Far-infrared investigation of class 0 sources: Line cooling

We have investigated with the Long Wavelength Spectrometer (LWS) of the Inf rared Space Observatory (ISO) the far-infrared spectra (43-197 mum) of a sa mple of 17 class 0 sources and their associated outflows. In addition to [O I] 63 mum, the pure rotational lines of abundant molecules such as CO, H2O , and OH are frequently observed in these sources, at variance with more ev olved young stellar objects. We found, in agreement with previous studies c onducted on individual sources, that the molecular line excitation arises f rom small regions, with typical sizes of 10(-9) sr, of warm (200 < T < 2000 K) and dense gas (10(4) < n(H2) < 10(7) cm(-3)), compressed after the pass age of shocks. In particular, we found slow, non-dissociative shocks as the main mechanism at the origin of the molecular gas heating, while the bulk of the [O I] 63 mum line emission is due to the dissociative J-shock compon ent arising from the Mach disk at the head of the protostellar jet, as test ified by the fact that this line emission happens to be a good tracer of th e source mass-loss rate. Large abundances of gas-phase H2O are commonly est imated, with values that appear to be correlated with the gas temperature. The total far-infrared (FIR) line cooling L-FIR = L(O I) + L(CO) + L(H2O) L(OH), which amounts to similar to 10(-2) to 10(-1) L., is roughly equal t o the outflow kinetic luminosity as estimated by means of millimeter molecu lar mapping. This circumstance demonstrates that the FIR line cooling can b e a valid direct measure of the power deposited in the outflow, not affecte d by geometrical or opacity problems like the determination of L-kin or by extinction problems like the near-infrared shocked H-2 emission. We finally remark that the strong molecular emission observed, and in particular H2O emission, is a peculiarity of the environments of class 0 sources. The pres ent analysis shows that the ratio between FIR molecular line luminosity and bolometric luminosity (L-mol/L-bol) is always larger than similar to 10(-3 ) in class 0 objects. We suggest that this parameter could be used as a fur ther criterion for identifying future class 0 candidates.