0
$\begingroup$

The pseudo-scalar Yukawa theory Lagrangian is

$$\mathcal{L}=\bar{\psi}(i\gamma ^\mu \partial_\mu - m)\psi -g\bar{\psi}i\gamma^5\phi\psi,$$ where $g$ is a coupling constant. How can I show it is invariant under a chiral transformation, $\psi\to e^{i\lambda \gamma_5}\psi$?

$\endgroup$
  • 2
    $\begingroup$ Um...plug the transformation in and see what happens? $\endgroup$ – ACuriousMind Nov 2 '14 at 16:08
2
$\begingroup$

(This is a largely a response prompted by your comment.)

You can get the answer by just remembering the commutation/anti-commutation properties of the $\gamma$ matrices, and the fact that ${\bar \psi} = \psi^{\dagger} \gamma^0$. To see the following, you would have to expand the exponential factor, up to linear order $e^{M} = I + M + \ldots$.

(I'm not going to do your homework, this is just a guide!)

1) The kinematic term $i \bar{\psi}\gamma ^\mu \partial_\mu\psi$ goes into itself, using $ \{\gamma^{\mu}, \gamma^5\} = 0$.

2) The Yukawa coupling term follows suite.

3) There is no such cancellation in the mass term $m \bar \psi \psi$, but the two factors reinforce each other. This term picks up an overall factor of $e^{2i\lambda \gamma_5}\psi$, i.e. two times either factor. Thus, the mass term is not invariant under this transformation, and breaks chiral symmetry.

B.T.W. This transformation is called the axial-vector transformation, since the corresponding conserved (in the m=0 limit) Noether current transforms like an axialvector $\bar \psi \gamma^{\mu} \gamma^5 \psi$.

Resolving into Weyl spinors $\psi_{L,R} = (1\mp \gamma^5)\psi/2$ is an alternative way of seeing this. With this, you will again have to use the $\gamma$ matrices' properties, and you will arrive at the result that only the mass term mixes up the two chiralities, i.e. becomes $m (\bar \psi_L \psi_R + \bar \psi_R \psi_L)$. The kinematic term would transform into $i \bar{\psi_L}\gamma ^\mu \partial_\mu\psi_L + i \bar{\psi_R}\gamma ^\mu \partial_\mu\psi_R$ and hence, it is like the the kinematic terms of two independent Lagrangians added up. No mixing. The two formulations are absolutely equivalent.

$\endgroup$
  • $\begingroup$ Thanks for your very clear guidance. I have a question though, while I am performing this transformation on Yukawa coupling term: I get the following:$$L'=-g\bar{\psi}i\gamma^5\phi\psi = -gi\bar{\psi}e^{i\lambda \gamma^5}\gamma^5 \phi e^{i\lambda\gamma^5}\psi$$ Why aren't the exponentials cancelling so it would be invariant.. $\endgroup$ – Fluctuations Nov 2 '14 at 19:09
  • 1
    $\begingroup$ In one of your exponentials there should be a minus sign, because $\bar{\psi}$ is a conjugate of $\psi$. Also, $\gamma^5$ commutes with its exponential (because it commutes with all terms in its Tailor expansion). So the two exponentials cancel each other. $\endgroup$ – Prof. Legolasov Nov 2 '14 at 20:31
  • 1
    $\begingroup$ @Fluctuations no, it is not true. But for every matrix $x$ (in your case, $x=i\gamma^5$), the following holds: $$ \left[ x, \exp x \right] = 0 $$ $\endgroup$ – Prof. Legolasov Nov 2 '14 at 21:38
  • 1
    $\begingroup$ Hindsight did most of the follow-up job for me, so thanks. Regarding the last point, @Fluctuations, as Hindsight already mentioned, they don't anticommute, they commute! To see this explicitly, expand the exponential in $$ [ x, \exp x ]$$ and exploit the linearity $$ [ x, (y+z) ] = [ x, y ] + [ x, z ]$$. Clearly, $x$ commutes with identity, with $x$, with $x^2$, and so on. :) $\endgroup$ – 299792458 Nov 3 '14 at 5:10
  • 1
    $\begingroup$ And @Hindsight, thanks for the follow-up job. Yes, chiral symmetry holds only in the $m=0$ limit. Finite mass explicitly breaks chiral symmetry. But if these mass terms are small, like $u$ and $d$ quarks (the $SU(2)$ case), chiral symmetry can be considered an approximate symmetry of the strong interactions. :) $\endgroup$ – 299792458 Nov 3 '14 at 5:16

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.