Griffiths 4.24: Solving Electric Field in Insulated Sphere

  • Thread starter ehrenfest
  • Start date
  • Tags
    Griffiths
In summary: Electric field is not constant at r=a. The lack of free charge implies that the divergence of *D* vanishes.The POTENTIAL must be constant in r = a, correct, since all of the metal sphere is an equipotential.In summary, the electric field in the dielectric is not constant, which means that the potential cannot be solved for in terms of its derivatives at r=a.
  • #1
ehrenfest
2,020
1

Homework Statement


This question refers to Griffiths E and M book.

An uncharged conducting sphere of radius a is coated with a thick insulating shell (dielectric constant \epsilon_r) out to radius b. This object is now placed in an otherwise uniform electric field [itex]\bfE_0[/itex]. Find the electric field in the insulator.

Homework Equations


The Attempt at a Solution


So, I wrote down the boundary conditions and I can't seem to find solutions that match them. Since the electric field vanishes inside of a conductor. The solution inside the dielectric must be constant at r=a. Its derivative must vanish at r=a since there is no free charge on that surface. I don't see how we can obtain that unless the potential is constant inside the conductor. But then how can you meet the boundary conditions for the r=b where the potential is not constant!
 
Last edited:
Physics news on Phys.org
  • #2
anyone?
 
  • #3
Why don't you show your attempt to solution? You seldom does that..
 
  • #4
I did actually. I just didn't put it in "The attempt at the solution" section.

EDIT: Now I did.
 
Last edited:
  • #5
I think you're supposed to use the eigenfunction equation for angular momentum (L^2 * psi = h_bar^2 * l(l+1) psi instead of the hamiltionian equation. Then scale the eigenvalue to get the measure in energy. Never done it myself, but thatd be my guess. Remember, the its not charge dependent, the particles could be neutrons.
 
Last edited:
  • #6
A statement of the question may also be useful...
 
  • #7
faen said:
I think you're supposed to use the eigenfunction equation for angular momentum (L^2 * psi = h_bar^2 * l(l+1) psi instead of the hamiltionian equation. Then scale the eigenvalue to get the measure in energy. Never done it myself, but thatd be my guess. Remember, the its not charge dependent, the particles could be neutrons.

I think you are looking at the wrong book.
 
  • #8
ehrenfest said:
I think you are looking at the wrong book.
Why not state the damn problem in your opening post to avoid such confusion?! :grumpy:
 
  • #9
I wrote this as the first line in my post: "This question refers to Griffiths E and M book." I think the confusion is not my fault.

This is not really something you could help me with unless you have the book (and a lot of people do have the book).
 
  • #10
ehrenfest said:
I wrote this as the first line in my post: "This question refers to Griffiths E and M book." I think the confusion is not my fault.

This is not really something you could help me with unless you have the book (and a lot of people do have the book).
That is of course your choice, you are more than welcome to wait for someone who has a copy of that text (with the same edition as you). I was simply offering to help. However, do you not wonder why so many of your threads go unanswered?
 
Last edited:
  • #11
See the opening post.
 
  • #12
See pages 97-100 of Electromagnetism by Grant & Philips :wink:
 
  • #13
Can you just tell me what is the flaw in my attempt at the solution?
 
  • #14
The obvious flaw is that the electric field is not constant at r=a. The lack of free charge implies that the divergence of *D* vanishes.
 
  • #15
The POTENTIAL must be constant in r = a, correct, since all of the metal sphere is an equipotential.
 
  • #16
ehrenfest said:
So, I wrote down the boundary conditions and I can't seem to find solutions that match them. Since the electric field vanishes inside of a conductor. The solution inside the dielectric must be constant at r=a. Its derivative must vanish at r=a since there is no free charge on that surface. I don't see how we can obtain that unless the potential is constant inside the conductor. But then how can you meet the boundary conditions for the r=b where the potential is not constant!

Not right. Why don't you try writing down what the general form of the solution to Laplace equation would be in the three regions?

This way, you'll see how many constants you need to evaluate, and the corresponding boundary conditions.
 
Last edited:
  • #17
I don't understand. The potential is constant in the metal sphere. Therefore it must be constant at the boundary.
 
  • #18
ehrenfest said:
I don't understand. The potential is constant in the metal sphere. Therefore it must be constant at the boundary.

Constant at the boundary? Do you mean continuous?

What I mean is this.

[tex] V(r,\theta) = 0, \quad r<a[/tex]
[tex] V(r,\theta) = ?? \quad a<r<b[/tex]
[tex] V(r,\theta) = ?? \quad r>b[/tex]

If you can find the functional form [tex]V(r,\theta)[/tex] (hint: soln of laplace eqn in spherical coordinates), then you can find the corresponding constants by evaluvating the boundary conditions (ie, continuity at r=a and r=b).
 
Last edited:
  • #19
What I am saying is that [tex]V(a,\theta) =\frac{\partial V}{\partial r} \left|_a= 0[/tex]. That is true, right?

That means that the solution inside the solution inside the dielectric cannot have any \theta dependence right or else either the derivative of the potential or the potential itself would have theta dependence at a.

genneth said:
The obvious flaw is that the electric field is not constant at r=a. The lack of free charge implies that the divergence of *D* vanishes.

There is NO WAY the electric is not constant at r=a! Then it would not be constant inside the conductor which is absurd.

Hmm. OK I was using the following formula (eqn 4.41) in Griffiths :

[tex]\epsilon_{above} \vec{E}_{above} \cdot \vec{n} - \epsilon_{below} \vec{E}_{below} \cdot \vec{n} = \rho_f[/tex]

at r=a. I am wondering whether that applies or not because my book says it applies here. Maybe that only applies at the boundary of two dielectrics and a metal is not considered a dielectric i guess?

If that equation does apply at r=a, then genneth must be wrong and the electric field must be constant (actually 0) just above the boundary at r = a.
 
Last edited:
  • #20
anyone?
 
  • #21
What's tau? And no, it's not true that the field inside the dielectric has no theta dependence. The fact that the potential's derivative in the theta direction vanishes means that the field doesn't *point* in the theta direction.
 
  • #22
genneth said:
What's tau?
Its an r not a tau.
 
  • #23
So here is the 6 million dollar question:

Do we know have a boundary condition for the derivatives of the potentials (i.e. the fields) at r = a?
 
  • #24
Yes. You want E parallel, D perpendicular to be continuous.
 
  • #25
genneth said:
Yes. You want E parallel, D perpendicular to be continuous.

Do we know anything about E perpendicular?
 
  • #26
No -- since there could be induced surface charges there.
 
  • #27
But that implies that

[tex]
\epsilon_{above} \vec{E}_{above} \cdot \hat{\vec{n}} - \epsilon_{below} \vec{E}_{below} \cdot \hat{\vec{n}} = \rho_f
[/tex]

where n is the normal is wrong. That is equation 4.40 in Griffiths! It is supposed to hold for all linear dielectrics!
 
  • #28
Genneth, would your answers change if the metal were instead a dielectric solid sphere (obviously not grounded)?
 
  • #29
ehrenfest said:
What I am saying is that [tex]V(a,\theta) =\frac{\partial V}{\partial r} \left|_a= 0[/tex]. That is true, right?

That means that the solution inside the solution inside the dielectric cannot have any \theta dependence right or else either the derivative of the potential or the potential itself would have theta dependence at a.

There is NO WAY the electric is not constant at r=a! Then it would not be constant inside the conductor which is absurd.

Hmm. OK I was using the following formula (eqn 4.41) in Griffiths :

[tex]\epsilon_{above} \vec{E}_{above} \cdot \vec{n} - \epsilon_{below} \vec{E}_{below} \cdot \vec{n} = \rho_f[/tex]

at r=a. I am wondering whether that applies or not because my book says it applies here. Maybe that only applies at the boundary of two dielectrics and a metal is not considered a dielectric i guess?

If that equation does apply at r=a, then genneth must be wrong and the electric field must be constant (actually 0) just above the boundary at r = a.

So here is the 6 million dollar question:

Do we know have a boundary condition for the derivatives of the potentials (i.e. the fields) at r = a?

To repeat my last post, just work out the math by finding the form of the solution. Once you do that, you will see how the boundary conditions at a and b can be applied, and hence what the answers are to your questions.
 
Last edited:
  • #30
I know what the form of the solutions is but I cannot apply boundary conditions if I do not know what they are. Please answer my questions.
 
  • #31
ehrenfest said:
I know what the form of the solutions is but I cannot apply boundary conditions if I do not know what they are. Please answer my questions.

Ok, if you type the form of the solution in your next post, it might help.

Because, if you do that,
(i) you will see how many constants you need to evaluate
(ii) you will be able to figure out what bc's to use (ex, what happens to V at r=b? What happens to its derivative)
(iii) once you find the solution, you will see what happens at r=a (ex, what is dV/dr at r=a)
 
Last edited:

Similar threads

  • Advanced Physics Homework Help
Replies
4
Views
2K
  • Advanced Physics Homework Help
Replies
1
Views
2K
  • Advanced Physics Homework Help
Replies
4
Views
3K
  • Advanced Physics Homework Help
Replies
1
Views
5K
  • Advanced Physics Homework Help
Replies
1
Views
2K
  • Advanced Physics Homework Help
2
Replies
49
Views
4K
  • Advanced Physics Homework Help
2
Replies
37
Views
10K
  • Introductory Physics Homework Help
Replies
17
Views
327
Replies
1
Views
1K
  • Advanced Physics Homework Help
Replies
1
Views
1K
Back
Top