Skip to content

Complex Analysis

1. Complex Numbers Review

1.1 Definition and Arithmetic

A complex number is z=a+biz = a + bi where a,bRa, b \in \mathbb{R} and i2=1i^2 = -1. We call a=Re(z)a = \mathrm{Re}(z) the real part and b=Im(z)b = \mathrm{Im}(z) the imaginary part.

Arithmetic: (a+bi)+(c+di)=(a+c)+(b+d)i(a + bi) + (c + di) = (a + c) + (b + d)i and (a+bi)(c+di)=(acbd)+(ad+bc)i(a + bi)(c + di) = (ac - bd) + (ad + bc)i.

Proposition 1.1 (Properties of Complex Arithmetic). For all z,wCz, w \in \mathbb{C}:

  1. z+w=w+zz + w = w + z and zw=wzzw = wz (commutativity)
  2. (z+w)+u=z+(w+u)(z + w) + u = z + (w + u) and (zw)u=z(wu)(zw)u = z(wu) (associativity)
  3. z(w+u)=zw+zuz(w + u) = zw + zu (distributivity)
  4. There exist additive identity 00 and multiplicative identity 11.
  5. Every z0z \neq 0 has a multiplicative inverse 1z=zˉz2\frac{1}{z} = \frac{\bar{z}}{|z|^2}.

Remark. The complex field C\mathbb{C} cannot be ordered: there is no total ordering on C\mathbb{C} Compatible with the field operations. In particular, i2=1i^2 = -1 precludes any such ordering.

1.2 The Complex Conjugate and Modulus

Definition. The complex conjugate of z=a+biz = a + bi is zˉ=abi\bar{z} = a - bi.

Proposition 1.2. For all z,wCz, w \in \mathbb{C}:

  1. z+w=zˉ+wˉ\overline{z + w} = \bar{z} + \bar{w} and zw=zˉwˉ\overline{zw} = \bar{z}\bar{w}
  2. zzˉ=z2z\bar{z} = |z|^2
  3. z+zˉ=2Re(z)z + \bar{z} = 2\,\mathrm{Re}(z) and zzˉ=2iIm(z)z - \bar{z} = 2i\,\mathrm{Im}(z)
  4. zˉˉ=z\bar{\bar{z}} = z

Definition. The modulus (or absolute value) of z=a+biz = a + bi is z=a2+b2|z| = \sqrt{a^2 + b^2}.

Proposition 1.3 (Modulus Properties). For all z,wCz, w \in \mathbb{C}:

  1. z0|z| \geq 0 with equality iff z=0z = 0
  2. zw=zw|zw| = |z||w|
  3. z+wz+w|z + w| \leq |z| + |w| (triangle inequality)
  4. zwzw\bigl||z| - |w|\bigr| \leq |z - w| (reverse triangle inequality)

Proof of (3). z+w2=(z+w)(zˉ+wˉ)=z2+zwˉ+zˉw+w2=z2+2Re(zwˉ)+w2z2+2zw+w2=(z+w)2|z + w|^2 = (z + w)(\bar{z} + \bar{w}) = |z|^2 + z\bar{w} + \bar{z}w + |w|^2 = |z|^2 + 2\,\mathrm{Re}(z\bar{w}) + |w|^2 \leq |z|^2 + 2|z||w| + |w|^2 = (|z| + |w|)^2. The inequality follows from Re(zwˉ)zwˉ=zw\mathrm{Re}(z\bar{w}) \leq |z\bar{w}| = |z||w|. \blacksquare

1.3 Polar Form

Every non-zero complex number can be written in polar form:

z=r(cosθ+isinθ)=reiθz = r(\cos\theta + i\sin\theta) = re^{i\theta}

Where r=z=a2+b2r = |z| = \sqrt{a^2 + b^2} is the modulus and θ=arg(z)\theta = \arg(z) is the argument.

Definition. The principal argument Arg(z)\mathrm{Arg}(z) is the unique θ(π,π]\theta \in (-\pi, \pi] Such that z=zeiθz = |z|e^{i\theta}. The argument arg(z)\arg(z) is multi-valued: arg(z)=Arg(z)+2πk\arg(z) = \mathrm{Arg}(z) + 2\pi k for kZk \in \mathbb{Z}.

Proposition 1.4. If z1=r1eiθ1z_1 = r_1 e^{i\theta_1} and z2=r2eiθ2z_2 = r_2 e^{i\theta_2}Then z1z2=r1r2ei(θ1+θ2)z_1 z_2 = r_1 r_2 e^{i(\theta_1 + \theta_2)} and z1/z2=(r1/r2)ei(θ1θ2)z_1/z_2 = (r_1/r_2)\, e^{i(\theta_1 - \theta_2)}.

Worked Examples: Polar Form Conversions

Solution

Problem. Convert z=1+3iz = -1 + \sqrt{3}\,i to polar form and find all arguments.

z=(1)2+(3)2=1+3=2|z| = \sqrt{(-1)^2 + (\sqrt{3})^2} = \sqrt{1 + 3} = 2.

Re(z)=1<0\mathrm{Re}(z) = -1 \lt 0 and Im(z)=3>0\mathrm{Im}(z) = \sqrt{3} \gt 0So zz is in the second quadrant.

θ=arctan ⁣(31)=2π3\theta = \arctan\!\left(\frac{\sqrt{3}}{-1}\right) = \frac{2\pi}{3} (adjusting to second quadrant).

Polar form: z=2e2πi/3z = 2\,e^{2\pi i/3}.

All arguments: arg(z)=2π3+2πk\arg(z) = \frac{2\pi}{3} + 2\pi k for kZk \in \mathbb{Z}.

Problem. Convert z=3eiπ/4z = 3e^{-i\pi/4} to rectangular form.

z=3(cos ⁣(π4)+isin ⁣(π4))=3(22i22)=322322iz = 3\left(\cos\!\left(-\frac{\pi}{4}\right) + i\sin\!\left(-\frac{\pi}{4}\right)\right) = 3\left(\frac{\sqrt{2}}{2} - i\,\frac{\sqrt{2}}{2}\right) = \frac{3\sqrt{2}}{2} - \frac{3\sqrt{2}}{2}\,i.

Problem. Express z=34iz = -3 - 4i in polar form.

z=9+16=5|z| = \sqrt{9 + 16} = 5.

Both real and imaginary parts are negative, so zz is in the third quadrant.

θ=arctan(4/3)+π=π+arctan(4/3)\theta = \arctan(4/3) + \pi = \pi + \arctan(4/3).

z=5ei(π+arctan(4/3))z = 5\,e^{i(\pi + \arctan(4/3))}.

1.4 Euler”s Formula and De Moivre’s Theorem

Euler’s formula: eiθ=cosθ+isinθe^{i\theta} = \cos\theta + i\sin\theta.

De Moivre’s theorem: (eiθ)n=einθ(e^{i\theta})^n = e^{in\theta}So

(cosθ+isinθ)n=cos(nθ)+isin(nθ)(\cos\theta + i\sin\theta)^n = \cos(n\theta) + i\sin(n\theta)

Proposition 1.5. De Moivre’s theorem holds for all integers nnIncluding negative values.

Proof. For n0n \geq 0It follows by induction from the multiplication law eiαeiβ=ei(α+β)e^{i\alpha}e^{i\beta} = e^{i(\alpha + \beta)}. For n<0n \lt 0Write n=mn = -m with m>0m \gt 0: (cosθ+isinθ)n=1(cosθ+isinθ)m=1cos(mθ)+isin(mθ)=cos(mθ)+isin(mθ)=cos(nθ)+isin(nθ)(\cos\theta + i\sin\theta)^n = \frac{1}{(\cos\theta + i\sin\theta)^m} = \frac{1}{\cos(m\theta) + i\sin(m\theta)} = \cos(-m\theta) + i\sin(-m\theta) = \cos(n\theta) + i\sin(n\theta). \blacksquare

Applications of De Moivre’s Theorem

Example. Compute (1+i)20(1 + i)^{20}.

1+i=2eiπ/41 + i = \sqrt{2}\,e^{i\pi/4}So (1+i)20=(2)20e20πi/4=210e5πi=1024eπi=1024(1 + i)^{20} = (\sqrt{2})^{20}\, e^{20\pi i/4} = 2^{10}\, e^{5\pi i} = 1024\,e^{\pi i} = -1024.

Solution

Problem. Express cos(5θ)\cos(5\theta) in terms of cosθ\cos\theta using de Moivre.

By de Moivre: cos(5θ)+isin(5θ)=(cosθ+isinθ)5\cos(5\theta) + i\sin(5\theta) = (\cos\theta + i\sin\theta)^5.

Expanding the right side by the binomial theorem and equating real parts:

cos(5θ)=cos5θ10cos3θsin2θ+5cosθsin4θ\cos(5\theta) = \cos^5\theta - 10\cos^3\theta\sin^2\theta + 5\cos\theta\sin^4\theta.

Using sin2θ=1cos2θ\sin^2\theta = 1 - \cos^2\theta:

cos(5θ)=cos5θ10cos3θ(1cos2θ)+5cosθ(1cos2θ)2\cos(5\theta) = \cos^5\theta - 10\cos^3\theta(1 - \cos^2\theta) + 5\cos\theta(1 - \cos^2\theta)^2 =cos5θ10cos3θ+10cos5θ+5cosθ10cos3θ+5cos5θ= \cos^5\theta - 10\cos^3\theta + 10\cos^5\theta + 5\cos\theta - 10\cos^3\theta + 5\cos^5\theta =16cos5θ20cos3θ+5cosθ= 16\cos^5\theta - 20\cos^3\theta + 5\cos\theta.

Problem. Show that k=0n1cos(kθ)=sin(nθ/2)sin(θ/2)cos ⁣((n1)θ2)\sum_{k=0}^{n-1} \cos(k\theta) = \frac{\sin(n\theta/2)}{\sin(\theta/2)}\cos\!\left(\frac{(n-1)\theta}{2}\right) For θ2πZ\theta \notin 2\pi\mathbb{Z}.

Consider S=k=0n1eikθ=1einθ1eiθS = \sum_{k=0}^{n-1} e^{ik\theta} = \frac{1 - e^{in\theta}}{1 - e^{i\theta}} (geometric series with r=eiθ1r = e^{i\theta} \neq 1).

S=einθ/2(einθ/2einθ/2)eiθ/2(eiθ/2eiθ/2)=ei(n1)θ/2sin(nθ/2)sin(θ/2)S = \frac{e^{in\theta/2}(e^{-in\theta/2} - e^{in\theta/2})}{e^{i\theta/2}(e^{-i\theta/2} - e^{i\theta/2})} = e^{i(n-1)\theta/2} \cdot \frac{\sin(n\theta/2)}{\sin(\theta/2)}.

Taking real parts gives the result.

1.5 Roots of Complex Numbers

Definition. An nn-th root of wCw \in \mathbb{C} is a complex number zz such that zn=wz^n = w.

Proposition 1.6. Every non-zero wCw \in \mathbb{C} has exactly nn distinct nn-th roots. If w=ρeiϕw = \rho\, e^{i\phi}Then

zk=ρ1/nei(ϕ+2πk)/n,k=0,1,,n1z_k = \rho^{1/n}\, e^{i(\phi + 2\pi k)/n}, \quad k = 0, 1, \ldots, n - 1

Where ρ1/n>0\rho^{1/n} \gt 0 is the positive real nn-th root of ρ\rho.

Proof. If zn=wz^n = wWrite z=reiθz = r\,e^{i\theta}. Then rneinθ=ρeiϕr^n e^{in\theta} = \rho\, e^{i\phi} So r=ρ1/nr = \rho^{1/n} and nθ=ϕ+2πkn\theta = \phi + 2\pi k. For k=0,1,,n1k = 0, 1, \ldots, n-1 these give distinct Values of θ\theta; for knk \geq n they repeat. \blacksquare

Remark. The nn-th roots of ww lie equally spaced on a circle of radius ρ1/n\rho^{1/n}Forming a Regular nn-gon.

1.6 Roots of Unity

The nn-th roots of unity are the solutions of zn=1z^n = 1:

zk=e2πik/n,k=0,1,,n1z_k = e^{2\pi i k / n}, \quad k = 0, 1, \ldots, n - 1

They form a regular nn-gon on the unit circle in the complex plane.

Proposition 1.7. If ω=e2πi/n\omega = e^{2\pi i/n} is a primitive nn-th root of unity, then k=0n1ωk=0\sum_{k=0}^{n-1} \omega^k = 0 and k=0n1ωjk=0\sum_{k=0}^{n-1} \omega^{jk} = 0 for any jj not divisible by nn.

Proof. The sum k=0n1ωk=1ωn1ω=111ω=0\sum_{k=0}^{n-1} \omega^k = \frac{1 - \omega^n}{1 - \omega} = \frac{1 - 1}{1 - \omega} = 0 Provided ω1\omega \neq 1. For jj not divisible by nn, ωj\omega^j is a non-trivial root of unity, So the same argument applies. \blacksquare

Solution

Problem. Find all cube roots of 8-8.

8=8eiπ-8 = 8\,e^{i\pi}. The cube roots are: zk=81/3ei(π+2πk)/3=2ei(π+2πk)/3z_k = 8^{1/3}\, e^{i(\pi + 2\pi k)/3} = 2\, e^{i(\pi + 2\pi k)/3} for k=0,1,2k = 0, 1, 2.

z0=2eiπ/3=2(12+i32)=1+i3z_0 = 2\,e^{i\pi/3} = 2\left(\frac{1}{2} + i\,\frac{\sqrt{3}}{2}\right) = 1 + i\sqrt{3}. z1=2eiπ=2z_1 = 2\,e^{i\pi} = -2. z2=2ei5π/3=2(12i32)=1i3z_2 = 2\,e^{i5\pi/3} = 2\left(\frac{1}{2} - i\,\frac{\sqrt{3}}{2}\right) = 1 - i\sqrt{3}.

Problem. Find all fourth roots of z=16iz = 16i.

16i=16eiπ/216i = 16\,e^{i\pi/2}. The fourth roots are: zk=161/4ei(π/2+2πk)/4=2ei(π/8+πk/2)z_k = 16^{1/4}\, e^{i(\pi/2 + 2\pi k)/4} = 2\, e^{i(\pi/8 + \pi k/2)} for k=0,1,2,3k = 0, 1, 2, 3.

z0=2eiπ/8z_0 = 2\,e^{i\pi/8}, z1=2ei5π/8z_1 = 2\,e^{i5\pi/8}, z2=2ei9π/8z_2 = 2\,e^{i9\pi/8}, z3=2ei13π/8z_3 = 2\,e^{i13\pi/8}.

Problem. Show that the nn-th roots of any non-zero ww are in geometric progression.

The roots are zk=ρ1/nei(ϕ+2πk)/n=z0(e2πi/n)k=z0ωkz_k = \rho^{1/n}\, e^{i(\phi + 2\pi k)/n} = z_0 \cdot \left(e^{2\pi i/n}\right)^k = z_0 \cdot \omega^k Where ω=e2πi/n\omega = e^{2\pi i/n} is a primitive nn-th root of unity. This is a geometric sequence With ratio ω\omega.

2. Complex Functions and Analyticity

2.1 Complex Functions

A complex function is a function f:DCCf : D \subseteq \mathbb{C} \to \mathbb{C}. We can write f(z)=u(x,y)+iv(x,y)f(z) = u(x, y) + iv(x, y) where z=x+iyz = x + iy and u,vu, v are real-valued functions.

Example. f(z)=z2=(x+iy)2=(x2y2)+i(2xy)f(z) = z^2 = (x + iy)^2 = (x^2 - y^2) + i(2xy). Here u=x2y2u = x^2 - y^2 and v=2xyv = 2xy.

Example. f(z)=zˉ=xiyf(z) = \bar{z} = x - iy. Here u=xu = x and v=yv = -y.

Example. f(z)=z2=x2+y2f(z) = |z|^2 = x^2 + y^2. Here u=x2+y2u = x^2 + y^2 and v=0v = 0.

2.2 Limits and Continuity

The limit limzz0f(z)=L\lim_{z \to z_0} f(z) = L means: for every ε>0\varepsilon \gt 0There exists δ>0\delta \gt 0 Such that 0<zz0<δ0 \lt |z - z_0| \lt \delta implies f(z)L<ε|f(z) - L| \lt \varepsilon.

Unlike the real case, zz can approach z0z_0 from any direction in C\mathbb{C}. This makes limits More restrictive.

Proposition 2.1. limzz0f(z)=L\lim_{z \to z_0} f(z) = L if and only if lim(x,y)(x0,y0)u(x,y)=a\lim_{(x,y) \to (x_0, y_0)} u(x, y) = a And lim(x,y)(x0,y0)v(x,y)=b\lim_{(x,y) \to (x_0, y_0)} v(x, y) = b where L=a+biL = a + bi.

Definition. ff is continuous at z0z_0 if limzz0f(z)=f(z0)\lim_{z \to z_0} f(z) = f(z_0).

Solution

Problem. Show that limz0zˉz\lim_{z \to 0} \frac{\bar{z}}{z} does not exist.

Let z=reiθz = re^{i\theta}. Then zˉz=e2iθ\frac{\bar{z}}{z} = e^{-2i\theta}. As z0z \to 0 along different Rays (θ=0,π/2,π/4\theta = 0, \pi/2, \pi/4Etc.), the ratio takes different values (1,1,i1, -1, -iEtc.). Since the limit depends on the direction of approach, it does not exist.

Problem. Determine whether f(z)=z21z1f(z) = \frac{z^2 - 1}{z - 1} is continuous at z=1z = 1.

For z1z \neq 1: f(z)=z+1f(z) = z + 1. The limit as z1z \to 1 is 22But f(1)f(1) is undefined (division by zero). If we define f(1)=2f(1) = 2Then ff becomes continuous at z=1z = 1.

2.3 The Derivative

Definition. ff is differentiable at z0z_0 if

f(z0)=limh0f(z0+h)f(z0)hf'(z_0) = \lim_{h \to 0} \frac{f(z_0 + h) - f(z_0)}{h}

Exists (and is independent of how h0h \to 0 in C\mathbb{C}).

Remark. The requirement that the limit be the same for all directions of approach of hh is what Makes complex differentiability far more restrictive than real differentiability.

2.4 Analytic Functions

Definition. A function ff is analytic (or holomorphic) on an open set UCU \subseteq \mathbb{C} if ff is differentiable at every point of UU. A function that is analytic On all of C\mathbb{C} is called entire.

Examples of entire functions: znz^n, eze^z, sinz\sin z, cosz\cos zPolynomials.

Example of a non-analytic function: f(z)=zˉf(z) = \bar{z} is nowhere differentiable (except at z=0z = 0 if we define it, but still not analytic there).

Solution

Problem. Show that f(z)=z2f(z) = |z|^2 is differentiable only at z=0z = 0.

f(z)=x2+y2f(z) = x^2 + y^2So u=x2+y2u = x^2 + y^2 and v=0v = 0. ux=2xu_x = 2x, uy=2yu_y = 2y, vx=0v_x = 0, vy=0v_y = 0. The Cauchy-Riemann equations require 2x=02x = 0 and 2y=02y = 0So x=y=0x = y = 0. Thus ff satisfies CR only at z=0z = 0.

At z=0z = 0: f(0)=limh0h2h=limh0hˉ=0f'(0) = \lim_{h \to 0} \frac{|h|^2}{h} = \lim_{h \to 0} \bar{h} = 0So ff is Differentiable at 00 but not analytic anywhere (no neighbourhood of 00 is analytic).

Problem. Show that f(z)=zzˉ+zf(z) = z\bar{z} + z is differentiable only at z=0z = 0.

f(z)=z2+z=(x2+y2+x)+iyf(z) = |z|^2 + z = (x^2 + y^2 + x) + iy. ux=2x+1u_x = 2x + 1, uy=2yu_y = 2y, vx=0v_x = 0, vy=1v_y = 1. CR equations: 2x+1=1x=02x + 1 = 1 \Rightarrow x = 0And 2y=0y=02y = 0 \Rightarrow y = 0. At (0,0)(0, 0): f(0)=limh0hhˉ+hh=limh0(hˉ+1)=1f'(0) = \lim_{h \to 0} \frac{h\bar{h} + h}{h} = \lim_{h \to 0} (\bar{h} + 1) = 1. So ff is differentiable at z=0z = 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 ff is a single-valued analytic function gg Defined on a domain DD such that g(z)f(z)g(z) \in f(z) for all zDz \in D.

The Complex Logarithm. We define logz=lnz+iarg(z)\log z = \ln|z| + i\arg(z)Which is multi-valued because arg(z)=Arg(z)+2πk\arg(z) = \mathrm{Arg}(z) + 2\pi k for kZk \in \mathbb{Z}. The principal branch is

Logz=lnz+iArg(z)\mathrm{Log}\, z = \ln|z| + i\,\mathrm{Arg}(z)

Defined on C(,0]\mathbb{C} \setminus (-\infty, 0]. The negative real axis is called the branch cut.

Proposition 2.2. The principal branch Logz\mathrm{Log}\, z is analytic on C(,0]\mathbb{C} \setminus (-\infty, 0] and ddzLogz=1z\frac{d}{dz}\,\mathrm{Log}\, z = \frac{1}{z}.

Complex Powers. For z,αCz, \alpha \in \mathbb{C} with z0z \neq 0:

zα=eαlogzz^\alpha = e^{\alpha \log z}

This is multi-valued . When α\alpha is rational with reduced form p/qp/qThere are exactly qq distinct values.

Solution

Problem. Find all values of (1)i(-1)^i.

(1)i=eilog(1)=ei(iπ+2πik)=eπ2πk(-1)^i = e^{i \log(-1)} = e^{i(i\pi + 2\pi i k)} = e^{-\pi - 2\pi k} for kZk \in \mathbb{Z}.

These are all positive real numbers: ,e3π,eπ,eπ,e3π,\ldots, e^{3\pi}, e^{\pi}, e^{-\pi}, e^{-3\pi}, \ldots. The principal value (using the principal branch) is eπe^{-\pi}.

Problem. Find all values of i1/2i^{1/2}.

i1/2=e(1/2)logi=e(1/2)(iπ/2+2πik)=eiπ/4+iπki^{1/2} = e^{(1/2)\log i} = e^{(1/2)(i\pi/2 + 2\pi i k)} = e^{i\pi/4 + i\pi k}.

For k=0k = 0: eiπ/4=22(1+i)e^{i\pi/4} = \frac{\sqrt{2}}{2}(1 + i). For k=1k = 1: ei5π/4=22(1+i)e^{i5\pi/4} = -\frac{\sqrt{2}}{2}(1 + i). These are the two square roots of ii.

Problem. Find the domain of analyticity of f(z)=Log(z2+1)f(z) = \mathrm{Log}(z^2 + 1).

Logw\mathrm{Log}\, w is analytic on C(,0]\mathbb{C} \setminus (-\infty, 0]So we need z2+1(,0]z^2 + 1 \notin (-\infty, 0].

z2+10z^2 + 1 \leq 0 when z21z^2 \leq -1I.e., z[i,0][0,i]z \in [-i, 0] \cup [0, i] (the imaginary axis Segment from i-i to ii). Also z2+1=0z^2 + 1 = 0 at z=±iz = \pm i.

Domain: C{z:z=iy,y[1,1]}\mathbb{C} \setminus \{z : z = iy,\, y \in [-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)f(z) = u(x, y) + iv(x, y) is differentiable at z=x+iyz = x + iyThen

ux=vy,uy=vx\frac{\partial u}{\partial x} = \frac{\partial v}{\partial y}, \quad \frac{\partial u}{\partial y} = -\frac{\partial v}{\partial x}

Proof. Compute the limit along the real axis (hRh \in \mathbb{R}, h0h \to 0):

f(z)=limh0u(x+h,y)u(x,y)h+ilimh0v(x+h,y)v(x,y)h=ux+ivxf'(z) = \lim_{h \to 0} \frac{u(x+h, y) - u(x, y)}{h} + i\lim_{h \to 0} \frac{v(x+h, y) - v(x, y)}{h} = \frac{\partial u}{\partial x} + i\frac{\partial v}{\partial x}

Compute along the imaginary axis (h=ikh = ik, kRk \in \mathbb{R}, k0k \to 0):

f(z)=limk0u(x,y+k)u(x,y)ik+ilimk0v(x,y+k)v(x,y)ik=iuy+vyf'(z) = \lim_{k \to 0} \frac{u(x, y+k) - u(x, y)}{ik} + i\lim_{k \to 0} \frac{v(x, y+k) - v(x, y)}{ik} = -i\frac{\partial u}{\partial y} + \frac{\partial v}{\partial y}

Equating real and imaginary parts: ux=vy\frac{\partial u}{\partial x} = \frac{\partial v}{\partial y} And vx=uy\frac{\partial v}{\partial x} = -\frac{\partial u}{\partial y}. \blacksquare

3.2 Sufficiency Condition

Theorem 3.2. If uu and vv have continuous first partial derivatives on an open set UU and Satisfy the Cauchy-Riemann equations on UUThen f=u+ivf = u + iv is analytic on UU.

Proof. Since ux,uy,vx,vyu_x, u_y, v_x, v_y are continuous on UU, uu and vv are (real) differentiable. Let Δz=Δx+iΔy\Delta z = \Delta x + i\Delta y. By real differentiability:

u(x+Δx,y+Δy)u(x,y)=uxΔx+uyΔy+ε1u(x + \Delta x, y + \Delta y) - u(x, y) = u_x\,\Delta x + u_y\,\Delta y + \varepsilon_1 v(x+Δx,y+Δy)v(x,y)=vxΔx+vyΔy+ε2v(x + \Delta x, y + \Delta y) - v(x, y) = v_x\,\Delta x + v_y\,\Delta y + \varepsilon_2

Where ε1,ε2=o(Δz)\varepsilon_1, \varepsilon_2 = o(|\Delta z|). Therefore

f(z+Δz)f(z)Δz=(ux+ivx)Δx+(uy+ivy)Δy+ε1+iε2Δx+iΔy\frac{f(z + \Delta z) - f(z)}{\Delta z} = \frac{(u_x + iv_x)\Delta x + (u_y + iv_y)\Delta y + \varepsilon_1 + i\varepsilon_2}{\Delta x + i\Delta y}

By CR: uy+ivy=vx+iux=i(ux+ivx)u_y + iv_y = -v_x + iu_x = i(u_x + iv_x). Substituting:

=(ux+ivx)Δx+iΔyΔx+iΔy+o(Δz)Δzux+ivx= (u_x + iv_x)\frac{\Delta x + i\Delta y}{\Delta x + i\Delta y} + \frac{o(|\Delta z|)}{\Delta z} \to u_x + iv_x

As Δz0\Delta z \to 0. \blacksquare

3.3 The Derivative in Terms of Partial Derivatives

When the Cauchy-Riemann equations hold:

f(z)=ux+ivx=vyiuyf'(z) = \frac{\partial u}{\partial x} + i\frac{\partial v}{\partial x} = \frac{\partial v}{\partial y} - i\frac{\partial u}{\partial y}

3.4 Harmonic Functions

Definition. A real-valued function ϕ(x,y)\phi(x, y) is harmonic if ϕxx+ϕyy=0\phi_{xx} + \phi_{yy} = 0 (Laplace’s equation).

Proposition 3.3. If f=u+ivf = u + iv is analytic, then uu and vv are harmonic.

Proof. From the Cauchy-Riemann equations: ux=vyu_x = v_y and uy=vxu_y = -v_x. Differentiating: uxx=vyxu_{xx} = v_{yx} and uyy=vxyu_{yy} = -v_{xy}. By equality of mixed partials, uxx+uyy=vyxvxy=0u_{xx} + u_{yy} = v_{yx} - v_{xy} = 0. Similarly for vv. \blacksquare

Definition. If uu and vv are harmonic on UU and satisfy the Cauchy-Riemann equations, then vv is the harmonic conjugate of uu.

Proposition 3.4. If UU is a connected domain and uu is harmonic on UUThen uu has A harmonic conjugate on UUUnique up to an additive constant.

Proof. Define v(x,y)=(x0,y0)(x,y)(uydx+uxdy)v(x, y) = \int_{(x_0, y_0)}^{(x, y)} (-u_y\, dx + u_x\, dy). The integrand is closed (since (uy)y=uyy=uxx=(ux)x(-u_y)_y = -u_{yy} = u_{xx} = (u_x)_x) and since UU is Connected, vv is well-defined (path-independent) by Green’s theorem. Then vx=uyv_x = -u_y and vy=uxv_y = u_xWhich are the CR equations. \blacksquare

Solution

Problem. Find the harmonic conjugate of u(x,y)=x33xy2u(x, y) = x^3 - 3xy^2.

Verify uu is harmonic: uxx=6xu_{xx} = 6x, uyy=6xu_{yy} = -6xSo uxx+uyy=0u_{xx} + u_{yy} = 0. \checkmark

By CR: vy=ux=3x23y2v_y = u_x = 3x^2 - 3y^2So v=3x2yy3+g(x)v = 3x^2 y - y^3 + g(x). Also vx=uy=6xyv_x = -u_y = 6xySo 6xy=6xy+g(x)6xy = 6xy + g'(x)Giving g(x)=0g'(x) = 0So g(x)=Cg(x) = C.

Harmonic conjugate: v(x,y)=3x2yy3+Cv(x, y) = 3x^2 y - y^3 + C.

Note: f(z)=u+iv=x33xy2+i(3x2yy3)=(x+iy)3=z3f(z) = u + iv = x^3 - 3xy^2 + i(3x^2 y - y^3) = (x + iy)^3 = z^3.

Problem. Show that u(x,y)=ln(x2+y2)u(x, y) = \ln(x^2 + y^2) is harmonic on R2{0}\mathbb{R}^2 \setminus \{0\} but Has no harmonic conjugate on R2{0}\mathbb{R}^2 \setminus \{0\}.

ux=2xx2+y2u_x = \frac{2x}{x^2 + y^2}, uxx=2(y2x2)(x2+y2)2u_{xx} = \frac{2(y^2 - x^2)}{(x^2 + y^2)^2}. uy=2yx2+y2u_y = \frac{2y}{x^2 + y^2}, uyy=2(x2y2)(x2+y2)2u_{yy} = \frac{2(x^2 - y^2)}{(x^2 + y^2)^2}. Δu=0\Delta u = 0. \checkmark

However, z=1(uydx+uxdy)=z=1ydx+xdyx2+y2=02π1dθ=2π0\oint_{|z|=1} (-u_y\, dx + u_x\, dy) = \oint_{|z|=1} \frac{-y\, dx + x\, dy}{x^2 + y^2} = \int_0^{2\pi} 1\, d\theta = 2\pi \neq 0.

Since R2{0}\mathbb{R}^2 \setminus \{0\} is not connected and this integral is non-zero, no Harmonic conjugate exists on this domain.

3.5 Worked Examples: Verifying CR Equations

Solution

Problem. Verify that f(z)=ezf(z) = e^z satisfies the Cauchy-Riemann equations and find f(z)f'(z).

Solution. ez=ex+iy=ex(cosy+isiny)e^z = e^{x+iy} = e^x(\cos y + i\sin y). So u=excosyu = e^x \cos y and v=exsinyv = e^x \sin y.

ux=excosyu_x = e^x \cos y, uy=exsinyu_y = -e^x \sin y, vx=exsinyv_x = e^x \sin y, vy=excosyv_y = e^x \cos y.

Cauchy-Riemann: ux=excosy=vyu_x = e^x \cos y = v_y and uy=exsiny=vxu_y = -e^x \sin y = -v_x. Both satisfied.

f(z)=ux+ivx=excosy+iexsiny=ezf'(z) = u_x + iv_x = e^x \cos y + ie^x \sin y = e^z. \blacksquare

Problem. Verify CR for f(z)=sinzf(z) = \sin z and find f(z)f'(z).

sinz=sin(x+iy)=sinxcoshy+icosxsinhy\sin z = \sin(x + iy) = \sin x \cosh y + i\cos x \sinh y.

u=sinxcoshyu = \sin x \cosh y, v=cosxsinhyv = \cos x \sinh y.

ux=cosxcoshyu_x = \cos x \cosh y, uy=sinxsinhyu_y = \sin x \sinh y. vx=sinxsinhyv_x = -\sin x \sinh y, vy=cosxcoshyv_y = \cos x \cosh y.

CR: ux=cosxcoshy=vyu_x = \cos x \cosh y = v_y \checkmark and uy=sinxsinhy=vxu_y = \sin x \sinh y = -v_x \checkmark.

f(z)=ux+ivx=cosxcoshyisinxsinhy=coszf'(z) = u_x + iv_x = \cos x \cosh y - i\sin x \sinh y = \cos z. \blacksquare

Problem. Show f(z)=1zf(z) = \frac{1}{z} satisfies CR on C{0}\mathbb{C} \setminus \{0\}.

1z=zˉz2=xiyx2+y2\frac{1}{z} = \frac{\bar{z}}{|z|^2} = \frac{x - iy}{x^2 + y^2}.

u=xx2+y2u = \frac{x}{x^2 + y^2}, v=yx2+y2v = \frac{-y}{x^2 + y^2}.

ux=y2x2(x2+y2)2u_x = \frac{y^2 - x^2}{(x^2 + y^2)^2}, vy=y2x2(x2+y2)2v_y = \frac{y^2 - x^2}{(x^2 + y^2)^2}. So ux=vyu_x = v_y. \checkmark

uy=2xy(x2+y2)2u_y = \frac{-2xy}{(x^2 + y^2)^2}, vx=2xy(x2+y2)2v_x = \frac{2xy}{(x^2 + y^2)^2}. So uy=vxu_y = -v_x. \checkmark

f(z)=ux+ivx=(x2y2+2ixy)(x2+y2)2=1z2f'(z) = u_x + iv_x = \frac{-(x^2 - y^2 + 2ixy)}{(x^2 + y^2)^2} = \frac{-1}{z^2}. \blacksquare

4. Complex Integration

4.1 Contours

A contour (or piecewise smooth path) in C\mathbb{C} is a continuous function γ:[a,b]C\gamma : [a, b] \to \mathbb{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)\gamma(a) = \gamma(b) and no other Self-intersections.

4.2 The Complex Integral

Definition. For a contour γ\gamma and a continuous function ff on γ\gamma:

γf(z)dz=abf(γ(t))γ(t)dt\int_{\gamma} f(z)\, dz = \int_a^b f(\gamma(t))\gamma'(t)\, dt

4.3 Basic Properties

Proposition 4.1. The complex integral is linear:

γ(af+bg)dz=aγfdz+bγgdz\int_\gamma (af + bg)\, dz = a\int_\gamma f\, dz + b\int_\gamma g\, dz

Proposition 4.2. Reversing orientation changes the sign:

γfdz=γfdz\int_{-\gamma} f\, dz = -\int_\gamma f\, dz

Proposition 4.3. Additivity over contours:

γ1+γ2fdz=γ1fdz+γ2fdz\int_{\gamma_1 + \gamma_2} f\, dz = \int_{\gamma_1} f\, dz + \int_{\gamma_2} f\, dz

4.4 ML Inequality

Proposition 4.4 (ML Inequality). If f(z)M|f(z)| \leq M for all zz on a contour γ\gamma of length LLThen

γf(z)dzML\left|\int_\gamma f(z)\, dz\right| \leq ML

Proof. abf(γ(t))γ(t)dtabf(γ(t))γ(t)dtMabγ(t)dt=ML\left|\int_a^b f(\gamma(t))\gamma'(t)\, dt\right| \leq \int_a^b |f(\gamma(t))||\gamma'(t)|\, dt \leq M \int_a^b |\gamma'(t)|\, dt = ML. \blacksquare

4.5 Worked Examples: Contour Integrals

Solution

Problem. Evaluate γz2dz\int_\gamma z^2\, dz where γ\gamma is the line segment from 00 to 1+i1 + i.

Solution. Parameterize γ(t)=t(1+i)\gamma(t) = t(1 + i) for 0t10 \leq t \leq 1. Then γ(t)=1+i\gamma'(t) = 1 + i.

γz2dz=01(t(1+i))2(1+i)dt=(1+i)301t2dt=(1+i)313\int_\gamma z^2\, dz = \int_0^1 (t(1+i))^2 (1+i)\, dt = (1+i)^3 \int_0^1 t^2\, dt = (1+i)^3 \cdot \frac{1}{3}

(1+i)3=(1+i)(1+i)2=(1+i)(2i)=2i+2i2=2i2=2+2i(1+i)^3 = (1+i)(1+i)^2 = (1+i)(2i) = 2i + 2i^2 = 2i - 2 = -2 + 2i.

γz2dz=2+2i3\int_\gamma z^2\, dz = \frac{-2 + 2i}{3}. \blacksquare

Problem. Evaluate γzˉdz\int_\gamma \bar{z}\, dz where γ\gamma is the unit circle traversed once Counterclockwise.

γ(t)=eit\gamma(t) = e^{it}, 0t2π0 \leq t \leq 2\pi, γ(t)=ieit\gamma'(t) = ie^{it}. zˉ=eit\bar{z} = e^{-it} on γ\gamma.

γzˉdz=02πeitieitdt=02πidt=2πi\int_\gamma \bar{z}\, dz = \int_0^{2\pi} e^{-it} \cdot ie^{it}\, dt = \int_0^{2\pi} i\, dt = 2\pi i.

Note: Since zˉ\bar{z} is not analytic, this result is non-zero, as expected.

Problem. Evaluate γdzz\int_\gamma \frac{dz}{z} where γ\gamma is the upper semicircle z=eiθz = e^{i\theta}, 0θπ0 \leq \theta \leq \pi.

0πieiθeiθdθ=0πidθ=iπ\int_0^\pi \frac{ie^{i\theta}}{e^{i\theta}}\, d\theta = \int_0^\pi i\, d\theta = i\pi.

Problem. Evaluate γzdz\int_\gamma z\, dz where γ\gamma consists of the line segment from 00 to 11 followed by the line segment from 11 to 1+i1 + i.

γ1(t)=t\gamma_1(t) = t, 0t10 \leq t \leq 1: 01t1dt=12\int_0^1 t \cdot 1\, dt = \frac{1}{2}.

γ2(t)=1+it\gamma_2(t) = 1 + it, 0t10 \leq t \leq 1: 01(1+it)idt=01(it)dt=i12\int_0^1 (1 + it) \cdot i\, dt = \int_0^1 (i - t)\, dt = i - \frac{1}{2}.

Total: 12+i12=i\frac{1}{2} + i - \frac{1}{2} = i.

Check: Since zz is entire, the integral from 00 to 1+i1 + i is 12(1+i)2=i\frac{1}{2}(1+i)^2 = i. Consistent. \blacksquare

4.6 ML Inequality Applications

Solution

Problem. Use the ML inequality to show that limRCReizzdz=0\lim_{R \to \infty} \int_{C_R} \frac{e^{iz}}{z}\, dz = 0 Where CRC_R is the upper semicircle z=R|z| = R, Im(z)0\mathrm{Im}(z) \geq 0.

On CRC_R: z=Reiθz = Re^{i\theta}, 0θπ0 \leq \theta \leq \pi. eiz=eiR(cosθ+isinθ)=eRsinθ|e^{iz}| = |e^{iR(\cos\theta + i\sin\theta)}| = e^{-R\sin\theta}.

CReizzdz0πeRsinθRRdθ=0πeRsinθdθ\left|\int_{C_R} \frac{e^{iz}}{z}\, dz\right| \leq \int_0^\pi \frac{e^{-R\sin\theta}}{R} \cdot R\, d\theta = \int_0^\pi e^{-R\sin\theta}\, d\theta.

By Jordan’s inequality sinθ2θπ\sin\theta \geq \frac{2\theta}{\pi} for θ[0,π/2]\theta \in [0, \pi/2]:

20π/2e2Rθ/πdθ=πR(1eR)0\leq 2\int_0^{\pi/2} e^{-2R\theta/\pi}\, d\theta = \frac{\pi}{R}(1 - e^{-R}) \to 0 as RR \to \infty. \blacksquare

Problem. Bound γdzz2+4\left|\int_\gamma \frac{dz}{z^2 + 4}\right| where γ\gamma is z=3|z| = 3.

On γ\gamma: z2+4z24=94=5|z^2 + 4| \geq |z|^2 - 4 = 9 - 4 = 5 (reverse triangle inequality). So 1z2+415\left|\frac{1}{z^2 + 4}\right| \leq \frac{1}{5}.

Length of γ\gamma: L=2π3=6πL = 2\pi \cdot 3 = 6\pi.

γdzz2+4156π=6π5\left|\int_\gamma \frac{dz}{z^2 + 4}\right| \leq \frac{1}{5} \cdot 6\pi = \frac{6\pi}{5}.

4.7 Antiderivative Method

When ff is analytic on a connected domain and has a known antiderivative FF with F=fF' = f:

γf(z)dz=F(γ(b))F(γ(a))\int_\gamma f(z)\, dz = F(\gamma(b)) - F(\gamma(a))

This follows from the fundamental theorem of calculus applied to F(γ(t))F(\gamma(t)).

Solution

Problem. Evaluate γcoszdz\int_\gamma \cos z\, dz where γ\gamma is any path from 00 to π+i\pi + i.

Since cosz\cos z is entire with antiderivative sinz\sin z:

γcoszdz=sin(π+i)sin(0)=sin(π+i)\int_\gamma \cos z\, dz = \sin(\pi + i) - \sin(0) = \sin(\pi + i).

sin(π+i)=sinπcosh1+icosπsinh1=isinh1\sin(\pi + i) = \sin\pi\cosh 1 + i\cos\pi\sinh 1 = -i\sinh 1.

So the integral equals isinh1-i\sinh 1.

Problem. Evaluate γe2zdz\int_\gamma e^{2z}\, dz where γ\gamma is any path from 11 to ii.

Antiderivative: 12e2z\frac{1}{2}e^{2z}.

γe2zdz=12(e2ie2)\int_\gamma e^{2z}\, dz = \frac{1}{2}(e^{2i} - e^{2}).

5. Cauchy’s Theorem

5.1 Statement

Theorem 5.1 (Cauchy’s Theorem). If ff is analytic on a connected domain DD and γ\gamma Is a simple closed contour in DDThen

γf(z)dz=0\int_\gamma f(z)\, dz = 0

Proof (for ff' continuous). By Green’s theorem in the plane, writing f=u+ivf = u + iv:

γfdz=γ(udxvdy)+iγ(vdx+udy)\int_\gamma f\, dz = \int_\gamma (u\, dx - v\, dy) + i\int_\gamma (v\, dx + u\, dy)

Applying Green’s theorem to each integral:

=D(vxuy)dA+iD(uxvy)dA=0= \iint_D (-v_x - u_y)\, dA + i\iint_D (u_x - v_y)\, dA = 0

By the Cauchy-Riemann equations. \blacksquare

5.2 Connected Domains

A domain DCD \subseteq \mathbb{C} is ** connected** if every simple closed contour in DD can Be continuously shrunk to a point within DD.

Cauchy’s theorem may fail on multiply connected domains. For example, γ1zdz=2πi\int_\gamma \frac{1}{z}\, dz = 2\pi i where γ\gamma is the unit circle (traversing a region that Excludes the singularity at z=0z = 0).

5.3 Path Independence

Corollary 5.2. If ff is analytic on a connected domain DDThen the integral z0z1f(z)dz\int_{z_0}^{z_1} f(z)\, dz is independent of the path from z0z_0 to z1z_1 in DD.

5.4 Antiderivatives

Theorem 5.3. If ff is analytic on a connected domain DDThen ff has an antiderivative FF in DD (i.e., F(z)=f(z)F'(z) = f(z)), and

γf(z)dz=F(z1)F(z0)\int_\gamma f(z)\, dz = F(z_1) - F(z_0)

Where z0z_0 and z1z_1 are the endpoints of γ\gamma.

5.5 Cauchy’s Theorem for Multiply Connected Domains

Theorem 5.4. If ff is analytic on a domain DD containing simple closed contours γ,γ1,,γn\gamma, \gamma_1, \ldots, \gamma_n where γ1,,γn\gamma_1, \ldots, \gamma_n Lie in the interior of γ\gamma and the region between γ\gamma and the γk\gamma_k is contained in DD And all contours are positively oriented, then

γf(z)dz=k=1nγkf(z)dz\int_\gamma f(z)\, dz = \sum_{k=1}^n \int_{\gamma_k} f(z)\, dz

5.6 Deformation of Contours

Theorem 5.5 (Deformation of Contours). If ff is analytic on a domain containing two simple Closed contours γ1\gamma_1 and γ2\gamma_2 where one can be continuously deformed into the other Within the domain of analyticity of ffThen

γ1f(z)dz=γ2f(z)dz\int_{\gamma_1} f(z)\, dz = \int_{\gamma_2} f(z)\, dz

Proof. This follows directly from Theorem 5.4 applied to the region between γ1\gamma_1 and γ2\gamma_2. \blacksquare

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 γdzz2\int_\gamma \frac{dz}{z - 2} where γ\gamma is the ellipse x24+y29=1\frac{x^2}{4} + \frac{y^2}{9} = 1.

Since z=2z = 2 is inside the ellipse and 1z2\frac{1}{z - 2} is analytic everywhere else, By deformation of contours we can replace γ\gamma with a small circle around z=2z = 2:

γdzz2=2πi\int_\gamma \frac{dz}{z - 2} = 2\pi i.

Problem. Evaluate γezzdz\int_\gamma \frac{e^z}{z}\, dz where γ\gamma is the square with vertices ±2±2i\pm 2 \pm 2i.

ezz\frac{e^z}{z} is analytic on and inside γ\gamma except at z=0z = 0. By deformation: γezzdz=z=rezzdz=2πie0=2πi\int_\gamma \frac{e^z}{z}\, dz = \int_{|z|=r} \frac{e^z}{z}\, dz = 2\pi i \cdot e^0 = 2\pi i.

Problem. Evaluate γdzz21\int_\gamma \frac{dz}{z^2 - 1} where γ\gamma is z=2|z| = 2.

1z21=12(1z11z+1)\frac{1}{z^2 - 1} = \frac{1}{2}\left(\frac{1}{z-1} - \frac{1}{z+1}\right).

Both z=±1z = \pm 1 are inside z=2|z| = 2.

γdzz21=12(2πi2πi)=0\int_\gamma \frac{dz}{z^2 - 1} = \frac{1}{2}(2\pi i - 2\pi i) = 0.

6. Cauchy’s Integral Formula

6.1 Statement

Theorem 6.1 (Cauchy’s Integral Formula). If ff is analytic on a connected domain Containing a simple closed positively oriented contour γ\gammaAnd z0z_0 is inside γ\gammaThen

f(z0)=12πiγf(z)zz0dzf(z_0) = \frac{1}{2\pi i}\int_\gamma \frac{f(z)}{z - z_0}\, dz

Proof. Let γε\gamma_\varepsilon be a small circle of radius ε\varepsilon around z0z_0. Since f(z)zz0\frac{f(z)}{z - z_0} is analytic on the region between γ\gamma and γε\gamma_\varepsilon

γf(z)zz0dz=γεf(z)zz0dz\int_\gamma \frac{f(z)}{z - z_0}\, dz = \int_{\gamma_\varepsilon} \frac{f(z)}{z - z_0}\, dz

On γε\gamma_\varepsilon: f(z)=f(z0)+(zz0)f(ζ)f(z) = f(z_0) + (z - z_0)f'(\zeta) for some ζ\zeta between zz and z0z_0.

=γεf(z0)zz0dz+γεf(ζ)dz=f(z0)2πi+0= \int_{\gamma_\varepsilon} \frac{f(z_0)}{z - z_0}\, dz + \int_{\gamma_\varepsilon} f'(\zeta)\, dz = f(z_0) \cdot 2\pi i + 0

Since γεdzzz0=2πi\int_{\gamma_\varepsilon} \frac{dz}{z - z_0} = 2\pi i (parameterize z=z0+εeiθz = z_0 + \varepsilon e^{i\theta}) and γεf(ζ)dz0\int_{\gamma_\varepsilon} f'(\zeta)\, dz \to 0 as ε0\varepsilon \to 0 by the ML inequality. \blacksquare

6.2 Derivatives

Theorem 6.2 (Cauchy’s Integral Formula for Derivatives). Under the same conditions,

f(n)(z0)=n!2πiγf(z)(zz0)n+1dzf^{(n)}(z_0) = \frac{n!}{2\pi i}\int_\gamma \frac{f(z)}{(z - z_0)^{n+1}}\, dz

Proof. We proceed by induction. The base case n=0n = 0 is Theorem 6.1. For the inductive step, Assume the formula holds for nn. Using the difference quotient:

f(n+1)(z0)=limh0f(n)(z0+h)f(n)(z0)h=limh0n!2πiγ1h[f(z)(zz0h)n+1f(z)(zz0)n+1]dzf^{(n+1)}(z_0) = \lim_{h \to 0} \frac{f^{(n)}(z_0 + h) - f^{(n)}(z_0)}{h} = \lim_{h \to 0} \frac{n!}{2\pi i}\int_\gamma \frac{1}{h}\left[\frac{f(z)}{(z - z_0 - h)^{n+1}} - \frac{f(z)}{(z - z_0)^{n+1}}\right] dz

=(n+1)!2πiγf(z)(zz0)n+2dz= \frac{(n+1)!}{2\pi i}\int_\gamma \frac{f(z)}{(z - z_0)^{n+2}}\, dz

Where we justified passing the limit inside the integral by uniform convergence of the integrand On compact subsets. \blacksquare

6.3 Consequences of Cauchy’s Integral Formula

Corollary 6.3. If ff is analytic, then ff is infinitely differentiable.

This is remarkable: a single complex derivative implies the existence of all derivatives.

Corollary 6.4 (Cauchy’s Estimates). If ff is analytic on and inside a circle zz0=R|z - z_0| = R And f(z)M|f(z)| \leq M on the circle, then

f(n)(z0)n!MRn|f^{(n)}(z_0)| \leq \frac{n!M}{R^n}

Proof. From the integral formula: f(n)(z0)=n!2πzz0=Rf(z)(zz0)n+1dzn!2πMRn+12πR=n!MRn|f^{(n)}(z_0)| = \frac{n!}{2\pi}\left|\int_{|z-z_0|=R} \frac{f(z)}{(z-z_0)^{n+1}}\, dz\right| \leq \frac{n!}{2\pi} \cdot \frac{M}{R^{n+1}} \cdot 2\pi R = \frac{n!M}{R^n}. \blacksquare

6.4 Liouville’s Theorem

Theorem 6.5 (Liouville’s Theorem). Every bounded entire function is constant.

Proof. If f(z)M|f(z)| \leq M for all zzThen by Cauchy’s estimates with RR arbitrarily large: f(z0)MR0|f'(z_0)| \leq \frac{M}{R} \to 0 as RR \to \infty. So f(z)=0f'(z) = 0 for all zzMeaning ff is Constant. \blacksquare

Corollary 6.6. If ff is entire and f(z)M|f(z)| \geq M for all zz (bounded away from zero), then ff is constant.

Proof. 1/f1/f is entire and bounded by 1/M1/MSo constant by Liouville. \blacksquare

6.5 Fundamental Theorem of Algebra

Theorem 6.7 (Fundamental Theorem of Algebra). Every non-constant polynomial p(z)C[z]p(z) \in \mathbb{C}[z] has a root in C\mathbb{C}.

Proof. Suppose p(z)p(z) has no root. Then f(z)=1/p(z)f(z) = 1/p(z) is entire. Since p(z)|p(z)| \to \infty as z|z| \to \infty, f(z)0f(z) \to 0So ff is bounded. By Liouville’s theorem, ff is constant, so pp Is constant, a contradiction. \blacksquare

Corollary 6.8. Every polynomial of degree n1n \geq 1 has exactly nn roots in C\mathbb{C} Counting multiplicities.

6.6 Worked Examples: Cauchy’s Integral Formula

Solution

Problem. Evaluate γezz1dz\int_\gamma \frac{e^z}{z - 1}\, dz where γ\gamma is z=2|z| = 2.

Solution. The function ezz1\frac{e^z}{z - 1} has a singularity at z=1z = 1Which lies inside γ\gamma. By Cauchy’s integral formula with f(z)=ezf(z) = e^z and z0=1z_0 = 1:

γezz1dz=2πif(1)=2πie1=2πie\int_\gamma \frac{e^z}{z - 1}\, dz = 2\pi i \cdot f(1) = 2\pi i \cdot e^1 = 2\pi i e. \blacksquare

Problem. Evaluate γz2+1(zi)3dz\int_\gamma \frac{z^2 + 1}{(z - i)^3}\, dz where γ\gamma is z=2|z| = 2.

By Cauchy’s formula for derivatives with f(z)=z2+1f(z) = z^2 + 1 and z0=iz_0 = i:

γf(z)(zi)3dz=2πi2!f(i)\int_\gamma \frac{f(z)}{(z - i)^3}\, dz = \frac{2\pi i}{2!}\,f''(i).

f(z)=2zf'(z) = 2z, f(z)=2f''(z) = 2. So f(i)=2f''(i) = 2.

γz2+1(zi)3dz=2πi22=2πi\int_\gamma \frac{z^2 + 1}{(z - i)^3}\, dz = \frac{2\pi i}{2} \cdot 2 = 2\pi i. \blacksquare

Problem. Evaluate γsinzz(zπ)dz\int_\gamma \frac{\sin z}{z(z - \pi)}\, dz where γ\gamma is z=4|z| = 4.

Singularities inside γ\gamma: z=0z = 0 and z=πz = \pi.

sinzz(zπ)=1π(sinzzπsinzz)\frac{\sin z}{z(z - \pi)} = \frac{1}{\pi}\left(\frac{\sin z}{z - \pi} - \frac{\sin z}{z}\right).

At z=0z = 0: by CIF, γsinzzdz=2πisin(0)=0\int_\gamma \frac{\sin z}{z}\, dz = 2\pi i \cdot \sin(0) = 0. At z=πz = \pi: by CIF, γsinzzπdz=2πisin(π)=0\int_\gamma \frac{\sin z}{z - \pi}\, dz = 2\pi i \cdot \sin(\pi) = 0.

γsinzz(zπ)dz=1π(00)=0\int_\gamma \frac{\sin z}{z(z - \pi)}\, dz = \frac{1}{\pi}(0 - 0) = 0.

Problem. Evaluate γe2z(z1)2(z+1)dz\int_\gamma \frac{e^{2z}}{(z - 1)^2(z + 1)}\, dz where γ\gamma is z=3|z| = 3.

By partial fractions: 1(z1)2(z+1)=1/4z+11/4z1+1/2(z1)2\frac{1}{(z-1)^2(z+1)} = \frac{1/4}{z+1} - \frac{1/4}{z-1} + \frac{1/2}{(z-1)^2}.

γe2z(z1)2(z+1)dz=142πie2142πie2+122πi1!2e2\int_\gamma \frac{e^{2z}}{(z-1)^2(z+1)}\, dz = \frac{1}{4} \cdot 2\pi i \cdot e^{-2} - \frac{1}{4} \cdot 2\pi i \cdot e^2 + \frac{1}{2} \cdot \frac{2\pi i}{1!} \cdot 2e^2

=πie22πie22+2πie2=πie22+3πie22= \frac{\pi i e^{-2}}{2} - \frac{\pi i e^2}{2} + 2\pi i e^2 = \frac{\pi i e^{-2}}{2} + \frac{3\pi i e^2}{2}.

7. Taylor and Laurent Series

7.1 Taylor Series

Theorem 7.1. If ff is analytic on zz0<R|z - z_0| \lt RThen

f(z)=n=0f(n)(z0)n!(zz0)nf(z) = \sum_{n=0}^{\infty} \frac{f^{(n)}(z_0)}{n!}(z - z_0)^n

And the series converges uniformly on compact subsets of zz0<R|z - z_0| \lt R.

Proof. For zz0<r<R|z - z_0| \lt r \lt RApply Cauchy’s integral formula on ζz0=r|\zeta - z_0| = r:

f(z)=12πiζz0=rf(ζ)ζzdζf(z) = \frac{1}{2\pi i}\int_{|\zeta - z_0| = r} \frac{f(\zeta)}{\zeta - z}\, d\zeta

Write 1ζz=1(ζz0)(zz0)=1ζz011(zz0)/(ζz0)\frac{1}{\zeta - z} = \frac{1}{(\zeta - z_0) - (z - z_0)} = \frac{1}{\zeta - z_0} \cdot \frac{1}{1 - (z - z_0)/(\zeta - z_0)} =n=0(zz0)n(ζz0)n+1= \sum_{n=0}^{\infty} \frac{(z - z_0)^n}{(\zeta - z_0)^{n+1}} (geometric series, convergent since zz0/ζz0<1|z - z_0|/|\zeta - z_0| \lt 1).

Substituting and integrating term by term gives the Taylor series. \blacksquare

Remark. The radius of convergence RR is the distance from z0z_0 to the nearest singularity of ff.

7.2 Common Taylor Series

ez=n=0znn!=1+z+z22!+e^z = \sum_{n=0}^{\infty} \frac{z^n}{n!} = 1 + z + \frac{z^2}{2!} + \cdots

sinz=n=0(1)nz2n+1(2n+1)!\sin z = \sum_{n=0}^{\infty} \frac{(-1)^n z^{2n+1}}{(2n+1)!}

cosz=n=0(1)nz2n(2n)!\cos z = \sum_{n=0}^{\infty} \frac{(-1)^n z^{2n}}{(2n)!}

11z=n=0zn,z<1\frac{1}{1 - z} = \sum_{n=0}^{\infty} z^n, \quad |z| \lt 1

ln(1+z)=n=1(1)n+1znn,z<1\ln(1 + z) = \sum_{n=1}^{\infty} \frac{(-1)^{n+1} z^n}{n}, \quad |z| \lt 1

7.3 Worked Examples: Taylor Series

Solution

Problem. Find the Taylor series of f(z)=1zf(z) = \frac{1}{z} centered at z0=1z_0 = 1.

1z=11+(z1)=n=0(1)n(z1)n\frac{1}{z} = \frac{1}{1 + (z - 1)} = \sum_{n=0}^{\infty} (-1)^n (z - 1)^n for z1<1|z - 1| \lt 1.

Radius of convergence: distance from z0=1z_0 = 1 to the singularity at z=0z = 0Which is 11.

Problem. Find the Taylor series of f(z)=1(1z)2f(z) = \frac{1}{(1 - z)^2} centered at z0=0z_0 = 0.

1(1z)2=ddz[11z]=ddzn=0zn=n=1nzn1=n=0(n+1)zn\frac{1}{(1-z)^2} = \frac{d}{dz}\left[\frac{1}{1 - z}\right] = \frac{d}{dz}\sum_{n=0}^{\infty} z^n = \sum_{n=1}^{\infty} nz^{n-1} = \sum_{n=0}^{\infty} (n+1)z^n for z<1|z| \lt 1.

Problem. Find the Taylor series of f(z)=ezsinzf(z) = e^z \sin z up to the z4z^4 term.

ez=1+z+z2/2+z3/6+z4/24+e^z = 1 + z + z^2/2 + z^3/6 + z^4/24 + \cdots sinz=zz3/6+z5/120\sin z = z - z^3/6 + z^5/120 - \cdots

ezsinz=(1+z+z2/2+z3/6+z4/24+)(zz3/6+)e^z \sin z = (1 + z + z^2/2 + z^3/6 + z^4/24 + \cdots)(z - z^3/6 + \cdots)

=z+z2+z3/2+z4/6+z3/6z4/6+= z + z^2 + z^3/2 + z^4/6 + \cdots - z^3/6 - z^4/6 + \cdots =z+z2+z3/3z4/30+= z + z^2 + z^3/3 - z^4/30 + \cdots

7.4 Laurent Series

Theorem 7.2 (Laurent Series). If ff is analytic on the annulus r<zz0<Rr \lt |z - z_0| \lt R Then

f(z)=n=an(zz0)n=+a2(zz0)2+a1zz0+a0+a1(zz0)+f(z) = \sum_{n=-\infty}^{\infty} a_n(z - z_0)^n = \cdots + \frac{a_{-2}}{(z - z_0)^2} + \frac{a_{-1}}{z - z_0} + a_0 + a_1(z - z_0) + \cdots

Where

an=12πiγf(z)(zz0)n+1dza_n = \frac{1}{2\pi i}\int_\gamma \frac{f(z)}{(z - z_0)^{n+1}}\, dz

For any simple closed contour γ\gamma in the annulus encircling z0z_0.

The principal part is n=1an(zz0)n\sum_{n=-\infty}^{-1} a_n(z - z_0)^n (negative powers). The analytic Part is n=0an(zz0)n\sum_{n=0}^{\infty} a_n(z - z_0)^n (non-negative powers).

7.5 Classification of Laurent Series

The Laurent series expansion depends on the annulus of convergence. A function may have different Laurent expansions in different annuli.

Proposition 7.3. The Laurent series expansion of ff in an annulus is unique.

7.6 Worked Examples: Laurent Series

Solution

Problem. Find the Laurent series of f(z)=1z(z1)f(z) = \frac{1}{z(z-1)} in 0<z<10 \lt |z| \lt 1.

Solution. Using partial fractions: 1z(z1)=1z11z\frac{1}{z(z-1)} = \frac{1}{z-1} - \frac{1}{z}.

In z<1|z| \lt 1: 1z1=11z=n=0zn\frac{1}{z - 1} = -\frac{1}{1 - z} = -\sum_{n=0}^{\infty} z^n.

So f(z)=n=0zn1z=z2z11zf(z) = -\sum_{n=0}^{\infty} z^n - \frac{1}{z} = \cdots - z^2 - z - 1 - \frac{1}{z}.

The principal part is 1/z-1/zSo z=0z = 0 is a simple pole. \blacksquare

Problem. Find the Laurent series of f(z)=1z(z1)f(z) = \frac{1}{z(z-1)} in 1<z<1 \lt |z| \lt \infty.

In z>1|z| \gt 1: 1z1=1z111/z=n=2zn\frac{1}{z - 1} = \frac{1}{z} \cdot \frac{1}{1 - 1/z} = \sum_{n=2}^{\infty} z^{-n}.

f(z)=n=2zn1z=1z2+1z3+f(z) = \sum_{n=2}^{\infty} z^{-n} - \frac{1}{z} = \frac{1}{z^2} + \frac{1}{z^3} + \cdots

Problem. Find the Laurent series of f(z)=ezz2f(z) = \frac{e^z}{z^2} in 0<z<0 \lt |z| \lt \infty.

ez=n=0znn!e^z = \sum_{n=0}^{\infty} \frac{z^n}{n!}So f(z)=n=0zn2n!=1z2+1z+12+z6+f(z) = \sum_{n=0}^{\infty} \frac{z^{n-2}}{n!} = \frac{1}{z^2} + \frac{1}{z} + \frac{1}{2} + \frac{z}{6} + \cdots

Residue at z=0z = 0: a1=1a_{-1} = 1.

Problem. Find the Laurent series of f(z)=1z2(z3)f(z) = \frac{1}{z^2(z - 3)} in 0<z<30 \lt |z| \lt 3.

1z3=13n=0zn3n\frac{1}{z - 3} = -\frac{1}{3}\sum_{n=0}^{\infty} \frac{z^n}{3^n}.

f(z)=n=0zn23n+1=13z219z127z81f(z) = -\sum_{n=0}^{\infty} \frac{z^{n-2}}{3^{n+1}} = -\frac{1}{3z^2} - \frac{1}{9z} - \frac{1}{27} - \frac{z}{81} - \cdots

Residue at z=0z = 0: a1=19a_{-1} = -\frac{1}{9}.

7.7 Residue at Infinity

Definition. The residue at infinity of ff is defined as

Res(f,)=12πiz=Rf(z)dz\mathrm{Res}(f, \infty) = -\frac{1}{2\pi i}\int_{|z|=R} f(z)\, dz

For sufficiently large RR (enclosing all finite singularities).

Proposition 7.4. For a function ff with finitely many singularities in C\mathbb{C}:

all finite zkRes(f,zk)+Res(f,)=0\sum_{\mathrm{all\ finite\ } z_k} \mathrm{Res}(f, z_k) + \mathrm{Res}(f, \infty) = 0

Proof. By the residue theorem applied to z=R|z| = R enclosing all finite singularities:

z=Rfdz=2πifiniteRes(f,zk)\int_{|z|=R} f\, dz = 2\pi i \sum_{\mathrm{finite} \mathrm{Res}(f, z_k)}.

But Res(f,)=12πiz=Rfdz\mathrm{Res}(f, \infty) = -\frac{1}{2\pi i}\int_{|z|=R} f\, dzSo the sum is zero. \blacksquare

8. Singularities and Residue Theory

8.1 Isolated Singularities

Let z0z_0 be an isolated singularity of ff (i.e., ff is analytic in a punctured neighbourhood of z0z_0).

Classification by Laurent series:

  1. Removable singularity: an=0a_n = 0 for all n<0n \lt 0. Can be removed by redefining f(z0)=a0f(z_0) = a_0.
  2. Pole of order mm: am0a_{-m} \neq 0 and an=0a_n = 0 for n<mn \lt -m. The principal part is finite.
  3. Essential singularity: infinitely many non-zero ana_n with n<0n \lt 0.

Proposition 8.1 (Riemann’s Removable Singularity Theorem). If ff is bounded near z0z_0Then z0z_0 is a removable singularity.

Proposition 8.2. z0z_0 is a pole of order mm if and only if limzz0(zz0)mf(z)\lim_{z \to z_0} (z - z_0)^m f(z) Exists and is non-zero.

Theorem 8.3 (Casorati-Weierstrass). If z0z_0 is an essential singularity of ffThen ff takes Values arbitrarily close to any complex number in every neighbourhood of z0z_0.

8.2 Classification with Worked Examples

Solution

Problem. Classify the singularities of f(z)=sinzzf(z) = \frac{\sin z}{z}.

z=0z = 0: sinz=zz3/6+\sin z = z - z^3/6 + \cdotsSo f(z)=1z2/6+f(z) = 1 - z^2/6 + \cdots. No negative powers, so z=0z = 0 is a removable singularity. f(0)=1f(0) = 1 by continuity.

Problem. Classify the singularities of f(z)=ez1z2f(z) = \frac{e^z - 1}{z^2}.

z=0z = 0: ez1=z+z2/2+e^z - 1 = z + z^2/2 + \cdotsSo f(z)=1z+12+f(z) = \frac{1}{z} + \frac{1}{2} + \cdots. Principal part is 1/z1/zSo z=0z = 0 is a simple pole with residue 11.

Problem. Classify the singularity of f(z)=e1/zf(z) = e^{1/z} at z=0z = 0.

e1/z=n=01n!zn=1+1z+12z2+e^{1/z} = \sum_{n=0}^{\infty} \frac{1}{n!\, z^n} = 1 + \frac{1}{z} + \frac{1}{2z^2} + \cdots

Infinitely many negative powers \Rightarrow z=0z = 0 is an essential singularity.

Problem. Classify the singularities of f(z)=z+1z3(z2+1)f(z) = \frac{z + 1}{z^3(z^2 + 1)}.

z=0z = 0: pole of order 33. z=iz = i: simple pole. z=iz = -i: simple pole.

Problem. Determine the type of singularity of f(z)=zsinzf(z) = \frac{z}{\sin z} at z=0z = 0.

sinz=zz3/6+\sin z = z - z^3/6 + \cdotsSo f(z)=11z2/6+=1+z26+f(z) = \frac{1}{1 - z^2/6 + \cdots} = 1 + \frac{z^2}{6} + \cdots.

No negative powers, so z=0z = 0 is a removable singularity with f(0)=1f(0) = 1.

8.3 Definition of the Residue

Definition. The residue of ff at an isolated singularity z0z_0 is the coefficient a1a_{-1} In the Laurent expansion:

Res(f,z0)=a1=12πiγf(z)dz\mathrm{Res}(f, z_0) = a_{-1} = \frac{1}{2\pi i}\int_\gamma f(z)\, dz

Where γ\gamma is a small positively oriented circle around z0z_0.

8.4 Computing Residues

For a simple pole at z0z_0:

Res(f,z0)=limzz0(zz0)f(z)\mathrm{Res}(f, z_0) = \lim_{z \to z_0} (z - z_0)f(z)

If f=g/hf = g/h where g(z0)0g(z_0) \neq 0, h(z0)=0h(z_0) = 0, h(z0)0h'(z_0) \neq 0:

Res(f,z0)=g(z0)h(z0)\mathrm{Res}(f, z_0) = \frac{g(z_0)}{h'(z_0)}

For a pole of order mm at z0z_0:

Res(f,z0)=1(m1)!limzz0dm1dzm1[(zz0)mf(z)]\mathrm{Res}(f, z_0) = \frac{1}{(m-1)!}\lim_{z \to z_0} \frac{d^{m-1}}{dz^{m-1}}\left[(z - z_0)^m f(z)\right]

Solution

Problem. Find the residue of f(z)=zz2+4z+3f(z) = \frac{z}{z^2 + 4z + 3} at each pole.

z2+4z+3=(z+1)(z+3)z^2 + 4z + 3 = (z + 1)(z + 3)So simple poles at z=1z = -1 and z=3z = -3.

At z=1z = -1: Res=limz1zz+3=12\mathrm{Res} = \lim_{z \to -1} \frac{z}{z + 3} = \frac{-1}{2}. At z=3z = -3: Res=limz3zz+1=32=32\mathrm{Res} = \lim_{z \to -3} \frac{z}{z + 1} = \frac{-3}{-2} = \frac{3}{2}.

Problem. Find the residue of f(z)=ez(z1)2(z2)f(z) = \frac{e^z}{(z - 1)^2(z - 2)} at each pole.

At z=1z = 1 (pole of order 22): Res=ddz[ezz2]z=1=ez(z2)ez(z2)2z=1=ee1=2e\mathrm{Res} = \frac{d}{dz}\left[\frac{e^z}{z - 2}\right]_{z=1} = \frac{e^z(z - 2) - e^z}{(z-2)^2}\Big|_{z=1} = \frac{-e - e}{1} = -2e.

At z=2z = 2 (simple pole): Res=e2(21)2=e2\mathrm{Res} = \frac{e^2}{(2-1)^2} = e^2.

8.5 The Residue Theorem

Theorem 8.4 (Residue Theorem). If ff is analytic inside and on a simple closed positively Oriented contour γ\gamma except for isolated singularities z1,z2,,znz_1, z_2, \ldots, z_n inside γ\gamma Then

γf(z)dz=2πik=1nRes(f,zk)\int_\gamma f(z)\, dz = 2\pi i \sum_{k=1}^{n} \mathrm{Res}(f, z_k)

Proof. For each singularity zkz_kDraw a small circle γk\gamma_k around it. By Cauchy’s theorem Applied to the multiply connected region between γ\gamma and the γk\gamma_k:

γfdz=k=1nγkfdz=k=1n2πiRes(f,zk)\int_\gamma f\, dz = \sum_{k=1}^n \int_{\gamma_k} f\, dz = \sum_{k=1}^n 2\pi i \cdot \mathrm{Res}(f, z_k). \blacksquare

8.6 Worked Examples: Residue Theorem

Solution

Problem 1. Evaluate γezz(z1)2dz\int_\gamma \frac{e^z}{z(z-1)^2}\, dz where γ\gamma is z=2|z| = 2.

Solution. Singularities inside γ\gamma: z=0z = 0 (simple pole) and z=1z = 1 (pole of order 22).

At z=0z = 0: Res=limz0ez(z1)2=1(1)2=1\mathrm{Res} = \lim_{z \to 0} \frac{e^z}{(z-1)^2} = \frac{1}{(-1)^2} = 1.

At z=1z = 1: Res(f,1)=ddz[(z1)2ezz(z1)2]z=1=ddz[ezz]z=1=ezzezz2z=1=ee1=0\mathrm{Res}(f, 1) = \frac{d}{dz}\left[(z-1)^2 \cdot \frac{e^z}{z(z-1)^2}\right]_{z=1} = \frac{d}{dz}\left[\frac{e^z}{z}\right]_{z=1} = \frac{e^z \cdot z - e^z}{z^2}\Big|_{z=1} = \frac{e - e}{1} = 0.

γfdz=2πi(1+0)=2πi\int_\gamma f\, dz = 2\pi i(1 + 0) = 2\pi i. \blacksquare

Problem 2. Evaluate γ1z4+1dz\int_\gamma \frac{1}{z^4 + 1}\, dz where γ\gamma is z=2|z| = 2.

Solution. The poles are the fourth roots of 1-1: zk=eiπ/4+ikπ/2z_k = e^{i\pi/4 + ik\pi/2} for k=0,1,2,3k = 0, 1, 2, 3. All four lie inside z=2|z| = 2.

Each is a simple pole with Res(f,zk)=14zk3\mathrm{Res}(f, z_k) = \frac{1}{4z_k^3}. Since zk4=1z_k^4 = -1: zk3=zkz_k^{-3} = -z_kSo the sum equals 14zk=0-\frac{1}{4}\sum z_k = 0.

γdzz4+1=2πi0=0\int_\gamma \frac{dz}{z^4 + 1} = 2\pi i \cdot 0 = 0. \blacksquare

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)f(x) = P(x)/Q(x) where deg(Q)deg(P)+2\deg(Q) \geq \deg(P) + 2 and QQ has no real roots, Then

f(x)dx=2πiIm(zk)>0Res(f,zk)\int_{-\infty}^{\infty} f(x)\, dx = 2\pi i \sum_{\mathrm{Im}(z_k) > 0} \mathrm{Res}(f, z_k)

Where the sum is over poles in the upper half-plane.

Proof. Integrate f(z)f(z) over the semicircular contour γR\gamma_R consisting of [R,R][-R, R] on the Real axis and the semicircle z=R|z| = R in the upper half-plane. As RR \to \inftyThe integral over The semicircle vanishes (since f(z)M/R2|f(z)| \leq M/R^2 and the length is πR\pi R). \blacksquare

9.3 Worked Example

Problem. Evaluate dxx2+1\int_{-\infty}^{\infty} \frac{dx}{x^2 + 1}.

Solution. f(z)=1z2+1f(z) = \frac{1}{z^2 + 1} has simple poles at z=±iz = \pm i.

Only z=iz = i is in the upper half-plane.

Res(1z2+1,i)=12zz=i=12i\mathrm{Res}\left(\frac{1}{z^2 + 1}, i\right) = \frac{1}{2z}\Big|_{z = i} = \frac{1}{2i}.

dxx2+1=2πi12i=π\int_{-\infty}^{\infty} \frac{dx}{x^2 + 1} = 2\pi i \cdot \frac{1}{2i} = \pi. \blacksquare

9.4 Integrals Involving Trigonometric Functions

For integrals of the form 02πR(cosθ,sinθ)dθ\int_0^{2\pi} R(\cos\theta, \sin\theta)\, d\thetaSubstitute z=eiθz = e^{i\theta}So dz=izdθdz = iz\, d\theta, cosθ=z+z12\cos\theta = \frac{z + z^{-1}}{2} sinθ=zz12i\sin\theta = \frac{z - z^{-1}}{2i}.

The integral becomes z=1f(z)dz\int_{|z|=1} f(z)\, dz where f(z)f(z) is a rational function.

9.5 Worked Example

Problem. Evaluate 02πdθ2+cosθ\int_0^{2\pi} \frac{d\theta}{2 + \cos\theta}.

Solution. Substitute z=eiθz = e^{i\theta}: dθ=dzizd\theta = \frac{dz}{iz} cosθ=z+1/z2\cos\theta = \frac{z + 1/z}{2}.

z=1dziz(2+z+1/z2)=z=12dzi(z2+4z+1)\int_{|z|=1} \frac{dz}{iz\left(2 + \frac{z + 1/z}{2}\right)} = \int_{|z|=1} \frac{2\, dz}{i(z^2 + 4z + 1)}

Poles: z2+4z+1=0z=2±3z^2 + 4z + 1 = 0 \Rightarrow z = -2 \pm \sqrt{3}.

z1=2+3=23<1|z_1| = |-2 + \sqrt{3}| = 2 - \sqrt{3} \lt 1 (inside). z2=23=2+3>1|z_2| = |-2 - \sqrt{3}| = 2 + \sqrt{3} \gt 1 (outside).

Res(1z2+4z+1,z1)=123\mathrm{Res}\left(\frac{1}{z^2 + 4z + 1}, z_1\right) = \frac{1}{2\sqrt{3}}.

02πdθ2+cosθ=2i2πi123=2π3\int_0^{2\pi} \frac{d\theta}{2 + \cos\theta} = \frac{2}{i} \cdot 2\pi i \cdot \frac{1}{2\sqrt{3}} = \frac{2\pi}{\sqrt{3}}. \blacksquare

9.6 Jordan’s Lemma

Theorem 9.2 (Jordan’s Lemma). If f(z)0f(z) \to 0 uniformly as z|z| \to \infty in the upper Half-plane and a>0a \gt 0Then

limRCReiazf(z)dz=0\lim_{R \to \infty} \int_{C_R} e^{iaz}f(z)\, dz = 0

Where CRC_R is the upper semicircle z=R|z| = R, Im(z)0\mathrm{Im}(z) \geq 0.

This allows evaluation of integrals of the form f(x)cos(ax)dx\int_{-\infty}^{\infty} f(x)\cos(ax)\, dx and f(x)sin(ax)dx\int_{-\infty}^{\infty} f(x)\sin(ax)\, dx.

9.7 Fourier-Type Integrals

Solution

Problem. Evaluate cosxx2+1dx\int_{-\infty}^{\infty} \frac{\cos x}{x^2 + 1}\, dx.

Consider eixx2+1dx=2πiRes ⁣(eizz2+1,i)\int_{-\infty}^{\infty} \frac{e^{ix}}{x^2 + 1}\, dx = 2\pi i \cdot \mathrm{Res}\!\left(\frac{e^{iz}}{z^2+1}, i\right).

Res ⁣(eizz2+1,i)=eii2i=e12i\mathrm{Res}\!\left(\frac{e^{iz}}{z^2+1}, i\right) = \frac{e^{i \cdot i}}{2i} = \frac{e^{-1}}{2i}.

eixx2+1dx=2πie12i=πe\int_{-\infty}^{\infty} \frac{e^{ix}}{x^2 + 1}\, dx = 2\pi i \cdot \frac{e^{-1}}{2i} = \frac{\pi}{e}.

Taking real parts: cosxx2+1dx=πe\int_{-\infty}^{\infty} \frac{\cos x}{x^2 + 1}\, dx = \frac{\pi}{e}.

Problem. Evaluate xsinxx2+a2dx\int_{-\infty}^{\infty} \frac{x \sin x}{x^2 + a^2}\, dx for a>0a \gt 0.

Consider zeizz2+a2dz\int_{-\infty}^{\infty} \frac{z\, e^{iz}}{z^2 + a^2}\, dz. Only z=iaz = ia is in the upper half-plane.

Res ⁣(zeizz2+a2,ia)=iaeiia2ia=ea2\mathrm{Res}\!\left(\frac{ze^{iz}}{z^2 + a^2}, ia\right) = \frac{ia \cdot e^{i \cdot ia}}{2ia} = \frac{e^{-a}}{2}.

xeixx2+a2dx=2πiea2=πiea\int_{-\infty}^{\infty} \frac{x\, e^{ix}}{x^2 + a^2}\, dx = 2\pi i \cdot \frac{e^{-a}}{2} = \pi i\, e^{-a}.

Taking imaginary parts: xsinxx2+a2dx=πea\int_{-\infty}^{\infty} \frac{x \sin x}{x^2 + a^2}\, dx = \pi\, e^{-a}.

Problem. Evaluate 02πcos2θ5+4cosθdθ\int_0^{2\pi} \frac{\cos 2\theta}{5 + 4\cos\theta}\, d\theta.

Substitute z=eiθz = e^{i\theta}: cosθ=(z+z1)/2\cos\theta = (z + z^{-1})/2, cos2θ=(z2+z2)/2\cos 2\theta = (z^2 + z^{-2})/2.

I=12iz=1z4+1z2(2z+1)(z+2)dzI = \frac{1}{2i}\int_{|z|=1} \frac{z^4 + 1}{z^2(2z + 1)(z + 2)}\, dz.

Poles inside z=1|z| = 1: z=0z = 0 (order 22) and z=1/2z = -1/2 (simple).

At z=0z = 0: Res=ddz[z4+1(2z+1)(z+2)]z=0=54\mathrm{Res} = \frac{d}{dz}\left[\frac{z^4 + 1}{(2z+1)(z+2)}\right]_{z=0} = -\frac{5}{4}.

At z=1/2z = -1/2: Res=17/163/4=1712\mathrm{Res} = \frac{17/16}{3/4} = \frac{17}{12}.

I=12i2πi(54+1712)=π6I = \frac{1}{2i} \cdot 2\pi i \left(-\frac{5}{4} + \frac{17}{12}\right) = \frac{\pi}{6}.

9.8 Improper Integrals and Principal Value

For integrals where the integrand has poles on the real axis, we use the Cauchy principal value:

PV ⁣f(x)dx=limε0+(aεf(x)dx+a+εf(x)dx)\mathrm{PV}\!\int_{-\infty}^{\infty} f(x)\, dx = \lim_{\varepsilon \to 0^+} \left(\int_{-\infty}^{a-\varepsilon} f(x)\, dx + \int_{a+\varepsilon}^{\infty} f(x)\, dx\right)

Solution

Problem. Evaluate PV ⁣sinxxdx\mathrm{PV}\!\int_{-\infty}^{\infty} \frac{\sin x}{x}\, dx.

Consider γeizzdz\oint_\gamma \frac{e^{iz}}{z}\, dz where γ\gamma consists of [R,ε][-R, -\varepsilon] [ε,R][\varepsilon, R] on the real axis, small upper semicircle CεC_\varepsilon around 00And large Upper semicircle CRC_R.

No poles inside the contour, so the integral is 00.

On CRC_R: vanishes as RR \to \infty by Jordan’s lemma. On CεC_\varepsilon (indenting above): Cεeizzdziπ\int_{C_\varepsilon} \frac{e^{iz}}{z}\, dz \to -i\pi as ε0\varepsilon \to 0 (half residue contribution).

0=PV ⁣eixxdx+(iπ)0 = \mathrm{PV}\!\int_{-\infty}^{\infty} \frac{e^{ix}}{x}\, dx + (-i\pi).

PV ⁣eixxdx=iπ\mathrm{PV}\!\int_{-\infty}^{\infty} \frac{e^{ix}}{x}\, dx = i\pi.

Taking imaginary parts: PV ⁣sinxxdx=π\mathrm{PV}\!\int_{-\infty}^{\infty} \frac{\sin x}{x}\, dx = \pi.

10. Conformal Mappings

10.1 Definition

Definition. An analytic function ff is conformal at z0z_0 if f(z0)0f'(z_0) \neq 0. A conformal Mapping preserves angles (both magnitude and orientation) between curves.

10.2 Geometric Interpretation

If f(z0)=reiθf'(z_0) = re^{i\theta}Then near z0z_0 the mapping ff acts as a rotation by θ\theta followed By a scaling by rr. The Jacobian determinant is f(z0)2>0|f'(z_0)|^2 \gt 0So orientation is preserved.

10.3 Common Conformal Mappings

MappingEffect
w=az+bw = az + b (a0a \neq 0)Translation, rotation, scaling
w=1/zw = 1/zInversion in the unit circle
w=z2w = z^2Squaring (doubles angles)
w=ezw = e^zExponential (maps strips to sectors)
w=za1aˉzw = \frac{z - a}{1 - \bar{a}z}Möbius (maps disk to disk)

10.4 Möbius Transformations

A Möbius transformation (or linear fractional transformation) is

T(z)=az+bcz+d,adbc0T(z) = \frac{az + b}{cz + d}, \quad ad - bc \neq 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(z_1) = w_1 T(z2)=w2T(z_2) = w_2, T(z3)=w3T(z_3) = w_3.

10.5 Cross-Ratio

Definition. The cross-ratio of four distinct points z1,z2,z3,z4z_1, z_2, z_3, z_4 is

(z1,z2,z3,z4)=(z1z3)(z2z4)(z1z4)(z2z3)(z_1, z_2, z_3, z_4) = \frac{(z_1 - z_3)(z_2 - z_4)}{(z_1 - z_4)(z_2 - z_3)}

Proposition 10.3. The cross-ratio is invariant under Möbius transformations: (Tz1,Tz2,Tz3,Tz4)=(z1,z2,z3,z4)(Tz_1, Tz_2, Tz_3, Tz_4) = (z_1, z_2, z_3, z_4).

Proposition 10.4. The unique Möbius transformation sending z10z_1 \mapsto 0, z21z_2 \mapsto 1 z3z_3 \mapsto \infty is

T(z)=(zz1)(z2z3)(zz3)(z2z1)T(z) = \frac{(z - z_1)(z_2 - z_3)}{(z - z_3)(z_2 - z_1)}

10.6 Classification of Möbius Transformations

A Möbius transformation T(z)=az+bcz+dT(z) = \frac{az + b}{cz + d} is classified by its fixed points (solutions of T(z)=zT(z) = z).

  1. Parabolic: Exactly one fixed point. Conjugate to w=z+kw = z + k.
  2. Elliptic: Two fixed points, T(z0)=1|T'(z_0)| = 1. Conjugate to a rotation w=eiθzw = e^{i\theta} z.
  3. Hyperbolic: Two fixed points, T(z0)R+T'(z_0) \in \mathbb{R}^+, T(z0)1T'(z_0) \neq 1. Conjugate to w=kzw = kz.
  4. Loxodromic: Two fixed points, T(z0)R{z:z=1}T'(z_0) \notin \mathbb{R} \cup \{z : |z| = 1\}. Conjugate to w=keiθzw = ke^{i\theta}z.
Solution

Problem. Find the Möbius transformation mapping 0i0 \mapsto i, 101 \mapsto 0, i\infty \mapsto -i.

T(z)=az+bcz+dT(z) = \frac{az + b}{cz + d} with T(0)=ib/d=ib=idT(0) = i \Rightarrow b/d = i \Rightarrow b = id. T(1)=0a=b=idT(1) = 0 \Rightarrow a = -b = -id. T()=ia/c=ic=dT(\infty) = -i \Rightarrow a/c = -i \Rightarrow c = d.

T(z)=idz+iddz+d=i(1z)z+1T(z) = \frac{-idz + id}{dz + d} = \frac{i(1 - z)}{z + 1}.

Problem. Show that T(z)=z1z+1T(z) = \frac{z - 1}{z + 1} maps the right half-plane to the unit disk.

If Re(z)>0\mathrm{Re}(z) \gt 0Then z1<z+1|z - 1| \lt |z + 1|So T(z)<1|T(z)| \lt 1.

Check boundary: T(i)=i1i+1=(i1)(i+1)(i+1)(i+1)=22=1T(i) = \frac{i - 1}{i + 1} = \frac{(i-1)(-i+1)}{(i+1)(-i+1)} = \frac{2}{2} = 1. T(i)=1|T(i)| = 1. \checkmark

Problem. Classify T(z)=2z+1z+2T(z) = \frac{2z + 1}{z + 2}.

Fixed points: z=2z+1z+2z2=1z=±1z = \frac{2z + 1}{z + 2} \Rightarrow z^2 = 1 \Rightarrow z = \pm 1.

T(z)=3(z+2)2T'(z) = \frac{3}{(z + 2)^2}. T(1)=1/3T'(1) = 1/3, T(1)=3T'(-1) = 3.

Both multipliers are real and positive (not equal to 11), so TT is hyperbolic.

10.7 The Riemann Mapping Theorem

Theorem 10.5 (Riemann Mapping Theorem). Let UU be a connected open proper subset of C\mathbb{C}. Then there exists a bijective conformal map from UU onto the unit disk D={z:z<1}\mathbb{D} = \{z : |z| \lt 1\}.

This is one of the most profound results in complex analysis, establishing that all connected Domains (other than C\mathbb{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|f(z)| \leq M for all zzThen by Cauchy’s estimates with RR arbitrarily large: f(z0)MR0|f'(z_0)| \leq \frac{M}{R} \to 0 as RR \to \infty. So f(z)=0f'(z) = 0 for all zzMeaning ff is Constant. \blacksquare

11.2 The Fundamental Theorem of Algebra

Theorem 11.2 (Fundamental Theorem of Algebra). Every non-constant polynomial p(z)C[z]p(z) \in \mathbb{C}[z] has a root in C\mathbb{C}.

Proof. Suppose p(z)p(z) has no root. Then f(z)=1/p(z)f(z) = 1/p(z) is entire. Since p(z)|p(z)| \to \infty as z|z| \to \infty, f(z)0f(z) \to 0So ff is bounded. By Liouville’s theorem, ff is constant, so pp Is constant, a contradiction. \blacksquare

11.3 The Maximum Modulus Principle

Theorem 11.3 (Maximum Modulus Principle). If ff is analytic and non-constant on a domain DD Then f|f| has no local maximum in DD.

Corollary 11.4. If ff is analytic on a bounded domain DD and continuous on Dˉ=DD\bar{D} = D \cup \partial DThen f|f| attains its maximum on D\partial D.

11.4 Minimum Modulus Principle

Theorem 11.5 (Minimum Modulus Principle). If ff is analytic and non-zero on a bounded domain DD And continuous on Dˉ\bar{D}Then f|f| attains its minimum on D\partial D.

Remark. If ff has zeros in DDThen f|f| attains its minimum of 00 at those zeros. The minimum modulus principle requires the non-vanishing hypothesis.

11.5 Schwarz Lemma

Theorem 11.6 (Schwarz Lemma). If f:DDf : \mathbb{D} \to \mathbb{D} is analytic with f(0)=0f(0) = 0 Then

f(z)zfor all zD|f(z)| \leq |z| \quad \mathrm{for\ all\ } z \in \mathbb{D}

And f(0)1|f'(0)| \leq 1. Equality in either case implies f(z)=eiθzf(z) = e^{i\theta} z for some real θ\theta.

Proof. Define g(z)=f(z)/zg(z) = f(z)/z for z0z \neq 0 and g(0)=f(0)g(0) = f'(0). Then gg is analytic on D\mathbb{D}. For z=r<1|z| = r \lt 1: g(z)=f(z)/z1/r|g(z)| = |f(z)|/|z| \leq 1/r. By the maximum modulus Principle, g(z)1/r|g(z)| \leq 1/r for zr|z| \leq r. Letting r1r \to 1: g(z)1|g(z)| \leq 1So f(z)z|f(z)| \leq |z|. Also f(0)=g(0)1|f'(0)| = |g(0)| \leq 1. If f(0)=1|f'(0)| = 1Then g|g| attains its maximum At an interior point, so gg is constant: g(z)=eiθg(z) = e^{i\theta}. \blacksquare

12. Argument Principle and Rouché’s Theorem

12.1 The Argument Principle

Theorem 12.1 (Argument Principle). If ff is meromorphic inside and on a simple closed contour γ\gamma with no zeros or poles on γ\gammaThen

12πiγf(z)f(z)dz=NP\frac{1}{2\pi i}\int_\gamma \frac{f'(z)}{f(z)}\, dz = N - P

Where NN is the number of zeros and PP is the number of poles of ff inside γ\gamma (counting Multiplicities).

12.2 Rouché’s Theorem

Theorem 12.2 (Rouché’s Theorem). If ff and gg are analytic inside and on a simple closed Contour γ\gammaAnd f(z)>g(z)|f(z)| \gt |g(z)| on γ\gammaThen ff and f+gf + g have the same number of Zeros inside γ\gamma.

Proof. On γ\gamma: g(z)/f(z)<1|g(z)/f(z)| \lt 1. The function h(z)=1+g(z)/f(z)h(z) = 1 + g(z)/f(z) satisfies h(z)1<1|h(z) - 1| \lt 1 on γ\gammaSo h(γ)h(\gamma) does not wind around 00. By the argument principle Applied to hh: 0=NhPh0 = N_h - P_hMeaning hh has the same number of zeros and poles inside γ\gamma. But h=(f+g)/fh = (f + g)/fSo zeros of hh are zeros of f+gf + g and poles of hh are zeros of ff. Therefore ff and f+gf + g have the same number of zeros. \blacksquare

12.3 Worked Example

Problem. Show that z4+6z+3z^4 + 6z + 3 has exactly one root in z<1|z| \lt 1.

Solution. On z=1|z| = 1: 6z=6>z4+3z4+3=4|6z| = 6 \gt |z^4 + 3| \leq |z|^4 + 3 = 4. By Rouché’s theorem with f(z)=6zf(z) = 6z and g(z)=z4+3g(z) = z^4 + 3: f+g=z4+6z+3f + g = z^4 + 6z + 3 has the same number of zeros in z<1|z| \lt 1 as f(z)=6zf(z) = 6zWhich has exactly one zero (at z=0z = 0). \blacksquare

Solution

Problem. Show that all roots of z4+z+1=0z^4 + z + 1 = 0 satisfy z<2|z| \lt 2.

On z=2|z| = 2: z4=16>z+13|z^4| = 16 \gt |z + 1| \leq 3. By Rouché with f(z)=z4f(z) = z^4 and g(z)=z+1g(z) = z + 1: z4+z+1z^4 + z + 1 has 44 zeros in z<2|z| \lt 2 (same as z4z^4).

Problem. Show that z5+3z2+1z^5 + 3z^2 + 1 has exactly two roots in z<1|z| \lt 1.

On z=1|z| = 1: 3z2+13z21=2>z5=1|3z^2 + 1| \geq |3z^2| - |1| = 2 \gt |z^5| = 1. By Rouché with f(z)=3z2+1f(z) = 3z^2 + 1 and g(z)=z5g(z) = z^5: z5+3z2+1z^5 + 3z^2 + 1 has the same number of zeros as 3z2+13z^2 + 1 in z<1|z| \lt 1. 3z2+1=0z=±i/33z^2 + 1 = 0 \Rightarrow z = \pm i/\sqrt{3}Both in z<1|z| \lt 1. So 22 zeros.

13. Analytic Continuation

13.1 Definition

Definition. If f1f_1 is analytic on D1D_1 and f2f_2 is analytic on D2D_2 with D1D2D_1 \cap D_2 \neq \emptyset and f1=f2f_1 = f_2 on D1D2D_1 \cap D_2Then f2f_2 is an analytic Continuation of f1f_1.

13.2 Identity Theorem

Theorem 13.1 (Identity Theorem). If ff and gg are analytic on a domain DD and agree on a set With a limit point in DDThen f=gf = g on all of DD.

Proof. Let E={zD:f(n)(z)=g(n)(z) for all n0}E = \{z \in D : f^{(n)}(z) = g^{(n)}(z) \mathrm{\ for\ all\ } n \geq 0\}. EE is Non-empty (it contains the limit point by continuity of derivatives). EE is closed (by continuity). If z0Ez_0 \in EThe Taylor series of ff and gg at z0z_0 coincide, so f=gf = g in a neighbourhood of z0z_0Giving EE open. Since DD is connected, E=DE = D. \blacksquare

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)f(z) = \exp(-1/z^4) extended by f(0)=0f(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\oint_{|z|=1} dz/z = 2\pi i. :::

:::caution Common Pitfall When computing residues at poles of order m2m \geq 2The formula involves Differentiation. A common error is forgetting the (m1)!(m-1)! in the denominator or differentiating (zz0)mf(z)(z - z_0)^m f(z) the wrong number of times. :::

:::caution Common Pitfall The residue at infinity is Res(f,)=Res(1/z2f(1/z),0)\mathrm{Res}(f, \infty) = -\mathrm{Res}(1/z^2 \cdot f(1/z), 0). It is NOT f()f(\infty). 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=z2w = z^2 is conformal at every z0z \neq 0But it doubles the angle between curves at each Point. At z=0z = 0It is not conformal because f(0)=0f'(0) = 0. :::

:::caution Common Pitfall The maximum modulus principle says that f|f| has no local maximum in the Interior, but the minimum can occur in the interior (e.g., f(z)=zf(z) = z on the unit disk has minimum f=0|f| = 0 at z=0z = 0). For the minimum principle, one needs the additional hypothesis that ff 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\mathrm{Log}\, z has a branch cut along (,0](-\infty, 0] And is undefined on this cut. :::

:::caution Common Pitfall When applying the ML inequality, make sure MM is a valid upper bound for f(z)|f(z)| on the entire contour. A common error is bounding f|f| on only part of the contour. Also, LL must be the arc length of the contour, not a diameter or radius. :::

15. Problem Set

Problem 1

Express z=3+iz = -\sqrt{3} + i in polar form and find all values of z1/3z^{1/3}.

Solution

z=3+1=2|z| = \sqrt{3 + 1} = 2. Since Re(z)<0\mathrm{Re}(z) \lt 0 and Im(z)>0\mathrm{Im}(z) \gt 0: arg(z)=ππ/6=5π/6\arg(z) = \pi - \pi/6 = 5\pi/6.

z=2e5πi/6z = 2\,e^{5\pi i/6}.

z1/3=21/3e(5π/6+2πk)/3z^{1/3} = 2^{1/3}\, e^{(5\pi/6 + 2\pi k)/3} for k=0,1,2k = 0, 1, 2.

z0=21/3e5πi/18z_0 = 2^{1/3}\, e^{5\pi i/18}, z1=21/3e17πi/18z_1 = 2^{1/3}\, e^{17\pi i/18}, z2=21/3e29πi/18z_2 = 2^{1/3}\, e^{29\pi i/18}.

If you get this wrong, revise: Section 1.5 (Roots of Complex Numbers).

Problem 2

Let f(z)=z2+zˉ2f(z) = z^2 + \bar{z}^2. Find where ff is differentiable and where it is analytic.

Solution

f(z)=(x+iy)2+(xiy)2=2(x2y2)f(z) = (x + iy)^2 + (x - iy)^2 = 2(x^2 - y^2). So u=2(x2y2)u = 2(x^2 - y^2), v=0v = 0.

ux=4xu_x = 4x, uy=4yu_y = -4y, vx=0v_x = 0, vy=0v_y = 0.

CR: 4x=0x=04x = 0 \Rightarrow x = 0, 4y=0y=0-4y = 0 \Rightarrow y = 0.

ff is differentiable only at z=0z = 0 and analytic nowhere.

f(0)=0f'(0) = 0 (verified by direct computation).

If you get this wrong, revise: Sections 2.4 and 3.1 (Analyticity and Cauchy-Riemann).

Problem 3

Verify that f(z)=1z2+1f(z) = \frac{1}{z^2 + 1} satisfies the Cauchy-Riemann equations on its domain and Find f(z)f'(z).

Solution

f(z)=1/(z2+1)f(z) = 1/(z^2 + 1) is a rational function with denominator non-zero away from ±i\pm iSo ff Is analytic on C{i,i}\mathbb{C} \setminus \{i, -i\}.

By the quotient rule: f(z)=2z(z2+1)2f'(z) = \frac{-2z}{(z^2 + 1)^2}.

Verify via CR at z=1z = 1: u=x2y2+1(x2y2+1)2+4x2y2u = \frac{x^2 - y^2 + 1}{(x^2 - y^2 + 1)^2 + 4x^2y^2} v=2xy(x2y2+1)2+4x2y2v = \frac{-2xy}{(x^2 - y^2 + 1)^2 + 4x^2y^2}.

ux(1,0)=1/2=f(1)u_x(1, 0) = -1/2 = f'(1). \checkmark

If you get this wrong, revise: Sections 3.1 and 3.3 (CR Equations).

Problem 4

Show that u(x,y)=x33xy2+3x23y2u(x, y) = x^3 - 3xy^2 + 3x^2 - 3y^2 is harmonic and find its harmonic conjugate.

Solution

uxx=6x+6u_{xx} = 6x + 6, uyy=6x6u_{yy} = -6x - 6. Δu=0\Delta u = 0. \checkmark

By CR: vy=ux=3x23y2+6xv_y = u_x = 3x^2 - 3y^2 + 6x. v=3x2yy3+6xy+g(x)v = 3x^2 y - y^3 + 6xy + g(x).

vx=uy=6xy+6yv_x = -u_y = 6xy + 6y. 6xy+6y=6xy+6y+g(x)g(x)=0g(x)=C6xy + 6y = 6xy + 6y + g'(x) \Rightarrow g'(x) = 0 \Rightarrow g(x) = C.

Harmonic conjugate: v(x,y)=3x2yy3+6xy+Cv(x, y) = 3x^2 y - y^3 + 6xy + C.

f(z)=u+iv=z3+3z2f(z) = u + iv = z^3 + 3z^2.

If you get this wrong, revise: Section 3.4 (Harmonic Functions).

Problem 5

Evaluate γ(z2+2z)dz\int_\gamma (z^2 + 2z)\, dz where γ\gamma is the upper half of the unit circle from z=1z = 1 to z=1z = -1.

Solution

Since z2+2zz^2 + 2z is entire, the integral is path-independent. Let F(z)=z3/3+z2F(z) = z^3/3 + z^2.

γ(z2+2z)dz=F(1)F(1)=2343=23\int_\gamma (z^2 + 2z)\, dz = F(-1) - F(1) = \frac{2}{3} - \frac{4}{3} = -\frac{2}{3}.

If you get this wrong, revise: Sections 4.5 and 4.7 (Contour Integrals).

Problem 6

Use the ML inequality to bound γezz2dz\left|\int_\gamma \frac{e^z}{z - 2}\, dz\right| where γ\gamma Is the circle z=1|z| = 1.

Solution

On γ\gamma: z=1|z| = 1So eze|e^z| \leq e and z21|z - 2| \geq 1.

ezz2e\left|\frac{e^z}{z - 2}\right| \leq e. L=2πL = 2\pi.

γezz2dz2πe\left|\int_\gamma \frac{e^z}{z - 2}\, dz\right| \leq 2\pi e.

If you get this wrong, revise: Section 4.6 (ML Inequality).

Problem 7

Evaluate γz+1z2zdz\oint_\gamma \frac{z + 1}{z^2 - z}\, dz where γ\gamma is z=2|z| = 2.

Solution

z+1z2z=z+1z(z1)\frac{z + 1}{z^2 - z} = \frac{z + 1}{z(z - 1)}. Simple poles at z=0z = 0 and z=1z = 1Both inside z=2|z| = 2.

At z=0z = 0: Res=limz0z+1z1=1\mathrm{Res} = \lim_{z \to 0} \frac{z + 1}{z - 1} = -1. At z=1z = 1: Res=limz1z+1z=2\mathrm{Res} = \lim_{z \to 1} \frac{z + 1}{z} = 2.

γz+1z2zdz=2πi(1+2)=2πi\oint_\gamma \frac{z + 1}{z^2 - z}\, dz = 2\pi i(-1 + 2) = 2\pi i.

If you get this wrong, revise: Sections 8.4 and 8.5 (Residues).

Problem 8

Classify all singularities of f(z)=e1/zz2+1f(z) = \frac{e^{1/z}}{z^2 + 1} and find all residues.

Solution

z=0z = 0: e1/ze^{1/z} has an essential singularity at 00So z=0z = 0 is an essential singularity of ff. z=iz = i: simple pole. z=iz = -i: simple pole.

At z=iz = i: Res=e1/i2i=ei2i\mathrm{Res} = \frac{e^{1/i}}{2i} = \frac{e^{-i}}{2i}. At z=iz = -i: Res=e1/(i)2i=ei2i\mathrm{Res} = \frac{e^{1/(-i)}}{-2i} = \frac{e^{i}}{-2i}.

At z=0z = 0: find the coefficient of 1/z1/z in e1/zz2+1\frac{e^{1/z}}{z^2 + 1}. 1z2+1=1z2+z4\frac{1}{z^2 + 1} = 1 - z^2 + z^4 - \cdots near z=0z = 0. e1/z=1+1/z+1/(2z2)+e^{1/z} = 1 + 1/z + 1/(2z^2) + \cdots. The 1/z1/z coefficient in the product: from 11/z=1/z1 \cdot 1/z = 1/zGiving residue 11.

If you get this wrong, revise: Sections 8.1 and 8.4 (Singularities and Residues).

Problem 9

Evaluate 02πcosθ5+4cosθdθ\int_0^{2\pi} \frac{\cos\theta}{5 + 4\cos\theta}\, d\theta.

Solution

Substitute z=eiθz = e^{i\theta}:

I=z=1(z+z1)/25+4(z+z1)/2dziz=12iz=1z2+1z(2z2+5z+2)dz=12iz=1z2+1z(2z+1)(z+2)dzI = \int_{|z|=1} \frac{(z + z^{-1})/2}{5 + 4(z + z^{-1})/2} \cdot \frac{dz}{iz} = \frac{1}{2i}\int_{|z|=1} \frac{z^2 + 1}{z(2z^2 + 5z + 2)}\, dz = \frac{1}{2i}\int_{|z|=1} \frac{z^2 + 1}{z(2z + 1)(z + 2)}\, dz.

Poles inside z=1|z| = 1: z=0z = 0 (simple) and z=1/2z = -1/2 (simple).

At z=0z = 0: Res=1(20+1)(0+2)=12\mathrm{Res} = \frac{1}{(2 \cdot 0 + 1)(0 + 2)} = \frac{1}{2}. At z=1/2z = -1/2: Res=1/4+1(1/2)(1+2)=5/41/2=52\mathrm{Res} = \frac{1/4 + 1}{(-1/2)(-1 + 2)} = \frac{5/4}{-1/2} = -\frac{5}{2}.

I=12i2πi(1252)=π(2)=π3I = \frac{1}{2i} \cdot 2\pi i\left(\frac{1}{2} - \frac{5}{2}\right) = \pi(-2) = -\frac{\pi}{3}.

If you get this wrong, revise: Section 9.4 (Trigonometric Integrals).

Problem 10

Evaluate dx(x2+1)(x2+4)\int_{-\infty}^{\infty} \frac{dx}{(x^2 + 1)(x^2 + 4)}.

Solution

f(z)=1(z2+1)(z2+4)f(z) = \frac{1}{(z^2 + 1)(z^2 + 4)}. Poles in upper half-plane: z=iz = i (simple) and z=2iz = 2i (simple).

At z=iz = i: Res=1(2i)(i2+4)=12i3=16i\mathrm{Res} = \frac{1}{(2i)(i^2 + 4)} = \frac{1}{2i \cdot 3} = \frac{1}{6i}. At z=2iz = 2i: Res=1(4i1)(4i)=14i(3)=112i\mathrm{Res} = \frac{1}{(4i - 1)(4i)} = \frac{1}{4i(-3)} = -\frac{1}{12i}.

f(x)dx=2πi(16i112i)=2πi112i=π6\int_{-\infty}^{\infty} f(x)\, dx = 2\pi i\left(\frac{1}{6i} - \frac{1}{12i}\right) = 2\pi i \cdot \frac{1}{12i} = \frac{\pi}{6}.

If you get this wrong, revise: Section 9.2 (Rational Function Integrals).

Problem 11

Find the Taylor series of f(z)=zz2+4f(z) = \frac{z}{z^2 + 4} centered at z0=0z_0 = 0 and state the radius Of convergence.

Solution

zz2+4=z411+z2/4=z4n=0(1)nz2n4n=n=0(1)nz2n+14n+1\frac{z}{z^2 + 4} = \frac{z}{4} \cdot \frac{1}{1 + z^2/4} = \frac{z}{4}\sum_{n=0}^{\infty} (-1)^n \frac{z^{2n}}{4^n} = \sum_{n=0}^{\infty} \frac{(-1)^n z^{2n+1}}{4^{n+1}}

For z<2|z| \lt 2. Radius of convergence: distance from 00 to nearest singularity (±2i\pm 2i), which is 22.

If you get this wrong, revise: Section 7.1 (Taylor Series).

Problem 12

Find the Laurent series of f(z)=1(z1)(z2)f(z) = \frac{1}{(z - 1)(z - 2)} in the annulus 1<z<21 \lt |z| \lt 2.

Solution

1(z1)(z2)=1z21z1\frac{1}{(z-1)(z-2)} = \frac{1}{z - 2} - \frac{1}{z - 1}.

For z>1|z| \gt 1: 1z1=1z111/z=n=0zn1\frac{1}{z - 1} = \frac{1}{z} \cdot \frac{1}{1 - 1/z} = \sum_{n=0}^{\infty} z^{-n-1}.

For z<2|z| \lt 2: 1z2=1211z/2=n=0zn2n+1\frac{1}{z - 2} = -\frac{1}{2} \cdot \frac{1}{1 - z/2} = -\sum_{n=0}^{\infty} \frac{z^n}{2^{n+1}}.

f(z)=n=0zn2n+1n=0zn1f(z) = -\sum_{n=0}^{\infty} \frac{z^n}{2^{n+1}} - \sum_{n=0}^{\infty} z^{-n-1}.

If you get this wrong, revise: Section 7.4 (Laurent Series).

Problem 13

Using Rouché’s theorem, determine the number of roots of z55z+1=0z^5 - 5z + 1 = 0 in z<1|z| \lt 1.

Solution

On z=1|z| = 1: 5z=5>z5+12|-5z| = 5 \gt |z^5 + 1| \leq 2.

By Rouché with f(z)=5zf(z) = -5z and g(z)=z5+1g(z) = z^5 + 1: z55z+1z^5 - 5z + 1 has the same number of zeros In z<1|z| \lt 1 as 5z-5zWhich has exactly one zero (at z=0z = 0).

So exactly one root in z<1|z| \lt 1.

If you get this wrong, revise: Section 12.2 (Rouché’s Theorem).

Problem 14

Find the Möbius transformation that maps 101 \mapsto 0, i1i \mapsto 1, 1-1 \mapsto \infty.

Solution

T(z)=(z1)(i(1))(z(1))(i1)=(z1)(i+1)(z+1)(i1)T(z) = \frac{(z - 1)(i - (-1))}{(z - (-1))(i - 1)} = \frac{(z - 1)(i + 1)}{(z + 1)(i - 1)}.

Simplify: i+1i1=(i+1)(i1)(i1)(i1)=i22i1i2+1=2i2=i\frac{i + 1}{i - 1} = \frac{(i+1)(-i-1)}{(i-1)(-i-1)} = \frac{-i^2 - 2i - 1}{-i^2 + 1} = \frac{-2i}{2} = -i.

T(z)=iz1z+1T(z) = -i \cdot \frac{z - 1}{z + 1}.

Verify: T(1)=0T(1) = 0 \checkmark, T(i)=ii1i+1=i(i)=1T(i) = -i \cdot \frac{i-1}{i+1} = -i \cdot (-i) = -1.

That gives 1-1Not 11. Let me recompute.

T(z)=(zz1)(z2z3)(zz3)(z2z1)T(z) = \frac{(z - z_1)(z_2 - z_3)}{(z - z_3)(z_2 - z_1)} with z1=1z_1 = 1, z2=iz_2 = i, z3=1z_3 = -1.

T(z)=(z1)(i+1)(z+1)(i1)T(z) = \frac{(z - 1)(i + 1)}{(z + 1)(i - 1)}.

T(i)=(i1)(i+1)(i+1)(i1)=1T(i) = \frac{(i - 1)(i + 1)}{(i + 1)(i - 1)} = 1. \checkmark

T(1)=0T(1) = 0. \checkmark. T(1)=T(-1) = \infty. \checkmark.

So T(z)=(z1)(i+1)(z+1)(i1)T(z) = \frac{(z - 1)(i + 1)}{(z + 1)(i - 1)}.

If you get this wrong, revise: Section 10.5 (Cross-Ratio).

Problem 15

Evaluate γz3z2+1dz\int_\gamma \frac{z^3}{z^2 + 1}\, dz where γ\gamma is z=2|z| = 2.

Solution

z3z2+1\frac{z^3}{z^2 + 1} has simple poles at z=±iz = \pm iBoth inside z=2|z| = 2.

At z=iz = i: Res=i32i=i2i=12\mathrm{Res} = \frac{i^3}{2i} = \frac{-i}{2i} = -\frac{1}{2}. At z=iz = -i: Res=(i)32i=i2i=12\mathrm{Res} = \frac{(-i)^3}{-2i} = \frac{i}{-2i} = -\frac{1}{2}.

γz3z2+1dz=2πi(1212)=2πi\int_\gamma \frac{z^3}{z^2 + 1}\, dz = 2\pi i\left(-\frac{1}{2} - \frac{1}{2}\right) = -2\pi i.

Alternatively: z3z2+1=zzz2+1\frac{z^3}{z^2 + 1} = z - \frac{z}{z^2 + 1}. γzdz=0\int_\gamma z\, dz = 0 (entire), and γzz2+1dz=2πi(1/2+1/2)=2πi\int_\gamma \frac{z}{z^2 + 1}\, dz = 2\pi i(1/2 + 1/2) = 2\pi i. So the integral equals 02πi=2πi0 - 2\pi i = -2\pi i. \checkmark

If you get this wrong, revise: Sections 8.4 and 8.5 (Residues).

Problem 16

Show that cos2xx2+1dx=πe2\int_{-\infty}^{\infty} \frac{\cos 2x}{x^2 + 1}\, dx = \frac{\pi}{e^2}.

Solution

Consider e2ixx2+1dx\int_{-\infty}^{\infty} \frac{e^{2ix}}{x^2 + 1}\, dx.

f(z)=e2izz2+1f(z) = \frac{e^{2iz}}{z^2 + 1} has a simple pole at z=iz = i in the upper half-plane.

Res ⁣(e2izz2+1,i)=e2ii2i=e22i\mathrm{Res}\!\left(\frac{e^{2iz}}{z^2 + 1}, i\right) = \frac{e^{2i \cdot i}}{2i} = \frac{e^{-2}}{2i}.

e2ixx2+1dx=2πie22i=πe2\int_{-\infty}^{\infty} \frac{e^{2ix}}{x^2 + 1}\, dx = 2\pi i \cdot \frac{e^{-2}}{2i} = \frac{\pi}{e^2}.

Taking real parts: cos2xx2+1dx=πe2\int_{-\infty}^{\infty} \frac{\cos 2x}{x^2 + 1}\, dx = \frac{\pi}{e^2}.

If you get this wrong, revise: Section 9.7 (Fourier-Type Integrals).

Problem 17

Find the residue of f(z)=sinzz4f(z) = \frac{\sin z}{z^4} at z=0z = 0.

Solution

sinz=zz3/6+z5/120\sin z = z - z^3/6 + z^5/120 - \cdots

f(z)=zz3/6+z5/120z4=1z316z+z120f(z) = \frac{z - z^3/6 + z^5/120 - \cdots}{z^4} = \frac{1}{z^3} - \frac{1}{6z} + \frac{z}{120} - \cdots

The coefficient of 1/z1/z is 1/6-1/6So Res(f,0)=16\mathrm{Res}(f, 0) = -\frac{1}{6}.

If you get this wrong, revise: Section 8.4 (Computing Residues).

Problem 18

Evaluate γdz(z1)2(z2)\int_\gamma \frac{dz}{(z - 1)^2(z - 2)} where γ\gamma is z1=1/2|z - 1| = 1/2.

Solution

Only z=1z = 1 is inside γ\gamma (a pole of order 22). z=2z = 2 is outside.

Res(f,1)=ddz[1z2]z=1=1(z2)2z=1=1\mathrm{Res}(f, 1) = \frac{d}{dz}\left[\frac{1}{z - 2}\right]_{z=1} = -\frac{1}{(z-2)^2}\Big|_{z=1} = -1.

γfdz=2πi(1)=2πi\int_\gamma f\, dz = 2\pi i \cdot (-1) = -2\pi i.

If you get this wrong, revise: Section 6.2 (CIF for Derivatives) and 8.4 (Residues).

Problem 19

Use the Cauchy-Riemann equations to show that f(z)=z2+2zˉf(z) = |z|^2 + 2\bar{z} is differentiable at Exactly one point and find f(z)f'(z) there.

Solution

f(z)=x2+y2+2x2iyf(z) = x^2 + y^2 + 2x - 2iy. So u=x2+y2+2xu = x^2 + y^2 + 2x, v=2yv = -2y.

ux=2x+2u_x = 2x + 2, uy=2yu_y = 2y, vx=0v_x = 0, vy=2v_y = -2.

CR: 2x+2=2x=22x + 2 = -2 \Rightarrow x = -2And 2y=0y=02y = 0 \Rightarrow y = 0.

ff is differentiable only at z=2z = -2.

f(2)=ux(2,0)+ivx(2,0)=(2(2)+2)+0=2f'(-2) = u_x(-2, 0) + iv_x(-2, 0) = (2(-2) + 2) + 0 = -2.

If you get this wrong, revise: Section 3.1 (Cauchy-Riemann Equations).

Problem 20

Evaluate γezsinz(zπ)3dz\int_\gamma \frac{e^z \sin z}{(z - \pi)^3}\, dz where γ\gamma is z=4|z| = 4.

Solution

Only z=πz = \pi is inside γ\gamma (a pole of order 33).

By CIF for derivatives: γf(z)(zπ)3dz=2πi2!f(π)\int_\gamma \frac{f(z)}{(z - \pi)^3}\, dz = \frac{2\pi i}{2!}\,f''(\pi) Where f(z)=ezsinzf(z) = e^z \sin z.

f(z)=ezsinz+ezcosz=ez(sinz+cosz)f'(z) = e^z \sin z + e^z \cos z = e^z(\sin z + \cos z). f(z)=ez(sinz+cosz)+ez(coszsinz)=2ezcoszf''(z) = e^z(\sin z + \cos z) + e^z(\cos z - \sin z) = 2e^z \cos z.

f(π)=2eπcosπ=2eπf''(\pi) = 2e^\pi \cos\pi = -2e^\pi.

γezsinz(zπ)3dz=πi(2eπ)=2πieπ\int_\gamma \frac{e^z \sin z}{(z - \pi)^3}\, dz = \pi i \cdot (-2e^\pi) = -2\pi i\, e^\pi.

If you get this wrong, revise: Section 6.2 (Cauchy’s Integral Formula for Derivatives).

Worked Examples

Example 1: Definite integration

Evaluate 02(3x2+2x)dx\displaystyle\int_0^2 (3x^2 + 2x)\,dx.

Solution:

(3x2+2x)dx=x3+x2+c\int (3x^2 + 2x)\,dx = x^3 + x^2 + c

[x3+x2]02=(8+4)(0)=12\left[x^3 + x^2\right]_0^2 = (8 + 4) - (0) = 12

Example 2: Integration by parts

Find xe2xdx\displaystyle\int x e^{2x}\,dx.

Solution:

Let u=x    dudx=1u = x \implies \frac{du}{dx} = 1 and dvdx=e2x    v=12e2x\frac{dv}{dx} = e^{2x} \implies v = \frac{1}{2}e^{2x}.

xe2xdx=x12e2x12e2xdx=xe2x2e2x4+c=e2x(2x1)4+c\int x e^{2x}\,dx = x \cdot \frac{1}{2}e^{2x} - \int \frac{1}{2}e^{2x}\,dx = \frac{x e^{2x}}{2} - \frac{e^{2x}}{4} + c = \frac{e^{2x}(2x - 1)}{4} + c

Summary

  • Holomorphic functions: complex differentiable on an open set; Cauchy-Riemann equations ux=vyu_x = v_y, uy=vxu_y = -v_x are necessary conditions.
  • Cauchy’s integral theorem: γf(z)dz=0\oint_\gamma f(z)\,dz = 0 for ff holomorphic inside and on γ\gamma; Cauchy’s integral formula evaluates f(a)f(a) and all derivatives via contour integrals.
  • Residue theorem: γf(z)dz=2πiRes(f,zk)\oint_\gamma f(z)\,dz = 2\pi i \sum \text{Res}(f, z_k); 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.

Cross-References

TopicSiteLink
Real AnalysisWyattsNotesView
Multivariable CalculusWyattsNotesView
Differential EquationsWyattsNotesView
Complex Analysis — MIT 18.04MIT OCWView