Local structural heterogeneity, mixing behaviour and saturation effects in the grossular?spessartine solid solution

Abstract

Local structural heterogeneities in crystals of the binary grossular-spessartine solid solution have been analyzed using powder IR absorption spectroscopy. Wavenurnber shifts of the highest energy Si-O stretching mode in spectra collected at room temperature are consistent with variations in Si-O bond length from structural data. They show a smaller positive deviation from linearity across the Join than is seen for the grossular-pyrope and grossular-almandine binaries. The effective line widths, Deltacorr, of three selected wavenumber regions all deviate positively from linear behaviour. An empirical calibration of this excess spectroscopic property, obtained by comparison with calorimetric enthalpy of mixing data, gives an estimate for the symmetric Margules parameter of W-spec(H) = 14.4(7) kJ mol(-1) in DeltaH(mix) = W-spec(H) XGrXSp. W-spec(H) values derived on the same basis four aluminosilicate garnet solid solutions analyzed by IR spectroscopy vary with DeltaV(2), where DeltaV represents the difference in molar volume between the end members of each binary system. Measurements of lattice parameters and IR spectra were made over a range of temperatures for seven samples with different compositions. Positive excess molar volumes of mixing at low temperature (30 K) may be larger than the excess molar volumes at room temperature. The saturation temperatures of the molar volumes show no correlation with composition, however, in contrast with what had been expected on the basis of data for the grossular-pyrope binary. Saturation temperatures for spectroscopic parameters and lattice parameters of samples with compositions Gr15Sp85 and Gr60Sp40 seem to be outliers in all experiments. It is concluded that the data hint at systematic changes in saturation temperatures across the solid solution, with implications for both the excess entropy of mixing and the excess volume of mixing, but more precise data or further sample characterization are needed to prove that this composition dependence is real in garnet solid solutions.