It is found that the grain boundaries in cast polycrystalline silicon material capable of enhancing diffusion always have strong recombination activities. Both phenomena could be related to the existence of dangling bonds at boundaries. Because the enhanced diffusion is an atomic transport phenomenon and the recombination is an electronic process, the relationship between the two phenomena is still not clear at this moment. The present study gives the first evidence that incoherent second order twins of {111/115} type are phosphorus diffusion-active.

Phosphorus implanted emitters are known to have a dislocation network formed to relieve the strain caused by the lattice contraction of solute phosphorus ir, silicon. The extension of the dislocation network into the emitter-base regior, has also been related to the degradation of leakage current and the generatiol, of excess noise in an NPN bipolar transistor. The penetration of the extendeot dislocation network is shown in this paper to be related to a shear force. The shear force is the resultant of the compressive force generated from the solute phosphorus and an induced surface force generated from the thick oxide layer over the emitter formed during high temperature heat treatments. This shear force can pull the dislocation network through the junction sidewall.

We have studied interface atomic structure in epitaxial cobalt and nickel disilicides on silicon using high-resolution transmission electron microscopy. By employing UHV techniques during deposition and reaction we have grown truly single-crystalline NiSi2 and CoSi2 films on (111) Si and in the former case on (100) Si. These films are shown to be continuous to below 10Å thickness. By close control over preparation conditions, afforded by UHV, we can greatly influence the nucleation and growth of these films to the extent, for example with NiSi2 on (111)Si, of yielding continuous single-crystal films with either of two orientations as desired. Whilst in the (111) NiSi2 on Si system the interfacial structure invariably appears to well-fit a model in which metal atoms nearest to the interface are 7-fold co-ordinated, for (111) CoSi2 on Si agreement is generally better with a model involving 5-fold co-ordination of these atoms. A misfit dislocation core is also imaged. Results are discussed in the light of silicide nucleation and growth. The structure and stability of the (100) NiSi2 on Si interface is also considered.

The electrical properties of ion irradiated metal-semiconductor contacts are investigated. Silicide contacts are fabricated by depositing Pt on chemically clean or slightly oxidized (∼14 Å SiO2) n+-and n-type <111> Si, followed by a Si ion irradiation (1014 − 6×1015 Si/cm2) through the metal-Si interface at various substrate temperatures, and a final thermal annealing in vacuum to form the silicide. Forward I-V measurements are employed for electrical characterization. Metal-Si interaction and substrate damage are measured by MeV ion backscattering and channeling, and interfacial oxygen monitored by nuclear 16O(d,α)14N reaction.

Platinum contacts prepared on clean n-type substrates are Schottky diodes with a barrier height øBn = 0.83 eV. After Si irradiation, the forward I(V) is a power law whose form is largely independent of the dose. Subsequent thermal annealing induces silicide formation, but at a reduced rate compared to irradiated samples. The do characteristics is roughly exponential again, but departures from the original Schottky characteristics remain and are largest for the highest Si doses. The effect is attributed to radiation damage in the Si that is not consumed by the silicide reaction.

Platinum contacts prepared on chemically oxidized samples behave differently for different substrate materials, although the total amount of interfacial oxygen is always the same. On n+-type samples, the silicide formation at 400°C is laterally uniform for Si doses ≥ 2 × 1014 cm−2, but is nonuniform for all doses (≤ 2 × 1015 Si/cm2) on n-type samples. For n+-type samples at 250°C, a dose of 2 × 1015 Si/cm2 is required to induce (uniform) silicide formation; the kinetics displays a time delay compared with that of clean n+ substrates. On oxidized n-type substrates, the I(V) characteristics of Pt contacts before irradiation is not Schottky-like, but power-law-type. After irradiation, the characteristic is the same as for the clean irradiated samples. Thermal annealing induces only incomplete recovery toward an exponential behavior.

These results demonstrate that radiation damage in the unreacted Si remains significant for the electrical behavior of all.

These results demonstrate that radiation damage determines the I(V) characteristics of as-irradiated Pt contacts to n-type Si regardless of the presence of an interfacial oxide layer. After annealing at 400°C for 30 min, radiation damage is still significant, but the oxidized samples recover less than the clean ones. The results are attributed to radiation damage in the unreacted Si substrate.

Two problems connected with the growth of Ti-silicide have been investigated. It is shown if a silicon dioxide step on a single crystal of silicon covered with titanium is annealed then, following vertical growth on the silicon part, lateral growth of Ti-silicide takes place over the oxide layer. We also studied the problems of Ti-silicide growthon samples implanted with high doses of Sb, As, P, Ar and O prior to Ti evaporation.

The effect of oxygen incorporation on the growth and microstructure of TiSi2 has been investigated. Cosputtered films, with Si/Ti ratios between one and three, were deposited on (100) Si substrates and reacted at temperatures from 650° to 1050°C. Both Auger electron spectroscopy, in conjunction with sputter profiling, and Rutherford backscattering spectrometry indicate that oxygen in the as-deposited films redistributes to the silicide-silicon interface upon heating. Cross sectional transmission electron microscope images show that the oxygen is present as an amorphous oxide.

The use of a Time-of-Flight Atom Probe in the analysis of silicon surfaces, and the interfaces between metals and silicon, promises to provide very accurate chemical analysis allied with structural information from Field Ion Microscopy images. This paper presents results on the analysis of silicon surfaces by this technique,showing that good spectra can be obtained without difficulty. Some preliminary experiments on the structure of such specimens after the deposition of thin layers of Pd and Ni will be described,concentrating on the analysis of the stoichiometry of the reacted layers.

A novel crystal growth technique for NiSi2 epitaxy is presented which utilizes thin silicide (<60Å) template layers to pin the subsequent growth under ultrahigh vacuum conditions. Single crystalline NiSi2 films can be grown with either type A or type B orientations on Si(111). Continuous single crystalline NiSi2 is grown on Si(100) with flat interface and uniform thickness.

Three types of structure-related problems in VLSI materials and the impact of these problems on device properties are discussed: substrate defects and device leakage/breakdown, metallization linewidth and device failure, and SiO2/polysilicon interface texture and device leakage/breakdown. In many cases transmission electron microscopy is the major source of information on structure.

Diode lasers and/or LEDs fabricated from the Ga1 − xAlxAs/GaAs or In1 −xGaxAsyP1−y alloy system provide the basis for rapidly developing optical communications systems. These devices are operated at very high optical and electrical power densities, and the inevitable less-than-perfect efficiency results in intense local thermal and athermal lattice excitation. This in turn can lead to generation, motion and growth of extended defect structures including dislocation networks, precipitation of impurities, or phase separation.

For the AlxGa1−xAs material, a predominant degradation effect is the development of so-called <100> “dark line defects” (DLDs). These have been shown to arise from recombination enhanced climb of threading dislocations giving rise to dipole loops of a primarily interstitial character. A less common form, the <110> DLDs, is associated with slip and typically arises from the strain associated with mechanical damage or careless handling or device processing.

The In1−xGaxAsyP 1−y device material affords less energy per recombina ion event by virtue of its smaller band-gaps, but is likely to have more lattice mismatch strain than Ga1−xAlxAs for typical compositions of device interest. Thus slip, <110> DLDs, and precipitates are the commonly observed features of In1−xGaxAsyP1−y degradation.

A reversible change in the flow stress with the illumination of a band gap light (photoplastic effect: PPE) is observed commonly in II-VI semiconducting compounds both with the zincblende and the wurtzite structures. After reviewing the experimental results accumulated so far concerning the usually observed PPE, detailed microscopic experiments on CdTe (partly on CdS) single crystals are described. In-situ TEM straining experiments with a laser illumination system clarified that the dislocation mobility, evidently controlled by the Peierls mechanism, is not affected by light illumination.The investigation of the dislocation glide behavior over a longer range using the etch pit method and the cathodoluminescence microscopy with a SEM revealed that the dislocations are apt to become immobilized after traveling a certain distance from the sources with a high velocity.The dislocation loops and cusps observed by TEM suggest that the immobilization is caused by jog formation along screw segments. Those results suggest that the positive PPE is caused by the decrease in the mean free path of multiplied dislocations due to the enhanced jog formation by illumination. This viewpoint is supported by the high density of dislocation debris left in specimens deformed under illumination, the enhanced work hardening rate under illumination and the reduction of the PPE when the dislocation mean free path becomes determined by other photoinsensitive factors such as forest dislocation cutting and specimen size. The microscopic mechanism of the enhanced jog formation is discussed in terms of electrostatic interaction between charged point defects and dislocation charge which is increased with the illumination.

Various degradation modes and features of crystalline defects associated with the degradation observed both in GaA1As/GaAs and InGaAsP/InP double heterostructure light emitting sources (LED's and Lasers) are reviewed, noticing similarities and differencies between those two material systems. Non-existence of rapid degradation in the quaternary caused by DLD formation (dislocation motion) will be discussed in terms of atomic rearrangements arouna the dislocation core.

Photon energy from laser beams can be used to rapidly heat and melt localized regions of semiconductors with a high degree of spatial and temporal selectivity. Pulsed lasers have been successfully used to anneal displacement damage and to remove other defects. However, the number density of trapped defects increases with velocity of solidification and finally thin layers turn directly amorphous after laser-melt quenching. Annealing characteristics are found to be a strong function of ion implantation variables, which determine optical properties of materials. By both solid- and liquid-phase crystallization, supersaturated solid solutions can be formed. Residual defects in SPE grown layers primarily consist of dislocation loops. Device applications utilizing these transient thermal processing techniques are reviewed briefly.

We have applied stress measurements using the cantilever beam technique and Raman spectroscopy to characterize the dose dependence of damage production for He+, C+, or Ar+ implants into GaP. Stress increases monotonically with dose until a speciesdependent critical dose is reached. Above that dose, the material yields at an integrated lateral stress of ∼2×105 dynes/cm2 corresponding to an expansion of ∼1% in the implanted volume. The dose dependence of stress scales well with the volume density of ion energy deposited into atomic collisions. Raman measurements indicate that the material is still crystalline when the yield stress is reached.

Formation of amorphous (α) layers in Si during ion implantation in the energy range 100 KeV–11MeV and temperature range liquid nitrogen (LN)-100°C has been investigated.Cross-sectional transmission electron microscopy (XTEM) shows that buried amorphous layers can be created for both room temperature (RT) and LN temperature implants, with a wider 100 percent amorphous region for the LN cooled case. The relative narrowing of the α layer during RT implantation is attributed to in situ annealing. Implantation to the same fluence at temperatures above 100°C does not produce αlayers. To further investigate in situ annealing effects, specimens already containing buried α layers were further irradiated with ion beams in the temperature range RT-400°C. It was found that isolated small α zones (< 50 Å diameter)embedded in the crystalline matrix near the two α/c interfaces dissolved into the crystal but the thickness of the 100 percent α layer was not appreciably affected by further implantation at 200°C. A model for in situ annealing during implantation is presented.

Non destructive analysis by Raman Scattering (RS) and Spectroscopic Ellipsometry (SE) is demonstrated on B+- and Se+ shallow implanted GaAs. Qualitative informations are obtained from 1st and 2nd order RS spectra. The former are analysed using the intensity ratio of the T0 and L0 modes, which defines a lattice potential perfection scale. The SE analysis of the E1, E1 +Δ1 structure in the imaginary part of the dielectric function confirms the RS results and its multilayer analysis yields the depth profile of the implanted ions.

The purpose of this study is to analyze ion implant damage profile using RBS and ellipsometry. Silicon wafers were implanted with 75As at 100, 200 or 300 keV; doses were chosen to generate constant peak impurity concentrations at each energy. The samples were then analyzed using RBS to obtain damage- depth profiles and ellipsometry to obtain Δ,ψ parameters. At light doses decreasingΔ values correspond to increased scattering yield; at higher doses ψ increases rapidly as the scattering yield approaches the random value. The higher energy implants shift the Δ-ψ curves to larger ψ values. Multilayer structures, that include lightly damaged silicon on either side of the project range as well as more damaged near the projected range, are required to model the ellipsometer parameters.

We have determined the dielectric function of silicon samples which were implanted with 100-150 KeV P, As, Si ions to doses of 2·1014-1·1016cm−2, by a rotating analyser Automated ellipsometer in the spectral range 1.77− 4.59 eV. These data have been analyzed using a simplified three layer model.The results are compared with an earlier ellipsometric investigation [2].

N-type, 1015 −1016 cm−3 doped, Fz, Silicon has been implanted with 1 to 4 × 1012 cm−2, 100 or 300 keV, As ions. The nature and concentration of the defects has been monitored using Deep Level transient Spectroscopy as a function of the thermal treatment (in the range 500–900°C) and of the energy of the pulse (15 ns) of a ruby (0.69,μm) laser (in the range 0.3 to 0.6 J cm−2). The defects resulting from annealing by the two treatments are found to be the same. Only, for energies higher than 0.5 J cm−2, the laser treatment introduced new defects (at Ec−0. 32 eV), presumably resulting from a quenching process. Thus a laser energy below the theshold for melting and epitaxial recrystallization is able to anneal the defects produced by implantation, demonstrating that the annealing process induced by the laser pulse is not a purely thermal process but probably involves an ionization enhanced mechanism.

Zr ions have been implanted at 300 keV (Rp= 1400Å) and doses of 3×1012 − 3×l015 Zr/cm2 into Si-implanted, amorphous Sip layers on (100) bulk Si and Sion- sapphire. Rutherford backscattering and channeling spectrometry was used to study the Zr distribution and lattice location during solid-phase regrowth of the Si layers. The regrowth at 500—550°C stops at 3.4хl020 Zr/cm3, and Zr exhibits interface trapping and surface segregation effects. In this temperature range, Zr is essentially non-substitutional, and inactive electrically.