|Flybys of Mars Express confirm the low density of Phobos with the derived value of 1.876 ± 0.02 g/cm3 (Andert et al., 2010, Witasse et al., 2013, Paetzold et al., 2013). Such a value strongly argues in favour of a Phobos formation from a disk of debris (Peale 2007) whether as a remnant of the formation of Mars (Safronov et al., 1986) or as the result of a collision between Mars and a large body (Craddock 1994, 2011; Singer 2007). Within this scenario a low density of the re-accreted material forming Phobos is expected, due to large interior porosity. Thermal emission spectra of Phobos suggest an ultramafic composition with the presence of phyllosilicates and feldspathoids in some regions (Giuranna et al., 2011). Such data would be consistent with formation of Phobos near its current location (1.4-1.7 AU) or in situ (Giuranna et al., 2011).
The difficulties in understanding the origin of Phobos arises from the fact that the 0.3-4.0 μm surface spectra taken from multiple areas of the body in more than 43 years of observations (Duxbury et al., 2013), show physical characteristics similar to low-albedo asteroids such as C-type (Masursky et al., 1972, Pang et al., 1980) or D-type (Murchie 1999, Rivkin et al., 2002, Lynch et al., 2007, Pajola et al., 2012). These data argue against an in-situ formation leading to an asteroidal capture scenario: this can be favoured by binary asteroid dissociation (Landis 2009) or by collisional capture in the Martian orbital region (Pajola et al., 2012).
Recent data (Schmedemann et al., 2014) suggest an ancient surface age for Phobos of ~ 4.3 – 3.7 Ga, dating back to a period where there was an intensification in the number of impactors in the inner Solar System (Gomes et al., 2005), supporting both the in-situ and the captured scenario.
Pajola et al. (2013) presented a mineralogical model composed of a mixture of Tagish Lake meteorite (TL) and Pyroxene Glass (PM80) to explain the surface reflectance of Phobos from 0.4 to 4.0 μm. Starting from the reasonable fit between the proposed model and Phobos spectra, we adopted the weighted TL and PM80 densities to investigate if low bulk density of Phobos could be matched by these components reconciling both inner properties and surface spectra. TL density is available from measurements by Hildebrand et al. (2006), but the density of PM80 (Jager et al., 1994) has not been measured. In order to overcome the lack of density data for the above mentioned pyroxene glass, we have considered density values of different pyroxene glasses from the literature (Karamanov and Pelino, 1999, and Smithsonian Physical Tables 1921) and the density of mafic-rich glasses with VNIR spectra similar to PM80 (Carli et al., 2014). The obtained results will be presented.