Effects of Mechanical Cycling Induced Damage on the Creep Response of SAC305 Solder

Abstract

In electronic packaging, lead-free solders often experience fatigue failures due to thermal-mechanical cyclic stress and strain caused by changing temperatures and mismatches in thermal expansion coefficients. As a result, damage accumulates in the solder joints including plastic deformation, crack initiation, crack propagation, and finally failure occur. In our previous work, changes in the mechanical behavior of SAC305 lead free solder due to prior damage accumulation was investigated. Circular cross-section solder specimens were first reflowed, and these samples were then mechanically cycled for various durations using a Micro-Mechanical tester. Monotonic stress-strain tests were subsequently conducted on the prior cycled samples to characterize the change in mechanical behavior occurring in the solder due to damage accumulation. Using the data from these tests, we were able to characterize and quantify the cycling induced damage through the observed degradations of several mechanical properties (initial elastic modulus, yield stress, and ultimate tensile strength) with the amount of prior cycling. In the current work, we have extended the experimental work in our prior studies on SAC305 to examine the evolution of the creep response due to prior damage accumulation. In the experimental testing, small uniaxial cylindrical samples of SAC305 solder were prepared and reflowed in a reflow oven. These specimens were then mechanically cycled under several different sets of conditions to induce various levels of damage in the samples. In particular, four levels of initial damage per cycle were considered (ΔW = 0.25, 0.50, 0.75 and 1.00 MJ/m3), as well as three cycling temperatures (T = 25, 100, and 125 °C). For each of these damage levels per cycle, various durations of cycling were applied (e.g., 0, 50, 100, 300, and 600 cycles). This test matrix generated a large set of prior damaged samples, where the damage had been accumulated at different rates (different damage amounts per cycle), different cycling temperatures, and for different durations. In this paper, selected results obtained for isothermal mechanical cycling at T = 25 °C will be presented in detail. Creep tests were performed on the prior damage samples at room temperature and several stress levels including σ = 10.0, 12.0, and 15.0 MPa. The changes in the steady state secondary creep rate were then evaluated and plotted versus the duration of cycling for the various applied levels of damage per cycle. Exponential empirical models were found to fit the material property degradations well for any one set of conditions. More importantly, it was found that the total energy dissipation that had occurred in the sample (sum of ΔW for all cycles) could be used as a governing failure variable independent of the damage level applied during each cycle. In particular, all of the creep rate data for a selected stress level were modeled well using a single degradation curve independent of that rate the damage was accumulated. Using the results of this study, we are working to develop better damage mechanics models and fatigue criteria for lead free solders that are subjected to variable temperature applications.