2
$\begingroup$

I tried to compare the phase and group velocity of mater waves, say a free electron. I did this once using the de Broglie equation and then using relativity and got different results: $$E = \hbar \omega = \frac{p^2}{2m} = \frac{\hbar^2}{2m}k^2,\ \textrm{where }\ \ p= \hbar k$$ $$\omega = \frac{\hbar}{2m} k^2$$ $$v_{ph} = \frac{\omega}{k} = \frac{\hbar}{2m}k \ \ \ \ \ \ \ \ \ \ \ v_{gr}= \frac{\partial \omega}{\partial k}=\frac{\hbar}{m}k $$ Now using relativity: $$E = \sqrt{p^2 c^2 + m_0^2c^4}$$ $$v_\textrm{ph} = \frac{\hbar \omega}{\hbar k} = \frac{E}{p} = \frac{\gamma m_0 c^2}{\gamma m_0 v} = \frac{c^2}{v} $$ $$v_\textrm{gr} = \frac{\partial E}{\partial p} = \frac{\partial \sqrt{p^2 c^2 + m_0^2c^4}}{\partial p} = \frac{c^2 p}{E}= v$$ So in the first case the phase velocity is half the group velocity and in the second case it is always bigger. If the first case works only for $v\ll c$ than it still should yield a much higher phase velocity. Which of the two is correct and why?

$\endgroup$
  • $\begingroup$ Hi Rosen, please don't misunderstand this comment (as I am not speaking about this particular post and I am way, way too old to worry about my rep points on this site : ), but many new users either don't know or plain forget that (when they acquire enough points, I don't know the actual number) they can also accept answers, as well as up vote them. I write this for your future questions, but wait a few days before acceptance of any of them, in case you get a better one. thanks. $\endgroup$ – user108787 Nov 6 '16 at 14:42
2
$\begingroup$

The dispersion relation for a free relativistic electron wave is $$ω(k) = c\sqrt {k^2 +k_e^2}$$ where $k_e =\frac {m_ec}{\hbar}$

Writing $E = \hbar ω$ and $p = \hbar k$.

A wave with frequency $f$ and wavelength $λ$ has a phase speed $v_p$, which can be defined as $v_p = fλ$.

From $λ = h/ p$ and $f = 2πω$, we can conclude that the phase speed $v_p$ of a free electron wave is $$v_p = ω(k) /k=c\left(1+\frac {k_e^2} {k^2}\right)^{1/2} \ge c$$

The group speed $v_g$ of a wave is defined as $v_g = \frac {dω(k)} {dk}$

This implies that the group $v_g$ of a free electron wave is $$v_g=c\left(1+\frac {k_e^2} {k^2}\right)^{-1/2} \le c $$

Then we can say $$ v_pv_g = c^2$$.

The group speed $v_g$ of a free electron wave is equal to the electron particle speed $v$ defined by $$p =\gamma m_ev$$ $$E =\gamma m_ec^2$$

That is, $$v =\frac {pc^2}{E}=c\left(1+\frac {k_e^2} {k^2}\right)^{-1/2} =v_g$$

$\endgroup$
  • $\begingroup$ Thanks for your answer. Only question I have now is, why is the dispersion relation not just equal to $\omega_{(k)} = \frac{\hbar}{2m} k^2$ or rather where did the relation you used come from? $\endgroup$ – Rosen Sofroniev Nov 5 '16 at 9:45
  • $\begingroup$ That , fingers crossed, should be the relativistic version of E. It came from my olde, oldes notes, the tattiest set of scraps of paper I am embarrassed to say. If you put hbar and c back in, as I am used to setting them to 1, you should be able to reconcile them. Sorry about that, I will google to see if I can get them in your form, but it's a good exercise (for both of us) to try and reconcile them. $\endgroup$ – user108787 Nov 5 '16 at 10:40
  • $\begingroup$ Just to finish off, please ignore my answer if it's too much hassle, there a million versions of this and I don't want you wasting your time on a mistake of mine. I think Schaum's book of exercises covers it. It did get 1 upvote, it's that not you then someone thinks it's OK. $\endgroup$ – user108787 Nov 5 '16 at 11:09
  • $\begingroup$ I tried to upvote your answer because it was helpful but unfortunately I couldn't because I have less than 15 reputation. $\endgroup$ – Rosen Sofroniev Nov 5 '16 at 13:29
  • $\begingroup$ Your anwser is probably the correct one because it agrees with the relativistic way of solving the problem and this one is in wikipedias article for phase velocity I think. I don't know though how to derive the dispersion relation and why it is the one we should use as opposed to the parabolic one ($\omega_{(k)} = \frac{\hbar}{2m} k^2$), that I've seen to be correct as well. I will look for the book you mentioned and if I find a solution I'll update the post. $\endgroup$ – Rosen Sofroniev Nov 5 '16 at 13:35

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.