1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Homework Help: Mandl & Shaw Problem 4.3

  1. Jun 21, 2011 #1
    1. The problem statement, all variables and given/known data

    The charge current density operator for the Dirac equation is defined as : [itex]s^\mu = - ec \bar{\psi}\gamma^\mu\psi[/itex].

    2. Relevant equations

    I need to show that the current density operator commutes when measured at two spacelike separated points :

    [itex][s^\mu(x),s^\nu(y)] = 0 [/itex] for [itex](x-y)^2 < 0[/itex].

    3. The attempt at a solution

    First, I inspired myself with the following post :


    However, I admit that I don't understand how this thread has been tagged "solved" since there is nothing there to help anyone... (or I don't see how the second post is a hint good enough to end the thread).

    So, I get that by microcausality, the field must anti-commute :

    [itex]\{ \psi(x), \bar{\psi}(y) \} = 0[/itex].

    The commutator can be obtained by a simple substitution :

    [itex][s^\mu(x),s^\nu(y)] = e^2 c^2( \bar{\psi}(x) \gamma^\mu \psi (x) \bar{\psi}(y) \gamma^\nu \psi(y) - \bar{\psi}(y) \gamma^\nu \psi(y) \bar{\psi}(x) \gamma^\mu \psi (x) )[/itex]

    My problem here is how to deal with the [itex]\gamma^\mu[/itex]. I know that [itex]s^\mu[/itex] is basically 4 numbers ([itex]\mu[/itex] = 0, 1, 2, 3 or 4). So I deduce that [itex]- ec \bar{\psi}\gamma^\mu\psi[/itex] must also be 4 numbers.

    I tried expressing this quantity using the matrix indices. I got the following expression:

    [itex]s^\mu = - ec \bar{\psi}_\alpha (\gamma^\mu)_{\alpha \beta} \psi_{\beta}[/itex] (The Einstein summation convention on repeated indices is used here).

    By using this formalism, I wrote the commutator and I got the following expression :

    [itex][s^\mu(x),s^\nu(y)] = e^2 c^2( \bar{\psi}_\alpha (x) \psi_\beta (x) \bar{\psi}_\delta (y) \psi_\epsilon (y) (\gamma^\mu)_{\alpha \beta} (\gamma^\nu)_{\delta \epsilon} - \bar{\psi}_\alpha (y) \psi_\beta (y) \bar{\psi}_\delta (x) \psi_\epsilon (x) (\gamma^\nu)_{\alpha \beta} (\gamma^\mu)_{\delta \epsilon} )[/itex]

    This is the point where I am stuck. I notice that if I just exchange the indices [itex]\alpha \leftrightarrow \delta[/itex] and [itex]\beta \leftrightarrow \epsilon[/itex] in the second term, the commutator goes to zero without any problem.

    My questions are :

    #1 Can I do this ? It seems to easy to be true...

    #2 I put the [itex]\gamma[/itex] at the end of my expression... Can I move them in this way? Since, they are numbers, I took the liberty of moveing them around. However, can I do this with the fields ?

    #3 I did not use the microcausality condition to get this result. Is it required or not ? And if so, ca I have a hint as to where I should include it ?

    Thanks for your help (I spent a whole day on this thing and I am really going nowhere)!
    Last edited by a moderator: Apr 26, 2017
  2. jcsd
  3. Jul 4, 2011 #2
    Nobody can help me ? :frown:
  4. Nov 17, 2011 #3


    User Avatar
    Gold Member

    You can't do that. If you change the order of the summations you are changing the order of the matrices.
    I think should demonstrate that the commutator is a number and not a matrix using the definition of the 4-density current and gamma properties.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook