Kinetic Monte Carlo Simulations for Recrystallization Process of Discrete Amorphous Regions in C3 H5-Molecular-Ion-Implanted Silicon Substrate Surface

The proximity gettering technique using ion-implantation-induced defects is attracting considerable attention to prevent leakage current which caused by contaminated transition metallic impurity in active regions of sensitive device such as complementary metal-oxide-semiconductor (CMOS) image sensors. Especially, defects formed by monomer carbon implantation are known to perform high gettering capability for transition metallic impurities contaminated in device active regions [1, 2]. On the other hand, we have developed C
3
H
5
-molecular-ion-implanted epitaxial silicon (Si) wafers to provide additional advantageous features for improving CMOS image sensors performance [3–9]. This novel concept proximity gettering Si wafers are manufactured by implanting C
3
H
5
molecular ions into a Czochralski (CZ)-grown Si substrate and then epitaxial Si layer growth on implanted Si substrate surface. The C
3
H
5
molecular ion projection ranges are located underneath device active regions in epitaxial Si layer and function well as powerful gettering sink for transition metallic impurities. This C
3
H
5
molecular ion projection ranges also have barrier effect against oxygen diffusing out from the CZ-grown Si substrate to the device active regions in epitaxial Si layer. In addition, in device fabrication process, hydrogen trapped in C
3
H
5
molecular ion projection ranges are diffused out and passivate dangling bonds at SiO
2
/Si interface in device structure and prevent leakage current. Moreover, in order to satisfy severe electrical characteristics of advanced CMOS image sensors, it is necessary to enhance the gettering capability for transition metallic impurities and the passivation effect own to hydrogen trapped in C
3
H
5
molecular ion projection ranges. One simple solution to enhance gettering capability and hydrogen passivation effect is high-dose implantation of C
3
H
5
molecular ions. High-dose implantation of C
3
H
5
molecular ions increases the density of defects consisting of carbon and interstitial Si atoms, which served as gettering sink [3]. Additionally, hydrogen trapped in the projection range also increase linearly with high-dose implantation condition of C
3
H
5
molecular ions [6]. However, in high-dose C
3
H
5
molecular ion implantation condition, it is reported that numerous discrete amorphous regions are formed in Si substrate surface due to overlapping of each molecular ion implantation damage [7]. Therefore, subsequent thermal annealing treatment must be applied to recrystallize discrete amorphous regions in C
3
H
5
-molecular-ion-implanted Si substrate surface prior to epitaxial Si layer growth [8, 9]. In order to achieve crystal perfection of Si substrate surface by optimizing thermal annealing treatment condition, it is necessary to clarify recrystallization behavior of C
3
H
5
-molecular-ion-implanted Si substrate surface. However, recrystallization process of discrete amorphous regions in C
3
H
5
-molecular-ion-implanted Si substrate surface is still not revealed. In addition, in terms of crystallinity of Si substrate surface after thermal annealing treatment, understanding of recrystallization dimension is crucial to anticipate which type of defect formation after recrystallization. In previous study, two-dimensional recrystallization of amorphous regions generate crystalline boundaries and corner defects are occasionally remained after three-dimensional recrystallization of amorphous regions [10–13]. Although, it is difficult to observe detailed recrystallization behavior of nanometer-scale discrete amorphous regions in Si substrate surface during several seconds thermal annealing treatment experimentally using conventional cross-sectional or plan-view transmission electron microscopy (TEM). Therefore, in this study, we investigate the recrystallization process of discrete amorphous regions in C
3
H
5
-molecular-ion-implanted Si substrate surface by Technology Computer-Aided Design (TCAD) process simulator using Kinetic Monte Carlo (KMC) approach [14]. Firstly, we calculate distribution and concentration of vacancy and interstitial Si atom induced by high-dose C
3
H
5
molecular ion implantation into Si substrate using Monte Carlo (MC) ion implantation simulations [15]. Then, from the distribution and concentration of vacancy and interstitial Si atom, amorphous regions of C
3
H
5
-molecular-ion-implanted Si substrate are obtained and compare with the cross-sectional TEM images. Furthermore, using KMC simulations, we calculate recrystallization via solid phase epitaxial growth (SPEG) from the amorphous/crystalline interface of each discrete amorphous regions in C
3
H
5
-molecula-ion-implanted Si substrate surface and compare with the plan-view TEM images when subsequent thermal annealing treatment is applied. After confirmation of correspondence between simulated models and experimental observations, we analysis the detailed recrystallization process of discrete amorphous regions in C
3
H
5
-molecular-ion-implanted Si substrate surface and discuss these mechanism.