Detection of water in the shocked gas associated with IC 443: Constraints on shock models

We have used the Submillimeter Wave Astronomy Satellite (SWAS) to observe the ground-state 1(10) --> 1(01) transition of ortho-H2O at 557 GHz in three of the shocked molecular clumps associated with the supernova remnant IC 443. We also observed simultaneously the 487 GHz line (3, 1 --> 3; 2) of O-2, the 492 GHz line (P-3(1) --> P-3(0)) of C I, and the 550 GHz line ( J = 5 --> 4) of (CO)-C-13. We detected the H2O, C I, and (CO)-C-13 lines toward the shocked clumps B, C, and G. In addition, ground-based observations of the J = 1 --> 0 transitions of CO and HCO+ were obtained. Assuming that the shocked gas has a temperature of 100 K and a density of 5 x 10(5) cm(-3), we derive SWAS beam-averaged ortho- H2O column densities of 3.2 x 10(13), 1,8 x 10(13), and 3.9 x 10(13) cm(-2) in clumps B, C, and G, respectively. Combining the SWAS results with our ground-based observations, we derive a relative abundance of ortho- H2O to CO in the postshock gas of between 2 x 10(-4) and 3 x 10(-3). On the basis of our results for H2O, published results of numerous atomic and molecular shock tracers, and archival Infrared Space Observatory (ISO) observations, we conclude that no single shock type can explain these observations. However, a combination of fast J-type shocks (similar to 100 km s(-1)) and slow C-type shocks (similar to 12 km s(-1)) or, more likely, slow J-type shocks (similar to 12 - 25 km s(-1)) can most naturally explain the postshock velocities and the emission seen in various atomic and molecular tracers. Such a superposition of shocks might be expected as the supernova remnant overtakes a clumpy interstellar medium. The fast J-type shocks provide a strong source of ultraviolet radiation, which photodissociates the H2O in the cooling ( T less than or equal to 300 K) gas behind the slow shocks and strongly affects the slow C-type shock structure by enhancing the fractional ionization. At these high ionization fractions, C-type shocks break down at speeds similar to 10 - 12 km s(-1), while faster flows will produce J-type shocks. Our model favors a preshock gas-phase abundance of oxygen not in CO that is depleted by a least a factor of 2, presumably as water ice on grain surfaces. Both freezeout of H2O and photodissociation of H2O in the postshock gas must be significant to explain the weak H2O emission seen by SWAS and ISO from the shocked and postshock gas.