In order to discuss stress-induced voiding (SIV) of Cu damascene interconnect structures in ultra-large-scale integrated circuits (ULSls), 3-D elastic-plastic finite element analysis (FEA) was carried out based on stress-temperature behavior of constituent thin films measured by a wafer curvature method. Two types of Cu interconnect geometries with either D-SiN cap/p-SiON etch-stop layers or D-SiCN : H cap/p-SiC : H etch-stop layers were analyzed. The effect of the Cu line width was also investigated. It was found from the FEA results that, regardless of the geometry, the hydrostatic tensile stress of Cu in the structure with D-SiCN : H/p-SiC : H layers was generally lower than that with p-SiN/p-SiON layers. It was also expected that, for the structure with p-SiN/p-SiON layers, SIV is most likely to occur near the via center while for the structure with p-SiCN : H/p-SiC : H layers, it is most likely to occur near the via bottom. In the structure with D-SiCN : H/p-SiC : H layers, it was considered that a larger line width is susceptible to voiding in the via due to a high hydrostatic stress gradient in the via and a high magnitude of equivalent plastic strain in the line.

View full abstract

A method of evaluation of six components of grain strain in polycrystalline metal was proposed. In the present method, the electron backscattered diffraction (EBSD) technique and the digital image processing were efficiently combined. Four components of grain strain, ε_x, ε_y, ε_z and γ_xy were evaluated based on the change in average aspect ratio of the number of pixels inside the grain which was obtained by an optical microscope and digital image analysis, where xy-plane is set on the specimen surface and z-axis is directed perpendicularly to the surface. The other two components, γ_yz, γ_zx, were calculated by taking into consideration the restriction of plastic deformation of each grain which was derived on the assumption that the plastic deformation occurred by crystallographic slipping on a single set of slip plane. An angle of the slip planes was 3-dimensionally identified by observation of the slip line by a scanning electron microscope and EBSD technique. Grain strain of polycrystalline copper during tension was evaluated by the proposed method and the results were discussed from the viewpoints of the distribution of grain strains and the deformation restriction by neighboring grains.

View full abstract

To select the best solder alloy for the electronic packaging, a constitutive model which can be applied to alternative solder alloys is needed for the reliability evaluation such as estimations of fatigue failure using FEM analysis. In this paper, to apply the plasticity-creep separate method using the constitutive model proposed by authors to the three solder alloys such as Sn-3.5 Ag-0.75 Cu, Sn-7.5 Zn-3.0 Bi and Sn-57 Bi-1.0 Ag, the basic experiments and numerical simulations are conducted. Pure tensile tests are conducted to determine material constants of the constitutive model using the plasticity-creep separate method. The material constants determined from the only pure tensile data are used for the simulation of creep and cyclic tension-compression loading. As a result, the plasticity-creep separate method can be applied to lead-free solder alloys as well as lead solder alloys, and the simulations of the both creep and cyclic loading are successfully conducted using the material constants determined from the pure tension.

View full abstract

Recently one of the authors has proposed a method for the analysis of stress intensity factor which uses the crack tip stress calculated by FEM. It is called the crack tip stress method. In this paper, the crack tip stress method is applied to the K_I analysis of an infinite body subjected to tension, which has an axisymmetrical ellipsoidal cavity or a disk-type cavity with a circumferential crack. The values of K_I are discussed based on the effective crack size (a+λ), where a is the radius of ellipsoid or disk and λ is the size of crack emanating from the notch root. Furthermore, based on the linear notch mechanics it is shown that the relation between F_I = K_I/(1.121 5σ_max √(π)λ) and λ/ρ is almost independent of ρ/a, where ρ is the notch root radius and σ_max is the maximum stress at the notch root when the crack dose not exist.

View full abstract

A new method of nondestructive estimation of the in-depth distribution of residual stresses is proposed by using monochromatic high-energy X-rays from synchrotron radiation source SPring-8. The X-ray with 72 keV energy was used for stress measurements of shot-peened steel samples, whose compressive zone extended a few hundred micrometers below the surface. The measurements were carried out by using the sin² φ method in a Ψ-diffractometer. The obtained results of sin² φ diagrams showed nonlinearity due to the in-depth distribution of the residual stress. The stress distribution below the surface was estimated from the analysis to the nonlinearity of sin²φ dragrams. The estimated stress distribution agreed well with that determined by the conventional sin² φ method with Cr-Kα radiation corrected by surface removal effect.

View full abstract

Fatigue tests of crack propagation from a circular notch in thin-walled tubular specimens made of a medium-carbon steel were performed under cyclic torsion with and without superposed static and cyclic axial loading. The propagation path of Mode I fatigue cracks followed the direction perpendicular to the maximum direction of the total range of the normal stress. The propagation rate was faster than the uniaxial data when compared at the same range of the stress intensity factor or the effective stress intensity factor. The negative nonsingular stress induced excessive plasticity ahead of the fatigue crack, which accelerated the fatigue crack. The J-integral range was proved to be an appropriate parameter for crack propagation with excessive plasticity under combined loading.

View full abstract

This paper describes the fatigue behaviour of a high V-Cr-Ni cast iron. Rotating bending fatigue tests have been performed using smooth specimens in laboratory air at ambient temperature. The obtained results were compared with the fatigue behaviour of three types of nodular cast irons, FCD450, FCD700 and ADI, and fracture mechanisms were discussed on the basis of crack initiation behaviour and fracture surface analysis. The high V-Cr-Ni cast iron showed superior fatigue strength to the nodular cast irons and the fatigue strength characterized in terms of fatigue ratio was also significantly higher than that of the nodular cast irons. Fatigue cracks initiated at inclusion, casting defect and cluster of VC carbides, of which cluster of VC carbides was the predominant crack initiation site. In small crack region following the crack initiation, cracks grew avoiding VC carbides, indicating that VC carbides acted as a barrier to the growth of small cracks.

View full abstract

For studying the effect of crystallographic orientation and stress concentration on fatigue properties at 773 K, high cycle fatigue tests were conducted on conventional casting superalloys MM 247, directionally solidified superalloy CM 247 LC and single crystal superalloy CMSX-2. Following conclusions were obtained. 1) Fatigue cracks were initiated at material defects in smooth specimen, and the ratios of fatigue limit to tensile strength were lower than 0.3. 2) Loading conditions made the difference in fatigue limits of smooth specimen, that is, the predicted values from notched fatigue > rotating bending fatigue > axial loading fatigue, which is considered to be caused by the difference of the extent that material defects could take part in fatigue crack origins. 3) Notched fatigue strength of CMSX-2 can be unifiedly expressed by resolved shear stress in fatigue limit diagram, which is independent of crystallographic orientation.

View full abstract

In this paper, a dynamic constitutive equation that is applicable for impact loading and inverse loading was investigated. The numerical calculation was done for aluminum using finite difference method under the boundary condition of measured axial and angular velocities. The calculated stress and strain agreed well with the experimental ones under torsional and inverse torsional loadings about the strain rate of 40 l/s. The comparison of calculated result with experimental one for the combined loadings of impact tension and torsion was also performed, and the validity of this constitutive equation was verified. Moreover, the appearing yield points were discussed based on some simulations under two-stage combined loadings of tension and torsion.

View full abstract

The present study reports the experimental research of particle velocity increasing process generated by elastic-plastic shock waves in 6061-T6 aluminum. Planar impact experiments are carried out by a single-stage light-gas gun. The gun has 2 m long launch tube and 40 mm muzzle diameter. Projectiles are made of ABS material, and accelerated by high-pressure gas (He). Impact velocities measured by the magnet flyer method are from 227 to 413 m/sec. A target plate is hold in a target ring with 4 tilt-pins. Tilt angles obtained by any shots are from 2.4 to 3.6 mrad. The velocity changes at the rear surface of target plate are measured by the velocity interferometer system (VISAR). High-resolution data focused in the plastic shock wave are obtained, and it is shown that those data can be compared to the results of estimating simulation of the shock wave profiles.

View full abstract

In this study, the relationships between porosity (P) and elastic modulus of sintered copper-nickel alloys made from mixed powder and mechanically alloyed powders were investigated by an acoustic pulse method. The elastic modulus of sintered alloy made from mixed powder differed from that made from mechanically alloyed powder below 50 mass% copper content. Especially, the difference of Poisson's ratio was clearly observed in the range above 0.05 of porosity. Irrespective of the copper contents or the powders mixed or mechanically alloyed, Young's modulus (E) and shear modulus (G) were described as E = (E₀-K_E·P)(1-P) and G = (G₀-K_G·P)(1-P), respectively, when E₀ and G₀ are E and G at P = 0, K_E and K_G are experimental constants for E and G, respectively. The E₀, G₀, K_E and K_G were cubic functionally decreased with increasing of the copper content. By using above relation, it was clarified that elastic modulus of the sintered copper-nickel alloy can be calculated from two parameters, copper content and porosity.

View full abstract

In order to make a new movement of the composite material design, the purpose of the present study is applying the solid-air composites derived from the load dispersion structure of living bones to investigate the static load dispersion function shown by the mechanical behavior of those constitutive elements. Internal hydrostatic pressures were controlled by making a hole in cell walls of the honeycomb core. Static indentation tests were carried out to examine the load dispersion effect. In-plane cell's deformations and stress distributions at the bottom surface of the specimens were measured optically. Experimental results illustrated that the distribution of the internal hydrostatic pressures greatly influenced on the deformation condition of the solid part. Consequently, it is considered that controlling the internal hydrostatic pressures is an effective design parameter of the load dispersion function.

View full abstract

The effects of various basic factors of combustion gas flow conditions on the recession behaviors of silicon carbide and alumina have been experimentally clarified. The exposure tests were performed under various gas flow conditions (T=1100∼1500°C, P=0.3∼0.7 MPa, V=45∼250 m/s, P_H2O=30∼120 kPa and P_O2=22∼43 kPa). The recession rate of silicon carbide mainly depended on water vapor partial presssure, temperature, and velocity in combustion gas flow. The pressure dependence was not found and the oxygen dependence was extremely low. Alumina also showed linear weight loss and the surface recession, and water vapor partial pressure, temperature, and velocity dependences were observed. It was considered that the recession of Al₂O₃ occurs in order to volatilize by forming gaseouse hydroxide species, as same as SiC. The recession rates of silicon carbide and alumina in combustion gas flow were experimentally expressed in the form exp(- E/RT)P_H2O^xV^y.

View full abstract