Abstract
The work is aimed at investigating the effects of charged particle bombardment on the preferential deposition of thin films so that knowledge may be gained regarding the feasibility of 'writing' interconnection patterns with programmed ion beams. Some theories of thin film nucleation are reviewed and comparisons made between two recent models that show their compatibility for the limited condition where the stable nucleus contains two atoms. Nucleation experiments involving silver on silicon oxide are described and the results compare well with theory. Values of surface energy are derived and estimates of fundamental vibration frequencies made. (E[alpha] ~ 0.3 eV E[d], ~ 0.1 eV and nu[1], and nu[c] ~ 10 sec[-1]). Critical condensation experiments are described that reinforce the value of obtained by the nucleation studies. Calculations of E[?] are made for a variety of elements on silicon oxide and their condensation energies are estimated and conclusions drawn regarding suitable condensates and dopants for preferential deposition sensitization. Development of an ion source and simple ion optical system for preferential deposition studies is described and results of ion energy spread measurements are given. The trapping probabilities of gold ions into silicon oxide are measured and compared with a theoretical model. These results and further experiments allow the minimum A[+] ion dose for inducing the preferential deposition of silver to be established over the energy range 70-350 eV. A model is proposed based on three dimensional cluster formation by extending already discussed and verified nucleation theory. The cluster formation is assumed to extend to the maximum range of the implanted ions and semi-quantitative agreement is found between this theory and experiment. Finally some experiments are described in which electrons are shown to promote preferential deposition on pure SiO[2] but not on evaporated SiO. A qualitative model based on the existence of neutral traps is proposed to account for the results. It is concluded that providing suitable ion optics can be developed the process of 'writing' interconnection patterns is feasible. Appendices deal with abortive condensation energy measurements, extension of data obtained to forecast the high energy implanted doses required for preferential deposition and the effects of ion bombardment on thin film adhesion.