Nickel precipitation in large-diameter Czochralski silicon

Influence of nickel precipitation on the formation of denuded zone in Czochralski silicon

“…Due to the sharp dependence of the solubility on temperature, the interstitial nickel atoms, which can diffuse into silicon wafers at high temperature, tended to agglomerate and form nickel silicide during the subsequent air cooling, the reaction is presented as follows [4,6,8,15]:…”

“…It was reported that the lattice parameter of NiSi 2 is less than that of silicon by 0.4% (0.5406 nm compared with 0.5428 nm), thus the volume change associated with nickel precipitation is small. Compared with the formation of copper silicide in silicon, it is reasonable to deduce that the stress generated around the NiSi 2 will not impose significant influence on the nickel precipitation [4,15,26].…”

“…Compared with the Cu precipitation, the local stress generated in the silicon matrix is negligible due to the small lattice misfit between nickel silicide and silicon matrix. It was reported that there are two preferred reaction paths of the supersaturated transition metals in Cz Si after in-diffusion and cooling down, either precipitation in the bulk or diffuse to its surface [8,14,15,26]. Considering the high diffusivity, for the samples A1 and M1, it can be concluded that the Ni I was introduced into the silicon bulk and reach its saturation concentration after the in-diffusion annealing at 1050 • C for 16 h. During the cooling in air with a cooling rate of about 30 K/s, the supersaturation level of Ni I builds up rapidly, leading to the significant increase of driving force.…”

“…On the contrary, the oxygen out-diffusion curve or vacancy depth profile formed after annealing at 1150 • C for 4 h or by RTA, at 1250 • C for 60 s, respectively, make the surface region deplete of oxygen precipitates. On the other hand, it was reported that the temperature at the surface decreased more quickly than that in the bulk during cooling, resulting in larger driving force than that in the bulk, leading to the formation of Ni-Si alloy on the surface [14,15,25]. Considering the difference of nickel solubility between silicon and Ni-Si alloy, it is obvious that the region of higher solubility acts as a sink for transition metals from the lower solubility region, leading to the gettering of another part of nickel atoms by the Ni-Si alloys on the surface.…”

“…Considering the difference of nickel solubility between silicon and Ni-Si alloy, Gay deduced that in the defect-free silicon, the behavior of nickel was mainly controlled by the segregation gettering, namely the nickel atoms is prefer to be absorbed by the Ni-Si alloy formed on the surface due to its stronger gettering ability [14][15][16]. On the other hand, oxygen is inevitably incorporated into the silicon directly from the quartz crucibles during Cz Si growth process [1,17,18].…”