Thermal stability of internal gettering of iron in silicon and its impact on optimization of gettering

Abstract

The redissolution behavior of gettered iron was studied in p-type Czochralski-grown silicon with a doping level of 2.5×1014 cm−3 and an oxide precipitate density of 5×109 cm−3. The concentrations of interstitial iron and iron–boron pairs were measured by deep level transient spectroscopy. It was found that the dependence of redissolved iron concentration on annealing time can be fitted by the function C(t)=C0[1−exp(−t/τdiss)], and the dissolution rate τdiss−1 has an Arrhenius-type temperature dependence of τdiss−1=4.01×104×exp[−(1.47±0.10) eV/kBT] s−1. Based on this empirical equation, we predict how stable the gettered iron is during different annealing sequences and discuss implications for optimization of internal gettering.

Imaging interstitial iron concentrations in boron-doped crystalline silicon using photoluminescence

“…[1][2][3][4] The concentration of interstitial iron ͓Fe i ͔ can laterally vary across a wafer 5 and along the length of an ingot 4 and can also dramatically change during cell processing due to precipitate formation or dissolution and impurity gettering. 2,[6][7][8][9][10][11] It is therefore of considerable interest to be able to map changes in the iron concentration both laterally and during processing.…”

Iron-rich particles in heavily contaminated multicrystalline silicon wafers and their response to phosphorus gettering

“…Therefore we conclude that the gettering process is dissolution limited. Note that previous work by Zhang et al [22] on dissolution of Fe particles formed at oxygen precipitates in Czochralski-grown silicon wafers found effective dissolution within 10 min at 800 • C. It is clear therefore that the dissolution energy of the Fe particles located at the grain boundaries in multicrystalline wafers is much higher than for Fe particles located at oxygen precipitates. Figure 5 also shows the result for a sample that was annealed in N 2 at 1000 • C for 100 min, but with no phosphorus gettering layer present.…”

On the mechanism of carrier scattering at oxide precipitates in Czochralski silicon

“…The cooling process of the drive-in anneal was similar to that described in Ref. [19], ensuring that the Cu atoms driven into the silicon host were substantially gettered by the existing oxygen precipitates. Taking the specimen with the drive-in anneal at 600°C for example, it was firstly cooled to 350°C in 60 min and was then idle for 50 min.…”