Icosahedral, decahedral, fcc, and defect‐fcc structural models for Ar_N clusters, N≳500: How plausible are they?

Different structural models for solid argon clusters with more than 500 atoms, that have been proposed previously on the basis of energy considerations, but whose origin and growth histories are not well understood, are compared with other models that have been devised mainly to overcome these difficulties, at the expense, however, of a smaller binding energy. Diffraction functions have been calculated for a variety of structures and sizes, and are compared with the observed electron‐diffraction patterns obtained by Farges et al. [Adv. Chem. Phys. 70, 45 (1988)]. Since the nonsplitting of the first peak in these patterns up to N∼3000 suggests the absence of the— energetically most favorable— decahedra, and other features reflect an increasing fraction of clusters with— at that size unfavorable— bulklike fcc structure, it seems appropriate to relax the energy criterion and to consider other properties as well, notably the rate of growth. Defect‐fcc clusters, obtained from perfect fcc crystallites by introducing at least two crossing stacking faults, derive this property from a modified surface structure, involving immobile, nonvanishing, stacking fault resisting steps. Calculated diffraction functions compare favorably with observed patterns for N∼600 and N∼3000. A growth mechanism is proposed which can explain the disappearance of multishell icosahedra and the appearance of (defect‐) fcc clusters at N∼500, with no structural transition involved.