We study the production of SiO in the gas phase of molecular outflows , through the sputtering of Si–bearing material in refractory grain cores , which are taken to be olivine ; we calculate also the rotational line spectrum of the SiO .
The sputtering is driven by neutral particle impact on charged grains , in steady–state C-type shock waves , at the speed of ambipolar diffusion .
The emission of the SiO molecule is calculated by means of an LVG code .
A grid of models , with shock speeds in the range 20 < v _ { s } < 50 km s ^ { -1 } and preshock gas densities 10 ^ { 4 } < n _ { H } < 10 ^ { 6 } cm ^ { -3 } , has been generated .
We compare our results with those of an earlier study ( Schilke et al .
1997 ) .
Improvements in the treatment of the coupling between the charged grains and the neutral fluid lead to narrower shock waves and lower fractions of Si ( \mathrel { \hbox { \hbox to 0.0 pt { \hbox { \lower 4.0 pt \hbox { $ \sim$ } } } \hbox { $ < $ } } } 10 % ) being released into the gas phase .
Erosion of grain cores is significant ( \mathrel { \hbox { \hbox to 0.0 pt { \hbox { \lower 4.0 pt \hbox { $ \sim$ } } } \hbox { $ > $ } } } 1 % ) only for C-type shock speeds v _ { s } > 25 km s ^ { -1 } , given the adopted properties of olivine .
More realistic assumptions concerning the initial fractional abundance of O _ { 2 } lead to SiO formation being delayed , so that it occurs in the cool , dense postshock flow .
Good agreement is obtained with recent observations of SiO line intensities in the L1157 and L1448 molecular outflows .
The inferred temperature , opacity , and SiO column density in the emission region differ significantly from those estimated by means of LVG ‘ slab ’ models .
The fractional abundance of SiO is deduced and found to be in the range 4 \times 10 ^ { -8 } \mathrel { \hbox { \hbox to 0.0 pt { \hbox { \lower 4.0 pt \hbox { $ \sim$ } } } % \hbox { $ < $ } } } n ( { SiO } ) / n _ { H } \mathrel { \hbox { \hbox to 0.0 pt { \hbox { \lower 4 % .0 pt \hbox { $ \sim$ } } } \hbox { $ < $ } } } 3 \times 10 ^ { -7 } .
Observed line profiles are wider than predicted and imply multiple , unresolved shock regions within the beam .