On the Question of the Nature of “Candoluminescence” of $\mathrm{Al_2O_3}$

V. A. Sokolov, A. N. Gorban

(Presented by Professor Doctor A. A. Vorob’ev)

In work [1], devoted to the study of the “candoluminescence” of calcium oxide, as a result of comparing the spectra of calcium oxidation and the spectra of the thermal radiation of CaO, the purely thermal nature was established of those features of the luminescence of this oxide which had been associated with ideas about candoluminescence.

Following this, we carried out similar studies of aluminum oxide, to which the ability for candoluminescence is also ascribed [2].

As in the preceding work, the luminescence of the oxide was excited by the flame of illuminating gas. As for the spectra of aluminum oxidation, in order to obtain them, despite many attempts, we were unable to ignite aluminum in an oxygen-filled quartz flask, as was done in obtaining the spectra of calcium oxidation [1]. The reason for this was evidently the aluminum oxide film, well known for its unusual stability; the temperature of the illuminating-gas flame by which the flask was heated, even when a blower was used, proved insufficient to ignite the aluminum.

It was possible to bring about the combustion of aluminum only in the following manner. An electric arc was excited between electrodes made of aluminum wire, and a stream of oxygen was then directed onto the arc. Owing to the very high temperature of the arc, the aluminum electrodes in the oxygen atmosphere flared up brightly and continued to burn even after the electric current was switched off. The aluminum flame thus produced was photographed already in the absence of an electric discharge with the aid of an ISP-51 spectrograph.

The spectrum of aluminum oxidation obtained as a result of photometering the photographs and allowing for the sensitivity of the plate is shown in Fig. 1. The same figure also shows the radiation spectrum (“candoluminescence”) of aluminum oxide when excited by the flame of illuminating gas. As can readily be seen, both spectra are continuous and reproduce one another rather accurately. This indicates that the luminescence of the aluminum flame during oxidation is due to the thermal radiation of the solid oxide particles formed in the process. The purely temperature character of both spectra in the investigated region of wavelengths, to which candoluminescence is ascribed, is evidenced by the fact that

both curves shown in Fig. 1 vary with frequency \(\left(\frac{1}{\lambda}\right)\) in proportion to the quantity \(\lambda^{-5} e^{-\frac{C_2}{\lambda T}}\). Indeed, as Fig. 2 shows, the experimental points corresponding to the values of \((\lg E + 5 \lg \lambda)\) as a function of \(\frac{1}{\lambda}\) lie relatively well (within the limits of scatter) on the theoretical straight lines expressing Kirchhoff’s law [1] in logarithmic form.

Fig. 1

Fig. 2

Thus, as in the case of CaO, from the experiments described we may draw the following conclusions.

1. Oxidation of aluminum cannot be the cause of the candoluminescence of its oxide, since that carried out in the purest form (go-

... in an oxygen atmosphere) it gives only the spectrum characteristic of purely thermal incandescence of the oxide.

1. The so-called candoluminescence of \( \mathrm{Al_2O_3} \) under the action of a flame is, at high temperatures, merely the ordinary thermal radiation of the given oxide, obeying Kirchhoff’s law, and is not appreciably connected with any forms of true luminescence.

REFERENCES

1. Sokolov V. A., Grozina I. S., Gorban A. N. On the nature of the candoluminescence of calcium oxide (present collection).

2. Nichols E. L., Howes H. L., Wilber D. T. Cathodoluminescence and Luminescence of Incandescent Bodies, Washington, 1928.