Thermal cycling, fatigue failure

Experience with PBCAs has shown that the relatively high die stiffness and the flexibility of the laminate substrate generally cause first thermal cycle fatigue failures in the ball closest to the die shadow. Despite the overall economy of elements in each strip model, selective mesh refinement was used to concentrate highly refined elements in the solder joints where failure was anticipated. 

On a physical level, the number of thermal cycles to failure is affected by the strain imposed on the Cu in each cycle and the fatigue resistance of the copper. These factors are in turn controlled by a number of environmental, material, and manufacturing parameters. Low-cycle metal fatigue, in which most of the strain is plastic strain, can be treated approximately with the Coffin-Manson relation ... 

A study (Ref 38) compared fatigue failure modes of lead-free solders to lead-tin solder joints, for a BGA package mounted on an or-gaiuc board. In the study, the assemblies were subjected to thermal cycling until failure, and a comparison of the failure modes for lead-free solder and lead-tin solder was made. For the case of Sn-Pb solder, the crack path lies in the bulk of the solder material near the component body (Fig. 10). For lead-free solders, however, two distinct fracture paths are observed. The first fracture path is similar to that of lead-fin solder passing through the bulk of the solder near the component side. This is shown for SAC solder (Fig. 11) and is also observed for Sn-Ag solder. The second fracture path is characteristic only of the Sn-Ag and SAC solders. This fracture path consists of very fine cracks with multiple fronts, has a shattered appearance, and is seen near both the component and the board sides (Fig. 12). 

Most solder fatigue models correlate the number of thermal cycles to failure with the parameters that quantify the failure mechanism. Because solder fatigue is one of the prominent mechanisms found in product field failures, significant efforts have been made to develop models that characterize creep-induced fatigue. In general, the number of cycles (Nf) to failure of a solder joint in low-cycle fatigue is described by the equation ... 

A uPVC rod of diameter 12 mm is subjected to an eccentric axial force at a distance of 3 ttun from the centre of the cross-section. If the force varies sinusoidally from — F to f at a frequency of 10 Hz, calculate the value of F so that fatigue failure will not occur in 10 cycles. Assume a safety factor of 2.5 and use the creep rupture and fatigue characteristics described in the previous question. Thermal softening effects may be ignored at the stress levels involved. 

Fatigue corrosion occurring as Thermal fatigue cracking (thermal effect corrosion Corrosion fatigue Cycles of thermally induced stress leads to metal failure. Results from a combination of thermal cycling stress and SCC or other corrosion process. 

Fatigue failure may occur when a specimen fractured into two parts, was softened or its stiffness significanfly is reduced by thermal heating and/or cracking. Sometimes, for different reasons, a large number of cycles elapses from the first formation of microscopic cracks to complete frac-mre. In this case, the fatigue failure is arbitrarily defined as having occurred when the specimen... 

Flexometer tests are used to determine thermal stability under dynamic straining conditions. Measurements include temperature rise after a specified period of cycling, set and creep, and in some instances the time or number of cycles to failure in the form of thermal runaway or test piece destruction. In contrast to fatigue cracking tests, heat buildup tests... 

The 10-layer board was subjected to thermal cycling from 0 to 100°C, in accordance with NASA specification NHB-5300.4. No failures in electrical continuity were observed after more than 500 cycles, indicating that the high z-direction CTE does not cause fatigue failure in the plated through holes. 

Service tests characterize the performance of materials in specific applications and should mimic the environment of the actual application as much as possible. This includes simulating the chemical composition of the atmosphere, the temperature and thermal cycle profiles, stress states, fatigue conditions, and design. These tests may be accomplished by exposing test racks to actual service conditions or through the use of prototype apparatus designed and constructed to duplicate the end-use application. Additionally, laboratory service tests designed to evaluate the effect of one or more critical aspects of the exposure conditions may be employed. Service test data enable one to determine failure mechanisms, component or material life, or to screen materials for an application. 

---