We're solving Poisson's equation in Electrodynamics, and one ODE we must solve is: $$\dfrac{d^2 R}{dr^2}+\dfrac{2}{r}\dfrac{dR}{dr}-\dfrac{l(l+1)}{r^2}R=0$$ where $l\in\mathbb{R}$. I already know that the solution is in the form $R=Ar^l+\dfrac{B}{r^{l+1}}$ but I want to solve it via power series (instead of the miracle Ansatz). I assumed that the solution takes the form: $$R=\sum_{i=0}^{\infty}a_ir^i$$ such that when I plug this in the ODE I get: $$\sum_{i=0}^{\infty}\left[i(i-1)+a_i+2ia_i-l(l+1) \right]r^{i-2}=0$$ and so $$a_i=\dfrac{l(l+1)}{i(i+1)}$$ My solution is then $$R=l(l+1)\sum_{i=0}^{\infty}\dfrac{r^i}{i(i+1)}=l(l+1)\left[\dfrac{1}{2}r+\dfrac{1}{6}r^2+\dfrac{1}{12}r^3+\dfrac{1}{20}r^4+\dots \right]$$ This solution doesn't seem to match with the actual solution. Where is my mistake?
Solving the radial part of Poisson's equation via power series method
331 Views Asked by Bumbble Comm https://math.techqa.club/user/bumbble-comm/detail AtThere are 3 best solutions below
On
Hint:)
For differential equation $$\dfrac{d^2 R}{dr^2}+\dfrac{2}{r}\dfrac{dR}{dr}-\dfrac{\ell(\ell+1)}{r^2}R=0$$ we see $p(x)=\dfrac{2}{r}$ and $q(x)=-\dfrac{\ell(\ell+1)}{r^2}$ (comparison with $y''+p(r)y'+q(r)y=0$) which show $r=0$ is regular point of this equation, then $$p_0=\lim_{r\to0}rp(r)=2~~~,~~~q_0=\lim_{r\to0}r^2q(r)=-\ell(\ell+1)$$ so the Ferobenius characteristic equation $\lambda(\lambda+1)+p_0\lambda+q_0=0$ is $$\lambda(\lambda+1)+2\lambda-\ell(\ell+1)=0$$ hence $\lambda=\ell$ and $\lambda=-\ell-1$, and two answers must be of the form $$R=x^\ell\sum_{r=0}^\infty a_nr^2~~~~~,~~~~~R=x^{-\ell-1}\sum_{r=0}^\infty a_nr^2$$ so you must consider these series as solutions.
On
Let's put the equation in the form $$ r^2R''(r)+r R'(r)-l(l+1)R(r)=0\tag 1 $$ We look for a series solution of the form (Frobenius Method) $$ R(r)=\sum_{n=0}^{\infty}a_nr^{n+s}\tag 2 $$ Substituting $(2)$ in $(1)$ we have $$ \sum_{n=0}^{\infty}(n+s)(n+s-1)a_nr^{n+s}+2\sum_{n=0}^{\infty}(n+s)a_nr^{n+s}-l(l+1)\sum_{n=0}^{\infty}a_nr^{n+s}=0\tag 3 $$ We then re-order the sums to contain the same power of $r$. We find the indicial equation by setting $n=0$ in all the sums of $(3)$ that bear the lowest exponent of $r$. Since all the sums in $(3)$ are to the same power of $r$ (i.e, $r^{n+s}$), all the sums contribute to the indicial equation and we obtain for $a_0\ne 0$: $$ s(s-1)+2s-l(l+1)=0\tag 4 $$ and solving $(4)$ we find $$ s=l\quad\text{and}\quad s=-(l+1)\tag 5 $$ We next find the recursion relation. Note that since all the sums already involve the same exponent of $r$, there is no need to do any re-indexing. That means that all the coefficients will remain expressed as $a_n$, and so we can write immediately: $$ \Big[(n+s)(n+s-1)+2(n+s)-l(l+1)\Big]a_n=0\tag 6 $$ Equation $(6)$ tells us that either the bracketed expression vanishes or the value of $a_n$ vanishes. We know that the bracketed expression vanishes for $n=0$. However, for any other value of $n$, the bracketed expression cannot vanish, meaning that $a_n = 0$ for all values of $n > 0$. This means the only term in the series expansion is the $a_0$ term, so our solution for equation $(1)$ becomes simply:
$$ R(r)=Ar^l+Br^{-(l+1)} $$
The problem is your third equation. The result of plugging your guess into the original ode gives you
\begin{eqnarray} \sum_i a_i\left[i(i-1) +2i -l(l+1) \right]r^{i-2} = 0 \end{eqnarray}