Modeling extended defect ({311} and dislocation) nucleation and evolution in silicon

Abstract

End of range ͑EOR͒ defects are the most commonly observed defects in ultrashallow junction devices. They nucleate at the amorphous-crystalline interface upon annealing after amorphization due to ion implantation. EOR defects range from small interstitial clusters of a few atoms to ͕311͖ defects and dislocation loops. They are extrinsic defects and evolve during annealing. Li and Jones ͓Appl. Phys. Lett., 73, 3748 ͑1998͔͒ showed that ͕311͖ defects are the source of the projected range dislocation loops. In this article, the same theory is applied to EOR dislocation loops to model the nucleation and evolution of the loops. The model is verified with experimental data and accurately represents the nucleation, growth, and Ostwald ripening stages of dislocation loop evolution. The density and the number of interstitials trapped by dislocation loops are compared with the experimental results for several annealing times and temperatures.

Efficient TCAD Model for the Evolution of Interstitial Clusters, {311} Defects, and Dislocation Loops in Silicon

“…However, the density of loops as predicted by the model is too low for these experiments, but is correct qualitatively. Figure 11 shows the experimental and simulation results for the defect evolution at 750°C for a Si + 1x10 15 cm -2 20 keV implant [21] and for a Si + 2x10 15 cm -2 40 keV implant [8]. In addition, figure 10 shows the improvements by the loop model to predict the interstitial retention in extended defects for longer anneals.…”

“…R DL is the average radius of the dislocation loops [8] defined by: R DL is the average radius of the dislocation loops [8] defined by:…”

Physically based modeling of dislocation loops in ion implantation processing in silicon

“…There are also some models to simulate the growth and ripening of DLs [11] without considering {3 1 1}-defects. Concerning the key point of the transformation of {3 1 1}-defects into DLs, it has been treated in two different ways: one assuming a transformation rate that does not depend on defect size [12] and another considering a fixed threshold size for a {3 1 1}-defect to unfault into DL [9]. The first one has proved to properly describe the nucleation and evolution of DL in amorphizing implants [12] but, as no size-dependence is considered, it would incorrectly predict DL formation also for low-dose implants, in contradiction with experiments [1,3].…”

“…Concerning the key point of the transformation of {3 1 1}-defects into DLs, it has been treated in two different ways: one assuming a transformation rate that does not depend on defect size [12] and another considering a fixed threshold size for a {3 1 1}-defect to unfault into DL [9]. The first one has proved to properly describe the nucleation and evolution of DL in amorphizing implants [12] but, as no size-dependence is considered, it would incorrectly predict DL formation also for low-dose implants, in contradiction with experiments [1,3]. In contrast, the second one can be used for both low and high doses, elucidating whether or not DLs nucleate and correctly describing the overall evolution of the {3 1 1} and DL populations [9].…”

Ultrafast Generation of Unconventional{001}Loops in Si

“…Particularly in Si, there has been substantial experimental [2][3][4][5][6][7] and theoretical [8][9][10][11][12][13][14] work aimed at unraveling both the structure and properties of defects and their interaction with dopants. It is now well accepted that ion implantation generates a large self-interstitial (I) supersaturation, and that Is tend to aggregate in defect clusters that follow an Ostwald ripening (OR) process driven by the reduction of defect formation energies [8].…”