On the Influence of a Magnetic Field on the Polarization of Resonance Radiation

From Current Literature

On the Influence of a Magnetic Field on the Polarization of Resonance Radiation.

R. W. Wood and A. Ellett. On the influence of magnetic fields on the polarisation of resonance radiation. Proceedings of the Royal Society A. 103, 396, 1923.

The phenomenon of polarization of fluorescence of vapors was discovered by Wood as early as 1911. Polarization was found in the fluorescence of vapors of iodine, sodium, potassium, lithium, and others; it was not observed only in the ultraviolet resonance radiation of mercury vapor. In recent years, however, in this case as well polarization was found by Jordan and Ellett. At the end of 1922 Wood, changing the experimental conditions, finally obtained strongly polarized resonance radiation of mercury vapor. After publishing a report on this (Phil. Mag. December 1922), Wood noticed, however, that the phenomenon was extremely capricious and would appear and disappear despite apparently identical experimental conditions. Further investigation revealed that the cause of this instability of the results was the different orientation of the magnetic vector of the exciting light beam with respect to the earth’s magnetic field. In the case when the magnetic vector of the light was parallel to the earth’s field, the polarization almost completely disappeared; it reached 90% when the earth’s field was compensated by the weak field of a solenoid. The optical investigation was carried out by means of a quartz double-refracting prism by a photographic method. The dependence of the effect on the strength of the field, in the case when the latter was parallel to the magnetic vector of the polarized exciting light, is shown in the following table:

It is clear from the table that the whole effect takes place almost completely within the limits of one gauss, and that a further increase in the field strength should remain practically without influence.

Having established this fact, the authors turned to the study of the influence of a magnetic field on the polarization of the resonance radiation of sodium vapor. In this case the polarization, even with excitation by polarized light, is very small, about 5%; the earth’s field, as the experiments showed, has no noticeable influence. To eliminate the polarization in the case of parallelism of the magnetic field to the magnetic vector of the exciting light (with observation in the direction of the magnetic vector), a field of strength about 100 gauss was required.

Changing, however, the direction of the magnetic field, the authors discovered a new phenomenon—a sharp increase in the degree of polarization of the resonance light.

In the experiments with sodium the light source was a discharge tube of the same type as the tube that Wood had used to excite the hydrogen spectrum with a large number of lines of the Balmer series¹. Metallic sodium was introduced from a side tube into a space with a pressure of about 1 mm, flushed with hydrogen. The part of the tube used—

appearing as the exciting source, was heated by a Bunsen burner. The light from a small section of the discharge tube passed through a diaphragm, a lens, a large Nicol, and a second lens, by means of which it was concentrated at the center of a glass sphere containing a small quantity of metallic sodium. The sphere was heated to approximately 180° by a current of hot air. The polarization was studied with the aid of a quartz wedge, a double-refracting prism, and a Babinet compensator.

Of special interest are the results obtained in the case when the magnetic field was perpendicular to the magnetic vector of the exciting polarized light. Let the electric vector of the latter be vertical, and let the observation be carried out along the magnetic vector, i.e. perpendicular to the direction of propagation of the exciting beam. Different results are obtained depending on what angle the external magnetic field makes with the electric vector of the exciting light \(H\). If, for example, the angle is 90°, then we shall denote the field by \(H_{90}\); if the angle is zero, by \(H_0\), and so on. In the absence of a field, as we have said, the degree of polarization is \(p = 5\%\); at \(H_{90}\) (100 gauss) \(p = 30\%\); at \(H_0\) equally strong polarization is obtained; at \(H_{45}\), \(p = 0\).

Let now the electric vector of the exciting light be horizontal, and let observation be carried out in its direction. In the absence of a field no polarization is observed. If the lines of force of the external magnetic field are perpendicular, in this case, to the line of observation (and to the electric vector), then for any position of the external field relative to the magnetic vector of the exciting polarized light, strong polarization is observed (field 100 gauss), approximately unchanged in magnitude, but at the same time the electric vector of the polarized resonance light rotates together with the field, lagging in phase by 90°. Thus, when the external field is vertical, the electric vector of the polarized resonance light is horizontal.

After the indicated rotation had been discovered for the horizontally positioned electric vector of the exciting light, the authors were able to establish a similar rotation also for the vertical position of the electric vector of the exciting light.

In the present case, however, the matter is complicated by the fact that, as we saw above, the degree of polarization changes when the external magnetic field is rotated, becoming zero at \(H_{45}\) and acquiring maximum values at \(H_0\) and \(H_{90}\).

The authors were unable, for sodium, to obtain polarization greater than 25–30% under any conditions. In connection with this there arose the supposition that, in mercury vapor, the corresponding effects of strengthening and weakening of polarization and of rotation of the electric vector of the resonance light could be masked by the very strong normal polarization of mercury. The authors again returned to experiments with mercury and, with the horizontal position of the exciting electric vector, found the appearance of polarization and a change in the intensity of the resonance light in the case when the magnetic lines of force of the external field were perpendicular to the line of observation, i.e. to the electric vector of the exciting light. Just as in the above-described experiments with mercury, these experiments were performed with an external field of strength 1 gauss, i.e. of the order of magnitude of the Earth’s field. It was found that in this case too the electric vector of the resonance light rotates together with the external field, just as in the experiments with sodium.

In connection with the formal explanation of the remarkable phenomena set forth, proposed by C. Darwin (Ch. Darwin), the authors also carried out observations of the longitudinal effect, measuring the polarization of the resonance light in the direction of propagation of the exciting beam. These experiments are extremely difficult, but nevertheless success was achieved in this case as well. The results of the experiments are reported briefly in an addendum to the article. The following experiment is extremely important. In the case when the external field was parallel to the magnetic vector of the exciting

of the light, and the electric vector of the latter was vertical, then, as we saw in transverse observation, all traces of polarization were absent. In longitudinal observation in the same case, strong polarization was detected. This exhausts the experimental data reported in the paper under review.

In addition, the article gives the following formal interpretation of the phenomenon, proposed by Ch. Darwin. It is assumed that, under the action of an external magnetic field, the radiating circular, plane orbits of the electrons of the corresponding atoms are set perpendicular to the field. Let us analyze from this point of view the first experiment, when the directions of observation, of the magnetic vector of the exciting light, and of the external field are parallel. According to the hypothesis, the circular plane orbits will set themselves perpendicular to the field, and each of them, under the influence of the vertical exciting vector, will send to the observer (transverse observation) a circularly polarized wave. Right- and left-handed rotations are equally probable; as a result, the observer, in agreement with the experiment, will see no polarization. Conversely, in longitudinal observation in the same case, the observer will see the same oriented orbits “in profile” and will receive from them (both from right- and left-rotating ones) linearly polarized light, again in agreement with the experiment.

When the electric exciting vector is vertical and the external magnetic field is rotated in the vertical plane perpendicular to the horizontal line of observation, coinciding with the direction of the magnetic exciting vector, then on the basis of the hypothesis just given, in agreement with experiment, the following results are obtained. For \(H_{90}\) the orbits are perpendicular to the external field and parallel to the electric exciting vector; the observer sees them vertically, “in profile,” and therefore receives vertical, linear vibrations; for \(H_0\), horizontal linear vibrations are observed analogously. For \(H_{45}\) the orbits are visible to the observer “in profile” at an angle of \(45^\circ\) to the horizon, but the vertical electric vector excites in them horizontal and vertical vibrations of equal amplitude, i.e. for the observer the light appears unpolarized.

If the electric exciting vector is horizontal and observation is made in its direction, while the external field is rotated in the vertical plane, then, as is not difficult to see, the orbits of the atoms, according to the hypothesis, will all the time be parallel to the electric vector and perpendicular to the field; the observer will see the light linearly polarized all the time to the same degree, but the electric vector of the resonance light will rotate as the field rotates, in agreement with the experiment.

In connection with Darwin’s theory, the experiments of Wood and Ellett must be compared with the experiments of Stern and Gerlach \(^{1}\), published last year. The discreteness of the orientation of atoms in a magnetic field is revealed in the new experiments with extraordinary clarity. As is known, this discreteness follows from the postulates of Bohr’s theory. Of course, many aspects of the phenomenon are still not explained by the theory set forth. In particular, the quantitative side and the relation of the polarization in a magnetic field to the normal polarization of fluorescence still remain to be taken into account.

S. Vavilov.

\(^{1}\) Cf. p. 301.

1. Cf. Uspekhi F. N. ↩