Zbigniew J. Radzimski

Recombination at clean and decorated misfit dislocations

The electrical activity of interfacial misfit dislocations in silicon has been examined using the electron beam induced current technique in a scanning electron microscope. Clean dislocations formed during high‐temperature Si(Ge) chemical vapor epitaxy were studied. These defects were subsequently decorated with known metallic impurities (Au and Ni) by diffusion at different temperatures from a backside evaporated layer. Differences in electrical activity are discussed in relation to the detection limits of electron beam induced current technique and energy levels anticipated for the clean or decorated dislocations.

Bulk and surface components of recombination lifetime based on a two‐laser microwave reflection technique

An algorithm for separating the bulk and surface components of recombination lifetime, tailored for contactless measurement techniques with laser excitation, is presented in the paper. In order to analyze the carrier decays and subtract the surface recombination term, two lasers operating at 910 and 830 nm are applied. A separation of carrier decay resulting from the different contribution of surface and bulk components due to difference in the light absorption is observed for such a case. This separation is a function of surface recombination velocity S. An experimental verification of the analysis is presented using microwave absorption/reflection measurements.

Separation of the bulk and surface components of recombination lifetime obtained with a single laser/microwave photoconductance technique

An algorithm for separating the bulk and surface components of recombination lifetime obtained via a contactless single laser excitation/microwave reflection decay measurement is presented. The surface recombination component of lifetime is determined by extrapolating the tail portion of the carrier decay curve to the carrier axis. Although the slope of this curve depends on both surface and bulk properties, it is shown that the y intercept depends only on the surface component of lifetime. A wide range of surface lifetimes, corresponding to surface recombination velocities from 102 to 105 cm/s, and bulk lifetimes from a few microseconds to several hundred microseconds can be measured. An experimental verification of the analysis is presented using microwave absorption/reflection measurements on silicon wafers representing a wide variety of bulk and surface lifetime components.

Minority-carrier lifetime analysis of silicon epitaxy and bulk crystals with nonuniformly distributed defects

Existing analyses of the pulsed response of an MOS capacitor for minority-carrier lifetime determination result in a lifetime value averaged over most of the depletion region width. The authors present an analysis of MOS capacitance-versus-time data that enables minority-carrier generation lifetime to be plotted as function of depletion-region depth. The technique is shown to be useful for samples with bulk or buried interfacial layer defects that have defect-free surfaces. Data are presented for intrinsically gettered bulk crystals and extrinsically gettered Si (2%Ge) epitaxial layers with misfit dislocations. For samples that do have uniform lifetimes, the measurement time required for determining carrier lifetime is reduced by more than an order of magnitude. >

Electrical activity of dislocations: Prospects for practical utilization

The electrical activity of interfacial misfit dislocations in silicon has been examined using the electron beam induced current technique (EBIC) in a scanning electron microscope. “Clean” misfit dislocations, i.e. no EBIC contrast, formed during high-temperature Si(Ge) chemical vapor epitaxy were studied. These defects were subsequently decorated with known metallic impurities (Au and Ni) by diffusion at 400° C to 1130° C from a back-side evaporated layer. Qualitative analysis of the electrical activity in relation to the energy levels anticipated for the clean or decorated dislocations is presented. Of particular interest is the case of defect-induced conductivity type inversion which occurred both at the top surface and at the buried dislocated interfaces of the multilayer. The prospects for using dislocations in a beneficial manner as active elements in electronic devices are discussed.

Gettering of Fe to below 1010 cm−3 in MeV self‐implanted Czochralski and float zone silicon

The effects of Si ion fluence and oxygen concentration on secondary defect formation and gettering of metallic impurities in MeV self‐implanted silicon have been studied for Czochralski (Cz) and float zone (FZ) silicon by means of deep level transient spectroscopy, secondary ion mass spectroscopy, transmission electron microscopy, and optical microscopy/chemical etching. We found that the density, depth distribution, and number of extended defects is strongly dependent upon both the Si ion fluence and the oxygen concentration. Effective gettering of iron to below 1010 cm−3 can be achieved in both FZ and Cz wafers at implantation doses of 1015 cm−2.

Noncontact energy level analysis of metallic impurities in silicon crystals

Noncontact laser/microwave deep level transient spectroscopy (LM‐DLTS) based on the measurement of microwave reflection power as a function of temperature has been developed and applied to Czochralski silicon crystals intentionally contaminated with selected metals during crystal growth. The energy levels related to these metallic impurities in p‐type silicon have been obtained on bare silicon for the first time without any electrode contact or special sample preparation. The data agree in very satisfactory fashion with results obtained by conventional DLTS.

Gated diode leakage and lifetime measurements of misfit dislocation gettered Si epitaxy

Gated diode leakage current and minority‐carrier lifetime are compared between Si wafers extrinsically gettered with epitaxial misfit dislocations and reference homoepitaxial material. The stability and gettering efficiency of the interfacial misfit dislocations have been verified by measuring leakage currents of less than 1 nA/cm2 for both gate depletion and accumulation for a large number of diodes. In either mode, the gated diodes with misfit dislocation gettering exhibited more than an order of magnitude lower leakage current than that produced by standard epi without misfit dislocations. In addition, minority‐carrier generation lifetimes greater than 2 ms were typical of material extrinsically gettered via misfit dislocations, while reference epi was two to three times lower. The lifetme was found to be uniform in the near‐surface region, but was drastically reduced in the immediate vicinity of the misfit locations, indicating that the defects may provide useful options in high‐speed devices, latch‐u...

Electron beam patterning of SiO2

Maskless or ideally ‘‘resistless’ patterning of semiconductors is essential for in situ processes which require regrowth of semiconductor layers. If a practical method of patterning SiO2 were viable, many processing problems related to tasks currently utilizing organic resists could be eliminated. With the knowledge that etching selectivity can be induced in SiO2 by electron beam exposure, experiments were performed with the goal of gaining an understanding of some of the practical aspects of electron beam exposure and etching of SiO2. Electron beam energy optimization considerations for dose minimization were addressed followed by e‐beam exposure and etching of 100 nm films of thermal and remote plasma enhanced chemical vapor deposition SiO2. The etching characteristics of the respective oxides are presented and discussed.

Two classes of recombination behavior as studied by the technique of the electron beam induced current: NiSi2 particles and misfit dislocations in Ni contaminated n‐type silicon

The recombination activity of well‐defined NiSi2 precipitates and of misfit dislocations in Ni contaminated Si samples has been investigated using the technique of the electron‐beam‐induced current in dependence on sample temperature and beam current. Individual NiSi2 precipitates are found to show a high recombination activity, increasing slightly with temperature and decreasing with increasing beam current. On the other hand, misfit dislocations are nearly inactive at room temperature and increase their activity upon cooling the sample. The experimental findings are discussed in terms of recombination activity controlled by either defect charging or shallow centers.