BURGESS AND SALE! PURITY OF PLATINUM WARE 287 



described, however, provides a delicate means of determining 

 whether or not the desired Hmitation of impurity, expressed in 

 terms of iridium content, and as measured thermoelectrically, 

 has been met by the manufacturer. For the experimental ar- 

 rangement here given, the amount of impurity is readily deter- 

 mined to 0.01 per cent, and this accuracy could be improved 

 upon if necessary. 



The method does not distinguish the various posgible impu- 

 rities from each other but is nevertheless a certain check on the 

 platinum purity.^ One of the most undesirable impurities often 

 found in commercial platinum ware is iron; this is readily de- 

 tected by ignition and subsequent washing with hot HCl and 

 applying the usual color test for Fe. Iron, if present in relatively 

 considerable quantity, will also discolor the crucible on ignition. 



It would probably be advantageous to substitute rhodium for 

 iridium in platinum crucibles when stiffness is desu'ed and when 

 exact weighings have to be taken before and after ignition, for 

 the reason that rhodium is far less volatile than iridium, and 

 somewhat less so than platinum, although it appears not yet 

 to have been proven that the more volatile metals alloyed in 

 small amounts with platinum retain their volatilitj^ independ- 

 ently of the presence of the platinum. 



The above described thermoelectric method alone will not 

 distinguish between small amounts of rhodium and iridium, but 

 a combination of the thermoelectric and loss of weight methods 

 might be devised that would be satisfactory. An electric dis- 

 charge method operated at high temperatures might possibly 

 be made sufficiently selective to differentiate between platinum 

 alloys of iridium and rhodium. There is evidently room for 

 much more work along these lines. 



' All the metals found associated with platinum, such as palladium, iridium, 

 rhodium, etc., when alloyed with platinum (up to 90 per cent only of palladiumj 

 give, at high temperatures, an E. M. F. of the same sign against pure platinum. 

 Therefore there is no ambiguity or balancing one impurity against another. 



