Seems to me that the size of the star would be a problem. Assuming the process of actually building stuff large enough to cover a star is solved, how do you control the polarizer to get any reasonable bandwidth? The light emitted from one side of the star will be shifted in time relative to the light emitted from the other side, in anything other than a very narrow beam. Targets outside of this beam will not really receive much of a signal due to interference.