Simple calculations using the delta-lattice-parameter (DLP) model indicate that the enthalpy of mixing is invariably larger than or equal to zero for III/V alloys, and increases with increasing difference in lattice constant for the constituent binary compounds. In terms of the regular solution model this suggests the occurrence of miscibility gaps. Solid phase immiscibility has indeed been observed in a number of systems. Nevertheless, such alloys can be grown by OMVPE, including the highly metastable alloys GaPSb and InPSb. Initially surprising was the occurrence of ordered structures in these same alloys. The regular solution model apparently specifically excludes immiscibility and ordering in the same system. However, when the positive enthalpy of mixing is due to strain energy effects, as in III/V alloys, Hume-Rothery recognized very early that such phenomena should be anticipated. This was later confirmed by detailed first principles calculations. In fact, the tendency for ordering is anticipated to increase as the difference in tetrahedral radii of the elements sharing a common sublattice increases. Thus, it is somewhat surprising that ordering was first observed in the AlGaAs system where Al and Ga have nearly equal sizes. Ordered structures have now been observed in several III/V alloy systems including the ternary systems GaAsSb and GalnP and the quaternaries GalnAsP, GalnAsSb, and AlGalnP. In this paper, ordering in other alloy systems such as GaPSb and InAsSb will be described. Surprisingly, no strong correlation between atomic size difference and the degree of ordering has been observed. Another unexpected observation is that the preferred ordered structure for most ternaries involves ordering along the {111}directions. Both first principles total energy and simple strain energy calculations indicate that such ordered structures are only slightly more stable than the disordered solid solution. Other ordered structures, including L10, where ordering occurs along the {100} directions, are considerably more stable. Both phenomena must be explained in terms of the surface kinetic processes occurring during epitaxial growth. Such an explanation is supported by the importance of kinetic parameters such as growth rate, temperature, and substrate orientation in determining both the degree of order and the specific ordered structures observed.

Ordered and modulated structures in InGaP alloy semiconductors grown on (001) GaAs substrates by metalorganic chemical vapor deposition, chloride-vapor phase epitaxy, and liquid phase epitaxy have been systematically studied by transmission electron microscopy. In InGaP grown by metalorganic chemical vapor deposition at 630°C, strong ordering of CuPt-type has been observed, which is associated with an abnormality in the photoluminescence peak energy. For crystals grown by chloride-vapor phase epitaxy, CuPt-type ordered structures have also been observed. However, the degree of ordering is weaker in the latter case and crystals grown at 576-740°C exhibit normal photoluminescence peak energies. On the other hand, in crystals grown by liquid phase epitaxy, no superstructure spots are found in the electron diffraction patterns and the crystals exhibit normal photoluminescence peak energies. Modulated structures do not depend on the growth method since they are observed in all crystals. From these results, it has been concluded that the ordered structures are not generated under thermal equilibrium conditions but rather by the diffusion and reconstruction of deposited atoms on the growth surface.

MBE InASySb1-y layers were grown at 370°C across the full composition range. TEM/TED examinations directly showed that separation into two phases had taken place for compositions 0.4<y<0.8 with plates 20 to 200nm thick occurring approximately parallel to the layer surface. The two phases in the individual specimens were tetragonally distorted and their compositions were deduced from the TED results. CuPt-type atomic ordering occurred and was most pronounced in the middle of the composition range. Alloy clustering was also present. Hall measurements showed that the carrier concentration was a maximum, and the mobility was a minimum, for the middle of the composition range.

We show in this study that RTP-induced defects analysed by Deep Level Transient Spectroscopy (DLTS) are related to residual impurities present in as-grown silicon wafers. For one particular material an activation of a specific residual metallic impurity was observed in the temperature range 800 - 1000°C. This impurity can be returned to an electrically inactive precipitated form by classical thermal annealing (CTA) with a slow cooling rate or neutralized by means of low-energy hydrogen ion implantation.

In order to study the iron-oxygen interaction and their redistribution in silicon, different doses of 100 keV iron ions have been implanted into CZ silicon substrates, and subsequently annealed. The redistribution of iron and oxygen upon thermal treatment in nitrogen atmosphere was monitored with SIMS and FTIR measurements. A significant gettering of iron as well as of oxygen into the layer close to the implanted surface has been found only for doses exceeding the critical fluence, therefore indicating that the structural disorder and associated lattice strain field are the main driving forces for gettering mechanism. It has been found, however, that for doses above the critical fluence not all iron ions have been trapped into the damaged region, but a significant part of them have diffused into the bulk of the material, unaffected by gettered oxygen.

Defects introduced into silicon by ion beam etching are investigated by low-temperature photoluminescence (PL) and Rutherford backseattering (RBS) measurements. The RBS results show that during the ion beam etch a highly damaged surface layer is formed which contains a large concentration of Ar atoms. The Ar atoms then diffuse out of the surface and into the crystalline bulk by some form of radiation enhanced diffusion. Annealing of the etched samples at 350°C results in the formation of noble gas defects known from previous PL studies of ion implanted silicon. When the samples are annealed at 650βC PL lines due to new defects are formed. Although little is known about their structure, we show that the new Ar defects getter small copper contaminations very effectively.

Angle-resolved X-ray photoelectron spectroscopy characterization of the surface region of high-dose Sb+ ion implanted silicon, after rapid thermal treatments over various temperatures, is reported. The results obtained are compared with the Rutherford backscattering data and the capacitance-voltage measurements on the metal-oxide-semiconductor mesa structures built on them. Rapid anneal at 1100 °C of the 1.4×1016 Sb+/cm2 samples showed an anomalous deep oxygen diffusion inside the implanted region.

Few non-destructive techniques are available which provide information regarding defect type, concentration and depth distribution in semiconductors. The variable-energy positron beam technique has recently demonstrated a sensitivity to near surface defects and impurities at low defect concentrations. In the present study, intrinsic silicon (100) epilayers of ~3000 Å thickness grown by MBE at different temperatures were examined by this method for evidence of changing defect concentration and type.

High voltage transmission electron microscope (JEM-1000) has been used to investigate the electron irradiated defects in in p-type FZ-Si and nitrogen doped p-type FZ-Si. It was found that when the irradiated conditions were the saie ,the irradiated defects were easier to be produced in the FZ-Si than in nitrogen doped FZ-Si in the temperature range 573-773 K. The defect density was higher in the foraer. The migration energy of the vacancies in the temperature range 573-773 K was 0.34 and 0.58 eV for FZ-Si and nitrogen doped FZ-Si respectively. It seems to indicate that there was some interaction between vacancies and nitrogen atoms in the nitrogen doped FZ-Si. The results proved that the nitrogen doped FZ-Si has excellent property against electron irradiation.

The mobility of dislocations in a model Silicon lattice is examined at an atomistic level using molecular dynamics. Straight and double-kinked 30° and 90° partial dislocation glide-set dipoles are modelled in a strain-free environment: reconstruction and antiphase defects are found to be present for 30° partial dislocations. The effects of applied shear strains and of temperatures up to the melting point are considered.

Germanium bicrystals containing dissociated low energy grain boundaries with a common [011] tilt axis were studied. The dislocation structure of a grain boundary with a misorientation of about 28° has been investigated in detail and could be interpreted as a near-coincidence Σ 27 - tilt grain boundary. Both conventional and high-resolution techniques were applied to analyze the Burgers vectors of the secondary dislocation network. In all cases non - primitive Burgers vectors were observed. The comparison with previous results indicates that the Burgers vectors of the dislocation network may depend on the degree of misorientation of the bicrystal.

A focused photon beam chopped at 2 kHz scans p-n junctions in a p-type Si wafer and ac photovoltages are capacitively measured in order to inspect homogeneities of the junctions. It is found that the ac photovoltages are high not only in the junction areas but also in the field oxide regions around the junctions when the junctions are leaky. This indicates that dense positively charged traps exist at the interface between the heavily boron implanted Si substrate and the field oxide layer around the high leakage junction, and that the traps cause the increase in both the junction leakage current and the ac photovoltage.

Conventional and high resolution electron microscopy (HREM) were used to study the structure of the {113} defects in high-dose oxygen implanted silicon. The defects are created with a density of 1011 cm-2 below the buried oxide layer in the substrate region. The {113} defects are similar to the ribbon-like defects in bulk silicon. Our HREM observations show that two crystalline phases are present in the defect. Portions of the defects exhibit the original cubic diamond structure which is twinned across {115} planes. The atomic model shows that the {115} interface is a coherent interface with alternating five- and seven-membered rings and no dangling bonds.

We report a quantitative investigation of silicidation in Ti/amorphous-Si thin-films using Differential Scanning Calorimetry (DSC), thin-film X-ray diffraction and Cross-sectional Transmission Electron Microscopy (XTEM). Multilayered thin films were used to facilitate calorimetric observation of the heat released or absorbed at many reacting interfaces. It is shown that calorimetric analysis, combined with structural analysis using X-ray diffraction and XTEM, is effective in providing both kinetic and thermodynamic information about interdiffusion reactions in thin films. The present paper describes experimental results for multilayers with an atomic concentration ratio of 1 Ti to 2 Si and modulation periods ranging from 10 to 60 nm. A thin amorphous titanium suicide layer was found to exist between the as-deposited Ti and a-Si layers. Heating the multilayer film caused the amorphous Ti-silicide to grow over a broad temperature range by an exothermic reaction. An endothermic relaxation occurs during the late stage of amorphous suicide growth. Heating to temperatures over 800K causes C49-TiSi2 to form at the a-si1icide/a-Si interface. Temperatures at which all the above structural transitions occur vary with modulation period. Analysis of the DSC data indicates an activation energy of 3.1 eV for the formation of C49-TiSio, which is attributed to both the nucleation and the early growth of the suicide. The heat of formation for C49-TiSi2 from a reaction of a-silicide and a-Si was found to be -30±5KJ/mol. Nucleation appears to be the controlling step in C49-TiSi2 formation.

We studied oxygen precipitation in quenched Czochralski silicon crystals. The larger the cooling rate and the higher the quenching temperature, the more oxygen precipitated. The defects enhancing oxygen precipitation are eliminated by annealing above 900°C. The defects are formed and removed repeatedly by quenching and slow cooling. We think that the aggregation of vacancies is related to those defects.

MeV ion implantation in GaAs is known to cause amorphization of the region at the end of the ion range. The near-surface region, however, is still crystalline albeit heavily compensated. We have carried out deep level transient spectroscopy (DLTS) studies of the defect levels in the near—surface region of n—GaAs samples implanted with different doses of 2.0 MeV 16O+ ions.

A comparison between the defect structures in the original and the implanted samples shows that implantation produced a broad range of defect levels ranging from 0.58 to 0.3 eV from the conduction band edge. This broad range of defects has an unusually large capture cross—section. The intensities of the DLTS peaks increase with the dose of 160+ ions. The presence of EL2, which was present in the original samples, was not observed in the implanted samples.

Results from measurements made on samples that have been implanted at 200°C and on implanted samples subjected to rapid thermal annealing will also be discussed.

We report the study of high-dose Si-implanted GaAs containing doses ranging from 1×1014 to 1×1015 cm-2 and with subsequent anneals at 850°C for 1 hour. At doses ≥ 3×1014 cm-2, a severe reduction of carrier concentration and anomalous Si diffusion are observed in the near-surface region. In the same region, small, near-spherical voids are found by transmission electron microscopy. In contrast, for samples implanted with doses ≤ 1×1014 cm-2, voids are not found, and both normal carrier activation and Si diffusion profiles are observed. The concurrent onset of these three phenomena in the same region in high-dose samples leads us to conclude that the severe reduction of carrier concentration and anomalous Si diffusion are attributable to the formation of voids.

Damage resulting from reactive ion etching (RIE) and wet etching of MBE grown GaAs and the defects generated in subsequent GaAs regrowth was evaluated by Transmission Electron Microscopy (TEM), photoluminescence (PL) , and Raman spectroscopy. The samples were reactively etched with HCl and Cl2 at two power densities; 0.3 W/cm2 and 0.6 W/cm2. A wet chemical (H2O:H3PO4:H2O2) etched sample was used for comparison. The LO phonon intensities measured by room temperature Raman spectroscopy showed a strong correlation to the relative PL intensities, where the HCl sample was found to be inferior to the others. TEM microstructural evaluation showed both HCl RIE samples to have significant interface roughness, with the sample etched at 0.6 W/cm2 showing the most structural damage in the regrowth region. Defect densities for the Cl2. and wet etched samples were two orders of magnitude lower than that of the HCl etched samples. The extent of disorder in the regrowth region was largest for the HCl sample with respect to the Cl2 and wet etched samples. These microstructural and optical results suggest that the H+ ion plays an important role in the generation of crystal defects during the RIE process.