390 BELL SYSTEM TECHNICAL JOURNAL 



iron, and magnesium in zinc-base alloys have been determined by 

 this method. 



4. Routine Analysis 



In certain cases the spectrochemical method is superior to wet 

 methods for routine quantitative analysis. In such cases dependence 

 is placed on a set of solid standards whose composition is known, a set 

 of standard solutions as in the foregoing section, or a standard working 

 curve prepared by plotting the logarithm of the relative intensity of 

 impurity lines as compared to lines of the base metal against the 

 logarithm of the concentration of the impurity. In the latter case 

 recourse is had to a photometric means of determining relative line 

 intensity. The most common methods are the logarithmic sector 

 method and the densitometer method (see next section). Both 

 require considerable work to prepare the standard curve, but if many 

 samples are to be analyzed periodically a saving in time results. The 

 precision in such cases may be better than ± 5 per cent of the amount 

 determined. 



A few cases follow in which the spectrochemical method has proved 

 superior for routine uses: (1) the analysis of zinc-base alloys, where it is 

 more rapid; (2) the determination of zinc in tin-base alloys and in 

 aluminum-base alloys, where no chemical or microchemical methods 

 are available for the smaller amounts; (3) the determination of silicon 

 in beryllium and aluminum oxides, where it is much more rapid (Fig. 

 3) ; (4) the determination of tin in lead-antimony cable sheath where it 

 is more rapid; and (5) the determination of magnesium in nickel alloys 

 where it is more economical of time and sample. 



5. Quantitative Analysis with High Precision 

 Of recent years much of the published research on spectrochemical 

 analysis has been concerned with attempts to improve the precision 

 of the method. Applications of the densitometer and of the logarithmic 

 sector have achieved something in this direction. Both of these 

 devices have been used with some success in this laboratory. We have 

 also developed a rotary electrode assembly (Fig. 5) which achieves a 

 fairly steady mean position of the arc by causing it to vibrate rapidly 

 about a point on the optical axis rather than to wander in a random 

 manner over the electrode surface. The apparatus is applicable to 

 the analysis of solutions in the graphite arc and to the analysis of alloys 

 which can themselves be used as electrodes. Salts and other loose 

 powders of course cannot be rotated at high speeds (600 R.P.M. is 

 recommended) . 



The highest precision has been claimed by Duffendack and his 



