Is it true that a field of conformal dimension $h$ (integer or half integer) has $1-2h$ zero-modes on the sphere, if $1-2h \geq0$.

This seems to be right for different ghost fields :

  • $c$ has weight $-1$ and 3 zero-modes
  • $\gamma$ has weight $-1/2$ and 2 zero-modes
  • $\xi$ has weight $0$ and 1 zero-modes
  • $\eta$ has weight $1$ and 0 zero-modes

I'm not sure how to compute the number of zero-modes on the sphere... Is it the argument given at p.167 in volume 1 of Polchinski's book ?


1 Answer 1


Yes, OP is right, cf. Ref. 1 and 2. Consider a generalized $bc$ ghost system. (This can be generalized to any system with the pertinent mathematical properties used below.) Here $b$ and $c$ are primary fields of conformal weights $h_b$ and $h_c$, respectively, where $h_b+h_c=1$.

Counting globally defined, holomorphic zero-modes of a (generalized) $bc$ system is related to the Riemann-Roch theorem. Here applied to the Riemann sphere $S^2$, i.e. Euler characteristic $\chi=2$, or equivalently, genus zero.

We basically have to check that a zero-mode on the complex plane $\mathbb{C}$ can be holomorphically extended to the point $z=\infty$ on the Riemann sphere $S^2=\mathbb{C}\cup \{\infty\}$. To describe the neighborhood of $z=\infty$, we perform a coordinate transformation $\tilde{z}=-1/z$. In the new coordinates the ghost field reads

$$\tilde{c}~=~\left(\frac{\partial\tilde{z}}{\partial z}\right)^{-h_c} c~=~z^{2h_c} c.$$

In order that a zero-mode doesn't blow up at $z=\infty$, it must be a polynomial in $z$ of order at most $-2h_c$. Recall that an $n$'th order polynomial in $z$ has $n+1$ coefficients. In other words, the number of globally defined, holomorphic zero-modes for $c$ is

$$ \max(1-2h_c,0).$$


  1. J. Polchinski, String Theory Vol. 1, 1998; p.167.

  2. M.B. Green, J.H. Schwarz and E. Witten, Superstring theory, Vol. 1, 1986; p. 161-163.

  • $\begingroup$ Thank you for the answer ! However you seem to keep focusing on ghosts, is there any reason for that ? The argument should be valid for any primary field, shouldn't it ? Also, do what you call a "generalized $bc$ ghost system" include the case of $\beta \gamma$ ghosts ? $\endgroup$
    – Antoine
    Apr 29, 2014 at 7:46
  • $\begingroup$ I updated the answer. $\endgroup$
    – Qmechanic
    Apr 30, 2014 at 13:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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