I'm trying to prove that the time average of the energy density $\omega$ is equivalent to

$$\langle \omega \rangle = \frac{1}{2}\epsilon \epsilon_0 \|\vec{E}_0\|^2$$

for a plane EMW. My approach is to use that

$$\langle f \cdot g \rangle = \frac{1}{2} Re(f \cdot g^*)$$

This comes from Griffiths' Introduction to Electrodynamics:

enter image description here

and also $$\omega = \frac{1}{2} ( \epsilon \epsilon_0 E^2 + \mu\mu_0 H^2)$$

  1. The time average distributes linearly so using $\langle f \cdot g \rangle $ with $f=g=E$ yields $$\langle E^2\rangle = \frac{1}{2}\left(\frac{E_0^2}{1}\right)$$

  2. Similarly for $H$ I get $$\langle H^2 \rangle = \frac{1}{2}\left(\frac{E_0^2}{c^2}\right)$$ using the fact that $\vec{H} =\frac{1}{c} (\hat{k} \times \vec{E_0}) e^{i(kx-wt)}$.

  3. When I sum this two quantities I get $$\omega = \frac{1}{2} E_0^2 \left(\frac{\epsilon\epsilon_0}{2} + \frac{\mu\mu_0}{2c^2}\right)$$ but this puzzles me. I know it shouldn't be this way; otherwise I would get $\mu_0^2 = \mu/\epsilon$ and that's not true from what I know.


1 Answer 1


$$\vec{H} = \frac{1}{c}(\hat{k}\times\vec{E}_0)e^{i(kx-\omega t)}$$ is incorrect. Perhaps you have $\vec{H}$ and $\vec{B}$ confused. The correct formulae are $$\vec{B} = \frac{1}{c}(\hat k\times\vec{E})$$ and $$\vec{H} = \frac{1}{\eta}(\hat k\times\vec{E})$$ where $\eta=\sqrt{\mu\mu_0/\epsilon\epsilon_0}$ is the wave impedance of the medium. With either one, the magnetic energy density $$w_m = \frac{1}{2}\mu\mu_0 H^2 = \frac{B^2}{2\mu\mu_0}$$ will give you the correct result.


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.