S-terminated InP(100) surfaces which are chemically clean and stable were obtained by sulfide solution treatments. X-ray photoelectron spectroscopy (XPS) and low energy electron diffraction (LEED) were used to investigate the thermal and chemical stability of the surfaces. Schottky diodes based on S-passivated InP(100) wafers were used as test devices tostudy the surface electrical properties.

X-ray photoelectron spectroscopy, X-ray photoelectron diffraction, and S K edge X-ray absorption near-edge structure were used to characterize the air-stable S-passivated InP(100)-(lxl) surface prepared in (NH4)2S solution. The results show that one monolayer of sulphur which is bonded only to In is adsorbed on the surface. The S is found to occupy bridge-bonded sites, and the orientation of the In-S bond is determined to be in the [011] azimuth, with an In-S-In bridge-bond angle of about 100°.

A low temperature preheating process is developed for metalorganic chemical vapor deposition (MOCVD) growth of GaAs on wet chemical pretreated Si substrates. NH4 OH/H2 O2 is found to be most effective in decreasing the preheating temperature among the chemicals we tried: NH4 OH/H2 O2, H2SO4 /H2O2, or hot HNO3. By using NH4OH/H2 O2, the preheating temperature is reduced from 1000°C to 875°C. X-ray diffraction measurements and surface observations with an atomic force microscope (AFM) show that the GaAs film quality obtained with the 875 °C preheating process is better than that obtained with 1000°C preheating.

We have investigated the effect of GaAs surface conditions prior to plasma enhanced chemical vapor deposition of a silicon nitride cap on the activation efficiency of implanted Si in GaAs. The oxygen plasma treatment improved the activation efficiency of implanted Si by ∼35% over (1:10) NH4OH:H2O treatment. X-ray photoelectron spectroscopy (XPS) analysis of the oxygen plasma treated GaAs surface indicated the formation of ∼25Å thick oxide layer consisting of Ga2O3, As2O3, As2O5 and elemental As. During the activation anneal, the arsenic-containing oxides react with the GaAs substrate to form Ga2O3 and elemental As. The presence of excess As between the GaAs and the nitride cap film increases the probability that the implanted Si incorporates in the Ga sites over the As sites, and thereby improves the activation efficiency. This surface-related mechanism suggests that the variation in activation efficiency is mostly attributed to variation in surface conditions, and may explain the wide variety of reported values of activation efficiency.

The efficacy of residual photoresist removal on the top surface of the InGaAs QWW grating and the effects of surface oxides on the optical property of quantum well wires (QWWs) were examined through atomic force microscopy (AFM) and photoluminescence (PL) spectroscopy. Different resist removal treatments, including acetone, ozone and diluted HC1 were evaluated. Both AFM and PL measurements reveal that with the surface cleaning processing we have developed, high luminescence efficiency from the QWWs is conserved after removal of the residual photoresist.

A passivation processes using Na2S and photochemical washing of GaAs (100) surfaces was studied in real time by a second-harmonic generation (SHG) technique. The intensities of surface-specific SHG signals were compared with those of photoluminescence (PL) signals. We found a remarkable similarity between the SHG and PL intensity changes during these processes. A band-bending model due to Fermi-level pinning at the surface has been applied in order to account for both the SHG and the PL intensity changes.

Contamination level of chemically cleaned (100) GaAs wafer surfaces were measured by Total reflection X-Ray Fluorescence (TXRF). Cu and Zn were detected as major metallic Impurities on GaAs surfaces. The Cu contamination level was in the range of 1011 atoms/cm2 on wafers treated with NH4 OH/H2O2 /H2O (80:1:400) mixture. On the other hand, Cu could be reduced to less than 5×1010 atoms/cm2 and Zn could be controlled below the detection limit by treating with a choline/H2O2 /H20/Na5P3O10 or SEMICO-CLEAN 23/H2O2 /Na5P3O10 solution. Surface cleanliness depended on the H2O2 concentration and whether or not Na5P3O10 was present in the organic base solutions.

The incorporation of triple crystal x-ray diffraction (TCD) under conditions of a nearly grazing incident x-ray beam provides a unique capability to characterize surface and subsurface structural damage. The effects of bromine-methanol chemical-mechanical (CM) polishing on the surface quality of GaAs were investigated by inclined Bragg plane triple crystal x-ray diffraction. GaAs samples polished under varying conditions of Br2 concentration, total polish time, polish wheel rotation speed, and force exerted on the samples were analyzed using TCD profiles from inclined (220) planes, with the samples oriented at both the exact Bragg condition, θB, and at deviations from the exact Bragg condition. As the deviation from the 220 reciprocal lattice point was increased, multiple peaks corresponding to both the dynamic and the kinematic components were observed. The peak positions and the relative intensities of the various peaks were used to characterize the effects of the CM polishing variables on the GaAs surfaces.

A unique process for texturing InP (100) wafers by anisotropic etching has been developed. The process produces irregular V-grooves on the surface, which reduce the surface reflectivity. The process does not require photolithography or masking. The etching characteristics depend on doping, with etching tending to proceed more rapidly on the more heavily doped samples. Reduced reflectivity surfaces formed using this process can be applied to solar cells, photodetectors, and other optoelectronic devices.

HF etches silicon oxide, removes metal impurities and completely terminates the silicon surface with hydrogen. The H-passivated Si surfaces are stable, but can be slowly and selectively etched in buffered HF solutions or in hot water, changing the initial morphology of the interface. Si(111) surfaces flatten, forming large, ideally H-terminated (111) planes. Si(100) surfaces roughen, ultimately developing H-terminated (111) facets.

Vapor phase cleaning of silicon wafers is reviewed. Particular emphasis is placed on oxide etching and removal, including the mechanisms involved in vapor HF etching of silicon oxides. Other items discussed include native oxide formation, impurity removal, and device applications. Future directions involving other chemistries and integrated processing are summarized.

State-of-the-art first-principles calculations allow detailed studies of the mechanisms by which hydrogen and fluorine interact with silicon. The results for hydrogen are presented in the form of an energy diagram which includes many different configurations. The theoretical values allow a discussion of issues such as hydrogen solubility, and desorption from a Si surface. For fluorine, we investigate the behavior as an interstitial impurity in the bulk, as well as Si-F interactions at or near the surface. A study of the insertion of F atoms into Si-Si bonds elucidates the microscopic mechanisms of etching, and the dependence of etch rate on doping. Thermodynamic aspects of HF etching are briefly discussed.

HF-treated Si surfaces and the oxidation kinetics in pure water or in clean room air have systematically been studied by x-ray photoelectron spectroscopy (XPS). The oxidation of heavily-doped n-type Si appears to proceed parallel to the surface, resulting in the layer-by-layer oxidation. The oxide growth rate in pure water for heavily-doped n-type Si is significantly higher than that of heavily-doped ptype Si. This is explained by the electron tunneling from the Si conduction band to adsorbed O2 molecules to form the O2 state. O2 ions easily decompose and induce a surface electric field, enhancing the oxidation rate. The growth rate of native oxide on heavily-doped n-type Si is less sensitive to the crystallographic orientations than the case of lightly doped Si where the steric hindrance against oxygen molecules significantly lowers the oxidation rate of the (110) and (111) surfaces. We suggest that the decomposed oxygen can penetrate into Si without steric hindrance. It is also found that the oxidation of heavily-doped n-type Si in pure water is effectively suppressed by adding a small amount (10 ∼ 3600 ppm) of HCI.

The characteristics of the HF-treated Si-surface are investigated as a function of dipping time in dilute HF solutions. It is found that the contact angle is a very sensitive measure for the degree of oxidation of the Si-surface. The importance of obtaining a perfectly passivated surface in order to reduce the particle deposition on the surface is shown. HF-last cleans are found to be beneficial in terms of metallic contamination and gate oxide integrity. The importance of the loading ambient in furnaces is investigated after HF-treatments and RCA-cleans.

Changes in the surface concentration of Fe, Ni, Cu and Zn on Si wafers processed in a NH 4OH / H2O2 / H2O (SC1) solution was investigated. The adsorption of Fe or Zn on Si wafers was not affected by the other impurities in the SC1 solution. However, the adsorption of Ni or Cu on Si wafers was obstructed by Fe, and was not obstructed by Zn.

The effects of dopant type and dopant concentration on the native oxide growth in air on the silicon surface were investigated. The oxide thickness was measured by X-ray photoelectron spectrometry (XPS). The oxide was thicker on n-type Si than on p-type Si in early oxidation. The oxide increased linearly with the dopant concentration. This enhancement of oxidation was assumed to be caused by vacancies near the surface in the silicon bulk.

Scanning tunneling microscopy (STM) was made in order to examine the surface structure and roughness in an atomic scale. The surfaces were prepared by several ways: NH4F (pH = 8) dipping just after RCA cleaning or after keeping in dry air for a few weeks; dipping into NH4F (pH = 8) solution or dipping into solutions with pH=10 just after boiling in HNO3. The STM images clearly showed that the surface structure and roughness are dependent on the sample treatments. The smooth surfaces with less defects were obtained for surfaces prepared by removing the HNO3-oxidized layer by NH4F (pH = 8) dipping.

The nature of light-emitting porous Si layers produced by non-anodic stain etching of p-type (100) Si substrates is studied. The layers were characterized by transmission electron microscopy as being amorphous in nature. X-ray photoelectron spectroscopy and electron spin resonance measurements show these layers to be composed mainly of a-Si. The formation mechanism of the a-Si is explored using by stain etching SiGe ‘marker’ layers within epitaxially grown Si films and by high temperature annealing. These experiments provide strong evidence for a spontaneous crystalline-amorphous phase transformation during the etching process.

The valence band and the Si 2p core level of electrochemically etched light emitting porous silicon samples prepared with different etching parameters and upon thermal annealing have been studied. The core level shows appreciable broadening as chemical etching time increases. Upon annealing, the core level linewidth decreases and the valence band develops spectral features. The photoluminescence intensity, also, decreases upon annealing. However, impurity species are found to be present only in trace amounts. Our data is consistent with a photoluminescence mechanism involving a silicon species that degrades, decomposes, or desorbs upon annealing.

Surface Charge Analysis (SCA), and ellipsometry have been used to study the stability over time of HF treated (100) silicon surfaces as a function of the post-HF rinse time. Using SCA, the electrical properties of the chemical terminating layer of these silicon surfaces were measured. The surfaces which remained native oxide free the longest (−10 hours) had very low Qox and Dit values on the order of 1.0 × 1011/cm2 and 5.0 × 1010 eV−lcm−2, respectively. A good correlation was found between Dit and the native oxide thickness measured by ellipsometry. This and other results are discussed in terms of the chemical bonding on the silicon surfaces.