No abstract

Solders are used extensively as electrical interconnects in microelectronics packaging. Because of environmental concerns, lead-based solders are being replaced by Sn/Ag and Sn/Ag/Cu based solder materials. Since the thermomechanical reliability of modern electronic devices depends on, to a large extent, the fatigue and creep behavior of the solder joints, it is imperative to understand the deformation behavior of these new lead-free solders. This study conducted extensive thermomechanical testing on several commercial lead-free solder alloys. Anand viscoplastic model was used to describe the behavior of these materials with new curve fitting techniques. A modified Anand models was proposed that can yield a more accurate description of lead-free solders.

No abstract

Solders are used extensively as electrical interconnects in microelectronics packaging. Because of environmental concerns, lead-based solders are being replaced by Sn/Ag and Sn/Ag/Cu based solder materials.Since the thermomechanical reliability of modern electronic devices depends on, to a large extent, the fatigue and creep behavior of the solder joints, it is imperative to understand the deformation behavior of these new lead-free solders. This study conducted extensive thermomechanical testing on several commercial leadfree solder alloys. Anand viscoplastic model was used to describe the behavior of these materials with new curve fitting techniques. A modified Anand models was proposed that can yield a more accurate description of lead-free solders.

In this study, quantitative microstructure studies were performed on multiple length scales to investigate the effect of lanthanum (La) doping on Sn-Ag leadfree solder materials. Factors considered in this paper include doping amount, aging temperature, and aging time. It was found that La doping reduces the grain size significantly, and the reduced grain size remains stable during thermal aging. The size of the Ag 3 Sn particles is also greatly reduced by La doping, and the particles coarsen during thermal aging, albeit at a much reduced rate than in the undoped alloy. The rate of particle coarsening can be described by a cubic-root law. Another observation is that the interparticle spacing remains unaffected by the doping. Therefore, higher La doping level leads to higher volume fraction of the eutectic region due to the increased total number of Ag 3 Sn particles.

In this paper, a user behavior based solder joint reliability modeling approach has been proposed to estimate the design and test requirements for the second level interconnect (SLI) reliability prediction. This approach uses a numerical tool to integrate solder joint creep damage during the actual use condition that was collected from a large user sample size. The resultant damage per time period was then input to the solder joint fatigue model to estimate equivalent damage to testing duration. The is a physics based approach and is expected to provide more accurate product life prediction and reliability performance demand for BGA package designs. IntroductionFor most electronic packaging failure mechanisms the user behavior plays a critical role in damage / degradation of the product. Currently, design requirements are based on a count of the number of times a system switches between power states and associated temperature ranges. A user survey is commonly used to collect user behavior data. The user survey result is then being translated to simple uniform temperature cycles based on a series of criteria (eg. power ON/OFF cycles/day counts). The translation is over conservative to account for the uncertainties in the user data.The first drawback of the current approach is the uncertainty of the user behavior. For electronic components, the heat generated by the silicon die is related to the work load. Furthermore, contemporary power saving features result in reduced temperatures when the component is in a standby or idle state. The "mini cycle" effect, the temperature fluctuation during normal use when the power is ON, should be carefully characterized. Currently, a simple assumption is that many small mini cycles are equal to several power cycles.The assumption may lead to conservative or unrealistic results based on the amplitude, duration and high or low temperature ranges in which the mini-cycle occurs.A significant drawback of current approach is not considering the temperature gradient of the component during use. Solder joint damage is caused by the thermal mismatch of the different materials in the packaging, so the temperature is the driving factor of the damage. The thermal cycle test condition uses a uniform temperature field for convenience and test efficiency. However, in the use condition, the silicon die is the heat source and the temperature gradient between the die-package and board will cause differential mechanical deformation, and a different solder joint damage distribution compared with the test condition. The empirically derived Norris-Landzburg (NL) equations cannot consider the effect of temperature gradients, transients or geometric factors; however, finite element modeling (FEM) can include the effect.

In this paper, a microstructure-dependent creep model is developed that accounts for the hierarchal microstructure at multiple length scales. The model considers three distinguishable phases in the solder alloy at two different length scales: at the larger scale Sn dendrites of micrometer size are embedded in a homogeneous eutectic region; at a much smaller length scale the eutectic region consists of submicron size Ag3Sn particles embedded in a homogeneous Sn matrix. The model predictions agree well with creep test data of lanthanum doped SnAg solders.

No abstract