Results in the literature indicate that C12 etches GaAs at room temperature but HCl etches GaAs at a measurable rate only at temperatures above ∼670 K. In this work, molecular beam scattering and surface analysis techniques have been applied to address the fundamental kinetic differences between these two systems. The results indicate that the onset of GaAs etching by C12 is determined by the kinetics of Ga-removal as GaC13 while etching by HCl is limited by As evaporation as As2. The results also suggest that HCl selectively etches gallium from GaAs at temperatures between 600 and 650 K.

The initial stage of the reaction of aluminum trichloride (AlCl3) with the Si(111)-(7×7) surface and the annealing effects of the adsorbed surface are investigated with a scanning tunneling microscope (STM). Reacted and unreacted sites manifest in the contrast of adatom sites on the AlCl3-exposed surface. An AlCl3 molecule dissociatively adsorbs onto the Si( 111)-(7×7) surface at room temperature, which yields C1 atoms. Preferential adsorption site is found to be the center adatom site rather than the corner adatom site, and Cl atom adsorbs onto the top site of the center Si adatom. The migration of adsorbed molecules (AlClx) in the (7×7) unit cell is observed at room temperature. The SiCl,, species desorb from the surface by thermal annealing at 600 'C, leaving vacancies behind. After annealing at 1200 0C, Al deposition occurs in a limited area and the surface shows the V/7×V7 reconstruction, which is a characteristic Al-induced surface structure.

This paper describes x-ray photoelectron spectroscopy (XPS) studies of etching of Si in high-density Cl2 plasmas. Polycrystalline Si films, masked with photoresist stripes, are etched and then transferred in vacuum to the XPS analysis chamber. Shadowing of photoelectrons by adjacent stripes and differential charging of the photoresist and poly-Si were used to separate contributions from the top of the mask, the side of the mask, the etched poly-Si sidewall, and the bottom of the etched trench. In pure Cl2 plasmas, surfaces are covered with about one monolayer of Cl. If oxygen is introduced into the plasma, either by addition of O2 or by erosion of the glass discharge tube, then a thin Si-oxide layer forms on the sides of both the poly-Si and the photoresist. Laser-induced thermal desorption (LITD) was used to study etching in real time. LITD of SiCI was detected by laser-induced fluorescence. These studies show that the Si-chloride layer formed during plasma etching is stable after the plasma is extinguished, so the XPS measurements are representative of the surface during etching. LITD measurements as a function of pressure and discharge power show that the etching rate is limited by the positive ion flux to the surface, and not by the supply of Cl2, at pressures above 0.5 mTorr and for ion fluxes of σ4× 1016cm−2 s−1.

The etching of polymer films in oxygen-based plasmas has been studied between 5 and 100 mTorr in a reactive ion etch reactor using Langmuir probe and optical actinometry measurements. Results for the etch yield (the number of carbon atoms removed per incident ion) are analyzed in terms of a surface-chemical model for ion-enhanced etching proposed by Joubert et al. (J. Appl. Phys. 65, 5096 (1989)). A proper description of the results requires that this model be modified by including a term due to direct reactive ion etching and physical sputtering. The contribution by direct reactive ion etching to the overall etching turns out to be significant under all conditions and even dominant at the lowest pressures. The modified model should be applicable to the etching of polymers in other types of reactors, especially highplasma- density reactors. The relationship between these results and the anisotropic patterning of polymer films is also discussed.

Dry etching of SiO2 insulation layer has been required in the Si semiconductor manufacturing process. The etching of SiO2/Si is chemically carried out by using HF solution. We successfully demonstrated a new method for exclusive etching of SiO2 using the nitrosyle fluoride (NFO) gas which was produced from the mixed gas of NF3 and O2 with an ArF excimer laser irradiation.

SiO2 and Si substrates were placed side by side in a reaction cell which was filled with 3% O2 gas in NF3 at the gas pressure of 380 Torr. ArF excimer laser beam was irradiated parallel to the substrates. The laser fluence was kept at lOOmJ/cm2. As soon as the mixed gas of NF3 and O2 was irradiated with the ArF laser beam, an intermediate product of NFO was produced. The chemical behavior of NFO was confirmed from the UV absorption spectrum with absorption in the 310 to 330nm wavelength region. In the presence of SiO2, the absorption of NFO diminished. The absorption of NO2, instead of NFO, appeared at 350nm. This indicates that the oxygen atoms of SiO2 were pulled out by NFO.

The etching reactions continued for 3 minutes after irradiation when the SiO2 and Si substrates were kept in an atmosphere of the reactant gases. As a result, not the Si but SiO2 substrate was etched with the depth of 2000Å.

Organosilicon precursors were used to deposit thin films of highly stable amorphous hydrogenated silicon carbide at low temperatures by plasma enhanced chemical vapor deposition. X-ray photoemission and Auger electron spectroscopy were employed to characterize the surface chemistry of air exposed and plasma etched films. Auger analysis of the as-deposited material shows its composition to be constant throughout its depth. RF plasma etching was performed in CF4/O2 and SF6/O2 gas mixtures. Etch rates in these atmospheres are similar to those reported in the literature. After plasma etching, the surface was converted to a highly fluorinated state. In addition to the expected SiC bond, Si-OF bonds were found after plasma etching. Fluorocarbon residue was not produced in this process. Chlorine etching of this low temperature PECVD film is described and is shown to be compatible with standard integrated circuit manufacture processing as an hermetic-like sealant.

Spatial confinement of mass-selected ions and extracting them to a solid surface at low kinetic energy is achieved by using rf ion trapping which is combined with multi-photon ionization by a KrF laser. SF5+ fragment ions of SF6 generated by photo-ionization are separated from other ionic species and then confined in the cylindrical ion trap cell for at least 100 ms. The stored ions are extracted onto the seniconductor surface. This technique is promising for clarifying surface reactions of semiconductors, which is the key to controlling surface structure on an atomic scale.

The state of the surface after energetic keV particle bombardment is investigated using molecular dynamics. The model utilizes a Ag{110} microcrystallite which is statically bombarded by Ar particles at normal incidence. After being bombarded at incident energy of 1 keV, the relocation probability is <0.3 for all the surface atoms initially residing within four lattice spacings from the target. The probability decreases exponentially as the initial distance of the substrate atom from the target is increased. The most probable distances of displacement from the lattice site also vary with the initial atomic distance from the target atom. The probable displacement of the surface atom, except the target atom, is less than one twentieth of the surface lattice spacing. An analytical formulae for the initial-distance dependence of the relocation probability is also proposed. The formula has three adjustable parameters which are determined by the least-squares method.

In this paper, we report our results on surface preparation for the growth of epitaxial Si films. Hydrogen passivated surfaces are currently being investigated for application in Si epitaxy to eliminate the high temperature in-situ bake necessary to remove the native oxide. Hydrogen passivation is obtained by a dilute HF dip before the substrate is loaded in the process chamber. However the passivation is partially lost when the HF dip is followed by a water rinse which results in oxygen absorption on the substrate. It was found that the peak oxygen concentration at the epitaxy substrate interface increase by an order of magnitude due to a five minute water rinse. We report here that oxygen and carbon at the epitaxy substrate interface can be desorbed during initial stage of epitaxial growth by reducing epitaxial growth rate. In this work, epitaxial Si films were deposited over a wide range of growth rates obtained by varying Si2H6 flow rates. The peak oxygen concentration decreases by an order of magnitude by changing the growth rate from 3000 to 700Å/kminute for a deposition temperature of 800°C. We believe that at higher growth rates Si overgrows on absorbed oxygen maintaining epitaxial alignment reflected in the good electrical quality of the epitaxial films. However, at low growth rates oxygen has sufficient time to desorb before overgrowth can take place, improving the epitaxy substrate interface quality.

The characteristic step coverage behavior which a given LPCVD process exhibits depends on the nature of the controlling gas phase and/or surface chemical reactions. Physically-based ballistic transport and reaction film profile evolution simulation has provided a structure wherein the origins of step coverage limitations can be understood in the context of the interaction of transport and the controlling chemistry. Based on comparisons of the simulations to literature and in-house experimental data, we have categorized LPCVD mechanisms into one of three types. In heterogeneous deposition, conformal step coverage can usually be found under at least some process conditions. Step coverage typically degrades with increasing deposition temperature. In homogeneous precursor-mediateddeposition, a reactive intermediate is formed in the gas-phase above the wafer surface, resulting in poor to moderate step coverage. Step coverage may or may not degrade with increasing temperature. In byproduct-inhibited deposition, a gas-phase byproduct generated via a surface reaction readsorbs on the growing film surface and slows the deposition rate, yielding a poor to moderate, relatively temperature-insensitive step coverage. Poor step coverage is manifested in a marked film thickness discontinuity at the feature mouth, with a relatively uniform film down the feature sidewalls.

The influence of temperature gradients on the partial pressures of a binary mixture in a cold wall low pressure chemical vapor deposition reactor was determined by Raman spectroscopy of the gaseous species in the reactor. It is demonstrated for the first time that the partial pressure of the heavy constituent in the hot region of a low pressure reactor is reduced by 35 % due to the Soret effect. Model calculations that included the Soret effect are in agreement with the experimental data.

Computational programs are often used for estimating molecular orbital energies and geometries. The results of these calculations can also provide information needed for investigation of adsorption and decomposition mechanisms on surfaces. In this study, ab initio methods are used to calculate properties of gas phase diethylgermane and GeHx-substituted silanes (x=1−3). The silanes are used as models for surface Ge hydrides. The calculated gas phase molecular orbital energies are then compared with experimentally determined molecular orbital energies for the adsorbed species. The surface species are prepared by germane, digermane, ethylbromide, or diethylgermane adsorption on the Si(100)-(2×1) surface and the molecular orbital energies are measured by photoelectron spectroscopy.

The low temperature epitaxial growth of Si / GaP / Si heterostructures is investigated with the aim using GaP as a dielectric isolation layer for Si circuits. GaP layers have been deposited on Si(100) surfaces by chemical beam epitaxy (CBE) using tertiarybutyl phosphine (TBP) and triethylgallium (TEG) as source materials. The influence of the cleaning and passivation of the GaP surface has been studied in-situ by AES and LEED, with high quality epitaxial growth proceeding on vicinal GaP(100) substrates. Si / GaP / Si heterostructures have been investigated by cross sectional high resolution transmission electron microscope (HRTEM) and secondary ion mass spectroscope (SIMS). These methods reveal the formation of an amorphous SiC interlayer between the Si substrate and GaP film due to diffusion of carbon generated in the decomposition of the metalorganic precursors at the surface to the GaP/Si interface upon prolonged growth (layer thickness > 300Å). The formation of twins parallel to {111} variants in the GaP epilayer are extended into the subsequently grown Si film with minor generation of new twins.

The epitaxial growth of an insulator, BaF2, and semiconductors of the II-VI and the IV-VI families on Si substrates were carried out. In-situ XPS analyses during the growth of the first monolayers were used to study the surface chemical reactions involved. The results point to a common ingredient in these growths: that the Ba atoms are involved in forming interfacial compounds that would facilitate the heteroepitaxies. In the case of BaF2/Si, a BaSi2 compound has been identified previously. In the case of PbTe and CdTe, the heteroepitaxies on Si are made possible with the BaSi2 buffer. As a result, the impinging semiconductor molecules are broken up, and the metallic elements are ejected from the BaSi2 surface. A new surface chemical, BaTe, is thereby formed. These surface Ba compounds appear to be the dominant factors as the crystal orientations of the BaF2, CdTe, and PbTe layers are independent of those of the Si substrates.

The paper presented discusses certain topographical effects being significant for the coverage of nonplanar structures with undoped and doped thin SiGe films. They are essential for new integrated heterojunction Si/SiGe devices. To investigate the coverage of topographical surfaces SiGe layers have been deposited on different trench structures in a Rapid Thermal Low Pressure single wafer CVD reactor (RTCVD) from the system SiH4, GeH4, and H2 using B2H6/ H2 and PH3/ H2 for the in-situ doping, respectively. Various deposition conditions and different film compositions have been used. The results especially the differences of the thickness distribution within trenches and on the surface have been discussed with regard to different CVD models taking into account distinguished reaction mechanisms.

A new damage-less patterning method of the photo-oxidized GaAs mask used for the selective-area growth of GaAs has been developed. We have found a new characteristic of the GaAs oxide: a metal Ga deposition onto the GaAs oxide surface lowers the desorption temperature of the oxide. The patterning method employed is based upon this characteristic. The GaAs oxide where 15 atomic layers (ALs) of Ga is deposited is locally removed at 540°C to form an opening area in the oxide mask. After forming this opening area, GaAs is selectively grown there by metal-organic molecular beam epitaxy (MOMBE).

Silicon nucleation on silicon dioxide and selective silicon epitaxial growth (SEG) were studied in an ultra high vacuum rapid thermal chemical vapor deposition (UHV-RTCVD) reactor. Experiments were performed using 10% Si2H6 in H2 in a pressure range of 10 - 100 mTorr at 760°C. Under these conditions, the growth rate ranged from 75 to 330 nm/minute. Loss of selectivity via Si island formation on SiO2 was studied using scanning electron microscopy (SEM) and atomic force microscopy (AFM) revealing a strong dependence on deposition pressure. Cross sectional transmission electron microscopy (XTEM) was employed to study the vertical oxide/epitaxy interface where faceting can occur. The incubation time for nucleation was found to increase from 10s to 70s as pressure is reduced from 100 mTorr to 10 mTorr, allowing thicker selective epitaxial film growth in spite of the reduced growth rates. This was attributed to the reduction in gas phase supersaturation of the Si containing species resulting in a lower density of adsorbed atoms on the SiO2 surface. This process shows a potential for chlorine free selective epitaxial growth and provides insight to the surface morphology of polycrystalline films deposited at low pressures.

Perfect selectivity in Si epitaxial growth on a Si/SiO2 substrate has been achieved by irradiating SR during molecular beam epitaxy (MBE) using disilane. This differs from conventional selective growth by MBE using disilane in that no Si nucleation occurs on SiO2 irrespective of growth time. Temperature above 700°C is necessary for SR -induced selective growth. This perfect selectivity might result from the elimination of adsorbed Si atoms induced by photo-stimulated evaporation of SiO2 which results in the perfect suppression of Si nucleation.

As the feature size of MOSFET devices shrink, issues such as thermal budget associated with controlling channel doping profiles and oxide growth kinetics raise concerns about using thermally grown furnace oxides for deep-submicron device applications. To address these concerns, we have developed a new RTCVD oxide process using a gas system of silane and nitrous oxide. The RTCVD oxides are deposited in a lamp-heated, cold wall, RTP system. Deposition rates ranging from 55 Å/min. to 624 Å/min. can be achieved at 800°C with silane nitrous oxide flow rate ratio of 2% and total pressure ranging from 3 to 10 Torr. The results indicate that this RTCVD process can be used to deposit both thin gate and thick isolation insulators for single wafer processing. Deposition rates of the RTCVD oxides exhibit a nonlinear dependence on the total deposition pressure. Electrical characterization of the as-deposited RTCVD oxides shows a mid-gap interface trap density of < 5×1010 eV−1 cm−2 and an average breakdown field of 13MV/cm. AES, RBS and TEM analyses have been used to study surface cleaning effects on the silicon-silicon dioxide interface quality and to determine the chemical composition of the RTCVD oxides. The results show that RTCVD oxides with stoichiometric composition and atomic flat silicon-silicon dioxide interface can be achieved using silane nitrous oxide flow rate ratio of <2%. I-V characteristics and transconductance degradation under hot carrier stress for MOSFET's using as-deposited RTCVD gate oxides have been found to be comparable to those of MOSFET's using thermal gate oxides.

Hydrogenated amorphous carbon (a-C:H) thin films were deposited onto various substrates at 200 °C by DC saddle-field glow-discharge dissociation of CH4. In situ plasma probe mass and energy spectroscopy was employed to systematically study the effects of the discharge parameters on the formation of reactive precursors. The ion current at the substrate holder consists mainly of CH3+ for pressures < 75 mTorr; higher pressures and lower discharge currents encourage the formation of C2Hn+ and C3Hn+ radicals in the discharge. The growth rate was largely independent of the total pressure, increasing approximately linearly with the discharge current. Under 476 nm photo-excitation, the a-C:H films exhibited room-temperature photoluminescence in the visible, near 1.9 eV, with an intensity that was strongly dependant on the discharge parameters.