0
$\begingroup$

A time dependent point charge $q\left ( t \right )$ at the origin $\rho\left ( \vec{r},t \right )=q\left ( t \right )\delta ^{3}\left ( \vec{r} \right ) $, is fed by a current $$\vec{J}\left ( \vec{r},t \right )=-\left ( \frac{1}{4 \pi} \right )\left ( \frac{\dot{q}}{r^{2}} \right )\hat{r}$$ where $\dot{q}=\frac{\mathrm{dq} }{\mathrm{d} t}$

I want to determine that $\nabla\cdot \vec{J}=-\frac{\partial \rho}{\partial t}$ but the integrals are giving me a terribly hard time.

Here's my attempt: $\vec{\nabla} \cdot \vec{J}=0$

$\frac{\partial \rho}{\partial t}$, as I've worked, is $\dot{q}\delta ^{3}\left ( \vec{r} \right )$

but it doesn't seem to satisfy the continuity equation. Is this a problem with the physics or the mathematics?

$\endgroup$
  • $\begingroup$ Be careful at the origin $\nabla\cdot\vec{J}\neq 0$. $\endgroup$ – James Rowland Aug 11 '16 at 4:59
2
$\begingroup$

We know that the two answers should agree, but indeed, you found that $\nabla\cdot J=0$ by taking derivatives everywhere BUT at the origin. At the origin the result is obviously infinite. However this infinity is special (we expect it to be a delta function at the origin).

One way to detect delta functions at the origin is by taking an integral on a sphere around the origin and apply the divergence theorem:

$$\int_{B=\{r\leq R>0\}}\nabla\cdot J ~dV= \int_{\partial B}\mathbf{J}\cdot d\mathbf{S}=\int_{r=R} R^2\sin\theta d\theta d\phi\Big(-\frac{\dot{q}}{4\pi R^2}\Big)=-\dot{q}$$

Then we let $R\to 0$. The only function that can provide a contribution to an integral as one shrinks the region of integration around a certain point, is a delta-function at that point.

We see that $\nabla\cdot J=-\dot{q}\delta^{3}(\mathbf{r})$ satisfies the equation above and therefore this is the solution you are looking for. A full mathematical treatment would go into using Schwarz class test functions to deal with the quantity of interest as a distribution. However, physicists that deal with electromagnetism consider a fact of life that the following identities hold:

$$\nabla^2\Big(\frac{1}{r}\Big)=4\pi\delta^{3}(\mathbf{r})~~~,~~~ \nabla\cdot\Big(\frac{\hat{\mathbf{r}}}{r^2}\Big)=-4\pi\delta^{3}(\mathbf{r})$$

$\endgroup$
0
$\begingroup$

When you work with Dirac delta functions, don't use radial or cylindrical coordinates. They tend to exclude the origin so that a Dirac delta at the origin would not contribute to the integral. Try to reformulate your problem in Cartesian coordinate that you can integrate from -infinity to infinity.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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