302 SOSMAN AND HOSTETTERI REDUCTION OF IRON OXIDES 



It also reacts with magnetite in the same way at 1600° and at 

 the usual atmospheric pressure of oxygen. (Hematite is not 

 stable in air at 1600°, but goes over into magnetite). On the 

 other hand, it is well known to analysts that platinum crucibles 

 in which Fe203 is ignited in air for weighing in analytical pro- 

 cedures take up no such amounts of iron as we have described in 

 previous paragraphs. The reason for these differences of be- 

 havior is readily found in the phase rule and the relations of iron 

 and platinum in their alloys. 



The system contains 3 components : platinum, iron, and oxygen. 

 Iron and platinum form a continuous series of solid solutions. 

 The oxidation of the iron causes it to separate from the platinum 

 as an oxide. If we now have present the three phases: iron- 

 platinum alloy, solid oxide, and gaseous oxygen, and assume a 

 certain concentration of iron in the platinum (say 0.01 per cent), 

 and a certain temperature (say 1200°) then there must be a 

 definite oxygen pressure in equilibrium with this system. If 

 the existing pressure of oxygen is less than this equilibrium pres- 

 sure, the oxide will dissociate and metallic iron will be absorbed 

 by the platinum. If the pressure of oxygen be greater, on the 

 other hand, iron from the platinum solution will be oxidized, the 

 oxide will separate on the surface of the metal, and the concen- 

 tration of the iron in the platinum will be reduced. 



It is evident from the fact that melted magnetite is reduced by 

 platinum in air that at 1600° the oxygen pressure of the air (about 

 152 mm) is less than the oxygen pressure in equilibrium with 

 dilute solutions of iron in platinum. Similarly, at 1200°, 18 mm 

 oxygen is less than the equilibrium oxygen pressure, and at this 

 temperature and pressure iron is absorbed by platinum from all 

 oxides between re203 and FcsO,. The 152 mm oxygen, however, 

 is greater than the oxygen pressure of any but the most extremely 

 dilute solution of iron in platinum at 1200°, and at this tempera- 

 ture, which is seldom exceeded in igniting iron oxide precipitates 

 for analysis, there is no appreciable reduction of FciOa by plati- 

 num in the open air. 



These considerations explain the very common occurrence of 

 small amounts of iron in platinum, since platinum will exercise 



