Technical Abstracts
In Case You Missed It
Artificial Intelligence
“APOLLO: An Automated Power Modeling Framework for Runtime Power Introspection in High-Volume Commercial Microprocessors”

Authors: Zhiyao Xie, et al.

Abstract: Computer engineers at Duke University have developed a new AI method for accurately predicting the power consumption of any type of computer processor more than a trillion times per second, while barely using any computational power itself. Dubbed APOLLO, the technique has been validated on real-world, high-performance microprocessors and could help improve the efficiency and inform the development of new microprocessors. (MICRO-54: 54th Annual IEEE/ACM International Symposium on Microarchitecture, 2021,

Assembly Materials Reliability
“Reliability Analysis of SnAgCu Lead-Free Solder Thermal Interface Materials in Microelectronics”

Authors: Mathias Ekpu

Abstract: Thermal interface materials (TIMs) are used in electronic devices to bridge the topologies that exist between a heat sink and the flip-chip assembly. Therefore, this study aims to investigate the reliability of SAC 405 and SAC 396 solder alloys in a microelectronics assembly. In this paper, SAC 405 and SAC 396 were used as the TIMs. The model, which comprises the chip, TIM and heat sink base, was developed with Ansys finite element analysis software and simulated under a thermal cycling load of between -40° and 85°C. The results were based on the total deformation, stress, strain and fatigue life of the lead-free solder materials. The analyses of the results showed SAC 405 is more reliable than SAC 396. This was evident in the fatigue-life analysis, where it was predicted it took about 85 days for SAC 405 to fail, whereas it took about 13 days for SAC 396 to fail. Therefore, SAC 405 is recommended as the TIM of choice compared to SAC 396. (Soldering & Surface Mount Technology, October 2021,

“Influence of Pad Surface Finish on the Microstructure Evolution and Intermetallic Compound Growth in Homogeneous Sn-Bi and Sn-Bi-Ag Solder Interconnects”

Authors: Yaohui Fan, et al.

Abstract: Low reflow temperature solder interconnect technology based on SnBi alloys is being considered as an alternative for SnAgCu (SAC) solder alloys to form solder interconnects at significantly lower melting temperatures than required for SAC alloys. Microstructural evolution after reflow and aging, especially of intermetallic compound (IMC) growth at solder/pad surface finish interfaces, is important to understanding fatigue life and crack paths in the solder joints. This study describes intermetallic growth in homogeneous solder joints of SnBi eutectic alloy and SnBiAg alloys formed with electroless nickel-immersion gold (ENIG) and Cu-organic surface protection (Cu-OSP) surface finishes. Experimental observations revealed that, during solid state annealing following reflow, the 50nm Au from the ENIG surface finish catalyzed rapid (Ni,Au)Sn4 intermetallic growth at the Ni-solder interface in both Sn-Bi and SnBiAg homogeneous joints, which led to significant solder joint embrittlement during fatigue testing. Intermetallic growth of (Ni,Au)Sn4 was decreased by Ag alloying of eutectic SnBi solder and was completely eliminated by changing the metallization from ENIG to Cu-OSP on the board side of the assembly. The reduction in (Ni,Au)Sn4 growth rate with Ag additions is attributed to changes in grain boundary wetting of the IMC by Bi with Ag alloying. (Journal of Electronic Materials, October 2021;

Others of Note
University of British Columbia researchers have created what could be the first battery that is both flexible and washable. In addition to watches and patches for measuring vital signs, the battery might also be integrated with clothing that can actively change color or temperature. (

Facebook revealed a haptic glove designed to give the wearer sensations that mimic the weight and feel of real objects when they are handled in virtual space. Slip on this glove, and you can be convinced you’re holding the real thing (or something close to it), even when the object is entirely digital. (

Washington University (St. Louis) researchers have designed a new processing-in-memory (PIM) circuit that brings the flexibility of neural networks to bear on PIM computing. The circuit has the potential to increase PIM computing’s performance by orders of magnitude beyond its current theoretical capabilities. (

This column provides abstracts from recent industry conferences and company white papers. Our goal is to provide an added opportunity for readers to keep abreast of technology and business trends.