Showing that path-independent parallel transport implies that the curvature tensor vanishes.

816 Views Asked by At

I am working through an exercise in the book 'Manifold Theory: An introduction for mathematical physicists' which asks me to show that path-independent parallel transport implies that the curvature tensor vanishes.

A hint given in the book says that I can use the fact that $$\tag{1} \partial_iX^k+\Gamma^k_{i\sigma}X^\sigma=0$$ is integrable. This equation comes from that fact that $X$ is undergoing parallel transport.

I believe I am almost there but I am unsure how exactly how I have used the fact that (1) is integrable/ path independent.

Ok so my solution went something like this;

Let $M$ be a differentiable manifold and let $X$ be smooth a vector field passing through a point $p \in M$. Let $\gamma:\mathbb{R}_+ \to M$ be a path for which $\gamma(0)=p$. Then by parallel transport we obtain a family of vector fields $S=\{X_{\gamma(t)}\}$ which have been obtained from $X$ by parallel transport. Accordingly we have that each vector in $S$ satisfies the parallel transport equations (1).

The Riemann curvature tensor in a coordinate expression is

$$R^\rho{}_{\sigma\mu\nu} = \partial_\mu\Gamma^\rho{}_{\nu\sigma} - \partial_\nu\Gamma^\rho{}_{\mu\sigma} + \Gamma^\rho{}_{\mu\lambda}\Gamma^\lambda{}_{\nu\sigma} - \Gamma^\rho{}_{\nu\lambda}\Gamma^\lambda{}_{\mu\sigma}$$ and hence we can compute the value of $R^\rho{}_{\sigma\mu\nu}X^\sigma_{\gamma(t)}$ at a particular point just by multiplying the above equation by $X^\sigma_{\gamma(t)}$. This is given by \begin{align} R^\rho{}_{\sigma\mu\nu} X^\sigma_{\gamma(t)}&= (\partial_\mu\Gamma^\rho{}_{\nu\sigma})X^\sigma_{\gamma(t)} - (\partial_\nu\Gamma^\rho{}_{\mu\sigma})X^\sigma_{\gamma(t)} + \Gamma^\rho{}_{\mu\lambda}(\Gamma^\lambda{}_{\nu\sigma}X^\sigma_{\gamma(t)}) - \Gamma^\rho{}_{\nu\lambda}(\Gamma^\lambda{}_{\mu\sigma}X^\sigma_{\gamma(t)})\\ &=(\partial_\mu\Gamma^\rho{}_{\nu\sigma})X^\sigma_{\gamma(t)} - (\partial_\nu\Gamma^\rho{}_{\mu\sigma})X^\sigma_{\gamma(t)} - \Gamma^\rho{}_{\mu\lambda}\left(\partial_\nu X^\lambda_{\gamma(t)}\right) + \Gamma^\rho{}_{\nu\lambda}\left(\partial_\mu X^\lambda_{\gamma(t)}\right)\\ &=(\partial_\mu\Gamma^\rho{}_{\nu\sigma})X^\sigma_{\gamma(t)} + \Gamma^\rho{}_{\nu\lambda}\left(\partial_\mu X^\lambda_{\gamma(t)}\right) - \left((\partial_\nu\Gamma^\rho{}_{\mu\sigma})X^\sigma_{\gamma(t)} + \Gamma^\rho{}_{\mu\lambda}\left(\partial_\nu X^\lambda_{\gamma(t)}\right) \right)\\ &=\partial_\mu\left(\Gamma^\rho{}_{\nu\sigma}X^\sigma_{\gamma(t)}\right)-\partial_\nu\left(\Gamma^\rho{}_{\mu\lambda}X^\lambda_{\gamma(t)}\right)\\ &=-\partial_\mu\partial_\nu(X^\rho_{\gamma(t)} )+\partial_\nu\partial_\mu(X^\rho_{\gamma(t)})\\ &=0 \end{align} Where I have used equation (1), some index gymnastics, the product rule, equation (1) again and equality of mixed partials, In that order.

Hence If we took the original vector field $X$ to be non-vanishing, it follows that $X_{\gamma(t)}$ is non vanishing, thus the curvature tensor is zero everywhere in which $X_{\gamma(t)}$ is defined. I believe now I must use the fact that I could of chosen a different path/ integrability of (1) to say that $X_{\gamma(t)}$ is defined everywhere and hence the curvature tensor vanishes. I am unsure on how to proceed.

Any hints of comments would be much appreciated.

1

There are 1 best solutions below

1
On

If you are aware of the exponential map, one can apply Gauss Bonnet in the following way.

Given $p \in M$, let $U \subset M$ be a (totally) normal neighbourhood of $p$. That is, $exp_p: V \subset T_pM \to U$ is a diffeomorphism for some set $V$ contained in the tangent space at $p$.

Pick $q \in U$ and let $exp_p(y)=q$, then one can consider any plane $\sigma \subset T_pM$ containing the vector $y$. The image of $\sigma$ under $exp_p$ is a surface contained in our set $U$. The Gaussian curvature $K$ of this surface is the sectional curvature $K(\sigma)$.

One can proceed as in the question here to establish that this sectional curvature is $0$. (note the curves that one is transporting along must be contained in the surface).

This is true for any plane $\sigma$. The Riemann curvature tensor is determined entirely by its sectional curvatures and hence must be $0$.