| Summary: | 博士 === 國立清華大學 === 材料科學與工程學研究所 === 85 === The solid phase epitaxial (SPE) regrowth of 5 keV, 5 x 10^15 /cm2P+ and As+ implanted (001) Si has been investigated by high resolution electron microscopy (HRTEM). Two significantly different SPE regrowth stages were found. In P+ implanted samples, the lower activation energy of 1.45 ±0.15 eV obtained for the initial SPE regrowth is attributed to the presence of the stress induced by a high density of excess interstitials near the original a/c interface. The activation energy of 2.5 ±0.4 eV, obtained in P+ implanted samples in the later stage, is close to those reported previously although the impurity concentrationsnd the SPE regrowth rate are much higher (0.6-4 at.% vs. 0.1-0.5 at.%), and lower (0.01-2 mn/min vs. several-tens of nm/min), respectively. In addition, the regrowth occurs much closer to the surface (15 nm vs. tens of nm)
In As+ implanted (001)Si samples, the SPE regrowth behaviors were found to be similar to that in P+ implanted samples although lower regrowth rate was observed. Average activation energy values of about 2.05 ±0.15eV and 2.5 ±0.3eV corresponding to initial and later regrowth stages, respectively, were determined. The defects with {113} habit planes and (001) tail were found to remain beneath the original a/c interface during the SPE regrowth temperature range of 400-600℃. In As+ implanted samples, smaller {113} defects than that in P+ implanted samples were observed. It is suggested that the proximity to the surface affects the stability and restrains the growth of the defects formed in the As+ implanted samples. Residual defects in 5 keV, 5x10^15/cm2 BF2+, P+, As+ and B+ implanted (001) Si samples at higher annealed temperature were also investigated. In BF+, P+ and As+ implanted samples, the residual defects were found to be mainly {113} interstitial defects and {001} defect clusters. The defects were generally distributed near the original a/c interface. In BF2+ implanted samples, in addition to those defects, fluorine bubbles were found to advance with the a/c interface front during annealing. In B+ implanted samples, the defects were found to tangle with each other. Instead of {113} defects, the major defects were Lomer-lock dislocations, 60° dislocations, twins and stacking faults.
The time and temperature for the removal of EOR damage formed by low energy implantation in As+ or P+ implanted samples were found to be much shorter and lower, respectively, than those reported previously (for medium or high energy implant). For the removal of the end-of-range defects, activation energies of about 4.5 ±0.4 eV and 5.0± 0.3eV were determined for BF2+ and P+ implanted samples, respectively. It is suggested that the free surface and fluorine bubbles (in BF2+ samples) are effective vacancy sources to accelerate the removal of EOR defects. variation in activation energy between BF2+ and P+ implanted samples is thought to relate to the difference in distance between the location of EOR defects and the free surface. For interstitial defects in P@ implanted samples, the effect of the surface as a major interstitial sink is not as critical as that in BF2+ implanted samples.
Smooth surface was observed in P+ and As+ implanted samples after the removal of defects. In contrast, in BF2+ implanted samples, rugged surface was found. The truncated surface is caused by the truncation of fluorine bubbles with the surface. The {113} defects were generated and grown in P+ and As+ samples irradiated by 400 keV electrons during the HRTEM observation. In contrast, no growth of {113} defects by irradiation was observed in BF2+ implanted samples. The fluorine bubbles were thought to attract the interstitials and suppress the generation and growth of the {113} defects. Moreover, small twins were found to remain near the surface for samples, such as As+ and BF2+ implanted samples, grown with low SPE growth rate.
|
|---|
