Kimiyuki Nishiguchi

A Numerical Analysis of Void Shrinkage Processes Controlled by Coupled Surface and Interface Diffusion

The shrinkage behaviour of voids arrayed on a bond-interface (grain boundary) is analyzed by computer simulations. It is assumed that surface and interface self-diffusion operate in series. The junction between the void-surface and bond-interface is treated as a tip with a dihedral angle θ. Two numerical models are proposed to discuss the effect of interface and surface tensions on the void shrinkage. One of them always restricts θ to the value of the equilibrium balance (model 1). In the other (model 2), the equilibrium balance is ignored, i.e. the change in θ (free dihedral angle) is considered during the void shrinkage. For discussing the rate controlling step, model 2 is essentially the same as model 1. But, when the void shrinkage is controlled by the interface diffusion, they give void shapes quite different from each other, i.e. when the diffusivity ratio ƒ ⪢ 1 (ƒ = 2δsDs/δbDb, where δsDs and δbDb are produced by the thickness of the diffusion layer and the self-diffusion coefficient for the surface and interface, respectively), model 1 gives the equilibrium void shape but model 2 only exhibits the stationary void shape. The void morphology is not only influenced by the prescription of the equilibrium balance but also by the diffusivity ratio ƒ. The stress and temperature dependence of the void shrinkage can roughly be expressed by 2δsDsδbDb/[2δsDs + δbDb)Pn], where P is the bonding pressure and n the stress exponent. If θ is prescribed by the equilibrium balance, n increases from −1 to −0.3 with decreasing P. But, if θ changes, n is about −1. It is verified by experiments that the dihedral angle is not always constant and changes as increasing net stress is applied to the bond-interface. It is suggested that model 2 more exactly predicts the void-shrinkage process than model 1 as the pressure is increased.

Identification of void shrinkage mechanisms

A method for identifying void shrinkage mechanism experimentally is presented. We treat the void shrinkage on the bonded interface during solid state diffusion bonding of similar metals. The voids are arranged at regular intervals. The experimental results were analysed by log(T/ts) vs (1/T) and log ts vs log P plots, where T is the absolute temperature, P the compressive stress, ts is the time taken to attain the void shrinkage of a certain volume ΔV. We obtained ts by measuring a certain growth of bonded area ΔS replaced by ΔV. Even if ΔS is adopted, we can identify the transition of the dominant mechanism by slopes of those plots. As a result, interface self-diffusion, volume self-diffusion and power law creep were experimentally identified as fundamental mechanisms which contribute to the void shrinkage process during diffusion bonding of metals.

Valence Auger analysis of the annealing effect on atomic interaction at titanium-sapphire, titanium-silica and silver-silica interfaces

Abstract The atomic interaction (chemical bonding) across the interface between deposited metal films and ceramic substrates is analysed using Auger electron spectroscopy. For this study the titanium-sapphire, titanium-silica and silver-silica systems were chosen. The annealed interfaces are compared with the as-deposited interfaces. It is suggested that AlTi bonds exist at the annealed titanium-sapphire interface but not at the as-deposited one. The results for the titanium-silica system suggest that SiTi bonds appear at both the as-deposited and annealed interfaces. No SiAg bonds are detected at either the as-deposited or the annealed silver-silica interface. Our results for the silver-silica and titanium-silica systems are in good agreement with thermochemical data. It is suggested that an interfacial excess energy induced by Fermi electrons may influence the atomic interaction at the titanium-sapphire interface.

Dissolution process of surface oxide film during diffusion bonding of metals

Two models of the dissolution process of surface oxide film in metals (copper and titanium) are described in order to estimate the dissolution time ts required for the oxide film to dissolve in the metals completely. One is based on impurity diffusion (model 1). The other is given by reaction diffusion (model 2). Model 1 was applied to estimate ts in copper and α-titanium but model 2 for ts of β-titanium because of the formation of α-phase. In these models, the dissolution process is assumed to be controlled by the oxygen diffusion in metals. Bonding tests were performed in order to verify the calculated results. The experimental results suggest that model 1 is valid in Cu-Cu bonding. Also, as for β-titanium, no retardation of the bonding process was observed. This was in agreement with model 2. However, for α-titanium below 1000 K, the retardation time was much longer than the dissolution time calculated by model 1, i.e. the retardation below 1000 K cannot be dictated by the diffusion-controlled process. A new model is therefore proposed. The dissociation rate of the oxide film is taken into account in the new model. This new model can explain the retardation of the bonding process.

A valence Auger analysis across the annealed interface between a deposited titanium film and sapphire

Abstract A titanium film was deposited on a sapphire substrate by electron beam deposition and a valence Auger analysis was performed on this after annealing. We mainly analysed the annealed interface between titanium and sapphire. The effect of argon ion beam and electron beam bombardment on sapphire is examined. Several characteristics of spectral shapes concerning TiO and AlO bonds were obtained from the surface-oxidized titanium film and Al 2 O 3 bulk respectively. Although the Ti-sapphire interface was affected by argon ion and electron beam bombardment, the interfacial spectra have characteristics of TiO and AlO bonds. However, it was difficult to find any evidence for the existence of TiAl bonds from the characteristics of the spectral features only.

Interfacial reaction of Ag/Ti thin films on a silica substrate

Abstract Double and triple layer films of AgTi system were deposited on silica (fused quartz) substrates and annealed at 800 K for 1 h under vacuum (10−4Pa). Atomic interaction between the film and the substrate due to the annealing was observed using Auger electron spectroscopy. The annealing segregates Ti atoms both at the film surface and the interface between metallic film and silica substrate. Directly deposited Ti atoms on silica decomposed SiO2 bonds without annealing but Ti atoms segregating at the interface did not. The behaviour of titanium can be explained in terms of the affinity for oxygen. This was suggested by the change in the peak height ratio of Ti L 23 M 23 M 45 Ti L 23 M 23 M 23 related to the valence Auger transition.

A valence Auger analysis across the interfaces between nitride ceramics and deposited metals

Abstract Specimens of Cr/AIN, TI/AIN, Ag/AIN, Cr/Si 3 N 4 and Ti/Si 3 N 4 interfaces were prepared by electron beam deposition. Auger spectra of as-deposited and annealed samples were obtained at their interfaces using JEOL JAMP10s. The overlapping spectra were separated by a subtraction technique. The subtraction was available for these systems. A change in Auger line shapes across the interface was observed and the interfaces were classified into two types: type A, the defect peak in the LVV Auger transition is constant across the interfaces and type B, the peak increases as the metal peak increases at the interface. The interfaces of Cr/AIN, Ag/AIN and Cr/Si 3 N 4 belong to type A. The interfaces of Ti/AIN and Ti/Si 3 N 4 belong to type B. AES results suggest that the substrate is reduced by the deposited metal at the Ti/AIN and Ti/Si 3 N 4 interfaces. This is supported by thermochemical data. It is suggested that the valence Auger analysis is applicable to the metal/ceramic systems as an evaluation method if the electronic dose is kept constant across the interface.

Arc Phenomena at Low Pressure Atmosphere (Report 1)

Systematic research has been made on arc phenomena at low pressure atmosphere to clarify the electrode mechanism. Distinguished features of low pressure arc were found in cathode and anode phenomena, which appear eminently in the arc characteristics and appearance. For example, pure graphite cathode, which used to behave as a typical thermionic cathode (hot cathode), exhibits so called cold cathode under the condition of low pressure and small current. This cold cathode is characterized by random movement of a constricted cathode spot and strong vapour jet from the spot area.This paper discusses the result of the systematic analysis on arc characteristics and their relation to transition of the cathode mode. The experiment was done under the condition of 10-150 Torr of pure Hydrogen atmosphere and 10-120 A of arc current. The electrode material was pure graphite rods (99.99%) of 6 mmφ.Conclusions obtained are as follows:(1) Transition of the cathode mode from the hot cathode to the cold cathode is mainly governed by arc current I, ambient gas pressure p and ionization p4tential Vi of the gas in the cathode fall region. Arc voltage suddenly falls by a few volts when the hot cathode changes to the cold one, though the average electric field of arc column Ep increases abruptly because of the shrinkage of column. This fact suggests that, in the cold cathode mode, Carbon rather than Hydrogen is ionized in the cathode fall region, and consequently the cathode drop Vσ reduces.(2) Sum of the anode and cathode drops, (VA+VC), increases gradually in the range of the hot cathode with the decrease of ambient gas pressure and arc current. This value of (VA+VC) suddenly falls by 3-4 volts under the critical condition of ambient gas pressure and arc current when the cathode mode changes to the cold cathode. This abrupt change of (VA+VC) is due to the decrease of the cathode drop VC.(3) In the range of the cold cathode, (VA +VC) decreases with the reduction of ambient gas pressure, while (VA+VC) remains almost constant against arc current.(4) Variation of the cathode drop VC in both cathode. modes is explained qualitatively by Slepians ionization mechanism in the cathode fall region.

A Consideration on the Mechanism of the Droplet from the Wire Tip in MIG Welding

The paper explains the reason why the wire tip takes a pencil-like form for the wire of low heat conductivity from the standpoint of the anode heating on the cylindrical wire surface. It explains also the reason why the pencil like-forming tendency is increased when the wire is preheated by .Joules loss.It is explained that the pencil-like forming of the solid part at the wire tip allows an easy detachment of the molten droplet and results in a lower temperature of the droplet arriving at the base metal.From the above mentioned idea, the following characteristics observed in MIG are welding are explained. For aluminum wire the heat content of the droplet increases as the current is increased until it reaches a nearly constant value in spray transfer range. The temperature is as high as the boiling point in spray transfer. The increase of the temperature is due to the quick heating of the droplet. For steel wire the heat content increases and reaches a maximum and then decreases as the current is increased under a constant wire extension The decrease of the heat content is related to the pencil forming.For any given current the heat content is smaller for longer wire extension because the pencil forming tendency is stronger. The decrease of the current corresponding to the maximum heat content above mentioned for longer wire extension is also attributed to the same cause.Lesnewich reports that the critical current in globular to spray transition decreases as the wire extension is increased. The reason is explained by the pencil forming. Salter reports that the melting speed is increased with decreasing of atmospheric pressure. The potential gradient of the arc column decreases in lower pressure and the arc foot climbs up along the wire. Therefore the pencil forming is increased and results in a decrease of the droplet temperature and an increase of the melting speed, because the equivalent anode melting voltage VmA=VA+VW+VT is presumed to be independent of the pressure which is not so much different from the atmospheric pressure.