The maskless ion implantation with the focused ion beam as a new method for ion beam synthesis of cobalt sulicide wires is presented. In order to perform the implantation a special achromatic mass separator was implemented into the ion column, liquid alloy ion sources for cobalt ions were developed and a substrate heating was built. Ion implantation was performed with 30 keV Co+ and 60 keV Co++ ions. The dose dependence for room temperature implantation and the influence of the substrate temperature were investigated.

Heteroepitaxial CoxNi1−xSi2 layers with good crystalline quality have been formed by ion beam synthesis. Rutherford Backscattering (RBS) - Channeling, Auger Electron Spectroscopy (AES), Transmission Electron Microscopy (TEM) and X-Ray Diffraction (XRD) have been used to study the buried layers. For a sample with x=0.66, we found that this ternary silicide layer contains 11% type B and 89% type A orientation. The TEM investigation reveals that the type B component is mainly located at the interfaces with a thickness of a few monolayers. XRD studies show that the strain of the type B component is smaller than that of the type A component, and this is probably the reason for such a unique distribution of the type B component in the epilayer.

The formation of a Si/IrSi3/Si heterostructurie by 1-MeV Ir ion implantation and subsequent annealing has been studied for different doses (0.1-2.25 × 1017 Ir/cm2), substrate temperatures (450°-600°C) and annealing temperatures (1000°-1200°C) using Rutherford backscattering spectrometry, ion channeling and cross-sectional transmission electron microscopy. The heterostructure formation is observed to depend strongly on the processing conditions. The best structure, nearly continuous and precipitate-free, is obtained by implanting 1.8-2.0× 1017 1r/cm2 at a substrate temperature of 550°C and annealing at 1100°C for 5 h. A stoichiometric IrSi3 layer can also be produced by furnace annealing at 1150°C for 1 h or by rapid-thermal-annealing at 1200°C for 3 min. Other substrate temperatures generally lead to a structure with a discontinuous IrSi3 layer frequently interrupted by large surface precipitates or islands. The origin of these islands, as well as the dependence of the heterostructure on processing parameters, is discussed.

In this paper optical absorption and photoluminescence (PL) techniques are used to study the optical properties of βFeSi2 layers fabricated by Ion Beam Synthesis (IBS). The way in which the band-gap varies when a small dose of Co is added before Fe implantation is also investigated. Our results indicate that the band-gap for βFeSi2 is direct with a value of 0.87eV. When a sufficiently high dose (1 ×10 16cm2) of Co is implanted prior to Fe we see the appearance of an additional band-gap after annealing at 1173K which we attribute to the formation of a ternary CoxFeySiz phase. We also show that by placing the sample in a hydrogen plasma it is possible to raise the temperature at which PL occurs in βFeSi2 to σ 200K.

<111> -Si was implanted with 250 keV Co ions at a target temperature of 350°C. The ion dose was varied between 1 × 1014 cm−2 and 2 × 1017 cm−2. The evolution of the defect and impurity profile was investigated by Rutherford Backscattering Spectrometry (RBS), channeling and transmission electron microscopy (TEM).

Up to a dose of 1 × 1015 Co cm−2 no defects can be detected. At higher Co doses, we find correlated defects in the center of the Co distribution and point defects in the region below. Moreover, damage accumulation at the surface is observed. The concentration of defects increases with increasing ion dose and reaches its level of saturation at a dose of 2 × 1016 cm−2.

The Co profiles of samples implanted at 350°C differ considerably from the Gaussian shape. The near surface and the back flank are parts of Gaussian distributions. However, the standard deviation of the near surface flank is always smaller than that of the back flank. Moreover, the distributions show tails into the substrate at depths > 320 nm. This proves that radiation damage acts as an effective center for the nucleation of CoSi2.

During annealing we find a redistribution of Co towards the defective regions for Co doses between 1 × 1016 cm−2 and 5 × 1016 cm−2.

A buried epitaxial IrSi3 layer has been produced in (111) oriented Si by high dose Ir+- implantation (1 MeV, 1.7× 1017 Ir+/cm2, 550°C) and subsequent annealing (5 h at 1100°C). Transmission electron microscopy and ion channeling show that the hexagonal IrSi3 prefers to form with its (2110) plane oriented parallel to Si(111). while its [0001] direction is parallel to either [112], [121], or [211]. Thus, the IrSi3 film is made up of three differently oriented crystals, each occupying about 1/3 of the silicide volume.

A dose of 3.5× 1016 1r+/cm2 leads to the formation of a band of large, isolated precipitates. They have the same epitaxial relationship to Si, but the Si(111) plane is no longer the only one to which the (2110) of IrSi3 is parallel. Instead precipitates are also observed oriented parallel to (111), (111), and (111) in Si in roughly equal numbers. Consequently, there are now 12 possible orientations of lrSi3 crystals. We assume that strain relief and interface area minimization are the main factors suppressing the additional orientations in the continuous lrSi3 films formed by the coalescence of precipitates at higher doses.

Metal Vapour Vacuum Arc (MEVVA) Ion Source was employed, for the first time, to synthesize metal silicides. Refractory metal silicides NbSi2, TaSi2, WSi2 and MoSi2; were successfully formed by implanting respective metal ions into Si(111) wafers. It was found that the high current density metal ion implantation not only provided the metal components but also caused a simultaneous annealing for various metal-silicide phases to grow. When the imp ytation was conducted at 40 kV, the Nb ion current density was up to 76 μA/cm at a dose of 4×1011/cm2, and the Ta, W and Mo ion current densities were up to 65 μA/cm2 at a dose of 5×1017 ions/cm2, hexagonal NbSi2, TaSi2, MoSi2 and WSi2 were formed. Post annealing transformed the WSi2 and MoSi2 phases from hexagonal to tetragonal structure, featuring lower resistance than that of the as-implanted ones.

By subsequent implantation of iron and cobalt into silicon buried layers of semiconductinβ-(Fe1−xCox)Si2 are formed. The band gap energy as well as the spreading resistance decrease with increasing cobalt content. The microstructure of the sulicide layers is characterized by XTFM images.

Epitaxial CoSi2 layers on Si substrates have been amorphized with Co and/or Si ion implantation. The influence of nonstoichiometry on the rate of solid-phase epitaxial growth (SPEG) of amorphized CoSi2 has been investigated with time-resolved reflectivity, Rutherford backscattering spectrometry and Mossbauer spectrometry, the latter with radioactive 57Co probes. A decrease in SPEG rate was apparent with an increase in nonstoichiometry. For a given ion dose, the decrease was greater following Co implantation. The means by which non-stoichiometry is accommodated in a crystalline CoSi2 lattice - either through phase separation or defect formation - has been considered. SPEG rate retardation was also evident in samples implanted with both Si and Co ions with a Si:Co dose ratio of 2:1. Additional mechanisms may thus also contribute to the observed SPEG rate reduction.

We have performed electrical transport measurements on thin epitaxial buried CoSi2 layers in (111) and (100)Si with thicknesses ranging from 11.5 to 110 nm. The resistivity as a function of temperature exhibits metallic behaviour. The increase of residual resistivity with decreasing thickness can be explained by quantum mechanical weak localization effects induced by the interface roughness. This is supported by magnetoresistance measurements which provide long phase coherence lengths of 1Փ, ≈ 0.75 μm in (111)Si and 1Փ, ≈ 2.3 μm in (100)Si at 4.2 K.

The PtSi/Si interface is of technological interest for Schottky barrier infrared detectors. We are studying PtSi/Si heterostructures using ballistic -electron-emission microscopy (BEEM), an STM-based technique that uses the STM tip to inject hot electrons at a particular energy into the metal overlayer. The BEEM technique allows imaging of the Schottky barrier with good spatial resolution (of the order of tens of nanometers) and allows the measurement of the hot electron attenuation length in the metal overlayer. Our results indicate a Schottky barrier of 0.87 eV for PtSi/Si n-type, and an attenuation length of 4 nm for electrons with an energy of 1 eV above the metal Fermi level. The attenuation length we measure is a convolution of the electron elastic and inelastic mean free path lengths.

We have also used an ac BEEM technique to observe inelastic scattering events at the metalsemiconductor interface in PtSi/Si(100) n-type. There are several features visible in the spectrum, including one at 1040 meV which we attribute to optical phonon-assisted electron-hole pair creation near the metal- semiconductor interface in analogy to a feature we have observed at the same energy in the Au/Si(100) ac BEEM spectrum. Higher-energy features appear at 1230 meV and 1300 meV. Similar features appeared in the Au/Si(100) spectrum at 1120 meV and 1230 meV.

We also suggest that the traditional assumption of momentum conservation parallel to the Schottky barrier interface is unnecessary to obtain a quadratic turn on of the BEEM current above the threshold. If electrons are elastically scattered at the interface so that momentum is not conserved, the increase in the ratio of the density of states in the semiconductor to those in the metal will also give a quadratic turn on even when the band structure is much more complicated than a nearly free electron model. This model also explains why the simple square-root dependence of the photoresponse on wavelength above threshold observed in all metal/semiconductor Schottky barriers despite the complications in band structure.

The kinetics and phase formation sequence of thin-film solid-state reactions, including silicide forming reactions, have frequently been considered to be controlled by interfacial kinetic reaction barriers. These are purely phenomenological quantities which describe the finite rate of the interfacial reaction in terms of limited particle fluxes crossing the respective interfaces. No atomic mechanisms that might be responsible for the action of such barriers have so far been indicated, with the exception of Schmalzried's formulation. The latter says that the interfacial barrier is due to the limited relaxation time needed by the particles to rearrange into the proper sublattice after having crossed the interface. We present correlated kinetic and structural observations during the 2Ni + Si → NiSi2 reaction on the Si(111) surface and discuss them with the help of a model involving the formation and lateral propagation of interfacial steps of different height. The model allows us to explain the kinetic observations by reaction barriers formed as a result of the crystallographic boundary conditions of the reaction.

This paper describes a high-resolution transmission electron microscope (HRTEM) study of the morphology and the atomic structure of CoSi2/Si(001) interfaces of continuous buried sulicide layers. These layers were produced by high-dose Co+ ion implantation and subsequent rapid thermal annealing. Planar interface regions of high perfection with domains of different atomic interface structure, and interface steps, frequently with I 111)I facets, were observed. There were significant differences in the interface structure between the upper and lower interfaces. Measuring Schottky barrier heights (SBH's) revealed different values for the upper (0.67eV) and lower (0.78eV) CoSi2/n-Si(001) interfaces. Possible correlations between the atomic interface structures and the resulting electronic properties are discussed.

A slow positron beam was used to investigate the solid state reaction of Co/Si and Co/Ti/Si. Variable-energy (0-20 keV) positrons were implanted into samples at different depths. The Doppler broadening of the annihilation -y-ray energy spectra, measured at a number of different incident positron energies were characterized a line-shape parameter “5”. It was found that the measured S parameters were sensitive to thin film reaction and crystalline characteristics. In particular, the S parameter of epitaxial CoSi2 formed by the ternary reaction was quit different from that of the polycrystalline CoSi2 formed by direct reaction of Co with Si.

We discuss atomistic aspects of the silicide contact reaction inferred primarily from STM observations of the prototype system Co/Si(1 11). For room temperature deposition and low coverage (0.01M1) we find that metal atoms exist as near-surface interstitials within the 7×7 reconstruction. Bond breaking associated with silicide formation occurs only at higher coverages. Deposition at 320C results in flat-topped triangular islands of epitaxial CoSi2 with a metastable 7-fold (111) interface, stabilized by the lateral silicon-silicide interface along the island edges. Some islands are covered with a 2×2 array of silicon adatoms. Very high temperature annealing (1200C) results in an “impurity stabilized 1×1” surface which is in fact a lattice gas of ring-clusters that appear like tiny donuts or bagels in empty states STM images. These structures phase-separate from the clean 7×7 structure upon cooling below 850C.

The interface structure and Schottky barrier height Of CoSi2/Si(111) interfaces may be controlled by manipulating the thin film growth conditions. Single crystal CoSi2 films on Si(111) were prepared by ultra-high vacuum processing, analyzed electrically by currentvoltage techniques, and characterized structurally by plan-view and cross-section high resolution transmission electron microscopy (HRTEM) and transm-ission electron diffraction (TED). Interfaces exhibiting n-type barrier heights ranging from 0.27 to 0.69 eV, and p-type barrier heights ranging from 0.43 eV to over 0.71 eV were prepared -by varying the processing conditions. HRTI3M and TED revealed the existence of a √3 × √3 interface reconstruction for the low barrier n-type/high barrier p-type samples. Possible models of the interface reconstruction are discussed.

Si L2,3 valence band soft x-ray emission spectrum (SXES) due to an e ectron excitation for silicides shows a clear modification from that for Si single crystal. Using this fact in combination with the incident angle variation(IAV) device, a non-destructive in-depth analysis of a Au(thin film)-Si(lll) contact is successfully carried out. Also, the SXES method has clarified the fact that a fair amount of the Si-s valence band density of state (yB-DOS) is included in the upper part of the yB-DOS for a Au-Si alloy, or Au-silicide, due to the Au-Si bond formation, which is a clear contrast to proposals given so far.

A laterally uniform formation of NiSi2 at low temperature had been reported for the SiaNi/Sic system. In this work, the interfacial structure of NiSi2 is studied by the high-resolution electron microscopy. We observe an amorphous layer (5–10 nm in thickness) between Sic and NiSi2 in a Sia(300nm)/Ni(5Onm)/Sic sample after annealing at 450°C. The NiSi2 phase in this sample has a polycrystalline structure. We also find that a contaminated amorphous-Si film can not promote a reaction with Ni. A SIMS depth profile shows that this amorphous-Si layer is high in oxygen and carbon. In this case, the silicide formation is similar to the Ni/Sic system.

We present results of electrical transport studies performed on thin films of ε1-Cu3Ge in the temperature range 4.2 – 300 K. It is found that ε1-Cu3Ge which has a long-range ordered monoclinic crystal structure, exhibits a remarkably low metallic resistivity of σ 6 μΩ cm at room temperature. The density of charge carriers, which are predominantly holes, is σ 8 × 1022/cm3 and is independent of temperature and film thickness. The Hall mobility at 4.2 K is σ 132 cm2/V s, considerably higher than in pure copper. The elastic mean free path is found to be σ 1200Å, which is surprisingly large for a metallic compound film. The results show that the residual resistivity is dominated by surface scattering rather than grain-boundary scattering. It is also found that by varying the Ge concentration from 0 to 40 at. % the resistivity exhibits anomalous behavior. This behavior is correlated with changes observed in the crystal structure of the thin-film alloys as the Ge concentration is increased. The resistivity remains close to that of the ε1-Cu3Ge phase over a range of Ge concentration which extends from 25 to 35 at. %

The electrical properties of Cu/Si(100) and Cu3Si/Si(100) interfaces have been studied using both n- and p-type silicon samples. Current-voltage and capacitance-voltage measurements were performed in the temperature range 80-295 K in order to monitor Schottky barrier formation and electrical carrier concentration profiles. Deep-level transient spectroscopy was employed to observe Cu-related energy levels in the forbidden band gap of Si, and different ion beam analysis techniques were applied to study the interfacial reaction between Cu and Si. Emphasis is put on determination of Schottky barrier heights and their variation with temperature, dopant passivation by Cu atoms and interaction of Cu with irradiation-induced point defects in silicon.