Electrodynamics: Dirac Delta function and Gauss' divergence theorem

AI Thread Summary
The discussion focuses on computing the Laplacian of the function 1/|x| in spherical coordinates, which yields zero except at the origin. To integrate the Laplacian over a spherical surface enclosing the origin, Gauss's divergence theorem is applied, requiring the substitution of ∇(1/|x|) for F in the integral. Participants confirm that this substitution is necessary to derive the Laplacian correctly. The conversation also explores deriving the general result for the gradient of the potential function, leading to the conclusion that ∇(1/|x - x'|) equals -4πδ(x - x'). A suggestion is made to shift the coordinate system to simplify the evaluation of the Laplacian at the charge's location.
lys04
Messages
144
Reaction score
5
Homework Statement
Show that $$ \nabla ^ {2} (\frac{1}{|\vec{x}|}) = -4\pi \delta(\vec{x}) $$
Relevant Equations
Gauss's divergence theorem, Laplacian in spherical coordinates
Computing the Laplacian of ##(\frac{1}{|\vec{x}|})## in spherical coordinates I get 0 everywhere except at 0.
Now I want integrate the Laplacian over a spherical surface that encloses the origin. I do this using Gauss's divergence theorem, i.e $$ \iint_S \vec{F}.\vec{dA} = \iiint_V \nabla . F dV $$

Do I need to use ∇F here in substitution for F so that when I take the divergence of the gradient of F I get the Laplacian of F which is what I need?
need?
 
Physics news on Phys.org
lys04 said:
Now I want integrate the Laplacian over a spherical surface that encloses the origin. I do this using Gauss's divergence theorem, i.e $$ \iint_S \vec{F}.\vec{dA} = \iiint_V \nabla . F dV $$

Do I need to use ∇F here in substitution for F so that when I take the divergence of the gradient of F I get the Laplacian of F which is what I need?
need?
Yes. Well, you need to sub ##\nabla \dfrac 1 {|\vec x|}## for ##\vec F## in your integral equation.
 
  • Like
Likes Albertus Magnus
TSny said:
Yes. Well, you need to sub ##\nabla \dfrac 1 {|\vec x|}## for ##\vec F## in your integral equation.
Yes. That's what I did.
And to derive the more general result, i.e $$ \nabla (\frac{1}{|\vec{x}-\vec{x'}|}) = -4 \pi \delta (\vec{x}-\vec{x'})$$, do I just let ##\vec{x}-\vec{x'}## equal to some other vector say ##\vec{r}## and then using my previous result it gives me $$ \nabla (\frac{1}{|\vec{r}|}) = -4 \pi \delta (\vec{r})$$ so $$ \nabla (\frac{1}{|\vec{x}-\vec{x'}|}) = -4 \pi \delta (\vec{x}-\vec{x'})$$ is correct?
 
lys04 said:
And to derive the more general result, i.e $$ \nabla (\frac{1}{|\vec{x}-\vec{x'}|}) = -4 \pi \delta (\vec{x}-\vec{x'})$$
do I just let ##\vec{x}-\vec{x'}## equal to some other vector say ##\vec{r}## and then using my previous result it gives me $$ \nabla (\frac{1}{|\vec{r}|}) = -4 \pi \delta (\vec{r})$$ so $$ \nabla (\frac{1}{|\vec{x}-\vec{x'}|}) = -4 \pi \delta (\vec{x}-\vec{x'})$$ is correct?
Typos: ##\nabla## should be ##\nabla^2## in the expressions above. Otherwise, yes.

One way to think about it is to imagine shifting the origin of the coordinate system to the location of the charge, ##\vec x'##. Then, ##\vec x - \vec x'## in the original system becomes ##\vec x## in the new system. After evaluating ##\nabla^2 \dfrac 1 { |\vec x|}## in the new system you can shift back to the original coordinate system.
 
  • Like
Likes Albertus Magnus
You can use "\cdot" to write your dot products correctly.
 
  • Like
Likes Gavran and lys04
TSny said:
Typos: ∇ should be ∇2 in the expressions above. Otherwise, yes.
Oh yes, sorry about that.
TSny said:
One way to think about it is to imagine shifting the origin of the coordinate system to the location of the charge, x→′. Then, x→−x→′ in the original system becomes x→ in the new system. After evaluating ∇21|x→| in the new system you can shift back to the original coordinate system.
Ah, thats a good way of thinking about it.
 
Back
Top