Although a number of cold dense cores are known where very substantial depletion, by a factor of about 10, of gas phase CO obtains (e.g. Willacy et al. 1998; Caselli et al. 1999; Kramer et al. 1999), it is possible that freeze-out is never total, as a variety of desorption processes should operate even in cold, dark regions (e.g. Williams 1993). In cold dense cores observed in Orion a few percent of the elemental carbon is retained in the gas phase in CO (Gibb & Little 1998). Sulphur is more highly depleted in many environments than other elements, but a few tenths of a percent or more of sulphur is always found in gas phase species within dense cores. Another explanation (cf. Ruffle et al. 1999) might account for some sulphur remaining in the gas phase even when it is highly depleted, but it is possible that a continuous desorption process may limit the sulphur depletion. Other studies of depletion (e.g. Bergin et al. 1997) always use some molecular tracers of the gas, implying that freeze-out on to dust is never total. We have taken the observations of Gibb & Little as a guide to depletions in very dense cores, and have assumed that freeze-out halts when a small fraction (3% in most models) of the elemental carbon is left in gas phase species; a good fraction of the gas phase carbon is in CO, but the halt in freeze-out also leaves similar fractional abundances of other more reactive species in the gas phase which can take part in shock-induced high-temperature chemistry. We explore the possibility that the passage of a shock through the core (which still contains at least some gas phase species) - either before it is warmed, or after - can be detected from the core's chemical composition.
In this section we present results for two simple models in which partial freeze-out has been assumed to take place. Model S(0)T(0)3% is for a core through which a shock passes after collapse has occurred; sputtering in the shock of all mantle material without subsequent freeze-out was assumed (so that the focus was on the shock processing of material). Model T(0)3% is of a core which is never shocked and in which instantaneous thermal evaporation of grain mantles takes place.
Figures 2 and 3 present the results for Models S(0)T(0)3% and T(0)3%. Table 1 gives results for the ratios of some abundances for each of the two models at three different times. At 5 104 years, the differences between the results for models S(0)T(0)3% and T(0)3% are fairly small. The most pronounced differences between values of measurable abundance ratios at 5 104 years for models S(0)T(0)3% and T(0)3% are for the ratios of NS to CS, SO to CS, and HCO to H2CO.
The main cause for the contrasts between the results for S(0)T(0)3%
and T(0)3% is the much higher early-time fractional abundance of
atomic oxygen in T(0)3%; O is a very reactive species which removes
many radicals but in shocks is reduced significantly in abundance by
the neutral-neutral sequence in which O and OH abstract H from H2to form H2O (Bergin et al 1998; O'neill & Williams 1999). For
example, SO/CS is about one order of magnitude smaller in the shocked
case in which the abundance of free oxygen is much reduced. This
result is to be expected: 
70% of SO is formed by the reaction
and therefore free oxygen is required to form SO from its
precursors such as HS. The reduction of the SO formation rate leaves
more sulphur available for the production of NS which is facilitated
by the fact that even at a temperature of 1000 K the reaction N with
H2 is slow.
Copyright ESO 2001