CHEMICAL ANALYSIS BY THE SPECTROSCOPE Qualitative Analysis.—Spectroscopic methods provide an important aid to chemical analysis, especially when an element is present only in small proportion, or when only a small quantity of a substance is available for examination. The principles in volved in qualitative analysis are embodied in what has been stated before, and it is only necessary to add a few supplementary remarks. In the first place, it should be understood that although refined measurements extending over the greatest possible range of spectrum, and including spectra emitted under very varied conditions, are necessary for exhaustive investigations of the structure of spectra, the procedure for purposes of chemical analysis is much simpler. Such analysis can, in fact, be carried on to a considerable extent without any attempt at precision in the determination of wave-lengths by either visual or photographic methods.
For visual observations with an ordinary spectroscope, the graphical method of interpolation is sufficiently accurate for the identification of many spectrum lines. It is more convenient, how ever, to use a wave-length spectrometer, in which the scale is so constructed that wave-lengths of fair accuracy can be read off directly. Then again, familiarity with the spectra of various ele ments often enables an experienced observer to identify lines at sight for, in addition to the help given by colour, lines frequently occur in groups which can be identified in the same way that constellations are recognized among the stars.
An iron arc spectrum is photographed beside the spectrum under investigation, and estimates of wave-length, accurate enough for purposes of identification, can be made by reference to the scale on a wave-length map of the iron spectrum. If the spectrum is a complicated one, and some of the constituent elements have been identified in this way, it is convenient to take additional photographs with these elements as comparisons. Measurements and calculations of wave-lengths are then necessary only for the lines which remain unidentified. As the most sensitive lines of many of the elements lie in the ultra-violet, a quartz spectro graph is usually the best instrument for chemical applications.
It is sometimes convenient, or necessary, to analyse solutions by the spectroscopic method. If the electric arc be employed, it is generally sufficient to soak the tips of the carbon poles in the solution before striking the arc. It is usually more satisfactory, however, to use the condensed spark, in which case the lower elec trode may be formed by a bundle of fine platinum wires stand ing in the liquid and projecting slightly above its surface, while the upper electrode is a stouter piece of platinum wire.
For the analysis of refractory substances, such as silicates, the method introduced by A. de Gramont gives excellent results. The substance is first fused with four or five times its weight of sodium or lithium carbonate, and is placed in a small platinum cup which forms the lower electrode for the passage of a con densed spark. The cup is heated by a bunsen flame while the dis charge passes. In photographic work, the spectrum of the fusion material would naturally be photographed separately, so that its contribution to the final spectrum could easily be recognised.
Abridged tables of spectra, including the brightest and most sensitive lines of the various elements, have been compiled to facilitate chemical identifications. (See Bibliography.) Quantitative Analysis.—.Several of the earlier workers with the spectroscope were attracted by the possibility of discovering methods by which spectrum analysis might yield quantitative results. Such methods as were proposed, however, were too restricted in their application, or subject to too many experimental uncertainties, to be of much practical service. The advance in the theory of spectra has somewhat improved the outlook, but even now the most successful methods are of limited application and of an entirely empirical character.
The most systematic efforts to establish a procedure for quanti tative analysis are probably those made at the United States Bureau of Standards by W. F. Meggers and others (Scientific Papers, No. 444). The method is based partly on the work of de Gramont, who made a prolonged study of the so-called raies ultimes of the elements, these being the last lines to disappear as the proportion of an element in a mixture is reduced towards the vanishing point. As an example of the procedure adopted, the method of estimating a small percentage impurity of zinc in tin may be considered. A graded series of alloys is first prepared, of tin containing zinc in the proportions o.00i, o.oi, o.i, i.o, and io.o per cent.. and their spark spectra are compared under iden tical conditions. The lines of zinc will obviously increase in intensity with respect to those of tin as the quantity of zinc increases, and the correlation of the intensities with the known percentage composition of the specimens provides a scale for the analysis of similar alloys of unknown composition. Similar methods have been successfully employed in the quantitative analysis of samples of gold and platinum, and give promise of useful application in numerous problems of practical importance.