f iirnol'of .gineering or power lisiiiiiei able, too, even though their work does little to settle which of the three basic mechanisms mentioned earlier dominate the mechanism of corrosion. Although research on external corrosion of wall tubes and superheater elements disclosed many years ago that catastrophic metal loss was most likely to occur when a liquid film is formed on the metal surface, the exact mechanism by which wastage occurred has been open to question. In the case of the alkali

more »

... on trisulfates, the most objectionable materials when burning coal, good correlations have been obtained between the presence of molten trisulfates and loss of metal, but at least two conflicting opinions exist on the mode of formation of the trisulfates. When burning oil, the most objectionable materials are vanadium compounds, and at least three conflicting opinions exist on why molten vanadium compounds attack a metal surface: (1) vanadates act as an oxygen carrier; (2) vanadates distort the normal stable lattice of the metal oxide; and (3) vanadates dissolve the normally protective oxide layer. All three explanations are supported by sound experimental evidence, and it is quite possible that all three occur simultaneously in some, or even in many cases. The authors here have conducted a skillful investigation, using a procedure that may appear to be quite different from field exposure. Their justification, of course, is that their data provide information on reaction kinetics not obtainable under less wellcontrolled conditions. The eventual problem, then, comes in defining conditions well enough in the field to apply such basic kinetic data to the variable and ill-defined parameters of temperature, time of exposure, and corrodent composition existing on a section of superheater elements in a very large boiler furnace. The fact that the dissolution rate of the author's rotating-disk specimens in molten V2O5 was found to depend on the square root of the rotational speed, as predicted from mass-transfer theory, lends considerable weight to these data. Most significant, probably, is the finding that the corrosion rate is dependent so largely on the partial pressure of oxygen over the melt. The author's explanation of the role of oxygen is reasonable, as is their assumption that the influence of sodium in the liquid vanadates is to alter the diffusion rate of oxygen through the liquid melt. Their conclusions,on the rate-controlling parameters are reason-