Extended D2CO emission:: The smoking gun of grain surface-chemistry

We present new observations of the H2CO and D2CO emission around IRAS16293-2422, a low mass protostar in the rho Ophiuchus complex. Bright H2CO and D2CO emission is detected up to 40" from the center, corresponding to a linear distance of similar to 5000 AU. The derived H2CO abundance profile has two jumps at tau less than or equal to 150 AU and r less than or equal to 700 AU, where the dust temperature reaches 100 K and 50 K respectively. The measured [D2CO]/[H2CO] abundance ratio in the envelope is between 0.03 and 0.16, an extremely high value. We demonstrate that the present new observations can only be explained if the D2CO (and H2CO) are formed during the previous cold pre-collapse phase, stored in the grain mantles, and released in the gas phase during the pre-collapse phase. We consider the two main competing theories for mantle formation, i.e. pure accretion against grain surface chemistry, and we conclude that the former theory cannot explain the present observations, whereas grain active chemistry very naturally does. We found that the mantles are evaporated because of the thermal heating of the grains by the central source and that in the outer cold envelope H2CO and D2CO molecules are embedded in CO-rich mantles which sublimate when the dust is warmer than 25 K. Finally, the present day H2CO and D2CO abundances very probably reflect the mantle composition. We argue that mantles have likely formed in an onion-like structure with the innermost ice layers more enriched in H2CO molecules and we give estimates of the CO hydrogenation efficiency across the envelope and/or in different ices.