Back to Table of Contents

42361-AC5
In-Situ 3-D Analysis of Nanoscale Catalysts

Nigel D. Browning, University of California (Davis)

One of the main aims of this program is to develop a method in scanning transmission electron microscopy (STEM) to quantify both the size and composition of nanoscale clusters in heterogeneous catalysts. The ability to quantify the size, shape and composition of individual nanostructures in the STEM is primarily determined by the size of the electron beam (probe).  The recent development of aberration correctors for STEM has vastly improved the spatial resolution possible with these instruments, taking resolution to below 0.1nm.  However, although a spherical aberration corrected STEM has exceptional spatial resolution, its intense probe can force nanoclusters to move on a support surface, making a quantitative measurement uncertain.  Additionally, the intense probe may also cause changes in the structure of the nanoclusters. Reducing the probe intensity (e.g., by changing the optical configuration of the aberration corrected microscope or by using a conventional non-aberration-corrected STEM) limits the cluster movement, but at the expense of resolution and signal intensity (the smallest clusters get lost in the noise of the image).   However, in work performed as part of this research program a new approach that permits the quantitative determination of size distributions of supported nanoclusters from conventional STEM HAADF images (low spatial resolution and signal levels) has been developed. The method described here produces accurate results even for the smallest clusters when the image has a low signal-to-noise ratio (S/N) by analyzing all the contributions to the final image intensity.
The method works by using a Gaussian blurring function to smooth out the effect of noise on the image.  By convoluting the image with a Gaussian it is possible to determine the root mean square (rms) size of the particles as a function of the width of the Gaussian.  By plotting rms size against the Gaussian width it is possible to extrapolate back to an unblurred size that is free from noise.  Initial analyses were performed on Os₁₀ clusters to test the accuracy of these measurements have found that the size distributions acquired from single STEM images (~30 particles total) agree well with those determined by extensive EXAFS measurements.  This approach to size determination does not result in electron beam modification of the size, shape and composition and can be extended to 3-D tomographic imaging.  In this measurement, one consideration that is taken to account for measuring the cluster sizes in dose tolerant samples is the application of a focal series for each image and selecting the optimum focus to increase the precision of the measurements. In this case, the effect of uneven support surface is avoided (each cluster is measured at optimum focus regardless of the curvature of the support surface).
During this year’s work, the focus of the experimental work was the analysis of Ta nanoclusters on SiO₂ supports.  The aim of the analysis is to confirm the uniformity of the cluster shape and size. In addition, using a unique vacuum transfer stage for the sample, the effect of oxidation on the cluster size distribution was determined.  As expected, the size measurement on non air-exposed sample showed the average size of supported clusters is smaller than the air-exposed sample, demonstrating the tendency for Ta to oxidize when exposed to air. Although the histogram of nanocluster sizes in the fresh Ta sample showed a broader distribution than the Os clusters in the Os/MgO system analyzed previously, there was a clear dominant size range present. Based on the fact that intensity in Z-contrast images is approximately proportional to the square of the atomic number of the elements contributing to the image (mass of elements), the contrast of a cluster in the z-contrast image gives a measure of the mass. Thus the relative mass for each cluster and the relation between the mass of each cluster and its size can be found. To find the number of Ta contributing to each cluster the Ta/SiO2 catalyst system was compared to Os/MgO catalyst as a reference. The cross-section for elastic scattering in supported Os clusters was calculated from the detector angle and the ratio between cross-section and mass was found. Applying this ratio to mass/cross-section of Ta clusters yielded the total cross-section for the total Ta atoms in a cluster.  In this way, structural models for the Ta clusters can be tested against other measurements, such as EXAFS. For the Ta system analyzed, the most prevalent cluster size and corresponding number of Ta atoms corresponds closely with the coordination number of Ta atoms determined by EXAFS. Using the STEM it is therefore possible to determine the cluster size and number of atoms/coordination (like EXAFS), and then use the images to determine the distribution of the cluster types across the support.  These results have been presented at the 2008 Microscopy and Microanalysis conference in New Mexico, NM and a publication is currently being prepared.

Back to top