STEM characterization of the synthesis of metallic core/shell nanoparticles using ALD MA Verheijen 1,2 , MJ Weber 1 , A. Mackus 1 , C van der Marel 2 and WMM Kessels 1 1. Plasma and Materials Processing, Department of Applied Physics, Eindhoven (The Netherlands) 2. Philips Innovation Services, High Tech Campus, Eindhoven, the Netherlands email.m.a.verheijen@tue.nl Keywords: Atomic Layer Deposition, Nanoparticles, Core/shell Noble metal nanoparticles (NPs) are known for being excellent materials for heterogeneous catalysis in applications such as hydrogen storage, sensing, automotive emissions catalytic conversion and fuel cells. Catalytic activity can be maximized when the size, structure and composition of these metallic NPs are optimized for a specific catalytic reaction. In this light, structured bimetallic core/shell NPs have attracted tremendous interest in the last few years. The lattice strain created in these core/shell NPs, as well as the hetero-metallic bonding interactions, modify the surface electronic properties of the NPs, enhancing their catalytic performance. Therefore, core/shell NPs show improved catalytic properties compared to their alloyed counterparts or to mixtures of monometallic NPs. Recently, Atomic Layer Deposition (ALD) has been used for the synthesis of supported metallic NPs. [1] Many metals have the tendency to form highly dispersed nanoparticles on oxide substrates during the initial stage of the ALD process, [2] because the interactions between the deposited metal atoms are stronger than those between the metal adatoms and the oxide substrate. This so-called Volmer-Weber (island) growth is therefore a natural and straightforward way to deposit metallic NPs. By tuning the ALD deposition parameters, such as partial gas pressures, temperature, process time and number of ALD cycles, the particle size and distribution can be tuned. HR-STEM studies of Pd and Pt single-metal nano-particles not only display a narrow size distribution (figure 1b-c), but also a significant density of single metal atoms on the oxide surface (figure 2a), providing evidence for the importance of surface diffusion during the ALD process. To deposit a Pt shell on Pd NPs, the Pt ALD process was tuned such that only selective growth on Pd NPs was obtained. The selective ALD process was based on Pt ALD from MeCpPtMe 3 and O 2 gas at 300°C, as developed by Aaltonen et al. and Knoops et al . [3,4] It has been shown that selective ALD growth of Pt can be obtained when using a low O 2 partial pressure during the O 2 pulse. For an O 2 pressure of ~7.5 mTorr, no Pt growth could be observed on Al 2 O 3 , even after 600 cycles. On the other hand, immediate growth occurred on a Pd substrate (figure 1d). In order to synthesize the supported Pd/Pt NPs, 50 cycles of Pt ALD were applied to selectively cover catalytic Pd cores. No vacuum break was applied between the two ALD processes. The resulting core/shell particles were characterized using TEM as shown in figure 1e. The diameter of these Pd/Pt NPs was 4.1±1.1nm. Compared to Pd NPs, a shift in the distribution of NPs towards larger particles and broader size dispersion was obtained after the deposition of the Pt shell as can be seen in Figure 1f. The density of the Pd/Pt NPs was 7.9×10 11 NPs/cm 2 which is virtually equal to the density of the Pd NPs. Furthermore, no particles with a diameter smaller than 1.5 nm can be distinguished. This suggests that the Pt ALD process only takes place on the Pd cores forming shells without creating new monometallic Pt NPs. EDX and EELS line scans provided evidence for the core/shell nature of the NPs. X-ray Photo-electron Spectroscopy (XPS) on the same samples showed that the particles are in their metallic state. This implies that the Pd/Pt NPs were not oxidized during processing, despite the use of O 2 in the Pt ALD process. Modeling of the XPS data assuming a semi-sphere geometry yielded a Pt:Pd ratio that is in good agreement with the TEM data. References