Is it non-trivial that the coupling constant $g$ in gluon self-interaction terms is the same as the coupling constant $g$ in gluon-fermion interaction term in Yang-Mills theory?

Pure Yang-Mills theory is well- defined itself. But it seems to me that it contains more information due to the same $g$ in the minimal coupling between the Yang-Mills field and fermions.

  • 4
    $\begingroup$ Posit different couplings, and observe gauge invariance failure. $\endgroup$ Commented Aug 17, 2023 at 13:42

1 Answer 1


In annihilation of fermion-antifermion into gluons, there are three diagrams, the usual two that appear in the analogous QED process (the $t$ and $u$ channels).

t and u channel diagrams

However, because of the existence of the three-gluon coupling, there is also a new $s$-channel diagram.

s channel

[Diagrams from Peskin and Schroeder, Introduction to Quantum Field Theory, chapter 16]

If we call the coupling between the fermions and the gauge bosons $g_{f}$ and the pure gauge coupling $g_{b}$, the first two diagrams are $\mathcal{O}(g_{f}^{2})$, while the third is $\mathcal{O}(g_{f}g_{b})$. To have a well-defined theory, these three diagrams have to add up to a gauge-invariant, unitary (meaning with no production of unphysical timelike gauge field modes with negative norm) matrix element $\mathcal{M}$. In order for this to happen, there have to be cancelations between the $\mathcal{O}(g_{f}^{2})$ and $\mathcal{O}(g_{f}g_{b})$ parts of $\mathcal{M}$, which is not generally going to be possible if $g_{f}$ and $g_{b}$ are independent parameters. This requires that the coupling to the fermions be through a representation of the gauge group, with the same coupling $g$ as in the pure gauge sector.

Incidentally, it is also possible to use a similar argument to show that the three- and four-gluon vertices also have to have the same couplings $g$ and $g^{2}$. In this case, there needs to be cancelations among these four diagrams in order to ensure unitarity.

gluon-gluon scattering


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.