392 Gooch and Phinney — Quantitative Determination 



Art. XLYIII. — The Quantitative Determination of Rubid- 

 ium by the Spectroscope; by F. A. Gooch and J. I. 

 Phinney. 



[Contributions from the Kent Chemical Laboratory of Yale College — XVIII.] 



In a recent paper issued from this laboratory* the possi- 

 bility of determining small amounts of potassium quantita- 

 tively by means of the flame spectrum was demonstrated. 

 The work which we are about to describe was performed in 

 the endeavor to see how far similar methods might be appli- 

 cable to the quantitative determination of rubidium. It was 

 shown in the former work that hollow coils of platinum wire 

 may be adjusted to hold definite amounts of liquid and that, 

 by taking care to pluuge the coil while hot into the liquid and 

 to remove it from the liquid with its axis inclined obliquely to 

 the surface, it is possible to take up constant amounts through 

 a long series of experiments. It thus becomes possible to 

 bring definite amounts of any soluble substance into a flame to 

 be viewed spectroscopically. The potassium salt best adapted 

 to spectroscopic use proved to be the chloride, and it was 

 found to be advantageous to dry the coils over a hot radiator 

 before introducing them into the flame. A large Muencke 

 burner gave the best sort of flame, an ordinary single prism 

 spectroscope provided with an adjustable slit (the width of 

 which we were able to fix by closing it upon wires of known 

 diameter) and an observing telescope moveable so that different 

 portions of the spectrum might be shut out at will, served 

 sufficiently well for the work. The coils were made of 'No. 28 

 platinum wire (0'32 mm in diameter) wound in about thirty turns 

 to a spiral l cm long by 2 mm in diameter and twisted together at 

 the ends to form a long handle. With the apparatus described 

 it was found possible, under the most favorable conditions, to 

 recognize the characteristic red line of the potassium spectrum 

 when only yoVo m 8'- °^ potassium, in the form of the chloride, 

 was brought to the flame. It appeared, furthermore, that a 

 comparison of the brightness of two spectra produced in close 

 succession could be made with all the accuracy to be antici- 

 pated in photometric measurements, and that therefore by 

 gradual dilutions and successive testings a solution of potas- 

 sium chloride of unknown strength could be brought to the 

 point of containing as much potassium to the coil-full (and so, 

 of course, to the cubic centimeter or any other chosen volume) 

 as a standard solution of known strength. By noting the final 



* Gooch and Hart ; this Journal, xlii, p. 448. 



