Abstract
Water on aluminium surface was liberated by stepwise heating of a sample in an IR radiation furnace, and determined by the coulometric titration method after conversion into ammonia. Adsorbed water on the oxide layer formed on aluminium surface was first removed at 373 K, and combined water in the layer was liberated between 673 and 873 K. No water liberation could be found at 473 and 573 K, but some hydrogen was generated at 573 K by the reaction between the water and aluminium. The amount of each kind of water increased appreciably by the storage of the samples in an hygrostat after their surface treatments. However, the amounts of water liberated between 673 and 873 K were nearly constant, when the samples were analysed immediately after their surface treatments. Therefore, the correction of combined water for the determination of hydrogen in the metal was partly possible, when the sample was heated by two steps. After removal of adsorbed water at 373 K, the sample was heated at 873 K for the determination of both hydrogen and water, and the amounts of 873 K water measured separately was deducted from the previous result. However the error due to the reaction between water and the metal found at 573 K cannot be corrected irrespective of the method of hydrogen determination, though it may be reduced by taking a large amount of the sample.
