Silicide Grains Research Articles

Surface morphology study of titanium silicide formed on polycrystalline silicon

TiSi2 can be used as a shunting layer to decrease the resistance of the poly-Si lines used as interconnections in integrated Si technology. In this work the correlation between the silicide thickness, the silicide surface roughness, the silicide grain size, the dopants in Si, and the crystallinity of Si was investigated on unpatterned wafers. The results will be used as reference in a further study of silicide formation on sub-micron poly-Si lines. After silicide formation at 700°C, the surface roughness has been found to be strongly dependent on the thickness of the silicide layer formed. The dopants in the Si did not only retard the formation of the C49 phase, but also hindered the transformation of TiSi2 from the C49 to the low resistive C54 phase. The surface roughness did not change when a second annealing step at 850°C was used to transform the silicide from the C49 phase to the C54 phase.

Room Temperature Oxidation of Silicon Catalyzed by Cu3Si

ABSTRACTThe dependence of room temperature oxidation of silicon catalyzed by Cu3Si on silicide thickness, silicide grain size and the substrate orientation has been investigated by transmission electron microscopy and x—ray diffractometry.The extent of oxidation was found to depend critically on the thickness of starting Cu3Si film. The oxidation was found to be more restricted on (lll)Si than that of (001)Si samples. The thickness of SiO2 layer was found to decrease with the average grain size of the starting Cu3Si layer. HRTEM revealed that the oxidation is initiated at the grain boundaries. Oxide film as thick as 4.5 μm was grown at room temperature over a period of two weeks in (001) samples. The growth of thick oxide film was achieved by minimizing the grain size of Cu3Si through a reaction between Cu and an intermediate amorphous silicon layer at 200 °C.

Influence of oxide thickness on ion-beam induced and thermal CoSi 2 formation

In this study, the influence of thin oxides on thermal and ion-beam induced CoSi 2 formation as well as the interaction between Co and thick isolation oxides during ion-beam mixing and rapid thermal annealing were investigated. It was shown that a native oxide layer between the Co and the Si has a strong influence on reaction temperature and properties of the formed silicide layer whereas ion-beam mixing with Ge doses at or above 5 × 10 14 cm -2 resulted reproducibly in a decreased reaction temperature and in formation of homogeneous layers. Further investigations revealed that silicide grain formation is possible on thin thermal oxides up to 10 nm. Using oxide layers thicker than 100 nm, no CoSi 2 formation could be detected after thermal treatment but a rather severe damage of the SiO 2 surface resulted. In contrast to this, ion-beam mixing with Ge or As ion doses at or above 5 × 10 14 cm -2 and subsequent annealing led to good adhesion of Co on SiO 2 without damage of the oxide surface.

Diffusion and precipitation of As from a CoSi 2 diffusion source

Arsenic ions were implanted into CoSi 2 thin layers and subsequently diffused into the underlying silicon substrate. Junctions in the range between 30 and 400 nm depth were obtained with a high dopant concentration (10 19-10 20 cm -3) at the silicide-silicon interface. Coherent Si-As precipitates were found at the silicide-silicon interface near the CoSi 2 grain boundaries. The precipitate formation has been associated with the high tensile stress induced by the silicide layer on the surface silicon region and to its influence on the dopant solid solubility. The precipitate density decreases if the silicide thickness is reduced, i.e. if the tensile stress produced by a silicide layer gets lower. These precipitates are easily dissolved during high temperature annealing after the silicide layer removal. A new delineation technique based on TEM analysis of selectively etched samples has evidenced that the intersections of silicide grain boundaries with a silicon surface act as diffusion sources.

Improved stability of thin cobalt disilicide films using BF2 implantation

The thermal stability of ∼50 nm CoSi2 and TiSi2 thin films after BF2+ implantation was investigated. The electrical characteristics of silicide films were evaluated after high temperature annealing as a function of implanted BF2+ energy. It was observed that implantation with a projected range near the silicide/silicon interface produced the most stable films. The silicide/silicon interface morphology was investigated using scanning tunneling microscopy, where with appropriate BF2 implantation conditions, smoother interfaces were seen after high temperature annealing. The stabilizing effect is attributed to fluorine segregation into the silicide grain boundaries and at the silicide/silicon interface.

The effect of oxygen impurity on growth of molybdenum disilicide and its distribution during rapid thermal annealing of co-sputtered MoSi x thin films

Thin films of MoSi x (with and without oxygen) were deposited by co-sputtering of molybdenum and silicon targets. As-deposited films were annealed in argon environment in the temperature range 700–1150 °C for 15–120 s. Rapid thermal annealing of oxygen-contaminated films at low temperature (700–800 °C) in an argon environment results in evolution of a mixture of MoSi 2, Mo 3Si 3, polycrystalline silicon and amorphous phases. Subsequent annealing at higher temperature (above 800 °C) results in further growth of Mo 5Si 3, MoSi 2 and polycrystalline silicon. At 1150 °C, a 15–120 s annealing time produces successive growth of tetragonal MoSi 2 phase at the expense of Mo 5Si 3 and silicon. Auger electron spectroscopy depth profiles of annealed films indicate the accumulation of oxygen impurity at the MoSi x Si(crystalline) interface, which is initially distributed uniformly throughout the MoSi x matrix. However, oxygen-free MoSi x films do exhibit the growth of MoSi 2 (both tetragonal and hexagonal) during rapid thermal annealing up to 1050 °C. At higher temperature (1150 °C) a single-phase MoSi 2 (tetragonal) is observed. The surface morphology of oxygen-free annealed films reveals a uniform distribution of silicide grains as compared with oxygen-containing films. The sheet resistance of the finally annealed films (1150 °C and 120 s) was found to be 9.5 Ω/◊ for oxygen-contaminated films and 7.6 Ω/◊ for oxygen-free films.

Aspects of the selective deposition of TiSi 2 by LRP-CVD for use in ULSI submicron technology

TiSi 2 layers are deposited in a short time cycle using a rapid thermal processing system at reduced pressure and temperature, called LRP for limited reaction processing. The use of TiCl 4 from a liquid or from a solid source gives selective layers with minimal or no substrate consumption. However, two main problems are studied in this paper which have a strong effect on a technological process: substrate doping behavior under silicide growth and post-annealing reactions and silicide grain size. We present some recent results on the latter as well as recall our interpretation of the TiSi 2 formation from a kinetic point of view. Some technological results on 4 inch wafers will also be presented.

Characteristics of LPCVD WSi/sub 2//n-Si Schottky contacts

Detailed current-voltage and capacitance-voltage characteristics of low-pressure chemical vapor deposited (LPCVD) WSi/sub 2//n-Si Schottky contacts are reported in the temperature range of 21 to 170 degrees C. The diode ideality factor n was found to decrease from a value of 1.46 to 1.15 as temperature was increased. Schottky barrier height phi /sub B/, on the other hand, was found to increase from 0.72 to 0.86 V with temperature. These results suggest that diode characteristics are affected by surface and bulk effects, especially at lower temperatures. High-resolution transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy analyses revealed isolated regions of oxynitride at the silicide/silicon interface that are predominantly located where silicide grain boundaries intersect the silicon surface.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Grain boundary diffusion and growth of titanium silicide layers on silicon

A kinetics study of titanium silicide formation is described. The results show that a fine grained precursor layer exist in between the well developed C-54 silicide layer and the unreacted titanium film. This layer is a mixture of C49-TiSiV2 and unreacted titanium. The fact that no C54-TiSi2 formed directly from the Ti-Si reaction suggests that the nucleation of C49-TiSi2 is easier than that of C54-TiSi2 under our annealing conditions. The silicide layer growth has a non-t1/2 dependence and is much better described by a grain boundary diffusion limited model giving different kinetics. This indicates that grain boundary diffusion is the major atomic transportation mechanism. The growth rate depends on both the grain boundary diffusion coefficient and the silicide grain growth rate.

Lattice and grain boundary self-diffusion in Ni2Si: Comparison with thin-film formation

The Ni2Si lattice and grain boundary self-diffusion parameters have been studied using radiotracers 63Ni for Ni diffusion, 68Ge for Si diffusion, conventional sectioning techniques, and bulk specimens. The results obtained show that Ni diffuses faster than Ge(Si) in the lattice (and in the grain boundaries) of Ni2Si, in agreement with the structure of this silicide. The activation energies for Ni lattice and grain boundary diffusion are respectively 2.48 eV (measured between 650 and 910 °C) and 1.71 eV (530–710 °C). These values are comparable to those obtained in pure metals of similar melting temperature. They suggest that Ni diffuses on its own sublattice by a vacancy mechanism. Grain boundary diffusion in this intermetallic compound appears as a very efficient process for mass transport at low temperature. A quantitative comparison with the kinetic of formation of Ni2Si (by solid state reaction between a Ni film and a Si substrate) indicates that the growth of thin films is controlled by the diffusion of Ni along the silicide grain boundaries. This mechanism explains the rapidity of the formation and its low activation energy (≊1.6 eV).

On The Nature of Multiple Orientation Relationships in Epitaxial Silicides

ABSTRACTThe Invariant Line Criteria (ILC) has been used to explain experimentally observed precipitation and epitaxial orientation relationships in a number of metallurgical systems. We propose that ILC may explain the crystallography of misoriented silicide grains in epitaxially grown silicide films. We test our hypothesis on the case of misoriented grains in CoSi2 on Si (001). Our analysis shows that the experimentally observed orientation relationship is predicted by the ILC.

Influence of Silicon Substrate Ion Implantation on the Subsequent Microstructure Evolution in Cobalt Silicide Films

ABSTRACTThe influence of amorphization of the Si substrate by Ge+ implantation on the microstructural evolution of CoSi2 thin films has been investigated by transmission electron microscopy (TEM), X-ray diffraction (XRD) and four-point-probe measurement. While the silicide films formed on unimplanted or recrystallized Si were similar in formation kinetics, microstructure and resistivity, unique features were found for silicide films formed on amorphized Si. The phase transformation to CoSi2 on amorphized Si was considerably accelerated, the silicide grain size was smaller, and the CoSi2/Si interface was much smoother on the amorphized Si. Direct evidence of grain boundary diffusion induced silicide/Si interface roughness was obtained. The origin of the difference in both the microstructure and kinetics on different Si substrates is explained by the latent energy stored in amorphous Si, increasing the silicide nucleation rate. Weak XRD peaks and different texture were found for the CoSi2 formed on amorphized Si. The facts that the dominant diffusing species during CoSi2 formation is Co, and the new phase was developed in an amorphous medium are thought to be responsible for these phenomena.

Properties of CoSi2 formed on (001) Si

Properties of CoSi2 films formed on (001) Si have been investigated for silicide films formed by solid-state reaction of either evaporated or sputtered Co with (001) Si substrates. It turns out that CoSi2 films formed from evaporated Co have a systematically higher resistivity than films formed from sputtered Co, which can be attributed to contaminants that were present in the evaporated metal film. Silicide films formed from evaporated Co develop a strong (110) texture, which becomes more pronounced with increasing annealing temperature. Microstructural analyses revealed that this is due to epitaxial growth of (110) CoSi2 upon (001) Si. This is a surprising phenomenon as the epitaxy of (001) CoSi2 on (001) Si would be expected in view of the similarity of the crystal structure and the small lattice mismatch of CoSi2 and Si. The expected epitaxy of (001) CoSi2 on (001) Si is also present but for a smaller fraction of the silicide grains. The film stress has been analyzed in terms of thermal and intrinsic stress. From the thermal stress contribution, a value of 140 GPa was calculated for the biaxial elastic constant of CoSi2. Using this value, it was also possible to interpret the intrinsic stress in terms of the density change upon silicide formation. Transmission electron microscopy and x-ray diffraction revealed that CoSi2 can not only grow epitaxially on (001) Si in a (001) and a (110) orientation but that there is also epitaxy close to these orientations with tilts of around 10° relative to the substrate normal. Simple geometrical calculations show that there is a good match for both (001) and (110) CoSi2 on (001) Si and for a slight tilt from these orientations, all in very good agreement with the observations.

Characteristics of Selective Lpcvd W Films By Silicon Reduction

AbstractThe characteristics of Selective LPCVD tungsten films produced by silicon reduction of tungsten hexafluoride are presented. The tungsten films deposited using Si(100), Si(111) and polysilicon undoped and doped substrates are analyzed by X-RAY, TEM, RBS, AES, SIMS and SEM. The as deposited bcc tungsten films are polycrystalline with a grain size 80 - 100Å. The effect of annealing temperature and time on the crystal structure of films was studied. Tungsten reacts to form tungsten silicide at 600°C. The silicide grain size is of the order of 100 - 200Å at 600°C and increases gradually to 400 - 500Å at 1000°C. The oxygen impurities in the film retard the silicide formation further at 1000°C. Silicon from the substrate out-diffuses to the film surface and reacts with the presence of oxygen impurities in the annealing ambient to form Si-O at 1000°C. As deposited film resistivities of 130-140 micro-ohm-cm are achieved reproducibly and reach 60-70 micro-ohm-cm after 1000°C annealing in nitrogen or argon ambient. The impurities H, C, O and F are found in the as deposited films.

The solution growth of silicide layers on molybdenum substrates

A saturated solution of silicon in molten tin is employed to grow a silicide layer on molybdenum substrates. The growth is carried out in a chamber analogous to a vertical liquid-phase epitaxy system by dipping the rotating substrate into the melt at temperatures in the range 900 to 1000° C, and using different dipping times of up to 60 min. By X-ray diffraction analysis, the deposits appear to contain only molybdenum disilicide (MoSi2). Examination by other characterization techniques such as optical microscopy, scanning electron microscopy and electron probe microanalysis give the layer thickness in the range 3 to 20μm, and the silicide grain size of the order of 0.1μm. Deposited layers appear as a rough-surfaced agglomeration of silicide grains with occasional cracks probably formed on cooling. The grains are generally equi-axed and efficiently packed into a film of fairly low porosity, there seems to be a linear correlation between layer thickness and deposition time. This suggests that the kinetics of MoSi2 formation by this technique may be controlled by the rate of the interface reaction at the metal-silicide interface.