Dispersion and Absorption of Water and Ethyl Alcohol in the Region of Short Electric Waves

Willy Möbius. Über die Dispersion von Wasser u Äthylalkohol zwischen 7 und 35 mm. Wellenlänge und Vorversuche zur Verwendung nach kürzerer elektrischer Wellen. Annalen der Physik. B. 62. p. 293, 1920.

The refractive index of water in the region of short electric waves, as the already quite old investigations of Lampa and later ones have shown, is close to the square root of the dielectric constant and lies around 8.

The refractive index of water in the region of long heat waves, 0.3 mm, is only slightly greater than in the visible spectrum. This compels one to assume, in the intermediate region, a band of anomalous dispersion, the discovery of which the author sets as the task of his investigation. The construction of the vibrator does not differ from earlier ones (Righi, Lebedev).

The receiver was always a comparatively long conductor, not tuned to resonance and possessing, owing to the inclusion of a thermoelement with a resistance of 250–400 Ω, considerable damping; such a system permits the use of one and the same receiver for different wavelengths.

The wavelength is measured by means of Boltzmann mirrors or a quartz interferometer (after Rubens and Hollnagel).

On the interference curves, the presence was found of a large number of incoming waves shorter than the fundamental wavelength of the vibrator and superposed upon it. Waves of 0.1; 0.11; 0.17; 0.25; 0.30; 0.50; 0.70; 1.0; 1.1; 1.2; 1.7; 1.8; 2.0; 2.5; 3.0; 3.8; 4.0; 5.0 mm were detected; waves of 2.0, 2.5, and 3.0 mm occurred most often. The incoming short waves have less damping than the fundamental one and are very variable, which makes working with them and isolating them difficult. Some vibrators give almost pure short waves, while the fundamental wave is not noticeable.

The occurrence of short waves may be explained partly by the presence of overtones of the vibrator; still shorter and variable waves may arise from oscillations occurring in metallic particles torn off by sparks; finally, it is possible to suppose oscillations of positive metal ions about negative ones; in this last case, calculation gives the emitted wavelength:

\[ \lambda = 0{,}005472 \frac{A}{\sqrt{d}} \ \text{mm}, \]

where \(A\) is the atomic weight, \(d\) the density of the metal, which for platinum gives:

\[ \lambda = 0{,}232 \ \text{mm}. \]

These considerations make it possible to suppose that different metals may give different constant wavelengths.

In one experiment it proved possible to isolate short waves by means of a mirror pasted with strips of tinfoil arranged in the form of a regular grating, similar to a crystal lattice; the strip length was 0.8 mm, the width 0.2 mm, the transverse spacing 0.8 mm, the longitudinal spacing 0.2 mm. With the vibrator’s own wavelength of 18.8 mm, a wave \(\lambda = 3.0\) mm was isolated; but after two days the same vibrator gave an entirely different wave, which shows the difficulty of using short waves for systematic investigations. For waves in the interval from 35 to 7 mm, measurements were made of absorption, dispersion, and reflection independently of one another. The absorption coefficient was measured from the change in the intensity of the oscillations on passage through a definite layer of liquid poured into a flat vessel; diffraction phenomena for such short waves can be entirely neglected. The coefficient of reflection was measured compa-

by varying the intensity of the reflected and direct ray. The refractive index was measured from the deviation in an ebonite prism filled with the liquid.

From the absorption coefficient \(k\) and the reflection coefficient \(R\), it is possible to calculate the refractive index by the optical formula:

\[ n=b\cos\varphi+\sqrt{(b^{2}-1)\cos^{2}\varphi-\eta^{2}}, \]

where

\[ b=\frac{1+R}{1-R}, \]

and \(\varphi\) is the angle of incidence.

Thus the curves show absorption and, independently of it, dispersion, measured directly and calculated from \(R\) and \(k\). The absorption curve has a sharp maximum at \(\lambda=22\) mm. At the same wavelength the refractive index has a sharp minimum \(n=5\); at \(\lambda=35\) mm, \(n=7.6\); at greater wavelengths, as is seen from many measurements by other authors, \(n\) reaches a value of 9; toward smaller wavelengths \(n\) increases to 7 at \(\lambda=10\) mm, and then again begins to fall; as the single measurement at \(\lambda=3\) mm showed, \(n=4.3\), which fits well with the continuation of the author’s curve, but disagrees with the old measurements of Lampa, who found in this region \(n\) about 9.

For ethyl alcohol a decrease of absorption into the region of short waves was found; the refractive index falls from 3.5 \((\lambda=60\ \mathrm{mm})\) to 2 \((\lambda=7\ \mathrm{mm})\).

The absorption band at \(\lambda=22\) mm found by the author enables him to propose a new method of isolating short incoming waves; namely, by giving the vibrator’s fundamental wave a length of 22 mm and placing before it a layer of water, it is possible completely to absorb the fundamental wave and to isolate the short incoming waves.

S. Rzhevkin.