Growth Of Stacking Faults Research Articles

Improvement to Bulk Oxygen Precipitate Density Necessary for Internal Gettering in Antimony Doped N/N+ Epitaxial Wafers Using the 3‐Step Technique and Ramped Nucleation

The addition of a ramped nucleation cycle, just after an initial oxidation, is shown to greatly increase the precipitation of interstitial oxygen during simulated processing on N/N+ antimony doped epitaxial wafers. When used with fully processed wafers, the ramped nucleation improves internal gettering. Oxygen precipitate density and bulk stacking fault density achieved are about a factor of 100 higher when using ramped nucleation. These results indicate that N/N+ antimony doped epitaxial wafers have an intrinsic gettering capability as good as that of lightly doped material when using an activation cycle (initial oxidation) and subsequent ramped nucleation. A method for calculating bulk stacking fault density is presented. It is shown that, at high concentration (above 107 cm−3), the square of the bulk stacking fault length is inversely proportional to their density. This finding indicates that the growth of bulk stacking faults is limited by the supply of interstitial silicon generated in the bulk during oxygen precipitate growth.

Use of ab etchant for clarifying the structural-impurity inhomogeneties in epitaxial layers of GaAs

In this paper we demonstrate the possibility of using Abrahams-Buiocchi etchant, widely used for revealing the dislocation structure of GaAs monocrystals, for studying submicroscopic impurity micro-inhomogeneities in epitaxial layers of GaAs. The required information is achieved by using etching patterns of electron microscopy replicas for the study. The presence of the “memory” effect during etching is demonstrated as an example of growth stacking faults. Impurity micro-inhomogeneities are visualized in the bulk of the epitaxial layers, caused by the presence of centers of step obstructions on the growth surface.

Structure of Au(100) and Au(111) single crystals surfaces prepared by flame annealing

The surfaces of flame annealed Au(111) and Au(100) single crystals have been investigated by synchrotron X-ray diffraction and reflectivity. Electropolished crystals were flame annealed for various lengths of time and studied for structural changes. Electropolishing of the Au(111) and Au(100) crystals produced smooth surfaces with no evidence for reconstructions. Upon flame annealing the Au(111) was slightly roughened, possibly due to stacking fault growth at the surface, while the Au(100) was smoothed with evidence for the hexagonal reconstruction.

A model of structural disorder and size distribution for palladium supported on pumice

A model of size distribution and structural disorder for palladium grown on pumice is developed. The model parameters are determined by fitting the simulated powder pattern to the experimental data. Three samples, characterized by 2.4, 1.14 and 0.86 wt.% Pd respectively were examined. The results point to nearly spheroidal crystallites, with sizes ranging from 40 to 20 Å, affected by stacking faults in the sequence of the close-packed (111) layers to an extent independent of size and by intralayer second-kind distortions. A comparison with the results obtained following classical procedures is supplied. In the Appendix a new approach for the determination of growth stacking faults from peak shifts is presented.

Determination of the Kinetic Coefficients of Silicon Self‐Interstitials from Oxygen Precipitation/Front‐Surface Stacking‐Fault Growth Experiments

An oxygen precipitation/surface stacking‐fault growth experiment has been carried out to determine the kinetic coefficients of silicon self‐interstitials. In this experiment, silicon self‐interstitials were injected by the precipitation of interstitial oxygen within the bulk of the silicon samples. The sample front surfaces were capped with oxide or nitride layers, and the concentration of self‐interstitials at the capped surfaces were monitored by the growth or shrinkage of surface stacking faults. Experimental results have been analyzed using steady‐state and transient models, based on the assumption that self‐interstitials dominate the kinetic processes of intrinsic point defects. From these analyses, estimates for the diffusivity the equilibrium concentration , and the surface‐reaction constant , and at nitride and oxide interfaces were obtained at 1125°C.

Determination of the Kinetic Coefficients of Silicon Self‐Interstitials from Back‐Side Oxidation/Front‐Surface Stacking‐Fault Growth Experiments

A series of back‐side oxidation/front‐side stacking‐fault growth experiments have been carried out to determine the kinetic coefficients of self‐interstitials in silicon. In these experiments, wet and dry oxidations of the back side of thinned silicon samples were used to inject self‐interstitials from the back surfaces. The sample front surfaces were capped with oxide or nitride layers, and the concentration of self‐interstitials at the capped surfaces were monitored by the growth or shrinkage of surface stacking faults. Experimental results have been analyzed using steady‐state and transient models, based on the assumption that self‐interstitials dominate the kinetic processes of intrinsic point defects. From these analyses, the relative recombination rates of self‐interstitials at oxide and nitride boundary layers have been obtained, with an oxide layer found to absorb self‐interstitials at about three times the rate of a nitride layer. The results also suggest that the surface recombination coefficients are time dependent rather than constant, as has been previously assumed.

Two-Time Three-Equation Method for Analysis of Oxidation-Enhanced and -Retarded Diffusions and Growth of Oxidation Stacking Faults in Silicon

For analysis of oxidation-enhanced and -retarded diffusion (OED and ORD) and oxidation stacking faults (OSF) in Si, two out of the three equations for OED, ORD and OSF were used. The equation for a local equilibrium between self-interstitials and vacancies was also used. Thus three equations were used. Two time values were taken in these three equations. In the previous paper (Jpn. J. Appl. Phys. 27 (1988) 967), the experimental results were modified in order to obtain physically reasonable solutions. But this is not good. In the present work, therefore, the equations were simultaneously solved without modifying experimental results. It was concluded that the equation of OSF should not be used.

The growth mechanism and kinetics on stacking fault reconsidered

Making use of the first nearest neighbour approximation it is shown that the sub-step created on a (111) face of an fcc crystal by stacking fault makes the barrier for the two-dimensional (2D) heterogeneous nucleation on the surface about twice as low as that predicted by Ming et al. [J. Crystal Growth 91 (1988) 11]. Thus, the stacking fault growth mechanism is much more effective for the step generation on the growing surface as it was considered earlier.

The role of silicon self-interstitial supersaturation in the retardation of oxygen precipitation in Czochralski silicon

The retardation phenomenon of oxygen precipitation in Czochralski silicon has been studied simultaneously with the growth of surface stacking faults under a silicon nitride capping layer. The surface faults were intentionally introduced to monitor the bulk self-interstitial supersaturation in the crystal during precipitate growth. It was observed that an increase in the low-temperature nucleation anneal time resulted in a reduced rate of oxygen precipitation and an enhanced rate of surface stacking fault growth at high temperatures. This indicates that as the nucleation anneal time increases, silicon self-interstitials, which must be generated for precipitation to proceed, reach a high level of supersaturation and cause the annihilation of most nuclei generated during the nucleation anneal.

Theoretical model for self-interstitial generation at the Si/SiO2 interface during thermal oxidation of silicon

Using a grating pattern of parallel nitride and oxide stripes on the silicon surface, self-interstitial concentration at the Si/SiO2 interface is accurately determined by means of oxidation-induced stacking fault growth observation. The results show that the interstitial concentration at the interface is found to be determined by the oxidation of interstitials in the oxide rather than self-interstitial diffusion in bulk silicon. A newly proposed physical model incorporates the balance among the generation rate of interstitials at the oxidizing interface, the annihilation rate of interstitials due to surface regrowth, and the rate at which they diffuse into the oxide and react with incoming oxidants. For higher oxidation rates, the model predicts that the interstitial concentration follows a 0.5 power dependence on oxidation rate. For slower oxidation rates, the concentration at the interface becomes lower than the equilibrium interstitial concentration, leading to the oxidation-retarded diffusion of dopants with a high fractional interstitialcy diffusion component.

Analysis of Oxidation Enhanced and Retarded Diffusions and Growth of Oxidation Stacking Fault in Silicon

In order to determine the fractional components of the interstitialcy mechanism for Sb, P and self-diffusions, dISb, dIP and dIsd, and the supersaturation ratios of vacancies and self-interstitials, sV and sI, from the experimental results of the oxidation-enhanced diffusion (OED) of P, oxidation-retarded diffusion (ORD) of Sb and growth of the interstitial-type stacking fault by oxidation (OSF), the equations of OED, ORD and OSF and of the special relation between sV and sI were solved simultaneously. The effect of the stacking fault energy upon growth of the OSF was taken into account in the OSF equation. As the experimental results of OED, ORD and OSF did not satisfy their equations exactly, nine kinds of solutions were obtained and three of them were shown. The errors caused by the lack of exact satisfaction were shown. A dIsd much smaller than 0.5 was obtained.

Ion Impiantation Doping of Siox with 31P and 69Ga

ABSTRACTImplantations of annealed SIMOX and Si wafers have been done using P and Ga to investigate the effect of excess oxygen and oxygen precipitates on amorphous layer regrowth and category II (end-of-range) dislocation loop elimination. Solid phase epitaxial regrowth of the amorphous silicon in both SIMOX and Si control wafers occurred at 550°C without the formation of category III defects and annealing at 900°C 16 hours resulted in complete removal of the category II defects. In an oxidizing ambient, the implanted SIMOX wafer again exhibited complete defect elimination whereas the Si control wafer showed growth and development of extrinsic stacking faults. It is speculated that the buried oxide may act as a sink for the Si. SIMS results indicate the dopant getters to the Si/SiO2 interfaces and that redistribution can be modelled reasonably well with SUPREME III.

THE STACKING FAULT GROWTH MECHANISM AND ITS KINETICS

The atomic configurations of sub-steps created on (111) growth surface of fee crystal by stacking faults with fault vectors 1/6〈112〉and 1/3〈111〉show that they can act as growth step sources. Growth kinetics of stacking fault growth mechanism has been analysed. It has been shown that the nucleation barrier at the sub-steps is always smaller than that of the 2-di-mensional nucleation.

Point Defect Generation during Phosphorus Diffusion in Silicon: I . Concentrations above Solid Solubility

The goal of the present research was to investigate the proposal that P diffusion at concentrations above solid solubility generates silicon self‐interstitials. Buried layers of As and Sb were created by either implanting Sb at 150 keV, or As at 100 keV, in ‘100’ Si substrates. After a 900°C, 30 min anneal in , an 8–10 μm thick epi layer was grown over the buried layers. Masking oxides were then created either by depositing a LPCVD layer (1 μm) or by thermal oxidation at 900°C in steam. Spreading resistance profiles were taken on all samples. SIMS was used to characterize the chemical P doping densities. Plan view and cross‐section TEM was used to look for precipitates and defects in the P‐diffused layers. Major results of this matrix of experiments are the following. 1. Phosphorus‐related precipitates 0.1–0.2 μm in size were observed in selected samples using TEM. In all cases their concentration is insufficient to account for the concentration of nonelectrically active P. 2. Measurable enhanced diffusion of As in the buried layers occurred over the temperature range 1000°–1150°C with an activation energy of ≈1.5 eV. This enhanced diffusion only occurred under the windows in which P was diffusing. 3. Measurable retarded diffusion of Sb in the buried layers occurred over the temperature range 1100°–1200°C with an activation energy of ≃6.6 eV. 4. Stacking fault growth in the buried As and Sb layers under the P‐diffused region implies that a silicon self‐interstitial supersaturation was produced in the P doped layer. Self‐interstitials feed the stacking fault growth and the enhanced As diffusion and retarded Sb diffusion. 5. The level of self‐interstitials generated cannot be accounted for by the sparse number of precipitates formed in the P‐diffused region.

X-ray analysis of strain in titanium nitride layers

Titanium nitride layers were deposited by d.c. reactive magnetron sputtering onto steel substrates maintained at 50 °C. The layers of thickness about 5 μm revealed preferred orientation of h.c.p. atomic planes parallel to the substrate surface. It was found that the major part of the X-ray line broadening is strain broadening probably caused by lattice defects and growth stacking faults. The strain decreases in the direction from the substrate to the free surface in textured coatings with low or medium nitrogen contents. A relatively strong crystallographic anisotropy of this strain was observed. The hardest coatings are nearly textureless, with the highest observed values of the strain but free from the strain gradient. Their average domain size is 23 nm.

Point-defect generation during oxidation of silicon in dry oxygen. II. Comparison to experiment

In the companion paper [J. Appl. Phys. 59, 2541 (1986)], we developed a model based on a continuity equation approach for the supersaturation of interstitials during oxidation in dry O2. The analysis resulted in an expression having the experimentally observed sublinear dependence of oxidation rate without requiring any assumptions of nonlinearity. Through observations of the dependence of the linear rate constant at the linear-parabolic model of oxidation on ambient oxygen pressure, predictions were made of how the interstitial supersaturation will vary with oxidation rate at a given temperature. In this paper, those predictions are compared to data for the enhanced diffusion of phosphorus in 〈100〉 silicon during oxidation at 900 and 1000 °C in both 100% O2 and O2-Ar mixtures as well as existing data for oxidation enchanced diffusion and stacking fault growth.

Observation of silicon self-interstitial supersaturation during phosphorus diffusion from growth and shrinkage of oxidation-induced stacking faults

The effects of phosphorus diffusion on the growth and shrinkage of oxidation-induced stacking faults (OSF) has been studied in the temperature range from 950 to 1100 °C. By using Wright etch on angle-lapped samples, OSF size dependence on depth has been obtained. OSF grow faster or shrink slower with increasing phosphorus dose, thus indicating a self-interstitial supersaturation. Moreover, it was found that this supersaturation is not the result of phosphorus precipitation. Silicon self-interstitial concentration normalized to its equilibrium value has been calculated from the OSF kinetics and was compared with time, phosphorus dose, and temperature. It was found that the interstitial supersaturation increases with phosphorus dose and decreases with annealing time. At the same phosphorus deposition condition, the interstitial concentration decreases with increasing temperature. Normalized self-interstitial concentration calculated from enhanced phosphorus diffusivity at the profile tail, assuming a pure interstitialcy phosphorus diffusion mechanism, yields values approximately twice as high as those calculated from OSF kinetics. Possible reasons for this discrepancy are discussed.

Dependence of Warpage of Czochralski-Grown Silicon Wafers on Oxygen Concentration and Its Application to MOS Image-Sensor Device

The resistance of precipitation-treated Czochralski-grown silicon wafers to warpage has been investigated using crystals containing oxygen at a concentration of 5.5-12.3×1017 atoms/cm3. The precipitation softening is effectively suppressed if the oxygen concentration is lower than a threshold value of about 8×1017 atoms/cm3. In wafers with oxygen concentrations of 5.5-8×1017 atoms/cm3, oxygen precipitation proceeds slowly, resulting in slow growth of bulk stacking faults and few dislocation sources. In MOS image-sensor devices, the wafers with such oxygen concentrations have few crystal defects within the area of the photodiodes after processing, which cause the fatal failure of white blemishes detected as white scratches on output pictures.

An X-ray fourier line shape analysis of hexagonal tellurium films III: Vacuum evaporated onto air-cleaved mica substrates at low temperature (133 K)

As a continuation of a series of studies of hexagonal tellurium films deposited onto glass at room temperature and low temperature (133 K), investigations were carried out on the growth structure of tellurium films 331–800 nm thick (for normal vapour incidence) and 260–1129 nm thick (for oblique vapour incidence) deposited by vacuum evaporation at various rates of deposition onto air-cleaved mica substrates at a temperature of 133 K. Since mica is a crystalline substrate it was used to study its influence on the growth structures and preferred orientation. Considering the cases of normal and oblique vapour incidence for both high and low rates of deposition, a small domain or crystalline size effect (about 115–160 Å), appreciable microstrains (2.2–5.0)×10 -3 in the domains and a near-isotropy in the size and strain parameters were observed from a detailed Fourier analysis of X-ray diffraction profiles. The results obtained from the analysis of fault-affected reflections ( H - K = 3 N±1, L 0 odd or even) show that the intrinsic stacking faults (probability α) are totally absent and the presence of growth stacking faults (probability β) is also negligible. The dislocation density ϱ is about 10 11 cm -2 and a c axis texture is found to develop in the films. The average residual internal stresses σ and lattice parameter changes Δ a and Δ c, obtained from line shift analysis, are small. From a comparison of the results obtained for the different cases it appears that the changes in the angle of deposition or the rate of deposition as well as the nature of the substrate (mica) do not markedly affect the growth structures and orientation of the films deposited at this lower temperature.

Kinetics of interstitial supersaturation and enhanced diffusion in short-time/low-temperature oxidation of silicon

The kinetics of the supersaturation of self-interstitials and the enhancement of impurity diffusivity in short-time/low-temperature oxidation of silicon is investigated analytically. It is found that, whereas in long-time/high-temperature oxidation the interstitial supersaturation and the diffusivity enhancement decrease with time as t−n (n≂0.2–0.3), in short-time low-temperature oxidation they start from zero and increase with time, until some characteristic time determined by the linear-parabolic oxide growth. This characteristic time increases rapidly with the decrease of temperature. This kinetic behavior has not been expected previously, mainly because there is no available data on stacking fault growth and on diffusion enhancement under the short-time/low-temperature condition which has become common in modern IC processings. The more general case of linear-parabolic oxidation and the effect of bulk recombination have also been analyzed.