1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Homework Help: Expansion of complex function

  1. Jun 13, 2012 #1
    I have an exercise that says the following:

    Expand the following function as a series:

    f(z) = [itex]\frac{1}{z-1}[/itex] + [itex]\frac{1}{2-z}[/itex] for 1<lzl<2

    The result is attached, but I don't really understand what has been done. Therefore tell me:
    How is that series generated? Initially I thought I should do a taylor expansion but then that is always expanded about a specific point, which is not given in the exercise.

    Attached Files:

  2. jcsd
  3. Jun 13, 2012 #2


    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper
    Education Advisor

    You're expanding about z=0. You can tell from the region for which you're asked to find the series. It's an annulus centered at z=0.

    Look up Laurent series.
  4. Jun 13, 2012 #3
    okay, I have done and I think I understand it mostly. But I still don't understand how you see that one MUST expand about z=0 - isn't that just done because it is the easiest?
  5. Jun 13, 2012 #4
    Hmm.. Now I've been looking a lot on this exercise, and you need to help me. I do understand that a function which is not analytic at certain singularities can be expanded as a laurent series about them. However, in this exercise f is only defined in the region 1<lzl<2 and is thus analytic everywhere. Therefore, why would a taylor series about a point z0 which satisfies 1<lz0l<2 not work?
    I don't really understand what is done to achieve the attached result. The solution notes says that the first expression (which I guess must be 1/(z-1)) is written as a series for z>1 and the second expression (which I again must guess refers to 1/(2-z)) is expanded for z<2. But what on earth is this supposed to mean? If the functions are expanded around z=0 the above can't be true, and either way you get terms for (1/z)k as though a laurent series is written.

    Can you please show how to do it, and I will learn from that rather than having to guess my way out of this (my book's chapter on Laurent series is VERY poor).
  6. Jun 13, 2012 #5
    1 < |z-0|< 2

    If it were centered anywhere else, the 0 would be replaced by that number. Remember that |z| < 2 => z in D(0;2)

    (I'm on my phone so Latex is a little time consuming to do).
  7. Jun 13, 2012 #6


    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper
    Education Advisor

    No. Such a series won't converge for all points in the annulus. Suppose we expanded 1/(1-z) about z=3/2, for example. First, do some algebra to get
    $$\frac{1}{1-z} = \frac{1}{1-(z-3/2)-3/2} = -\frac{1}{1/2} \left(\frac{1}{1+2(z-3/2)}\right).$$ Then use the fact that
    $$\frac{1}{1+w} = 1-w+w^2-w^3+\cdots$$ with w=2(z-3/2) to obtain the series
    $$\frac{1}{1-z} = -2\left[1-2(z-3/2)+4(z-3/2)^2-8(z-3/2)^3+\cdots\right] = -2+4(z-3/2)-8(z-3/2)^2+\cdots.$$ (You could also find the Taylor series by taking derivatives, but this method is much faster.) This series will converge only when |w|<1, which means that |z-3/2| < 1/2. In other words, it'll only converge for points in the complex plane contained in a circle of radius 1/2 centered at z=3/2. Most of points in the annulus 1<|z|<2 don't meet this criterion, e.g. z=-3/2.

    Consider the Taylor series for 1/(1-z).
    $$\frac{1}{1-z} = 1+z+z^2+z^3+\cdots.$$ This series is valid if |z|<1; it won't converge if |z|>1. So for |z|>1, you expand in powers of 1/z. That series will converge because |1/z| will be less than 1.
    $$\frac{1}{1-z} = \frac{1}{z(1/z-1)} = -\frac{1}{z}\left(\frac{1}{1-1/z}\right) = -\frac{1}{z}\sum_{n=0}^\infty \left(\frac{1}{z}\right)^n.$$ For your original function, the two singularities divide the complex plane into three regions: |z|<1, 1<|z|<2, and |z|>2. For each region, you expand each term in either powers of z or powers of 1/z so that they'll both converge.

    Last edited: Jun 13, 2012
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook