Authored By:
Jagadeesh Radhakrishnan, Sunny Lu - Quanta Shangha, Al Molina, Olivia H. Chen, Wu Jin Chang, Xin Wang, Kok Kwan Tang, Scott Mokler, Raiyo Aspandiar
Intel Corporation
Summary
With each new generation, the complexity in the design of flip chip devices, as exemplified by thinner package stack-ups, larger device sizes, and multiple die configurations, is increasing significantly. This is creating new challenges in their surface mount manufacturing and their solder joint reliability. To improve surface mount solder joint reliability under mechanical stresses, such as those imposed under shock, drop, and vibration during transportation and end user handling, the use of polymeric materials to provide added reinforcement to the second level solder interconnects on flip chip ball grid arrays (FCBGA)and package-on-package (POP) solder joints has been proposed as a solution.
Some of the common examples of such polymeric reinforcement applications in manufacturing include, but are not limited to, corner glue edge bonding, underfill (UF) application, and epoxy-containing solder pastes. However, as the solder joints' pitch size and height decreases, control of the extent and uniformity of polymeric encapsulation of second level solder joints becomes more challenging. As a result, solder joint encapsulation materials (SJEM) have been developed to provide a better controlled and localized application process. Unlike other polymeric materials in use today, these SJEMs do not require an additional step for cure, since they are applied before the reflow soldering process step and cure during the reflow soldering process step.
Based on past studies on polymer reinforced solder joints, mechanical shock performance generally improved with the application of the polymer reinforcement and was less sensitive to the polymeric material properties as long as the material has acceptable application, curing, adhesion and fracture strength properties. However, thermal cycling reliability is more sensitive to certain material properties of the reinforcing polymer.The glass transition temperature (Tg) and the coefficient of thermal expansion (CTE) are two such properties [1-2]. Materials with a low Tg and a high CTE often lead to accelerated solder joint failure from thermal fatigue. Therefore, the material properties of SJEM and the extent of material coverage on the solder joint both play important roles in optimizing solder joint reliability performance under mechanical and thermal conditions.
In this paper, two SJEM material application methods, dispensing and dipping, will be studied for the extent and uniformity of their encapsulation of high density BGA solder joints. The solder joint mechanical shock and thermal cycling reliability from these two SJEM dispensing techniques, which correspond to different encapsulation coverage, will also be analyzed and discussed.
From the assembly and reliability test results, both SJEM materials showed process feasibility to be applied, reflowed, and cured with SAC305 solder paste. Both SJEM application methods showed promising mechanical shock and temperature cycle reliability. These materials can be considered as a solution to replace underfills and corner glues for smaller, finer pitch components in the future.
Conclusions
This study set out to determine the processability, mechanical and temperature cycle reliability of SAC BGA solder joints reinforced by encapsulation using SJEMs. A total of 5 different SJEMs from two suppliers were evaluated. Three of these SJEMS were applied by dipping the BGA package balls into a reservoir containing these SJEMs and two others were applied by dispensing along one edge the package after it was initially reflow soldered. The solder joint reliability enhancement provided by the SJEMs was compared to a selected underfill material.
From the process assembly results, all SJEMs types can be successfully applied and cured with a typical SAC305 reflow process without solder joint defects. When comparing the two application methods, the dipping method provided a better control on SJEM coverage through in-line dipping height, but the number of solder balls covered by the SJEM in the BGA array was very sensitive to the room temperature coplanarity of the component. The dipping method requires an additional feeder to carry out the application process. The dispensing method ensured sufficient SJEM material coverage each solder joint and was less sensitive to component warpage. However, this method requires additional dispensing equipment to carry out the application process.
The SJEMs covered solder joints formed by the dipping method contained slightly higher voiding when compared to the SJEM covered solder joints formed by the dispensing method. The dispensing method resulted in more volume of SJEM around the solder joints as well as better uniformity of this volume across the solder joint array for a package. However, this larger volume of SJEMs surrounding the solder joint can potentially make reworking the component more difficult.
Based on in-situ failures recorded during the shock event which the POP components were subjected to, solder joints reinforced with all SJEMs showed better mechanical shock margin when compared to the solder joints without any SJEMs (the control leg). This confirmed that use of SJEMs do enhance the shock reliability of POP components. Among the various SJEM materials evaluated, the Dis3 SJEM and the underfill material, performed the best and did not show any in-situ shock failures. SJEMs applied by the dipping method showed slightly better margin than those applied by the dispensing method. Based on the observed characteristic life on the1st and 2nd row NCTF nets, the ranking of the SJEMs according to the shock resistance of the solder joints formed when using them is: Bare board < Dis1 < Dis2 =< Dip2 < Dip1
Based on in-situ failures recorded during temperature cycling of the POP components, solder joints at the package corners formed when using dipping SJEMs showed improved temperature cycle performance then the control sample which had no SJEM reinforcement (the control leg). But, solder joints applied with dispensing SJEMs, did not show any significant difference when compared to the control leg. The underfill material evaluated in this study showed a negative impact to temp cycle reliability, with the failure occurring even before the control leg. Based on the observed characteristic life on the 1st and 2nd row corner NCTF nets, the ranking of the SJEMs according to the enhancement they provide in improving the solder joint temp cycle failure resistance is: Dipping > Dispensing ≥ Bare Board > UF. Dispensing SJEMs showed thicker IMC formation on the PCB side compared to the rest of the DOE legs.
Initially Published in the IPC Proceedings
|
