Average structure of an An₄₈ plagioclase from the Hogarth Ranges

The average structure (C1) of a volcanic plagioclase megacryst with composition An₄₈ from the Hogarth Ranges, Australia, has been determined using three-dimensional, singlecrystal neutron and X-ray diffraction data. Least-squares refinements, incorporating anisotropic thermal motion of all atoms and an extinction correction, resulted in weighted R factors (based on intensities) of 0.076 and 0.056, respectively, for the neutron and X-ray data. Very weak e reflections could be detected in long-exposure X-ray and electron diffraction photographs of this crystal, but the refined average structure is believed to be unaffected by the presence of such a weak superstructure. The ratio of the scattering power of Na to that of Ca is different for X-ray and neutron radiation, and this radiation-depen-dence of scattering power has been used to determine the distribution of Na and Ca over a split-atom M site (two sites designated M′ and M″) in this An₄₈ plagioclase. Relative peak-height ratios M′/M″, revealed in difference Fourier sections calculated from neutron and X-ray data, formed the basis for the cation-distribution analysis. As neutron and X-ray data sets were directly compared in this analysis, it was important that systematic bias between refined neutron and X-ray positional parameters could be demonstrated to be absent. In summary, with an M-site model constrained only by the electron-microprobe-determined bulk composition of the crystal, the following values were obtained for the M-site occupancies: Na_M′ = 0.29(7), Na_M′ = 0.23(7), Ca_M′ = 0.15(4), and Ca_M′ = 0.33(4). These results indicate that restrictive assumptions about M sites, on which previous plagioclase refinements have been based, are not applicable to this An₄₈ and possibly not to the entire compositional range. T-site ordering determined by (T-O) bond-length variation–t₁o = 0.51(1), t₁m ≃ t₂o ≃ t₂m ≃ 0.32(1)–is weak, as might be expected from the volcanic origin of this megacryst.