Nitrogen reaction with silicon: Investigation of Si undercooling and Si3N4 growth

One-dimensional model of the equiaxed grain formation in multi-crystalline silicon

“…electromagnetic levitation set-up has been developed in order to measure these thermo-physical properties [19]. It is found that the Si undercooling is a function of C concentration [20,21]. -The law used for spherical Si grain growth is a simplified approximation of the model published for the case of parabolic thermal metallic dendrites.…”

Growth velocity and grain size of multicrystalline solar cell silicon

“…The conclusion was that the nucleation of equiaxed silicon grains during growth was triggered by Si 3 N 4 particles rather than SiC [7]. It should be kept in mind, however, that the experiments by Beaudhuin et al [6,7] were levitation drop experiments without any crucible or other solid substrates present. The source of carbon leading to nucleation can be either the furnace or the silicon feedstock.…”

“…Si 3 N 4 particles can, again, form due to the dissolution and re-precipitation of nitrogen from the Si 3 N 4 coating or the furnace atmosphere [4]. Beaudhuin et al [6] showed that the nucleation undercooling decreased from 300 K for pure silicon to 1 K for silicon with a nitrogen concentration of 600 ppmw. With carbon as the main impurity, the nucleation undercooling went from 300 K, for pure silicon, to a few Kelvin with increasing impurity concentration.…”

“…The source of carbon leading to nucleation can be either the furnace or the silicon feedstock. Nucleation undercooling for silicon is strongly dependent on both the carbon and the nitrogen concentration in the melt [6,7]. It is therefore believed that compensated silicon could give rise to an increased nucleation, since the carbon concentration is higher than in the polysilicon feedstock used today.…”

Undercooling measurement and nucleation study of silicon droplet solidification

“…In bibliography there are some research works dealing with undercooling measurements in an electromagnetic levitation setup either concerning pure silicon or silicon containing a controlled percentage of impurities. Those kinds of experiments give information about the role that the presence of nucleation centers, inside the droplet, has on the final microstructure of silicon but do not take under consideration the role that the substrate (crucible or mould walls) plays on the nucleation of silicon .…”