The band spectrum of water vapour

Abstract

1. Experimental Evidence on the Nature of the Emitter . Experiments on the spectrum of burning hydrogen or hydrogen compounds by Liveing and Dewar led to the discovery of the now well-known band spectrum of water vapour in the ultra-violet. Its discovery was published simultaneously by Huggins. The presence of water vapour, or of both its constituent elements in flames, in an electric discharge or in an arc, always gives rise to this spectrum, which consists of several well-marked bands. With oxygen, even when it has been subjected to the most powerful means of drying, it is not possible to eliminate completely the water-vapour bands, and con­sequently it was believed that these bands were due to oxygen. Eder and Valenta showed, however, that the addition of a small quantity of hydrogen to the oxygen in a discharge tube greatly enhanced the intensity of the bands. Their appearance with dry oxygen is presumably due to the emission by the electrodes of small traces of hydrogen, which combine with the oxygen under the influence of the discharge. Grebe and Holtz, in 1912, came to the con­clusion that the emitter was some combination of oxygen and hydrogen, although not necessarily H 2 0. Reiss obtained the spectrum with cyanogen burning in very dry oxygen, and in 1914 Fortrat attributed the spectrum to oxygen. More recent experiments have been carried out by Watson, who found that when water vapour was excited by a weak electrodeless discharge, the series lines α, β, and γ of hydrogen made their appearance along with the water-vapour bands. This indicates that even with weak excitation the water molecule may be broken up, at least into OH and H. Under similar conditions he obtained the benzene bands without any evidence of the breaking up of the complex benzene molecule. When water vapour was subjected to a powerful disruptive discharge, the characteristic oxygen bands appeared, which showed that under these conditions complete dissociation of the water molecules was taking place. Watson concluded that the bands were probably due to the OH ion, and in support of this view he gave further evidence, which will be referred to later, based on an estimate of the moments of inertia of the emitter. 2. Bands λ 3064 and λ 3122. Among the earlier measurements of the water-vapour spectrum may be mentioned those of Liveing and Dewar, Meyerheim, and Grebe and Holtz. Some of the lines have been arranged into series by Meyerheim, and Deslandres and D. Azambuja, but to Heurlinger is due the credit of giving a complete arrangement into series of the lines measured by Grebe and Holtz. From these measurements Heurlinger picked out two bands with heads at λ 3064 and λ 3122. With the exception of a group of closely packed lines probably forming the head of the band 3122, for which region complete details are not given by Heurlinger, all but 10 of the 260 lines measured by Grebe and Holtz are included. Fortrat succeeded in resolving some of the lines given as single by Grebe and Holtz and gave the wave-numbers of 7 lines forming the head of the band 3122. Photograph No. 1 of Plate 6 shows the appearance of the bands 3064 and 3122. The band 3122 is much weaker than the band 3064, by which it is entirely overlaid, so that it is not obvious as a separate band in the photograph. In fact, Heurlinger considered these two as forming a single band, but the work of Dieke and of the author shows that they must be treated as quite distinct.