Atomistic Structures of 25 000-Atom Oxide Nanoparticles Supported on an Oxide Substrate Dean C. Sayle* ,² and Graeme W. Watson ‡ Department of EnVironmental and Ordnance Systems, Cranfield UniVersity, Royal Military College of Science, ShriVenham, Swindon, U.K. SN6 8LA, and Department of Chemistry, Trinity College, Dublin 2, Ireland ReceiVed: May 30, 2002; In Final Form: July 27, 2002 A simulated amorphization and recrystallization strategy has been used to generate atomistic models of CaO and SrO nanoparticles, approximately 13 × 13 × 3 nm in size, supported on an MgO(001) substrate. The (single crystal) CaO nanoparticle exhibits a cubic “slab” morphology in contrast to the SrO nanoparticle, which has a convex-lens or “droplet” morphological appearance with a height of 3.5 nm and a diameter of about 14-15 nm. In addition, the SrO nanoparticle comprises 10 interconnecting misoriented crystallites. The epitaxial relationships that exist between the nanoparticle and substrate are identified and correlated with the lattice misfit together with the interfacial structures and complex dislocation networks that evolve within the nanoparticles. The atomistic structure of various screw-edge dislocation cores (in CaO), grain-boundaries, and grain-junctions (in SrO) that form within the nanoparticles are presented graphically. In addition, the structures of the nanoparticles are compared with previous simulations performed on the analogous CaO and SrO thin films supported on MgO(001). Introduction When an oxide thin film is supported on a lattice misfitting oxide substrate, the overlying oxide evolves various structural features to maximize the interfacial interactions, while minimiz- ing the strain associated with the incommensurate nature of the system. These structural modifications may include grain- boundaries, 1,2 dislocation arrays, 3-5 lattice slip 6 defects including vacancies, interstitials, and substitutionals, and defect clustering particularly at the interfacial region, 7 where there is likely to be a reduction in the ionic density. 8 These structural features influence considerably the chemical, physical, and mechanical properties of a material. Consequently, in designing devices, which comprise thin films, such as supported catalysts, 9 superconductors, 10 sensors 11 and recording media, 12 it is desir- able to understand the nature of these structural features and the corresponding changes in material properties they give rise to. Ultimately, one requires the ability to tailor such structural features in order that the material properties are optimized for the particular application. A further area of intense research, similar to that of supported thin films, is clusters and nanoparticles. 1,13 In particular, as the size of the material reduces to the nanometer scale, the properties change uniquely in comparison with the bulk characteristics of the parent material. 14 The origins of which may be attributed to the dimensions of the particles being comparable to the length scales of basic quanta in solids (phonon wavelengths, de Broglie wavelengths of electrons). Moreover, because most of the ions comprising the material are located at surface or near-surface regions, surface effects dominate the thermodynamics and energetics of the particle (crystal structure, morphology, reactiv- ity, etc). If one then were to support a nanoparticle on a lattice misfitting substrate, further structural modifications will arise. Here, we aim to explore the influence of the substrate on the morphological and structural features of supported oxide nano- particles, using atomistic simulation techniques. The area of nanoscience has recently enjoyed explosive growth, which can perhaps be attributed to two primary reasons: first, to exploit the remarkable properties of such materials (together with the envisaged plethora of associated applications) and, second, methods of controllable nanoparticle synthesis are now available (for example see ref 12). However, experimental elucidation of the complete three-dimensional atomistic structure of such systems, which include, for example, dislocations and grain-boundaries, is difficult. 7,15 Although there has been much effort directed at metal clusters and nanoparticles supported on an oxide substrate, 13 there is less data pertaining to oxide nanoparticles supported on oxide substrates 16 because of the problems associated with the insulating nature of many oxide materials. 7 Accordingly, in this study, atomistic simulation techniques are used to generate models of supported oxide nanoparticles. In particular, atomistic simulation has been employed in a predictive capacity to explore how structural modifications induced within the supported oxide are related to the underlying substrate and the associated lattice misfit of the system. Indeed, with the increase in speed of modern computers, all of the ions comprising the nanoparticle can be treated explicitly without resorting to continuum methods. Atomistic simulation appears at first sight an ideal approach with which to explore such features. For example, one might expect that when an oxide is placed on top of an oxide substrate all of the structural features will evolve as dynamical simulation or energy minimization directs the ions into low energy configurations. However, within crystalline oxide materials, because the energy barriers for ionic migration are high, the time scales accessible to dynamical simulation (typically nanoseconds) are insufficient to enable such structures to evolve. Similarly, for energy minimization, the simulation is likely to * To whom correspondence should be addressed. E-mail: sayle@ rmcs.cranfield.ac.uk. ² Cranfield University. ‡ Trinity College. 10793 J. Phys. Chem. B 2002, 106, 10793-10807 10.1021/jp021311u CCC: $22.00 © 2002 American Chemical Society Published on Web 09/25/2002