Howard M. Berg

Properties of Die Bond Alloys Relating to Thermal Fatigue

The differing performances of hard soldered and soft soldered power devices under cycling stresses are well understood. Hard soldered dice are often under considerable stress and may even fracture in extreme cases, but the solder itself does not degrade by fatigue after many temperature cycles. This behavior is due to the very high mechanical strengths of the hard solders between -55 and 150°C. On the other hand, soft solders transmit very little stress to the die but exhibit considerable degradation from thermal fatigue during temperature or power cycling. Both behaviors are related to the very low mechanical strengths and high ductilities of the soft solders. In this study the mechanical properties of the eutectic hard solders, Au-Sn, Au-Ge, and Au-Si, and two common soft solders, Pb-5%Sn and Sn-3.5%Ag-l.5%Sb, have been measured between -55 and 200°C. Several physical properties were also measured. An ahoy was then developed with mechanical properties intermediate to those of hard and soft solders and with a power cycling performance that approaches that of the Au eutectics. A model is presented which relates the fatigue behavior of hard, soft, and intermediate solders to their mechanical properties.

Enhancing Ultrasonic Bond Development

To enhance ultrasonic bond development, an improved understanding is needed of the processes taking place at the bond interface. This work is designed to determine the role that bonding parameters play in contributing to bond development. Of particular concern are the effects created by the machine variables, i.e., load, power, and time. Silicon was selected as a bond surface for aluminum wire bonds. The brittle nature of silicon provides a permanent record of the bonding history. With both the crystal orientation and defects of the silicon well characterized prior to bonding, features such as the location of residual bonding strains in the silicon were determined. The pattern of partially bonded material exposed by peeling underdeveloped bonds simulates a torus (or doughnut) with an unbonded central region. Features of aluminum wire bonds to aluminum, glass, beryllium, and silicon were compared to show that a common mechanism exists independent of the bond surface material. The ability to bond to silicon varies with wire composition. For example, both Al-0.5% Mg and pure Al wires bond readily, while Al-l% Si wire does not. Two modes of material flow characterize interfacial behavior. Ultrasonic energy promotes a material softening which, in conjunction with the applied load, results in a gross flow to expose fresh material for bonding. In the second stage of material flow, a wave form is propagated through the wire to the periphery of the wire-silicon interface. This wave form is observed as a periodic cutting action into the silicon perpendicular to the pulsing direction. A fine ball-like formation in the grooves of the wave region at the bond zone was a feature common to the different bonding surfaces. The wavelength and wave amplitude vary linearly with the applied power as does the tip-to-tip displacement of the wedge. The groove spacings are of the same magnitude as the wedge displacement. For constant power and time, increased load increases the size of the central bond region that does not experience the wave action. For constant load and power, the width of the wave affected periphery increases toward the center of the bond with time. The method of thermally induced stacking faults by steam oxidation was used to characterize the residual bond strains in the silicon. Faulting was found in the peripheral region of the bond where the wave action was operable. This faulting correlated to stresses generated in the pulsing direction and not to the directions of gross material flow. Reliable bonding depends upon a proper control of the gross and wave flow processes by optimizing material properties as well as machine variables. A model has been developed to qualitatively relate the influence of these variables and the manner in which a change in one parameter affects the response of the remaining variables.

Thermal Characteristics of 16- and 40-Pin Plastic DIP's

The thermal performance of 16- and 40-pin plastic dual in-line packages (DIPs) is described in light of current trends towards higher device powers and ambient temperatures. Extensive experimental and computer modeling studies on both package types have revealed that extraordinary improvements in thermal characteristics can result from proper selection of package materials and cooling parameters. In particular, judicious choices of leadframe material, molding compound, heatspreader, and mode of external cooling can lead to excellent thermal performance. To accomplish this, component manufacturers and system users must work closely to properly select materials and conditions to maximize effectiveness. In the best case, thermal resistances of 38°C/W and 26°C/W were found for 16- and 40-pin plastic DIPs using forced air cooling. The importance of both device power level and board temperature rise above ambient on theta JA is discussed in some detail. Suggestions are made to both component manufacturers and system users on how to minimize variations in measured theta JA .

Measurement of Silicon Strength as Affected by Wafer Back Processing

The distribution of strength of silicon chips has been measured by bending a chip between a soft ball and pad. The chip strength is controlled by flaws introduced by processing and the critical flaws can be located and analyzed after the chip breaks. The flaw sizes can be predicted reasonably well from the fracture toughness of silicon. Wafer back etching strengthens the chip, but the response depends on whether the wafer was lapped or ground.

Micro-Corrosion of Al-Cu Bonding Pads

Aluminum metallization films with copper additions are found to exhibit highly localized pitting in the presence of moisture. Galvanic action of aluminum surrounding Al2Cu theta phase particles causes localized aluminum corrosion. The thin layer of aluminum hydroxide corrosion product on the bonding pad creates an effective barrier to high quality wire bonding.

Chip corrosion in plastic packages

Abstract The corrosion performance of both unencapsulated and plastic encapsulated parts has been studied under temperature-humidity-bias (THB) stresses using a patterned test chip with aluminum metallization. The effect of encapsulation materials and ionic contaminants purposely introduced on the chip surface has been determined for both accelerated THB and severe, real life conditions. Unencapsulated test chips exhibited very different THB performance characteristics from encapsulated parts including a near-zero induction time and a shallow post-induction slope on the % failures vs time curve. The induction time and mean-time-to-failure (MTF) for encapsulated devices were strong functions of both the polymeric class of encapsulant and the particular compound employed. At least part of the difference in performance between molding compounds has been attributed to impurities inherent in the compounds. Impurities contributed from conductive epoxies used for die attach also adversely affect the devices MTF, as do normal contaminants present on as-plated leadframes. The environmental factors of temperature, relative humidity and bias greatly affect the performance of encapsulated devices. In this study, the MTF decreased dramatically with bias between 5 and 30 volts, the range over which most devices operate. The acceleration factors relating device performance in an 85°C/85%RH environment to severe, real life conditions were surprisingly small and somewhat dependent on the encapsulant.

Effects of Intermetallics on the Reliability of Tin Coated Cu, Ag, and Ni Parts

The temperature dependence of the growth rates of intermetallic compounds in the Cu-Sn, Ag-Sn and Ni-Sn systems was determined between 100°C and 213°C for Sn-dipped and Sn-plated samples. Below 175°C the fastest growing intermetallic compound was Ag3Sn. The Ni-Sn compoknd, Ni3Sn, was the slowest growing phase below 150°C, but the fastest growing phase above 175°C. The two Cu-Sn intermetallic phases, Cu3Sn and Cu6Sn5, had a combined growth rate which increases more slowly with temperature than the single intermetallic phases observed in the Ni-Snand Ag-Sn systems. The growth rate data plotted against reciprocal temperature satisfies the Arrhenius relationship and yields a range of apparent activation energies between 14.8 Kcal/mole for Ag3Sn and 37.6 Kcal/mole for Ni3Sn4. The growth of intermetallic compounds can affect the reliability of electronic parts through a reduction in lead solderability or a decrease in the mechanical strength of soldered connections. Lap shear testing of metal strips bonded with Sn demonstrated that strengths of both bonded Ni and Cu strips decrease as the thickness of the intermetallic compounds increase during annealing at 213°C. In the case of Ni-Sn, a 50% reduction in joint strength occurred after only one day-at temperature. For Cu-Sn, the reduction in joint strength occurred at a slower rate, and only a slight decrease was observed in the Ag-Sn lap shear samples. The degradation observed in the Ni-Sn and Cu-Sn systems is related to the growth rate of brittle intermetallics forming at the interface (i.e. Ni3Sn4 and Cu3Sn).

Thermal Studies of a Plastic Dual-in-Line Package

The major factors affecting heat flow in a 16 pin plastic dual-in-line package (DIP) are investigated using thermal resistance measurements of packages with various materials and design permutations. Factors explored include the molding compound material, leadframe material and design, die size, wire size, and die bond material. Of these, the leadframe material and molding compound most dramatically impact \theta JA while the leadframe design plays a secondary role. Other factors are of minor importance. The users external cooling conditions are very important to properly utilizing a packages thermal capabilities. By optimizing material selection in a standard 16 pin plastic DIP while using natural convection cooling, its thermal performance can he potentially increased by a factor of three. A factor of seven improvement is possible when using both optimum cooling conditions and superior packaging materials.

Role of materials evolution in VLSI plastic packages in improving reflow soldering performance

Cracking of surface mounted plastic packages worsens with increasing die sizes and thinner packages, both are recent trends in packaging. The precursor to failure, delamination at a leadframe to polymer interface, suggests that improvements in mold compounds, die attach adhesives and leadframe surface finishes are key elements in a solution. Identifying which specific materials properties must be improved and to what degree is a major task, followed by working with vendors to supply improved materials. In this study, the strategy is to improve all weak interfaces in parallel, rather than simply strengthen the weakest link. An excellent measurement method capable of detecting small improvements in the measured reflow soldering performance of a test package (148 PQFP) quantified both its delamination and cracking performance. These studies identified a general weakness in polymer to Ag die pad interfaces, implying that improving the adherend is mandatory.<<ETX>>

Organic contamination in IC package assembly and it's impact on interfacial integrity

To build reliable plastic packages, its essential to achieve excellent adhesion at all internal interfaces. One common inhibitor to good interfacial adhesion in plastic packages is surface contamination, most commonly organic in nature, accumulated prior to molding and causing poor adhesion with the mold compound. The most vulnerable and important surface is the chips topside passivation where organic contaminants are often responsible for delamination observed after molding or after stressing. Organic contamination also causes degradation of wire bond quality and in some instances pad corrosion failures. To successfully study adhesion or interfacial problems, its important to ensure the adherends (i.e., substrate) surface cleanliness prior to depositing the adhesive (mold compound). This paper discusses a low-cost, fast, and quantitative method for characterizing surface cleanliness by combining UV/ozone cleaning with contact angle measurements. The level of organic contamination on silicon from various assembly processes was quantified using this methodology. To demonstrate the impact of organic contamination, delamination performance of the silicon-mold compound interface is determined as a function of cleanliness (contact angle) on the silicon surface. The critical contact angle to achieve Level 1 crackfree performance in 148 PQFP with the window flag leadframe is determined.