0
$\begingroup$

I've been trying to follow this post on deriving the Biot-Savart law from Maxwell's Equations but am getting stuck on this step:

$$-\frac{\mu_0}{4\pi}\iiint{\nabla\times\frac{J}{|r-r'|} d^3r'}.$$

Mainly the curl of current density part which would expand to:

$$\nabla\times\frac{J}{|r-r'|}=\frac{\nabla\times J}{|r-r'|}+\nabla\frac{1}{|r-r'|}\times J$$

The only way this would reduce to $\frac{|r-r'|\times J}{|r-r'|^3}$ would be if $\nabla\times J=0$.

So I tried showing that $\nabla\times J = 0$ in a steady-state situation, however, I'm getting stuck.

Assumption: Steady-state current which implies that for all of space and time $\frac{\partial B}{\partial t}=0$

We start with Ampere's law,

$$\nabla\times B=\mu_0J+\mu_0\epsilon_0\frac{\partial E}{\partial t}$$

Curl of both sides

$$-\nabla^2B=\mu_0\nabla\times J+\mu_0\epsilon_0\frac{\partial(\nabla\times E)}{\partial t}$$

And because $\nabla\times E = -\frac{\partial B}{\partial t}$

$$-\nabla^2B=\mu_0\nabla\times J-\mu_0\epsilon_0\frac{\partial^2 B}{\partial t^2}$$

$$\mu_0\epsilon_0\frac{\partial^2 B}{\partial t^2}=\mu_0\nabla\times J+\nabla^2B$$

However, this is where I'm getting stuck, there are two scenarios for $\partial^2 B / \partial t^2 = 0$: Either $\nabla^2 B$ is zero meaning that $\nabla\times J$ would also be zero, or that $\nabla^2 B$ is non-zero meaning that $\nabla\times J=-\nabla^2 B$.

So now I would need to show that $\nabla^2 B=0$ which I'm not sure is true (For example, wouldn't it be $\infty$ at the wire - assuming thin wire?)

So my question is: is $\nabla\times J=0$ in steady-state? If so, where am I going wrong in my proof?

$\endgroup$

1 Answer 1

4
$\begingroup$

This is a consequence of bad notation.

In writing the formula:

$$\nabla × \vec{A} = \vec{B}$$

You need to be very careful about which variables you are taking the curl of.

You should write:

$$\nabla_{\vec{r}} × \vec{A}(\vec{r}) = \vec{B}(\vec{r})$$

This is because we are taking the curl with respect to the variables $\vec{r} = x\hat i + y\hat j + z\hat k$

When we apply the $\nabla_{\vec{r}} ×$ operator on $\vec{J}$ we need to keep track of what $\vec{J}$ is a function of.

Your confusion lies in the fact that yes, in maxwells equations, $\vec{J}$ and $\vec{B}$ are both functions of $\vec{r}$.

However, when solving poissons equation, $\vec{J}(\vec{r})$ changes to $\vec{J}(\vec{r}')$

Meaning,

$\nabla_{\vec{r}} × \vec{J}(\vec{r}') = 0$

As $\vec{J}(\vec{r}')$ is independant on $\vec{r}$

So your conclusions you've drawn are presumably correct if $\nabla_{\vec{r}'} × \vec{J}(\vec{r}') = 0$, but that's not what we are saying when deriving biot savart.

$\endgroup$
5
  • $\begingroup$ Oh wow, I'm a newbie to this whole area of math and physics, and I keep making dumb mistakes like this, thanks so much for your help :) So curl of J is zero in the Biot-Savart law derivation because the curl with respect to r on a function of r' is zero. Math is hard ;-; $\endgroup$
    – nreh
    Commented Jun 4, 2022 at 12:47
  • $\begingroup$ Yup, which is why notation is very important! The curl is basically a fancy derivative, d/dx (x') = 0 as x' is a constant as far as x is concerned! $\endgroup$ Commented Jun 4, 2022 at 12:52
  • $\begingroup$ Hey I ran through the problem again and now I'm confused about something else, I used the rule for 'curl of a scalar field times a vector field' to expand $\nabla_r \times J(r')/|r'|$. However, because $J(r')$ is a constant as far r is concerned, it would become a scalar field times a constant vector. Wouldn't this make $\nabla_r \times J(r')/|r'|=0$? Which isn't done by the other derivations I've seen. $\endgroup$
    – nreh
    Commented Jun 4, 2022 at 14:40
  • $\begingroup$ I'm not sure I follow.you are forgetting the expression we are taking the curl of has 1/|r-r'| in the denominator, which IS a function of r, and not 1/|r'| $\endgroup$ Commented Jun 4, 2022 at 14:51
  • $\begingroup$ Sorry yeah you're right, I got the equation wrong. Thanks a lot for your help. $\endgroup$
    – nreh
    Commented Jun 4, 2022 at 15:01

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.