Publication: Warpage behavior of thin fcbga package and prediction of its first interconnect snag solder joint shape
Date
2020-02-01
Authors
Lim, Shaw Fa
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The influence of substrate copper density distribution, substrate bump
coplanarity, stiffener attach process, and substrate clamping by magnetic boat during
die attach towards Flip Chip Ball Grid Array (FCBGA) assembled package warpage
were evaluated. The substrate warpage behavior throughout the package assembly
process was characterized using shadow moiré. In this study, it was found that a
balanced substrate copper density distribution (50/50 ratio), pre-stiffener substrate
before flip chip bump reflow, and substrate clamping during reflow able to reduce flip
chip solder bridging fall-out. The decrease in solder bridging <1% was due to the lower
substrate warpage seen during die attach. In particular, solder bridging fall-out was
well-correlated to die attach area warpage. Substrate with and without clamping during
reflow has met the package reliability requirement of temperature cycle 1200 condition
G (–40 °C to +125 °C). Simulation works through FEA (ANSYS) on the bare substrate
and package warpage was carried out and correlated to experiment data. Various
material properties and package designs was evaluated from the correlated FEA model
and its respective warpage behavior was understood. With understading of warpage
data through FEA, SnAg solder joint shape and its solder bridging can be understood
through Surface Evolver. The effect of solder volume, gap height and Under Bump
Metalization (UBM) size towards solder joint was evaluated. Higher solder volume
and smaller gap height led to higher occurrence of solder bridging. Solder joint
formation through solder cap copper pillar onto copper trace was predicted through
Surface Evolver. It can be shown that copper pillar solder joint geometry can be
successfully simulated and agreed with experiment. The relationship between various
solder joint influencing factors such as die bump diameters, copper pad geometry,
solder height and solder volume were established. The optimum die bump to copper
pad width ratio can be obtained through this simulation work. The information can be
used to estimate critical volume of solder needed for new, smaller pitch die bump and
copper pad design which help to save cost and time by avoiding a large number of
experiments prior to mass production.