I know that the density and potential (in spherycals) of a charged ring is, respectively,:

$$ \rho(\textbf{r}) = \frac{\lambda}{a} \delta(r-a)\delta(\theta-\tfrac{\pi}{2}) $$

$$ \varphi(\textbf{r})= \frac{2\pi a \lambda}{r_>} \left[ 1+ \sum_{n=1}^\infty (-1)^n \frac{(2n-1)!!}{(2n)!!}\left(\frac{r_<}{r_>}\right)^{2n}P_{2n}(\cos\theta) \right] $$

Where $P_{2n}$ is the $2n$-th Legendre Polynomial, and $r_<=\min\{a,r\},r_>=\max\{a,r\}$. If I evaluate $\mathbf r$ in the ring ($r=a,\theta=\tfrac{\pi}{2}$):

$$ \varphi(\mathbf r)\,\propto\, \left[ 1+ \sum_{n=1}^\infty (-1)^n \frac{(2n-1)!!}{(2n)!!} \right] \to\infty $$

So this is a problem (I suppose).

  • $\begingroup$ Actually, I'm rethinking my answer. Are you sure that your sum for the energy does not also have a term like $P_{2n}(0)$? My answer was assuming that the sum was as you stated, but in trying to derive your result, I seem to have gotten something different. $\endgroup$
    – Philip
    Jun 4, 2020 at 12:38
  • $\begingroup$ @Philip Thank you so much for your answer, you have a recurrence relation for the Polynomials Legendre $(n+1)P_{n+1}=(2n+1)xP_n(x) - nP_{n-1}(x)$, and $P_0(0)=1,\quad P_1(0)=0$, so you can say $P_{2n}(0)=(-1)^n\frac{(2n-1)!!}{(2n)!!}$ and $P_{2n+1}(0)=0$ for $n>0$ $\endgroup$ Jun 4, 2020 at 17:22
  • 1
    $\begingroup$ Oh great! I never paid as much attention as I should have to recurrence relations. But in that case, would the formula for the potential $\varphi(\mathbf{r})$ also have the same double factorial terms in it? I'm afraid that given your formula for $\varphi$, I can't derive your formula for $U_E$. It might be a typo? Also, does my answer answer your question? :) $\endgroup$
    – Philip
    Jun 4, 2020 at 17:43
  • $\begingroup$ It is the continous form of the energy for discrete charges. You answer my question, thank you. $\endgroup$ Jun 4, 2020 at 21:41

1 Answer 1


This is actually a fun question, I learnt something new about double factorials while trying to answer it!

I don't see why that term diverges. Using the identities on Wikipedia for the "double factorial", we have that for even integers $k$,

$$\int_0^{\pi/2} \sin^{k}(x)\text{d}x = \frac{(k-1)!!}{(k)!!}\frac{\pi}{2}.$$

We can use this to calculate the sum term you have explicitly.

$$\sum_{n=1}^\infty (-1)^n \frac{(2n-1)!!}{(2n)!!} = \frac{2}{\pi}\sum_{n=1}^\infty (-1)^n \int_0^{\pi/2} \sin^{2n}(x)\text{d}x = \frac{2}{\pi}\int_0^{\pi/2} \text{d}x \sum_{n=1}^\infty (-1)^n \sin^{2n}(x).$$

Where in the last step I've interchanged the sum and the integral. This particular sum is quite easy to do, and I'll leave it as an exercise to show that

$$\sum_{n=1}^\infty (-1)^n \sin^{2n}(x) = -\frac{\sin^2(x)}{1+\sin^2(x)}.$$

We can now perform the integral and show that $$-\frac{2}{\pi}\int_0^{\pi/2} \frac{\sin^2(x)}{1+\sin^2(x)} \text{d}x = \frac{-2 + \sqrt{2}}{2}.$$

Thus, $$\sum_{n=1}^\infty (-1)^n \frac{(2n-1)!!}{(2n)!!} = \frac{-2 + \sqrt{2}}{2} < \infty,$$

which should solve your problem.

  • $\begingroup$ Thank you very much, it works to me!! $\endgroup$ Jun 4, 2020 at 21:37
  • $\begingroup$ I'm going to change the question, because I have an error. $\endgroup$ Jun 9, 2020 at 15:00
  • $\begingroup$ Here is the new question $\endgroup$ Jun 9, 2020 at 15:02
  • $\begingroup$ @Elborito I think it might be better to just edit this question, perhaps? I think repeating questions is frowned up. Also, I am not sure how you got the $$\left(\frac{(2n-1)!!}{(2n)!!}\right)^2$$ term, since I only got one which came from the Legendre Polynomial recurrence relations. Perhaps you could edit the question with a little bit more of your working? $\endgroup$
    – Philip
    Jun 9, 2020 at 15:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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