P. Debye has predicted that x-rays diffusely scattered from simple cubic crystals should decrease in intensity by a factor ($1-e^{-M}$) as the temperature decreases. $M$ is a quantity depending on the characteristics of the crystal, wave-length of the x-rays used, and the angle of scattering. Later, Waller's modification of the Debye factor, namely, that the exponent should be $-2M$, enjoyed some experimental success in regular reflection. Quite recently, Jauncey and Harvey have announced an entirely new formula (in which diffuse scattering and regular reflection are shown to be complementary) for predicting the change in intensity of diffuse scattering with temperature, from the known temperature effect on atomic structure factors.

The present research was undertaken: first, to establish the effect of temperature on diffuse scattering from rocksalt as an experimental fact over a wide range of temperature from 120°K to 750°K; second, to attempt to distinguish among the three predicted values. An ionization method was first used, but was found to give erratic values at the larger scattering angles. A photographic method gave the following ratios of intensities between 120°K and 300°K: 28°, 0.58; 35°, 0.64; 45°, 0.68; 58°, 0.79. Between 750°K and 300°K the ratios were: 28°, 1.20; 35°, 1.26; 45°, 1.23. At the smaller scattering angles, these ratios agreed with those obtained by the ionization method. The effect of temperature was thus established as an experimental fact. On comparison with the various theories, it was found that the Debye theory is definitely precluded. At low temperatures, the agreement appears to be with the Waller modification of Debye's theory, when zero point energy for the atoms of the crystal is included. At the high temperature, the agreement is with the formula of Jauncey and Harvey. A discussion of these results leads to the conclusion that the agreement with the Waller theory at low temperatures is probably fortuitous, and that the formula of Jauncey and Harvey probably predicts more closely the ratios for sylvine, to which the formula is more strictly applicable than to rocksalt.

• Received 12 June 1931

DOI:https://doi.org/10.1103/PhysRev.38.604