Chemical Vapor Deposited Tungsten Silicide Research Articles

Stress changes in chemical vapor deposition tungsten silicide (polycide) film measured by x-ray diffraction

Stresses in chemical vapor deposited polycide tungsten silicide (poly-Si/WSi2) were evaluated at each stage of the fabrication. The individual layers of the Si/SiO2/poly-Si/WSi2/poly-Si multilayered structure were deposited sequentially on separate wafers and subjected to x-ray diffraction analysis in the as deposited and annealed conditions to determine the changes in strain occurring in WSi2. Samples cut from wafers containing all the layers were either heat treated at 1193 K for 30 min or capped with a 25 nm thermal oxide and the strain in the WSi2 film was also analyzed by x-ray diffraction. The change in strain of the WSi2 layer, following each step of the fabrication process, was evaluated by the lattice parameter variation of the c axis. The stress is affected by the layers of the multilayered film. An as deposited poly-Si layer on top of WSi2 reduces its stress, since it introduces a compressive component. The stress in WSi2 is further reduced upon annealing. Poly-Si also maintains a Si supply at the poly-Si/SiO2 interface thus eliminating Si out-diffusion during heat treatment in an oxygen containing ambient. Capping the system by a thin oxide layer modifies the stress pattern of the WSi2 which becomes compressive.

AES determination of compositional variation in LPCVD tungsten silicide films in confined spaces

AbstractCompositional variation in chemical vapor deposited tungsten silicide films in confined spaces has been studied using scanning Auger electron spectroscopy. ‘Infinitely’ long silicon trenches with aspect ratios ranging between 1.5 and 5 were used to define the confined space. Composition changes down the sidewall of a partially filled trench were determined by cleaving the sample along the length of a trench and analyzing specific points down the sidewall. The samples were lightly sputtered before analyzing to remove the native oxide and reach a uniform composition of the silicide. The sample was carefully positioned during sputtering to minimize any redeposition of material onto the analysis area due to sputtering of adjacent surfaces. Using this method the silicon content of the silicide was shown to increase as a function of distance from the surface. This spatial variation in film composition qualitatively agrees with the predictions of a computer simulation of the deposition process.

Plasma Etching of Tungsten Polycide Structures Using NF 3 ‐ Mixed Halocarbon Etchants

The feasibility of utilizing halocarbon etchants to anisotropically etch tungsten polycide structures has been investigated. Chemical vapor deposited tungsten silicide and low‐pressure chemical vapor deposited polysilicon have been etched using various mixtures of , , and with . Anisotropy, undercutting, and notching were evaluated using cross‐sectional scanning electron microscopy. Etch rates and profiles were controlled by utilizing halocarbon‐rich (>50%) binary halocarbon etchants. The etch rate data indicate that the volatility of the tungsten halide formed during etching influences the silicide etch rate and profile while the polymer forming ability of the halocarbon controls the etch characteristics of the polysilicon.

A study of structural, electrical, and optical properties of low-pressure chemical vapor deposition tungsten silicide films processed at elevated temperatures

Tungsten silicide films deposited by the low-pressure chemical vapor deposition technique on polysilicon are studied at elevated processing temperatures. It is shown that annealing of as-deposited silicon-rich tungsten silicide films in nitrogen at process temperatures exceeding 1000 °C results in severe silicide surface morphology. Silicide capping using a layer of dielectric is shown to be attractive to circumvent this problem and to avoid silicide peeloff. Structural, electrical, and optical properties of tungsten silicide films are studied as a function of annealing parameters. It is shown that silicide annealing ambient and activation temperature result in a considerable modification of silicide thin-film properties. Correlation among structural, electrical, and optical silicide film properties is made based on extensive experimental data.

Competitive reactions of fluorine and oxygen with W, WSi2, and Si surfaces in reactive ion etching using CF4/O2

We have studied CF4/O2 based plasma and reactive ion etching of dc magnetron-sputtered tungsten, chemical vapor deposited tungsten silicide and single-crystal silicon. Etch rates and optical emission intensities of atomic fluorine and oxygen in the gas phase were determined as a function of gas composition. Etched surfaces were characterized using in situ x-ray photoelectron spectroscopy (XPS) and low-energy ion scattering spectrometry. All three materials exhibit an etch rate maximum as a function of oxygen addition. For silicon (7% O2) and WSi2 (10% O2) it occurs for a smaller percentage of added oxygen than the optical emission maximum of atomic fluorine (15% O2), whereas for W films (35% O2) it occurs at a greater percentage of added oxygen. In postplasma XPS analysis it is shown that for all materials the etch rate decrease coincides with the onset of significant oxidation. From the analysis of the degree of fluorination of the top surface, the oxygen/fluorine ratio of the reaction layer, and the oxygen percentage of the etch rate maximum relative to the fluorine emission maximum, we conclude that the etching of Si and WSi2 is limited by a reduction of fluorine adsorption because of competitive oxygen adsorption. For tungsten, which shows a very low degree of fluorination of the top surface, both the drop in etch rate and the formation of tungsten oxide can be explained by a reduced arrival rate of fluorine atoms to the tungsten surface when etching in oxygen-rich gas mixtures.

Characterization and optimization of low-pressure chemical vapor deposited tungsten silicide using screening and response surface experimental designs

Low-pressure chemical vapor deposited WSiX has been characterized and optimized through the use of screening and modeling experimental designs. In the screening study, the effects of temperature, pressure, gas flow rates, and a predeposition treatment were determined for a list of properties which included deposition rate, resistivity, stress, stoichiometry, oxidation rate, particle generation, and resistance to chemical attack. Over the process domain covered, temperature, WF6 flow rate, and SiH4 flow rate were found to affect deposition rate, resistivity, film stress, and stoichiometry while stoichiometry was also affected by pressure. WSiX films produced in the screening study ranged in stoichiometry from X=2.1 to 3.0. The oxidation rate of these samples was found to be insensitive to deposition conditions and unrelated to the as-deposited stoichiometry. As-deposited and annealed film stress were found to decrease linearly with increasing silicon content of the as-deposited films. Similarly, the as-deposited and postannealed resistivity of WSiX films were found to be a function of the initial stoichiometry. Manufacturability related properties such as particle generation, chemical resistance, and uniformity measures involving sheet resistance, resistivity, and film thickness were found to be largely insensitive to the process factors studied. Response surface models were developed for deposition rate, annealed resistivity, annealed stress, and stoichiometry and were used to guide the final process optimization. The optimized process window was relatively large and yielded WSiX films with an annealed (950 °C, 30 min) resistivity of ∼75 μΩ cm, an as-deposited Si/W atom ratio of ∼2.5, annealed tensile stress of ∼1.2×1010 dyne/cm2, and deposition rates of 9 to 10.5 Å/s.

Effect of excess silicon on behavior of LPCVD WSi x films on silicon

This paper addresses the important effects of excess silicon in low pressure chemical vapor deposition tungsten silicide films on resistivity, adhesion, and stress as a function of annealing temperature with physical interpretation. Excess silicon can be incorporated in the films to lower stress, but a consequence of this is that the resistivity is increased. On annealing such fimls, excess silicon tends to diffuse out of the silicide increasing stress and lowering resistivity. Results are presented of both the stress in the films and their resistivity as a function of annealing conditions. The annealing ambient also plays a strong role in the rate at which excess silicon migrates to the surface. Oxygen greatly enhances the rate compared with argon or nitrogen.

Changes in Resistivity and Composition of Chemical Vapor Deposited Tungsten Silicide Films by Annealing

Chemical vapor deposited (CVD) tungsten silicide films were formed by a cold wall reactor. These films were annealed in to investigate changes in resistivity, composition, thickness, and impurity. The change in resistivity after 1000°C annealing becomes larger as the film reaches the stoichiometric value. A composition change occurs in a film whose composition Si/W is more than 2.6. Excess Si in the films is segregated in the boundary between and poly‐Si. A thickness change of about 15% occurs after 1000°C annealing at on ; this value is smaller than the calculated value. F and H, which are impurities in films decrease gradually and diffuse into gate after 1000°C annealing.

Analysis of the effects of annealing on resistivity of chemical vapor deposition tungsten–silicide films

Chemical vapor deposition WSix films were formed with varying compositions and film thicknesses. The were films annealed in N2 for 30 min and their resistivities were measured. The maximum resistivity was obtained for annealing temperatures between 500 and 600 °C if the film was comparatively richer in tungsten and was thick. The film was analyzed by x-ray diffraction and secondary ion mass spectrometry. The maximum resistivity is considered to be related to crystallization, which changes amorphous WSix into a mainly hexagonal structure, and to the difference in the conduction mechanism of electrons between amorphous, hexagonal, and tetragonal WSi2. It is clear that impurities in the film do not contribute to the maximum resistivity. There is also no relationship between resistivity and the number of grains of tetragonal WSi2 per unit length for annealing below 600 °C.