Bookmark and Share

Letter to the editor

Dear Bill,

The Arizona electron-diffraction group has recently discussed 'three substantial points' in a 'Letter to the Editor' ( [1] IUCr Newsletter 2001, 9(1):2).

1. Accuracy of methods and published results: The reliability of the experimental and theoretical raw data may indeed be better than 1%. However, the validity of the published Figs. generated therefrom, e.g. the difference electron density plots, also depends on the applied approximation models. For instance, the good quality of the QCBED raw data does not automatically prove the existence of the hypothetically deduced novel type of Cu+--Cu+ bonding, no safe evidence of which could be found so far during a decade of serious experimental and theoretical searches (including the band structure calculation of [1]).

2. Attaching the word 'orbital' to difference densities: While the difference density under consideration exhibits the sign pattern of a g-function with 12 extrema (6 positive and 6 negative ones) in a plane around the Cu, it is stated in [1] that 'it would be perverse not to mention the resemblance of this shape to a d-orbital charge density', although the latter is characterized by 4 positive extrema (and 4 zeros), not to mention the shape of 'text-book orbitals'.

3. The significance of their difference density: The WIEN band structure calculations [1] resulted in an atomic orbital population on Cu of 3dz21.6 4s0.2 4p0.1. Our Mo calculations resulted in a 3dz21.5 4s0.75 4p0.25 MULLIKEN population. We note i) the different physical normalizations of different partial charge definitions (see J. Phys. Chem. 98 (1994) 8245), and ii) the 4s/4p WIEN shorter range and MULLIKEN longer range population ratios of 2/1 and 3/1, respectively, which are not too different.

Summarizing, while most of the basic (exp. and theor.) data and their interpretations seem to agree with each other, the main discrepancy seems to concern the interpretations of the data advocated for by the Arizona group.

W.H.E. Schwarz