AREA CONSORTIUM REPORTS

2024 REPORTS

Design Mitigation of VIPPO Hot Tear Through Partial Stacked Microvias
Author:  Jeff Kurish

Abstract:
Circuit board design complexity has led to the wider adoption of the Via-In-Pad Plated over (VIPPO) geometry. A specific fault mechanism plagues multiple reflowed mixed arrays of VIPPO and non-VIPPO pads. Before melting, the bulk solder of a VIPPO pad is separated from the IMC by the expansion of the PCB in the non-VIPPO pad. Herein, the effect of this fault in accelerated thermal cycling (ATC) studies has been evaluated. Further, the effectiveness of partial vias in mitigating the fault was probed Faults were shown to be inversely proportional to thickness of partial via A relationship between the location of the partial via within the z-direction of the board was shown.

Avoiding VIPPO Hot Tear via Nitrogen Reflow: Review Addendum
Author:  Jeff Kurish

Abstract:
Use of nitrogen atmosphere in reflow of PCB products has been increasing with the increasing degree of complexity of reflowed products. Though often dismissed due to cost limitations, the value of reduced oxidation can have plethora benefits, including reduction in via in pad plated over (VIPPO) hot tear fault formation. This addendum is intended to generate a full picture specifically about the information available from literature sources as well as from AREA consortium work regarding the effectiveness of inert atmosphere in preventing VIPPO hot tear.

Durafuse LT Shear Testing on Gold-Containing Surface Finishes
Author:  Alyssa Yaeger

Abstract:
Durafuse LT is a lower melting temperature, mixed alloy solder system containing Sn, Ag, Cu, and In. Previous experiments testing the shear and drop performance of Durafuse LT with both OSP and ENIG surface finishes indicated a potential reaction between the Durafuse LT alloy and the ENIG surface finish. It was suspected that the gold from the surface finish may have been reacting with the alloy to create or exacerbate the black pad-type defects seen in the fracture surfaces. In order to better understand this phenomenon, this experiment tests the shear strength of three gold-containing surface finishes: ENIG, Direct Immersion Gold (DIG), and a novel Ni-less ENIG (LILO). This experiment was done in combination with the liloTree surface finish shear test evaluation in “Evaluation of liloTree Nickel-less ENIG PremiumTM Surface Finish” and will reference the results of other alloys with these surface finishes in order to understand the difference seen with Durafuse LT.

Evaluation of Lilotree Nickel-Less Enig PremiumTM Surface Finish
Authors: Joseph Brown, Nancy Wang, and Alyssa Yaeger

Abstract:
As the world shifts to higher frequency circuit applications, the insertion loss of the NiP layer in the Electroless Nickel Immersion Gold (ENIG) finish has become a concern. One possible solution is to replace the NiP layer of a typical ENIG finish with a Cu-organic layer (Ni-Less ENIG PremiumTM). Without the poor conductivity of Ni, the organic layer should cause less loss while also retarding copper migration through the gold layer resulting in a longer shelf life and reduced intermetallics (IMCs). Insertion loss testing of Ni-less ENIG (LILO) showed losses between ENIG and DIG up to 26.5GHz. We also conducted individual ball shear tests before and after accelerated aging on Direct Immersion Gold (DIG), ENIG, and novel Ni-less ENIG finishes alongside microstructure analysis to determine the growth of IMCs and their effects on mechanical reliability for each of the finishes. We apply the same analysis for Sn67Pb33, SAC305, and Sn/3.6Ag/.6Cu/.1Ni/1.3Sb/2.8Bi (Innolot) solders. It was found that solder joints on the Ni-less ENIG finish were much weaker than both DIG and ENIG finishes and had similar IMC growth to the DIG finish. Planar voiding was found in all Ni-less ENIG joints which we attribute to the inability of the organics to escape the solder matrix during the reflow process. Surface oxidation and copper diffusion analysis were performed to test the shelf-life and thermal stability of each surface finish. Surface finishes were exposed to a variety of thermal excursions ranging from room temperature shelf life to accelerated aging as in the ball shear tests. Each surface finish was etched with argon gas in order to remove surface contaminants and expose the top few nanometers of the surface. The Ni-less ENIG finish retained surface contaminants after significantly more etching than ENIG and DIG surface finishes, possibly due to the roughness of the finish. Ni-less ENIG also had the highest ratio of copper to gold, and the most persistent copper oxides.

Electromigration of Durafuse LT Solder
Author:  Alyssa Yaeger

Abstract:
Electromigration of SnBi solders are heavily studied due to concerns of bismuth buildup at solder joint interfaces causing resistance increase and decreased mechanical performance. There is some literature indicating that indium is also subject to electromigration, which may also increase joint resistance over time. This experiment explores current stress comparisons between Sn-37Bi (B37), Durafuse LT, and SAC305. LGA joints were temperature and current stressed to 2.5kA/cm2 and 90oC. Durafuse LT has an initial resistance similar to SAC305 and showed slight joint resistance increase over 1500 hours, in comparison to a large resistance increase for B37 joints and a slight decrease for SAC305 joints. Current stress also affected the Durafuse LT microstructure, redistributing Ag-In-Sn IMCs throughout the joint rather than the typical clustered groups seen in Durafuse LT microstructure.

Isothermal Electromigration of SnBi Solders
Author:  Alyssa Yaeger

Abstract:
This study focuses on accelerated current stressing to encourage electromigration of Bi and understand the implications of electromigration on SnBi containing solder joints. Package-to-board level BGA solder joints were tested under a joint temperature of 90 oC and a constant current density of either 1.7kA/cm2 or 2.5kA/cm2. A near-eutectic SnBi (L29) and lower Bi SnBi alloy (B37) were used to form both homogeneous SnBi joints and hybrid joints with SAC305 spheres and SnBi paste. SAC joints were also tested with a high current density of 2.5kA/cm2 for reference.

Effects of Electromigration Pre-Stress on Thermal Cycling Lifetime
Author: Alyssa Yaeger

Abstract:
Electromigration of SnBi solders remains a concern not only for joint resistance increases but for mechanical reliability. To determine the effects of electromigration on thermal cycling lifetime, SAC, B37 (SnBi), and B37/SAC solders were subjected to current and temperature stress prior to thermal cycling. These were compared with matching samples which had not experienced current stress in order to determine the effects on the thermal cycling lifetime.

Without electromigration pre-stress, B37 outperformed SAC, and B37/SAC had the lowest lifetime for 0/100oC thermal cycling. Electromigration pre-stress caused a buildup of Bi at the interfaces for the Bi-containing alloys and led to a decrease in lifetime for SAC and B37 and an increase in lifetime for B37/SAC hybrid samples. The hybrid solder joints failed due to ball drift with and without electromigration pre-stress. Although B37 joints showed a reduced lifetime with electromigration pre-stress, they performed as well as pre-stressed SAC. This is in part due to the degradation of the pre-stressed SAC samples caused by high temperature exposure mimicking annealing.

Effects of Temperature and Current Stressing on Low Temperature Solder BGA Drop Performance
Authors: Alyssa Yaeger, Michael Meilunas and Sufyan Tahat

Abstract:

BGA components constructed with low temperature solder (LTS) SnBiSbNi spheres were assembled using a matching alloy paste and then subjected to simultaneous direct current and temperature stressing to generate electromigration of the bismuth atoms. The samples were then drop tested to evaluate solder joint lifetime performance and the results were compared to non-stressed control samples. Solder joint failure analysis was performed by cross-sectional examination using scanning electron microscopy. The experiment described indicates that the current densities, joint temperatures, and stress durations examined could result in a significant reduction in drop reliability – even when minimal bismuth electromigration was observed in cross-section. The extent of the reliability reduction was highly dependent upon the solder joint temperature with joint temperatures of 85C and 100C resulting in rapid joint degradation while joint temperatures of 75C showed little or no reduction in reliability relative to the control samples – even when the 75C samples were subjected to relatively high current densities.

Key words: electromigration, drop/shock, LTS, bismuth, BGA.

Thermal Cycle Reliability Assessment of Hybrid SAC305_Cyclomax BGA
Author:  Michael Meilunas

Abstract:
The experiment described in this report is an accelerated thermal cycle reliability analysis of lead-free SAC305 BGA componentry assembled using Cyclomax solder paste supplied by Accurus. Such assemblies are referred to as “Hybrid” due to the differing solder alloy compositions. Multiple stencil aperture designs were included in the evaluation in order to study the effect of Cyclomax solder paste volume on the reliability of six BGA component types. SAC305 samples were produced using the BGA components and a SAC305 paste for general baselining purposes. BGA assemblies were constructed using a single stencil aperture design for each BGA type. Samples were subjected to −40 to 125C accelerated thermal cycling and the failure data was analyzed using Weibull plots. Failure analysis was performed on each test population and included electrical failure mapping, x-ray examination, dye penetrant analysis and cross-sectioning. The results indicate that the Cyclomax solder paste produced very good characteristic lifetimes relative to the SAC305 baseline populations. Correlation between solder paste volume and reliability was observed among/within some populations. It is proposed that the volume effects were primarily due to differences in solder joint shape and size more so than alloy compositional changes.

High Reliability Solder Device Shear on Aluminum-Clad PCBs
Author:  Alyssa Yaeger

Abstract:
High reliability or automotive alloys are often tested for thermal cycling optimization, but the use of high reliability solders in LED applications require slightly different testing. High-power LEDs have steady operating temperatures for long periods of time, and are often mounted on aluminum-clad PCBs in order to effectively spread and remove heat. The high CTE of aluminum can lead to design and fabrication issues as well as high shear stress during thermal or power cycling due to the low CTE of the ceramic LED body. Shear testing at both room temperature and elevated temperatures can help determine the best solders for this application.

Effects of Solder Joint Solidification Temperature on BGA Failure Behavior During Accelerated Thermal Cycling|
Author:  Michael Meilunas

Abstract:
Accelerated thermal cycle (ATC) testing is often used to compare the relative performances between different solder alloys such as SnPb and SAC with the assumption that the alloy that performs better in thermal cycling will likely perform better in actual usage environments.

Exploring The Effects of Solder Solidification Temperature on BGA Reliability Using Finite Element Modeling
Authors: Michael Meilunas and Mohammed Abdel Razzaq

Abstract:
AREA Consortium research has repeatedly found that otherwise identical BGA assemblies of differing solder alloys can produce significantly different behaviors in accelerated thermal cycling experiments. These behaviors can include:
1.) Alloy “A” performing better than Alloy “B” when exposed to one thermal cycle condition, but worse than Alloy “B” when exposed to a different thermal cycle condition
2.) Alloy “A” and Alloy “B” failing at significantly different solder joint positions
3.) Alloy “A” and Alloy “B” exhibiting different failure modes for the same thermal cycle condition
4.) Possible failure mode changes with changing thermal cycle parameters

Often, the alloys being compared have significantly different solidification temperatures and it has been proposed by the authors that the behaviors listed previously may be due, in whole or in part, to the different “residual stress” states that are created when the BGA body and PCB are united during the solidification phase of the assembly process. It is also proposed that BGA temperature dependent warpage characteristics are often the dominant driver of the “residual stress” conditions and therefore, the effect of solder joint solidification temperature on the BGA stress state is more readily apparent among samples that tend to warp significantly during the reflow process.

The following report describes a simple procedure used to physically compare the effects of solder joint solidification temperature on a BGA component and presents a novel approach that was used to construct a finite element model of the system in an effort to simulate solder joint stress shortly after solder solidification and during accelerated thermal cycling. The results of the physical examination show that components assembled with solder alloys that melt at approximately 220, 183 and 140C have significantly different body shapes and solder joint heights after solidification while the finite element model shows that there are significant differences in stress distribution and stress levels for all three solidification temperatures evaluated immediately after reflow and throughout simulated thermal cycling.

Laser Reflow Thermal Limitations
Author:  Jeff Kurish

Abstract:
Despite a presence in SMT assembly processes for multiple decades, growth in usage of laser reflow (LR) instrumentation remains sluggish. Packaging processes remain tied to robotic, selective, or hand soldering methods largely due to fears of LR inefficiencies in poor reflow, poor surface wetting and PCB damage. In this work, the interaction of 980 nm laser light with various base materials of SMT processes, copper, SAC305, and FR4 boards is investigated. Subsequently these interactions are modeled to predict temperatures in laser reflow profiles. These models are then used to probe the difficulties hampering LR processes. Solutions are offered for this space and PCB design rules are generated to make LR a more viable solution.