50 On the Preparation of the Bubidium Compounds. 



has been already mentioned, when dissolved in aqua regia, may 

 at once be used for the next precipitation. In this way, by using 

 only 30 grms. of platinum, which can be regained at the end of 

 the operation with very slight loss, more than a quarter of a 

 pound of chloride of rubidium can be obtained, containing only 

 from about 3 to 4 per cent, of the chlorides of potassium and 

 caesium. In order to separate these impurities, the 36 grms. of 

 the salt and the 30 grms. of platinum as chloride are each mixed 

 with a kilogramme of water, and the solutions mixed when boil- 

 ing. On cooling down to 40° C, a yellow sandy precipitate is 

 deposited, and this is washed by decantation with water of the 

 temperature of about 40° C. The chloride of rubidium obtained 

 from this precipitate by reduction in a current of hydrogen is 

 dissolved in water, and the chloride of potassium separated by 

 a repetition of the foregoing operation, until a small quantity of 

 the salt examined in the spectrum apparatus shows no trace of 

 the red potassium line. 



The salt likewise contains a small quantity of chloride of 

 caesium, the spectrum of which, however, is not visible until 

 the potassium has been removed. For the purpose of separating 

 the caesium, the chlorides are converted into sulphates, and the 

 sulphuric acid separated by the addition of a slight excess of 

 baryta water, the resulting hydrated oxide of rubidium being 

 evaporated to dryness with carbonate of ammonium in a silver 

 basin. The carbonate of rubidium, which is separated by pre- 

 vious nitration from a trace of carbonate of barium, is completely 

 dried, and the dry powder extracted twenty to thirty times with 

 boiling absolute alcohol. The carbonate of caesium is thus dis- 

 solved, leaving behind pure carbonate of rubidium. 



The salt is pure as soon as no trace of the caesium lines are 

 seen in the spectrum. The few grammes of chloride of caesium 

 contained in it may be obtained by the evaporation of the alcoholic 

 solution. 



It is not advisable to separate the chlorides of sodium and 

 potassium by preliminary crystallization of the original salt. 

 This is seen by the following experiment made by M. Wander : — 

 200 grms. of the salt were dissolved in water and the solution 

 evaporated by boiling; as soon as about one-third of the salt 

 had separated out in the form of crystals, the mother-liquor was 

 poured off and evaporated to dryness. The residue, completely 

 dried, weighed 32*955 grms. j 2*7975 grms. of this salt, when 

 treated with chloride of platinum, yielded 2*0465 grms. =A of 

 a mixture of the chlorides of rubidium and potassium, which, 

 precipitated with nitrate of silver, gave 2*724 grms. = B of 

 chloride of silver. Hence the salt consists of 



