This thesis consists of a study of several qualification techniques for SI LEC GaAs and the application of these techniques to various ingots. For use on the starting material before any doping procedures, the technique of studying the semi-insulating properties by monitoring the activation energy of dark resistivity with temperature was investigated. Experiments were performed on both ring dot as well as cloverleaf samples. Different activation energies for the dark resistivity were observed for temperatures above and below 290 K. Also, ingots with different background impurity concentrations were tested. Another technique applicable to the undoped starting material is Optical Transient Current Spectroscopy (OTCS). The occurrence of 'negative' peaks was simulated using a depletion layer model. The results showed that under certain conditions a recombination centre can produce a positive peak, a negative peak, or both a positive and a negative peak. Further analysis of the negative peaks led to the formulation of a field enhanced injection model to explain their occurrence. More than one negative peak was observed experimentally. In addition, the effect of different electrode structures on OTCS experiments was investigated. The effect of polarity on negative peaks was studied using ring dot structures and was found to agree with the proposed model. Some peculiar anomalies which were observed in investigating OTCS led to the discovery of a photocurrent memory effect with decay time constants of the order of minutes at a temperature of 266 K. This memory effect was found to be associated with surface modifications. The effects of surface passivation with Na₂S were investigated. The method of normalizing the OTCS peak height with photocurrent was investigated. A microscopic spatial analysis tool, scanning OTCS, with a spot size of about 2 µm was developed in order to probe the spatial variation of deep levels and compare with that of dislocations or other defects. An experiment on an abraded surface was performed using the scanning OTCS and showed that the negative peak does indeed correlate with mechanical damage. Wafer performance during implantation doping is an important qualification test. Comparisons between standard furnace annealing and rapid thermal annealing were performed. A comparison of the estimated percentage activation using C(V) measurements with that from Hall measurements, with and without a correction for the surface depleted region, was performed. The C(V) analysis technique, used in the industry to obtain doping profiles of implanted wafers, was studied. The effect of using serial and parallel measurement modes was investigated. Simulations of C(V) measurements on implanted devices by solving the Poisson-Boltzmann equation for the charge distribution under different biases were performed. The limitation of the C(V) profiling technique in detecting sharp dopant profiles was investigated. A system for quick analysis of the percentage of activation using a mercury probe was designed. The effect of serial and parallel analysis of the impedance measured by the mercury probe on the estimated dopant profile was investigated. The effect of different electrode structures (Schottky to Schottky as compared to Schottky to Ohmic) on estimated doping profiles was studied. The mobility profile as a tool for qualification was investigated. The effect of surface states on mobility was studied. A crucial factor in wafer qualification is the uniformity of transistor characteristics across the wafer. In order to test this on a wafer, thousands of transistors have to be measured. A technique of perforating measurements automatically with consistency is needed. An automatic probing station for measuring large arrays of transistors was engineered. Tests on arrays of transistors were performed to investigate the effect of different fabrication processes, in particular the amount of surface etch, on the uniformity of threshold voltage. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

This study examines the effect of furnace annealing and rapid thermal annealing (RTA) on the room-temperature cathodoluminescence (CL) image of liquid encapsulated Czochralski (LEC) GaAs substrates. Furnace annealed samples were heated in a tube furnace for 20 minutes at temperatures from 600 to 950 °C. RTA annealed samples were heated in a commercial RTA furnace for 5 s at temperatures from 650 to 950 °C. The times used for both methods are typical of those used for post-ion implantation annealing although selected samples were RTA annealed at times from 10 to 160 s. The temperature range examined has been extended beyond typical post-ion implantation anneal temperatures to investigate the effect of temperature on the substrate. Examination of RTA annealed GaAs using CL has not been reported previous to this investigation. The CL images of the annealed samples are compared with those of the as-received (un-annealed) material. The CL images of the LEC GaAs wafers prior to annealing have dark spots which correspond to the location of dislocations with regions of higher CL intensity surrounding them. These regions of higher CL intensity are referred to as 'halos'. The remaining material has a CL intensity lower than these halos. The dislocations in LEC GaAs form into cellular networks to reduce the strain energy in the crystal. When viewed at low magnification the overlap of the halos makes the cell walls appear bright and the cell interiors appear dark in a CL image. The furnace annealed substrates show an increase in CL intensity in the interior of the cells. The halos are still present at the cell walls but a region of low CL intensity persists outside the halos making the cell walls appear dark with a brighter interior. This behaviour was seen in all the furnace annealed material although the contrast decreases with increasing anneal temperature. The RTA anneal samples show similar behaviour to the furnace annealed samples at temperatures below 800 °C. Above this temperature the halos are no longer noticeable in the surface CL images, although the region of lower CL intensity can be seen along some cell walls. When a cleaved cross-section is examined using CL, there are regions of higher CL intensity adjacent to both surfaces. These regions typically extend from 100 to 200 μm in from the surface and are nearly uniform in depth. The centre region of the sample appeared the same as the as-received material with dislocation spots, surrounding halos and low CL intensity in the interiors of the cells. The bright regions seen in the cross-section CL images of the RTA samples were examined using photoluminescence at liquid helium temperatures. This investigation found a correlation between the bright regions and the presence of Cu. In addition, the amount of residual Cu on the surface has a significant affect on the depth of the bright bands. Low residual Cu levels results in shallower band depths than high residual Cu levels. It is proposed that the presence of the Cu acts to increase the recombination rate thus increasing the brightness of the CL image. / Applied Science, Faculty of / Materials Engineering, Department of / Graduate

A flash evaporation system is constructed in order to deposit thin films of Gallium Antimonide. The system includes a substrate heater-holder, a film thickness monitoring device and a powdered evaporant feeder. Thin films of GaSb, 2 to 4 microns thick are deposited on glass, sapphire and silicon substrates maintained at temperatures up to 365°C. The deposited films are observed to have many structural defects. The films exhibit semiconductor-like properties. Optical studies of the fundamental absorption edge at liquid-nitrogen temperature show that the films deposited on heated sapphire substrates have a polycrystalline structure and a band gap of 0.78 eV. The nature of the films is dependent on the type and temperature of the substrate. / Science, Faculty of / Physics and Astronomy, Department of / Graduate

Although the organo hydride derivatives of boron and aluminum are well characterized, little work has been reported on the corresponding gallium systems. The present study was initiated to determine the relative stabilities and reactivity of organo gallium hydride derivatives as compared with the stabilities and reactions of the corresponding compounds of boron and aluminum. Various preparative routes to this new class of gallium compounds have been investigated. These include the use of organo-mercury, organo-lithium and lithium hydride derivatives in reactions with gallium hydride and gallium alkyl compounds and their halogen substituted derivatives: Me₃NGaH₃ + HgR₂ → Me₃NGaH₂R + 1/2Hg + 1/2H₂ Me₃NGaH₂Cl + LiR → Me₃NGaH₂R + LiCl Me₃NGaR₂Cl + LiH → Me₃NGaHR₂ + LiCl A fourth preparative method involves disproportionation reactions between gallium hydride compounds and organo gallium compounds to yield the mixed organo hydride derivatives. Alkyl-hydride disproportionation reactions were also examined using organo and hydride derivatives of different Group IIIB elements in order to obtain a better understanding of the exchange process. Both infrared and proton NMR spectroscopy have been used extensively in following the progress of these reactions and in the characterization of the products. / Science, Faculty of / Chemistry, Department of / Graduate

The reaction of gallium arsenide {100} with molecular and atomic bromine was studied at temperatures between 100 and 225°C and at pressures of bromine between 0.1 and 40 Torr. Samples of GaAs were placed on a silicon platform within a Pyrex reactor flow system and the etch rate was determined by profilometry or weight change of the sample. Atomic bromine was produced by a 2450 MHz microwave discharge and the samples were etched downstream. The atomic concentration was measured by an isothermal calorimetric detector. Pressure dependence studies for molecular Br₂ etching showed that below 1-2 Torr of bromine, a first order reaction was rate-limiting whereas above this pressure a half order reaction was rate-limiting. Temperature dependence studies for the low pressure and high pressure regimes gave activation energies and pre-exponential values for the two respective rate controlling reactions. The first order reaction was found to have an activation energy of 29.2 ±4.0 kJ mol⁻¹ and a pre-exponential value of (3.4 ± 4.4) x 10²¹ molecule cm⁻² s⁻¹ Torr⁻¹. The activation energy for the half order reaction was found to be 8.4 ± 0.7 kJ mol⁻¹ with a pre-exponential of (6.4 ± 1.3) x 10¹⁸ molecule cm⁻² s⁻¹ Torr⁻¹⁄². The activation energy for atomic etching was calculated to be 12.9 ± 0.9 kJ mol⁻¹ and the pre-exponential, (7.1 ± 2.0) x 10²⁰ atom cm⁻² s⁻¹ Torr⁻¹. / Science, Faculty of / Chemistry, Department of / Graduate

A numerical analysis of radiative heat transfer in a liquid encapsulant Czochralski gallium arsenide crystal puller is developed. The heat transfer and equivilent ambient temperature of each surface element are calculated using the Gebhart radiative model. The effective ambient temperature, to which each surface element is radiating, is found to vary indicating that assuming a constant ambient temperature for all surfaces (simplified radiative model) is incorrect. The importance of including the middle and top cylinders of the growth chamber in numerical analysis of radiative heat transfer in the system is evaluated in the study. The upper section could be replaced by one isothermal surface without significant change of the effective ambient temperature distribution. Fluid flow and heat transfer in the GaAs melt, crystal and encapsulant are calculated using a three dimensional axisymmetric finite difference code which includes the detailed radiative model. The mathematical modelling of the fluid and heat flow describes steady state transport phenomena in a three dimensional solution domain with latent heat release at the liquid/solid interface. The predicted flow and temperature fields using the detailed radiative model differ considerably from the predicted fields using the simplified model. The simplified model shows high axial and low radial temperature gradients in the crystal near the encapsulant region; the axial gradient decreases and the radial gradient increases with increasing distance from the encapsulant top. The detailed model shows a high radial temperature gradient in the crystal near the crystal-encapsulant-ambient junction and nearly flat isotherms in the top half of the crystal. / Applied Science, Faculty of / Mechanical Engineering, Department of / Graduate

Work is reported on three topics relating to problems which hold back the development of GaAs integrated circuits. These topics are deep trapping levels in the starting semi-insulating GaAs, the effect of dislocations on device characteristics, and backgating. (The latter is the influence of voltages on nearby contacts on device performance). Deep trapping levels in undoped semi-insulating liquid-encapsulated— Czochralskl GaAs grown in the <100> direction were characterized using photocurrent deep level transient spectroscopy (photocurrent-DLTS). Three electron levels were found using photocurrent-DLTS in the temperature range 200K to 400K. By using Cr electrodes, instead of Au-Ge electrodes, it became possible to extend the experiment to a higher temperature range than previously used in this laboratory, and hence to observe the trap known as EL2. This trap has not previously been observed in undoped liquid-encapsulated-Czochralski GaAs by using photocurrent-DLTS. The possibility of an effect of dislocations on device characteristics was investigated using a dislocation etch procedure and measurements on an array of MESFET. Due to problems in controlling the fabrication processes, the scatter was such that no correlation between device characteristics and the distance to nearest dislocation would be established. However, scatter of threshold voltage was larger for devices fabricated on areas of honeycomb-like dislocations network as opposed to areas with unconnected wavy lines of dislocation. Backgating (which causes unwanted communication between devices) was investigated in conjunction with substrate conduction measurement. A model was proposed for the effect as present in the devices used in this experiment. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

No description available.

Note:

No description available.