In string theory, the Polyakov string action is given by \begin{align} S = -\frac{T}{2}\int d^2\sigma\:\sqrt{-\gamma}\gamma^{ab}\eta_{\mu\nu}\frac{\partial X^{\mu}}{\partial \sigma^a}\frac{\partial X^{\nu}}{\partial \sigma^b}, \end{align} where $\gamma^{ab}$ is an independent worldsheet metric that has a residual conformal symmetry, and $\gamma = \text{det}\gamma_{ab}$. By performing an infinitesimal worldsheet reparameterisation of the form $\tilde{\sigma}^a = \sigma^a - \epsilon^a(\sigma)$, I have been told that when the variation of the action vanishes, I should find \begin{align} \partial^a(T_{ab}\epsilon^b) = 0. \end{align} $T_{ab}$ is the stress-energy tensor given by \begin{align} T_{ab} = \eta_{\mu\nu}\frac{\partial X^{\mu}}{\partial \sigma^a}\frac{\partial X^{\nu}}{\partial \sigma^b} - \frac{1}{2}\gamma_{ab}\gamma^{cd}\eta_{\mu\nu}\frac{\partial X^{\mu}}{\partial \sigma^c}\frac{\partial X^{\nu}}{\partial \sigma^d}. \end{align}

I've not been able to prove this so far, and at this point I'm stuck. I was wondering if anyone would be able to show me how it is done?

My attempt:

First calculate \begin{align} \tilde{\gamma}^{ab}(\tilde{\sigma}) &= \frac{\partial\tilde{\sigma}^a}{\partial \sigma^c} \frac{\partial\tilde{\sigma}^b}{\partial \sigma^d}\gamma^{cd}(\sigma) \nonumber\\ &= (\delta^a_c - \partial_c\epsilon^a(\sigma))(\delta^b_d - \partial_d\epsilon^b(\sigma))\gamma^{cd}(\sigma) \nonumber\\ &= \gamma^{ab}(\sigma) - \partial_c\epsilon^a(\sigma)\gamma^{cb}(\sigma) - \partial_d\epsilon^b(\sigma)\gamma^{ad}(\sigma)\nonumber\\ &= - \partial^b\epsilon^a(\sigma) - \partial^a\epsilon^b(\sigma), \end{align} and so \begin{align} \delta\gamma^{ab} = - \partial^b\epsilon^a(\sigma) - \partial^a\epsilon^b(\sigma). \end{align} Use the result \begin{align} \delta \sqrt{-\gamma} = -\frac{1}{2}\sqrt{-\gamma} \gamma_{ab}(\sigma)\delta\gamma^{ab}. \end{align} Vary the action with respect to the reparameterisation: \begin{align} \delta S =& \delta\left( -\frac{T}{2}\int d\sigma^2\:\sqrt{-\gamma}\gamma^{ab}(\sigma)\eta_{\mu\nu}\frac{\partial X^{\mu}}{\partial \sigma^a}\frac{\partial X^{\nu}}{\partial \sigma^b}\right) \nonumber\\ =& -\frac{T}{2}\int d\sigma^2\:\left[\delta\sqrt{-\gamma}\gamma^{ab}(\sigma) + \sqrt{-\gamma}\delta\gamma^{ab}\right]\eta_{\mu\nu}\frac{\partial X^{\mu}}{\partial \sigma^a}\frac{\partial X^{\nu}}{\partial \sigma^b} \nonumber\\ =& -\frac{T}{2}\int d\sigma^2\:\left[-\frac{1}{2}\sqrt{-\gamma} \gamma_{cd}(\sigma)\delta\gamma^{cd}\gamma^{ab}(\sigma) + \sqrt{-\gamma}\delta\gamma^{ab}\right]\eta_{\mu\nu}\frac{\partial X^{\mu}}{\partial \sigma^a}\frac{\partial X^{\nu}}{\partial \sigma^b} \nonumber\\ =& -\frac{T}{2}\int d\sigma^2\:\sqrt{-\gamma} \underbrace{\left[ -\frac{1}{2}\gamma_{ab}(\sigma)\gamma^{cd}(\sigma)\eta_{\mu\nu}\frac{\partial X^{\mu}}{\partial \sigma^c}\frac{\partial X^{\nu}}{\partial \sigma^d} + \eta_{\mu\nu}\frac{\partial X^{\mu}}{\partial \sigma^a}\frac{\partial X^{\nu}}{\partial \sigma^b}\right]}_{T_{ab}(\sigma)} \delta\gamma^{ab} \nonumber\\ = & -\frac{T}{2}\int d\sigma^2\:\sqrt{-\gamma}T_{ab}(\sigma)(- \partial^b\epsilon^a(\sigma) - \partial^a\epsilon^b(\sigma)) \nonumber\\ =& T\int d\sigma^2\:\sqrt{-\gamma}T_{ab}(\sigma)\partial^a\epsilon^b(\sigma), \end{align} where in the last line I used the symmetry of $T_{ab}(\sigma)$.

Now using the continuity equation $\partial^aT_{ab}(\sigma) = 0$ for the stress-energy tensor, \begin{align} \delta S_P =& T\int d\sigma^2\:\sqrt{-\gamma}T_{ab}(\sigma)\partial^a\epsilon^b(\sigma)\nonumber\\ =& T\int d\sigma^2\:\sqrt{-\gamma}\left[\partial^a(T_{ab}(\sigma)\epsilon^b(\sigma)) - \underbrace{\partial^aT_{ab}(\sigma)}_{0}\epsilon^b(\sigma)\right]\nonumber\\ =& T\int d\sigma^2\:\sqrt{-\gamma}\partial^a(T_{ab}(\sigma)\epsilon^b(\sigma)). \end{align}


1 Answer 1


To derive the result, we can use the fact that $T_{ab} \propto \frac{ \delta S}{\delta \gamma^{ab}}$.

Given the transformation $\sigma \mapsto \sigma - \epsilon$, we can derive the change in the worldsheet metric, $\delta \gamma_{ab}$. Then, we compute the variation of the action under this change $\delta S = \int d^{2} \sigma \frac{\delta S}{\delta \gamma^{ab}} \delta \gamma^{ab}$, which turns out to be proportional to $\int d^{2} \sigma T_{ab} \partial^{a} \epsilon^{b}$. Your quoted result follows after an integration by parts and ensuring that the variation $\delta S$ vanishes.

  • $\begingroup$ Thanks for your answer, I've given it another go with what you have said, and I think I am close. There are a few things that I'm still not sure about, so I am going to put in my answer and hopefully you can point out my mistakes. $\endgroup$
    – Bedge
    Jan 6, 2022 at 13:17
  • $\begingroup$ So in my attempt, there are a few things I'm unsure about. Firstly, is the factor of $\sqrt(-\gamma)$ in the integrand a problem? Secondly, if my work is correct, why didn't I need to consider variations of the derivatives $\frac{\partial X^{\mu}}{\partial \sigma^a}$? Thanks for any help you can give me. $\endgroup$
    – Bedge
    Jan 6, 2022 at 13:30

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.