I had heard from a professor saying that "Polyakov and ’tHooft discover the magnetic monopoles in $SU(2)$ gauge theory with scalar fields [Georgi-Glashow model]." And he cited two references:

’t Hooft G., Nucl.Phys. B79 (1974), 276.

A. Polyakov, Pis’ma v ZhETF 20 (1974), 430.

My question is that isnt that $$\pi_1(SU(2))=0$$ so in the $SU(2)$ gauge theory , there is no magnetic object (magnetic monopole as the cut end of ’t Hooft line or Dirac string)? So why/how/what do Polyakov and ’t Hooft discover the magnetic monopoles in $SU(2)$ gauge theory?

p.s. Maybe providing some of your personal view on ’t Hooft and Polyakov works will be nice.

  • 7
    $\begingroup$ Where did the $\pi_1$ come from? Aren't monopoles classified by $\pi_2(G/H)$ for a $G\to H$ symmetry breaking, so in this case $\pi_2(SU(2)/U(1))=\pi_1(U(1))=\mathbb Z$? $\endgroup$ Jul 20, 2021 at 14:41
  • $\begingroup$ $𝜋_1$ is associated with the magnetic line operator, which is the ’tHooft magnetic line operator. And the magnetic monopole as the cut end of ’tHooft line or Dirac string. $\endgroup$ Jul 20, 2021 at 14:48
  • 3
    $\begingroup$ possible duplicate: physics.stackexchange.com/q/512570/84967 $\endgroup$ Jul 20, 2021 at 14:57
  • $\begingroup$ The magnetic monopoles in the Georgi-Glashow model are monopoles with respect to the unbroken $U(1)$ subgroup. There are no $SU(2)$ monopoles. $\endgroup$
    – fewfew4
    Jul 20, 2021 at 15:36
  • $\begingroup$ @fewfew4, yes that clarifies. It counts a partial answer. $\endgroup$ Jul 20, 2021 at 16:30

1 Answer 1


When a gauge group $G$ is Higgsed down to a subgroup $H\subset G$, there are different notions of magnetic charges that we should consider.

  • The first notion is the one you mentioned, particles which carry non-trivial $\pi_1(G)$ charge. These are called $G$-monopoles.

  • Then there are $H$-monopoles, which don't have $\pi_1(G)$ charge, but carry non-trivial $\pi_1(H)$ charge. First, we note that one can always find group homomorphism $f:\pi_1(H)\to\pi_1(G)$. Even if $\pi_1(G)=0$, there can still be $H$-monopoles, as long as its image onto $\pi_1(G)$ is trivial. Physically this means the magnetic charge looks non-trivial with respect to $H$, but is trivial when looking at the full group $G$. The type of elements of $\pi_1(H)$ which satisfy this make up $\text{ker}f$. Making contact with @NiharKarve 's comment, we note that $\text{ker}f=\pi_2(G/H)$. Also note that $\pi_1(G)$ does not need to be trivial for this definition to work.

In the case of the Georgi-Glashow model, $\pi_2(SU(2)/U(1))=\pi_1(U(1))=\mathbb{Z}$, as @NiharKarve mentions. $U(1)$ monopoles are allowed to exist in this theory, and it was shown by 't Hooft that they do.


Your Answer

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

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