TESTS OF SHOCK CHEMISTRY IN IC 443G

A multitransition single-dish study has been made of the molecular species CO, HCO+, HCN, CN, SiO, CS, SO, and H2CO toward the "shocked" clump IC 443G associated with the supernova remnant IC 443. Maps of each molecule show source sizes less-than-or-equal-to 40" and essentially uniform brightness. A BIMA interferometer map of the 1-0 transition of HCO' shows four small clumps, whose combined flux is only 4.5% of the single-dish flux, so that they represent only small enhancements of an otherwise smooth distribution. A BIMA map of the 86 GHz transition of SO detected no emission to a level of 0.518 Jy beam-1, indicating no clumping on size scales up to 21", the beamwidth of the single-dish observations. The observations have been analyzed by both large velocity gradient and microturbulent radiative transfer models, from which densities, temperatures, and abundances are inferred. These models used collision rates calculated in detail for every species except CN and SO, including new collision rates calculated for SiO at low temperatures (40-300 K) and for CS at higher temperatures (100-300 K) than previously existed. We find that the eight species may be divided into two groups. Group 1 species (CO, HCO', HCN, CN) arise in region 1 with density n approximately (2-8) x 1O(5) cm-3, temperature T less-than-or-equal-to 100 K, and linear extent L almost-equal-to 7 x 10(17) cm, While group 2 species (SiO, CS, SO, H2CO) arise in the hotter (T = 200-300 K), denser (n greater-than-or-equal-to 10(6) cm-3) region 2 with smaller L. While abundances of group 1 species are consistent with ion-molecule chemistry, the observed abundances of group 2 species, if they are coextensive with group 1 species, are much lower than predicted ion-molecule abundances, and in addition the observed line widths are all much larger than those of quiescent clouds where ion-molecule processes are believed to dominate. We find that dissociative (D) shock models of Neufeld & Dalgarno (1989) agree well with observed abundances of group 1 species, while nondissociative (ND) shock models (Mitchell 1984a, b; Hartquist, Oppenheimer, & Dalgarno 1980) explain group 2 abundances well. Specifically, both observations and D shock models show that abundances of group 1 species are in no cases enhanced by shocks over quiescent cloud conditions, while both observations and ND shock models show that abundances of group 2 species are all reduced below quiescent cloud (ion-molecule) abundances as a result of shocks unless they occupy a region approximately 100 times smaller than group 1 species, in which case their fractional abundances are similar to those of ion-molecule chemistry. In general, observed line widths vary inversely with derived excitation density, while centroid velocities of all species are essentially identical. This requires that the shock velocity v, in the clump varies along the path z, rising quasi-symmetrically from a small value at z = 0 (rear face of the clump) to a maximum value in region 1, then decreasing to a value consistent with a ND shock in region 2 (near face of clump). The density gradient necessary to produce the upsilon(c), function must have preceded the shock. All physical and chemical properties can be explained satisfactorily on this picture of the IC 443G clump, but it is ad hoc, and more detailed models of shocks in nonhomogeneous media are needed.