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!

Deriving potential distribution between concentric spherical electrode

  1. Sep 26, 2013 #1
    1. The problem statement, all variables and given/known data

    For some work I am doing I wish to be able to define the potential distribution as a function of the radius (ρ) between two concentric electrodes.


    2. Relevant equations

    One solution (from reliable literature) defines the varying radial potential as:

    V(ρ)=2V0(ρ0/ρ -1)

    Where V0 and ρ0 are the potential and radius of the optic axis respectively.


    3. The attempt at a solution

    Taking the laplace equation in spherical coordinates I find the potential is given by:

    [itex]\frac{∂^{2}V}{∂ρ^{2}}[/itex]=[itex]\frac{1}{ρ^{2}}[/itex][itex]\frac{∂}{∂ρ}[/itex][itex](ρ^{2}[/itex] [itex]\frac{∂V}{∂ρ}[/itex][itex])=0[/itex]

    This assumes there is no variance in potential in θ or [itex]\varphi[/itex] which I'm confident is correct providing there is a uniform surface charge distribution across both electrodes. However the steps after this become pretty complex involving Legendre polynomials. I'm certainly no where near deriving the above equation.

    Any help or suggestions anyone may have would be very welcome!
     
  2. jcsd
  3. Sep 26, 2013 #2
    Hello.
    It is possible to solve this equation without using Legendre polynomials. Just remember that:
    [tex]\displaystyle{\frac{\partial}{\partial \rho }\left(\rho ^2\frac{\partial V}{\partial \rho } \right)=0\Rightarrow \rho ^2\frac{dV}{d\rho }=C}[/tex]
    which is easy to solve. You have to find two constants using two known values of the potential. I am not sure which are these values at your exercise, but I can find a solution which is the same as yours.
     
  4. Sep 27, 2013 #3
    Thats a great help thank you.

    So I rerrange and integrate to give:

    [itex]\frac{dV}{dρ}=[/itex][itex]\frac{C}{ρ^{2}}[/itex]

    [itex]\int\frac{C}{ρ^{2}}dρ=\frac{-C}{ρ}+D[/itex]

    So:

    [itex]V(ρ)=\frac{-C}{ρ}+D[/itex]

    I know that at the optic axis (ρ0), V(ρ)=V0 which leaves me with:

    [itex]V_{0}=\frac{-C}{ρ_{0}}+D[/itex]

    I'm struggling to work out what the other boundary may have been used for the quoted solution in my previous post. If I examine ρ=∞ where V=0, then I get:

    [itex]0=0+D[/itex]

    Thus V(ρ) is simply:

    [itex]V(ρ)=\frac{V_{0}ρ_{0}}{ρ}[/itex]

    But this clearly doesn't match my quoted solution:

    [itex]V(ρ)=2V_{0}(\frac{ρ_{0}}{ρ}-1)[/itex]

    Thanks for the reply Stealth, very helpful!

    P.S. I'm unsure as to where the [itex]1/ρ^{2}[/itex] has gone in your previous post Stealth
     
    Last edited: Sep 27, 2013
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted