If I take a $(d+1)$ dimensional Einstein Hilbert Lagrangian $L_{d+1}=\sqrt{-\hat{g}} \hat{R}$ and perform a standard Kaluza Klein dimensional reduction by periodically identifying one direction, let's say $z$, by $z \sim z + 2 \pi r$, I arrive at a $d$ dimensional Lagrangian $L_d=\sqrt{-g}(R - \frac{1}{2} (\partial \phi)^2 - \frac{1}{4} e^{-2(d-1) \alpha \phi} F^2)$.

We can see that the Kaluza Klein vector in the $(d+1)$ dimensional metric manifests itself as a $d$ dimensional gauge field in the lower dimensional system. This gauge field has some associated electric charge and I would like to know how, and why, this gets quantized as a result of the identification $z \sim z + 2 \pi r$.

Thanks very much.


2 Answers 2


The Kaluza-Klein equations of motion (the geodesic equations) for a particle moving in the 5D spacetime contain the equations of motion of a particle in 4D spacetime under influence of electromagnetism if and only if one identifies $p^5 = mU^5 = \frac{1}{\sqrt{G}}cq$, i.e. relates the momentum in the fifth dimension $p^5$ to electric charge $q$. (And yes, the momentum in the fifth dimension is a constant of motion, so this is allowed.)

Now, for the Kaluza-Klein theory on a 5D cylinder $\mathbb{R}^{3,1}\times S^1$, the fifth coordinate is that of a circle, and since position and momentum are Fourier transforms of each other in the quantum theory, this means the allowed momenta $p^5$ are discrete, yielding discretization of the electric charge.

  • $\begingroup$ Thanks for the reply. Can you explain why the momentum is to be identified with the specific quantity $\frac{1}{\sqrt{G}} cq$? Or possibly provide a reference? Intuitively, shouldn't the charge have some dependence on the size of the circle? $\endgroup$
    – user11128
    Commented Jan 20, 2016 at 21:14
  • $\begingroup$ @user11128: Yes, the charge has someting to do with the size of the circle. Use the deBroglie relation $p = \hbar / \lambda$ to get the allowed momenta in terms of the circle length (the allowed wavefunctions are standing waves on the circle with circumference $L$). This gives a naive bound on the size of the fifth dimension $\sim 10^{-30}\mathrm{m}$, iirc. $\endgroup$
    – ACuriousMind
    Commented Jan 21, 2016 at 2:27

You can read lecture notes by Malcolm Perry on the Applications of Differential Geometry to Physics. A detailed answer to your question is given in pages 37 & 38. I am providing the link for the (unofficial) lecture notes


  • $\begingroup$ Thanks for the notes. Usually in KK reductions we truncate to the massless sector ($S^1$ needs to be small enough so as not to be visible forcing the masses to be very high - too high to detect). If we truncate the massive modes then 5.47 is telling me $k^2=-\frac{m^2}{h^2}$. What is this $m$? and how can $k^2<0$ since above 5.46 he just required $k^2>0$? Thanks. $\endgroup$
    – user11128
    Commented Jan 21, 2016 at 10:01

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.