If $S_F(x-y)$ is the Green's function for the Dirac operator $(i\gamma^\mu\partial_\mu-m)$, that is, I assume the following matrix equation holds: $$ (i\gamma^\mu\partial_\mu-m)S_F(x-y)=i\delta(x-y) $$

The adjoint dirac equation is: $$ -i\partial_\mu\bar{\psi}\gamma^\mu -m\bar\psi=0 $$

I am looking for the Green's function of the above equation, in terms of $S_F(x-y)$, that is, if $F[S]$ is some function of $S$ (may include complex conjugation, transposing, etc), then I am looking for such an $F[S]$ such that this equation holds: $$ -i\partial_\mu F[S]\gamma^\mu -mF[S]=i\delta(x-y) $$

What I have done so far:

  1. Dagger this equation: $(i\gamma^\mu\partial_\mu-m)S_F(x-y)=i\delta(x-y)$ and get $$(-i\partial_\mu S_F(x-y)^\dagger \gamma^{\mu\dagger} - S_F(x-y)^\dagger m)=-i\delta(x-y)$$
  2. Multiply by $\gamma^0$ on the right to get: $$ (-i\partial_\mu [-\bar{S_F}(x-y)] \gamma^{\mu} - [-\bar{S_F}(x-y)] m)=i\delta(x-y)\gamma^0 $$
  3. So $[-\bar{S_F}(x-y)]$ almost solves this equation, except you get a factor of $\gamma^0$ on the right which I am not sure how to handle.
  • $\begingroup$ $(\gamma^0)^2 = 1$, no? $\endgroup$ – Robin Ekman Apr 6 '14 at 19:24
  • $\begingroup$ Yes, but if you multiply again by $\gamma^0$ on the right you get on the first term $\gamma^\mu \gamma^0=2\delta^{\mu 0}-\gamma^0 \gamma^\mu$ which is not proportional to $\gamma^\mu$. $\endgroup$ – PPR Apr 6 '14 at 19:34
  • $\begingroup$ You're missing a $\gamma^0 $ on the left of $\gamma^\mu $ in the adjoint of the latter. Your final expression for $ S $ should absorb this and one on the left to eliminate the one on the right-hand side. $\endgroup$ – Emilio Pisanty Apr 6 '14 at 22:32
  • 1
    $\begingroup$ people.math.osu.edu/gerlach.1/math/BVtypset/node81.html $\endgroup$ – yess Apr 6 '14 at 22:57

The solution to the problem can be found in the nature of the propagator of the Dirac equation: it is a matrix with two spinor indices, i.e.


In step 2, you have treated the propagator as if it was a spinor with a single index, which is not correct. Avoiding this, but multiplying the equation by $\gamma^0$ from the left leaves you with the result

$$F(S)=\gamma^0 S^\dagger \gamma^0=\bar{S}.$$

The last equality sign is consistent with the reference given in one of the comments and with the definition of the adjoint of a matrix discussed in this question.

  • $\begingroup$ Thanks for your solution. How do you show that $\gamma^0 S^\dagger \gamma^0 = S$ from $ (i\gamma^\mu\partial_\mu-m)S_F(x-y)=i\delta(x-y) $? $\endgroup$ – PPR Apr 6 '14 at 23:27
  • $\begingroup$ I mixed something up, I have edited my answer. $\endgroup$ – Frederic Brünner Apr 7 '14 at 2:55
  • $\begingroup$ So, the conclusion is that $F(S)\neq S$? $\endgroup$ – PPR Apr 7 '14 at 8:11
  • 1
    $\begingroup$ Yes, that is my understanding of the answer. $\endgroup$ – Frederic Brünner Apr 7 '14 at 10:14

After thinking for a while I think it's easiest to treat this problem in momentum space. Then $$S(p) = \frac{p\!\!\!/ + m}{p^2 - m^2 + i\varepsilon}$$ and satisfies $$(p\!\!\!/ - m)S(p) = 1.$$ Note that $S$ carries spinor indices.

Then from dagger on the above equation, $$S(p)^\dagger (p\!\!\!/ - m)^\dagger =1.$$ Now, $p\!\!\!/ = p_\mu \gamma^\mu$ and $p_\mu$is real, so $$p\!\!\!/^\dagger = p_\mu (\gamma^\mu)^\dagger = p_\mu \gamma^0 \gamma^\mu \gamma^0 = \gamma^0 p\!\!\!/ \gamma^0.$$ Therefore, $$S(p)^\dagger(\gamma^0 p\!\!\!/ \gamma^0 - m) = S(p)^\dagger \gamma^0(p\!\!\!/ +m)\gamma^0 = 1$$ so $S(p)^\dagger\gamma^0$ solves the adjoint Dirac equation with unit source.

  • $\begingroup$ The last few steps are not clear. What happened to the minus sign? Shouldn't you have a gamma on the left of your final solution? $\endgroup$ – Emilio Pisanty Apr 6 '14 at 22:28

Can we just multiply from the left with $\gamma^{0}$ then we would get , \begin{equation} \left[-\gamma^{0}S_{F}(x-y)\gamma^{0}\right]\left(-i\partial_{\mu}\gamma^{\mu}-m\right)=i\delta(x-y) \end{equation}


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.