Optical interference studies of graphitized layers formed in diamond by ion implantation and annealing provided the data on the depth and the thickness of the layers as well as their optical parzmeters. The latter were found to be close to those of dispersed graphite. Cathodoluminescince analysis of implanted diamond samples gave evidence of vacancy migration over macroscopic distances (∼1001μ).

In this contribution we present a study aimed at comparing results of positron-lifetime and Doppler-broadening measurements on the wide-band-gap compound semiconductors ZnS, ZnSe, and ZnTe. To investigate the basic properties of intrinsic and radiation induced defects the samples were irradiated either with 3 MeV protons or 1 MeV electrons. The isochronal annealing was performed in an Ar atmosphere. It was found that electron and proton irradiation cause different changes in the positron annihilation characteristics. Several annealing stages were observed, related to the annealing of variously sized vacancy complexes.

Epitaxially grown GaAs (p- and n-type) and n-Si were bombarded with low energy Ar-ions. Current voltage measurements on Schottky barrier diodes fabricated on the sputtered p-GaAs (Sc) and n-Si (Pd) showed that the series resistance and ideality factor were increased as the Arion dose was increased. The respective increase and decrease in barrier heights of Sc/p-GaAs and Pd/n-Si diodes were attributed to the presence of donor-type surface states in the bombarded material. The barrier heights of Au Schottky diodes made on n-GaAs changed nonmonotonically with Ar-ion sputter voltage. Variations of barrier height in the 0-1 kV range were explained by the introduction of donor-type defects. We demonstrated that the introduction of high concentrations of continuous level defects above 1 kV resulted in Fermi level pinning to become the dominant mechanism for controlling the effective barrier of current transport. Our results have shown that Schottky barrier properties could be changed by controlled amounts by varying the bombarding ion dose or sputter voltage.

Deep level transient spectroscopy (DLTS) was used to examine the metastability of a defect configuration in epitaxially grown boron-doped p-type Si. We report the detection of a new metastable defect Hα2 in p-Si following room temperature alpha particle irradiation. DLTS measurements coupled with bias-on/bias-off cooling cycles were used to study the annealing and introduction kinetics of this metastable defect. After removing Hα2 by zero-bias annealing at room temperature, it was re-introduced by reverse bias annealing in the 240-265 K temperature range under predominantly first order kinetics. The energy level and apparent capture cross section, as determined by DLTS, were E,+ 0.43 eV and 1.4 × 10−15 cm2, respectively.

Two deep levels, located at Ev+0.26eV and Ec-0.44eV, in Al-implanted n-type samples and one at Ev+0.48eV in p-type samples have been observed by the deep level transient spectroscopy. The level of is identified as the shallower aluminum-acceptor. The 1.7 MeV electron-irradiation, used as a probe to distinguish the implantation induced deep-levels, induces at least six electron traps in the n-SiC and one hole-trap in the p-type material. The peak positions of these deep-levels in DLTS spectra are quite different from those induced by Al-implantation. This result suggests that various damages are formed after heavy ion (Al) and light particle (e) irradiation.

Ever since the introduction of the metal-oxide-silicon field-effect-transistor (MOSFET), the nature of mobile and trapped charge in the oxide layer has been studied in great detail. For example, contamination with alkali ions such as sodium, causing instability of the flat-band voltage, was a major concern in the early days of MOS fabrication. Another SiO2 impurity of particular interest is hydrogen, because of its beneficial property of passivating charge traps. In this work we show that annealing of Si/SiO2/Si structures in forming gas (Ar:H2; 95:5) above 400 °C can introduce mobile H+ ions into the SiO2 layer. These mobile protons are confined within the oxide layer, and their space-charge distribution is well controllable and easily rearrangeable by applying a gate bias, making them potentially useful for application in a reliable nonvolatile MOSFET memory device. We present speed, retention, endurance, and radiation tolerance data showing that this non-volatile memory technology can be competitive with existing Si-based non-volatile memory technologies such as Flash.

The chemical kinetics of mobile-proton reactions in the SiO2 film are also analyzed in greater detail. Our data show that the initial buildup of mobile protons during hydrogen annealing is limited by the rate of lateral hydrogen diffusion into the buried SiO2 films. The final density of mobile protons is determined by the cooling rate which terminates the annealing process and, in the case of subsequent anneals, by the temperature of the final anneal. To explain the observations, we propose a dynamical equilibrium model. Based on these insights, the incorporation of the proton generation process into standard semiconductor process flows is discussed.

We have extended by five the number of deep-level traps known to create dark current in charge-coupled device (CCD) image sensors. These include Mn, Pt, and three much weaker traps that are as yet unidentified. Using dark current spectroscopy (DCS) we show that the generation rates at 55°C range from 6400 electrons/s for Mn to only 2 electrons/s for the weakest trap, which lies 0.27 eV off mid-gap. These weak traps determine the bandwidths and resolution of the trap peaks seen in the dark current spectra.

Based on the charge redistribution effect, as observed by the present authors, and the earlier reported large lattice relaxation and persistent photoconductivity phenomena associated with the EL6 defect seen in doped, undoped, semiinsulating(SI) and low temperature grown GaAs LT-GaAs), it is suggested that this defect be classified as a DX-center. A tentative unified atomic model is proposed for all the native defects EL2, EL3, EL5, and EL6 observed in GaAs.

We report results of studying some natural properties of highly disordered mesoscopic systems which seem to be promising for elevating quantum device work temperatures up to 77-300 K. They are FET-type Si-MNOS (metal-nitride-oxide-semiconductor) structures with built-in charge concentrations being so high that the systems remain to be strongly disordered even at room temperature. Disorder of studied structures could be controlled by varying charged traps concentration at the SiO2-Si3N4 interface that induce strong potential fluctuations. Important feature of the structures is the possibility to vary the built-in charge over a wide range (up to 1013 CM−2) that results in varying the disorder range. The conductance of such a system is shown to be controlled by the single small quantum-sized region with a ballistic transport which is a saddle-point region of the fluctuation relief. Narrowness of that “bottle neck” (comparable with the electron wavelength) results in quantizing conductance of the structure, and if the disorder is high enough, the conductance for some gate voltages shows a real tendency to reach a plateau at the quantum value e2/h. What is important, that tendency occurs a∼t high temperatures (77-300 K).

We have investigated the surface quality of thick 10μm Silicon-On-Insulator (SOI) wafers by comparing CMOS device performance and reliability of SOI wafers with those of bulk wafers. Large scale CMOS circuits, microcontrollers with 12Kbyte ROM and 512byte EEPROM, were fabricated on thick SOI wafers and on bulk wafers, and device performance and reliability were evaluated by means of functional tests and acceleration tests. A significant difference was observed only in the ROM bit-map failure analysis in the functional tests, in which the number of single bit failures (defined as a chain of up to 4 adjacent bit failures) was more evident on ROM fabricated on the thick SOI than that fabricated on bulk wafers. The result indicates that the crystal defects introduced by the initial oxygen concentration and excess film stress of the SiO interlayer may cause single bit failures. The SOI wafers were prepared by Direct Wafer Bonding 2(DWB) and a mechanical polishing technique to form a 10μm active silicon layer and a 625μm silicon substrate, which sandwiches a 0.25μm SiO2 interlayer. Thick SOI wafers such as these are becoming a promising material for one-chip integration of CMOS devices and Micro Electrical Mechanical System (MEMS) devices.

Solid C60/GaAs contacts were fabricated by growing solid C60 films on both n-type and p-type GaAs(100) substrates through vacuum deposition. The electronic states at C60/GaAs interfaces were studied. Both C60/n-GaAs and C60/p-GaAs contacts were found to be strong rectification junctions with a rectifying ratio higher than 106 at a bias of ± 1.OV. Two distinct traps were also observed at the C60/GaAs interfaces with deep level transient spectroscopy (DLTS); one is the electron trap at 0.35 eV below the GaAs conduction band and the other is a hole trap at 0.45 eV above the GaAs valence band.

Improvements in the techniques for the growth of Si substrates, used for commercial solar cells, have yielded wafers that exhibit low average defect density × typically less than 105 cm−2. We have observed that low defect density leads to the formation of defect clusters. This defect configuration influences the device performance in a unique way × by primarily degrading the voltage-related parameters. We discuss the nature of the defect clusters and show that they constitute regions of high carrier recombination. Network modeling is used to show that, in a device, these regions act as shunts that dissipate power generated within the cell.

This paper reports the effects of oxygen in a Si substrate on dryetching damage during gate-electrode formation and side-wall formation. Different types of deep level traps were created in Czochralsi (CZ) and in p/p+-epitaxial wafers during side-wall formation. In comparison to the CZ wafers, a deep energy state and large hole capture cross section for the electrically active trap appears in the p/p+-epitaxial wafers. The depth profile of the deep trap in CZ wafers is shallower than that in p/p+-epitaxial wafers. This result indicates that the trap diffusivity in a CZ wafer is smaller than that in a p/p+-epitaxial wafer. Therefore, it suggests that creation of the trap in CZ wafers is influenced by the oxygen within the silicon. From results, the origin of the trap was identified to be carbon-oxygen complexes in the CZ wafer and carbon complexes in the p/p+-epitaxial wafer. Thus, the oxygen in the Si substrate plays a role in the formation of carbon-oxygen complexes and in suppressing the trap introduced into the Si substrate during SiO2 etching.

We report on the electrical properties of defects introduced in epitaxially grown n-Si by 1 keV He-, Ne-, and Ar-ion bombardment. Epitaxial layers with different O contents were used in this study. We demonstrate using deep level transient spectroscopy that the low energy ions introduced a family of similarly structured defects (DI) with electronic levels at ∼0.20 eV below the conduction band. The introduction of this set of identical defects was not influenced by the presence of O. Ion bombardment of O-rich Si introduced another family of prominent traps (D2) with levels close to the middle of the band gap. Both sets of defects were thermally stable up to ∼400 °C, and their annealing was accompanied by the introduction of a family of secondary defects (D3). The “D3” defects had levels at ∼0.21 eV below the conduction band and were thermally stable at 650 °C. We have proposed that the “DI”, “D2”, and “D3” defects are higherorder vacancy clusters (larger than the divacancy) or complexes thereof.

CF4, SF6 and CF 4/O2 plasma-induced damage in silicon-doped n-type GaAs crystals have been investigated by using Hall measurements, and SIMS analysis. We observed that the carrier density degraded with post annealing after plasma exposure. The degree of deactivation of carrier density with post annealing was strongly dependent on the plasma chemistries and the self-bias Vdc of the plasmas. From SIMS analysis, fluorine contamination was observed in the surface region of GaAs crystals after CF4 plasma exposure. The internal diffusion of fluorine atoms and the localization of fluorine atoms in n-type GaAs layers were observed after post annealing at 400°C. From these results, we proposed a mechanism of plasma induced damage in which internally diffused fluorine atoms neutralize the donor silicon by forminq a Si-F bond. We determined the effective diffusion coefficient of fluorine in GaAs as 1.5 × 10−11 cm2/s (at 400°C).

Low energy bombardment of silicon with noble gas ions generates a family of defects with characteristic low-temperature photoluminescence spectra. The defect distribution up to one gtm below the sample surface indicates an unusually enhanced migration mechanism during the bombardment of the specimen. Annealing the samples above 500 °C leads to the disappearance of the photoluminescence spectra and the formation of new, more extended defects, which are excellent gettering sites for transition metals, in particular copper. Systematic calculations of noble-gas vacancy interactions imply a model for the photoluminescing defect where the noble gas atom is trapped by a divacancy.

In the process of etching contact and via openings with a downstream oxide etcher, plasma etch byproducts form polymer that collects on process chamber components. Polymer build up over time will begin to separate from chamber components and deposit on the wafer, blocking the etch and creating incomplete contact and via openings, which result in device failure. When this phenomena occurs, complete disassembling and cleaning of the chamber (wet clean) is required before particle levels are acceptable again for processing wafers. Over half of the time wet cleans are performed at or before 1500 wafers processed.

Experimentation was performed to isolate one particular chamber component causing particle deposition on the wafer. Once a chamber component was isolated, possible solutions were designed so that original process parameters were not altered. The first experiment was a maintenance procedure to disassemble and clean one chamber component at set intervals between wet cleans. The second was a no wafer, high concentration Oxygen plasma chamber clean.

It was found that changing a particular chamber component after 1250 wafers processed, and then again after 2500 wafers, made it possible to run to 3750 wafers before performing a wet clean. Overall particle levels were also reduced by 30% with the plasma cleans after every 100 wafers processed. The combined solutions resulted in a greater than 30% reduction in overall particle counts. Also, maintenance down time for particle failures was reduced from 57% to 19%. During the time frame of implementation of these solutions, overall factory defectivity of finished product wafers was reduced up to 25%, which relates to a 15 to 25% yield increase. The reduction of defects on the downstream oxide etcher is a contributing factor to device yield improvements.

GaAs is treated with remote PH3 and N2 plasmas. Electron traps induced by plasma treatments are investigated by isothermal capacitance transient spectroscopy measurements. The EL2 trap is detected in the as-grown GaAs. The TP1 trap(Ec-0.26eV) is generated in GaAs phosphidized for 10min, while the TN1 trap(Ec-0.66eV) is induced in GaAs nitrided for 30min. It is found that the TP1 trap is changed to the another trap with an energy level as shallow as 0.16eV below the conduction band edge and a capture cross section as small as 1.8×10−21cm2 by treating with N2 plasma subsequently after PH3 plasma treatment.

We review our two models for the impurity excitation of the oxygen in silicon. The first model aims to explain the energies and intensities of the infrared absorption lines. The coupled excitation is formulated for the first time under the D3d symmetry. Taking into account the lowenergy excitation of the oxygen, the A2u local mode, and the anharmonic coupling between them, we quantitatively describe the absorption lines in the 30-, 1100-, and 1200-cm−1 bands, including those of the 18O isotope. We next take into account the A1g local mode, the Eu resonant mode, and the couplings between the excitation elements. This enables us to qualitatively explain the origin and characteristic aspects of other absorptions, e.g., the 500- and 1700-cm−1 bands. The second model is a simplification of the first one with the coupling to the band (extended) phonons included. The model describes the strong temperature dependence of the line widths observed for both of the phonon-resonant 30-cm−1 band and the phonon-off-resonant 1100-cm−1 band. The life-time limiting process causing the finite line-widths is shown to be the one-phonon transitions in the energy-level ladders of the low-energy excitation each belonging to the ground and the first excited states of the A2u local mode. The vibrational excitations of the similar impurity systems in Si, Ge, Ge xSi1−x crystals are analyzed.

A method for measuring denuded zone depth from lifetime measurements is proposed and compared to observations by microscopy techniques and to Electron Beam Induced Current (EBIC) analyses. The surface photovoltage (SPV) method is chosen for measuring lifetime, as it is found to be extremely sensitive to oxygen precipitation. Samples prepared according to three typical thermal hystories are used as test vehicles.

Correlating the results given by the different techniques, this study points out a few relevant features from the point of view of wafer processing. For instance, processes starting with a low temperature treatment are poorly reproducible for what concerns oxygen precipitation and denuding, as shown by SPV data, and do not guarantee a defect-free denuded zone, as seen in EBIC images. Additional high temperature treatments during the process increase the denuded zone depth and improve uniformity.

Techniques which rely on a selective etching for defect detection, such as the Secondary Electron method of Scanning Electron Microscopy (SEM) and Atomic Force Microscopy (AFM), are limited by etching sensitivy. On the other hand, the Transmission Electron Microscopy (TEM) statistics is too poor for this sort of observations. Instead, as a technique based on the recombination activity of defects, EBIC has roughly the same sensitivity as SPV. Both these techniques detect the formation of a denuded zone even when other methods fail, and can be considered as complementary tools.