Heat treatment of "shell" molybdena-alumina catalyst pellets results in macroscopic redistribution of the molybdena. Redistribution is minimal at 625°C but extensive at 750°C. After 6 hours at 750", a sharp maximum is observed in the radial distribution. This is a result of opposing factors: radial diffusion inward from the initial shell, and preferential sintering of the alumina support in the shell region.Loss of molybdena by vaporization is not appreciable at 750°C but is serious at 875°C. A computer program has been developed to convert SEM-EDAX (scanning electron microscopy energydispersive analysis of x-ray) data for unpolished pellet sections to a quantitative, smooth profile of Mo concentration as a function of reduced radial position. Conventional preparation of these catalysts by impregnation of pelleted alumina with ammonium heptamolybdate is known to lead to a peripheral distribution of molybdena ("shell" catalysts). In contrast to supported metal catalysts, which sinter by growth of metal crystallites but without macroscopic redistribution of the metal, molybdena-alumina suffers gross change of the Mo distribution within the pellet and, ultimately, loss of molybdena by vaporization. Superimposed on these effects is a molybdena-catalyzed sintering of the alumina support.The changes in radial distribution of Mo in the pellet can be followed crudely by optical microscopy and more exactly by energydispersive analysis of x-rays (EDAX) with a scanning electron microscope (SEM). Previous work of this kind on other systems (especially chromia-alumina), with use of an electron probe microanalyzer (EPMA), has employed metallurgically polished sections of catalyst pellets; the radial concentration profiles deduced from the EPMA data have shown marked undulations, even after data smoothing procedures. Of interest in the present SEM-EDAX work are the questions: (1) Is it possible to achieve semiquantitative concentration profiles from SEM-EDAX data on rough (unpolished) surfaces of sectioned catalyst pellets? and (2) Do the correction procedures lead to plausible profiles without undulations? If so, it becomes possible to follow the time-temperature history of the Mo redistribution in molybdena-alumina catalysts with some degree of convenience and assurance.
CONCLUSIONS AND SIGNIFICANCEThe changes in Mo distribution within a molybdena-alumina pelleted catalyst on heat treatment in air have been followed both by optical microscopy and, in much greater detail, by a scanning electron microscope equipped with an energydispersive x-ray analyzer (SEM-EDAX). Rough, unpolished surfaces of catalyst sections can be satisfactorily analyzed if the SEM-EDAX data are subjected to proper corrections; these include corrections for peak/background (to accommodate rough surfaces), and for x-ray backscattering and electron stopping by the matrix (a, x-ray absorption (A), and fluorescence by nonanalyte elements (F). A computer program has been developed which incorporates these corrections and includes a systematic procedure for ter...