The composition dependence of crystallization of amorphous Co-Si alloys has been investigated. Crystallization temperature decreased slowly from 400°C for a Co0.42Si0.58 alloy to 300°C for a CO0.2Si0.8 alloy. However, abrupt decrease in the crystallization temperature was observed for alloys with the monosilicide and disilicide compositions. The crystallization of alloys compositionally close to the disilicide was a one-step process with an activation energy of about 1.3 eV. The heat release and the volume contraction during crystallization of the stoichiometric Co0.33Si0.67 alloy were measured to be 0.118 eV/atom and 0.6%, respectively. In the case of the CO0.2Si0.8 alloy, complete crystallization was achieved in two steps with an activation energy of 2.1 eV for the primary crystallization.

CoSi2, or TiSi2 formation on gate polysilicon can degrade the current-voltage and capacitance-voltage characteristics of MOS capacitors. Degradation of the gate oxide breakdown field is much more severe for capacitors with TiSi2 than for those with COSi2 TEM reveals evidence for a reaction at the interface between TiSi2 and SiO2, whereas there is no observable reaction between COSi2 and SiO2- The low breakdown fields for devices with TiSi2 may be due to thinning of the gate oxide by the interfacial reaction or mechanical deformation. A high density of electron traps and a small reduction in the breakdown field is observed when COSi2 contacts the gate, possibly due to a compressive stress in the oxide exerted by the suicide. In addition, an increase in the interface state density at the Si-SiO2 interface is seen for all samples exposed to a rapid thermal anneal (RTA) at 800°C, possibly due to the release of H from dangling bonds.

The formation of a 12nm thick, continuous and thermally stable COSi2 layer was described in our previous work [MRS Proc. 238, 587 (1992)]. Interdiffusion in the Co/Ti-Si multilayer system has been further studied and the initial Ti(O) thickness is shown to be a critical parameter in controlling its effectiveness as a diffusion barrier, and in modulating the Co-Si and Ti-Si compctctive reactions. Three Ti(O) and three Co layers with thickness from ∼5nm 20nm were deposited sequentially, with Ti(O) as the first layer, on Si-(100) substrates by dual source thermal evaporation. The morphology of the CoSix/Si interface was strongly influenced by Ti(O) thickness from ∼5nm to ∼10nm, and a 12nm thick uniform CoSi2 layer with ∼28μΩ-cm resistivity was produced as decribed previously. When the initial Ti(O) thickness was increased to ∼20nm and the Co thickness set at -10nm, Co diffusion was suppressed and Ti reacted with Si yielding an ∼10nm amorphous TiSix layer at 550°C. This amorphous layer transformed to a 15nm thick uniform C-54 TiSi2 layer after selective removal of upper layers and a 750°C plus 800°C annealing. A flat silicide/Si interface and a ∼58μΩ-cm resistivity were obtained. The significance of both thermodynamic and kinetic factors in the compctetive reactions is discussed.

This study explores the nucleation and surface and interface morphology of TiSi2 films formed on Si (100) and (111) surfaces. Titanium suicide films were formed on atomically clean Si(111) and (100) surfaces by UHV deposition of Ti followed by in situ anneals. The TiSi2 formation process was characterized is itu using AES and LEED, and the interface structure was determined post preparation by HRTEM. The metastable C49 and stable C54 TiSi2 phases have been identified from Raman spectroscopy measurements. The islands were partially recessed into the substrate, and the shape and degree of recession was different for the islands on (111) and (100) substrates. On (111) substrates epitaxial islands were observed which exhibited different interface structures and epitaxial alignments. The contact angles of the islands were measured and the surface and interface energies of the different configurations were deduced. The mechanisms pertaining to island formation and interface morphology are described in terms of the wetting properties of a liquid-liquid system, with the effects of interface structures causing deviations from the model.

In this work we have studied the effect of polysilicon doping on titanium disilicide (TiSi2) formation using extensive physical and electrical characterization.

Thermal wave imaging can be used as a rapid non-destructive in-line probe to monitor defects in metals. In the study of the Ti-polysilicon system, this technique has been used to quantify the degradation of the suicide film by monitoring the formation of a second phase during BPSG annealing. The amount of second phase formed was found to vary with Ti preclean, dopant concentration and temperature of BPSG anneal on both blanket and patterned wafers. The second phase has been identified using TEM and AES to be a silicon rich phase. Digitized thermal wave images were used to obtain information regarding the size distribution of precipitates. These distributions have been correlated with resistivity changes in submicron suicide lines.

Picosecond ultrasonics is employed to study the titanium silicide formation sequence for evaporated Ti films on silicon substrates annealed at temperatures between 300 and 800 °C. The measurements show significant differences in the ultrasonic echo pattern before and after the structural phases C49 and C54 are formed, thus indicating that picosecond ultrasonics is a sensitive non-destructive probe of silicide formation. The longitudinal sound velocity has been found to be (8.3 ± 0.2) × 105 cm/sec for C49 TiSi2, and about 5% lower for the C54 phase.

The interaction of titanium thin films sputter-deposited onto single crystal silicon was studied by in situ heating experiments within the TEM. Reactions at the Ti-Si interface including amorphization, crystallization, allotropie phase transformations and agglomeration have been observed in real time and recorded on videotape. Interpretation of these recordings can yield a wealth of information on the silicidation process.

To minimize the bridging problem In the self-aligned titanium slllcide (SALICI DE) processes, a titanium nitride (TiN) cap has been suggested. The objective of this work was to study the impact of such a TiN cap layer on a Ti-based SALICI DE process. The slllcide formation was carried out by Rapid Thermal Annealing (RTA). The Ti-Si Interaction, the dopant loss during silicide formation, the contact resistance between the formed TiSi2 and SI substrates, the leakage current, and the lateral diffusion of SI atoms over the spacers of MOS devices were investigated with and without a TiN layer on top of Tl films. The dependence of these factors on the annealing temperature was also investigated. The results showed that: a) An increased dopant loss was a direct result of more silicide formation, due to a larger SI substrate consumption. This was observed both with and without TiN cap. b) The specific contact resistance increased with annealing temperature, regardless the TiN cap. c) Sheet resistance, specific contact resistance and leakage current were not sensitive to the selective etch procedure. And d) A lower yield for bridging test structures was found where a TiN cap was employed, Indicating that the TiN layer did not suppress the Si lateral diffusion.

IBS of buried α and β iron suicide layers was achieved by the implantation of 2 MeV 56Fe+ ions into (100) single crystal silicon substrates over a dose range of 3 × 1017 to 1 × 1018 cm“-2followed by a high temperature anneal. No photoluminescence was observed from the as-implanted samples which contained a discontinuous layer of βFeSi2 precipitates approximately 1.5 μm below the silicon surface. Upon annealing at 700°C, a 200 nm polycrystalline βFeSi2 layer was formed which gave a PL signal centred at 1.55 μm. After a 900°C anneal, the layer transformed to αFeSix with a resistivity of approximately 280μΩcm.

We present a new analytical theory of electronic transport for inhomogeneous Schottky contacts. Our model combines a novel method to solve Poisson's equation for the space charge region of an inhomogeneous contact with thermionic emission theory. We explain quantitatively the temperature and voltage dependences of the ideality factor n of current/voltage characteristics, and we are able to extract from experimental data information on the characteristic size of the inhomogeneities. Our measurements on polycrystalline and epitaxial Schottky contacts yield values in the range of 200–500nm as the typical size of inhomogeneities, covering a fraction of 0.1–1% of the total interface area.

The morphology of zirconium suicide grown on Si(111) is investigated and compared to that of titanium suicide on Si(111) and (100). Zirconium was deposited on atomically clean silicon in UHV, and the suicide was formed by postdeposition annealing. ZrSi2 films formed from ≤100 Å of Zr deposited on Si(111) remained continuous for anneals of 400°C to 700°C. A film formed from 50 Å of Zr on Si(111) remained continuous for anneals up to 800°C, at which point the reappearance of a partial 7×7 LEED pattern indicated island formation. The interfaces of the films, studied by TME, show roughening after suicide formation. These results are compared and contrasted with the islanding behavior of titanium suicides, and analyzed in terms of a liquid-liquid model previously used with TiSi2.

In this study, a novel model has been proposed to describe reactions in thin metal film/silicon diffusion couples. A reactive interface in a metal-Si diffusion couple is considered to be a reaction region, and a reaction process in this region is divided into three steps. Several physical quantities have been defined to describe each of these steps, i.e., the diffusion flux of moving reactant to the reaction region, J; the release rate of nonmoving reactant, r, and the formation rate of the growing phase, F. The relationship between these quantities is demonstrated by means of a reaction process plot, which is also developed in this study. A semiquantitative calculation of the reaction process plot has been done for Mn-Si and Ti-Si systems. The results show that the plots can be successfully used to predict suicide formation sequences.

n+-n and n+-p junctions were formed on n-type and p-type Si using a thin film of phosphorus obtained from a simple vapor source, and driving in the dopant atoms in a rapid thermal processing (RTP) system. The vapor treatment consisted of heating powdered red phosphorus in a nitrogen ambient and allowing the resulting phosphorus vapor to deposit on the Si samples. This was done in an inexpensive apparatus constructed from flasks and test tubes. Following the vapor treatment, an SiOxfilm was sputtered over the phosphorus coating in order to serve as a capping layer during subsequent RTP drive-in that forms the junction. The junction properties were characterized by spreading resistance and electrical (IV and CV) measurements after deposition of metal contacts layers.

Copper ball bonding is a new technology which is expected to replace the traditional gold ball bonding and paid attention in woled recently, the technology is very important to reduce cost and improve reliability of microelectronic components. In this paper, the copper wire ball bonding processes have been studied by means of the MW-EFO metal wire ball forming device and the JWYH-2 thermosonic ball bonder. The bond strength of the ball bond under varied ultrasonic power and bonding time have been tested. The tested results show that the maximal strength of copper ball bond can be over 20g, it has advantage over the gold wire ball bonding. In this paper, the deformation process of the copper ball bond has been analyzed and viewed by SME.

In the development of air bridges for low capacitance crossovers, resists of 3–6μm thickness a e needed, which calls for special alignment techniques. A technique has been developed for an air bridge tec nology, which is realized entirely by electron beam lithography. Such an all-electron-beam technique for sub-micron devices and ultrahigh frequency circuits, results in a fast turn-around-time in the development of integrated circuits.

Gold plated air bridges with 20 to 120 pm span, 5–60 μm width, 2.6 pm height, and 2–4 μm thickness, were realized on GaAs and InP.

The application of ion beam mixing and ion implantation techniques in the formation of ohmic contacts to n-Hg1-xCdxTe (x = 0.6, 0.7) has been investigated. For experiments in ion beam mixing, an indium layer (15nm thick) on Hg1-xCdxTe wasbombarded with 45keV Te+ ions at doses ranging from 1 × 1013 to 1 × 1016 cm2. Minimum values of specific contact resistance, pc, of 4 × 10-4Ωcm-2 were measured for a dose of 1 × 1016 cm-2 after 200°C annealing; pc was independent of dose in this range. Analysis of the interfaces by Auger depth profile and Rutherford backscattering spectrometry has shown that ion beam mixing produced an enhanced indiffusion of indium. In comparison, samples which were ion implanted with In+ at 50keV prior to metalisation showed a reduction in pc which was strongly dependent on dosage.

In-line thickness monitoring of CoSi2 films is usually carried out by four-point probe resistivity measurements on blanket samples. This method, however, reveals no information about the thickness variation and stoichiometry of the CoSi2 films. Transmission Electron Microscopy (TEM) can provide this information but the technique is destructive, whereas Raman spectroscopy is non-invasive and offers a technique for patterned wafers compatible with current VLSI technology.

The CoSi and Co2Si phases exhibit very characteristic Raman signatures which, with glancing angle X-ray diffraction (GA-XRD), can be used to investigate the silicidation sequence. Although CoSi2 does not produce a Raman spectrum, the silicon local mode can be easily detected through the film and is very sensitive to variations in thickness. The intensity of the silicon local mode was measured on a series of ten samples of CoSi, on silicon and the precise CoSi, film thickness subsequently determined by cross sectional TEM. From these measurements, the absorption coefficient of CoSi2 at 500nm was determined.

The effect of the defective nature of the substrate on the alloying behaviour of Xe implanted Au(55 ran)/n-GaAs system was studied using cross sectional transmission electron microscopy.

Low dose Xe implantation (700 keV, 1*1014 ions/cm2) caused the formation of about SO nm thick polycrystalline region of GaAs beneath the gold layer. Annealing the implanted sample at 450°C gold diffused through the polycrystalline GaAs region and formed large pits of Au(Ga) solid solution in the defective area of GaAs having stacking faults and twins. The formation of a regrown GaAs covering layer was observed on the top of the reacted metallization simultaneously.

High dose implantation of Xe++ ions resulted in the formation of amorphous GaAs layer with a thickness of about 750 nm. Twinned regions of GaAs were observed at the amorphous - crystalline GaAs interface by high resolution electron microscopy. Ion beam caused phase transition was observed in this sample. The amorphous GaAs region recrystallized to single crystalline GaAs due to annealing at 400°C.

We studied Mo and Ni contacts to p-type CuInSe2(112) with angular-resolved synchrotron radiation ultraviolet photoemission spectroscopy in the range 40 eV hv < 120 eV. By varying the photon energy, we determined that emission from the CuInSe2 Г-point in the Brillouin zone occurs for hv = 60 eV, and identified emission from the valence band maximum. After depositing Ni and Mo, we found that the conduction band minimum of the CuInSe2 very nearly aligned to the Fermi edges of both metals, giving Schottky barriers Vsb = 1.05 ±0.1 eV. The interfaces are not atomically abrupt, as seen by the outdiffusion of Se into the refractory metal overlayers and the formation of In-Mo and In-Ni alloys.