S. D. Brotherton

Defect production and lifetime control in electron and γ‐irradiated silicon

A study of the effect of 1‐ and 12‐MeV electron and Co60 γ irradiation has been made on power p‐i‐ n diodes and Schottky barrier diodes fabricated on the same starting material. A comparison of the results from these two types of structures illustrated the influence of device processing on the type of defects formed by subsequent irradiation. Detailed electrical characterization of the defects demonstrated good consistency between certain elements of the structural nature of the defect, inferred from these measurements, and those already obtained from electron spin resonance (ESR) measurements. Lifetime measurements on the p‐i‐n diodes indicated that both the A center and the divacancy were active recombination centers. Finally, data are presented on defect and lifetime annealing.

Electrical properties of platinum in silicon

The energy levels produced by platinum in silicon have been measured in p‐ and n‐type material using deep‐level transient spectroscopy and constant‐capacitance thermal‐emission‐rate measurements. Within the sensitivity of the apparatus only two levels have been detected having thermal‐emission activation energies of EC−ET=0.231 eV and ET−EV=0.321 eV. The majority‐carrier capture cross sections of these levels have also been measured, giving σnA =7×10−15 cm2 and σpD=1.5×10−15 exp(−7.3 ×10−3/kT). Using these values together with thermal‐emission data and assumed degeneracy factors of gA=0.5 and gD=2 in the equation of detailed balance, it was concluded that both levels were fixed to the conduction band edge.

Iron and the iron‐boron complex in silicon

Iron has been diffused into p‐ and n‐type silicon containing various concentrations of carbon and oxygen. Apart from the established iron interstitial level and the iron‐boron complex, no new centers were detected involving iron complexing with either the carbon or the oxygen. The iron‐boron level was shown to dissociate by a recombination‐enhanced mechanism and a deep acceptor level of this complex was detected at Ec −0.29 eV, which must be the recombination level rather than the well‐established level at Ev +0.1 eV.

The Electrical Properties of Sulphur in Silicon

Hall effect and diode C‐V type measurements (including DLTS) have been used to study the several levels introduced by sulphur into the upper half of the silicon band gap. Where possible, the electron capture cross section of each level has also been measured. The results are interpreted in terms of two well characterized and reproducibly occurring centers sulphur I (0.18 and 0.38 eV) and sulphur II (0.32 and 0.53 eV). A third center, or variety of close centers, with energy levels of <0.1 eV were also found, but their occurrence was determined by the starting material and diffusion conditions. The centers sulphur I and II were both assumed to be double donors.The observed variations in concentration of sulphur I and sulphur II with diffusion temperature have enabled us to fit previously published results into a consistent pattern.

Electrical observation of the Au‐Fe complex in silicon

Electron paramagnetic resonance (EPR) and diode capacitance measurements, including deep level transient spectroscopy (DLTS), have been used to identify the levels in the silicon band gap associated with the Au‐Fe complex. Two deep levels at Ec−0.354 eV and Ev+0.434 eV were found with properties consistent with those of the complex; the role of gold in the formation of the complex was confirmed by the observation that during low‐temperature annealing (<350 °C) the changes in the concentration of the center were accompanied by equal and opposite changes in the concentration of the normal gold acceptor and donor levels.

Measurement of minority carrier capture cross sections and application to gold and platinum in silicon

Two techniques have been examined for the direct measurement of minority carrier capture cross sections of deep‐lying impurity states in silicon. Both involved the use of a bipolar transistor structure with which minority carrier flow through the device could be controlled. In one technique the carrier capture transient due to the flow of this minority current was directly monitored, and in the other case the current was simply used to empty centers in the minority carrier half of the band gap of majority carriers prior to employing the conventional diode short‐circuit technique. It was shown by the former technique that the capture cross sections of gold and platinum in silicon are strongly field dependent. In addition, some evidence is presented to suggest that the donor and acceptor levels of platinum in silicon may not be different charge states of the same center. Data are reported on the temperature dependence of the minority carrier cross sections of gold and platinum in silicon. Lifetime measureme...

Photocurrent deep level transient spectroscopy in silicon

Photocurrent deep level transient spectroscopy (PCDLTS) is investigated as a technique for the detection of centers in the minority half of the band gap. The photocurrent and resulting peak shapes are analyzed and compared with experimental results obtained from gold‐doped silicon diodes illuminated with different wavelength sources. It is shown that by suitable selection of the photon wavelength, reliable detection of levels in the minority half of the band gap is possible in long lifetime material. However, for centers with a large ratio of majority carrier capture cross section to minority carrier capture cross section, their detection in short lifetime material is likely to be more difficult, unless the material is thinned down and used with an appropriately matched photon source.

Measurement of concentration and photoionization cross section of indium in silicon

Variable‐temperature Hall‐effect measurements, capacitance–voltage (C–V) measurements and flameless atomic absorption have been used to assess the concentration of indium in indium‐doped silicon (Si:In) crystals. It was found that the flameless atomic absorption and C–V results were in very good agreement, but the extraction of an indium concentration from Hall‐effect data was not a straightforward exercise due to the fact that the indium did not enter the ‘‘exhaustion’’ region. However, by analyzing the Hall data using a recent determination of the hole effective mass, and by using a physically reasonable value for the Hall scattering factor, the Hall data gave indium concentrations in very good agreement with the other two techniques. The Si:In substrates were also characterized optically using infrared absorption. These results, together with the indium concentrations determined above, yielded a peak indium photoionization cross section of (1.55±0.25)×10−16 cm2 at a 4‐μm wavelength.

Annealing kinetics of the gold‐iron complex in silicon

Gold and iron are known to interact in silicon at temperatures below ∼400 °C to form gold‐iron pairs with band‐gap energy levels of Ev +0.434 eV and Ec −0.354 eV. In this work, the details of the formation and dissociation of these pairs were examined and from the equilibrium concentrations a binding energy of 1.22±0.02 eV was deduced. The activation energy of the gold‐iron pair formation process was found to be 0.8 eV; when corrections were made to extract the activation energy for the iron diffusion coefficient from the data, a value of 0.42 eV was obtained. This is substantially smaller than the values of 0.7–0.9 eV obtained from iron precipitation and iron‐boron pairing studies. It was noted during the higher‐temperature dissociation process that there was concurrent precipitation of the iron which caused the dissociation to appear to be a two‐stage process. This precipitation also resulted in substantial irreversibility of successive dissociation and pair formation anneals. Finally, it is pointed out...

Photoionization cross section of electron irradiation induced levels in silicon

Transient photocapacitance measurements have been made on 12‐MeV electron irradiated silicon diodes. Both p‐ and n‐type substrates were used in order to investigate the centers in both halves of the band gap. Photoionization cross sections were measured on two divacancy levels at EV+0.20 eV and Ec−0.23 eV, the A center (Ec−0.17 eV), and further levels at Ec−0.09 eV and EV+0.35 eV. The only center likely to be a useful counterdopant for a 3–5‐μm extrinsic silicon imaging device is the divacancy at EV+0.20 eV.