One of the most important findings of the IRAS survey has been the discovery of a new class of starburst galaxies which radiate up to 95% of their energy in the far infrared. These galaxies have luminosities much larger than normal galaxies (up to 10E14 solar luminosities) and are often associated with the interaction or merging of two galaxies. A very important question is the role of these phases in the evolution of galaxies and especially in the early stages of their evolution. At a redshift of two, their emission peaks around 300 microns. The sensitivity of the proposed mission allows the detection of about 70000 galaxies.
The almost complete sky survey will thus provide information on the number evolution with redshift of the most luminous sources, such as IRAS 10214+47. For lower luminosity sources, the question of their cosmological evolution is better answered by deeper surveys of more limited areas of the sky. Such a survey can only be done by a large submillimeter observatory. The COBRAS-SAMBA survey will nevertheless provide constraints on the weaker sources counts from correlation studies of the small scale fluctuations between the highest frequency channels.
It has been stressed in section 1.2 that the mapping of cold dust with a different distribution from the warmer dust mapped by IRAS is necessary to reach the required sensitivity for the cosmological objectives. This is very valuable in itself. The sky survey will provide a systematic search for dense molecular cores in all the nearby interstellar clouds. The formation mechanism and the life-time before they are destroyed or before they form stars are still unknown. It is known today that such cores were not detectable by the IRAS survey. They are found in continuum surveys at millimeter wavelengths but the areas surveyed are very small. The systematic survey provided by the COBRAS/SAMBA mission will be very valuable for studies of low mass star formation.
Interstellar clouds have been found to exhibit structure on all observable scales and the mid infrared emission has been shown to be extremely variable using the IRAS data. This variability is associated with variation in the size distribution of the smallest particles. It is expected that the size spectrum of large dust grains will also be affected by the processes of grain coagulation, grain sputtering, and mantle-formation and evaporation; the COBRAS/SAMBA survey will also extend to longer wavelengths the IRAS survey with similar angular resolution, and will permit the study of the evolution of the grain size distribution from the smallest to the largest particles.
The lowest frequency channel will be dominated by synchrotron and free-free emission. The separation of these two components is difficult because their spectra are not very different. The combination of the COBRAS/SAMBA maps with other data such as H_alpha maps will help in this task. High sensitivity maps of the synchrotron emission at these frequencies with 0resolution could be very useful for the investigations of the distribution of magnetic fields and cosmic ray electrons in interstellar clouds.
The small scale structure of magnetic fields in molecular clouds may have a strong influence on the star formation process, and these maps will serve to elucidate this still poorly known process. All-sky maps of the free-free emission will contain information on the distribution of the ionized component of the interstellar medium; the COBRAS/SAMBA maps will allow to estimate the filling factor of this phase of the ISM, as well as to study the still unknown source of ionization.