Steps toward interstellar silicate mineralogy. II. Study of Mg-Fe-silicate glasses of variable composition.

In this paper, two series of Mg-Fe silicate glasses prepared as laboratory analogues of interstellar and/or circumstellar silicate dust are presented: glasses with pyroxene stoichiometry following the formula Mg(x)Fe(1-x)SiO_3_ with x=0.4, 0.5, 0.6, 0.7, 0.8, 0.95, and 1.0 and glasses with olivine stoichiometry according to the formula Mg(2y)Fe(2-2y)SiO_4_ with y=0.4 and 0.5. Preparation and analytical characterization followed the lines described in Paper I (Jaeger et al. 1994). From the reflectance and ellipsometric measurements on polished surfaces of the samples which were supplemented by transmittance measurements of thin slabs in sufficiently transparent spectral regions, consistent sets of optical constants in the wavelength range from 190 nm to 500μm have been derived. The seven sets of optical data of pyroxene glasses represent a unique data base that permitted to study correlations between diagnostic spectral parameters in the range of the vibrational bands (peak positions, band widths, band strengths' ratio, depth of the trough between the bands) and the iron content in great detail. The olivine data followed the same trends found in the pyroxene data. For the first time, the influence of the iron content on the UV/VIS/NIR absorption of silicate dust analogues could be demonstrated quantitatively in an experimental way. With the new data, the astrophysical discussion started in Paper I was continued and focussed on the comparison of the olivine glass data with IRAS LRS spectra of O-rich circumstellar envelopes around Mira stars. The generally accepted fact that amorphous silicates with olivine stoichiometry can reproduce the observed bands of evolved stars was confirmed by the new data on the whole. However, in the range of the trough between the bands, the new laboratory analogues are apparently too clean in order to yield sufficiently large absorption coefficients. Possibilities to remove this discrepancy are discussed. The new olivine glasses point to the interesting consequence that the temperature for silicate dust grains at the inner edge of circumstellar envelopes must be considerably higher than the values often given in the literature. Thus, the new data would reduce the well-known discrepancy between the low dust temperatures formerly derived from the observations and the much higher condensation temperatures computed by theorists studying the dust formation process in the envelopes.