Micross

Abstract Preview

Here is the abstract you requested from the dpc_2018 technical program page. This is the original abstract submitted by the author. Any changes to the technical content of the final manuscript published by IMAPS or the presentation that is given during the event is done by the author, not IMAPS.

Effective Method for Wire Bonds Rework using Conductive Epoxy
Keywords: Wire bond, Conductive epoxy, Rework
Behind every new electronic device, hidden hours of challenging work and resources were invested in its development and packaging. Thus, it is no surprise that companies are looking for innovative means to lower their costs. In the case of wire bonded chips, its development requires to go through multiple designs to achieve the best packaging performances. Nevertheless, some of the designs may not be tested due to defaults such as damaged traces or insufficient thickness or lack of bonding metallurgy, preventing the wire bonds to be effectively completed. Unless an effective way to rework such failed assemblies is feasible, the invested resources in the design and manufacture are lost. The MiQro Innovation Collaborative Center (C2MI) has developed a method to rework such assemblies. We will present the results of such a method using two silver filled conductive epoxies to repair failed gold wire bonds compared to successfully realized gold wire bonds. Experiments were conducted on two identical test chips with gold pads on each substrate. A matrix of 1 mil diameter gold wire bonds was realized on a K&S Iconn ProCu Plus ball bonder. The bonding quality of the reference wires onto the test chip pads was assessed prior to the experiment by ball shears and wire pulls. Then, on each substrate, a first group of wire bonds was used as reference and the others were sheared at the stitch bond and reworked with the two-different silver filled conductive epoxies. The rework method consists of reattaching the sheared end of the wire to the landing pad on which a drop of conductive epoxy has been deposited. Each group is at least composed of 40 samples for statistical analysis of the method reproducibility. The resistance of reference and reworked wire bonds were measured on a two probes station to measure the effectiveness of the conductive epoxy rework. The quality of the connection between the reworked wire, the conductive epoxy and the gold pad was investigated through cross-sections and the mechanical reliability by wire pull. On both samples, reworked wire bonds with the first epoxy gives similar resistance than the reference ones, with a mean value of 90 mohm for the reference gold wires and a mean value of 96 mohm for the reworked wires with epoxy one. A visual inspection of the cross section of the reworked area shows no visual defect. We can also see that the deposited epoxy drop creates a spheroid shape onto the landing pad traducing its good capability to easily wet both the wire and the pad. The second epoxy was much more complicated to use as the deposited drop did not well wet the gold wire and was stretching, making it hard to complete repeatable connections and increasing the risks of shorting with a neighbor wire. Moreover, although wires could be re-attached to the landing pad using the second epoxy, the measured resistance was more than twice the one of reference wire bonds, with a mean value of 253 mohm. The lower capability of the second epoxy to be easily used for the method can be seen by visual inspection of the cross section of the reworked area. The wetting of both the pad and the wire is incomplete, explaining the higher resistance of this group of reworked wires. However, the pull tests revealed that all the wires, reference and reworked with epoxy one and two, pulled at the same force values, around 12 g, and broke mostly at the neck of the first ball bond. We demonstrate that failed wire bonds could be easily and effectively re-attached to the substrate using the appropriate conductive epoxy. Indeed, the electrical measurement of such reworked interconnections is comparable to the reference value of non-failed ones. Furthermore, we demonstrated that the method is robust and reproducible. That simple solution enables to reduce development costs by reworking junked considered assemblies to avoid the waste of resources and delays in the run to market for the new product.
Catherine Marsan-Loyer,
MiQro Innovation Collaboration Center (C2MI)
Bromont, Quebec
Canada


CORPORATE PREMIER MEMBERS
  • Amkor
  • ASE
  • Canon
  • EMD Performance Materials
  • Honeywell
  • Indium
  • Kester
  • Kyocera America
  • Master Bond
  • Micro Systems Technologies
  • MRSI
  • NGK NTK
  • Palomar
  • Plexus
  • Promex
  • Qualcomm
  • Quik-Pak
  • Raytheon
  • Specialty Coating Systems