The radial metric function does not identify null surfaces
Yi-Hsiung Hsu, Will Barker, Michael Hobson, Anthony Lasenby
TL;DR
The paper tackles the problem of reliably identifying null hypersurfaces across coordinate systems by using EF-like and Kruskal-like transformations. It shows that for static spacetimes the physically correct criterion is $g_{tt}=0$ (not $g^{rr}=0$) and derives the conditions under which Kruskal-like coordinates recover the Vollick proxy, while acknowledging cases where such coordinates are unavailable. The approach generalises to axisymmetric stationary spacetimes, where the null condition is captured by det$(h_{ab})=0$ on the induced metric, and is demonstrated through applications to Ellis drainhole and traversable wormholes (including rotating Teo-type geometries). Overall, the method provides a physically intuitive link between null surfaces and photon geodesics and offers a coordinate-robust framework for horizons in diverse spacetimes.
Abstract
We investigate the conditions under which a hypersurface becomes null through the use of coordinate transformations. We demonstrate that, in static spacetimes, the correct criterion for a surface to be null is $g_{tt} = 0$, rather than $g^{rr} = 0$, in agreement with the results of Vollick. We further show that, if a Kruskal-like coordinate exists, the proxy condition $g^{rr} = 0$ is equivalent to $g_{tt} = 0$ if $\partial_r g_{tt} \neq 0$ and both $g^{rr}$ and $g_{tt}$ vanish at the same rate near the horizon. Our method extends naturally to axisymmetric stationary spacetimes, for which we demonstrate that the condition $\det\big(h_{ab}\big) = 0$ for the induced metric on a null hypersurface is recovered. By contrast with the induced metric approach, our method provides a physical perspective that connects the general null condition with its underlying relationship to photon geodesics.
