In QCD, the gauge-invariant Lagrangian under the transformation

$$ \psi \to \psi' = e^{ig T^a \theta^a(x)} \psi$$

is written as:

$$\mathcal{L} = \bar{\psi}(i\gamma^\mu D_\mu - m)\psi - \frac{1}{4}G^a_{\mu\nu}G_a^{\mu\nu}$$

where the covariant derivative is:

$$D_\mu = \partial_\mu - ig T^a G^a_\mu$$

and the field strength tensor is defined as:

$$ G^a_{\mu\nu} = \partial_\mu G^a_\nu - \partial_\nu G^a_\mu + g f_{abc} G^b_\mu G^c_\nu $$

If I impose the gauge-invariance, I find that the gauge field transforms as:

$$ G^a_\mu \to G'^a_\mu = G^a_\mu + \partial_\mu \theta^a $$

Am I correct? I think I am, but if I look at how the field strength transforms, I expect it to remain invariant, but instead I find an extra term:

$$ G^a_{\mu\nu} \to G^a_{\mu\nu} + g_s f_{abc}(\partial_\mu \theta^b \partial_\nu \theta^c + \partial_\mu \theta^b G^c_\nu + \partial_\nu \theta^c G^b_\mu) $$

Does this term vanish? If does then Why? Or am I totally wrong on the transformation of the gauge field?

  • $\begingroup$ For nonabelian gauge fields the gauge field should transform as $A^a_\mu \rightarrow A^a_\mu + (D_\mu \theta)^a$. $\endgroup$
    – DJBunk
    Commented Jan 12, 2013 at 0:39

1 Answer 1


The Gauge field transforms as $$ G_{\mu}^{i}\frac{\lambda^{i}}{2} \rightarrow G_{\mu}' = uG_{\mu}u^{-1}+\frac{i}{g}u\partial_{\mu}u^{-1} $$ with $u \in SU(3)$ such that $$ u^{-1}(x)=\exp (-i\alpha^i \lambda^i /2) $$ this can be expanded in a series for infinitesimal transformations.


expanding $u^{-1}$ $$ u^{-1} \approx 1 -i\alpha^i \lambda^i/2 + \mathcal{O}(\alpha^2) $$ you can preform this expansion for both $u$ and it's inverse to first order and make sure to keep everything to first order in $\alpha$, it will include a commutator.

  • $\begingroup$ Can you elaborate the expansion? Is the usual expression for the derivative of an exponential not valid for matrix exponentials? $\endgroup$
    – gbaso
    Commented Jan 11, 2013 at 22:32
  • $\begingroup$ @Ganondolf see the update. $\endgroup$
    – kηives
    Commented Jan 11, 2013 at 23:18
  • $\begingroup$ Uhm, I think I sorted out... Most probably I missed a commutator in the interaction term, which accounts for an extra term in the transformation of G. More important, as you suggested I should keep everything to first order, that should be the way to transform correctly the field strength tensor. I still don't know how exactly, but now I should be on the right way... Dropping the problem now for lack of time, but at least I saw the light at the end of the tunnel. Thank you. $\endgroup$
    – gbaso
    Commented Jan 12, 2013 at 0:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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