Iron detection in polished and epitaxial silicon wafers using generation lifetime measurements

Abstract

For iron detection in silicon, minority carrier diffusion length measurements are often used. These methods analyze only the bulk properties of the wafer and therefore failed in the case of epitaxial wafers due to the disturbing influence of the highly doped substrate. A technique is presented for the determination of the iron content in epitaxial as well as in polished wafers using generation lifetime measurements on metal-oxide-silicon capacitors. The influence of interstitial iron and iron-boron pairs on generation lifetime was investigated for samples diffused with iron in boron-doped silicon. The difference in reciprocal generation lifetime before and after a thermal anneal of 200 °C for three minutes was calibrated using deep level transient spectroscopy to detect the iron-boron pair concentration. For a typical metal-oxide-silicon capacitor fabrication process this means a detectability limit for iron-boron concentrations of about 8×1010 at/cm3. The advantage of this method is that it detects iron in a surface near region in contrast to the often used diffusion length measurements. This gives us the opportunity of controlling the iron content in epitaxial wafers as well as in wafers after an internal gettering process. In addition, the iron-boron pairing reaction was investigated in the temperature range between 21 and 50 °C. The activation energy for this reaction was determined to be 0.70 eV.

The implementation of temperature control to an inductive‐coil photoconductance instrument for the range of 0–230°C

“…9,19 Table I presents the results, showing that the best-fit value of N t agrees well with N t determined from the dose and the cross-over point at room temperature. 20,21 The results of these experiments support the accuracy of the system. While this temperaturedependent analysis of t eff has been conducted at a single carrier concentration, the system is, of course, also compatible with injection-dependent analyses.…”

Formation rates of iron-acceptor pairs in crystalline silicon

“…Furthermore, we have only shown data for which the methodology used would be expected to generate reliable results, hence excluding studies based on mid-or high-injection techniques that did not account for changing excess carrier densities. The techniques used include SPV, 8 deep-level transient spectroscopy, 6,7 generation lifetime measurements, 11 dark conductivity, 5 and internal quantum efficiency measurements of solar cells. 10 All of these techniques ensure low-injection conditions prevail, at least in the probed region of the samples.…”

“…This can be achieved by measuring the characteristic time constant of Fe-acceptor pair formation after dissociation by either illumination or thermal activation. 2 Such association times have been measured for FeB pairs in numerous studies, [5][6][7][8][9][10][11][12] the most widely used results being those of Zoth and Bergholz, 2 who determined an expression for the FeB pair association time constant as a function of boron concentration and temperature based on a model of Reiss et al 13 If the pair-formation process is limited by diffusion of Fe i , as is suggested by measurements of the activation energy for pair formation, the same expression should be valid for other p-type dopant species also. However, this has not been confirmed experimentally.…”

Effect of Ge doping on the kinetics of iron–boron pair association and dissociation in photovoltaic silicon

“…It is known that the value of N FeB is an exponential function of time, and therefore the association rate (R Ass ) of FeB pairs can be derived from N FeB ¼ N tot ð1ÀexpðÀR Ass tÞÞ [22], where N tot is the total concentration of Fe, including the interstitial Fe and FeB pairs. Fig.…”