Irradiation with excimer laser pulses is demonstrated as a technique to fill contact/via holes for integrated circuits with Al-(1 wt.%)Cu. We discussed the influence of the size, shape and aspect ratio of the contact hole, and the metal thickness on the process (attitude for laser planarization. The most dominant parameters found to influence the planariza-tion process are the shape of the contact, and the shape of the as-deposited metal surface (“step coverage”). The application of a Cu anti-reflection coating increased the process window substantially. Although a planar Al-Cu surface was easily formed with laser irradiation, void free filling of contact holes was a more demanding task. Conditions for which completely filled contacts can be obtained were discussed.

In this study we investigate the formation mechanisms and morphology of TiSi2 formed by deposition of Ti on atomically clean silicon substrates. Ti films of 50–400 Å thickness were deposited in ultra-high vacuum on Si (111) wafers and annealed to temperatures between 500–900°C. Films were monitored in situ with AES and LEED, and post preparation characterization was accomplished with SEM, TEM and Raman scattering. The results show that for films of thickness ≤, 100 Å the C49 TiSi2 phase is stable over the entire 600-800°C temperature range. However, for films of 200-400 Å thickness, the C49 to C54 phase transition occurs at temperatures varying from 700 to 800°C dependent upon film thickness. The high temperature annealing results in flat interface structures, and island formation is observed for all films with the C54 structure. The interface morphology and the mechanisms of TiSi2 island and phase formation are discussed in terms of surface and bulk free energies considerations based on nucleation theory.

The effect of interfacial oxides and impurities left by aqueous chemical cleaning procedures on the n-type Schottky barrier heights of various metal-GaAs contacts have been examined as a function of annealing temperature. The as-deposited barrier heights of metals which are expected to be reactive with respect to the native oxides of GaAs (Mn, Cr, Al and Ti) were the most sensitive to variations in the residual oxide thickness. Omitting the final oxide etch from the cleaning sequence resulted in a 50 to 60 meV reduction in the as-deposited barrier height of these metals. However annealing these contacts at temperatures as low as 175 °C results in a barrier height that is independent of the initial surface clean. In contrast, for metals which are expected to be unreactive with respect to the native oxides of GaAs (Ni, Au, Cu and Ag) the pre-deposition cleaning procedure had little effect on either the as-deposited or the annealed Schottky barrier heights.

A structure of surface and near surface layers in inorganic glasses used as substrated in metallization was studied. The analysed samples were made of melted quartz, technical and optical quartz glass and sodium-silicate glass.

The sample surface were at first treated mechanically ground and polished and then chemically etched.

In the investigation we applied the following methods:

- X-ray and electron/mean voltages 35…65 kV/diffraction,

- electron emission microscopy,

- optical/UV-VIS/and infrared/IR/microspectrophotometry and microscopy.

For the studied series of quartz based/SiO2/glasses we determined the size of inhomogeneous. regions and the distortion degree of the elementary tetrahedra/basic structural elements/and estimated the thickness of surface and near-surfaoe layers.

The obtained structural parameters enabled us to propose structural models of the surface of real glasses uaed in elec-trotechnology, electronics and microelectronics.

A New method of the low resistivity α-Ta deposition technique has been developed. The sputtered Ta film deposited on TaMo alloy has bcc structure (α-Ta) in contrast to tetragonal of the Ta mono layer (β-Ta), and shows the resistivity as low as 22μΩ cm. The mechanism of this transformation is not explained by simple epitaxial growth. X-ray diffraction analysis and RBS analysis indicated that the TaMo alloy layer, due to Mo diffusion from the under layer, acts as a seed plane of growth. The Ta/TaMo layered film is suitable for Large Area Electronic devices for its low resistivity.

An excimer laser (XeCl, 308 nm) has been used to locally melt and reflow aluminum alloy films into contacts of VLSI device structures resulting in contact filling and local planarization. This technique has the potential to be the answer to the ‘poof’ aluminum step coverage problem. The first order problems are that of ablation and voids. Voids have been categorised as ‘line voids’ and ‘contact voids’. Various processing conditions of the barrier, Al-Si-Cu alloy and Al-Cu alloy metallizations were explored to obtain the optimum process with the desired film characteristics.

30-keV focused Ga+ ion beam was used for induced deposition of small-area tungsten thin films from W(CO)6 on Si and SiO2. Deposition yield, calculated assuming pure tungsten depositions, depends on dwell time (beam diameter/scan speed) and beam current density. High current density and/or long dwell time are known to cause low deposition yield because of the depletion of adsorbed gas molecules during ion beam irradiation. Based on a model taking this effect into account, numerical fitting was carried out. The reaction cross-section was estimated to be 1.4 × 10−14 cm2. For doses below 1017 ions/cm2, film resistivity decreases with increasing dose. This was confirmed for several dwell times. However, for doses above 1017 ions/cm2, film resistivity remains independent of dose. In this “high”-dose range, variation of beam current density has little effect on film resistivity. AES analyses revealed a consistency between film composition and resistivity. For a “high”-dose film with a resistivity of 190 μΩ-cm, the approximate tungsten content was 50 at%.

Refractory metal compounds, such as nitrides or borides, are attractive candidates for diffusion barrier between silicon and aluminium in VLSI technology. We studied tungsten boride films deposited on silicon (100) by ion beam sputter deposition (IBSD).

The tungsten boride films were prepared by sputtering a W2B5 target by argon ions with energy ranging from 0.5 to 2keV. The substrate temperature was varied from room temperature to 630°C. Finally, the films were patterned by selective wet etching in order to characterize the resulting Schottky diodes. We observed that a boron loss occurs during deposition, probably due to the backscattering of sputtered boron on previously deposited W atoms. By in situ AES analysis, we verified that a 5 nm thick layer acts as a diffusion barrier for silicon up to about 630°C, for all deposition conditions. The films properties were found to depend weakly on the primary ion energy and on the substrate temperature. All the films have resistivity at room temperature in the range of about 250 µΩ cm. The measured density, in the range of 12 g/cm3, is very close to that of WB2 bulk material, while the intrinsic stress of the films remains compressive and in the range of -lGPa. This value is notably lower than what we measured for pure tungsten prepared under similar deposition conditions.

Laser-induced Chemical Vapor Deposition of tungsten on Si(100) using WF6 and H2 has been investigated using a high-vacuum system comprising a cold-wall reactor. The activation source is a pulsed ArF-excimer laser. The deposition rate depends linearly on the repetition rate, when H2 is used as a reducing agent. When no H2 is used the laser radiation suppresses deposition.

At deposition temperatures down to 200 °C laser deposited layers have resistivities better than 20 μΩ cm. Thick layers have resistivities down to 8 μΩ cm. There is a direct relation between layer thickness and resistivity. X-ray diffraction revealed the layers to consist of α-tungsten. β-tungsten was only obtained for those thermally deposited layers where growth was slower than expected.

Nuclear reaction analysis of fluorine showed that most fluorine is present near the W-Si interface, and that the amount of fluorine relative to the amount of tungsten in the layer decreases markedly with deposition temperature.

A modified line-of-sight model for transport and deposition during LPCVD is used to predict step coverage and film composition uniformity of tungsten silicide barrier layers. Predictions are compared with experimental results for 2 μim wide by 6 μm deep trenches with barrier layers of 0.2 μm nominal thickness. Model predictions are in quantitative agreement with those of a diffusion-reaction model and are in qualitative agreement with experiment.

The effect of stress on W-deposition at low pressure by Si-reduction of WF6 has been studied on patterned wafers utilizing cross-section SEM and cross-section TEM. Si-consumption and encroachment formation at the Si/Si02 interface was analyzed as a function of the oxide thickness, which changes the residual stress of the Si-substrate beneath the SiO2-layer. At the edge of the SiO2-layer, significant enhancement of W-deposition was observed from the early stage of W-metallization. Penetration of W into the Si-substrate beneath the edge showed strong dependency upon the oxide thickness, while penetration along the SiO2/Si interface (i.e. encroachment) did not. It has been analyzed that early stage of W-deposition was strongly affected by residual stress in the Si-substrate, and contact between WF6 and Si will be much easier through porous W-film and the W/Si interface, where the existence of microvoids is expected.

The reaction between Al and CVD-W films has been studied. Al/α-W/Si and Al/β-W/Si structures were prepared by deposition on different Si substrates by changing deposition conditions using silane reduction of tungsten hexafluoride, followed by Al-Si-Cu alloy film sputter deposition. The sheet resistance of Al/α-W/Si structure is higher than that of Al/β-W/Si structure after 500°C annealing. RBS measurements show that the W diffusion into Al occurs in both structures after annealing, and the reaction between α-W and Al takes place easily compared with that between β-W and Al. This causes the sheet resistance difference. The activation energies for the W diffusion into Al, however, are almost the same in both structures. When CVD-W films are exposed to air after removal from the reactor, the sheet resistance of β-W film increases according to the exposure time, while that of α-W film does not. AES measurements indicate that the β-W film absorbs more oxygen than the α-W because of the difference of grain structures. The resistance increase of β-W film is caused by the oxygen that is absorbed from air. Our results indicate that the oxygen in the β-W layer suppresses the W diffusion into Al. Once the reaction begins, however, the diffusion into Al does not depend on crystalline phases of CVD-W.

WSix thin films were deposited on SiO2/Si substrates by Low Pressure Chemical Vapor Deposition (LPCVD) using WF6 and SiH4 gases. The deposition mechanism has been studied by measuring the thickness, resistivity and composition of the films by varying deposition temperature and gas flow rate at a constant total reactant gas pressure. Below 300°C, the surface chemical reaction was the rate-limiting process and the deposition rate increased exponentially with temperature having a thermal activation energy of 3.2 kcal/mol. Meanwhile, above 300°C the reaction was governed by the mass transfer step in the gas. The deposition rate in this range is insensitive to the deposition temperature but shows dependence of the flow rate of reactant gases. AES and RBS analyses were performed to determine the stoichiometry of WSix thin film. The Si content in film gradually increased as the deposition temperature increased. The resistivity of as-deposited WSix film has dependence on both deposition temperature and Si/W ratio, and exponentially increased with a moderate slope. However, temperature insensitive behavior of resistivity appeared in the mass transfer controlled region. Such resistivity changes with temperature were discussed with the Si/W ratio and the microstructure of films.