New J. Phys. 18, 023009 (2016)

Almost 90 years ago, Richard Tolman, Paul Ehrenfest and Boris Poldolsky published a paper titled 'On the gravitation field produced by light' (Phys. Rev. 37, 602–615; 1931), which suggested that, via weak-field general relativity theory, a cylindrical pulse, or 'pencil', of light causes no acceleration to a co-propagating test ray. One consequence of the analysis is that a light pulse does not self-interact with its own gravitational field. However, they noted that if the test ray is instead counter-propagating it can bend the trajectory of light, an effect similar to the bending of light by gravitational fields of celestial bodies. Now, Dennis Rätzel, Martin Wilkens and Ralf Menzel from the University of Potsdam, Germany, have theoretically proposed that the gravitational field of a linearly polarized light pulse is 'modulated' by the electric field strength. However, in the case of circularly polarized light, no modulation is expected. The team theoretically investigated the gravitational effect on different test particles showing that both attraction and repulsion effects from the pulse trajectory may occur.