In Summer 2010, I was selected to take part in the National Science Foundation’s National Nanotechnology Infrastructure Network’s Research Experience for Undergraduates (NSF NNIN REU) program, which supports undergraduate students to work on mentored research projects at 14 NNIN sites across the country. I was 1 of 5 interns selected to participate at NNIN REU’s Stanford University site, hosted by the Stanford Nanofabrication Facility.
I spent the summer with Professor Reinhold Dauskardt’s research group in Stanford’s Department of Materials Science and Engineering, working on an Intel-funded project to improve adhesion in flip chip devices with mentorship from Dr. Taek-Soo Kim.
PROJECT BACKGROUND
As electronic devices become increasingly complex, advances in packaging technology — the ways in which devices are connected electrically and mechanically to the circuit board — become increasingly critical. Optimal packaging technology not only reduces the overall size of the package, but also accommodates an increasing number of more complex connections in ways that reduce thermo-mechanical operational stress.
Flip chip packaging is a common packaging method in which the chip is connected to its substrate with solder balls. An underfill epoxy resin matrix surrounds the solder ball array to offer mechanical support and thermal protection.
The goal of our research was to optimize the adhesive strength of these underfill epoxies in order to develop more structurally reliable flip chip devices.
METHODOLOGY
To test the adhesion of epoxy to silicon substrates, we created double cantilever beam (DCB) specimens, which essentially consist of silicon panels sandwiched together by epoxy:
I spent my summer fabricating DCB test specimens under different experimental conditions, and then running fracture mechanic tests on them to determine the fracture energy — or, the amount of energy per unit area required to propagate a crack through a material — of the inner epoxy.
RESULTS
What were the various experimental conditions we were subjecting the DCBs to, and how did they affect the final adhesive strength of our devices? To answer that, let’s take a look at some of our original research questions:
Does the addition of flexibilizers affect epoxy durability?
We tested the adhesive strength of epoxies both with and without flexibilizers, which are functional molecules that can be introduced to the epoxy matrix for increased molecular toughness. Our research indicated that yes, adding flexibilizers can increase the fracture energy of the epoxy — and higher fracture energy means more energy was required to break the DCB specimens, which indicates stronger adhesion. This is likely due to the increased ductility that flexibilizers introduced to the epoxy.
Does the addition of a hybrid film layer improve epoxy adhesion?
Bonding at organic (e.g. epoxy) / inorganic (e.g. silicon) interfaces has long posed challenges in applied engineering. However, we can leverage the multifunctional properties of hybrid materials that contain cross-linked organic and inorganic networks to address these challenges. We applied hybrid zirconium sol-gels — which are graded hybrid films featuring both organic epoxy groups and inorganic zirconium oxides that could respectively bind to organic epoxy surfaces and inorganic metal oxides — to our DCBs in an attempt to improve adhesion right at the epoxy-silicon interface.
Overall, our research indicated that yes, the addition of hybrid zirconium sol-gels can improve the adhesive strength of epoxies to silicon. Furthermore, we were able to experiment with different sol-gel preparation methods and found that fracture energy increased (i.e. adhesion was stronger) if the hybrid solution was stirred during the aging portion of its preparation, and cured (this refers to the step after which epoxy and sol-gel layers are applied to the DCB and everything is put in the oven to essentially “bake”) for longer. Such prep conditions likely allowed the hybrid layer to achieve a more condensed state with greater network cross-linking, and thus higher fracture energy.
WHERE CAN YOU READ MORE?
At the end of the program, I presented our research findings both to my lab mates as well as the broader Stanford REU program. I also gave a broadcasted oral presentation of our results and presented this research poster at the end-of-summer NNIN Convocation at the University of Minnesota.
Here’s my research as published in the 2010 NNIN REU Research Accomplishments booklet and as presented by my Stanford lab mates at TECHCON2011.
GILLIAN LUI 