The effect of a transverse magnetic field on the propagation of light in vacuo

Abstract

In modern physics there have developed two complementary—and apparently mutually contradictory—modes of description of radiation processes and of the motion of molecules, atoms, electrons and protons. How far can the parallelism in description of photons (light quanta) and members of the second group of entities be carried ? It is now well known that the physical effects by which an entity is recognised are, in all cases, atomic and individual, e. g ., photoelectric and Compton effects, Scintillation on a fluorescent screen, effect on a photographic plate. Such effects are quite naturally correlated with the concept of a moving particle with energy and momentum. On the other hand, the motion of these particles can only be completely described by the wave method. Except in the case of very long wireless waves, the waves are never observed directly as waves with periodic physical effects in space and time, and even in the case of the exception it seem probable that the periodicity observed is only a large scale phenomenon (for comparatively large numbers of particles). The development of the analogy between photons and entities of the second class (atoms, electrons, etc.) has already reached the stage where it is possible to give a wave description of the motion of the particles in all cases and to assign to the particles an energy-momentum four-vector within the limits of Heisenberg's principle. There remains, however, a fundamental distinction in current theory. All the entities in the second class have electro-magnetic particle properties while none have been assigned to the photon. Electrons, ions, and protons have electric charge which is quantised in integral multiples of the electronic charge ± e . Atoms possess the electromagnetic particle properties, magnetic moment and possibly electric moment, while molecules have magnetic and electric moments and mechanical moments of inertia. Since the photon is assumed to be electromagnetic in origin, and can produce electromagnetic effects, it is necessary to assign to it some electro-magnetic character. The simplest particle properties which one can postulate are those of electric moment and magnetic moment; free electric charge is excluded by the fact that light is not deflected in a uniform electric or magnetic field. The present investigation was carried out with the object of detecting, if possible, the existence of the magnetic moment of a photon. The Stern-Gerlach method of the non-uniform field involving the deflection of particles moving with velocity of light, presents obvious difficulties. The method actually adopted depends for its sensitiveness on the interference properties of light, and the principle was the following. Light was passed through a Fabry-Perot étalon placed in a strong magnetic field which was perpendicular to the direction of propagation of the light. A particle with a magnetic moment μ parallel or antiparallel to the field H, would undergo a change in energy ΔE = ± μH on entering the field and in accordance with the principles of the quantum theory would experience a change in frequency Δ v = ±μH/ h . This would involve an effective change in wave-length Δλ = ±μHλ 2 / hc and a change in the interference pattern formed by the interferometer of the type observed in the normal Zeeman effect.