Proving Lorentz Condition for Retarded Potentials in Griffith 10.8

  • Thread starter Thread starter stunner5000pt
  • Start date Start date
  • Tags Tags
    Potentials
stunner5000pt
Messages
1,443
Reaction score
4

Homework Statement


Griffith's problem 10.8
Show that retarded potentials satisfy the Lorentz condition. Hint proceed as follows
a) Show that
\nabla\cdot\left(\frac{J}{R}\right)=\frac{1}{R}\left(\nabla\cdot\vec{J}\right)+\frac{1}{R}\left(\nabla '\cdot\vec{J}\right)-\nabla '\cdot\left(\frac{J}{R}\right)
b) Show that \nabla\cdot\vec{J}=-\frac{1}{c}\frac{\partial\vec{J}}{\partial t_{r}}\cdot(\nabla R)
c) Note that \vec{J}=\vec{J}\left(\vec{r'},t_{r}\right)
\nabla '\cdot J=-\frac{\partial \rho}{\partial t}-\frac{1}{c}\frac{\partial J}{\partial t_{r}}\cdot (\nabla ' R)

where \vec{R}=\vec{r}-\vec{r'}

2. The attempt at a solution

I managed to do the first and second parts but its the third part that i am unable to prove.
Ok so i know that \vec{J}=\vec{J}\left(\vec{r'},t_{r}\right)=\vec{J}\left(\vec{r'},t-\frac{\vec{r}-\vec{r'}}{c}\right)

To make it simpler for me to understand let's do it for one dimension.
\frac{\partial J_{x}}{\partial x'} = \frac{\partial J_{x}}{\partial t_{r}}\frac{\partial t_{r}}{\partial x'}
But \frac{\partial t_{r}}{\partial x'}=\frac{1}{c}\frac{\partial R}{\partial x'}
so \frac{\partial J_{x}}{\partial x'} = \frac{1}{c}\frac{\partial J_{x}}{\partial t_{r}}\frac{\partial R}{\partial x'}

THat explains the second term which i need to get in the proof. But how do i get the first term?

Also is it supposed to be \nabla\cdot J=-\frac{\partial \rho}{\partial t}
or is it supposed to be \nabla'\cdot J=-\frac{\partial \rho}{\partial t}

Thanks for your help!
 
Physics news on Phys.org
Use the product rules:
\vec \nabla \cdot \left(\frac{\vec J}{R}\right) = {1\over R}\left(\vec{\nabla} \cdot \vec J\right) + \vec J \cdot \vec{\nabla} \left({1\over R}\right)

\vec{\nabla}' \cdot \left(\frac{\vec J}{R}\right) = {1\over R}\left(\vec{\nabla}' \cdot \vec J\right) + \vec J \cdot \vec{\nabla}' \left({1\over R}\right)

Where:

\vec R = \vec r -\vec{r}'

\vec \nabla ({1\over R}) = -\vec{\nabla}' ({1\over R})
 
Last edited:
i cna use that part for the first two
but i cannot get the second part to work

while i was asleep i thought of something though

does this work? I have clearly forgotten how to apply chain rule...

\frac{\partial J_{x}}{\partial x'}=\frac{\partial J_{x}}{\partial x'}+\frac{\partial J_{x}}{\partial t_{r}}\frac{\partial t_{r}}{\partial x'}\frac{\partial x}{\partial x'}

is this correct??
 
stunner5000pt said:
i cna use that part for the first two
but i cannot get the second part to work

while i was asleep i thought of something though

does this work? I have clearly forgotten how to apply chain rule...

\frac{\partial J_{x}}{\partial x'}=\frac{\partial J_{x}}{\partial x'}+\frac{\partial J_{x}}{\partial t_{r}}\frac{\partial t_{r}}{\partial x'}\frac{\partial x}{\partial x'}

is this correct??

I will show you the steps for \vec \nabla \cdot \vec J

\frac{\partial t_r}{\partial x} = -{1\over c}\frac{\partial R}{\partial x}

So,
\vec \nabla \cdot \vec J = \frac{\partial J_x}{\partial x} + \frac{\partial J_y}{\partial y} + \frac{\partial J_z}{\partial z}

= \frac{\partial J_x}{\partial t_r}\frac{\partial t_r}{\partial x} + \frac{\partial J_y}{\partial t_r}\frac{\partial t_r}{\partial y} + \frac{\partial J_z}{\partial t_r}\frac{\partial t_r}{\partial z}

= -{1\over c} \frac{\partial \vec {J}}{\partial t_r}\cdot (\vec \nabla R)

Can you prove the same for Div J'?
 
Last edited:
Reshma said:
I will show you the steps for \vec \nabla \cdot \vec J

\frac{\partial t_r}{\partial x} = -{1\over c}\frac{\partial R}{\partial x}

So,
\vec \nabla \cdot \vec J = \frac{\partial J_x}{\partial x} + \frac{\partial J_y}{\partial y} + \frac{\partial J_z}{\partial z}

= \frac{\partial J_x}{\partial t_r}\frac{\partial t_r}{\partial x} + \frac{\partial J_y}{\partial t_r}\frac{\partial t_r}{\partial y} + \frac{\partial J_z}{\partial t_r}\frac{\partial t_r}{\partial z}

= -{1\over c} \frac{\partial \vec {J}}{\partial t_r}\cdot (\vec \nabla R)

Can you prove the same for Div J'?

I got it now thanks a lot
 
Hi, I had an exam and I completely messed up a problem. Especially one part which was necessary for the rest of the problem. Basically, I have a wormhole metric: $$(ds)^2 = -(dt)^2 + (dr)^2 + (r^2 + b^2)( (d\theta)^2 + sin^2 \theta (d\phi)^2 )$$ Where ##b=1## with an orbit only in the equatorial plane. We also know from the question that the orbit must satisfy this relationship: $$\varepsilon = \frac{1}{2} (\frac{dr}{d\tau})^2 + V_{eff}(r)$$ Ultimately, I was tasked to find the initial...
The value of H equals ## 10^{3}## in natural units, According to : https://en.wikipedia.org/wiki/Natural_units, ## t \sim 10^{-21} sec = 10^{21} Hz ##, and since ## \text{GeV} \sim 10^{24} \text{Hz } ##, ## GeV \sim 10^{24} \times 10^{-21} = 10^3 ## in natural units. So is this conversion correct? Also in the above formula, can I convert H to that natural units , since it’s a constant, while keeping k in Hz ?
Back
Top