A clean measurement of the hydrogen retardation of the rate of solid phase epitaxy in silicon

The rate retarding effects of the impurity hydrogen on solid phase epitaxy (SPE) in silicon have yet to be completely understood. Existing measurements of this behaviour do not coincide exactly, however, several features have attained prominence. Firstly, a linear decrease in the SPE rate is detected up until a certain concentration of hydrogen. Subsequent to this point the rate remains almost constant at around half the intrinsic rate. It is conjectured that the hydrogen bonds to and passivates the defects whose agency enables the incorporation of atoms from the amorphous phase to the crystalline. This rate reduction increases until the defect population is saturated. At this point the reduction in rate ceases. Secondly, a dependence on temperature has not been consolidated, in contrast with the trends observed with the doping species. Here a method is proposed for producing a controlled concentration of hydrogen for the advancing amorphous/crystalline interface to encounter during epitaxy. A bubble layer is formed in crystalline silicon approximately 0.6μm beneath the surface through the implantation of hydrogen at 65 keV with fluences of 4 x 1016/cm2 and 3 x 1016/cm2 and annealing for 1 hour at 850 deg C in dry argon. The anneal doesn't out gas all the introduced hydrogen, leaving a remnant gas pressure in the bubbles. The hydrogen implants at the two fluences should yield two samples with different amounts of hydrogen trapped in the bubbles. A buried amorphous layer is created to encompass the bubble layer containing this residual contaminant through silicon self implantation at appropriate energies and fluences. The progress of the front interface of the buried amorphous layer is monitored by time resolved reflectivity (TRR) as SPE is effected at various temperatures