As stated above, I want to use $\chi$=$\chi'$-i$\chi''$ to prove that E$\cdot$J is related to $\chi''$ and independent of $\chi'$.

I need to do this using a monochromatic field E(t)=E($\omega$)*exp(i$\omega$t) and then taking the absorption time-average <E$\cdot$J> (Note when using phasor notation temporal averaging is given by =$\frac{1}{2}$Re[A$\cdot$B$^*$]).

Also using the relation $J_p$=$\partial_t$P (the polarization) and assuming no external currents.

I plugged in both E and Jp and tried using things like the definition of susceptibility which is given by P(t)=$\epsilon_0$$\int$$\chi$(t-$\tau$)$\cdot$E(t)d$\tau$ where the integral goes from $-\infty$ to $\infty$ or the expression in the frequency domain P($\omega$)=$\epsilon_0$$\chi$($\omega$)E($\omega$) but it just got more complicated.

I'd really appreciate some help here.

Thanks in advance.

  • $\begingroup$ Technically, you shouldn't be using a time-average on any oscillating function that has not asymptotic mean. You could use a spatial or temporal ensemble average, which is not the same as a time-average in this case. $\endgroup$ Apr 12, 2016 at 19:53

1 Answer 1


You just need the phasor representation of a time derivative.

$P e^{i \omega t} \rightarrow \vec{P}$

$\partial_{t}(P e^{i \omega t}) \rightarrow i \omega \tilde{P}$

Now, since as you said $J=\partial_{t}P$, we have:

$\tilde{J}=i\omega \chi \tilde{E}$

$\langle \vec{E} \cdot \vec{J} \rangle = \frac{1}{2}\text{Re}(i \omega (\chi' + i\chi'')\tilde{E}\tilde{E}^{*}) = -\frac{1}{2}\text{Re}(\omega \chi'' |\tilde{E}|^{2})$.

Physically, this is power dissipation, the same as $\vec{F} \cdot \vec{v}$. The charge movement can be described by either $\vec{P}$ or $\vec{J}$, but either way it's the component of velocity (current) that's in phase with the field that corresponds to power expended. Because $J$ is 90 degrees out of phase with $P$, and $\chi$ describes polarization response:

$\cdot \chi'$ results in polarization in phase with $E$ (current 90 degrees out of phase)

$\cdot \chi''$ results in current in phase with $E$ (polarization 90 degrees out of phase)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.