Paper MS-ThM7
Structural Fingerprinting of Nanocrystals on the Basis of High Resolution Transmission Electron Microscopy and Open-Access Databases

Thursday, October 18, 2007, 10:00 am, Room 615

It is well known that many nanocrystals can not be identified from their powder X-ray diffraction pattern as it is customary for micrometer sized crystals. When nanocrystals are involved, i.e. the kinematic scattering approximation is sufficiently well satisfied for fast electrons; a new strategy for lattice-fringe fingerprinting from Fourier transforms of high-resolution phase contrast transmission electron microscopy (HRTEM) images is feasible. This strategy relies on crystal structure information that is transferred to HRTEM images in the weak-phase object approximation. Such information is contained in (i) the projected reciprocal lattice geometry, (ii) the phase angle distribution (in the imaginary part of the Fourier transform of HRTEM images), and (iii) the relative intensities at the positions of reciprocal lattice points. Systematic intensities close to zero at certain lattice points may suggest kinematical absences of diffracted beams, aiding the structural fingerprinting. Nanocrystal structure specific limitations to the application of this strategy are discussed. The first tests of this strategy have been promising and the whole procedure could be automated in the current generation of computer controlled HRTEMs. When automated, mixtures of nanocrystals could be analyzed both qualitatively and quantitatively. Since each nanocrystal would be identified individually and thousands of nanocrystals could be processed automatically, the detection limits of lattice fringe fingerprinting could readily be pushed to levels that are by far superior to those of traditional powder X-ray diffraction fingerprinting. The Crystallography Open Database (COD, http://crystallography.net) and its mainly inorganic subset (http://nanocrystallography.research.pdx.edu/CIF-searchable/cod.php) are discussed because the whole lattice-fringe fingerprinting concept is only viable if there are comprehensive databases to support the identification of unknown nanocrystals. While the COD contains the atomic coordinates, space group, lattice parameters, and other crystallographic information for more than 50,000 compounds, we provide at the research servers of Portland State University in addition interactive 3D structure visualizations and theoretical 2D lattice-fringe fingerprint plots for approximately 10,000 compounds for the mainly inorganic subset of the COD.