Universality of citation distributions: Toward an objective measure of scientific impact