1
$\begingroup$

I am learning from Jackson (3r edition), where I found one concept very confusing, that is Taylor expansion of charge density. (This is given in section "1.7 Poisson and Laplace equations" p.n:35)

I will write some equations first.

$$ {\Phi}_a(x) = \frac{1}{4{\pi}{\epsilon_0}}\int \frac{{\rho}(x')}{\sqrt{(x - x')^2 + a^2}}d^3x' $$

Now we want to find out potential such that a tends to zero. $$ \nabla^2{\Phi}_a(x) = \frac{1}{4{\pi}{\epsilon_0}}\int{\rho}(x') \nabla^2\frac{1}{(r^2 + a^2)^\frac{1}{2}}d^3x' $$

$$ \nabla^2{\Phi}_a(x) = -\frac{1}{4{\pi}{\epsilon_0}}\int{\rho}(x') \frac{3a^2}{(r^2 + a^2)^\frac{5}{2}}d^3x' $$

I understood this until above step, but now I didn't got the next step

$$ \nabla^2{\Phi}_a(x) = -\frac{1}{{\epsilon_0}}\int_0^R \frac{3a^2}{(r^2 + a^2)^\frac{5}{2}} \left[{\rho}(x) + \frac{r^2}{6}\nabla^2{\rho} + ....... \right]r^2 dr + O(a^2) $$

Jackson says that we gonna expand ${\rho}(x')$ around x' = x., but the expansion of ${\rho}(x')$ should also contain the first order derivative of ${\rho}(x')$ like $\nabla{\rho}$, also the taylor expansion of the second term should contain 2 at the denominator but it's 6, and how the last term of $O(a^2)$

So what I am thinking is that Taylor expansion should be $\left[{\rho}(x) + r\nabla{\rho} + \frac{r^2}{2}\nabla^2{\rho}\right]$

I know I am wrong but I don't know what is the answer. Any help is appreciated.

$\endgroup$
9
  • 4
    $\begingroup$ I think that will help: physics.stackexchange.com/q/441818 $\endgroup$ Commented Aug 14, 2020 at 8:38
  • 1
    $\begingroup$ The second term in your proposed Taylor expansion is dimensionally inconsistent. It can't be $r^{\frac{1}{2}}$. $\endgroup$
    – Philip
    Commented Aug 14, 2020 at 8:39
  • $\begingroup$ @Philip but in taylor's expansion the second term is $(x-x')\nabla{\rho}$, so if this is not, then what should be? $\endgroup$
    – anbhadane
    Commented Aug 14, 2020 at 8:41
  • 1
    $\begingroup$ $x-x' \neq r^\frac{1}{2}$ ;) What I expect you mean is $r \nabla \rho$... $\endgroup$
    – Philip
    Commented Aug 14, 2020 at 8:42
  • 1
    $\begingroup$ @anbhadane This is explained in the paragraph just above the equation that's bothering you in Jackson. I would suggest that you first read that paragraph carefully. If you still have trouble understanding it, I would suggest including it in your question. That would also explain why this question is essentially a duplicate of the one posted by Nathanael. :) $\endgroup$
    – Philip
    Commented Aug 14, 2020 at 8:54

1 Answer 1

1
$\begingroup$

Firstly, your form of Taylor expansion is wrong. The right form is here: https://en.wikipedia.org/wiki/Taylor_series#Taylor_series_in_several_variables . It seems like a special condition with a parameter $R$, and the integrals of some Taylor terms are zero. As it is not mentioned in the book, these zero integrals should be analyzed by symmetry.

$\endgroup$
4
  • $\begingroup$ So is there any article about such zero integrals? $\endgroup$
    – anbhadane
    Commented Aug 14, 2020 at 12:55
  • 1
    $\begingroup$ Sorry, I have not read this book. Do you mention that the coordinate system has been changed from Cartesian coordinates to spherical coordinates and the space angle parts, $\theta$ and $\phi$, have been integrated. Be careful with the parts which should be zero after the progress of space angle integral. $\endgroup$
    – xyzrggong
    Commented Aug 15, 2020 at 13:15
  • $\begingroup$ And if everything have been checked, just continue the calculation. Then compare the conclusions. Maybe the "6" below $r^2$ is wrong. Can you send the pages related to the problem? Then I can calculate it. $\endgroup$
    – xyzrggong
    Commented Aug 15, 2020 at 13:22
  • $\begingroup$ I have calculated the answer and it has "6" below $r^2$. If you want proof then it's so long that it couldn't be posted here, but need to be done on another site such as imgur? anyways Jackson's equation is accurate. We think it with our common perspective but it's not that way. At the top there is link in comment which leads to another post where hint is given for this problem you could try it. $\endgroup$
    – anbhadane
    Commented Aug 15, 2020 at 14:21

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