Extended abstract of a paper presented at Microscopy and Microanalysis 2008 in Albuquerque, New Mexico, USA, August 3 – August 7, 2008

Zeolite supported noble metal catalysts provide simultaneously both hydrogenation active sites and acid sites for hydrocracking reaction, which have been widely used in petroleum hydroprocessing processes. The size, location and distribution of the metal particles, as well as the pore structure and crystallinity of the zeolite support are the key factors affecting the activity and selectivity of the catalyst. To enhance the activity and durability of the catalyst, it is of great importance to obtain highly dispersed noble metal clusters within zeolite support. It is of particular importance to introduce small metal particles into the confines of the zeolite micropores. While the conventional ion exchange method works only for inserting metal ion precursors into larger pore zeolite such as zeolite Y, it is difficult to add metal precursors into the cavity of zeolite A due to the small pore size of NaA. So far, much effort has been devoted to introducing Pt into zeolite A by adding a platinum precursor to the synthesis mixture containing both Al and Si source [1,2]. Although preliminary results of chemisorption and shape-selective catalytic test have indicated that the platinum clusters located inside the larger cavity (α cage) of zeolite A [1,3-5], the direct information on the location and distribution of Pt particles, and the microstructure of zeolite support after loading is still in shortage. Transmission electron microscopy (TEM) can provide insight into the structure (atomic/nanoscopic) , crystallography and chemical composition of solid catalysts. Our SEM and Xray diffraction (XRD) study has shown that both morphology and crystallinity of the synthesized zeolite changed when a platinum precursor was incorporated into the initial synthesis mixture (see Fig. 1) [5]. Here, we present the experimental results of the Pt nanoparticles and the microstructure of the aforementioned samples in TEM (JEOL 2200FS). To obtain ultrathin sections (~40 nm) of the zeolite catalysts, samples were first embedded in epoxy resin, and then microtomed and collected onto a carbon coated Cu-grid. Figure 2 presents the high angle annular dark field (HAADF) STEM image for the high Pt loaded sample, showing that very small (less than 2 nm) Pt particles are evenly distributed within the support. The diffraction rings can be attributed to Face Centered Cubic (FCC) platinum, and no obvious reflections from zeolite were observed. Figure 3 shows the STEM image taken from 0.69 wt% Pt loaded sample. Selected area electron diffraction (SAED) patterns recorded from different particles indicate that the large particles (μm scale in Fig. 1) are crystalline while the smaller ones with rough surfaces (see inset of Fig. 1b) are amorphous. The sharp diffraction rings originate from Pt. Electron Energy Loss Spectra (EELS) were also taken for local structural and chemical analysis. Figure 4 shows a comparison of EELS spectra of O-K edge between NaA and high Pt loaded sample after background subtraction and Fourier-ratio deconvolution [6]. The dramatic...

Extended abstract of a paper presented at Microscopy and Microanalysis 2008 in Albuquerque, New Mexico, USA, August 3 – August 7, 2008

There is much interest in fabricating metallic contacts on semiconductor surfaces by a straightforward class of electrochemical reactions called galvanic displacement [1,2]. The chemistry is carried out via contact of the desired semiconductor with the metal ion solution (generally aqueous); the semiconductor acts as the source of electrons, reducing the metal ions in situ to metal. In the case of gold deposition on silicon, we have evidence that the growth of Au nanoparticles on silicon is epitaxial [3]. In order to better understand the nature of these interfaces, we carried out a series of plan-view and cross-section transmission electron microscopy (TEM) studies. TEM samples were prepared by mechanical grinding followed with ion milling at cryogenic temperatures (below -100°C). TEM images and diffraction patterns were recorded using a 200kV JEOL 2200FS TEM/STEM instrument. In order to complement the work described above on flat silicon, we have also investigated Au nanoparticles deposited on Si nanowires. Au particles have a strong tendency to grow on certain facets, such as (110) for particles on a <112> nanowire; this is demonstrated in Fig. 4. By using a nano beam probe, TEM images and electron diffraction patterns were acquired from the interface area, allowing us to measure relative orientation of Si and Au lattices. While straightforward to carry out, the underlying mechanism of galvanic deposition is complicated and thus the reason for the selectivity is not yet known at this time [2]; nevertheless, the preferential growth may be dictated by the interfacial interaction of the Au/Si epilayers. More control experiments and theoretical calculations are under way to further investigate this issue [6].

We discuss several applications of Fourier-ratio deconvolution [1,2,3] to the low-loss region of the electron energy-loss spectrum and demonstrate its advantages over simple subtraction procedures.1. If PS is the spectrum recorded from a particle on a substrate (or embedded in a matrix) and S is the spectrum recorded (under identical conditions) from nearby bare substrate of uniform thickness, the spectrum P of the particle alone is given by Z*PS = P*S, where Z is the zero-loss peak recorded without a specimen and * represents convolution. Therefore P can be obtained by Fourier-ratio deconvolution with PS as the data, S as the kernel (deconvolving function) and Z or its Gaussian equivalent as a reconvolution function. This procedure assumes that the particle and substrate are thick enough that surface-mode losses can be ignored. An example is chosen in Fig.1 below. 2. If uniform-thickness films of two different materials (A and B, giving spectra SA and SB) are mounted together in the TEM but with a small gap between, the resulting spectrum S(A+B) is given by Z*S(A+B) = SA*SB even if surface modes are important. This allows an experimental test of the principles involved; compare the brown and green curves in Fig.2. 3. If material A is directly deposited on material B, the spectrum S(AB) of the double-layer film is given by Z*S(AB) = SA*SB if surface modes are negligible. If not, comparison of S(AB) with S(A+B) yields information about the free-surface and interface modes.4. If S1 is the spectrum of a film of thickness t 1 and S2 the spectrum of the same material with smaller thickness t 2 , S1 = S2*B where B is a bulk-only spectrum corresponding to a film thickness t 1 -t 2 with surface modes absent. The surface-only spectrum can be obtained by deconvolving or subtracting B from S1; see Fig.3. The procedure assumes that t 2 is large enough (> 10 nm) that the plasmon modes on opposite surfaces do not couple significantly.

Extended abstract of a paper presented at Microscopy and Microanalysis 2008 in Albuquerque, New Mexico, USA, August 3 – August 7, 2008