I have derived the Yang-Mills equation and its generalization coupled to a current of a scalar field $\phi$ by extremalizing the action describing a $\mathrm{SU}(2)$ scalar field gauge theory:

$$\partial_\mu F^{\mu\nu} +ig\left[ A_{\mu},F^{\mu\nu}\right] = j^\nu$$

where, $\phi$ is a two components scalar field,

$$j^\nu = -ig\left[ (D_\nu \phi)^\star T^a\phi -\phi^\star T^a(D_\nu \phi)\right]T^a$$

where $D_\nu = \partial_\nu + ig A_\nu$. But when I take a gauge transformation:

$$\phi' = e^{-i\omega_a T_a}\phi = U\phi, \quad A'_\mu = UA_\mu U^{-1} -\frac{i}{g}U\partial_\mu U^{-1}$$

I find I can not take same formalism from $j'^{\nu}$ as the original $j^{\nu}$. I think there must be something wrong with my calculation, because the current should be gauge invariant. My question is therefore whether the Yang-Mills equation and its generalization is gauge invariant and how one would show this invariance.

More about my calculation, please comment @ACuriousMind, Thanks you help and analysis.Now I will write down more of my calculation: When I calculate the $j^{\nu}$', I find that the term : $${\phi}^{*\alpha}{\partial _{\upsilon}\phi}^{\alpha }$$ always contains the quality: $${ U }^{ -1 }{ T }^{ a }U$$ as $${\phi}^{*\alpha}{\partial _{\upsilon}\phi}^{\alpha }{ U }^{ -1 }{ T }^{ a }U$$ the ${ U }^{ -1 }{ T }^{ a }U$can not be cancelled by the commutator calculation. Base on your above answer, can I think this phenomena is correct? I'm never work for QFT, and I learning the classical gauge field theory by my-self, please point out my mistake,Please, theanks.


1 Answer 1


No, the gauge current need not be gauge invariant, since it carries a group index in non-abelian theories. You should recall that both sides of the Yang-Mills equation (and therefore the current itself) are Lie-algebra valued and therefore transform in the adjoint representation. Not even the field strength $F^a_{\mu\nu}$ is gauge invariant, but transforms in the adjoint representation of the gauge group, which is why your action should (hopefully) contain only its trace as $\mathrm{Tr}(F_{\mu\nu}F^{\mu\nu})$.

It is to be noted that, since the current is not invariant, it is not observable.

  • $\begingroup$ en, thanks very much! Is your means that this motion equation is not gauge invariant? Please continuously focus this post, I will write down another thing $\endgroup$ Oct 15, 2014 at 12:01
  • $\begingroup$ @alxandernashzhang: Not really. The equation of motion itself is gauge invariant, since both sides transform in the same representation, so the transformation can be cancelled. Your current should transform as $j_\mu \mapsto Uj_\mu U^{-1}$, if it doesn't, it's the wrong current. Also, please don't post comments as answer - either edit the additional information into the question, or use these comments. $\endgroup$
    – ACuriousMind
    Oct 15, 2014 at 12:26
  • $\begingroup$ ,you mean,if we take a gauge transformation,the new current $j^{\nu}\prime$ must equal to $Uj^{\nu}U^{-1}$? $\endgroup$ Oct 15, 2014 at 12:35
  • $\begingroup$ @alxandernashzhang: Yes. $\endgroup$
    – ACuriousMind
    Oct 15, 2014 at 12:36
  • $\begingroup$ ,you mean,if we take a gauge transformation,the new current $j^{\nu}\prime$ must equal to $Uj^{\nu}U^{-1}$? And I always have another question about the trace in the energy density gauge field: Can I seem the trace as a standard Matrix trace act on the matrix of $F_{\mu\nu}F^{\mu\nu}$, and how to get the relation: $\frac { 1 }{ 2 } { F }_{ \mu \upsilon }^{ a }{ F }_{ a }^{ \mu \upsilon }=tr({ F }_{\mu \upsilon }{ F }^{ \mu \upsilon })$? I teach myself to learn gauge field theory, my supervisor always not investigate the area, so I have some basic problem like this.thanks $\endgroup$ Oct 15, 2014 at 12:45

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.