3
$\begingroup$

I need to integrate, $\int\limits_{|z| = R} \frac{|dz|}{|z-a|^2}$ where $a$ is a complex number such that $|a|\ne R$.

So first I tried polar coordinates, which gives something I cannot continue.

Then I tried to write $|dz| = rd\theta = dz/ie^{i\theta}$ and I have $$\int\limits_{|z| = R} \frac{dz}{ie^{i\theta}(z-a)(\overline{z}-\overline{a})}$$

which makes me want to use cauchy's integral formula, but I'm not sure if it has a pole at $z = a$ or not.

How to I calculate this integral?

$\endgroup$
5
$\begingroup$

First sub $z=R e^{i \phi}$, $dz = i R e^{i \phi} d\phi$, $|dz|=R d\phi$. Then realize that

$$|z-a|^2 = R^2 + |a|^2 - 2 R |a| \cos{\phi}$$

(I set an arbitrary phase to zero - it won't matter for the integration.)

The integral then becomes

$$R \int_0^{2 \pi} \frac{d\phi}{R^2 + |a|^2 - 2 |a| R \cos{\phi}}$$

Now - and this might seem weird - we go back to a complex representation so we may evaluate the integral using the residue theorem. That is, set $\zeta = e^{i \phi}$, $d\phi = -i d\zeta/\zeta$ and get that the integral is equal to

$$i R \oint_{|\zeta|=1} \frac{d\zeta}{|a| R \zeta^2 - (|a|^2+R^2) \zeta + |a| R} $$

To evaluate via the residue theorem, we find the poles of the integrand, which are at $\zeta=|a|/R$ and $\zeta=R/|a|$. Clearly, the analysis depends on whether $|a|$ is greater than or less than $R$. For example, when $|a| \lt R$, the integral is, by the residue theorem,

$$i 2 \pi (i R) \frac{1}{2 |a| R (|a|/R) - |a|^2-R^2} = \frac{2 \pi R}{R^2-|a|^2}$$

The analysis is similar for $R \lt |a|$. The end result is that

$$\oint_{|z|=R} \frac{|dz|}{|z-a|^2} = \frac{2 \pi R}{\left|R^2-|a|^2 \right|}$$

$\endgroup$
  • $\begingroup$ How do you get this: $|z-a|^2 = R^2 + |a|^2 - 2 R |a| \cos{\phi}$? I only get $|z-a|^2 = R^2+|a|^2-z\overline{a}-\overline{z}a$ $\endgroup$ – mez Oct 15 '13 at 20:10
  • $\begingroup$ $z \bar{a} + \bar{z} a = 2 \Re{(z \bar{a})}$. Then write $z$ and $a$ in polar form. $\endgroup$ – Ron Gordon Oct 15 '13 at 20:11
  • 1
    $\begingroup$ Wow, what made you so fast in spotting that you should substitute $\zeta = e^{i\phi}$ back? $\endgroup$ – mez Oct 15 '13 at 20:39
  • 1
    $\begingroup$ @mezhang: it is simply one approach to evaluating the integral. In this case, any rational function of sine and cosine integrated from $0$ to $2 \pi$ may be evaluated using that substitution and the residue theorem fairly easily, as you can see. $\endgroup$ – Ron Gordon Oct 15 '13 at 20:40
1
$\begingroup$

You may assume $a=|a|>0$. On $\gamma:=\partial D_R$ one has $\bar z={R^2\over z}$; furthermore the parametrization $$\gamma:\quad \phi\mapsto z=R e^{i\phi}\qquad(0\leq\phi\leq2\pi)$$ gives $dz=iR e^{i\phi}\ d\phi$ and therefore $$|dz|=Rd\phi=-i{R\over z}\ dz\ .$$ (Complexifying the real $|dz|$ in this way is a trick that will enable us to use the residue theorem later on.)

Altogether it follows that your integral ($=: J$) can be written as $$J=\int\nolimits_\gamma{-i R/z\over(z-a)\bigl({R^2\over z}-a\bigr)}\ dz={iR\over a}\int\nolimits_\gamma{dz\over(z-a)\bigl(z-{R^2\over a}\bigr)}\ .$$ The last integral can now be evaluated by means of the residue theorem. When $a<R$ we have a single pole in $D_R$ at $z=a$, and standard computation rules tell us that $$J=2\pi i\cdot {iR\over a}{1\over a-{R^2\over a}}={2\pi R\over R^2-a^2}\ .$$ I leave the case $a>R$ to you.

$\endgroup$

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.