A smooth Ge(001) surface can be severely roughened by chemical etching with oxygen gas and then returned to its original smoothness by exposure to hydrogen. The rate of recovery of the surface depends strongly on temperature, as well as on whether the hydrogen is introduced as a low energy ion beam or simply as a gas. The roughness induced in the surface by etching is “locked-in” after removal of the oxygen gas through pinning of ledges by residual contamination. This pinning prevents the surface from smoothening thermally. The introduction of hydrogen to the surface promotes a chemical reaction which frees the pinned sites and allows thermal smoothening to proceed.

A detailed investigation into the mechanisms of the ion-assisted etching of TiSi2 using 1000 eV argon ions and chlorine gas is reported. X-ray photoelectron spectroscopy (XPS) was used to investigate surface product distributions and modulated ion beam mass spectroscopy (MIBMS) was used to investigate gas phase product distributions. Identical experiments were performed on silicon and titanium single crystal substrates. Information on chemical mechanisms during ion-assisted etching was obtained by comparing the product distributions of the three substrates. For each substrate, the di and tetrachlorides were the major gas phase products. These products were emitted with high thermal velocities following a surface residence process. TiCl and TiCl2 products were found on the titanium surface after simultaneous exposure to chlorine and argon ions, but not on the TiSi2 surface. These observations are discussed in terms of specific chemical mechanisms which are critical during ion-assisted etching.

High quality SiO2 films are obtained by spin-coating wafers with a sol/gel of silicic acid in either a 2, 3, or 4-carbon linear-aliphatic alcohol. Some properties of the deposited film depend upon the solvent: such as density, tensile stress, and infrared spectrum. However, Rutherford-back-scattering analysis indicates the O:Si ratio (2.00±05) to be independent of the solvent. The infrared spectrum of the oxide exhibits Si-OSi absorption in the range 1070–1080 cm.−1 depending on the curing temperature and solvent system. (The weaker Si-OSi band is found at 804–810 cm.−1) In addition, low-temperature-cured (<500 °C) films show Si–OH absorption. Films hot-plate baked at 150–350 °C are stable but not fully cured. Films from propanol baked at 400 °C have a refractive index of 1.41–1.42 and a wet-etching rate in 30:1 BOE of ˜1250 Å/min. Films cured at 900 °C have a refractive index of 1.42–1.43, a wet-etching rate of ˜430 A/min, and are more dense by a factor of ˜1.25. Dry-etching with CHF3/O2 occurs at rates comparable to those of CVD oxides. Multiple applications lead to crack-free films as thick as 0.6–0.8 μ m. Deposition over aluminum-patterned topography results in a smoothing of the surface and suppression of hillock growth in the aluminum even after a 450 °C cure. SOX adheres to silicon, aluminum, and silicon dioxide. A boron-doped SOX is readily prepared.

This paper describes an analysis of the physical properties of the sol using several complementary light scattering techniques. Polymerization and aggregation kinetics were followed through time-dependent changes in the size, shape, and density of the sol particles. The sot growth rate was controlled by choice of solvent and silicon concentration. Changes in viscosity and pH were small during the reaction period. Three different particle-growth regimes exist in which either the particle density increased, decreased, or remained the same. The addition of boron, hydrofluoric acid, or water accelerated the reaction. The sol experimental data correlate with the density and wet-etching rate of the cured films. After curing, high-density films were obtained from sols with three common characteristics: an average particle diameter >450 Å, a relatively high polydispersity, and a low particle density. These criteria were generally satisfied by solutions one to two days old.

We studied the behavior of mobile charges, which we believe are hydrogen ions, in a novel spin-on oxide, SOX. Our analysis is based on the Triangular Voltage Sweep (TVS) technique, which is normally used to detect the presence of mobile alkali ions in silicon dioxides. We found that the TVS traces of protons exhibit several unique features that allow us to distinguish a proton signal from that of other common contaminants, such as sodium. We suggest a model to explain these unique features. We show that the presence of hydrogen in the SOX material is due to the high water retention of this porous oxide.

UV excimer laser light-induced prenucleation of various surfaces from spin-on films in air is described. Good results were obtained using palladium acetate on substrate materials ranging from ceramics to polymers and even paper. Virtually any surface material can be prenucleated in this way. The Pd-prenucleated planar and curved surfaces can be metallized by a vast variety of metals using electroless plating solutions (Cu, Ni, Au, Pt, Co, Sn, and even alloys). Results are described in detail for copper from a standard electroless solution (Doduco or Shipley). Contact metallization through via holes using this process was aachieved with excellent contact properties.

A new method is described for initiating electroless copper deposition onto dielectric substrates. The method employs a thin poly(acrylic acid) film, applied to the substrate by dip-coating, to bind Pd+2 from PdSO4 solution. Uptake of Pd+2 takes place by H+/Pd+2 ion exchange at the carboxylic acid sites of the polymer. Upon immersion of a substrate treated with Pd+2/poly(acrylic acid) into an electroless copper plating bath, reduction of Pd+2 to Pd0 takes place, forming an active catalyst for the initiation of electroless copper plating.

The surface conductivity of PTFE-copper oxide filled composites containing between 10 and 25% of the oxides increases significantly after treatment in a low pressure hydrogen plasma at temperatures below 100°C. The plasma-reduction of copper oxides (CuO and CuO/Cu2O mixture) was investigated and compared with thermal reduction at different hydrogen partial pressures.

A comparative investigation of polyimide/Cu interface degradation has been carried out for ultrathin photosensitive, non-photosensitive, and preimidized polyimide precursor films cured while in contact with a Cu substrate. The role of curing byproducts and environmental conditions on interface degradation has been elucidated by means of X-ray photoelectron spectroscopy. Immediately after curing, we observed some oxidation of the Cu in contact with non-photosensitive and photosensitive polyimide overlayers. On the other hand, only negligible oxidation was observed for the preimidized polyimide/Cu interface. Experiments in which samples were stored in vacuum, air and a humidity chamber show a dependence of the oxidation kinetics on air/moisture exposure. Preimidized and photosensitive polyimide/Cu interfaces, stored in air, became more extensively oxidized with time relative to identical samples stored in vacuum. Moreover, all three polyimide/Cu interfaces exhibited significantly more oxidation after 3 days in a humidity chamber than after 19 days of storage in air. Taken together, these data clearly demonstrate that absorbed water and its interaction with curing byproducts are key factors in the extent of Cu oxidation at the interface.

The electrical resistivity, optical band gap and activation energy for electrical conduction have been determined as a function of preparation conditions. The operating conditions for the glow discharge reactor have been interpreted in terms of ion energy and reactive species production. The change in the electrical properties could not be explained as a percentage of [SP3] versus [SP2] bonding ratio. Rather, these two species are embedded in an amorphous medium which determines the materials electrical properties.

The interaction between an aromatic molecule and the graphite basal plane is constructed using a Lennard-Jones 6–12 potential. The corrugation of the potential across the surface, the binding energies and vibrational frequencies of the aromatic molecules on the basal plane have been calculated and compared with experimental data. We have also calculated the interaction between a rare-gas atom and an aromatic molecule using two independent models, one based on the pairwise sum of the Lennard-Jones potential and the other on the Effective Medium Theory (EMT) for the repulsive part and a Van der Waals dispersion term for the attractive part. Our results are then compared with available experimental data.

Thin alumina films are grown on a silicon substrate, in vacuum, at a rate of about 10 Å per second, and at substrate temperatures below 1500C. Alumina is dissolved in supercritical water and the resulting solution expanded in a supersonic free jet, which is directed at the substrate. The films are characterized by ESCA and FTIR.

The self-limiting effect during Low Pressure Chemical Vapor Deposition of tungsten is manifested by a sudden interruption of the reduction of tungsten hexafluoride WF6 by silicon, so that only very thin (self-limited) films of silicon-reduced tungsten can be grown. It has been shown that the self-limiting effect is caused by formation of the non-volatile subfluoride WF4. The temperature dependence of the self-limiting thickness of the tungsten films grown in a hot-wall reactor exhibits a characteristic maximum at temperatures near 350°C, which indicates that at this temperature the rate of formation of WF4 is lower than at temperatures above and below. In the present paper we attempt to explain this peculiar dependence. We suggest a mechanism responsible for formation of the blocking agent WF4. We demonstrate that the presence of a hot tungsten surface is essential for the strong self-limiting effect. This leads us to the discussion of the proper selection of the reactor type (hot-wall vs. cold-wall) for different process requirements.

Antimony is a useful n-type dopant for buried layers, due to its low rate of lateral diffusion in subsequent processing steps. To date, no solid planar diffusion source has been available. Reported here is the development of such a source.

The source is made by high temperature synthesis in a hot press. A mixture of antimony trioxide and silicon powders is placed in the press, and heated under pressure. The reaction

results in a solidified product. Wafers are sliced from this product to form the actual planar diffusion sources.

The sources are used in a two-step doping process: a predeposition followed by a drive-in. Oxygen is required in both steps, due to the chemistry involved. Typical properties of the doped silicon are: sheet resistance, 10 Ω/square, junction depth, 4–6 microns, and antimony surface concentration. 3–8 × 1019/cm3.