We propose to take 5-37 micron IRS spectra of heavily obscured transition objects evolving from the Asymptotic Giant Branch (AGB) to the Planetary Nebula (PN) stage. The sample contains a complete flux-limited sample selected by IRAS colors. Their strong obscuration makes them optically invisible and some of them are not even detectable in the near infrared (< 3 micron), so that they are almost inaccesible from the ground, except at millimeter and radio wavelengths. These ``hidden post-AGB'' stars are expected to represent the higher mass fraction of former AGB stars, which were not covered by previous studies of (usually optical bright) post-AGB stars. The proposed Spitzer observations will allow: a) to determine the dominant chemistry of the circumstellar shells (O-rich vs. C-rich) for a major part of the sample b) to determine the composition of the dust in terms of their structure (amorphous vs. crystalline) c) to test the ISO result that disc-like geometries of the dust are connected with large fractions of crystalline dust, and d) to identify young infrared PN and peculiar sources with mixed chemistry. We expect that the properties of the solid-state features present in the mid-infrared spectral energy distributions will allow to order the sample into an evolutionary sequence. This will be a major step to improve our understanding of the fast developing processes occuring during the``hidden phase'' of post-AGB evolution.