Abstract

Due to a lack of z-sensitivity, standard high resolution transmission electron microscopy has been unable, so far, to determine the full 3-diniensional shape of small (less than 3 nm in diameter) metal particles supported on planar oxide carriers. The projected average particle size is often the only parameter that can be reliably deduced from standard electron micrographs, and below -1 nm even this becomes very questionable because of image interferences caused by the background structure of the support. Also, great care must be taken when interpreting TEM information because of possible damages of the sample either by electron radiation, well documented by Poppa et al. [1], or exposure to atmospheric pressure during the transfer of the sample from the preparation chamber to the microscope. Therefore, an alternative UHV-compatible technique which allows for in situ particle size estimation is most desirable. Chemisorption of a probe gas like CO followed by temperature programmed desorption is certainly the simplest method since it gives a measure of the percentage of metal exposed, E, which can be directly related to a characteristic dimension of the particle. This characteristic length is a function of particle shape. In some cases [1], the analysis of TPD peak shapes can also supply additional information about the crystallographic shape of metal particles; only in exceptional cases can similar information be obtained directly by specialized TEM techniques (profile-TEM). The usefulness of X-ray photoelectron spectroscopy for particle size determination [2] has also been investigated but the relationship between measured quantities and particle size is much more complex and still controversial.

© 1989 Optical Society of America