ORIENTATIONAL RELATIONSHIPS, INTERFACES AND GRAIN GROWTH IN NANOCRYSTALLINE PALLADIUM BY HIGH-RESOLUTION AND DARK-FIELD ELECTRON MICROSCOPY
W.M. Straub1, F. Phillipp2, and H.E. Schaefer1
1Universitat Stuttgart, Institut für
Theoretische und Angewandte Physik, Pfaffenwaldring 57, D-70550
Stuttgart, Germany
2Max-Planck-Institut fur
Metallforschung, Institut fur Physik, Heisenbergstr.1, D-70569
Stuttgart, Germany
The physical properties of nanocyrstalline solids are essentially determined by their small grain size and their high number of interfaces which can be characterized by the orientational relationship between adjacent nanocrystallites and the orientation of the interfaces. For cubic structures, interfaces with special properties exist between crystals near a coincidence site (S) orientational relationship. In face-centered cubic metals, low-energy interfaces are symmetrical tilt boundaries between crystallites with a S3 or S11 misorientation [1]. Therefore, intercrystalline orientational relationships are of pivotal interest for a general understanding of the physical properties of nanocrystalline solids.
By high-resolution transmission electron microscopy (HRTEM), the orientational relationship between nanocrystals simultaneously aligned along a zone axis can be determined. The atomic-resolution electron microscope JEOL ARM 1250 with a point resolution of 0.115 nm [2]allows the imaging of higher-indexed zone axes (up to <125> in fcc-Pd) and hence the determination of even general orientational relationships. To configurations of adjacent crystallites imaged under less-specific conditions an exclusion method can be applied to rule out special orientational relationships.
We report on the determination of distributions of zone-axis orientations of individual crystallites, their orientational relationships and of the grain size distributions in nanocrystalline palladium prepared by the inert-gas cluster-condensation technique [3] utilizing HRTEM and dark-field (DF) electron microscopy. The former distribution functions take into account appropriate statistical weigths derived from electron-microscopical detection probabilities calculated within a dynamical two-beam approximation. Samples were examined both in the as-prepared state and after isochronal annealing at temperatures up to Ta=773 K.
In the as-prepared state, the distribution of zone-axis orientations of individual crystals in parallel to the electron beam is consistent with a randomly oriented ensemble of nanocrystallites. Between the crystallites, both special S 3, 5, 7, 11, 13b, 21b, 39b and random orientational relationships exist. Again, within statistical accuracy, the distribution function matches that expected for randomly oriented crystals.
In all cases, between S3-misoriented nanocrystals, a low-energy coherent twin boundary is observed. In contrast, the boundaries between nanocrystals close to a S11 orientational relationship in most cases significantly deviate from a low-energy symmetrical tilt boundary. Moreover, no low-angle grain boundaries are found in the as-prepared state. From these observations we conclude that coherent S3 twins are essentially the only low-energy boundaries in nanocrystalline palladium.
Upon isochronal annealing the overall distributions of zone-axis types and orientational relationships do not change significantly. For Ta >= 473 K, after the onset of crystal growth, low-angle grain boundaries appear.
Starting from log-normal grain-size distribution functions [4] in the as-prepared state, grain-growth in nanocrystalline palladium r ~ 0.85 rth results in a few anomalously growing grains within a slowly normally-growing matrix of small crystallites with a essentially log-normal size distribution. These findings indicate strong interface pinning mechanisms, e.g. solute or Zener drag, caused by light impurity atoms or nanopores [5]. Within a model of cube-shaped crystallites with pores localized in the interfaces, the experimental results can be matched assuming nV ~ 1.3 pores per interface with a size in the few-nanometre range.