2013A&A...553A...5S

We have analysed far-infrared spectra of 32 circumstellar disks around Herbig Ae/Be and T Tauri stars obtained within the Herschel key programme Dust, Ice and Gas in Time (DIGIT). The spectra were taken with the Photodetector Array Camera and Spectrometer (PACS) on board the Herschel Space Observatory. In this paper we focus on the detection and analysis of the 69µm emission band of the crystalline silicate forsterite. This work aims at providing an overview of the 69µm forsterite bands present in the DIGIT sample. We use characteristics of the emission band (peak position and FWHM) to derive the dust temperature and to constrain the iron content of the crystalline silicates. With this information, constraints can be placed on the spatial distribution of the forsterite in the disk and the formation history of the crystalline grains. The 69µm forsterite emission feature is analysed in terms of position and shape to derive the temperature and composition of the dust by comparison to laboratory spectra of that band. The PACS spectra are combined with existing Spitzer IRS spectra and we compare the presence and strength of the 69 µm band to the forsterite bands at shorter wavelengths. A total of 32 disk sources have been observed. Out of these 32, 8 sources show a 69µm emission feature that can be attributed to forsterite. With the exception of the T Tauri star AS 205, all of the detections are for disks associated with Herbig Ae/Be stars. Most of the forsterite grains that give rise to the 69µm bands are found to be warm (∼100-200K) and iron-poor (less than ∼2% iron). AB Aur is the only source where the emission cannot be fitted with iron-free forsterite requiring approximately 3-4% of iron. Our findings support the hypothesis that the forsterite grains form through an equilibrium condensation process at high temperatures. The large width of the emission band in some sources may indicate the presence of forsterite reservoirs at different temperatures. The connection between the strength of the 69 and 33 µm bands shows that at least part of the emission in these two bands originates fom the same dust grains. We further find that any model that can explain the PACS and the Spitzer IRS observations must take the effects of a wavelength dependent optical depth into account. We find weak indications of a correlation of the detection rate of the 69 µm band with the spectral type of the host stars in our sample. However, the sample size is too small to obtain a definitive result.