Simon C. Hogg

Wire Bond Reliability for Power Electronic Modules - Effect of Bonding Temperature

In this paper, thermal cycling reliability along with ANSYS analysis of the residual stress generated in heavy-gauge Al bond wires at different bonding temperatures is reported. 99.999% pure Al wires of 375 mum in diameter, were ultrasonically bonded to silicon dies coated with a 5mum thick Al metallisation at 25degC (room temperature), 100degC and 200degC, respectively (with the same bonding parameters). The wire bonded samples were then subjected to thermal cycling in air from -60degC to +150degC. The degradation rate of the wire bonds was assessed by means of bond shear test and via microstructural characterisation. Prior to thermal cycling, the shear strength of all of the wire bonds was approximately equal to the shear strength of pure aluminum and independent of bonding temperature. During thermal cycling, however, the shear strength of room temperature bonded samples was observed to decrease more rapidly (as compared to bonds formed at 100degC and 200degC) as a result of a high crack propagation rate across the bonding area. In addition, modification of the grain structure at the bonding interface was also observed with bonding temperature, leading to changes in the mechanical properties of the wire. The heat and pressure induced by the high temperature bonding is believed to promote grain recovery and recrystallisation, softening the wires through removal of the dislocations and plastic strain energy. Coarse grains formed at the bonding interface after bonding at elevated temperatures may also contribute to greater resistance for crack propagation, thus lowering the wire bond degradation rate.

Unusual Observations in the Wear-Out of High-Purity Aluminum Wire Bonds Under Extended Range Passive Thermal Cycling

This paper reports on the reliability of ultrasonically wedge-bonded 99.99% (4N) and 99.999% (5N) pure aluminum wires under different passive thermal cycling ranges, namely, -40°C to 190°C, -60°C to 170°C, -35°C to 145 °C, and -55°C to 125°C. The rate of bond strength degradation during cycling was found to be more rapid in the wire bonds subjected to lower peak temperatures (Tjmax) and lower temperature ranges (ΔT) for both wire types. This observed effect of ΔT cannot be described by the commonly accepted empirical relationships based on damage accumulation, such as the Coffin-Manson law. In addition, the 4N wire bonds were found to degrade more rapidly than the 5N bonds under the cycling ranges investigated. Microstructural characterization and nanoindentation of the bond interfaces indicated differences in microstructural restoration in wires subjected to the different cycling ranges. These differences have been attributed to annealing phenomena occurring in the wires during the high-temperature phase of cycling, which are believed to remove some of the damage accumulated during the low-temperature phase. A model is proposed for the prediction of wire bond wear-out rate, which incorporates both damage accumulation and damage removal mechanisms. We conclude that the rate of annealing during cycling varies exponentially with temperature; the annealing effects which occur can reduce damage accumulation and therefore influence wire bond reliability.

Characterisation of microstructure and creep predictions of alloy IN740 for ultrasupercritical power plants

Abstract In the context of ultrasupercritical power plants, Ni base alloys are prime candidate materials for long term, high temperature applications, such as boilers, operating at temperatures and pressures as high as 750°C and 35 MPa. This necessitates the investigation of their microstructural evolution as a function of thermal treatment and simulated service conditions out to longer times at forecasted service temperatures, coupled with modelling activities able to predict the microstructural evolution under these new conditions. The lack of widespread microstructural data for most commercial nickel base alloys in these time–temperature regimes makes this type of investigation even more important. In this study, the microstructural evolution of IN740, a Ni–Cr–Co–Mo–Nb–Ti–Al superalloy, is characterised as a function of aging conditions. A method for phase identification is described that can be confidently used to gather relevant information for modelling activity, such as: phase identity, volume fraction, size distribution and interparticle spacing. The microstructural evolution of IN740 is investigated at temperatures of 700 and 750°C for aging times up to 10 000 h. The data obtained experimentally from the aged specimens as well as from literature are used both as input for and validation of a microstructurally based continuum damage mechanics (CDM) model for forecasting creep properties. Using this predictive model, discussion is made of possible approaches needed to optimize creep performance.

Spray Forming of Al-Fe-Cr-Ti and Al-Si-Li Alloys

This work presents an investigation of the spray forming and downstream processing of Al alloys that are difficult to produce in bulk by conventional solidification processing: Al-Fe-Cr-Ti alloys for intermediate temperature applications and Al-Si-Li alloys for high stiffness, low density applications in fast moving machinery. For the Al-Fe-Cr-Ti alloys, spray forming is being investigated to allow the scale-up of alloy compositions previously explored only as ribbons or powders in traditional rapid solidification routes. For Al-Si-Li alloys, spray forming is used to provide globular primary AlLiSi in a fully divorced AlLiSi/α-Al eutectic structure. For both alloys, the as spray formed and downstream processed microstructure of 20kg billets has been investigated by scanning electron microscopy, electron probe microanalysis, and X-ray diffractometry. Preliminary mechanical properties have also been investigated.

Analysis of ferrite formed in 321 grade austenitic stainless steel

Abstract A significant fraction of ferrite has been identified in a 321 grade austenitic stainless steel in the solution heat treated condition. The microstructures were analysed using electron backscatter diffraction, energy dispersive X-ray spectroscopy and X-ray diffraction (XRD) and the stability of the ferrite investigated using heat treatments in a tube furnace, dilatometry and high temperature XRD. The ferrite dissolved ∼800°C, then formed again on cooling at temperatures under 200°C. Thermodynamic predictions showed a significant ferrite content at room temperature under equilibrium conditions, and the DeLong diagrams predict an austenite+martensite microstructure in the cast condition. Sensitivity analysis on the DeLong diagram has shown that the nitrogen content had a large effect on the austenite stability. The instability of the austenite and the subsequent transformation to ferrite on cooling can be attributed to low nitrogen content measured in the as received material. It was found that thermal aging of the material caused further transformation of austenite to ferrite as well as the formation of sigma phase that appears higher in nitrogen than the matrix phases. The diffusion of nitrogen into sigma phase may cause instability of the austenite, which could cause further transformation of austenite to ferrite on cooling from the aging temperature. The transformation of austenite to ferrite is known to be accompanied by an increase in volume, which may be of relevance to components made with tight dimensional tolerances.

Evolution of sigma phase in 321 grade austenitic stainless steel parent and weld metal with duplex microstructure

Abstract Samples of 321 stainless steel from both the parent and welded section of a thin section tube were subjected to accelerated ageing to simulate long term service conditions in an advanced gas cooled reactor (AGR) power plant. The initial condition of the parent metal showed a duplex microstructure with approximately 50% ferrite and 50% austenite. The weld metal showed three distinct matrix phases, austenite, delta ferrite and ferrite. This result was surprising as the initial condition of the parent metal was expected to be fully austenitic and austenite+delta ferrite in the weldment. The intermetallic sigma phase formed during the accelerated ageing was imaged using ion beam induced secondary electrons then measured using computer software which gave the particle size as a function of aging time. The measurements were used to plot particle size, area coverage against aging time and minimum particle spacing for the parent metal. During aging the amount of ferrite in the parent metal actually increased from ∼50 to ∼80% after aging for 15 000 h at 750°C. Sigma has been observed to form on the austenite/ferrite boundaries as they may provide new nucleation sites for sigma phase precipitation. This has resulted in small sigma phase particles forming on the austenite/ferrite boundaries in the parent metal as the ferrite transforms from the austenite.

Thixoforming of Stellite Powder Compacts

Thixoforming involves processing metallic alloys in the semi‐solid state. The process requires the microstructure to be spheroidal when part‐solid and part‐liquid i.e. to consist of solid spheroids surrounded by liquid. The aim of this work was to investigate whether powder compacts can be used as feedstock for thixoforming and whether the consolidating pressure in the thixoformer can be used to remove porosity from the compact. The powder compacts were made from stellite 6 and stellite 21 alloys, cobalt‐based alloys widely used for e.g. manufacturing prostheses. Isothermal heat treatments of small samples in the consolidated state showed the optimum thixoforming temperature to be in the range 1340°C–1350°C for both materials. The alloys were thixoformed into graphite dies and flowed easily to fill the die. Porosity in the thixoformed components was lower than in the starting material. Hardness values at various positions along the radius of the thixoformed demonstrator component were above the specificat...

Corrosion of an Advanced Al-Cu-Li Alloy for Aerospace Applications

This paper discusses two Al-Cu alloys for aerospace applications, one which has a high concentration (>1.8 wt.%) of Li. These alloys are AA2024-T3 (Al-Cu) and AA2099-T8E77 (Al-Cu-Li) and are both in the plate format. Anodic polarisation and immersion in a 3.5 wt.% NaCl solution have been carried out and a comparison of the corrosion mechanisms have been made. Both alloys showed extensive corrosion over the surface, AA2024-T3, however, had areas of pitting that had joined together forming large corrosion regions, whereas AA2099-T8E77 had numerous corrosion pits that had not merged. The pits on AA2099-T8E77 propagated to a depth ~120 µm whereas on AA2024-T3 the maximum depth was ~85 µm. Intergranular and transgranular corrosion was also observed on the AA2024-T3 alloy following anodic polarisation. AA2099-T8E77 had reduced corrosion resistance with regards to Epit-Ecorr values. For both alloys, dissolution of matrix material surrounding the Cu-rich intermetallic particles was observed, highlighting the particles’ cathodic nature.

High-temperature microstructural evolution and quantification for alloys IN740 and IN740H: comparative study

In ultra-supercritical power plants, Ni-base alloys are candidate materials for long-term, high-temperature applications, operating at temperatures and pressures as high as 750°C and 35 MPa. Alloy IN740 and its modification, alloy IN740H, are considered for such applications. Their microstructural evolution, at 750°C for times ranging between 3000 and 5000 hours, has been investigated by means of scanning electron microscopy, electron back-scattered diffraction, energy dispersive X-ray analysis and phase quantification. All phases were identified and quantified allowing comparison between the two microstructures, their evolution and stability. Particular attention was paid to γ′, η and G phases. The results are used within a broader investigation aimed at improving and further developing a predictive creep model based on continuous damage mechanics.

Modelling of creep and fracture properties of nickel based alloys

This paper reviews the differences between two nickel based alloys, Alloy 740 and Alloy 740H. Microstructural evolution models are used to forecast the changes in volume fraction and interparticle spacing of both grain boundary and intra-granular precipitates in the alloys. These data are then employed in continuum damage mechanics models to forecast creep curves, and in fracture mechanics models to forecast Charpy impact energies/energy. The results reveal the key microstructural features that control secondary and tertiary creep rate as well as the time dependence of Charpy impact energy after high temperature exposure.