Optical constants of non-stoichiometric amorphous magnesium-iron silicates are determined. It is demonstrated that these can well reproduce the observed mid-infrared emission spectra of evolved stars. Stoichiometric and non-stoichiometric amorphous magnesium-iron silicate films are fabricated by pulsed laser deposition. Transmittance and ellipsometry measurements are performed in the wavelength range between 2 and 200 $\mu$m and 1.7 and 33 $\mu$m, respectively. Optical constants are derived from transmittance and ellipsometric $\Psi$ and $\Delta$ spectra by means of oscillator models. The optical constants are applied in radiative transfer models for examining reproducibility of the observed spectral features of circumstellar dust shells around supergiants. The spectra of four selected supergiants are dominated by amorphous silicate dust emission in the wavelength range of 9 and 25 $\mu$m. To obtain a good fit to the observed spectra, we take into account amorphous corundum and metallic iron particles as additional dust components to account for dust emission at $\lambda<8 \mu$m and in between the two silicate features. For each of the objects, a set of model parameters (dust mass, condensation temperature) is derived by an automated optimization procedure which well reproduces the observation. Consequently, our model spectra using the new optical data find that the silicate bands at $\approx 10$ and $\approx 18 \mu$m depend on the magnesium and iron ratio in the silicate system, and that a good fit requires a significant iron content of the amorphous silicate dust component to reproduce the observed peak positions and shape of the silicate bands.

We study whether the condensation of silicate dust in Mira envelopes could be caused by cluster formation by the abundant SiO molecules. For a simplified model of the pulsational motions of matter in the the outer layers of a Mira variable which is guided by a numerical model for Mira pulsations, the equations of dust nucleation and growth are solved in the co-moving frame of a fixed mass element. It is assumed that seed particles form by clustering of SiO molecules. The calculation of the nucleation rate is based on the experimental data of Nuth and Donn (1982). The quantity of dust formed is calculated by a moment method and the calculation of radiation pressure on the dusty gas is based on a dirty silicate model. Dust nucleation occurs in the model at the upper culmination of the trajectory of a gas parcel where it stays for a considerable time at low temperatures while subsequent dust growth occurs during the descending part of the motion and continues after the next shock reversed motion. It is found that sufficient dust forms that radiation pressure exceeds gravitational pull of the stars such that the mass element is finally driven out of the star. Nucleation of dust particles by clustering of the abundant SiO molecules could be the mechanism that triggers silicate dust formation in Miras.

Clustering of the abundant SiO molecules has been discussed as a possible mechanism of seed particle formation for silicate dust in stellar outflows with an oxygen rich element mixture. Previous results indicated that condensation temperatures based on this mechanism are significant lower than what is really observed. This negative result strongly rests on experimental data on vapour pressure of SiO. New determinations show the older data to be seriously in error. Here we aim to check with improved data the possibility that SiO nucleation triggers the cosmic silicate dust formation. First we present results of our measurements of vapour pressure of solid SiO. Second, we use the improved vapour pressure data to re-calibrate existing experimental data on SiO nucleation from the literature. Third, we use the re-calibrated data on SiO nucleation in a simple model for dust-driven winds to determine the condensation temperature of silicate in stellar outflows from AGB stars. We show that onset of nucleation under circumstellar conditions commences at higher temperature than was previously found. Calculated condensation temperatures are still by about 100 K lower than observed ones, but this may be due to the greenhouse effect of silicate dust temperatures. The assumption that the onset of silicate dust formation in late-type M stars is triggered by cluster formation of SiO is compatible with dust condensation temperatures derived from IR observations.

Silicate minerals belong to the most abundant solids in space. Their formation becomes difficult at the transition from the oxygen rich chemistry of M-stars to the carbon rich chemistry of C-stars. In the intermediate type S-stars oxygen and carbon are consumed by CO and SiO molecule formation, and left-over oxygen to build the SiO4-tetrahedra of silicates becomes scarce. Then SiO molecules may directly condense into solid SiO. The IR absorption spectrum of solid SiO differs from that of silicates by the absence of Si-O-Si bending modes at 18 mum while the absorption by Si-O bond stretching modes at 10 mum is present. Such characteristics are observed in a number of S-star spectra. We suggest that this observation may be explained by formation of solid SiO as a major dust component at C/O abundance ratios close to unity. We determine the IR absorption properties of solid SiO by laboratory transmission measurements of thin SiO films produced by vapour deposition on a Si(111) wafer. From the measured spectra the dielectric function of SiO is derived. The results are used in model calculations of radiative transfer in circumstellar dust shells with solid SiO dust. Comparison of synthetic and observed spectra shows that reasonable agreement is obtained between the main spectral characteristics of emission bands due to solid SiO and an emission band centred on 10 mum, but without accompanying 18 mum band, observed in some S-stars. We propose that solid SiO is the carrier material of this 10 mum spectral feature.