8  Concluding Remarks

The IMPETUS II code is intended to be a useful computational tool for engineers engaged in using SIMS or SNMS techniques. The software is very versatile in that any number of material systems may be attached. During the execution of IMPETUS II the software gives an on-screen display of the redistribution of material within the solid as well as the yield history. The yield-dose and yield-depth curves may be viewed on the linear or logarithmic scales. The results from the run are reproduced in a file. The yield curves produced by IMPETUS II aids the engineer in deciding whether a given sputter profile is compatible with the presumed original material structure.
In this paper the methods that are central to the IMPETUS II code have been described in detail and results given. The data that describes the material system is derived from the results of other software packages such as TRIM or it is empirical. Discrepancies between the results from SIMS or SNMS experiments could be as a result of insufficiently accurate data. For some material systems the model itself may need extending since the present model only includes the fundamental bombardment-related processes. Nevertheless, the atomic mixing model provides us with a realistic representation of the underlying process in SIMS and SNMS for many material systems under a range of experimental conditions.
This work has focussed on improving the efficiency and consistency of the computational solution of the original atomic mixing model. The replacement of the explicit finite difference method in the original IMPETUS by an implicit method is the main reason for the dramatic improvement in the execution of IMPETUS II in comparison to the original codes. Moreover, in the implicit method there is not a strong connection between fineness of the spacial mesh and the dose step length, as there is with the explicit method. This means that if the spacial mesh requires refinement in the implicit method (in the case, for example, of a thin layer in the structure), it is not necessary to also severely reduce the dose step.
There are still many challenging areas in the area of the computational modelling of the atomic mixing process. For example the original mathematical model is likely to need extending to include the effect of the chemical bonding characteristics of some species. The inverse problem to the one considered in this paper, that of recovering the original material structure from the sputter yield curves, is of great importance and would be a powerful tool for SIMS engineers. Clarifying the computational modelling, improving the methods and hence significantly decreasing run-time and having one central program serviced by material system libraries, consist an important step in making IMPETUS II an efficient and versatile code. Such improvements are also important before extending the model and tackling the inverse problem.