Skip to Content
Math416Complex Variables (Lecture 7)

Math416 Lecture 7

Review

Exponential function

ez=ex+iy=ex(cosy+isiny)e^z=e^{x+iy}=e^x(\cos y+i\sin y)

Logarithm Reviews

Definition 4.9 Logarithm

A logarithm of aa is any bb such that eb=ae^b=a.

Branch of Logarithm

A branch of logarithm is a continuous function ff on a domain DD such that ef(z)=exp(f(z))=ze^{f(z)}=\exp(f(z))=z for all zDz\in D.

Continue on Chapter 4 Elementary functions

Logarithm

Theorem 4.11

log(z)\log(z) is holomorphic on C{0}\mathbb{C}\setminus\{0\}.

Proof

We proved that zez=0\frac{\partial}{\partial\overline{z}}e^{z}=0 on C{0}\mathbb{C}\setminus\{0\}.

Then ddzez=xez=0\frac{d}{dz}e^{z}=\frac{\partial}{\partial x}e^{z}=0 if we know that eze^{z} is holomorphic.

Since ddzez=ez\frac{d}{dz}e^{z}=e^{z}, we know that eze^{z} is conformal, so any branch of logarithm is also conformal.

Since exp(log(z))=z\exp(\log(z))=z, we know that log(z)\log(z) is the inverse of exp(z)\exp(z), so ddzlog(z)=1elog(z)=1z\frac{d}{dz}\log(z)=\frac{1}{e^{\log(z)}}=\frac{1}{z}.

We call ff\frac{f'}{f} the logarithmic derivative of ff.

Definition 4.16

I don’t know if this material is covered or not, so I will add it here to prevent confusion for future readers

If aa and cc are complex numbers, with a0a\neq 0, then by the values of aca^c one means the value of eclogae^{c\log a}.

For example, 1i=ei(2πni)1^i=e^{i (2\pi n i)}

If you accidentally continue on this section and find it interesting, you will find Riemann zeta function

z(s)=n=11nsz(s)=\sum_{n=1}^{\infty}\frac{1}{n^s}

And analytic continuation for such function for number less than or equal to 11.

And perhaps find trivial zeros for negative integers on real line. It is important to note that the Riemann zeta function has non-trivial zeros, which are located in the critical strip where the real part of ss is between 0 and 1. The famous Riemann Hypothesis conjectures that all non-trivial zeros lie on the critical line where the real part of ss is 12\frac{1}{2}.

Chapter 5. Power series

Convergence

Necessary Condition for Convergence

If n=0cn\sum_{n=0}^{\infty}c_n converges, then limncn=0\lim_{n\to\infty}c_n=0 exists.

Geometric series

Let cc be a complex number

n=0Ncn=1cN+11c\sum_{n=0}^{N}c^n=\frac{1-c^{N+1}}{1-c}

If c<1|c|<1, then limNn=0Ncn=11c\lim_{N\to\infty}\sum_{n=0}^{N}c^n=\frac{1}{1-c}.

otherwise, the series diverges.

Proof

The geometric series converges if cN+11c\frac{c^{N+1}}{1-c} converges.

(1c)(1+c+c2++cN)=1cN+1(1-c)(1+c+c^2+\cdots+c^N)=1-c^{N+1}

If c<1|c|<1, then limNcN+1=0\lim_{N\to\infty}c^{N+1}=0, so limN(1c)(1+c+c2++cN)=1\lim_{N\to\infty}(1-c)(1+c+c^2+\cdots+c^N)=1.

If c1|c|\geq 1, then cN+1c^{N+1} does not converge to 0, so the series diverges.

Theorem 5.4 (Triangle Inequality for Series)

If the series n=0cn\sum_{n=0}^{\infty}c_n converges, then n=0cnn=0cn\left|\sum_{n=0}^\infty c_n\right|\leq \sum_{n=0}^{\infty}|c_n|.

Definition 5.5

n=0cn\sum_{n=0}^{\infty}c_n

converges absolutely if n=0cn\sum_{n=0}^{\infty}|c_n| converges.

Note: Some other properties of converging series covered in Math4111, bad, very bad.

Definition 5.6 Convergence of sequence of functions

A sequence of functions fnf_n converges pointwise to ff on a set GG if for every zGz\in G, ϵ>0\forall\epsilon>0, N\exists N such that for all nNn\geq N, fn(z)f(z)<ϵ|f_n(z)-f(z)|<\epsilon.

(choose NN based on zz)

A sequence of functions fnf_n converges uniformly to ff on a set GG if for every ϵ>0\epsilon>0, there exists a positive integer NN such that for all nNn\geq N and all zGz\in G, fn(z)f(z)<ϵ|f_n(z)-f(z)|<\epsilon.

(choose NN based on ϵ\epsilon)

A sequence of functions fnf_n converges locally uniformly to ff on a set GG if for every zGz\in G, ϵ>0\forall\epsilon>0, r>0\exists r>0 such that for all zB(z,r)z\in B(z,r), nN\forall n\geq N, fn(z)f(z)<ϵ|f_n(z)-f(z)|<\epsilon.

(choose NN based on zz and ϵ\epsilon)

A sequence of functions fnf_n converges uniformly on compacta to ff on a set GG if it converges uniformly on every compact subset of GG.

Theorem 5.7

If the subsequence (or partial sum) of a converging sequence of functions converges (a), then the original sequence converges (a).

The N-th partial sum of the series n=0fn\sum_{n=0}^\infty f_n is n=0Nfn\sum_{n=0}^{N}f_n

You can replace (a) with locally uniform convergence, uniform convergence, pointwise convergence, etc.

Corollary from definition of aba^b in complex plane

We defined ab={ebloga}a^b=\{e^{b\log a}\} if bb is real, then aba^b is unique, if bb is complex, then ab=ebloga{e2kπikb},kZa^b=e^{b\log a}\{e^{2k\pi ik b}\},k\in\mathbb{Z}.

Power series

Definition 5.8

A power series is a series of the form n=0cn(zz0)n\sum_{n=0}^{\infty}c_n(z-z_0)^n.

Definition 5.9 Region of Convergence

For every power series, there exists a radius of convergence rr such that the series converges absolutely and locally uniformly on Br(z0)B_r(z_0).

And it diverges pointwise outside Br(z0)B_r(z_0).

Proof

Without loss of generality, we can assume that z0=0z_0=0.

Suppose that the power series is n=0cn(z)n\sum_{n=0}^{\infty}c_n (z)^n converges at z=reiθz=re^{i\theta}.

We want to show that the series converges absolutely and uniformly on Br(0)\overline{B_r(0)} (closed disk, I prefer to use this notation, although they use D\mathbb{D} for the disk (open disk)).

We know cnrneinθ0c_n r^ne^{in\theta}\to 0 as nn\to\infty.

So there exists McnrneinθM\geq|c_n r^ne^{in\theta}| for all nNn\in\mathbb{N}.

So zBr(0)\forall z\in\overline{B_r(0)}, cnzncnznM(zr)n|c_nz^n|\leq |c_n| |z|^n \leq M \left(\frac{|z|}{r}\right)^n.

So n=0cnzn\sum_{n=0}^{\infty}|c_nz^n| converges absolutely.

So the series converges absolutely and uniformly on Br(0)\overline{B_r(0)}.

If z>r|z| > r, then cnzn|c_n z^n| does not tend to zero, and the series diverges.

We denote this rr captialized by te radius of convergence

Possible Cases for the Convergence of Power Series

  1. Convergence Only at z=0z = 0:

    • Proof: If the power series n=0cn(zz0)n\sum_{n=0}^{\infty} c_n (z - z_0)^n converges only at z=0z = 0, it means that the radius of convergence R=0R = 0. This occurs when the terms cn(zz0)nc_n (z - z_0)^n do not tend to zero for any z0z \neq 0. The series diverges for all z0z \neq 0 because the terms grow without bound.
  2. Convergence Everywhere:

    • Proof: If the power series converges for all zCz \in \mathbb{C}, the radius of convergence R=R = \infty. This implies that the terms cn(zz0)nc_n (z - z_0)^n tend to zero for all zz. This can happen if the coefficients cnc_n decrease rapidly enough, such as in the exponential series.
  3. Convergence Within a Finite Radius:

    • Proof: For a power series with a finite radius of convergence RR, the series converges absolutely and uniformly for zz0<R|z - z_0| < R and diverges for zz0>R|z - z_0| > R. On the boundary zz0=R|z - z_0| = R, the series may converge or diverge depending on the specific series. This is determined by the behavior of the terms on the boundary.
Last updated on