﻿116 Penfield and Harper — Washing of Aluminum. 



proportion the loss which would result from HOT grams BeO 

 the quantity used in our first experiments we find 



•4270 : -1107 = -0108 : x. x = -0027 



that is, in the quantity of BeO which we used in making our 

 experiments, the BeO was too heavy by about *0027 grams, rep- 

 resenting probably some soda compound which could not be 

 removed by washing. The excess in the A1 2 3 is due to the 

 BeO which was not precipitated by boiling. If we take the 

 sum of the A1 2 3 and BeO in the average of our determina- 

 tions -2133 and deduct from it *0027 we have -2106 against 

 •2109, the sum of A1 2 3 and BeO in our standard. The follow- 

 ing determinations made from our standard solutions, the boiling 

 of the soda solutions being done in platinum, confirm us in 

 this. First weight of BeO *1087 after solution and reprecipita- 

 tion 1068, weight of A1 2 3 1011. -1068 + 1041 = '2109, the 

 weight of mixed oxides in our standard. We conclude there- 

 fore that the separation is not as perfect as our results would 

 indicate. The error is however very constant, the amount of 

 BeO which is not precipitated being almost exactly equal to 

 the amount of soda which it seems impossible to remove from 

 the precipitate by washing. 



When phosphoric acid is present alumina may be separated 

 from beryllium by boiling the solution of the mixed chlorides 

 with barium hydroxide. The alumina goes readily into solu- 

 tion while the precipitate containing the barium phosphate and 

 beryllium is easy to filter and wash. After dissolving the pre- 

 cipitate and separating the barium with sulphuric acid, a beryl- 

 lium phosphate can be precipitated with ammonia. After 

 weighing this the P 2 5 may be determined by means of am- 

 monium molybdate and the BeO by difference or the precipi- 

 tate may be fused with sodium carbonate and the fusion soaked 

 out with water which gives almost a complete separation of 

 phosphoric acid from beryllium. If phosphoric acid is also to 

 be determined it must be borne in mind that very perceptible 

 quantities of it will be found in the barium sulphate precipi- 

 tate. 



We have also observed that when A1 2 3 is ignited over a 

 ring burner till a constant weight is obtained, then over a blast 

 lamp there is a slight loss, amounting in the quantities which 

 we have used in making the above experiments to about *0010 

 grams. What this loss represents we are not able to say. 

 Usually in practical work this error can not be avoided, because 

 it is necessary to dissolve the ignited A1 2 3 to test for traces of 

 Si0 2 or iron, and it is almost impossible to dissolve A1 2 3 which 

 has been strongly ignited over a blast lamp. The same is true 

 for ignited BeO. After strong ignition over a ring burner 



