C.-D. Lien

Kinetics of CoSi2 from evaporated silicon

Abstract2 MeV4He+ backscattering spectrometry and CuKα x-ray diffraction were used to study CoSi2 formed by annealing at temperatures between 405° and 500 °C from CoSi with evaporated Si films. A laterally uniform layer of CoSi2 forms, in contrast to the laterally nonuniform CoSi2 layer that is obtained on single crystal Si substrates. The thickness of the CoSi2 film formed is proportional to the square root of time at a fixed temperature. The activation energy of this reaction is about 2.3 eV.

Growth of Co-silicides from single crystal and evaporated Si

We have investigated the reaction of a thin Co film with a (100) Si (Sic) or an evaporated Si (Sie, which is amorphous) substrate during thermal annealing. On either substrate, Co2Si and CoSi form simultaneously and the growth of each phase has a square root of time dependence. Both silicides grow faster on Sic than on Sie. A model is proposed to calculate the effective diffusion constant in each silicide from the growth data of the silicides. The activation energies of the effective diffusion constants in Co2Si and CoSi grown on Sic are 1.7±0.1 eV and 1.8±0.1 eV, respectively; while those on Sie are 1.85±0.1 eV and 1.9 ±0.1 eV, respectively. The differences observed for the two substrates are tentatively attributed to the presence of impurities in Sie and to the microstructural differences of the silicides formed on either substrate.

Electrical properties of thin Co2Si, CoSi, and CoSi2 layers grown on evaporated silicon

MeV4He+ backscattering spectrometry and CuKα x-ray diffraction were used to measure the thicknesses and phases of Co-silicides grown from samples with a thin Co layer deposited on a thin Si layer evaporated on a SiO2 substrate (SiO2/Sie/ Co). Uniform layers of Co2Si, CoSi, and CoSi2 were formed sequentially in a layer-by-layer manner. Electrical properties of these silicides were studied with four-point probe measurements. The results show that Co2Si and CoSi2 are p-type conductors and CoSi is n-type conductor with resistivity close to the reported values. Results from measurements that were performed on samples with two phases of silicide present (e.g. SiO2/ Sie/CoSi2/CoSi) are explained very well by a simple model which assumes two layers of silicides connected electrically in parallel.

An inert marker study for palladium silicide formation: Si moves in polycrystalline Pd2Si

A novel use of Ti marker is introduced to investigate the moving species during Pd2Si formation on 〈111〉 and 〈100〉 Si substrates. Silicide formed from amorphous Si is also studied using a W marker. Although these markers are observed to alter the silicide formation in the initial stage, the moving species can be identified once a normal growth rate is resumed. It is found that Si is the dominant moving species for all three types of Si crystallinity. However, Pd will participate in mass transport when Si motion becomes obstructed. off

A structure marker study for Pd2Si formation: Pd moves in epitaxial Pd2Si

A sample with the configuration Si (111)/single crystalline Pd2Si/polycrystalline Pd2Si/Pd is used to study the dominant moving species during subsequent Pd2Si formation by annealing at 275 °C. The interface between monocrystalline and polycrystalline Pd2Si is used as a marker to monitor the dominant moving species. The result shows that Pd is the dominant moving species in the monocrystal. off

Profile of tracer Si in silicide when Si diffuses by vacancy mechanism

A mathematical model is constructed to interpret the profiles of radioactive Si tracers during silicide formation. This model assumes that only Si moves in the silicide during silicide formation and that the moving Si diffuses in the Si sublattice of the silicide in terms of vacancy mechanism. Analytical solutions of the model for long‐time annealing (i.e., asymptotic profiles) are given. The analytical asymptotic profiles are very accurate for the annealing period generally used in experiments. It is shown that the profiles of the Si tracer in the silicide are almost flat. This thus proves that self‐diffusion of the tracer atoms cannot be neglected as assumed in some published papers. In fact, several experimental tracer profiles are found to be flat in the silicide. Some numerical solutions for short‐time annealing are also given to show how the tracer profile evolves. The result given here can also be used for many intermetallic reactions.

Mathematical model for a radioactive marker in silicide formation

A mathematical model is constructed to interpret the profiles of radioactive (^31)Si tracers in a computer simulation proposed by R. Pretorius and A. P. Botha [Thin Solid Films 91, 99 (1982)]. This model assumes that only Si moves in the silicide, that the Si moves interstitially and convectively, and that the moving Si can exchange sites with the stationary Si in the silicide lattice. An analytical solution of this model is given and confirms the published computer simulation data. However, it is shown that the model is physically inadequate. Solutions of another model which assumes that metal, instead of Si, is the moving species for silicide formation (either interstitially, or substitutionally, or both), with self-diffusion of (^31)Si in the silicide during silicide formation. Almost all the experimental data can be fitted by solutions of both models. These examples demonstrate that radioactive tracer experiments alone are insufficient to determine the moving species when a solid binary compound film forms by reaction of adjacent elemental layers. Both inert marker and tracer data are needed to identify the moving species and the mechanisms.