Page 48 - Chip Scale Review_May June_2022-digital
P. 48
pitches (55µm and below) have been
demonstrated for flip-chip devices, they
are only done where a flip-chip device is
mounted onto a silicon substrate or onto a
substrate having a similar low coefficient of
thermal expansion (CTE) and are inherently
planar. Tighter array pad pitches will
require smaller diameter solder bumps with
much lower solder height. On-wafer FIWLP
devices can be fabricated with bump pitches
down to 50µm and below when used on a
silicon or other low coefficient of thermal
expansion (CTE) substrate. Molded-wafer
FIWLP devices are not dimensionally-
stable enough to support these low solder
bump pitches. Incorporation of Cu pillars
on array pads would allow FIWLP devices
to reduce solder bump height and, therefore,
array pitch. Although most FIWLP devices
Figure 5: Typical molded-wafer FIWLP device processing steps, from forming a molded-wafer and applying
RDL structures. are mounted on fine-line circuit boards, a
growing number will be assembled onto
of these next-generation semiconductor reduced chip placement tolerances and chip a SiP along with flip-chip devices and 3D
devices while maintaining high yields and movement; and 6) adaption of via locations chip stacks. These will utilize a silicon or
lowering costs, FIWLP fabricators need to and RDL routing. glass substrate permitting small pitch solder
develop the capability for: 1) smaller solder Smaller solder bump pitches. bump attach.
bump pitches; 2) tighter RDL line widths Increasing chip I/O counts requires Reducing RDL line widths and
and spacing; 3) higher thermal conductivity; shrinking array pad pitches of FIWLP spacing. Increasing the I/O count on a
4) reduced RDL interconnect parasitics; 5) devices. Although smaller solder bump FIWLP will also require RDL line widths
and line spacing to decrease proportionally.
Metallization and metal patterning
techniques particularly for molded-wafer
processing, need to shift from subtractive
metal patterning (standard PC panel
processing) to semi-additive metallization
techniques (standard semiconductor
processing) providing rectangular line cross-
sections, finer line width control and lower
interconnect resistance. Molded wafers have
a mix of low-CTE silicon chips and high-
CTE molding material and RDL dielectrics.
This combination can cause warpage, poor
planarity and variable and nonuniform
molded-wafer shrinkage. Molded-wafer
FIWLP has chip location issues related
to chip placement tolerances and chip
movement after placement. All of these
make going to finer features on molded-
wafer FIWLP devices problematic.
Higher power dissipation. Higher power
dissipation chips can cause over heating of
the chips and softening of organic materials
unless a low thermal resistance cooling
path is provided. Higher power dissipation
can also cause device hot spots that can
exacerbate solder fatigue in smaller solder
bumps. Since all fan-in devices have at
E-Tec Interconnect AG, Mr. Pablo Rodriguez, Lengnau Switzerland least one organic layer overlying the chip
Phone : +41 32 654 15 50, E-mail: p.rodriguez@e-tec.com surface, cooling a higher power dissipation
chip through its top surface or through its
solder bumps is less efficient than cooling
46 Chip Scale Review May • June • 2022 [ChipScaleReview.com]
46