Sergei V. Koveshnikov

Impurity gettering to secondary defects created by MeV ion implantation in silicon

Impurities in MeV-implanted and annealed silicon may be trapped at interstitial defects near the projected ion range, Rp, and also at vacancy-related defects at approximately Rp/2. We have investigated the temperature dependence of impurity trapping at these secondary defects, which were preformed by annealing at 900 °C. The binding energies of Fe, Ni, and Cu are greater at the vacancy-related defects than at extrinsic dislocation loops. During subsequent processing at temperatures up to 900 °C, the amount of these impurities trapped at Rp/2 increases with decreasing temperature while the amount trapped at Rp decreases, with most of the trapped metals located at Rp/2 in samples processed at temperatures ≲ 700 °C. However, intrinsic oxygen is trapped at both types of defects; this appears to have little effect on the trapping of metallic impurities at extrinsic dislocations, but may inhibit or completely suppress the trapping at vacancy-related defects.

Mechanism of iron gettering in MeV Si ion implanted epitaxial silicon

The gettering properties of the two damage regions, namely, a near-surface vacancy-rich region, Rp/2 and a buried layer of extended defects near the MeV Si ion projected range, Rp, have been quantitatively determined by means of secondary ion mass spectrometry and deep level transient spectroscopy techniques. To isolate the simultaneous and competitive gettering between iron and oxygen, the MeV defects were introduced into epitaxial silicon doped with boron. Czochralski Si samples were also used for comparison. Experiments on isochronal and isothermal annealing of Fe contaminated samples in conjunction with the rapid quenching procedure allowed us to establish the dominating gettering mechanisms in the two damage regions. Thus, gettering in the Rp region was driven by relaxation of supersaturated Fe, while the segregation-induced gettering mechanism was operative in the Rp/2 zone. As a result, the concentration of Fe, which was remained ungettered at elevated temperature, was well below the solid solubili...

Gettering of Cu and Ni in mega-electron-volt ion-implanted epitaxial silicon

Gettering of Cu and Ni by 2.3 MeV Si ion-implantation-induced defects has been investigated in epitaxial silicon as a function of annealing temperature, time, and cooling rate. Secondary ion mass spectrometry revealed two distinct gettering regions, the position of which correlated with the ion projected range Rp and approximately half of Rp. Gettering experiments performed on samples with low metal impurity concentration have shown that capture of Cu and Ni in the two gettering regions occurred during high-temperature annealing, indicating a segregation-induced gettering mechanism. The binding energies of Cu and Ni are higher in the shallow Rp/2 region than in the Rp region.

Iron diffusivity in silicon: Impact of charge state

The effect of iron charge state on its diffusion in p‐type Si near room temperature has been investigated by C–V and depth dependent deep‐level transient spectroscopy measurements. The migration enthalpies of both positively charged and neutral iron were found to be 0.92 and 0.56 eV, respectively. Disagreement between the theoretical predictions and experimentally observed trend in iron diffusion is briefly discussed in terms of diffusion path, elastic strain, and Coulombic interaction versus impurity charge state.

Test Methods for Measuring Bulk Copper and Nickel in Heavily Doped p-Type Silicon Wafers

The goal of this work was to optimize the existing test methods for measuring bulk copper and nickel impurities in the heavily doped p-type silicon wafers and to assess their sensitivity, recovery rate, and correlation. Three test methods were studied; low-temperature out-diffusion, polysilicon ultratrace profiling, and wafer digestion. The bulk copper and nickel recovery rates of low-temperature out-diffusion were improved by one order of magnitude in the concentration around 2 X 10 12 atoms/cm 3 . Among the three test methods, polysilicon ultratrace profiling and wafer digestion were found to be most sensitive and the method detection limit of 5 X 10 11 atoms/cm 3 or better was achieved after optimization. The optimized wafer digestion and polysilicon ultraprofiling also correlated well across six laboratories with a bias less than ±50% for both bulk copper and nickel.

Gettering issues using MeV ion implantation

Abstract MeV implantation has generated interest as a gettering option in silicon since it was established that under the appropriate conditions of irradiation fluence, substrate oxygen content, and annealing treatment, at least two distinct regions, denoted as R p and R p /2, getter impurities (M. Tamura, T. Ando, O.K., Nucl. Instrum. Methods Phys. Res. B 59/60 (1991) 572; A. Agarwal, K. Christensen, D. Venables, D.M. Maher, G.A. Rozgonyi, Appl. Phys. Lett. 69 (1996) 3899). In the present work we report on the thermal stability of Fe gettered at R p due to MeV ion implantation-induced dislocation loops in Fe contaminated CZ Si wafers, and an activation energy for the release of gettered Fe of 0.80 eV. It has also been determined that, the gettering mechanism in the R p /2 region in epitaxial Si for high MeV ion doses (≈5×10 15 Ge cm −2 ) is vacancy-related, since TEM analysis revealed 4–11-nm diameter voids. Under the annealing conditions studied, these voids were found to be more efficient at Fe capture than R p defects.

Gettering of iron in silicon-on-insulator wafers

Gettering of Fe in silicon-on-insulator material has been investigated on both the bonded and separation by implantation of oxygen (SIMOX) platforms. Reduction of electrically active iron in intentionally contaminated and annealed wafers has been measured by deep level transient spectroscopy. These data, coupled with structural characterization techniques, such as transmission electron microscopy and preferential chemical etching, provide evidence that structural postimplantation damage below the buried oxide (BOX) in SIMOX wafers is an effective site for gettering of iron with the iron gettering efficiency varying with the SIMOX processing. Gettering was not observed in bonded wafers, and the lower BOX interface did not provide any iron gettering in either bonded or SIMOX wafers.

Lateral Gettering of Fe on Bulk and Silicon‐on‐Insulator Wafers

Laterally displaced gettering sites have been studied as an alternative to traditional internal gettering and back-side gettering sites. Fe was diffused laterally and captured, first by coulombic pairing with B in p-type Si, and then by strategically placed ion implantation induced dislocation loops. This localization of Fe was tracked by both deep level transient spectroscopy and capacitance-voltage measurements. As proof of the viability of the gettering technique, laterally displaced gettering sites were formed adjacent to capacitors on various silicon-on-insulator (SOI) substrate types. Both implantation induced dislocation loops and P diffusion were used for gettering. An improvement in gate oxide integrity was observed for capacitors with lateral gettering on all SOI types studied.

Metastable electrical activity of misfit-dislocation-associated defects in Si/Si(Ge) heteroepitaxial structures: EBIC/DLTS correlation

Abstract Thermally induced metastable electrical activity with buried heteroepitaxial layers in Si/Si(Ge) has been investigated using deep-level transient spectroscopy (DLTS) and temperature-dependent electron-beam-induced current and preferential chemical etching. Examination of diodes located on regions with different misfit dislocation densities on the same wafer indicate that the concentration of the metastable electronic defects decreases with increasing dislocation density. Depth-dependent DLTS reveals two metastable deep level traps, one of which is associated with unrelaxed interfacial strain. A discussion of the possible physical sources of these two defect levels is presented.

The influence of an in-situ electric field on H{sup +} and He{sup +} implantation induced defects in silicon

The influence of in-situ electronic perturbations on defect generation during 150-keV proton implantation into biased silicon p-n junctions has been investigated. The concentration and spatial distribution of the deep traps were characterized using a modification of the double corelation deep level transient spectroscopy technique (D-DLTS). With the in-situ electric field applied, a decrease in concentration of vacancy-related, as well as H-related, traps was observed. 500 keV He{sup +} implantation was also performed to supplement the above studies and to differentiate any passivation effects due to hydrogen. A model based on the charge states of hydrogen and vacancies was used to explain the observed behaviour.