A complex number is z=a+bi where a,b∈R and i2=−1. We call a=Re(z) the real part and b=Im(z) the imaginary part.
Arithmetic:(a+bi)+(c+di)=(a+c)+(b+d)i and (a+bi)(c+di)=(ac−bd)+(ad+bc)i.
Proposition 1.1 (Properties of Complex Arithmetic). For all z,w∈C:
z+w=w+z and zw=wz (commutativity)
(z+w)+u=z+(w+u) and (zw)u=z(wu) (associativity)
z(w+u)=zw+zu (distributivity)
There exist additive identity 0 and multiplicative identity 1.
Every z=0 has a multiplicative inverse z1=∣z∣2zˉ.
Remark. The complex field C cannot be ordered: there is no total ordering on C Compatible with the field operations. In particular, i2=−1 precludes any such ordering.
1.2 The Complex Conjugate and Modulus
Definition. The complex conjugate of z=a+bi is zˉ=a−bi.
Proposition 1.2. For all z,w∈C:
z+w=zˉ+wˉ and zw=zˉwˉ
zzˉ=∣z∣2
z+zˉ=2Re(z) and z−zˉ=2iIm(z)
zˉˉ=z
Definition. The modulus (or absolute value) of z=a+bi is ∣z∣=a2+b2.
Proposition 1.3 (Modulus Properties). For all z,w∈C:
∣z∣≥0 with equality iff z=0
∣zw∣=∣z∣∣w∣
∣z+w∣≤∣z∣+∣w∣ (triangle inequality)
∣z∣−∣w∣≤∣z−w∣ (reverse triangle inequality)
Proof of (3).∣z+w∣2=(z+w)(zˉ+wˉ)=∣z∣2+zwˉ+zˉw+∣w∣2=∣z∣2+2Re(zwˉ)+∣w∣2≤∣z∣2+2∣z∣∣w∣+∣w∣2=(∣z∣+∣w∣)2. The inequality follows from Re(zwˉ)≤∣zwˉ∣=∣z∣∣w∣. ■
1.3 Polar Form
Every non-zero complex number can be written in polar form:
z=r(cosθ+isinθ)=reiθ
Where r=∣z∣=a2+b2 is the modulus and θ=arg(z) is the argument.
Definition. The principal argumentArg(z) is the unique θ∈(−π,π] Such that z=∣z∣eiθ. The argumentarg(z) is multi-valued: arg(z)=Arg(z)+2πk for k∈Z.
Proposition 1.4. If z1=r1eiθ1 and z2=r2eiθ2Then z1z2=r1r2ei(θ1+θ2) and z1/z2=(r1/r2)ei(θ1−θ2).
Worked Examples: Polar Form Conversions
Solution
Problem. Convert z=−1+3i to polar form and find all arguments.
∣z∣=(−1)2+(3)2=1+3=2.
Re(z)=−1<0 and Im(z)=3>0So z is in the second quadrant.
θ=arctan(−13)=32π (adjusting to second quadrant).
Both real and imaginary parts are negative, so z is in the third quadrant.
θ=arctan(4/3)+π=π+arctan(4/3).
z=5ei(π+arctan(4/3)).
1.4 Euler”s Formula and De Moivre’s Theorem
Euler’s formula:eiθ=cosθ+isinθ.
De Moivre’s theorem:(eiθ)n=einθSo
(cosθ+isinθ)n=cos(nθ)+isin(nθ)
Proposition 1.5. De Moivre’s theorem holds for all integers nIncluding negative values.
Proof. For n≥0It follows by induction from the multiplication law eiαeiβ=ei(α+β). For n<0Write n=−m with m>0: (cosθ+isinθ)n=(cosθ+isinθ)m1=cos(mθ)+isin(mθ)1=cos(−mθ)+isin(−mθ)=cos(nθ)+isin(nθ). ■
Definition. An n-th root of w∈C is a complex number z such that zn=w.
Proposition 1.6. Every non-zero w∈C has exactly n distinct n-th roots. If w=ρeiϕThen
zk=ρ1/nei(ϕ+2πk)/n,k=0,1,…,n−1
Where ρ1/n>0 is the positive real n-th root of ρ.
Proof. If zn=wWrite z=reiθ. Then rneinθ=ρeiϕ So r=ρ1/n and nθ=ϕ+2πk. For k=0,1,…,n−1 these give distinct Values of θ; for k≥n they repeat. ■
Remark. The n-th roots of w lie equally spaced on a circle of radius ρ1/nForming a Regular n-gon.
1.6 Roots of Unity
The n-th roots of unity are the solutions of zn=1:
zk=e2πik/n,k=0,1,…,n−1
They form a regular n-gon on the unit circle in the complex plane.
Proposition 1.7. If ω=e2πi/n is a primitive n-th root of unity, then ∑k=0n−1ωk=0 and ∑k=0n−1ωjk=0 for any j not divisible by n.
Proof. The sum ∑k=0n−1ωk=1−ω1−ωn=1−ω1−1=0 Provided ω=1. For j not divisible by n, ωj is a non-trivial root of unity, So the same argument applies. ■
Solution
Problem. Find all cube roots of −8.
−8=8eiπ. The cube roots are: zk=81/3ei(π+2πk)/3=2ei(π+2πk)/3 for k=0,1,2.
Problem. Show that the n-th roots of any non-zero w are in geometric progression.
The roots are zk=ρ1/nei(ϕ+2πk)/n=z0⋅(e2πi/n)k=z0⋅ωk Where ω=e2πi/n is a primitive n-th root of unity. This is a geometric sequence With ratio ω.
2. Complex Functions and Analyticity
2.1 Complex Functions
A complex function is a function f:D⊆C→C. We can write f(z)=u(x,y)+iv(x,y) where z=x+iy and u,v are real-valued functions.
Example.f(z)=z2=(x+iy)2=(x2−y2)+i(2xy). Here u=x2−y2 and v=2xy.
Example.f(z)=zˉ=x−iy. Here u=x and v=−y.
Example.f(z)=∣z∣2=x2+y2. Here u=x2+y2 and v=0.
2.2 Limits and Continuity
The limit limz→z0f(z)=L means: for every ε>0There exists δ>0 Such that 0<∣z−z0∣<δ implies ∣f(z)−L∣<ε.
Unlike the real case, z can approach z0 from any direction in C. This makes limits More restrictive.
Proposition 2.1.limz→z0f(z)=L if and only if lim(x,y)→(x0,y0)u(x,y)=a And lim(x,y)→(x0,y0)v(x,y)=b where L=a+bi.
Definition.f is continuous at z0 if limz→z0f(z)=f(z0).
Solution
Problem. Show that limz→0zzˉ does not exist.
Let z=reiθ. Then zzˉ=e−2iθ. As z→0 along different Rays (θ=0,π/2,π/4Etc.), the ratio takes different values (1,−1,−iEtc.). Since the limit depends on the direction of approach, it does not exist.
Problem. Determine whether f(z)=z−1z2−1 is continuous at z=1.
For z=1: f(z)=z+1. The limit as z→1 is 2But f(1) is undefined (division by zero). If we define f(1)=2Then f becomes continuous at z=1.
2.3 The Derivative
Definition.f is differentiable at z0 if
f′(z0)=limh→0hf(z0+h)−f(z0)
Exists (and is independent of how h→0 in C).
Remark. The requirement that the limit be the same for all directions of approach of h is what Makes complex differentiability far more restrictive than real differentiability.
2.4 Analytic Functions
Definition. A function f is analytic (or holomorphic) on an open set U⊆C if f is differentiable at every point of U. A function that is analytic On all of C is called entire.
Examples of entire functions:zn, ez, sinz, coszPolynomials.
Example of a non-analytic function:f(z)=zˉ is nowhere differentiable (except at z=0 if we define it, but still not analytic there).
Solution
Problem. Show that f(z)=∣z∣2 is differentiable only at z=0.
f(z)=x2+y2So u=x2+y2 and v=0. ux=2x, uy=2y, vx=0, vy=0. The Cauchy-Riemann equations require 2x=0 and 2y=0So x=y=0. Thus f satisfies CR only at z=0.
At z=0: f′(0)=limh→0h∣h∣2=limh→0hˉ=0So f is Differentiable at 0 but not analytic anywhere (no neighbourhood of 0 is analytic).
Problem. Show that f(z)=zzˉ+z is differentiable only at z=0.
f(z)=∣z∣2+z=(x2+y2+x)+iy. ux=2x+1, uy=2y, vx=0, vy=1. CR equations: 2x+1=1⇒x=0And 2y=0⇒y=0. At (0,0): f′(0)=limh→0hhhˉ+h=limh→0(hˉ+1)=1. So f is differentiable at z=0 only, hence nowhere analytic.
2.5 Branch Cuts and Multi-Valued Functions
Many important functions in complex analysis are inherently multi-valued. To work with them as Single-valued functions, we must restrict the domain.
Definition. A branch of a multi-valued function f is a single-valued analytic function g Defined on a domain D such that g(z)∈f(z) for all z∈D.
The Complex Logarithm. We define logz=ln∣z∣+iarg(z)Which is multi-valued because arg(z)=Arg(z)+2πk for k∈Z. The principal branch is
Logz=ln∣z∣+iArg(z)
Defined on C∖(−∞,0]. The negative real axis is called the branch cut.
Proposition 2.2. The principal branch Logz is analytic on C∖(−∞,0] and dzdLogz=z1.
Complex Powers. For z,α∈C with z=0:
zα=eαlogz
This is multi-valued . When α is rational with reduced form p/qThere are exactly q distinct values.
Solution
Problem. Find all values of (−1)i.
(−1)i=eilog(−1)=ei(iπ+2πik)=e−π−2πk for k∈Z.
These are all positive real numbers: …,e3π,eπ,e−π,e−3π,…. The principal value (using the principal branch) is e−π.
Problem. Find all values of i1/2.
i1/2=e(1/2)logi=e(1/2)(iπ/2+2πik)=eiπ/4+iπk.
For k=0: eiπ/4=22(1+i). For k=1: ei5π/4=−22(1+i). These are the two square roots of i.
Problem. Find the domain of analyticity of f(z)=Log(z2+1).
Logw is analytic on C∖(−∞,0]So we need z2+1∈/(−∞,0].
z2+1≤0 when z2≤−1I.e., z∈[−i,0]∪[0,i] (the imaginary axis Segment from −i to i). Also z2+1=0 at z=±i.
Domain: C∖{z:z=iy,y∈[−1,1]}.
3. The Cauchy-Riemann Equations
3.1 Statement
Theorem 3.1 (Cauchy-Riemann Equations). If f(z)=u(x,y)+iv(x,y) is differentiable at z=x+iyThen
∂x∂u=∂y∂v,∂y∂u=−∂x∂v
Proof. Compute the limit along the real axis (h∈R, h→0):
Equating real and imaginary parts: ∂x∂u=∂y∂v And ∂x∂v=−∂y∂u. ■
3.2 Sufficiency Condition
Theorem 3.2. If u and v have continuous first partial derivatives on an open set U and Satisfy the Cauchy-Riemann equations on UThen f=u+iv is analytic on U.
Proof. Since ux,uy,vx,vy are continuous on U, u and v are (real) differentiable. Let Δz=Δx+iΔy. By real differentiability:
By CR: uy+ivy=−vx+iux=i(ux+ivx). Substituting:
=(ux+ivx)Δx+iΔyΔx+iΔy+Δzo(∣Δz∣)→ux+ivx
As Δz→0. ■
3.3 The Derivative in Terms of Partial Derivatives
When the Cauchy-Riemann equations hold:
f′(z)=∂x∂u+i∂x∂v=∂y∂v−i∂y∂u
3.4 Harmonic Functions
Definition. A real-valued function ϕ(x,y) is harmonic if ϕxx+ϕyy=0 (Laplace’s equation).
Proposition 3.3. If f=u+iv is analytic, then u and v are harmonic.
Proof. From the Cauchy-Riemann equations: ux=vy and uy=−vx. Differentiating: uxx=vyx and uyy=−vxy. By equality of mixed partials, uxx+uyy=vyx−vxy=0. Similarly for v. ■
Definition. If u and v are harmonic on U and satisfy the Cauchy-Riemann equations, then v is the harmonic conjugate of u.
Proposition 3.4. If U is a connected domain and u is harmonic on UThen u has A harmonic conjugate on UUnique up to an additive constant.
Proof. Define v(x,y)=∫(x0,y0)(x,y)(−uydx+uxdy). The integrand is closed (since (−uy)y=−uyy=uxx=(ux)x) and since U is Connected, v is well-defined (path-independent) by Green’s theorem. Then vx=−uy and vy=uxWhich are the CR equations. ■
Solution
Problem. Find the harmonic conjugate of u(x,y)=x3−3xy2.
Verify u is harmonic: uxx=6x, uyy=−6xSo uxx+uyy=0. ✓
By CR: vy=ux=3x2−3y2So v=3x2y−y3+g(x). Also vx=−uy=6xySo 6xy=6xy+g′(x)Giving g′(x)=0So g(x)=C.
Harmonic conjugate: v(x,y)=3x2y−y3+C.
Note:f(z)=u+iv=x3−3xy2+i(3x2y−y3)=(x+iy)3=z3.
Problem. Show that u(x,y)=ln(x2+y2) is harmonic on R2∖{0} but Has no harmonic conjugate on R2∖{0}.
ux=(x2+y2)2y2−x2, vy=(x2+y2)2y2−x2. So ux=vy. ✓
uy=(x2+y2)2−2xy, vx=(x2+y2)22xy. So uy=−vx. ✓
f′(z)=ux+ivx=(x2+y2)2−(x2−y2+2ixy)=z2−1. ■
4. Complex Integration
4.1 Contours
A contour (or piecewise smooth path) in C is a continuous function γ:[a,b]→C that is differentiable except at finitely many points, with a Continuous derivative everywhere it exists.
A simple closed contour is a contour with γ(a)=γ(b) and no other Self-intersections.
4.2 The Complex Integral
Definition. For a contour γ and a continuous function f on γ:
∫γf(z)dz=∫abf(γ(t))γ′(t)dt
4.3 Basic Properties
Proposition 4.1. The complex integral is linear:
∫γ(af+bg)dz=a∫γfdz+b∫γgdz
Proposition 4.2. Reversing orientation changes the sign:
∫−γfdz=−∫γfdz
Proposition 4.3. Additivity over contours:
∫γ1+γ2fdz=∫γ1fdz+∫γ2fdz
4.4 ML Inequality
Proposition 4.4 (ML Inequality). If ∣f(z)∣≤M for all z on a contour γ of length LThen
Check: Since z is entire, the integral from 0 to 1+i is 21(1+i)2=i. Consistent. ■
4.6 ML Inequality Applications
Solution
Problem. Use the ML inequality to show that limR→∞∫CRzeizdz=0 Where CR is the upper semicircle ∣z∣=R, Im(z)≥0.
On CR: z=Reiθ, 0≤θ≤π. ∣eiz∣=∣eiR(cosθ+isinθ)∣=e−Rsinθ.
∫CRzeizdz≤∫0πRe−Rsinθ⋅Rdθ=∫0πe−Rsinθdθ.
By Jordan’s inequality sinθ≥π2θ for θ∈[0,π/2]:
≤2∫0π/2e−2Rθ/πdθ=Rπ(1−e−R)→0 as R→∞. ■
Problem. Bound ∫γz2+4dz where γ is ∣z∣=3.
On γ: ∣z2+4∣≥∣z∣2−4=9−4=5 (reverse triangle inequality). So z2+41≤51.
Length of γ: L=2π⋅3=6π.
∫γz2+4dz≤51⋅6π=56π.
4.7 Antiderivative Method
When f is analytic on a connected domain and has a known antiderivative F with F′=f:
∫γf(z)dz=F(γ(b))−F(γ(a))
This follows from the fundamental theorem of calculus applied to F(γ(t)).
Solution
Problem. Evaluate ∫γcoszdz where γ is any path from 0 to π+i.
Since cosz is entire with antiderivative sinz:
∫γcoszdz=sin(π+i)−sin(0)=sin(π+i).
sin(π+i)=sinπcosh1+icosπsinh1=−isinh1.
So the integral equals −isinh1.
Problem. Evaluate ∫γe2zdz where γ is any path from 1 to i.
Antiderivative: 21e2z.
∫γe2zdz=21(e2i−e2).
5. Cauchy’s Theorem
5.1 Statement
Theorem 5.1 (Cauchy’s Theorem). If f is analytic on a connected domain D and γ Is a simple closed contour in DThen
∫γf(z)dz=0
Proof (for f′ continuous). By Green’s theorem in the plane, writing f=u+iv:
∫γfdz=∫γ(udx−vdy)+i∫γ(vdx+udy)
Applying Green’s theorem to each integral:
=∬D(−vx−uy)dA+i∬D(ux−vy)dA=0
By the Cauchy-Riemann equations. ■
5.2 Connected Domains
A domain D⊆C is ** connected** if every simple closed contour in D can Be continuously shrunk to a point within D.
Cauchy’s theorem may fail on multiply connected domains. For example, ∫γz1dz=2πi where γ is the unit circle (traversing a region that Excludes the singularity at z=0).
5.3 Path Independence
Corollary 5.2. If f is analytic on a connected domain DThen the integral ∫z0z1f(z)dz is independent of the path from z0 to z1 in D.
5.4 Antiderivatives
Theorem 5.3. If f is analytic on a connected domain DThen f has an antiderivative F in D (i.e., F′(z)=f(z)), and
∫γf(z)dz=F(z1)−F(z0)
Where z0 and z1 are the endpoints of γ.
5.5 Cauchy’s Theorem for Multiply Connected Domains
Theorem 5.4. If f is analytic on a domain D containing simple closed contours γ,γ1,…,γn where γ1,…,γn Lie in the interior of γ and the region between γ and the γk is contained in D And all contours are positively oriented, then
∫γf(z)dz=∑k=1n∫γkf(z)dz
5.6 Deformation of Contours
Theorem 5.5 (Deformation of Contours). If f is analytic on a domain containing two simple Closed contours γ1 and γ2 where one can be continuously deformed into the other Within the domain of analyticity of fThen
∫γ1f(z)dz=∫γ2f(z)dz
Proof. This follows directly from Theorem 5.4 applied to the region between γ1 and γ2. ■
Remark. This theorem is enormously useful: we can replace a complicated contour with a simpler one (a small circle around each singularity) without changing the value of the integral.
Solution
Problem. Evaluate ∫γz−2dz where γ is the ellipse 4x2+9y2=1.
Since z=2 is inside the ellipse and z−21 is analytic everywhere else, By deformation of contours we can replace γ with a small circle around z=2:
∫γz−2dz=2πi.
Problem. Evaluate ∫γzezdz where γ is the square with vertices ±2±2i.
zez is analytic on and inside γ except at z=0. By deformation: ∫γzezdz=∫∣z∣=rzezdz=2πi⋅e0=2πi.
Problem. Evaluate ∫γz2−1dz where γ is ∣z∣=2.
z2−11=21(z−11−z+11).
Both z=±1 are inside ∣z∣=2.
∫γz2−1dz=21(2πi−2πi)=0.
6. Cauchy’s Integral Formula
6.1 Statement
Theorem 6.1 (Cauchy’s Integral Formula). If f is analytic on a connected domain Containing a simple closed positively oriented contour γAnd z0 is inside γThen
f(z0)=2πi1∫γz−z0f(z)dz
Proof. Let γε be a small circle of radius ε around z0. Since z−z0f(z) is analytic on the region between γ and γε
∫γz−z0f(z)dz=∫γεz−z0f(z)dz
On γε: f(z)=f(z0)+(z−z0)f′(ζ) for some ζ between z and z0.
=∫γεz−z0f(z0)dz+∫γεf′(ζ)dz=f(z0)⋅2πi+0
Since ∫γεz−z0dz=2πi (parameterize z=z0+εeiθ) and ∫γεf′(ζ)dz→0 as ε→0 by the ML inequality. ■
6.2 Derivatives
Theorem 6.2 (Cauchy’s Integral Formula for Derivatives). Under the same conditions,
f(n)(z0)=2πin!∫γ(z−z0)n+1f(z)dz
Proof. We proceed by induction. The base case n=0 is Theorem 6.1. For the inductive step, Assume the formula holds for n. Using the difference quotient:
Where we justified passing the limit inside the integral by uniform convergence of the integrand On compact subsets. ■
6.3 Consequences of Cauchy’s Integral Formula
Corollary 6.3. If f is analytic, then f is infinitely differentiable.
This is remarkable: a single complex derivative implies the existence of all derivatives.
Corollary 6.4 (Cauchy’s Estimates). If f is analytic on and inside a circle ∣z−z0∣=R And ∣f(z)∣≤M on the circle, then
∣f(n)(z0)∣≤Rnn!M
Proof. From the integral formula: ∣f(n)(z0)∣=2πn!∫∣z−z0∣=R(z−z0)n+1f(z)dz≤2πn!⋅Rn+1M⋅2πR=Rnn!M. ■
6.4 Liouville’s Theorem
Theorem 6.5 (Liouville’s Theorem). Every bounded entire function is constant.
Proof. If ∣f(z)∣≤M for all zThen by Cauchy’s estimates with R arbitrarily large: ∣f′(z0)∣≤RM→0 as R→∞. So f′(z)=0 for all zMeaning f is Constant. ■
Corollary 6.6. If f is entire and ∣f(z)∣≥M for all z (bounded away from zero), then f is constant.
Proof.1/f is entire and bounded by 1/MSo constant by Liouville. ■
6.5 Fundamental Theorem of Algebra
Theorem 6.7 (Fundamental Theorem of Algebra). Every non-constant polynomial p(z)∈C[z] has a root in C.
Proof. Suppose p(z) has no root. Then f(z)=1/p(z) is entire. Since ∣p(z)∣→∞ as ∣z∣→∞, f(z)→0So f is bounded. By Liouville’s theorem, f is constant, so p Is constant, a contradiction. ■
Corollary 6.8. Every polynomial of degree n≥1 has exactly n roots in C Counting multiplicities.
6.6 Worked Examples: Cauchy’s Integral Formula
Solution
Problem. Evaluate ∫γz−1ezdz where γ is ∣z∣=2.
Solution. The function z−1ez has a singularity at z=1Which lies inside γ. By Cauchy’s integral formula with f(z)=ez and z0=1:
∫γz−1ezdz=2πi⋅f(1)=2πi⋅e1=2πie. ■
Problem. Evaluate ∫γ(z−i)3z2+1dz where γ is ∣z∣=2.
By Cauchy’s formula for derivatives with f(z)=z2+1 and z0=i:
∫γ(z−i)3f(z)dz=2!2πif′′(i).
f′(z)=2z, f′′(z)=2. So f′′(i)=2.
∫γ(z−i)3z2+1dz=22πi⋅2=2πi. ■
Problem. Evaluate ∫γz(z−π)sinzdz where γ is ∣z∣=4.
Singularities inside γ: z=0 and z=π.
z(z−π)sinz=π1(z−πsinz−zsinz).
At z=0: by CIF, ∫γzsinzdz=2πi⋅sin(0)=0. At z=π: by CIF, ∫γz−πsinzdz=2πi⋅sin(π)=0.
∫γz(z−π)sinzdz=π1(0−0)=0.
Problem. Evaluate ∫γ(z−1)2(z+1)e2zdz where γ is ∣z∣=3.
By partial fractions: (z−1)2(z+1)1=z+11/4−z−11/4+(z−1)21/2.
Problem. Find the Laurent series of f(z)=z2(z−3)1 in 0<∣z∣<3.
z−31=−31∑n=0∞3nzn.
f(z)=−∑n=0∞3n+1zn−2=−3z21−9z1−271−81z−⋯
Residue at z=0: a−1=−91.
7.7 Residue at Infinity
Definition. The residue at infinity of f is defined as
Res(f,∞)=−2πi1∫∣z∣=Rf(z)dz
For sufficiently large R (enclosing all finite singularities).
Proposition 7.4. For a function f with finitely many singularities in C:
∑allfinitezkRes(f,zk)+Res(f,∞)=0
Proof. By the residue theorem applied to ∣z∣=R enclosing all finite singularities:
∫∣z∣=Rfdz=2πi∑finiteRes(f,zk).
But Res(f,∞)=−2πi1∫∣z∣=RfdzSo the sum is zero. ■
8. Singularities and Residue Theory
8.1 Isolated Singularities
Let z0 be an isolated singularity of f (i.e., f is analytic in a punctured neighbourhood of z0).
Classification by Laurent series:
Removable singularity: an=0 for all n<0. Can be removed by redefining f(z0)=a0.
Pole of order m: a−m=0 and an=0 for n<−m. The principal part is finite.
Essential singularity: infinitely many non-zero an with n<0.
Proposition 8.1 (Riemann’s Removable Singularity Theorem). If f is bounded near z0Then z0 is a removable singularity.
Proposition 8.2.z0 is a pole of order m if and only if limz→z0(z−z0)mf(z) Exists and is non-zero.
Theorem 8.3 (Casorati-Weierstrass). If z0 is an essential singularity of fThen f takes Values arbitrarily close to any complex number in every neighbourhood of z0.
8.2 Classification with Worked Examples
Solution
Problem. Classify the singularities of f(z)=zsinz.
z=0: sinz=z−z3/6+⋯So f(z)=1−z2/6+⋯. No negative powers, so z=0 is a removable singularity. f(0)=1 by continuity.
Problem. Classify the singularities of f(z)=z2ez−1.
z=0: ez−1=z+z2/2+⋯So f(z)=z1+21+⋯. Principal part is 1/zSo z=0 is a simple pole with residue 1.
Problem. Classify the singularity of f(z)=e1/z at z=0.
e1/z=∑n=0∞n!zn1=1+z1+2z21+⋯
Infinitely many negative powers ⇒z=0 is an essential singularity.
Problem. Classify the singularities of f(z)=z3(z2+1)z+1.
z=0: pole of order 3. z=i: simple pole. z=−i: simple pole.
Problem. Determine the type of singularity of f(z)=sinzz at z=0.
sinz=z−z3/6+⋯So f(z)=1−z2/6+⋯1=1+6z2+⋯.
No negative powers, so z=0 is a removable singularity with f(0)=1.
8.3 Definition of the Residue
Definition. The residue of f at an isolated singularity z0 is the coefficient a−1 In the Laurent expansion:
Res(f,z0)=a−1=2πi1∫γf(z)dz
Where γ is a small positively oriented circle around z0.
Problem. Find the residue of f(z)=z2+4z+3z at each pole.
z2+4z+3=(z+1)(z+3)So simple poles at z=−1 and z=−3.
At z=−1: Res=limz→−1z+3z=2−1. At z=−3: Res=limz→−3z+1z=−2−3=23.
Problem. Find the residue of f(z)=(z−1)2(z−2)ez at each pole.
At z=1 (pole of order 2): Res=dzd[z−2ez]z=1=(z−2)2ez(z−2)−ezz=1=1−e−e=−2e.
At z=2 (simple pole): Res=(2−1)2e2=e2.
8.5 The Residue Theorem
Theorem 8.4 (Residue Theorem). If f is analytic inside and on a simple closed positively Oriented contour γ except for isolated singularities z1,z2,…,zn inside γ Then
∫γf(z)dz=2πi∑k=1nRes(f,zk)
Proof. For each singularity zkDraw a small circle γk around it. By Cauchy’s theorem Applied to the multiply connected region between γ and the γk:
∫γfdz=∑k=1n∫γkfdz=∑k=1n2πi⋅Res(f,zk). ■
8.6 Worked Examples: Residue Theorem
Solution
Problem 1. Evaluate ∫γz(z−1)2ezdz where γ is ∣z∣=2.
Solution. Singularities inside γ: z=0 (simple pole) and z=1 (pole of order 2).
At z=0: Res=limz→0(z−1)2ez=(−1)21=1.
At z=1: Res(f,1)=dzd[(z−1)2⋅z(z−1)2ez]z=1=dzd[zez]z=1=z2ez⋅z−ezz=1=1e−e=0.
∫γfdz=2πi(1+0)=2πi. ■
Problem 2. Evaluate ∫γz4+11dz where γ is ∣z∣=2.
Solution. The poles are the fourth roots of −1: zk=eiπ/4+ikπ/2 for k=0,1,2,3. All four lie inside ∣z∣=2.
Each is a simple pole with Res(f,zk)=4zk31. Since zk4=−1: zk−3=−zkSo the sum equals −41∑zk=0.
∫γz4+1dz=2πi⋅0=0. ■
9. Applications of Contour Integration
9.1 Evaluation of Real Integrals
Contour integration is a powerful tool for evaluating definite integrals.
9.2 Integrals of Rational Functions over the Real Line
Theorem 9.1. If f(x)=P(x)/Q(x) where deg(Q)≥deg(P)+2 and Q has no real roots, Then
∫−∞∞f(x)dx=2πi∑Im(zk)>0Res(f,zk)
Where the sum is over poles in the upper half-plane.
Proof. Integrate f(z) over the semicircular contour γR consisting of [−R,R] on the Real axis and the semicircle ∣z∣=R in the upper half-plane. As R→∞The integral over The semicircle vanishes (since ∣f(z)∣≤M/R2 and the length is πR). ■
9.3 Worked Example
Problem. Evaluate ∫−∞∞x2+1dx.
Solution.f(z)=z2+11 has simple poles at z=±i.
Only z=i is in the upper half-plane.
Res(z2+11,i)=2z1z=i=2i1.
∫−∞∞x2+1dx=2πi⋅2i1=π. ■
9.4 Integrals Involving Trigonometric Functions
For integrals of the form ∫02πR(cosθ,sinθ)dθSubstitute z=eiθSo dz=izdθ, cosθ=2z+z−1sinθ=2iz−z−1.
The integral becomes ∫∣z∣=1f(z)dz where f(z) is a rational function.
Consider ∮γzeizdz where γ consists of [−R,−ε][ε,R] on the real axis, small upper semicircle Cε around 0And large Upper semicircle CR.
No poles inside the contour, so the integral is 0.
On CR: vanishes as R→∞ by Jordan’s lemma. On Cε (indenting above): ∫Cεzeizdz→−iπ as ε→0 (half residue contribution).
0=PV∫−∞∞xeixdx+(−iπ).
PV∫−∞∞xeixdx=iπ.
Taking imaginary parts: PV∫−∞∞xsinxdx=π.
10. Conformal Mappings
10.1 Definition
Definition. An analytic function f is conformal at z0 if f′(z0)=0. A conformal Mapping preserves angles (both magnitude and orientation) between curves.
10.2 Geometric Interpretation
If f′(z0)=reiθThen near z0 the mapping f acts as a rotation by θ followed By a scaling by r. The Jacobian determinant is ∣f′(z0)∣2>0So orientation is preserved.
10.3 Common Conformal Mappings
Mapping
Effect
w=az+b (a=0)
Translation, rotation, scaling
w=1/z
Inversion in the unit circle
w=z2
Squaring (doubles angles)
w=ez
Exponential (maps strips to sectors)
w=1−aˉzz−a
Möbius (maps disk to disk)
10.4 Möbius Transformations
A Möbius transformation (or linear fractional transformation) is
T(z)=cz+daz+b,ad−bc=0
Proposition 10.1. Möbius transformations are conformal (where defined) and map circles and lines To circles and lines.
Proposition 10.2. Three points determine a unique Möbius transformation: T(z1)=w1T(z2)=w2, T(z3)=w3.
10.5 Cross-Ratio
Definition. The cross-ratio of four distinct points z1,z2,z3,z4 is
Both multipliers are real and positive (not equal to 1), so T is hyperbolic.
10.7 The Riemann Mapping Theorem
Theorem 10.5 (Riemann Mapping Theorem). Let U be a connected open proper subset of C. Then there exists a bijective conformal map from U onto the unit disk D={z:∣z∣<1}.
This is one of the most profound results in complex analysis, establishing that all connected Domains (other than C itself) are conformally equivalent.
Remark. The Riemann mapping theorem is an existence theorem; it does not provide an explicit Formula for the conformal map .
11. Liouville’s Theorem and the Maximum Modulus Principle
11.1 Liouville’s Theorem
Theorem 11.1 (Liouville’s Theorem). Every bounded entire function is constant.
Proof. If ∣f(z)∣≤M for all zThen by Cauchy’s estimates with R arbitrarily large: ∣f′(z0)∣≤RM→0 as R→∞. So f′(z)=0 for all zMeaning f is Constant. ■
11.2 The Fundamental Theorem of Algebra
Theorem 11.2 (Fundamental Theorem of Algebra). Every non-constant polynomial p(z)∈C[z] has a root in C.
Proof. Suppose p(z) has no root. Then f(z)=1/p(z) is entire. Since ∣p(z)∣→∞ as ∣z∣→∞, f(z)→0So f is bounded. By Liouville’s theorem, f is constant, so p Is constant, a contradiction. ■
11.3 The Maximum Modulus Principle
Theorem 11.3 (Maximum Modulus Principle). If f is analytic and non-constant on a domain D Then ∣f∣ has no local maximum in D.
Corollary 11.4. If f is analytic on a bounded domain D and continuous on Dˉ=D∪∂DThen ∣f∣ attains its maximum on ∂D.
11.4 Minimum Modulus Principle
Theorem 11.5 (Minimum Modulus Principle). If f is analytic and non-zero on a bounded domain D And continuous on DˉThen ∣f∣ attains its minimum on ∂D.
Remark. If f has zeros in DThen ∣f∣ attains its minimum of 0 at those zeros. The minimum modulus principle requires the non-vanishing hypothesis.
11.5 Schwarz Lemma
Theorem 11.6 (Schwarz Lemma). If f:D→D is analytic with f(0)=0 Then
∣f(z)∣≤∣z∣forallz∈D
And ∣f′(0)∣≤1. Equality in either case implies f(z)=eiθz for some real θ.
Proof. Define g(z)=f(z)/z for z=0 and g(0)=f′(0). Then g is analytic on D. For ∣z∣=r<1: ∣g(z)∣=∣f(z)∣/∣z∣≤1/r. By the maximum modulus Principle, ∣g(z)∣≤1/r for ∣z∣≤r. Letting r→1: ∣g(z)∣≤1So ∣f(z)∣≤∣z∣. Also ∣f′(0)∣=∣g(0)∣≤1. If ∣f′(0)∣=1Then ∣g∣ attains its maximum At an interior point, so g is constant: g(z)=eiθ. ■
12. Argument Principle and Rouché’s Theorem
12.1 The Argument Principle
Theorem 12.1 (Argument Principle). If f is meromorphic inside and on a simple closed contour γ with no zeros or poles on γThen
2πi1∫γf(z)f′(z)dz=N−P
Where N is the number of zeros and P is the number of poles of f inside γ (counting Multiplicities).
12.2 Rouché’s Theorem
Theorem 12.2 (Rouché’s Theorem). If f and g are analytic inside and on a simple closed Contour γAnd ∣f(z)∣>∣g(z)∣ on γThen f and f+g have the same number of Zeros inside γ.
Proof. On γ: ∣g(z)/f(z)∣<1. The function h(z)=1+g(z)/f(z) satisfies ∣h(z)−1∣<1 on γSo h(γ) does not wind around 0. By the argument principle Applied to h: 0=Nh−PhMeaning h has the same number of zeros and poles inside γ. But h=(f+g)/fSo zeros of h are zeros of f+g and poles of h are zeros of f. Therefore f and f+g have the same number of zeros. ■
12.3 Worked Example
Problem. Show that z4+6z+3 has exactly one root in ∣z∣<1.
Solution. On ∣z∣=1: ∣6z∣=6>∣z4+3∣≤∣z∣4+3=4. By Rouché’s theorem with f(z)=6z and g(z)=z4+3: f+g=z4+6z+3 has the same number of zeros in ∣z∣<1 as f(z)=6zWhich has exactly one zero (at z=0). ■
Solution
Problem. Show that all roots of z4+z+1=0 satisfy ∣z∣<2.
On ∣z∣=2: ∣z4∣=16>∣z+1∣≤3. By Rouché with f(z)=z4 and g(z)=z+1: z4+z+1 has 4 zeros in ∣z∣<2 (same as z4).
Problem. Show that z5+3z2+1 has exactly two roots in ∣z∣<1.
On ∣z∣=1: ∣3z2+1∣≥∣3z2∣−∣1∣=2>∣z5∣=1. By Rouché with f(z)=3z2+1 and g(z)=z5: z5+3z2+1 has the same number of zeros as 3z2+1 in ∣z∣<1. 3z2+1=0⇒z=±i/3Both in ∣z∣<1. So 2 zeros.
13. Analytic Continuation
13.1 Definition
Definition. If f1 is analytic on D1 and f2 is analytic on D2 with D1∩D2=∅ and f1=f2 on D1∩D2Then f2 is an analytic Continuation of f1.
13.2 Identity Theorem
Theorem 13.1 (Identity Theorem). If f and g are analytic on a domain D and agree on a set With a limit point in DThen f=g on all of D.
Proof. Let E={z∈D:f(n)(z)=g(n)(z)foralln≥0}. E is Non-empty (it contains the limit point by continuity of derivatives). E is closed (by continuity). If z0∈EThe Taylor series of f and g at z0 coincide, so f=g in a neighbourhood of z0Giving E open. Since D is connected, E=D. ■
14. Common Pitfalls
:::caution Common Pitfall The Cauchy-Riemann equations are necessary but not sufficient for Differentiability. The partial derivatives must also be continuous. For example, f(z)=exp(−1/z4) extended by f(0)=0 satisfies the Cauchy-Riemann equations at the origin But is not differentiable there. :::
:::caution Common Pitfall Cauchy’s theorem requires a connected domain. On a multiply Connected domain, the integral of an analytic function around a closed contour may be non-zero. The Classic example is ∮∣z∣=1dz/z=2πi. :::
:::caution Common Pitfall When computing residues at poles of order m≥2The formula involves Differentiation. A common error is forgetting the (m−1)! in the denominator or differentiating (z−z0)mf(z) the wrong number of times. :::
:::caution Common Pitfall The residue at infinity is Res(f,∞)=−Res(1/z2⋅f(1/z),0). It is NOT f(∞). For A function that is analytic everywhere in the finite plane except for finitely many singularities, The sum of all residues (including the residue at infinity) is zero. :::
:::caution Common Pitfall A conformal mapping preserves angles but not necessarily distances. The Mapping w=z2 is conformal at every z=0But it doubles the angle between curves at each Point. At z=0It is not conformal because f′(0)=0. :::
:::caution Common Pitfall The maximum modulus principle says that ∣f∣ has no local maximum in the Interior, but the minimum can occur in the interior (e.g., f(z)=z on the unit disk has minimum ∣f∣=0 at z=0). For the minimum principle, one needs the additional hypothesis that f has No zeros in the domain. :::
:::caution Common Pitfall The complex logarithm is multi-valued. When a problem asks for “logarithm” without specifying a branch, you must either compute all values or explicitly state which Branch you are using. The principal branch Logz has a branch cut along (−∞,0] And is undefined on this cut. :::
:::caution Common Pitfall When applying the ML inequality, make sure M is a valid upper bound for ∣f(z)∣ on the entire contour. A common error is bounding ∣f∣ on only part of the contour. Also, L must be the arc length of the contour, not a diameter or radius. :::
15. Problem Set
Problem 1
Express z=−3+i in polar form and find all values of z1/3.
Solution
∣z∣=3+1=2. Since Re(z)<0 and Im(z)>0: arg(z)=π−π/6=5π/6.
If you get this wrong, revise: Section 3.4 (Harmonic Functions).
Problem 5
Evaluate ∫γ(z2+2z)dz where γ is the upper half of the unit circle from z=1 to z=−1.
Solution
Since z2+2z is entire, the integral is path-independent. Let F(z)=z3/3+z2.
∫γ(z2+2z)dz=F(−1)−F(1)=32−34=−32.
If you get this wrong, revise: Sections 4.5 and 4.7 (Contour Integrals).
Problem 6
Use the ML inequality to bound ∫γz−2ezdz where γ Is the circle ∣z∣=1.
Solution
On γ: ∣z∣=1So ∣ez∣≤e and ∣z−2∣≥1.
z−2ez≤e. L=2π.
∫γz−2ezdz≤2πe.
If you get this wrong, revise: Section 4.6 (ML Inequality).
Problem 7
Evaluate ∮γz2−zz+1dz where γ is ∣z∣=2.
Solution
z2−zz+1=z(z−1)z+1. Simple poles at z=0 and z=1Both inside ∣z∣=2.
At z=0: Res=limz→0z−1z+1=−1. At z=1: Res=limz→1zz+1=2.
∮γz2−zz+1dz=2πi(−1+2)=2πi.
If you get this wrong, revise: Sections 8.4 and 8.5 (Residues).
Problem 8
Classify all singularities of f(z)=z2+1e1/z and find all residues.
Solution
z=0: e1/z has an essential singularity at 0So z=0 is an essential singularity of f. z=i: simple pole. z=−i: simple pole.
At z=i: Res=2ie1/i=2ie−i. At z=−i: Res=−2ie1/(−i)=−2iei.
At z=0: find the coefficient of 1/z in z2+1e1/z. z2+11=1−z2+z4−⋯ near z=0. e1/z=1+1/z+1/(2z2)+⋯. The 1/z coefficient in the product: from 1⋅1/z=1/zGiving residue 1.
If you get this wrong, revise: Sections 8.1 and 8.4 (Singularities and Residues).
Holomorphic functions: complex differentiable on an open set; Cauchy-Riemann equations ux=vy, uy=−vx are necessary conditions.
Cauchy’s integral theorem: ∮γf(z)dz=0 for f holomorphic inside and on γ; Cauchy’s integral formula evaluates f(a) and all derivatives via contour integrals.
Residue theorem: ∮γf(z)dz=2πi∑Res(f,zk); residues computed via Laurent series coefficients.
Laurent series: generalises Taylor series to annular regions; classifies singularities as removable, poles, or essential.
Conformal mappings: holomorphic functions with non-zero derivative preserve angles; applications in fluid dynamics and electrostatics.