We have used in situ Auger spectroscopic analysis to investigate the composition of InP surfaces cleaned in rf H2 plasmas and etched in rf H2/CH4/Ar plasmas. In general agreement with previous results, hydrogen plasma treatment is found to remove surface carbon and oxygen impurities but also leads to substantial surface phosphorus depletion if not carefully controlled. Low plasma exposure times and rf power settings minimize both phosphorus depletion and surface roughening. Surfaces etched in H2/CH4/Ar plasmas can show severe phosphorus depletion in high density plasmas leading to etch rates of ∼ 700 Å/min, but this effect is greatly reduced in lower density plasmas that produce etch rates of 30–400 Å/min.

In this work, we studied the removal of the native oxide on polycrystalline CuInSe2 thin films by KCN and effect of subsequent chemical sulfurization with (NH4)2Sx solution on these films. As a result of the treatment, a portion of the selenide film was transformed into Culn(S,Se)2 The Auger Electron Spectroscopy and X-ray photoelectron spectroscopy studies showed that KCN removed the oxygen and the sulfurization prevented regrowth of the oxides. The optical bandgap of the sulfurized films increased about 0.27 eV. From these experiments, we concluded that sulfur atoms are incorporated in the CuInSe2 to form a stable and higher bandgap layer, Culn(S,Se)2 which may passivate the CuInSe2 and improve the performance of polycrystalline photovoltaic cells made from it.

The effect of arc plasma jet treatment (APJT) of silicon surface used for pre-gate oxidation cleaning on the electrophysical parameters of MOS structures (Si/SiO2/Si*/Al) has been studied- We show that APJT etching cleaning considerably improves the constant current charge to breakdown of MOS structures in comparison with conventional wet chemical cleaning. We have analyzed the effect of plasma cleaning conditions on the quality of gate oxide and SiO2/Si interface.

ZnxCd1-xTe is a widely used substrate for the epitaxial growth of HgCdTe, which is used in infrared detectors. Results of the effect of sputtering of ZnxCd1-xTe single crystals with low energy Ar beam are reported in this paper. X-ray photoelectron spectroscopy (XPS) and photoluminescence (PL) techniques were used to measure the concentration of Zn in these crystals. Selective sputtering of Zn atoms has been observed from freshly cleaved crystals using XPS studies. Sputtering is a common method of cleaning ZnxCd1-xTe crystals in their device preparation and our studies show that this method of cleaning alters the surface which may introduce lattice mismatch on the surface. Surface morphology before and after cleaving the crystals is studied using Atomic Force Microscopy (AFM).

Surface properties of GaAs passivated with (NH4)2Sx solution have been compared with HCl-treated GaAs using X-ray photoelectron spectroscopy, Sulfur treatment on GaAs surface results in the formation of S-Ga and S-As bonds, which remain after successive rinsing for I minute in DI water. The evolution of Ga 2p3 and As 3d peaks in the sulfidation treated GaAs was monitored with the exposing time to air. After 10 days exposure to air, the Ga-O and As-O bonds slightly increased, but maintained almost constant for further exposure. The increase of Ga-O and As-O bonds induces the partial decomposition of sulfur bonds. Decomposition and evaporation behaviors of sulfur and oxygen were observed through the heat treatment of sulfidation treated GaAs under ultra high vacuum (less than l × 10-9 torr). After anneal at 350 – 450 °C, slight decrease of sulfur and oxygen due to the decomposition of As-O bond were observed. No more sulfur was found after anneal at 550 – 650 °C, where the decomposition of Ga-O bond was completed.

By liquid phase epitaxy (LPE) we have grown silicon layers on silicon and partially masked silicon at temperatures below 450 °C from Ga and Ga-In solutions. Oxidation of the cleaned silicon substrate surfaces before epitaxial growth has been prevented by a buffered hydrofluoric acid treatment. The epitaxial layers reached a thickness of 7 jim and were free of extended defects.

Low growth temperatures make it possible to grow silicon layers also on pre-treated glass substrates. The amorphous glass is first coated with a thin nano-crystalline silicon layer which is deposited by plasma processes from a mixture of SiH4/H2 gas. The grains in the silicon layers grown from Ga solution on glass have reached sizes up to 100 μm.

Different substrate cleaning procedures were used before fabrication of pin diodes by silicon molecular beam epitaxy (MBE). We investigated the quality of these diodes in order to demonstrate the superior quality of a low energy plasma cleaning in an ultra-high vacuum ( UHV). This plasma cleaning by hydrogen makes a wet-chemical cleaning or a high-temperature desorption step unnecessary. Moreover, the plasma-cleaned substrates are so strongly hydrogen passivated, that they can be transported through air and processed in another MBE chamber without any additional cleaning steps.

This study demonstrates the cleaning of Si(100) surfaces with a remote H2/SiH4 plasma. The surfaces were prepared with a chemical oxide the remains after an RCA clean. The plasma cleaning process was designed to remove contaminants such as C, F, and SiO2. The key to successful removal of the oxide is to have the plasma chemistry in a neutral deposition regime. The neutral deposition process regime is a balance between the deposition of Si by SiH4 and the etching of the deposited Si by atomic H. During the neutral deposition mode the SiO2 was removed without deposition of Si on the SiO2 surface. Once the SiO2 layer is removed, theunderlying Si surface is exposed to the H2/SiH4 plasma and a thin epitaxial film may be deposited. The final Si surface configuration after plasma cleaning is a 2×1 hydrogen terminated surface. The characterization of the interface and epitaxial film were investigated using Auger electron spectroscopy (AES) and transmission electron microscopy (TEM).

Fluorine contamination was removed from a Si(100) surface by an atomic H flux. The surface was intentionally contaminated to approximate the residual fluorine concentration remaining after a concentrated HF last process. By dipping the wafers in concentrated HF the thin oxide was removed and replaced with a hydrogen and fluorine terminated surface. This surface was then either vacuum annealed or exposed to a 20 Watt rf excited H-plasma at 50 mTorr, in order to achieve an atomically clean surface. The substrate temperature during the Hplasma exposure and vacuum anneal was 450°C. The surface chemistry was characterized with x-ray photoemission spectroscopy (XPS), auger electron spectroscopy (AES), and angle-resolved UV photoemission spectroscopy (ARUPS). The surface symmetry was characterized with low energy electron diffraction (LEED). Before the H-plasma exposure, the XPS spectra indicated Si-F bonding, and a l×1 LEED diffraction pattern was observed. Immediately following the Hplasma exposure, the fluorine concentration was reduced below detection limits of XPS, and the surface showed a 2×1 reconstruction. A mechanism is proposed by which molecular HF results from atomic hydrogen interactions with fluorine on the surface.

In situ scanning tunneling microscopy (STM) was used to examine the etching process of n-Si(11) and Si(001) electrodes in dilute NH4F under cathodic potential control. For Si(111), time-dependent STM images have revealed the pronounced effect of the microscopic structure of surface Si atoms on their dissolution rates. The multiple hydrogen bonded Si atoms at the kink sites and dihydride steps eroded faster than the monohydride terminated Si. Presumably, the higher polarity at these defect sites is responsible for the difference. Steric consideration further favors the higher activity at the more open kink sites. The monohydride terminated Si(111) surface represents the most stable surface structure, which guides the dissolution of the Si(001) surface to the formation of {111} facets. The initial stage {111} facet formation on a Si(001) surface was revealed by in situ STM.

The use of Conducting Probe Atomic Force Microscopy to give nm scale electronic characterisation of surfaces is reviewed. Local conductance, Kelvin Probe work function measurements, Fowler-Nordheim tunnelling and local C-V characterisation techniques are outlined. The principle results of these and their applications to the semiconductor surface and thin film characterisation are discussed. We present tunnelling data from silicon through varying oxide thickness using conducting AFM and scanning Kelvin Probe measurements from sub micron MOS capacitors. The F-N tunnelling technique has also been used on epitaxial silicon surfaces with atomically flat topography.

The inherent problems associated with quantitative, reproducible measurements are outlined, and the potential applications of the measurements to surface and thin film technology are discussed.

Microroughness is a critical parameter in ULSI device interface reliability and has been shown to effect several critical MOS electrical properties. The atomic force microscope (AFM) has become the instrument of choice for silicon surface microroughness analysis. The parameters usually specified to characterize roughness are average and root mean square roughness. However, these parameters are spatial averages and can have the same value for two significantly different surfaces. Spectral analysis using the Fast Fourier Transform (FFT) has been applied as a powerful tool to analyze AFM data by looking at roughness as a function of spatial wavelength. The Fast Hartley Transform, being a real transform, is faster than the FFT and is better suited for this analysis. It has been used here to derive spectral information from the AFM height data. Before evaluating the transform, cancellation of any tilt or warp in the AFM data is done to remove frequency components which interfere with other spectral information. A PC-based computer program to determine the transform and its magnitude will be described. The application of this method to analyze data from Si and SiO2 surfaces as a function of pre-oxidation cleaning chemistry will be presented. Significantly better insight into the spatial distribution of roughness is obtained, when compared to previous implementations.

The carrier recombination rate in high-quality FZ and Cz silicon substrates is studied by contactless infrared and microwave absorption techniques. Different surface treatments covering a wide range of surface recombination velocity have been used for the separation of bulk and surface recombination components and evaluating of the efficiency of passivation. Limitations of effective lifetime approach are analyzed specific for low and high injection level. Sensitivity limits of the techniques for iron contamination are discussed

The technique of external reflection infrared (IR) spectroscopy is used to study silicon surface chemistry. External reflection is enhanced by implanting a buried cobalt silicide layer in silicon to act as an infrared reflector. The preparation of clean well-ordered surfaces from the ion implanted substrates is demonstrated. The reactions of water and ethanol with Si(100) are investigated.

In this work, a new commercial system allowing non-contact measurement of the surface charge is used to monitor the condition of the silicon surface following HF/water etch. Results obtained demonstrate that by monitoring changes of surface charge using this system, a truly non-invasive, instant and easy to carry out characterization of Si surfaces after HF/water etch can be accomplished. The results show that HF/water exposure adds positive charge to the silicon surface. Change in the surface charge, considered to be indicative of the change in the electro-chemical condition of the surface, appears to precede initiation of the oxide etching process, and is proposed to be a factor in initiating etching reactions that involve mainly negatively charged species.