now let’s invert back
ˆ ∞
1 2
f (x) = eikx dk
2π −∞ k2 + 1

This is not an integrand we can find a primitive to. So we cannot solve the problem using our calculus 1, 2,
or 3 knowledge. It is easy to do however with contour integration. We consider 2 cases:
1. If x > 0, take the upper semi-circle as contour. By Jordan’s lemma, its contribution vanishes and
−x
1
f (x) = 2π 2πi e i = e−x
2. If x > 0, take the lower semi-circle as contour. By Jordan’s lemma, its contribution vanishes and
1 ex
f (x) = − 2π 2πi −i = ex

Therefore, f (x) = e−|x| .

Example 8.1.2 Consider

2
f (x) = e−x
We find
ˆ ∞
2
fˆ(k) = e−x e−ikx dx
−∞

There are no poles and the integrand takes large values up and down the complex plane.

The neat idea here is to complete the square in the exponent. We obtain
ˆ ∞
ik 2 k2
ˆ
f (k) = e−(z+ 2 ) − 4 dz
−∞
2
ˆ ∞
ik 2
= e − k4
e−(z+ 2 ) dz
−∞
´∞ 2 √
Recall from calculus 2 that −∞
e−z dz = π Thus
√ k2
fˆ(k) = πe− 4

8.2 Laplace transform

We will not motivate it here, but it is central in Electrical Engineering and other engineering fields. It is a
great way to solve linear ODEs on t ∈ [0, ∞)
Definition ˆ ∞
F̂ (s) = f (x)e−sx dx (8.1)
0
8.2. LAPLACE TRANSFORM 137

We will use the hat notation to refer to the transform and s as the independent variable to distinguish the
Laplace transform from the Fourier transform. Question: How do we get back from F̂ (s) to f (x)?
In classes where complex variables are not used, there are 2 approaches.

1. Recognize the transform in a large tabulated collection of known cases. In addition, people use some
extra techniques to modify formulas for F̂ (s) into combinations of other functions.

2. Use the horrendous general formula

   k+1
(−1)k dk

k k
f (x) = lim F̂
k→∞ k! dxk x x

With complex variables, we can replace the above formula with a delightfully simple formula.

The idea of this formula is to first extend f (x) to x < 0 by then setting f (x) = 0 for x < 0. Then
ˆ ∞
F̂ (s) = f (x)e−sx dx (8.2)
−∞

Let’s do a change of variables to get what will look like a Fourier transform, which we know how to invert.
Thus, let s = c + ik, where c and k are real. Note that c is important for the convergence of the integral. A
sufficiently large c will suffice. Thus
ˆ ∞
f (x)e−cx e−ikx dx


F̂ (s = c + ik) =
−∞
| {z }
Formula for the Fourier transform of f (x)e−cx

So,
ˆ ∞
1
f (x)e−cx = F̂ (c + ik)eikx dx
2π −∞

Let’s swap back variables to s = c + ik and ds = idk and multiply everything by ecx . We obtain
ˆ c+i∞
1
f (x) = F̂ (s)esx
2πi c−i∞

We will choose c so that all of the singularities of F̂ (s) are on the left of Re z = c. Once again, we find that
the resulting integrals are extremely well suited for contour integration tools.
Let’s do 2 examples

Example 8.2.1 Consider

f (x) = sin x

We find
ˆ ∞
F̂ (s) = sin xe−sx dx
0
ˆ ∞
= Im eix−sx dx
0
ix−sx ∞

e
= Im 
i − s 0
1
= 2
s +1
138 CHAPTER 8. TRANSFORMS

Let’s invert back

ˆ c+i∞ sx
1 e
f (x) = ds
2πi c−i∞ 1 + s2
ˆ
1 esx
= ds
2πi Γ 1 + s2
 
 sx 
1  X e
= Res , sj 
2πi j
1 + s2
= sin x

where the contour Γ is displayed in Figure 8.1

Im s

i
R

c Re s

Figure 8.1: Contour in Example 8.2.1.

We can indeed apply Jordan’s lemma to show that the contribution of the integral around the semi-circle
vanishes as its radius R → ∞.

Example 8.2.2 Now consider

log s
F̂ (s) =
1 + s2
We are asked to find f (x) back. Consider the keyhole contour illustrated in Figure 8.2. Listed below are the
contributions from the contour and from the residues
π ix
1. Res(f, s = i) = 4e
π −ix
2. Res(f, s = −i) = 4e
´ ´ 1
´
∞ log(r)+iπ −xr ´∞ log(r)−iπ −xr
 ´∞ e−xr
3. →
f (z)dz + ←
f (z)dz = 2πi 0 1+r 2 e dr − 0 1+r 2 e dr = 0 1+r 2 dr
8.3. HILBERT TRANSFORM 139

´
4. |  f (z)dz| ≤ M · L ≈ 2π log  → 0 as  → 0
´
5. R f (z)dz → 0 as R → ∞ by Jordan’s lemma
Therefore, adding all the contribution, we obtain
ˆ ∞
π e−xr
f (x) = cos x − dr
2 0 1 + r2
´∞ e−xr
where 0 1+r 2 dr cannot be expressed in terms of elementary functions in closed form.

Im s

i
R

c Re s

Figure 8.2: Contour in Example 8.2.2.

8.3 Hilbert transform

8.4 z -transform
8.5 Select proofs
8.6 Exercises
140 CHAPTER 8. TRANSFORMS
Chapter 9

Special functions defined by ODEs

One of the core resources of applied mathematics and theoretical physics is their collection of standard and
special functions. Many of these ‘special functions’ can be introduced as solutions to linear ODEs with
variable coefficients (as discussed below in Sections 9.1-9.3) or to nonlinear ODEs (notably the Painlevé
equations, discussed in Section 9.5). A typical form for a linear ODE in this context is

u00 (z) + p(z) u0 (z) + q(z) u(z) = 0. (9.1)

Already if p(z) and q(z) are linear functions in z, it is rare that solutions u(z) can be found in closed form
using only the standard functions from calculus. Since ODEs of this type arise in many applications (such
as from separation of variables, when studying second order PDEs), they have become a rich source of
important special functions. These special functions are generally analytic, as follows from a very general
theorem due to Cauchy:

Theorem 9.0.1 Let u(z) satisfy an ODE of the form

dn u du dn−1 u
n
= F (u, , . . . , n−1 ; z) (9.2)
dz dz dz
n−1
where F is locally analytic in each of its arguments. For any initial conditions on u, du d u
dz , . . . , dz n−1 at some
point z0 , the solution to (9.2) then becomes an analytic functions in some neighborhood of z0 .

The proof for this result can be found in standard ODE text books.

9.1 Airy’s equation

9.1.1 Real-valued independent variable
The constant√ coefficient √ODE y 00 − αy = 0 has for α > 0 exponentially growing
√ or decaying√ solutions
y(x) = C1 e+ α x + C2 e− α x and for α < 0 oscillatory solutions y(x) = C1 cos( −α x) + C2 sin( −α x). If
we replace the constant α with the independent variable x, we arrive at Airy’s equation

y 00 (x) − x y(x) = 0. (9.3)

Unsurprisingly, its solutions are are of exponential-like character for x positive, and oscillatory for x negative.
This equation is of great importance in many applications, since it forms the simplest description of (and
leading order approximation to) many phenomena that change character between being oscillatory or not.
Since it is a second order linear homogeneous ODE, its general solution can be written as a linear combination
of two independent ones, traditionally called Ai(x) and Bi(x), respectively, looking as shown in Figure 9.1.
In order to understand this equation better, questions that need to be answered include:

1. What is the unique ratio of initial conditions u0 (0)/u(0) that gives solutions that decay for x → +∞?

141
142 CHAPTER 9. SPECIAL FUNCTIONS DEFINED BY ODES

Ai(x)
1 Bi(x)

0.5

-0.5
-10 -5 0 5
x

Figure 9.1: The Ai and Bi functions along the real axis.

2. Just how fast does the Ai-function grow and the Bi-function decay for x → +∞ (presumably at some
‘super-exponential’ rates)?
3. Can we approximate the envelope and oscillation rates (including the phase angle) for the oscillations
as x → −∞?

An important first step is to express the Ai(x) and Bi(x) functions as integrals. For the bounded Ai-function,
this can be done by a Fourier transform (FT). Recalling the rules for the FT given in Section ??, the transform
dŷ 3
of (9.3) becomes (iω)2 ŷ(ω) − i dω = 0, which can be solved in closed form: ŷ(ω) = C ei ω /3 . Inversion back
´ ∞ i ω3 /3 i ωx
to physical space gives y(x) = C −∞ e e dω, which can be rearranged into
ˆ ∞  3 
ω
y(x) = 2C cos + ωx dω. (9.4)
0 3
However, this integral is quite impractical as it stands, being rapidly oscillatory and not absolutely conver-
gent. It also only represents the bounded Ai-function and not the unbounded Bi-function. As in so many
other cases, the simplest path towards progress turns out to go via the complex plane.

9.1.2 Complex-valued independent variable

The complex counterpart to (9.3) becomes

u00 (z) − z u(z) = 0. (9.5)

If we are given initial conditions for u(0) and u0 (0), the method of unknown coefficients (Section ??) will
recursively provide any number of coefficients in the everywhere convergent Taylor expansion of u(z). How-
ever, Taylor expansions can rarely give any insights about functional behaviors at large distances. Figures
9.2 and 9.4 show that the two independent solutions Ai(z) and Bi(z) then take on very distinctive features.
Techniques to understand these will be described in Section 9.4 and Chapter 10. The first of these sections
describes a more general approach to obtain integral relations than the FT based one we used above, and
Chapter 10 discusses asymptotic analysis for exploring far-field behaviors. For both of these tasks, the exten-
sion to complex variables is essential. Following these discussions, we will then return to the Airy functions
as an example of applying these methods.

9.2 Bessel’s equation

9.3 Hypergeometric functions

9.3. HYPERGEOMETRIC FUNCTIONS 143

Figure 9.2: Real and imaginary parts of Ai(z).

Figure 9.3: Real and imaginary parts of Bi(z).

144 CHAPTER 9. SPECIAL FUNCTIONS DEFINED BY ODES

(a) Abs (Ai(z)) (b) Abs (Bi(z))

Figure 9.4: The magnitude of the Ai(z) and Bi(z) functions.

9.4 Converting linear ODEs to integrals

9.4.1 Fourier-Laplace method
Given a variable coefficient linear ODE, the general idea is to look for solutions of the form
ˆ
u(z) = F (t) ez t dt, (9.6)
C

where both the integration contour C and the function F (t) remain to be determined. A couple of specific
examples explains the typical steps in the procedure.

9.4.1.1 Airy’s equation

´
Substituting (9.6) into the Airy equation (9.5) gives C (t2 − z)F (t) ez t dt = 0. We next want to manipulate
d
this expression so that z will appear in the exponent only. Noting that z ez t = dt (ez t ) and then carrying
out integrating by parts gives
ˆ  
dF
+ t2 F ez t dt − F (t) ez t C = 0.
 
dt
C

This will be true for all z if it holds that (i) dF 2

dt + t F = 0 and (ii) F (t) e
zt
takes the same value at both
3
ends of the contour C. Starting with item (i): the general solution of dt + t2 F = 0 is F (t) = c e−t /3
dF

(where c is some constant which can be omitted). Regarding item (ii) we thus need to study the expression
3
G(t) = F (t) ez t = e−t /3+z t in a complex t-plane (temporarily viewing z as a complex constant). G(t) is
clearly an entire function of t, which far out in the t-plane rapidly decreases to zero within the three sectors
(valleys) (V1 ) − π6 < arg z < π6 , (V2 ) π2 < arg z < 5π 7π 3π
6 , (V3 ) 6 < arg z < 2 , and with high ‘mountain ridges’
zt
in between. Letting C form a closed loop will make [F (t) e ]C vanish, but this is useless as u(z) (as defined
in ((9.6)) will then also vanish, resulting only in the trivial u(z) = 0 solution. A much better idea in this case
is to run the contour C from one valley to another. Figure 9.5 illustrates the magnitude of the integrand,
3
|e−t /3+z t |, in the case of z = 1 (with this specific z-value only causing a slight deformation of the surface
near the origin in the complex t-plane). Denoting by C1 , C2 , C3 contours from valley V2 to V3 , V2 to V1 ,
and V1 to V3 , respectively, we obtain the two standard Airy functions as
ˆ ˆ ˆ
 
1 3 1  3
Ai(z) = e−t /3+z t dt, Bi(z) = − −  e−t /3+z t dt. (9.7)
2πi 2π
C1 C2 C3
9.5. THE PAINLEVÉ EQUATIONS 145

5
V2 V1
4

-1
V3
-2

y
-3 C1 2
3
4
5

1
-4 0
-1
-2
-5 -3
-4 x
-5

3
Figure 9.5: The function |e−t /3+z t |, displayed in a complex t-plane in the case of z = 1. As usual, the real
axis is marked by a thick red line. The green curve shows the lowest possible choice for the contour C1 .
While any contour from V2 to V3 would give the correct value for Ai(z), this specific type of contour will
play a key role in the analysis in Section 10.3.

The details in these definitions are motivated

´∞ by (i) ensuring that the two solutions are independent and
real along the real axis, (ii) making −∞ Ai(x)dx = 1, and (iii) making the two functions share the same
envelope for their oscillations as x → −∞ (further analyzed in Section ??). We can also note that deforming
the contour C1 to run along the imaginary axis (making its convergence properties much worse) produces
(9.4).

9.4.1.2 Bessel’s equation

9.4.2 Euler’s method

Several options are available for choosing the integrand in (9.6). In Euler’s version, we start instead with
ˆ
u(z) = F (t) (t − z)µ dt (9.8)
C

and proceed then in the same way - substitute into the ODE and choose F (t), C, and now also the parameter
µ so that the ODE becomes satisfied for all values of z.

9.4.2.1 Legendre’s equation

9.5 The Painlevé equations

With often centuries of history behind them and, more recently, with the immense power provided by
modern computers, most of the special functions of applied mathematics have by now become thoroughly
explored and understood. The Painlevé equations and their solutions, known as the Painlevé transcendents,
form an exception in this regard. They were first considered just over 100 years ago, initially from a very
theoretical perspective. In spite of proving to be particularly difficult to analyze theoretically and also
unusually challenging to compute numerically (in both cases partly due to their commonly occurring vast
pole fields in the complex plane), their range of applications kept expanding, to the point that they now are
146 CHAPTER 9. SPECIAL FUNCTIONS DEFINED BY ODES

considered among the most important types of special functions. Recent numerical advances have finally
made them much more accessible than previously.
Based on a level of intuition that is hard to imagine, Paul Painlevé (later Prime Minister of France)
2
and some collaborators of his decided to consider all ODEs of the form ddzu2 = F z, u, du


dz where F is an
arbitrary rational function of its three arguments z, u, and du
dz (algebraic function such that both numerator
and denominator are polynomials). Typically, solutions to nonlinear ODEs of such a general form would
be expected to feature movable branch points (i.e. branch points at locations that depend on the ODE’s
two initial conditions). Remarkably, they found that exactly 50 genuinely different equations of this form
were possible if one imposes the Painlevé Property: “The only movable singularities in the complex plane
are either poles or essential singularities” (i.e. no movable singularities are allowed that give rise to multi-
valuedness). It further transpired that essential singularities could not arise in this context, so all movable
singularities to these equations are in fact poles. Out of these 50 equations, it was also discovered that
the solution to 44 of them could be expressed in terms of elementary or traditional special functions. The
remaining six equations are now known as the Painlevé equations:
d2 u
PI dz 2 = 6u2 + z
d2 u
PII dz 2 = 2u3 + zu + α
d2 u

2 1 du αu2 +β
PIII dz 2 = u1 du dz
− z dz + z + γu3 + uδ
d2 u 1 du 2 β
PIV dz 2 = 2u dz +32 u3 + 4zu2 + 2(z 2 − α)u +u 
2 2
d2 u (u−1)
1 1 du
− 1 du + αu + βu + γu + δu(u+1)


PV dz 2 = 2u + u−1 dz 2
  z dz
2 z  z u−1  
d2 u u(u−1)(u−z) βz γ(z−1) δz(z−1)
PV I dz 2 = 12 u1 + u−1 1
+ u−z 1 du
dz − 1
z + 1
z−1 + 1
u−z
du
dz + 2
z (z−1)2 α+ u + (u−1)2 + (u−z)2

Apart from two initial conditions (ICs), the equations feature different numbers of additional free pa-
rameters, ranging from none for PI to four for PIII , PIV and PV I . While most of the analytic functions we
have considered so far have not had any free parameters (beyond the independent variable z, for ex. sin z
or Γ(z)), some have involved one or more parameters (for ex. the two periods ω1 and ω2 for the Weierstrass
℘(z)-function). Here (including the two ICs), we have from 2 to 6 parameters. This obviously adds greatly
to the challenge of exploring their full solution spaces.
We will below limit ourselves to showing a few highlights of the analytic functions that arise as solutions
to the PI , PII and PIII equations.

9.5.1 The PI equation

We noted in Chapter 6 (equation (6.8)) that the Weierstrass ℘-function satisfies the nonlinear ODE
1
℘00 (z) = 6℘(z)2 − g2 , (9.9)
2
where the parameter g2 could be chosen arbitrarily. For all ICs, solutions to (9.9) are doubly periodic
functions, with double poles repeating completely regularly across the full complex z-plane. The difference
to the PI -equation
u00 (z) = 6u(z)2 + z (9.10)
is that, in the last term, the constant − 21 g2 has been replaced by the independent variable z. By local Laurent
expansions, one sees that all singularities to (9.10) are again double poles of strength one and residue zero
(cf. the chapter exercises). Far away from the origin, z becomes locally like a large constant, and the pole
fields become locally similar to those of (9.9). However, as z grows, they generally get denser further out. A
main difference however is caused by the fact that the variable changes u → ζ 3 u, z → ζz with ζ 5 = 1 leaves
(9.10) again satisfied, causing pole fields far out to differ between five sectors with boundaries between them
following the directions arg z = 1π/5, 3π/5, 5π/5, 7π/5, 9π/5. These sectors are visible in the PI solution
pictures in Figure 9.6, which show the pole locations in four illustrative cases.
The NIST Handbook (2010) [4] was written before Painlevé solutions could readily be computed across
the complex plane [??]. It contains in its Chapter 32, on the Painlevé equations, an illustration similar to
Figure 9.7, showing along the real axis how the solution changes when u(0) = 0 and u0 (0) is varied in close
vicinity of a critical value for which the solution is smooth and non-oscillatory along the negative real axis.
9.5. THE PAINLEVÉ EQUATIONS 147

Figure 9.6: Four pole field plots for the PI equation, displayed over the region[−50, 50] × [−50, 50] in the
complex plane. (a) Case with the ICs u(0) = u0 (0) = 0; the sector boundaries are here narrow pole free
strips, (b) ICs u(0) = u0 (0) = 2; the pole field structures are far out similar to the previous case, but
the transitions between the sectors have become gradual instead of sharply defined, (c) The tritronqueé
solution; the unique case for which four of the five sectors have become completely pole free; its ICs are
u(0) ≈ −0.1875543083404949, u0 (0) ≈ 0.3049055602612289 and (d) A solution with ICs very close to the
tritronqueé case (here u(0) = −0.1875, u0 (0) = 0.3049). Little has changed in the right pole field, but a pair
of additional pole fields have very rapidly moved in from above left and below left, respectively.
148 CHAPTER 9. SPECIAL FUNCTIONS DEFINED BY ODES

Figure 9.7: Illustration along the real axis of the PI solutions with u(0) = 0 and with u0 (0) varied as shown
in the figure.

Corresponding illustrations over the complex plane are far more revealing, cf. Figures 9.8 (a,b) (with the
matching real axis intervals marked by thick black curves). For u0 (0) = 1.8518, the oscillations for x negative
are seen to be caused by a pole field that symmetrically surrounds the negative real axis, but without any
poles on the axis itself. As u0 (0) increases slightly, this left pole field recedes rapidly out to minus infinity,
and then moves back in again, however shifted vertically so there now have become an infinity of poles on
the negative real axis. The right pole field has remained essentially stationary.
Theoretical work on the PI equation have to a large extent been limited to asymptotic analysis of solutions
far out in the different sectors, and on connection formulas linking these results between sectors. For solution
features at finite distances, computational methods are nowadays the dominant source of information.

9.5.2 The PII equation

In contrast to the PI equation, the remaining Painlevé equations have a few instances of closed form solutions
(still like drops in a bucket only, compared to their full solution spaces). Key to obtaining these are Bäcklund
transformations - relating solutions between different parameter choices. In the case of the PII equation,
the key such relation is
2α + 1
u(α + 1; z) = −u(α; z) − 0 . (9.11)
2u (α; z) + 2u(α; z)2 + z
4−2z 3
Starting from the trivial α = 0 solution u(0; z) = 0, one thus finds u(1; z) = − z1 , u(2; z) = 4z+z 4 , u(3; z) =
3z 2 (160+8z 3 +z 6 )
320−24z 6 z 9 ,
etc. For each integer α, these form a single closed form solution in the two-dimensional
solution space for that particular α-value.
Another sequence of solutions is known for α half-integer ( 21 , 32 , 52 , . . .). With φ(z) = cos θ2 Ai 2−z



1/3 +
sin θ2 Bi 2−z 0 1



1/3 and Φ(z) = φ (z)/φ(z), one obtains the one-parameter (0 ≤ θ ≤ 2π) solutions: u( 2 ; z) = −Φ,
3 4 3 2 2 3
+6Φ +4z Φ +3zΦ+z −1
u( 23 ; z) = 2Φ2Φ+zΦ−1
2 +z , u( 25 , z) = 4zΦ (4Φ 3 +2zΦ−1)(2Φ2 +z) , etc. Letting the free parameter θ move through
its period [0, 2π] causes poles to move as seen for the case of α = 52 in Figure 9.9. When θ = 0, all poles
are located in a band surrounding the positive real axis. As θ increases, a curved band of poles enters from
the left, becomes part of two different entirely symmetric solution configurations at θ = π3 and θ = 5π 3 , and
finally moves back out left towards minus infinity as θ approaches 2π, leaving for θ = 2π the same pole
distribution as for θ = 0. However, if one looks carefully at the pole field along the positive real axis, one
can see that this has moved steadily to the left during this process. The band of poles that came in from the
left picked, during the interactions, up five poles from the field along the positive real axis, included then in
the curved band, and then brought theses out to minus infinity.
9.5. THE PAINLEVÉ EQUATIONS 149

Figure 9.8: The magnitude of same solutions as the bottom and top cases in Figure 9.7, here displayed over the
complex plane rather than only along the real axis: (a) u(0) = 0, u0 (0) = 1.8518, (b) u(0) = 0, u0 (0) = 1.8519.
The crossing of the thick black lines marks the origin z = 0.
150 CHAPTER 9. SPECIAL FUNCTIONS DEFINED BY ODES

Figure 9.9: The Airy family of PII solutions in the case of α = 52 . All poles to PII are simple, with residues
+1 or -1. In this and the next two figures, these are colored blue and yellow, respectively.

No other closed form solutions are known to the PII equation than the two classes just noted above, for
α integer and for α half-integer, respectively.
We conclude our PII equation summary by showing two other types of solutions to the PII equation
that are particularly important in applications. Hastings-McLeod solutions are smooth and non-oscillatory
along the entire real axis, cf. Figure 9.10. There are two different such solutions for each α in the range
|α| < 12 , otherwise only one. For each α-value in the range |α| < 12 there also exists a one-parameter family
of Ablowitz-Segur solutions. These are smooth and bounded along the entire real axis, oscillatory to the left
and decaying to the right. Figure 9.11 shows a typical such solution in the case of α = 0.

9.5.3 The PIII equation

In contrast to the PI , PII and PIV equations, solutions to the PIII , PIV and PV I equations need not be
single-valued. Focusing on PIII , this is a consequence of the term − z1 du dz in its RHS. Had this been the only
term in the RHS, u(z) = ln z would have been a solution. This multi-valuedness does not contradict the
Painlevé property, since the location of the branch point is fixed (at z = 0), and is not ‘movable’ with the
choice of ICs. Analysis will reveal that solutions to PIII , for different ICs and parameter values (for α, β, γ, δ)
will always have either 1, 3, or infinitely many Riemann sheets. It further transpires that one can make
variable changes to reduce the effective number of parameters (including the ICs) from 6 to 4. Nevertheless,
its solution space is so rich of features that we cannot here attempt any sort of survey of it, and we will limit
ourselves to the single illustration in Figure 9.12.
We have noted several times earlier that what can be seen about an analytic function along the real
axis contains only a small fraction of the information that can be gained by inspecting it (and utilizing its
9.5. THE PAINLEVÉ EQUATIONS 151

Figure 9.10: Examples of Hastings-McLeod solutions, shown along the real axis, and their associated pole
fields.

Figure 9.11: Example of an Ablowitz-Segur solution in the case of α = 0.

152 CHAPTER 9. SPECIAL FUNCTIONS DEFINED BY ODES

Figure 9.12: Phase plane portrait of a PIII solution that is pole free in the right half-plane of its primary
Riemann sheet. The verical axis is here used only to distinguish the different sheets from each other.

properties) over the complex plane. In the case of PIII , it turns out that intriguing phenomena can occur
on different Riemann sheets even in cases when little noteworthy seems to occur on the primary sheet.

9.6 Select proofs

9.7 Exercises
Chapter 10

Asymptotic Analysis of Integrals

Asymptotic analysis is a far reaching and important topic, which we will give only a very brief introduction
to. We will limit ourselves to the case of approximating integrals when some parameter goes large, causing
the integrand to either become locally spiked, or increasingly oscillatory - cases where analytic functions
play a crucial role. More generally, asymptotic analysis, be it for approximating solutions to ODEs, PDEs
(such as boundary layers in fluids), integrals, etc., becomes most effective just in those situations where
typical numerical methods encounter their greatest difficulties (and vice versa). Several very good books on
asymptotic methods are available, such as [??,??,??,??].
The integrals we will consider below may have arisen directly from applications, or via conversion from
ODEs, as discussed in Section 9.4. We start our discussion by some brief comments on what is meant by an
asymptotic (in contrast to a convergent) expansion. After that, in Section 10.2.1, we focus on a purely real-
valued case in which an integrand becomes increasingly spiked. Utilizing the freedom of choosing integration
paths in the complex plane for analytic functions, the method of steepest descent (Section 10.3) turns the
real valued result into a tool of far greater generality.

10.1 Asymptotic vs. convergent expansions

P∞ n PN o
An infinite sum S = k=0 ak converges if S − k=0 ak → 0 as N → ∞. This readily generalizes to the
case when the terms are functions of some parameter x. The sum
∞
X
S(x) = ak (x) (10.1)
k=0
n PN o
converges for a fixed x if S(x) − k=0 ak (x) → 0 as N → ∞. Given a value for x, we can reach any
accuracy we want by just letting N → ∞. For an asymptotic series, the roles of x and N can be seen as
reversed. In place of (10.1), we write
∞
X
S(x) ∼ ak (x) (10.2)
k=0
if it holds, for any fixed N , that
PN
S(x) − k=0 ak (x)
→0 (10.3)
aN (x)
when x goes to some limit (such as x → ∞). Both (10.1) and (10.2) are very useful for computation.
The utility of (10.1) is obvious, especially if the terms ak (x) are easy to compute, and S(x) is a lot more
complicated (with Taylor expansions a typical example). The utility of (10.2) is somewhat more subtle. The
example below is typical for this case:
Example 10.1.1
´x Compare the accuracies of the Taylor and the asymptotic expansions for the error function
2
erf(x) = √2π 0 e−t dt.

153
154 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS

1.5

0.5

erf(x) 0

-0.5

-1

-1.5
-5 -4 -3 -2 -1 0 1 2 3 4 5
x

Figure 10.1: The function erf(x) displayed along the real x-axis.
50 50

20

0
10

-5
-10
-15
-20

-30
10

-60
5
45 45
-5
-20

40 40
-15

20

35 35
-10

5
-30

30 30

20

0
10

-5
10

-10

-15

-20

-30
-5

5
N

-60
25 25
-20

20
0
-10 -15

20 20

15 15
5

10
10 10
0

-5
-5

-10

-3
-2
-15

0
0
5 5 5
5
0

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
x x

(a) Taylor expansion (b) Asymptotic expansion

Figure 10.2: log10 of the errors (shown numerically along the contour lines) when erf(x) is approximated at
location x using N terms in the Taylor expansion (10.4) and the asymptotic expansion (10.5), respectively.
The areas shaded gray indicate where the accuracy is better than 10−15 .

The error function transitions smoothly from -1 to +1, as seen in Figure 10.1. Since this is an entire function,
its Taylor expansion
x3 x5 x7
 
2
erf(x) = √ x− + − + −... (10.4)
π 3 · 1! 5 · 2! 7 · 3!
is everywhere convergent. Repeated integration by parts of the defining integral produces an asymptotic
expansion
2 
e−x

1 1·3 1·3·5
erf(x) = 1 − √ 1− + − + − . . . (10.5)
x π (2x2 )1 (2x2 )2 (2x2 )3
which, in contrast, never converges (since the factorial-type growth in the numerators always ‘wins out’
over the geometric rate in the denominators). Nevertheless, this expansion is the superior one for numerical
computing as long as x & 5, as illustrated in Figure 10.2 (b). For x-values increasing further still, the figure
shows that the accuracy increases extremely rapidly, even if the number of terms N is kept quite low. The
fact that (10.5) always diverges if we keep x fixed and let N → ∞ is irrelevant if our goal is to estimate
erf(x) for x large.

While the region of convergence for a Taylor expansion always is the largest singularity-free circle sur-
rounding the expansion point, applicability of asymptotic expansions (of the form of inverse powers, such
10.2. LAPLACE INTEGRALS 155

+3π/4 +π/4

-3π/4 -π/4

(a) Sector of validity for (10.5) (b) Sector of validity for (10.6)

Figure 10.3: The sectors of validity, shaded gray, for the two expansions (10.5) and (10.6), respectively.

as (10.5)) occurs in sectors of the complex plane. Generalizing x to be a complex variable, (10.5) is valid
within the sector shaded gray in Figure 10.3 (a), whereas a version with the leading “1” swapped for “-1”
2
e−x
 
1 1·3 1·3·5
erf(x) = −1 − √ 1− + − + −... (10.6)
x π (2x2 )1 (2x2 )2 (2x2 )3

is valid in the sector shown in part (b) of the figure. The expression Stokes Phenomenon refers both to sector
structures of this type, and to the fact that sectors of validity can overlap. In the overlapping regions, the
2
factor e−x grows massively large as x moves away from the origin, making the leading constant irrelevant.
A more extensive example of Stokes sectors will be given in Section 10.3.3, when we use the method of
steepest descent to analyze the far field behavior of solutions to Airy’s equation (illustrated earlier in Figures
9.2-9.4).

10.2 Laplace integrals

We start by considering real-valued integrals along the real x-axis, of the Laplace integral form

ˆb
I(s) = f (x) es φ(x) dx, (10.7)
a

in the limit of the parameter s → ∞. For the three Examples 10.2.1-10.2.3 considered later in this section,
the top row of subplots in Figure 10.4 illustrates the φ(x) functions, and the bottom row shows the inte-
grands f (x) es (φ(x)−φ(c)) for the s-values s = 1, 3, 10, 30. For improved visibility, we have for the integrands
included a (x-independent) factor e−s φ(c) , where x = c is the location of the maximum for φ(x). In all cases,
the asymptotic expansions for s → ∞ will come from the immediate vicinity of these x = c locations. In the
third example, we split the interval into two, [0, 1] and [1, ∞], and for the first one change variable x → −x.
Hence we can, without loss of generality, proceed by assuming that the maximum value of φ(x) over the
integration interval [a, b] occurs at its left boundary, denoted by x = a. We next distinguish two main cases:

Case 1: φ0 (a) < 0, and

Case 2: φ0 (a) = 0 and φ00 (a) < 0.

Case 2 can be generalized to having several leading derivatives zero at x = a. The next key step in ob-
taining the asymptotic expansion for (10.7) will be to make a change of variable in such a way that we then
can apply Watson’s lemma:
156 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS

Z π/2 Z 2 Z
−s sin2 x 1 −s arctan x ∞
1 −s(−x+log x)
e dx e dx e dx
0 1 1+x 1 x
0 -0.6 0

-0.8 -1
φ(x)

-0.5
-1 -2

-1 -1.2 -3
0 0.5 1 1.5 1 1.5 2 0 1 2 3

1 0.6 3
f(x)es(φ(x)−φ(c))

s=1
s=3 0.4 2
0.5 s = 10
s = 30 0.2 1

0 0 0
0 0.5 1 1.5 1 1.5 2 0 1 2 3
x x x

Figure 10.4: The three columns of subplots correspond to the three Examples 10.2.1-10.2.3, respectively,
that are discussed in Section 10.2.2. The top row of subplots show the functions φ(x) that are multiplied
by s in the exponents, and the bottom row of subplots shows how the integrands f (x) es φ(x) then vary
across the respective intervals (scaled by the factor e−s φ(c) , in order to display an s-independent value at
the location x = c where φ(x) has its maximum). In all cases, it is visually apparent that, as s → ∞, the
integral will become increasingly well described by its properties in the immediate vicinity of x = c.
10.2. LAPLACE INTEGRALS 157

10.2.1 Watson’s lemma

Theorem 10.2.1 Watson’s lemma: If

ˆb
I(s) = f (ξ) e−s ξ dξ , (b > 0) (10.8)
0

and
∞
X
f (ξ) = ξ α an ξ β n , (ξ & 0) (10.9)
n=0

then
∞
X an Γ(α + βn + 1)
I(s) ∼ , (s → ∞) (10.10)
n=0
sα+βn+1

Proof (Sketch): The two key observations behind demonstrating this result are: (i) In I(s), as defined by
(10.8), the actual value of the upper limit b is irrelevant, allowing us to replace it by b = ∞, and (ii) when
then substituting (10.9) into (10.8), all integrals that arise can be evaluated in closed form by means of
´∞
the Γ-function relation 0 ξ z e−s ξ dξ = Γ(z+1)
sz+1 (Re z > −1, Re s > 0; In the special case of n integer, this
´ ∞ n −s ξ n!
becomes 0 ξ e dξ = sn+1 ) 

While some other approaches (such as integration by parts) might work in certain cases, Watson’s lemma
is particularly powerful in that it works well also when (10.9) involves fractional powers of ξ (as will happen
in Case 2).
To find the appropriate change of variables so that Watson’s lemma can be applied to (10.7), we first
rewrite this integral (arranged so that maxt∈[a,b] φ(x) occurs at x = a) as

ˆb
s φ(a)
I(s) = e f (x) es (φ(x)−φ(a)) dx, (10.11)
a

and then change variable

−ξ(x) = φ(x) − φ(a) (10.12)
so that the exponent in the integral becomes −s ξ (as in (10.8)). The integral now takes the form

ˆ0+ 
f (x)
I(s) = e s φ(a)
− 0 e−s ξ dξ, (10.13)
φ (x)
0

where the upper limit n

of the integral
o (denoted 0+) will become irrelevant as s → ∞. The key step that now
f (x)
remains is to express − φ0 (x) as a power series expansion in ξ, and Watson’s lemma will then give the
complete asymptotic expansion.

10.2.2 Applications of Watson’s lemma

´ π/2 2
Example 10.2.1 Find the complete asymptotic expansion for I(s) = 0
e−s sin x
dx as s → +∞ (cf. the
first column of subplots in Figure 10.4).

n o
The integral is of the form (10.7) with φ(x) = − sin2 x and f (x) = 1, i.e. − φf0(x) 1
(x) = 2 sin x cos x . The
function φ(x) has its maximum over [0, π2 ] at x = 0, so the variable change
n (10.12)
o −ξ(x) = φ(x) − φ(0)
2 1/2 f (x)
gives ξ = sin x, and therefore sin x = ξ , from which follows that − φ0 (x) , expressed in ξ, becomes
158 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS
n o ´1
− φf0(x)
(x) = √ 1
. ξ −1/2 (1 − ξ)−1/2 e−s ξ dξ, which is in the perfect
The integral has become I(s) = 1
2 0
2 ξ(1−ξ)
P∞ Γ(n+ 1 ) ξn
form for Watson’s lemma after using the binomial theorem expansion (1 − ξ)−1/2 = n=0 n! Γ(2 1 ) . There-
2
P∞ [Γ(n+ 1 )]2
fore, I(s) ∼ 21 n=0 n! Γ( 1 ) s2n+1/2 for s → +∞. 
2
n o
In most cases, one cannot write down a closed form expression for − φf0(x)
(x) in terms of ξ(x). One can
then instead proceed by using Taylor expansions, starting from (10.12) and then expanding φ(x) around the
0 00
critical point x = a (explicitly as d1 = φ 1!(a) , d2 = φ 2!(a) , etc. or by a method from Section ??):

−ξ(x) = φ(x) − φ(a) = d1 (x − a) + d2 (x − a)2 + . . . (10.14)

Case 1: φ0 (a) 6= 0. We can invert (10.14) to obtain a new set of coefficients {ai }, i = 1, 2, . . . so that

(x − a) = a1 ξ + a2 ξ 2 + a3 ξ 3 + . . . , i.e. x = a + a1 ξ + a2 ξ 2 + a3 ξ 3 + . . . , (10.15)

as discussed
n in Section ?? (i.e. by substituting (10.15) into (10.14) and equating coefficients). Substituting
o
f (x)
into − φ0 (x) gives a Taylor expansion in ξ around ξ = 0:
 
f (x)
− 0 = b0 + b1 ξ + b2 ξ 2 + . . . , (10.16)
φ (x)

and Watson’s Lemma can be applied.

Case 2: φ0 (a) = 0; The first term in the RHS of (10.14) is now missing. The series can again be in-
verted, but the form of (10.15) will have to be changed to also include half-integer powers:

(x − a) = a1 ξ 1/2 + a2 ξ 2/2 + a3 ξ 3/2 + . . . . (10.17)

The coefficients {ai } can again be obtained by substituting the expansion into (10.14) and then equating
coefficients. The counterpart to (10.16) takes now the form
 
f (x)
− 0 = b−1 ξ −1/2 + b0 + b1 ξ 1/2 + b2 ξ 2/2 + b3 ξ 3/2 + . . . ,
φ (x)

which again works just fine for Watson’s lemma. If this case arose from a situation with a max point of
φ(x) occurring at an interior location and the interval was split in two sub-intervals, the signs for the even
numbered coefficients b0 , b2 , b4 , . . . get reversed when the other sub-interval (with the critical point at its
right edge) is analyzed. Then, we simply ignore these coefficients, and double the contributions from the
odd numbered ones b−1 , b1 , b3 , . . . We next give one example for each of these Cases 1 and 2, respectively.
´2
Example 10.2.2 Find some leading terms for I(s) = 1
1 1+x
e−s arctan x dx (cf. the second column of subplots
in Figure 10.4).

1
This example features f (x) = 1+x , and φ(x) = − arctan x. The variable change (10.12) gives ξ = arctan x− π4
and
1 + x2
 
f (x) 1 7 10 62
− 0 = = = 1 + ξ + 2ξ 2 + ξ 3 + ξ 4 + ξ 5 − . . . (10.18)
φ (x) 1+x cos ξ (sin ξ − cos ξ) 3 3 15

With φ(1) = − π4 giving a leading factor e−s π/4 , Watson’s lemma produces
 
1 1 4 14 80 496
I(s) = e−s π/4 + 2 + 3 + 4 + 5 + 6 + ... .
s s s s s s
10.2. LAPLACE INTEGRALS 159
n o
In most cases, we would not be able to express − φf0(x)
(x) explicitly like this as a function of ξ. We would
then instead do Taylor series manipulations as follows (most conveniently and reliably by using a symbolic
algebra package such as Mathematica, which has each of the steps as a built-in command). First expanding
around x = a = 1 gives
π 1 1 1 1
ξ = arctan x − = (x − 1) + (x − 1)2 − (x − 1)3 + 0 (x − 1)4 − (x − 1)5 + . . . ,
4 2 4 12 40
which can be inverted
8 10 64 8 10 64
(x − 1) = 2ξ + 2ξ 2 + ξ 3 + ξ 4 + ξ 5 + . . . , i.e. x = 1 + 2ξ + 2ξ 2 + ξ 3 + ξ 4 + ξ 5 + . . . ,
3 3 15 3 3 15
n o 2
Substituted into − φf0(x)
(x) = 1+x
1+x , this gives the same expansion in ξ as obtained above in (10.18). 

´∞ 1 s (−x+log x)
Example 10.2.3 Find some leading terms for I(s) = 0 xe dx (cf. the third column of subplots
in Figure 10.4).

The function φ(x) = −x + log x has its maximum at x = 1, with φ(1) = −1. Hence, we consider first
the integral over [1, ∞], and make the change of variable −ξ(x) = φ(x) − φ(1) = −x + 1 + log x, with Taylor
expansion
1 1 1 1 1 1 1 1 1
ξ= (x−1)2 − (x−1)2 + (x−1)4 − (x−1)5 + (x−1)6 − (x−1)7 + (x−1)8 − (x−1)9 + (x−1)10 +. . .
2 3 4 5 6 7 8 9 10
giving when inverted
√ 2ξ ξ 3/2 2ξ 2 ξ 5/2 4ξ 3 139ξ 7/2 2ξ 4 571ξ 9/2
x=1+ 2ξ 1/2 + + √ − + √ + − √ + − √ + ...
3 9 2 135 540 2 8505 340200 2 25515 73483200 2
(where we havenincluded o unusually many terms, since half of them soon will drop out). Substituting this
f (x) −1
expansion into − φ0 (x) = 1−x gives

ξ −1/2 ξ 1/2 ξ 3/2 4ξ 2 139ξ 5/2 8ξ 3 571ξ 7/9

 
f (x) 1 4ξ
− = √ − + √ − + √ + − √ + − √ + ...
φ0 (x) 2 3 6 2 135 216 2 2835 97200 2 25515 16329600 2

As described under “Case 2” above, the expansion for the sub-interval x ∈ [0, 1] will become identical apart
from a switch of sign for all terms with integer powers of ξ. When adding the results from the two sub-
intervals, all these terms vanish, and we double the coefficients for the remaining ones (with half-integer
powers of ξ). Recalling the leading factor esφ(a) in (10.11), Watson’s lemma gives
r  
−s 2π 1 1 139 571
I(s) ∼ e 1+ + − − + ... .  (10.19)
s 12 s 288 s2 51840 s3 2488320 s4

As a follow-up to this´∞ example, we can note that the change of variables s x = u in the original integral
for I(s) gives I(s) = s1s 0 us−1 e−u du = s1s Γ(s). We thus obtain the famous Stirling’s formula for Γ(z) by
multiplying (10.19) by ss . Since s! can be defined as Γ(s + 1) also for non-integer s-values, it follows (having
re-named s to n) that
 n n √ 
1 1 139 571

n! ∼ 2πn 1 + + − − + . . . . (10.20)
e 12 n 288 n2 51840 n3 2488320 n4

This expansion (10.20) can be re-cast in different ways, with

1 1 1 1 1
log n! ∼ n log n − n + log(2πn) + − 3
+ 5
− + ... (10.21)
2 12 n 360 n 1260 n 1680 n7
160 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS

50 50
-4

0
15

20
0

15
-10

-5
10

-45
-25

10

-35
-25
-15
-15

5
-30
5

-50
45 45

-3
-5

-40
-30
-20

-20
-10
5
40 40

35 35

-30
30 -2 30 -45

0
5

-5
0

-4

-15
-10

-25

-3
-2 0
-15

-10
N

5
25 0 25

-20
-5

-30
-25
20 20
-20 -35
15 15 -30
-15
-25
-10 -15
10 10 -20 -25
-15
-10 -1 -20
-5 -10 0
5 5 -15
-5 -5 -15
-10 -10
-5
0 0
2 4 6 8 10 12 14 16 18 20 4 6 8 10 12 14 16 18 20
n n

(a) Asymptotic expansion (10.20) (b) Asymptotic expansion (10.21)

Figure 10.5: log10 of the relative errors (shown numerically along the contour lines) when the expansions
(10.20) and (10.21) are applied to different values of n using N terms in the respective the asymptotic
expansions. The areas shaded gray indicate where the accuracy is better than 10−15 .

particularly attractive in that all even powers of n have vanished.

The sector of validity for both (10.20) and (10.21) can be shown to be the entire complex plane with the
exception of the negative real axis. Both approximations are remarkably accurate already for quite modest
values of n, as seen in Figures 10.5 (a,b). We display here the relative errors that the expansions produce for
positive values of n, with -15 roughly corresponding to the regular computer machine rounding level. This
is for (10.20) reached already for n (= s) around 6, at which point the bottom right subplot in Figure 10.4
shows that the integrand is still far from being sharply peaked. Nevertheless, the local expansion around
x = 1 suffices for typical machine precision accuracy.

10.2.3 Leading order approximations

In case that we only want the leading term of the asymptotic expansion, the Taylor series procedure becomes
simple enough that we can write down the end result explicitly. For the Laplace integral (10.7):

Case 1: φ0 (a) < 0:

1 f (a) s φ(a)
I(s) ∼ − e . (10.22)
s φ0 (a)

Case 2: φ0 (c) = 0, φ00 (c) < 0, where the critical point c is located inside [a, b]:
√
2πf (c) s φ(c)
I(s) ∼ p e . (10.23)
−s φ00 (c)

For one-sided cases (c at either edge of [a, b]), the result should be halved.

This last Case 2 is a special case of φ(k) (c) = 0, k = 1, 2, . . . , p − 1 and φ(p) (c) < 0, with p even, for which
(assuming c is located inside [a, b]; else halve the result):

2Γ(1/p)(p!)1/p f (c) s φ(c)

I(s) ∼ e . (10.24)
p [−s φ(p) (c)]1/p
10.3. STEEPEST DESCENT 161

10.3 Steepest descent

10.3.1 Concept and background
In the description above of Laplace integrals, we made no use of two key opportunities provided by the
theory of analytic functions:

(i) The integration path can be altered freely (as long as we account for residues when crossing
singularities), and

(ii) The real and the imaginary parts of analytic functions are tightly coupled through the Cauchy-
Riemann (CR) equations (??).

Focusing first on item (ii), we consider an analytic function φ(z) and decompose it in its real and imaginary
parts
φ(z) = u(x, y) + i v(x, y) ,

where z = x + i y. The CR equations ux = vy , vx = −uy imply:

1. Both u(x, y) and v(x, y) are harmonic functions (satisfying uxx + uyy = 0 and vxx + vyy = 0, respec-
tively), so they cannot have any local maxima or minima,
 T  
ux vx
2. The CR equations also imply that = 0, telling that the gradient vectors (directions of
uy vy
steepest descent) for u(x, y) and v(x, y) are orthogonal to each other,

3. For any smooth functions in x, y, level curves are orthogonal to gradient directions.

Combining the observations just above, we next conclude:

4. Along the steepest descent paths of u(x, y), the function v(x, y) is constant, and vice versa.

5. A point z with φ0 (z) = 0 (and φ00 (z) 6= 0) will be a saddle point for both u(x, y) and v(x, y). The
function u(x, y) will at such point have two orthogonal steepest descent/ascent directions, with a local
maximum along one and a local minimum along the other, and with v(x, y) constant along both (and
equivalent for v(x, y)).

Figures 10.6 and 10.7 illustrate the general features described above in the special case of

φ(z) = z − log z . (10.25)

This function features a branch cut along the negative real axis (which is irrelevant for our present discussion).
The red dot marks the only location at which φ0 (z) = 0, which occurs at z = 1. As seen in Figure 10.6
(a), this is a saddle point for Re φ(z) - a local minimum if we follow the real axis near z = 1, and a local
maximum if we follow the (green) steepest descent path through that point. These same paths are constant
level paths for Im φ(z), c.f. part (b) of the figure.
Figure 10.7 displays φ(z) somewhat differently over the same region in the complex plane. Instead of
highlighting just one steepest descent path for Re φ(z) (level path for Im φ(z)), we illustrate several of them,
as solid black curves. The dashed curves are similarly the level paths for Re φ(z), and therefore also steepest
descent paths for Im φ(z). The two curve sets (solid and dashed) are everywhere orthogonal to each other,
apart from at the saddle point z = 1 and at the singularity z = 0. In the next subsection, we will see that
steepest descent path for Re φ(z) through the saddle point and its local tangent line there play key roles in
the steepest descent method for approximating Laplace integrals.
162 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS

5 5

0 0

-5 -5
3 3
2 2
1 1
0 3 0 3
2 2
-1 1 -1 1
0 0
-2 -1 -2 -1
-2 -2
-3 -3 -3 -3

(a) Real part of φ(z) = z − log z (b) Imaginary part of φ(z) = z − log z

Figure 10.6: Real and imaginary parts of φ(z) = z − log z, with the saddle point at z = 1 marked with a red
dot, and the steepest descent path through it for Re φ(z) shown in green. The same path becomes a level
curve for Im φ(z).

-1

-2

-3
-3 -2 -1 0 1 2 3

Figure 10.7: Steepest descent paths for Re φ(z) (equal to level curves for Im φ(z)) shown as solid curves,
and level curves for Re φ(z) (equal to steepest descent curves for Im φ(z)) shown as dashed curves. The
blue line segment is a tangent at the saddle point to the steepest descent curve (green) through this point.
10.3. STEEPEST DESCENT 163

10.3.2 Application of steepest descent to Laplace integrals

10.3.2.1 Integrals that become increasingly peaked
We consider again the Laplace integral (10.7):
ˆb
I(s) = f (z) es φ(z) dz, (10.26)
a

but denote the integration variable by z instead of x, as we will utilize that the path need not be the straight
line from a to b (nor do a and b need to be on the real axis).
1
Example 10.3.1 Determine an asymptotic expansion for Γ(s) as s → ∞, based on (??)
ˆ
1 1
= ez z −s dz , (10.27)
Γ(s) 2πi C
where C is a Hankel contour, as illustrated in Figure ?? (i.e. it enters from infinity in the third quadrant,
goes around the branch point at z = 0 and exits to infinity in the second quadrant).

The first step is to convert (10.27) to the Laplace form (10.26), which can be done with the change of
variable z → s z, which gives ˆ
1 1
= es (z−log z) dz , (10.28)
Γ(s) 2πi ss−1 C
with f (z) = 1, φ(z) = z − log z and with the same type of Hankel contour C. This function φ(z) has already
been illustrated above in Figures 10.6 and 10.7. Of all possible paths from the third to the second quadrant,
going around the branch point at the origin, we note that the highlighted steepest descent curve (green) has
some unique features. With Im φ(z) constant along this path, es Im(φ(z)) is also a constant, which thus can
be factored out of the integral, in the present case just a factor of “1”. We thus replace (10.28) by
ˆ
1 1
= es Re (z−log z) dz . (10.29)
Γ(s) 2πi ss−1 C
With a purely real-valued factor now multiplying s in the exponent, we can apply the insights from Sections
10.2.1-10.2.3. There are now two main options for generating the asymptotic expansion:
1. Integrate along the steepest descent path and then apply Watson’s lemma, and
2. Return to (10.28) and follow the straight line segment that is tangent to the steepest descent curve at
the saddle point (marked blue in Figures 10.6 and 10.7).
In either case, the complete asymptotic expansion can be deduced from an infinitesimal vicinity of the saddle
point. The second approach might appear to be simpler, but intermediate expansions may require quite a
large number of terms. We compare the two approaches below:

Method 1: Follow steepest descent curve: With φ(z) = z − log z = x + iy − 12 log(x2 + y 2 ) − i arctan xy ,
we obtain the steepest descent curve through the saddle point (located at z = 1) bysolvingIm φ(z) = 0,
y sin y dx dx
giving x = y/ tan y. Along this curve, Re φ(z) = tan y + log y and dz = dx + i dy = dy + i dy. Since dy
is an odd function of y, it will not contribute to the integral, and we obtain
ˆ ˆ 0+
1 1 s Re (z−log z) 1
= e dz ∼ f (y) es φ(y) dy , (10.30)
Γ(s) 2πi ss−1 C 2π ss−1 0−
y sin y
with f (y) = 1 and φ(y) = tan y + log y . Apart from the naming of the integration variable as y instead
of x, the problem is now in the standard form for making the key variable change (10.12) followed by using
Watson’s lemma as in Example 10.2.3. The remaining steps become
y sin y y2 y4 y6 y8 y 10 691y 12
−ξ = φ(y) − φ(0) = + log −1=− − − − − − − ... ,
tan y y 2 36 405 4200 42525 294698250
164 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS

ξ2 139ξ 3 571ξ 4 163879ξ 5

 
p ξ
y = 2ξ 1 − + + − − + ... ,
18 1080 680400 146966400 67898476800
ξ2 139ξ 3 571ξ 4
   
f (y) y 1 ξ
− 0 = =√ 1− + + − + ... ,
φ (y) 1 − 2y cot y + y 2 csc2 y 2ξ 6 216 97200 16329600
and therefore (by Watson’s lemma)
ˆ 
∞

2 es  e s r s 
 
1 f (y) −s ξ 1 1 139 571
∼ − 0 e dξ ∼ 1− + + − + ... .
Γ(s) 2π ss−1 φ (y) s 2π 12s 288s2 51840s3 2488320s4
0

It was here possible to obtain closed form expressions for the steepest descent curve. However, that was quite
unnecessary, as a local Taylor expansion of φ(z) around the saddle point would have sufficed (cf. Example
10.3.2, part “Path returning to z = 1”).

Method 2: Follow tangent at saddle point: Thus, let z = 1 + it, and consider t along an infinitesimally
short stretch around t = 0, along a real t-axis. The integral in (10.28) can then be replaced by
ˆ ˆ 0+
1 1 s Re (z−log z) −1
= e dz ∼ es (1+it−log(1+it)) dt . (10.31)
Γ(s) 2πi ss−1 C 2π ss−1 0−
The key idea this time is to Taylor expand the exponent, and separate off the constant and the quadratic
terms
2
   3 4 5 6 7 8
  3 4 5 6 7 8

s 1− t +s i t + t −i t − t +i t + t +... st2 s i t + t −i t − t +i t + t8 +... =
es (1+it−log(1+it)) = e 2 3 4 5 6 7 8
= es e− 2 e 3 4 5 6 7
ist3 st4 ist5 s(3 + s)t6 is(12 + 7s)t7 s(60 + 47s)t8
 
2
s − st2
=e e 1+ + − − + + + ... . (10.32)
3 4 5 18 84 480
st2
If our aim is just to obtain the leading term, we don’t need anything beyond the factor es e− 2 , making this
approach then very convenient. Pursuing more terms, we note that, following the leading “1” in the last
expansion above, the next three terms are of size O(s), then three of size O(s2 ), etc.. Knowing that only a
vicinity of t = 0 matters, we next swap the integration interval from [−ε, ε] to [−∞, ∞], after which we can
do term-by-term integration in closed form thanks to the relation (α > 0):
ˆ∞  n+1
−α t2 n Γ( n+1
2 )/α
2 if n even
e t dt = (10.33)
0 if n odd
−∞
´∞ 2 n+1
(which generalizes for one-sided cases to 0 e−α t tn dt = 12 Γ( n+1
2 )/α
2 , n > −1). Since all terms with odd

powers of t vanish, we obtain from (10.32), using even powers up through t12 , that
r
1 −1 s π 1 3 1 5 3+s 7 60 + 47s
= e ( 2 1/2 + − + (10.34)
Γ(s) 2π ss−1 2 s 2 s3/2 3 s5/2 16 s7/2
3 1260 + s(1377 + 175s) 11 90720 + s(120564 + 7s(4437 + 80s))
− 9/2
+ + . . .)
20 s 576 s11/2
This needs to be re-sorted into inverse powers of s. The increasing powers of s in the numerators slows up
the process, but the powers of s in the denominators grow faster in the long run. When including still one
more term in (10.34), we arrive at

1  e s r 1  1 1 139

= 1− + + + . . . .  (10.35)
Γ(s) s 2πs 12s 288s2 51840s3
The need for a quite large number of terms in the internal expansions is not unusual, and can significantly
increase the amount of work that is required when one chooses to follow the tangent instead of the steepest
descent curve. However, following the tangent can be very convenient if we only want the leading term.
Also, in some cases, it can provide closed form expressions for complete expansions (as we will see for the
Airy function in Section 10.3.3).
10.3. STEEPEST DESCENT 165

10.3.2.2 Integrals that become increasingly oscillatory

The key observation here is that an analytic function that is highly oscillatory in one direction in the complex
plane may become rapidly decaying in another direction. The steepest descent approach may then again be
effective in obtaining asymptotic expansions, as illustrated in the next example:

´1
Example 10.3.2 Find the asymptotic expansion for I(s) = 0
cos sx2 dx as s → ∞.

Figure 10.8 (a) illustrates this integrand for s = 10 and s = 100. The cost of direct numerical integration
increases rapidly with s, since the the whole interval will then need to become increasingly finely resolved.
The integrand cos sx2 grows rapidly in size when analytically continued away from the real axis, ´ 1so an imme-
2
diate change of integration path is unlikely to help. However, we can rewrite I(s) as I(s) = Re 0 es (−ix ) dx,
and the form is now that of a Laplace integral. Figures 10.8 (b) and 10.9 illustrate the corresponding function
φ(z) = −i z 2 in the complex plane. If we, instead of the original path [0, 1] (marked blue) follow the green
path starting at z = 0 and then return via the other green path ending at z = 1, the value of the integral has
not changed, but we now have two steepest descent paths. Only immediate vicinities of z = 0 and z = 1 will
thus matter for the asymptotic expansion. Had the value of Re φ(0) been different from that of Re φ(1), we
could for the asymptotic expansion have ignored the path with the lower value, but Re φ(0) = Re φ(1) = 0, so
we need to consider both paths separately. We write I(s) = Re (J(s)+K(s)) corresponding to the two paths.
´ 0+ 2
Path originating at z = 0: The variable change z = ei π/4 t gives dz = ei π/4 dt and J(s) ∼ ei π/4 e−s t dt ∼
´∞ 2 i π/4 p
0
ei π/4 0 e−s t dt = e 2 π 1
pπ
s , i.e. Re J(s) = 2 2s .

Path returning to z = 1: As an alternative to first splitting φ(z) in real and imaginary parts and
determining the steepest descent path, we immediately apply (10.12): −ξ ´ = φ(z) −1/2− φ(1) = i z 2 − i
−1/2 i is ∞
1/2
and obtain z = (1 + iξ) , dz = 2 (1 + iξ) i
dx, and K(s) ∼ 2 e 0 (1 + iξ) e−ξ s dξ. With
n 1 n 1
∞ (−i ξ) Γ(n+ ) ∞ (−i) Γ(n+ )
(1 + i ξ)−1/2 = n=0 , Watson’s lemma gives K(s) ∼ − 2i ei s n=0 Γ( 1 ) sn+12 .
P 2
P
n! Γ( 1 )2 2

Accuracy of the expansion: In the same style as Figures 10.2 and 10.5, Figure 10.10 shows the er-
ror in the asymptotic expansion

ˆ1 " ∞
#
i i s X (−i)n Γ(n + 21 )
r
2 1 π
I(s) = cos sx dx ∼ − Re e
2 2s 2 n=0
Γ( 21 ) sn+1
0

as function of s and when using terms up through n = N in the expansion.The right edge (s = 100)
corresponds to the red curve in Figure 10.8 (a). 

10.3.3 Steepest descent analysis of Ai(z)

The Airy function Ai(z) was introduced´ ∞ (and illustrated) in Section 9.4.1.1 as the unique bounded solution
to u00 (z) − z u(z) = 0, normalized by −∞ Ai(z)dz = 1, Based on its integral representation (9.7)
ˆ
1 3
Ai(z) = e−t /3+z t
dt (10.36)
2πi
C

we will next use steepest descent to find its asymptotic expansions for z → +∞ and z → −∞ (leaving the
case of z → ∞ in an arbitrary direction from the origin to Exercise ??). The integration path C in (10.36)
connects ’Valley V2 ’ to ’Valley V3 ’, as illustrated in Figure 9.5. The first step is to change variables to bring
(10.36) to standard Laplace form with the exponential factor es φ(z) , where s the parameter that goes to plus
infinity and z the integration variable.
166 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS

1.5

0.5
0.5

0 0

-0.5
-0.5

-1

-1 s = 10
s = 100

-1.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -1 -0.5 0 0.5 1 1.5 2

(a) The integrand cos sx2 . (b) Integration paths.

Figure 10.8: (a) The integrand cos sx2 displayed for s = 10 and s = 100, (b) The steepest descent paths
for Re φ(z) with φ(z) = i z 2 (black, solid) and level curves (black, dashed). The original integration path is
here shown in blue and the two steepest descent paths are shown in green.

8 4

6 3
4
2
2
1
0
0
-2
-1
-4

-6 -2

-8 -3
1.5 1.5
1 1
0.5 2 0.5 2
1.5 1.5
0 0
1 1
-0.5 0.5 -0.5 0.5
-1 0 -1 0
-0.5 -0.5
-1.5 -1 -1.5 -1

(a) Real part of φ(z) = i z 2 . (b) Imaginary part of φ(z) = i z 2 .

Figure 10.9: Real and imaginary parts of φ(z) = i z 2 , with the two ends of the original integration interval
marked with a red dots, and the steepest descent paths for Re φ(z) shown by green curves (becoming level
curves for Im φ(z)).
10.3. STEEPEST DESCENT 167

50

3 40

-10
-35

-5
40

10 20 0
-25 -30
30

-15
N

20 -25
-20
0

-20
10 -15
-10 -15
-5 -10 -10
0
10 20 30 40 50 60 70 80 90 100
s

Figure 10.10: log10 of the relative errors (shown numerically along the contour lines) when the expansions
(10.20) and (10.21) are applied to different values of n using N terms in the respective the asymptotic
expansions. The areas shaded gray indicate where the accuracy is better than 10−15 .

10.3.3.1 Ai(z) for z → +∞

Method 1: Follow steepest descent curve: In order for (10.36) to take the form of a Laplace integral,
z should be a factor in the exponent −t3 /3 + z t in (10.36), suggesting the change of variable t = z 1/2 τ with
1
´ 1/2 +z3/2 (−τ 3 /3+τ )
dt = z 1/2 dτ , giving Ai(z) = 2πi C
z e dτ . We next arrive at the standard form by further
re-naming z 3/2 → s and then τ → z:
ˆ
s1/3 3
I(s) = Ai(s2/3 ) = es(−z /3+z)
dz . (10.37)
2πi
C

With φ(z) = −z 3 /3 + z, solving dφ

dz = 0 shows that the only two saddle points are z = ±1. The ’lowest’ path
between the valleys V2 to V3 will thus go through
p the saddle at z = −1, cf. Figures 10.11 (a,b). The
p steepest
descent path becomes Im φ(z) = 0, i.e. x = − 1 + y 2 /3 and, along it, Re φ(z) = − 92 (3 + 4y 2 ) 1 + y 2 /3.
The standard Watson’s lemma variable change gives

2 p 2 5y 4 7y 6 y8 11y 10 91y 12
−ξ = φ(y) − φ(0) = − (3 + 4y 2 ) 1 + y 2 /3 + = y 2 + − + − + + ... ,
9 3 36 648 576 31104 1119744
which becomes, when inverted

77ξ 2 2431ξ 3 1062347ξ 4 14003665ξ 5 1168767425ξ 6

 
p 5ξ
y = ξ 1− + − + − + + .... .
72 3456 248832 214990848 515978352 743008370688
 
As in Example 10.3.1, Method 1, dz = dx + i dy = dx dy + i dy, where dx
dy can be omitted since it is an
p
odd function that will vanish when integrating. With f (y) = 1 and φ(y) = − 29 (3 + 4y 2 ) 1 + y 2 /3, we get
n o √
3 9+3y 2
− φf0(y)
(y) = 18y+8y 3 , i.e.

ξ −1/2 5ξ 1/2 385ξ 3/2 17017ξ 5/2 1062347ξ 7/2 154040315ξ 9/2 15193976525ξ 11/2
 
f (y)
− = − + − + − + +. . .
φ0 (y) 2 48 6219 497664 47775744 10319560704 1486016741376

and then, by Watson’s lemma

e−2s/3
 
5 385 85085 37182145 5391411025 5849680962125
I(s) = √ 1/6 1− + − + − + + . . . .
2 πs 48s 4608s2 663552s3 127401984s4 6115295232s5 1761205026816s6
168 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS

30 1

20

10
0
0

-10

-1
-20

-30
3
2 -2
1
0
3
-1 2
1
-2 0 -3
-1
-3
-3
-2 -3 -2 -1 0 1 2 3
3
(a) Real part of φ(z) = − z3 + z. (b) Steepest descent paths
3
Figure 10.11: Displays of the functionφ(z) = − z3 + z, with its saddle points at z = ±1 marked with red
dots, and the steepest descent path through the one at z = −1 shown in green. The same path is a level
curve for Im φ(z), and is therefore the appropriate integration path for the present problem. (a) Surface
plot of Re φ(z) and (b) steepest descent paths of Re φ(z) (solid) and level curves of Re φ(z) (dashed).

The expansion for Ai(z), z → +∞ follows now by substituting s → z 3/2 .

3 3
Method 2: Follow tangent at saddle point: With z = −1+i t, we obtain φ(t) = − z3 +z = − 23 −t2 − i 3t
and therefore
ˆ  3
 ˆ
s− 32 −t2 − i 3t 2s 2 i s t3
e dt ∼ e− 3 e−s t e 3 dt {substitute u = s1/2 t, i.e. t = s−1/2 u, dt = s−1/2 u}
C
C
ˆ 0+ u3
 
2s 2 i√
∼ e− 3 e−u e 3 s
i s−1/2 du
0−
2s ˆ ∞ ∞ n
e− 3 1 i u3

−ut2
X
∼ i 1/2 e √ du {note that only even terms contribute}
s −∞ n=0
n! 3 s
2s ∞ m
e− 3 Γ(3m + 21 )

X 1
∼i − .
s1/2 m=0
(2m)! 9s

Therefore, for z → +∞:

∞ m
z −1/4 − 2 z3/2 X Γ(3m + 12 )

1
Ai(z) ∼ √ e 3 − 3/2 , (10.38)
2 π m=0
(2m)! 9z

and we have obtained the complete expansion in closed form.

10.3.3.2 Ai(z) for z → + − ∞

Instead of (10.37), we need now to consider the s → +∞ limit of
ˆ
2/3 s1/3 3
I(s) = Ai(−s ) = es(−z /3−z) dz , (10.39)
2πi
C
10.3. STEEPEST DESCENT 169

for which φ(z) = −z 3 /3 − z. The change of sign of the last term of φ(z) will have no bearing on the character
of the valleys V2 and V3 , but it has caused the saddle points to move from z = ±1 to z = ±i. The lowest
path between the two valleys will now be a composite of two steepest descent sections, with equally high
saddles, as shown in Figure 10.12 (a,b). Both Methods 1 and 2 will work once we sum the contributions
from the two saddles. Omitting the algebra, it will transpire that Method 2 again will produce a closed form
result:
"  ∞ 2m
(−z)−1/4 π X (−1)m Γ(6m + 21 )
 
2 1
Ai(z) ∼ √ cos (−z)−3/2 − − + (10.40)
π 3 4 m=0 (4m)! Γ( 12 ) 9(−z)3/2
 X ∞ m Γ(6m + 7 )
 2m+1 #
2 π (−1) 1
+ sin (−z)−3/2 − 2
−
3 4 m=0 (4m + 2)! Γ( 12 ) 9(−z)3/2

Figure 10.13 shows that both this expansion (10.40) for z negative and (10.38) for z positive can give excellent
accuracies for |z| greater than about 10 or 15. In contrast, it can easily be verified that the Taylor expansion
of Ai(z) around the origin barely is useful beyond |z| around two.
The form of (10.40) does not convey very clearly that the Ai(z) function for z → −∞, while decaying in
magnitude, oscillates with a slowly changing frequency (as was seen in Figure 9.1). Hence, it is preferable to
re-express (10.40) in terms of magnitude and phase angle. Denotingpthe two sums by S1 and S2 , respectively,
we thus apply the trigonometric identity (cos α) S1 + (sin α) S2 = S12 + S22 sin(α + arctan SS21 ) to obtain

Ai(z) = M (z) sin θ(z)

where, for z → −∞,

∞
1 X 1 · 3 · 5 · . . . · (6k − 1) 1
M 2 (z) ∼
π(−z)1/2 k!(96)k z3
k=0

and  
π 2 3/2 5 1105 82825 1282031525
θ(z) ∼ + (−z) 1+ + + + + ...
4 3 32 z 3 6144 z 6 65536 z 9 58720256 z 12
Although no practical closed form expressions are available for either M (z) or θ(z), that is of little practical
consequence, since this re-formulation into magnitude and phase angle was mainly to provide qualitative
insight. In general, closed form expressions for asymptotic expansions are rarities in all but the simplest
cases. What is important is that we have systematic methods available to generate some moderate number of
leading terms in asymptotic expansions. Although steepest descent provides a very powerful mathematical
approach, symbolic algebra packages are essential for utilizing this most effectively.
170 CHAPTER 10. ASYMPTOTIC ANALYSIS OF INTEGRALS

20 1

15

10

5 0
0

-5

-10
-1
-15

-20
3
2 -2
1
0
3
-1 2
1
-2 0 -3
-1
-3
-3
-2 -3 -2 -1 0 1 2 3
3
(a) Real part of φ(z) = − z3 − z. (b) Steepest descent paths
3
Figure 10.12: Displays of the functionφ(z) = − z3 − z, with its saddle points at z = ±i marked with red
dots, The appropriate lowermost path passes through both saddles, and is shown in green. (a) Surface plot
of Re φ(z) and (b) steepest descent paths of Re φ(z) (solid) and level curves of Re φ(z) (dashed).

20 20
-3 -35
0
5

-15
15

5
10

-4
-10

-20
10

-5
0

-25

18 18
-10

5
-3
5
-25

16 -40 16
-15

20
-5

-30
14 0
-3 14
15
0

-25
-20

-35
12 12
-2
0
5

-1
10

-30
-10

-25
5
-5
N

10 10
-10
5

-20
-15

8 8
-25
-20
-5

-15
6 6
-20
0

-15 -15
4 -10 4
-10
-15
-10
-10 -5
2 -10 2
-5
-5 -5
0 0
-30 -25 -20 -15 -10 -5 5 10 15 20 25 30
x x

(a) Asymptotic expansion, z ∈ [−30, −1]. (b) Asymptotic expansion z ∈ [1, 30].

Figure 10.13: log10 of the relative errors (shown numerically along the contour lines) when Ai(x) is approx-
imated at locations z along the real axis when using N terms in each of the expansions in (10.40) and N
terms in the expansion (10.38), respectively. The areas shaded gray indicate where the accuracy is better
than 10−15 .
Appendix A

Illustrations of Conformal Mappings

A.1 Elementary functions

A.2 Special functions

171
172 APPENDIX A. ILLUSTRATIONS OF CONFORMAL MAPPINGS
Bibliography

[1] Copson E.T., An Introduction to the Theory of Functions of a Complex Variable, Clarendon Press, 1935.

[2] Borwein J. and Devlin K., The Computer as a Crucible, A K Peters Ltd, Wellesley, MA, 2009.
[3] Ahlfors L.V., Complex Analysis, McGraw-Hill, New York, 1966.
[4] Clarkson P.A., NIST Handbook of Mathematical Functions, ch. 32, pp. 723 – 740, U.S. Dept. Commerce,
2010.

173
Index

A periodic functions
Ablowitz-Segur solution, 150 doubly periodic, 113
Airy function, 165 simply periodic, 113
Airy’s equation, 141, 144 Picard’s theorem, 118
Asymptotic analysis, 153
R
B Riemann sheets, 150
Bäcklund transformations, 148
Bessel’s equation, 142, 145 S
Schwartz reflection principle, 118
C steepest descent, 161
Cauchy-Riemann equations, 161 Stirling’s formula, 159
Stokes phenomenon, 155
E
elliptic functions, 113 W
entire function, 118 Watson’s lemma, 157
error function, 153 Weierstrass theorem, 118
Euler’s method, 145 Weierstrass ℘-function, 114, 146
F
Fourier-Laplace method, 144

G
gamma function, 159

H
harmonic functions, 161
Hastings-McLeod solution, 150
hypergeometric functions, 142

L
Laplace integral, 155
Laplace integrals, 155, 163
Laurent expansion, 117
Legendre’s equation, 145
Liouville’s theorem, 113, 118

M
modular function, 118
Morera’s theorem, 119

O
ODE, 141
linear, 141
nonlinear, 115, 145

P
Painlevé equations, 145

174