I am quite a newbie at Electromagnetism and I have been reading Intro to Electrodynamics by Griffiths. I have a doubt and I would highly appreciate any help :)

The author states that for the Electric Displacement vector $D = \epsilon_0E+P$ obeys its own Gauss Law: $\nabla \cdot D= \rho_{\text{free charge}}$ but that there is not any Coulomb's Law for D since $\nabla \times D \stackrel{\nabla \times E = 0}{=} \nabla \times P \neq 0\text{ (generally)}$ (so D it is not like E).

However, when we find ourselves inside a linear isotropic homogeneous dielectric ($P = \text{constant}$) we know that $\nabla \times P = 0 = \nabla \times D$ (the same as E). Later he concludes that

$D = \epsilon_0 E_{\text{vacuum}}$ where $E_{\text{vacuum}} =$ the electric field without the dielectric (generated by $\rho_{\text{free charge}}$) and I can't imagine how he reached there (that's my question).

Thus we DO have a Coulomb's Law for D in this particular case using the free charge to get the field.

I understand that $\nabla \times D = 0 = \epsilon_0\nabla \times E$ and $\nabla \cdot D = \rho_{\text{free charge}}\stackrel{\text{Gauss Law}}{=} \epsilon_0\nabla \cdot E_{\text{vacuum}}$ but as far as I am concerned that is not sufficient to state that $D=\epsilon_0E_{\text{vacuum}}$ (it could be proved if D was parallel to E though). How can I prove this equality?

Thanks in advance, some relevant images:

Picture 1 Picture 2


2 Answers 2


This is all about the structure of the equations. The electrostatic field $\mathbf{E}$ satisfies the two equations

$$\nabla\cdot \mathbf{E}=\dfrac{\rho}{\epsilon_0},\quad \nabla\times \mathbf{E}=0,$$

together with boundary conditions on crossing charged surfaces, for example.

What the author means is the following. We know that $\mathbf{D}$ satisfies the equations

$$\nabla\cdot \mathbf{D}=\rho_{\text{free}},\quad\nabla\times\mathbf{D}=\nabla\times \mathbf{P}.$$

Now when $\nabla\times \mathbf{P}=0$, for example, when $\mathbf{P}$ is constant, we have that

$$\nabla \cdot \mathbf{D}=\rho_{\text{free}},\quad \nabla \times \mathbf{D}=0.$$

As a mathematical problem of solving a differential equation this is the same as the electrostatic problem I said above, with the identification that the charge density should be $\rho = \epsilon_0\rho_{\text{free}}$.

To see this, divide by $\epsilon_0$ the first equation and notice that

$$\nabla\cdot \left(\dfrac{1}{\epsilon_0}\mathbf{D}\right)=\dfrac{\rho_\text{free}}{\epsilon_0},\quad \nabla \times \mathbf{D}=0.$$

Thus since these are the same equations for the electrostatic field in vacuum generated by a charge density $\rho_{\text{free}}$, which we shall call $\mathbf{E}_{\text{vac}}$ you have that


which gives your equation $\mathbf{D}=\epsilon_0 \mathbf{E}_{\text{vac}}$.

  • 2
    $\begingroup$ Thanks I think I've got it. As I see it, the key is that they share the same "boundary conditions" over a surface, thus, the electrostatics uniqueness theorem states they must be the same. $\endgroup$
    – J. Doe
    Commented Nov 3, 2017 at 7:38
  • $\begingroup$ If two functions have the same divergence and curl then it isn't necessary that the functions are the same. $\endgroup$
    – Kashmiri
    Commented Oct 4, 2020 at 16:51

Ah homogeneous linear dielectric is defined by the property

$\textbf{P} = \epsilon_0 \chi \textbf{E}$

so that

$\textbf{D} = \epsilon_0 \textbf{E} + \textbf{P} = \epsilon_0 \textbf{E} + \epsilon_0 \chi \textbf{E} = \epsilon_0 (1 + \chi) \textbf{E} \overset{def}{=} \epsilon_0 \textbf{E}_{vac}$

I think it is just a definition.


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.