I'm trying to show that

$$\log(z-z') = \log(-(z'-A)) - \sum_{n=1}^\infty \frac 1 n \left(\frac{z-A}{z'-A}\right)^n $$ where $A$ is a constant, and $\left|\dfrac{z-A}{z'-A}\right| < 1$.

If it matters, this question comes from a vector calculus problem where $A$ is the centroid of a region containing $z$, and $z'$ is from another region. Anyways, my attempt is:

We can split up the $\log$ term by adding and subtracting $A$: $$\log(z-z') = \log(z-A - (z' - A)) = \log(- (z' - A)) + \log(z-A)$$ so all that needs to be shown is that $$\log(z-A) \overset{\text{?}} = - \sum_{n=1}^\infty \frac 1 n \left(\frac{z-A}{z'-A}\right)^n \quad, $$ which is where I'm stuck. I can't find any expressions for power series of $\log$ which have a $1/n$ term like this.

  • $\begingroup$ Your splitting of the $\log$ term is incorrect. We don't have $\log (u+v) = \log u + \log v$, but with suitable choice of the branches of the logarithm, we have $\log (u\cdot v) = \log u + \log v$. $\endgroup$ – Daniel Fischer May 10 '17 at 15:54
  • $\begingroup$ @DanielFischer You're right, thanks. What a rookie mistake! $\endgroup$ – Alex May 10 '17 at 15:57
  • $\begingroup$ @Dr.MV Oh. They're complex. $\endgroup$ – Alex May 10 '17 at 16:11

Recall that the Taylor series (See Here) for the complex function $f(z)=\log(1-z)$ is

$$\log(1-z)=-\sum_{n=1}^\infty \frac{z^n}{n}\tag 1$$

for $|z|<1$ and the branch cut does not intersect any point inside the unit disk.

Using $(1)$, we can write

$$\begin{align} \log(z-z')&=\log((z-A)-(z'-A))\\\\ &=\log(A-z')+\log\left(1-\frac{z-A}{z'-A}\right) \tag 1\\\\ &=\log(A-z')-\sum_{n=1}^\infty \frac{\left(\frac{z-A}{z'-A}\right)^n}{n} \end{align}$$

for $\left|\frac{z-A}{z'-A}\right|<1$ as was to be shown!

Note that $(1)$ is interpreted as a set equivalence (See Here). For any branch of $\log(z-z')$, there is some branch of $\log(A-z')$ and some branch of $\log\left(1-\frac{z-A}{z'-A}\right)$ for which $(1)$ holds. Conversely, it means for any branch of $\log(A-z')$ and any branch of $\log\left(1-\frac{z-A}{z'-A}\right)$, there is some branch of $\log(z-z')$ for which $(1)$ is true.

  • $\begingroup$ Alex, please let me know how I can improve my answer. I really want to give you the best answer I can. -Mark $\endgroup$ – Mark Viola May 23 '17 at 0:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.