I've been having difficulty finding a source that lists all the properties of the spinor bundle of a string worldsheet explicitly, so I've had a go at creating my own description. I'd really appreciate it if someone could tell me if the following is true:

Take the worldsheet to be some 2d pseudo-Riemannian orientable manifold $M$.

One can associate with each point $x \in M$ a 2d tangent space $TM_{x}$. The disjoint union of $TM_{x}$ at all $x$ defines the total space $TM$ of a tangent bundle ($TM$, $\pi_{TM}$, $M$) whos projection is given by:

\begin{equation} \pi_{TM}: TM \rightarrow M \end{equation}

The worldsheet $M$ is the base space of the tangent bundle and each $TM_{x}$ is a fibre.

Since the tangent space is 2d, the bases that exist in each $TM_{x}$ are 2d also. Since the base space is pseudo-Riemannian, so is the tangent space and the ordered bases (frames) that exist on each $TM_{x}$ are 'pseudo-orthonormal'. This would mean that the bases transform under an $O(1,1)$ group. However, Since the base space $M$ is orientable, so is each $TM_{x}$ and that means that the frames are oriented pseudo-orthonormal and transform under $SO(1,1)$ instead.

This allows the oriented orthonormal frame bundle (a specific sub-class of principal bundle) to be written as $(F_{SO(1,1)}(M), \pi_{F}, M, SO(1,1))$, where the projection acts as:

\begin{equation} \pi_{F}: F_{SO(1,1)}(M) \rightarrow M \end{equation}

The fibre $F_{x}$ of this frame bundle at a point $x$ on $M$ is the set of all frames of $TM_{x}$ at the same point $x$. $F_{x}$ is homeomorphic to the gauge group $SO(1,1)$ and is said to be an $SO(1,1)$-torsor.

However, now one can define a lift of the group $SO(1,1)$ to $Spin(1,1)$. The corresponding frame bundle is now $(P, \pi_{P}, M, SO(1,1))$ with projection:

\begin{equation} \pi_{P}: P \rightarrow M \end{equation}

The fibre $P_{x}$ of this frame bundle at a point $x$ on $M$ is the set of all frames of $TM_{x}$ at the same point $x$. $P_{x}$ is homeomorphic to the gauge group $Spin(1,1)$ and is said to be an $Spin(1,1)$-torsor.

How can the set of all frames in $TM_{x}$ be homeomorphic to both $SO(1,1)$ and $Spin(1,1)$?

The spinor bundle can then be defined to be given by $(S, \pi_{S}, M, \Delta_{(1,1)} Spin(1,1))$, with projection that acts as:

\begin{equation} \kappa: S \rightarrow M \end{equation}

Here $S$ is given by:

\begin{equation} S = P \times_{\kappa} \Delta_{(1,1)} = (P \times \Delta_{(1,1)})/Spin(1,1) \end{equation}

The fibre is given by $\Delta_{(1,1)}$ which is the Hilbert space of all spinor states. Each section of this bundle then corresponds to a particular Majorana-Weyl spinor field configuration on the worldsheet.


1 Answer 1


It is nearly correct. First, $F_x$ is the space of oriented and orthonormal frames of $T_x M$. Furthermore, you gave the correct definition of the $Spin(1,1)$ bundle but confused a little bit its meaning. The $Spin(1,1,)$ bundle is defined as the lift of the $SO(1,1)$ bundle to $Spin(1,1)$. This means we have a principal $Spin(1,1)$-bundle $P$ and a principal bundle map $\phi: P \to F$ which is at the same time a covering (this literally lifts the double covering of $SO(1,1)$ by $Spin(1,1)$ to the bundle picture). For more details see http://en.wikipedia.org/wiki/Spin_structure#Spin_structures_on_Riemannian_manifolds. Now $P$ is not a frame bundle. More precisely, for every oriented, orthonormal frame of $T_xM$ there exists two points $p$ and $p'$ in the fiber $P_x$ which yield this frame by $\phi(p) = \phi(p')$. Thus $P_x$ is not homeomorphic to the frames on $T_xM$.

  • $\begingroup$ Thanks for the reply Tobias. Wouldn't $F_{x}$ be the space of oriented and pseudo-orthonormal frames of $T_{x}M$, just to be a little more precise? (Since $T_{x}M$ has Lorentzian metric signature). And I understand the $Spin(1,1)$-bundle now, thank you. $\endgroup$ Mar 4, 2015 at 10:49
  • 1
    $\begingroup$ You are welcome. Yes of course, it is pseudo-orthonormal. I was just too lazy to write it ;-). But note that the same argumentation goes through for the Riemannian case and in arbitrary dimension (with the obvious modifications of the groups, i.e. you get a $Spin(n)$ bundle). $\endgroup$ Mar 4, 2015 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.