No Arabic abstract
This is the second paper in a series of three in which we present an exhaustive inventory of the 49 solid state emission bands observed in a sample of 17 oxygen-rich dust shells surrounding evolved stars. Most of these emission bands are concentrated in well defined spectral regions (called complexes). We define 7 of these complexes; the 10, 18, 23, 28, 33, 40 and 60 micron complex. We derive average properties of the individual bands. Comparison with laboratory data suggests that both olivines (Mg(2x)Fe(2-2x)SiO(4)) and pyroxenes (Mg(x)Fe(1-x)SiO(3)) are present, with x close to 1, i.e. the minerals are very Mg-rich and Fe-poor. This composition is similar to that seen in disks surrounding young stars and in the solar system comet Hale-Bopp. A significant fraction of the emission bands cannot be identified with either olivines or pyroxenes. Possible other materials that may be the carriers of these unidentified bands are briefly discussed. There is a natural division into objects that show a disk-like geometry (strong crystalline silicate bands), and objects whose dust shell is characteristic of an outflow (weak crystalline silicate bands). In particular, stars with the 33.5 micron olivine band stronger than about 20 percent over continuum are invariably disk sources. Likewise, the 60 micron region is dominated by crystalline silicates in the disk sources, while it is dominated by crystalline H(2)O ice in the outflow sources. We show that the disk and outflow sources have significant differences in the shape of the emission bands. This difference must be related to the composition or grain shapes of the dust particles. The incredible richness of the crystalline silicate spectra observed by ISO allows detailed studies of the mineralogy of these dust shells, and is the origin and history of the dust.
We have carried out a quantitative trend analysis of the crystalline silicates observed in the ISO spectra of a sample of 14 stars with different evolutionary backgrounds. We have modeled the spectra using a simple dust radiative transfer model and have correlated the results with other known parameters. We confirm the abundance difference of the crystalline silicates in disk and in outflow sources, as found by Molster et al. (1999, Nature 401, 563). We found some indication that the enstatite over forsterite abundance ratio differs, it is slightly higher in the outflow sources with respect to the disk sources. It is clear that more data is required to fully test this hypothesis. We show that the 69.0 micron feature, attributed to forsterite, may be a very suitable temperature indicator. We found that the enstatite is more abundant than forsterite in almost all sources. The temperature of the enstatite grains is about equal to that of the forsterite grains in the disk sources but slightly lower in the outflow sources. Crystalline silicates are on average colder than amorphous silicates. This may be due to the difference in Fe content of both materials. Finally we find an indication that the ratio of ortho to clino enstatite, which is about 1:1 in disk sources, shifts towards ortho enstatite in the high luminosity (outflow) sources.
This is the first paper in a series of three where we present the first comprehensive inventory of solid state emission bands observed in a sample of 17 oxygen-rich circumstellar dust shells surrounding evolved stars. The data were taken with the Short and Long Wavelength Spectrographs on board of the Infrared Space Observatory (ISO) and cover the 2.4 to 195 micron wavelength range. The spectra show the presence of broad 10 and 18 micron bands that can be attributed to amorphous silicates. In addition, at least 49 narrow bands are found whose position and width indicate they can be attributed to crystalline silicates. Almost all of these bands were not known before ISO. We have measured the peak positions, widths and strengths of the individual, continuum subtracted bands. Based on these measurements, we were able to order the spectra in sequence of decreasing crystalline silicate band strength. We found that the strength of the emission bands correlates with the geometry of the circumstellar shell, as derived from direct imaging or inferred from the shape of the spectral energy distribution. This naturally divides the sample into objects that show a disk-like geometry (strong crystalline silicate bands), and objects whose dust shell is characteristic of an outflow (weak crystalline silicate bands). All stars with the 33.6 micron forsterite band stronger than 20 percent over continuum are disk sources. We define spectral regions (called complexes) where a concentration of emission bands is evident, at 10, 18, 23, 28, 33, 40 and 60 micron. We derive average shapes for these complexes and compare these to the individual band shapes of the programme stars.
We have observed the 8-13 $mu$m spectrum (R$sim$250) of the Vega-like star candidate HD145263 using Subaru/COMICS. The spectrum of HD145263 shows the broad trapezoidal silicate feature with the shoulders at 9.3 $mu$m and 11.44 $mu$m, indicating the presence of crystalline silicate grains. This detection implies that crystalline silicate may also be commonly present around Vega-like stars. The 11.44 $mu$m feature is slightly shifted to a longer wavelength compared to the usual 11.2-3 $mu$m crystalline forsterite feature detected toward Herbig Ae/Be stars and T Tauri stars. Although the peak shift due to the effects of the grain size can not be ruled out, we suggest that Fe-bearing crystalline olivine explains the observed peak wavelength fairly well. Fe-bearing silicates are commonly found in meteorites and most interplanetary dust particles, which originate from planetesimal-like asteroids. According to studies of meteorites, Fe-bearing silicate must have been formed in asteroidal planetesimals, supporting the scenario that dust grains around Vega-like stars are of planetesimal origin, if the observed 11.44 $mu$m peak is due to Fe-bearing silicates.
We present a mechanism for the crystalline silicate production associated with the formation and subsequent destruction of massive fragments in young protostellar disks. The fragments form in the embedded phase of star formation via disk fragmentation at radial distances ga 50-100 AU and anneal small amorphous grains in their interior when the gas temperature exceeds the crystallization threshold of ~ 800 K. We demonstrate that fragments that form in the early embedded phase can be destroyed before they either form solid cores or vaporize dust grains, thus releasing the processed crystalline dust into various radial distances from sub-AU to hundred-AU scales. Two possible mechanisms for the destruction of fragments are the tidal disruption and photoevaporation as fragments migrate radially inward and approach the central star and also dispersal by tidal torques exerted by spiral arms. As a result, most of the crystalline dust concentrates to the disk inner regions and spiral arms, which are the likely sites of fragment destruction.
Silicate carbon stars show the 10 micron silicate emission, despite their carbon-rich photospheres. They are considered to have circumbinary or circum-companion disks, which serve as a reservoir of oxygen-rich material shed by mass loss in the past. We present N-band spectro-interferometric observations of the silicate carbon star BM Gem using MIDI at the Very Large Telescope Interferometer (VLTI). Our aim is to probe the spatial distribution of oxygen-rich dust with high spatial resolution. BM Gem was observed with VLTI/MIDI at 44--62 m baselines using the UT2-UT3 and UT3-UT4 baseline configurations. The N-band visibilities observed for BM Gem show a steep decrease from 8 to ~10 micron and a gradual increase longward of ~10 micron, reflecting the optically thin silicate emission feature emanating from sub-micron-sized amorphous silicate grains. The differential phases obtained at baselines of ~44--46 m show significant non-zero values (~ -70 degrees) in the central part of the silicate emission feature between ~9 and 11 micron, revealing a photocenter shift and the asymmetric nature of the silicate emitting region. The observed N-band visibilities and differential phases can be fairly explained by a simple geometrical model in which the unresolved star is surrounded by a ring with azimuthal brightness modulation. The best-fit model is characterized by a broad ring (~70 mas across at 10 micron) with a bright region which is offset from the unresolved star by ~20 mas at a position angle of ~280 degrees. This model can be interpreted as a system with a circum-companion disk and is consistent with the spectroscopic signatures of an accretion disk around an unseen companion recently discovered in the violet spectrum of BM Gem.