THE RUSTING OF IRON AND STEEL. 
257 
salts on the formation of rust. The action of a large number of 
soluble salts not directly attacked by metallic iron has been examined. 
Some were found to accelerate or to produce no effect on rusting. 
Among salts of this class may be mentioned, sodium chloride, sodium 
sulphate, ammonium sulphate, magnesium chloride, magnesium 
sulphate, potassium chlorate, calcium chloride, and calcium sulphate. 
Certain salts, however, either greatly retarded or altogether prevented 
rusting of iron. Among these may be named, sodium carbonate, 
sodium hydroxide, sodium phosphate, sodium borate, ammonium 
carbonate, sodium nitrite (partially), calcium hydroxide, and barium 
hydroxide. 
A consideration of the two facts that liquid water is necessary for 
rusting, and that all salts having an alkaline re-action prevent it, 
suggested the possibility of hydrogen peroxide (H 2 0 2 ) being concerned 
in the process. This is an unstable substance, acting as a powerful 
oxidising agent, which is only formed in presence of liquid water and 
is decomposed by alkalis. Experiments were next tried to ascertain 
whether all substances which decompose hydrogen peroxide, and 
would therefore stand in the way of its formation, will also prevent the 
rusting of iron. The action of the following substances was therefore 
tried, potassium bichromate, potassium chromate, potassium ferro- 
cyanide, and sodium nitrate, with the striking result that in all cases 
rusting was either entirely or nearly entirely prevented. The inference 
that the hydrogen peroxide is in some intimate way concerned in 
rusting is thus greatly strengthened, especially by the action of the 
potassium bichromate, which, as itself an oxidising agent, might be 
expected to assist the rusting of iron ; this substance is, however, 
powerfully destructive of hydrogen peroxide. In order to illustrate 
the action of these salts in promoting the decomposition of hydrogen 
peroxide, I will make one experiment by taking a dilute solution of 
hydrogen peroxide and mixing with it a solution of potassium ferro- 
cyanide when almost immediately the peroxide begins to break up into 
oxygen and water. 
The direct action of hydrogen peroxide on iron is remarkable in at 
once giving an oxyhydroxide of iron identical with rust, as you will see 
from the experiment in progress, whilst when iron is oxidised under 
other conditions, another oxide—the magnetic oxide—is usually 
produced. The rust formed from pure iron is remarkably uniform in 
composition as is seen from the analysis recorded on the diagram. 
Its composition may be expressed by the formula Fe 2 0 2 ( OR) 2 and it 
is a curious coincidence, if nothing more, that this is equivalent to 
2 Fe 0 + E 2 0 2 . 
I consider that the results of these experiments justify the conclusion 
that hydrogen peroxide is intimately concerned in the aerial oxidation 
of iron, and that substances which prevent the formation of hydrogen 
peroxide also prevent rusting. If this is true, it seems probable that 
unless a metal i^ capable of being oxidised by hydrogen peroxide it 
will not rust. 
