par Bunster, Claudio;Henneaux, Marc 
Référence Physical Review D - Particles, Fields, Gravitation and Cosmology, 83, page (045031)
Publication Publié, 2011
Référence Physical Review D - Particles, Fields, Gravitation and Cosmology, 83, page (045031)
Publication Publié, 2011
Article révisé par les pairs
| Résumé : | There exists a formulation of the Maxwell theory in terms of two vector potentials, one electric and one magnetic. The action is then manifestly invariant under electric-magnetic duality transformations, which are rotations in the two-dimensional internal space of the two potentials, and local. We ask the question: Can duality be gauged? The only known and battle-tested method of accomplishing the gauging is the Noether procedure. In its decanted form, it amounts to turning on the coupling by deforming the Abelian gauge group of the free theory, out of whose curvatures the action is built, into a non-Abelian group which becomes the gauge group of the resulting theory. In this article, we show that the method cannot be successfully implemented for electric-magnetic duality. We thus conclude that, unless a radically new idea is introduced, electric-magnetic duality cannot be gauged. The implication of this result for supergravity is briefly discussed. © 2011 American Physical Society. |
