|
|
|
Solder Joint Reliability on Mixed Ball Grid Array Solder Joints
Production Floor |
|
Authored By:Olivia H Chen, Al Molina Intel Corporation, Folsom, CA, USA James Gao, Tim C.C. Pan Wistron InfoComm (Kunshan) Co., Ltd., Kunshan, PRC Kok Kwan Tang Intel Corporation Kulim, Malaysia Raiyo Aspandiar, Ph.D., Kevin Byrd, Bite Zhou, Scott Mokler Intel Corporation Hillsboro, OR, USA SummaryDue to the decreasing size of consumer electronic products such as smart phones, tablets, and personal computers, ultrathin flip chip ball grid array (FCBGA) packages are needed to meet the demand of lower z-heights for slimmer form factors. However, packages used in consumer electronics are commonly assembled on printed circuit boards (PCB) with lead-free SnAgCu (SAC) solder paste at peak reflow temperatures in the 240C to 260C range. Assembly challenges are observed at these peak reflow temperatures due to dynamic warpage of the component substrates of ultra-thin FCBGAs, as well as the PCB. The Bi-Sn low temperature solder metallurgical system has been proposed as an alternative to the SAC metallurgical system to overcome these package substrate and PCB warpage iduced assembly challenges. Besides improving solder joint yields on ultra-thin FCBGAs, the lower melting point of this Bi-Sn metallurgy also enables manufacturing cost savings and environmental benefits. However, based on previous literature studies, the presence of Bi in Sn-based low temperature solder has exhibited solder joint embrittlement and thereby decreased the mechanical shock resistance of such solder joints. In order to strengthen the solder joint, low temperture solder pastes have been developed containing resin, which by flowing around the solder joint and curing during the reflow process it provides polymeric encapsulation reinforcement at the solder joint level. In a previous study [16], this type of low temperature joint reinforced paste (JRP) had shown improved mechanical shock resistance on mixed BiSn+SnAgCu FCBGA solder joints when compared to those without the polymeric encapsulation. However, the effect of thermal fatigue on the resin reinfroced BiSn+SnAgCu BGA solder joint when subjected to thermal cycling needed to be better understood. In this paper, the investigation centered on both the mechanical shock and the thermal cycling performance of the mixed BiSnAg (BSA)+SAC solder joints assembled with JRP pastes on a Flip Chip Molded Package BGA (FCMB) and compared the data to those assembled with SAC305 paste. Reliability failure rate characterized with Weibull distributions, failure modes and locations within the solder joint stack-up were determined. Results indicated that the mixed SAC+BiSn solder joints with polymeric reinforcement when using two different JRPs were less shock resistant than the SAC solder joints for both JRPs. More development is therfore necessary with JRPs to enable low temperature Bi-Sn solder joints to become comparable with un-reinforced SAC BGA solder joints under mechanical shock. There was significant improvement in the thermal cycling performance for mixed SAC+BiSn solder joints formed using one of the JRPs, but this improvement was for the solder joints located at the package corners only. The level of encapsulation of the solder joints by the reinforcing resin was inconsistent across the FCMB array and this could have caused this inconsistent temperature cycling resistance. Key words: BGA solder joints, low temperature solder, Bi-Sn metallurgy, Mechanical Shock Reliability, Temperature Cycle Reliability, Polymeric reinforcement ConclusionsThis study set out to determine the mechanical shock and temperature cycle resistance of mixed SAC-BiSn FCMB BGA solder joints, which were polymerically reinforced by using two Bi-Sn JRPs, JRP1 and JRP2. Their resistance to shock and thermal fatigue was compared with that of SAC solder joints without any reinforcement. For mechanical shock, results indicated that based on in-situ failures recorded during the shock event, solder joints formed with both JRP1 and JRP2 showed lower number of drops to 63.2% failure when compared to SAC305. CTF failures were seen on both JRP legs, but not on SAC or SAC + Corner fill legs. This indicated that these mixed SAC+BiSn solder joints with polymeric reinforcement were less shock resistant than the SAC solder joints. Based on observed characteristic life on both NCTF nets and CTF nets, the ranking of the solder pastes according to the shock resistance of the solder joints formed when using them is JRP1 < JRP2 < SAC305 < SAC305 + Corner fill. For thermal fatigue, as determined under temperature cycling, results indicated that solder joints formed when using JRP1 showed significantly improved temperature cycle performance on the solder joints at the package corners, but for solder joints on inner rows of the package there was not significant difference when compared to that of the SAC solder paste leg. Based on observed characteristic life on 1st and 2nd row corner NCTF and the POP DC package inner net, Temp cycle resistance: JRP1 >> SAC305 on corner most NCTF; JRP1 ≈ SAC305 on F26 POP DC net, the ranking of the solder pastes, according to their resistance to thermal fatigue under temperature cycling is: JRP1 >> SAC305 on corner most NCTF; JRP1 ≈ SAC305 on F26 PoP DC net. From these results, the conclusion that can be surmised is that, although this resin reinforced solution has shown shock performance improvement from mixed SAC+BiSnAg solder joints formed when using traditional BSA solder pastes [3], the two JRP pastes in this study do not show comparable performance to SAC305 paste when subjected to test condition in accordance to JESD22-B111 standard to assess mobile product drop capabilities. Though there was significant improvement in the thermal fatigue resistance of the solder joints formed with one of the JRP pastes evaluated, this improvement was not consistent across all solder joints in the FCMB package array. The encapsulation levels of the solder joints with the reinforcing resin varied widely across the package solder joint array, and this could be one reason for this inconsistency in thermal fatigue improvement. Hence, further work is necessary to understand the effect of resin coverage of solder joints on their temperature cycle performance needs to be better understood. Initially Published in the SMTA Proceedings |
|
Comments
|
|
|
|
|
|
|