Download Adhesion strength and contact resistance of flip chip on flex packages

Microelectronics Reliability 44 (2004) 505–514
www.elsevier.com/locate/microrel
Adhesion strength and contact resistance of flip chip on
flex packages––effect of curing degree of
anisotropic conductive film
M.A. Uddin, M.O. Alam, Y.C. Chan *, H.P. Chan
Department of Electronic Engineering, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong
Received 17 March 2003; received in revised form 5 May 2003
Abstract
The use of anisotropic conductive adhesives film (ACF) as an interconnect materials for flip-chip joining technique is
increasingly becoming a vital part of the electronics industry. Therefore, the performances of the ACF joint turn into an
important issue and depend mostly on the curing condition of the ACF matrix. ACF is a thermosetting epoxy matrix
impregnated with small amount of electrically conductive particles. During component assembly, the epoxy resin is
cured to provide mechanical connection and the conducting medium provides electrical connection in the z-direction.
Therefore, the cure process is critical to develop the ultimate electrical and mechanical properties of ACF. The purpose
of the present work is to investigate optimum curing conditions to achieve the best performance of ACF joints. Differential scanning calorimeter was used to measure the curing degree. Adhesion strength was evaluated by 90° peeling
test. The contact resistance has also been studied as a function of bonding temperature and curing degree. Results show
a strong dependence of curing condition on the electrical and mechanical performances. Adhesion strength increases
exponentially with the curing degree. Whereas the contact resistance decreases with the curing degree and achieve the
minimum value at 87% of curing. Co-relation of the curing degree of the ACF was also studied through the detailed
investigation of the fracture surfaces under scanning electron microscopy.
Ó 2003 Elsevier Ltd. All rights reserved.
1. Introduction
One of the most promising technologies of todayÕs
chip-level interconnection is flip-chip technology, which
has emerged as a high density, and high performance
interconnection method for integrated-circuit chips. A
potential revolutionary technology of attaching devices
by flip chip without solder bumps/balls and underfill
epoxy is to use of conductive adhesive. They have
recently gained much attention as an environmentally
friendly alternative to tin/lead (Sn/Pb) solders [1].
Conductive adhesive not only avoid the toxicity and
environmental concerns from lead and chlorofluorocar-
*
Corresponding author. Tel.: +852-2788-7130; fax: +8522788-7579.
E-mail address: [email protected] (Y.C. Chan).
bon-based flux cleaners, but also have the following
technological advantages over their counterparts: (1) the
lower curing temperature required for adhesive reduces
joint fatigue and stress cracking problems enabling the
use of heat sensitive or non-solderable materials; (2)
fewer processing steps enable an increase in production
throughput; (3) the higher flexibility and the closer match
in coefficient of thermal expansion enable a more compliant connection and minimize failures; (4) the smaller
filler particle size facilitates finer line resolution. Adhesive could also provide most of the needs for flip-chip
technology by themselves: create short electrical path,
ensure good horizontal gap insulation, reduce joint stress
and provide strong mechanical adhesion. Furthermore,
no cleaning/flux is required, secondary underfill is not
necessary, and placement of adhesives is not critical.
Therefore, more recently, conductive adhesives are
playing an increasingly important role in the design and
production of electronic packages application [2].
0026-2714/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0026-2714(03)00185-9
506
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
Two critical reliability issues of conductive adhesive
application are contact resistance shift and poor adhesion strength. Extensive studies have been done for the
former subjected to various environmental stresses [3,4]
and slightly work on the latter. High adhesion strength
is a critical parameter of fine pitch interconnection that
fragile to shocks encountered during assembly, handling
and lifetime. Initial efforts at this aspect mostly focused on isotropic conductive adhesive (ICA) for surface
mount lead attachments, which involve structure–property–performance study [5] and dropping mounted chip
and board assemblies onto hard surfaces from certain
height [6].
Anisotropic conductive adhesive (ACA) is a special
variety of ICA which conduct electricity only in one
direction. ACA contains less conductive particles (lower
percentage of metal fillers in volume) than that of ICA.
The concentration of particles is controlled in such a
way that just enough particles are present to assure reliable electrical conductivity in the z-direction (normal to
the plan of adhesive film) while concentration is far
below a critical value to achieve percolation conduction
in the x–y plane [3]. Thus it has been attracted much
interest because of ultra-fine capability [7].
Because of the anisotropic property, ACA can be deposited over the entire contact region as film (known as
anisotropic conductive film, ACF), thus it is predicted
to increase mechanical reliability [7,8]. However, there
is still some sort of uncertainty of using ACF considering proper formulation and optimized curing profile
which lead to unstable contact resistance, poor adhesion
strength etc. Due to poor adhesion with the flex, void
may nucleate at the ACF/flex interface, which might
propagate to the interconnection to loss the electrical
conduction [9]. The physical, electrical and mechanical
properties of the cured conductive adhesives depend to a
large extent on the degree of cure of the epoxy composition of the conductive adhesive [10]. Successful bonding
involves the selection of proper bonding parameter during which chemical reactions proceed to completion, in
order for it to develop its service strength [10]. In our
previous piece of research, we studied on the peel strength
of chip-on-flex using ACF and tried to improve adhesion
through plasma cleaning of the flex substrate [9]. There is
little work done elsewhere on the adhesion strength of flip
chip, especially with the curing degree of ACF matrix.
Current work focuses on the understanding of the curing
degree that affects the peel strength and contact resistance
of chip on flex (COF) in terms of fracture microstructures.
Firstly we investigated the effect of different bonding
temperature on the contact resistance of the joint. After
measuring contact resistance, we measured the peel
strength of the flip-chip bond and correlated it with the
curing degree of the ACF. The results of this study would
allow the development of ACF joints using fine pitch flip
chips on flexible substrate with more reliable product.
2. Experiments
2.1. Materials
The FCOF packages are made up of three different
materials, namely silicon chip, ACF and flexible substrate.
2.1.1. Flexible substrate
It contains polyimide film as a base materials, Cu
trace as conductor and an adhesive in between. This
adhesive is an epoxy-based polymer. The thickness of
the polyimide base film, adhesive and Cu trace is 50, 10,
and 5 lm. The thickness of Ni coating and Au flash on
Cu trace was about 1–2 and 0.1–0.4 lm, respectively.
2.1.2. Silicon chip
The dimension of the test chip have a size of 11 3
mm2 , with 46 46 lm2 square shaped opening around
the periphery. These openings are of aluminum metal
with 1% silicon limited by chip-passivation layer. There
was not any bumping on the Al metal pad. Thus this
kind of chip is known as bumpless chip. There are 368
Al pads in the chip. Among them, 300 pads run parallel
to the length of the chip for electrical connection. These
Al pads are arranged in sets of five as a group; with two
adjacent bumps for measuring insulation resistance and
three for contact resistance. There are a total of 60 sets
of these daisy-chained bumps within the chip––30 sets in
each side.
2.1.3. Anisotropic conductive adhesives film
The type of ACF used in this study was consists of an
epoxy layer and filled with conductive particles. The
conductive particles are made up of polymers plated
with a thin layer of nickel and gold followed by a thin
insulation layer. The thickness of ACF is 35 lm and
particle diameter is 3.5 lm. Concentration of the conductive particles is about 3.5 million/mm3 . The glass
transition temperature (Tg ) of the ACF is 130 °C.
2.2. Bonding process
The procedures for flip-chip mounting included several steps: ACF placement, pre-bonding, IC placement
and final bonding. A manual flip-chip bonding machine
(Karl Suss 9493 Mauren) was used to carry out the prebonding, i.e. placement of ACF on flex. The pre-bonding pressure was 1 MPa, while the temperature and time
were 100 °C and 7 s respectively. A semi-automatic flipchip bonding machine (Toray SA2000) was used to
conduct the final bonding. The substrate pattern and the
position of the chip bumps were aligned automatically
by the flip-chip bonder. Finally, the chip is bonded to
the substrate by applying heat and pressure simultaneously. For final bonding, the pressure and time were
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
100 MPa and 10 s respectively, while the bonding temperature was varied from 150–230 °C. The alignment
accuracy is 2 lm. Fig. 1(a) and (b) shows the schematic
507
of the bonding process and the appearance of the assembly after bonding.
2.3. Contact resistance measurement
The contact resistance of FCOF assemblies was
measured by using the four-point probe method and the
schematic circuitry is shown in Fig. 2. The circuitry has
been improved comparing with the circuitry illustrated
in Ref. [11] and there was no need to measure the trace
resistance before bonding. In the test, 1 mA constant
DC current was applied to the circuit and the voltage
was read from the HEWLETT PACKARD multimeter.
Then the contact resistance was obtained simply by
using OhmÕs Law, R ¼ V =I.
2.4. Adhesion measurements
The 90° peel test has been widely used to measure the
adhesion strength of multilayer films. To measure the
adhesion strength of the ACF, flexible substrate were cut
from the side of the chip to make it to 3.5 mm width and
then peeled out from the bonded sample keeping the
chip fixed. Special accessory was prepared for smallscale peel strength measurement. Fig. 3(a) and (b) shows
the schematic arrangement and prepared sample respectively for measuring the peel strength. The test was
carried out by the Instron Model Mini 44 Tester, (Intertek testing Services, England) with a cross-head speed
of 10 mm/min. Instrument was adjusted deliberately in
order to peel out the flex from the whole chip. Five
samples were tested per given condition.
2.5. Failure analysis
After the peel test, the fracture surfaces of the samples (both the chip side and substrate side) were examined by optical microscopy and scanning electron
microscopy (SEM) to study the fractography characteristics of both chip and substrate sides. 30° tilting of
the stage was used to get the perspective view of the
surface morphology.
2.6. Differential scanning calorimeter tests
Fig. 1. (a) Schematic of the bonding process, (b) appearance of
the assembly after bonding.
Curing degree of ACF was examined with a Modulated differential scanning calorimeter (DSC) instrument
(by TA Instruments, Model 2910). It measures both the
heat capacity of endothermic and exothermic transitions
of a sample and provides quantitative information regarding the enthalpic changes. Dynamic curing experiments were performed from 50–250 °C with ramp rates
of 10 °C per min. In order to obtain accurate heat flow
or experimental data, a base line for each experiment
conditions was recorded by using empty sample pans
before actually conducting the cure reaction in the
508
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
Fig. 2. Schematic circuitry to measure the contact resistance using the four-point probe method.
amount of adhesive was taken form the fractured surface after the peel test and prepared the samples for DSC
tests. The reaction was considered complete when the
rate curve levels off to the base line.
3. Results
3.1. Curing degree
The DSC has been widely utilized to elucidate the key
cure process parameters such as the extent and rate of
chemical conversion. It quantifies both the heat capacity
of endothermic and exothermic transitions of a sample
and provides quantitative information regarding the
enthalpy changes. The method assumes that for a cure
process the measured heat flow is proportional to the
conversion rate. This assumption is valid only for
chemical reaction and no other enthalpic events, such as
evaporation, enthalpy relaxation, or significant changes
in heat capacity with conversion. In practice however, it
has proven to be a reasonably good assumption. A
commonality of all thermosetting systems is the liberation of heat accompanying cure [10],
DHE
Reactants ! Products;
Fig. 3. (a) Schematic arrangement for measuring the peel
strength, (b) prepared sample for peel strength measurement.
sample pan. To avoid temperature gradients small
sample quantities (<10 mg) were placed in aluminum
hermetic pans and were run on MDSC. In this way, two
kind of samples (1) the raw ACF and (2) cured ACF
after peel out the bonding were examined. For raw
ACF, it was removed from the refrigerator and allowed
to warm up to room temperature for one hour before
conducting the experiments. For cured ACF, small
where DHE is the exothermic heat, expressed as heat per
mole of reacting groups (kcal mol1 or kJ mol1 ) or per
mass of materials (cal g1 or J g1 ). In this case, during
the curing of ACF, the exothermic reaction of epoxy
with primary and secondary amines was occurred [10].
Fig. 4 shows the typical DSC traces of ACF for both
uncured and cured at different temperature. At the initial
stage, the reaction rate increases with time and temperature due to the lower viscosity. However, after reaching
a certain degree of curing, the reaction rate goes down
due to the increased viscosity. Viscosity increases in
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
5
509
1.5
Peel strength (N/mm)
Heat Flow-Endo up (W/g)
4
3
Uncured
Cured at
Cured at
Cured at
Cured at
Cured at
2
1
150˚C
170˚C
190˚C
210˚C
230˚C
0
50
100
150
200
250
Temperature (˚C)
0.5
0
130
150
170
190
210
230
250
Bonding temperature (˚C)
Fig. 4. Dynamic DSC scans for uncured and different temperature bonded ACF.
proportion with the cross-linking and hence reduces the
mobility of the polymer molecules for further reaction.
As a result, both cured and uncured ACF showed a
broad exothermic cure peak at a temperature range from
80 to 210 °C. They all had the same peak temperature,
approximately 130 °C. Based upon the assumption that
the exothermic heat evolved during the cure is directly
proportional to the extent of cure, the curing degree a
can be expressed by the relationship between the total
exothermic heat of raw material QT and the residual or
exothermic heat of previously cured material QR . The
relationship can be written as follows [10]:
a¼
1
QT QR
QT
Fig. 6. Peel strength of different bonding temperature.
sult shows that the ACF was only 24% cured, when the
bonding temperature was 150 °C. In contrast when the
bonding temperature was 230 °C, the ACF becomes 98%
cured. During bonding process, temperature was maintained to increase the fluidity of ACF and then to
ð1Þ
QT & QR were measured from dynamic DSC scan of the
sample of uncured and cured at different temperature
respectively. Fig. 5 shows the curing degree of ACF for
different bonding condition. It is obvious that the curing
degree depends upon the bonding temperature. The re-
100
Curing Degree (%)
80
60
40
20
0
130
150
170
190
210
230
250
Bonding Temperature (˚C)
Fig. 5. Curing degree at different bonding temperature.
Fig. 7. Fracture surfaces. (a) Flex side and (b) chip side, after
the peel strength test for bonding temperature of 150 °C.
510
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
Fig. 8. Fracture surfaces. (a) Flex side and (b) chip side, after
the peel strength test for bonding temperature of 210 °C.
Fig. 9. Fracture surfaces. (a) Flex side and (b) chip side, after
the peel strength test for bonding temperature of 230 °C.
quickly cure when maintaining constant pressure on the
chip [11]. In this thermal curing process of the epoxybased adhesive, higher temperature initiates and accelerates the cross-linking reaction by providing higher
active energy. Therefore, the curing degree of ACF also
increases with the increase of curing temperature.
debonding occurred at the flex/ACF layer interface. At
the flex side only some uncured monomer of adhesive
was observed. ACF was found on the chip side. It could
be said that the adhesion between ACF and flex is
5
In this work, peel strength were measured for studying the adhesion of the COF packages. The results of
the peel strength measurement are plotted in Fig. 6. The
curve shows that the peel strength increases with the
bonding temperature. As bonding temperature increases, the curing degree of this epoxy-based adhesive
increases resulting in stronger chemical bonding at the
interface and leading to increase in adhesion. For understanding the fracture mode, it is essential to study the
fracture surface along with the peel test data. Fig. 7
shows the fracture surfaces of the substrate side and chip
side after the peel strength test for the sample cured
at 150 °C. Fractured surface at both flex and chip sides
are very smooth. It is clear from the SEM images that
Contact Resistance (Ohm)
3.2. Adhesion study
4
3
2
1
0
130
150
170
190
210
230
250
Bonding Temperature (ºC)
Fig. 10. Contact resistance for different bonding temperature.
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
3.3. Contact resistance
The contact resistance of FCOF assemblies was
measured by using the four-point probe method. Fig. 10
shows the variation of the contact resistance with the
bonding temperature for bumpless FCOF. From the
curve, it is clear that the contact resistance of the
packages decreases with the increase of bonding temperature up to 210 °C and then increases when the
bonding temperature was used above 210 °C. FCOF
packages assembled at 210 °C showed the best contact
resistance when compared to those assembled at other
different temperatures. For the bumpless FCOF packages assembled at 210 °C, the average contact resistance
was 0.12 X.
4. Discussion
4.1. Adhesion strength
The curing degree of ACF plays an important role in
determining the reliability of the ACF joints. As the
curing of ACF proceeds, the linear polymer chain in the
epoxy resin grows and branches to form cross-links [12].
At the lower degree of cross-linking, the polymer chains
are capable of moving relatively easily and exhibit less
adhesion strength. Again, as the bonding temperature
increases the degree of cross-linking as well as adhesion
also increases. Because, the polymer chains locked together and their movement become consequently
somewhat restricted. Cross-linked polymer chains are
chemically bound together to give a three-dimensional
1.5
Peel Strength (N/mm)
weaker than that of the ACF/chip. This is because, ACF
at chip side become more cured than ACF at flex side,
as heat is applied from chip side. It also indicates that
this curing temperature is not sufficient and needs to be
set at a temperature higher than 150 °C for optimum
bonding of ACF for reliable flip-chip bonding.
However, the failure mode changes at high-temperature bonding. Fig. 8 depicts the failed surface of the
interfaces when the bonding temperature was 210 °C. In
the flex side, some ACF materials are visible. At least,
ACF adhered well with some portion of the flex. Interdiffusion of the ACF to the flex is clear at the periphery
of the ACF patches. Fig. 9 depicts the failed surface of
the interfaces when the bonding temperature was 230
°C. The ACF was found in both sides and the fractured
surfaces are very rough. Extensive deformation and
smearing of the ACF was noticed for such high-temperature bonding. It is also interesting to get the conductive particles in the both sides. Thus it is confirmed
that fracture occurred through the ACF. It reveals that
the adhesion between flexible substrate and ACF improved substantially.
511
1
0.5
0
0
20
40
60
80
100
Curing degree (%)
Fig. 11. Dependence of the peel strength of the joints on the
curing degree.
‘‘chicken wire’’ structure [13]. So, the higher the curing
degree, the stronger the chemical bonding and the better
adhesion strength at the ACF interface. High temperature also promotes the physical adsorption and diffusion
between adhesive and adherent. Moreover, high temperature is beneficial to wetting and flowing of the ACF
due to lowering its surface energy. As a result peel
strength increases significantly with the increase of
bonding temperature (Fig. 6).
Fig. 11 shows the dependence of the peel strength of
the joints on the curing degree of ACF. It was observed
that peel strength increases at first slowly with curing
degree and after 75% curing, it starts to increase quickly.
This is a very common phenomenon reported elsewhere
that partially cured polymer shows weaker strength and
later it increases with the degree of curing [14]. However,
in this experiment, after 87% curing of ACF (i.e. bonded
at 210 °C), peel strength increases at an unusual manner
than the typical polymer. It implies that other mechanism is also contributing to increase the strength.
Fracture mode for these higher cured samples is also
different than that of the expectation. Fracture surface,
either at flex or chip side, shows very rough morphology
with the ACF material in both sides (Fig. 9). Fracture
occurred through the ACF, i.e. adhesion strength of
ACF to flex substrate is higher than the intrinsic
strength of the ACF matrix. It is really surprising why
highly cross-linked ACF at higher curing degree failed
through the ACF matrix. Some strong reactions must
occur between ACF and the substrate, which is responsible for such contradictory fracture mode. However, the fracture mode study of the sample bonded at
210 °C revealed that in some area of flex, ACF patches
adhered well with the flex. High magnification study at
the interface between the ACF patches and flex prove
some inter-diffusion and/or reaction between ACF
epoxy and the flex (Fig. 12). It is anticipated that these
reactions between flex substrate to the ACF started at
210 °C. The possible reaction mechanism is described
below.
512
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
However, the Tg (glass transition temperature) of this
adhesive on the PI is comparatively low (80 °C), which
might be responsible to the interfacial reaction of the
ACF at higher temperature. During the bonding process at a temperature of 230 °C, substantial heat passes
through the ACF to flex substrate. Therefore, the temperature of this laminated adhesive reaches far more than
its Tg . As a result, it melts, inter-diffused and/or reacts with
the ACF. After cooling, it re-solidifies and strengthened
[17] the bond between ACF to flex substrate. Thus peel
strength increases extensively for this high temperature
bonded sample. From this evidence, it is proposed that at
higher temperature the adhesion strength increases because of two folds reaction phenomenon:
Fig. 12. ACF patches and flex express some inter-diffusion and/
or reaction.
In a traditional polyimide flex substrate, an adhesive is
used to bond copper foil to a flexible substrate and this
adhesive becomes cured during the Cu lamination process [15,16]. The most important feature of the flexible
substrate used in this experiment, is that this cured adhesive remains intake and adhered with the polyimide
substrate even after the development of the Cu circuitry.
That means, the base polyimide film is covered by a thin
adhesive layer, which is now acting as a barrier layer
between the PI and the ACF. Fig. 13 reveals the existence
of such adhesive underneath the Cu traces of the substrate after the peel test. So, ACF material would virtually be bonded with this residue adhesive in the substrate
side [9]. As this adhesive is already cured during Cu
lamination and passes through some other process (such
as Cu etching, dry film stripping, rinsing etc.) during Cu
circuitry development, it no longer has any affinity to
form firm bonding with the ACF. Moreover, the surface
morphology of this adhesive is very smooth. Quite naturally, it renders mechanical interlocking between the
flex and ACF. Therefore, lower adhesion strength was
found at lower bonding temperature.
Fig. 13. Existence of adhesive underneath the Cu traces.
1. Higher cross-linking of the ACF.
2. Interfacial diffusion and/or reaction between the ACF
and the previous thin adhesive.
Thus, it is concluded that adhesion strength increases
at an unusual manner when we are considering only
curing degree of the ACF. However, the inter-diffusion
and/or reaction are contributing much more than the
cross-linking reaction at 230 °C.
4.2. Contact resistance
Contact resistance of ACF joints mainly depends on
followings [18]
1. The number of conductive particles trapped between
bumps and pads,
2. the successful removal of insulation layer and
3. contact area of deformed particles.
The first one depends on the rate of curing of ACF
and temperature of the bonding process. Third point, i.e.
deformation of the particles depends on the applied
pressure during bonding. Insulation layer remove during
the early stage of the bonding time. When pressure is just
applied, entrapped particles moves between the pads, by
the way due to friction, insulation layer wear out and
conductive metal coating come in contact with the pads.
During the early stage of the bonding process, the
ACF becomes soft and rubbery and this transformation
allows the conductive particles to move within the ACF.
When the curing process is completed, the ACF becomes
hardened and the mobility of the conductive particles is
lost. Higher bonding temperature results in higher curing
degree shortly, and the ACF becomes stiffer with higher
modulus. After the bonding process, the conductive particles only recoil a little for highly cured ACF. Thus the
higher contact area of the deformed particles remains stable to the pads and the contact resistance is relatively low.
At the curing temperature of 150 °C, the ACF was
mostly uncured and the conductive particles carried
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
No. of early failure joint
10
8
6
4
2
0
130
150
170
190
210
230
250
Bonding Temperature (ºC)
Fig. 14. Number of early failure joints for different bonding
temperature.
within the epoxy layer are able to move. Therefore, there
is a possibility of the conductive particles to squeeze out
form the pads interface due to the applied pressure
through the soft adhesive. As a result, in some joints,
higher and unstable contact resistance and even sometime open joints were found. Fig. 14 shows the number
of early failure or open joint with the bonding temperature among 60 joints.
At the bonding temperature of 210 °C, the ACF becomes cured adequately (87%) and the conductive particles within the epoxy layer remains stable throughout
the ACF joints. A reliable interconnects should have
metal-to-metal contact through out the die pad-to-the
deformed particles and deformed particles-to-the Cu
pad of the substrate without any oxide layer [19]. 210 °C
is optimum in respect of the contact area requirement
having less possibility of the oxidation than the bonding
temperature of 230 °C. Therefore, this bonding temperature creating the best contact and show the low
contact resistance. However, when the bonding temperature was 230 °C, the ACF was cured much (98%)
but the contact resistance becomes high. This may be
due to the heat being absorbed by the FCOF packages
and acting as a catalyst for oxidation of the Al pads.
Aluminium atoms easily contribute its electrons to
oxygen atoms to form aluminium oxide film. The
aluminium oxide layer acts as a barrier between the Al
pad–conductive particle and conductive particle–Cu
pad, making difficulties for the electrons to flow through
this Al pad–conductive particle–Cu pad path [11]. That
leads to reduction in number of free electrons on its
surface, hence, increases its electrical resistance.
5. Conclusions
A systematic study was carried out to clarify the effects of the bonding temperature on the cross linking of
the ACF matrix which in turn effects the performance of
513
the interconnection. The results show that the high
temperature cured samples have higher reaction rates
and a greater curing degree as well as high adhesion
strength compared with low temperature cured samples.
Contact resistance was also found to be strongly dependent on the curing degree.
At temperatures of 210 °C, the curing degree was
about 87% with lesser extent of oxide formation on Al
pad and moderate adhesion strength. In contrast, if the
bonding temperature is 230 °C, 98% curing of the ACF
is achieved with higher extent of oxide formation on Al
pad and highest adhesion strength. Hence, the ACF
joints of the FCOF packages assembled at 210 °C performed better with lower contact resistance values when
compared to those assembled at other different temperatures. It is therefore concluded that the performance of
the ACF interconnects is greatly influenced by the
bonding temperature as well as curing degree during the
assembly of FCOF packages. In this study, the optimum
temperature for bonding FCOF with ACF was concluded to be at 210 °C. Fracture modes were studied by
SEM after the adhesion strength measurement. From
the fracture surface images, it was found that epoxy of
the ACF matrix start to react and inter-diffuse with the
thin adhesive layer on the polyimide substrate at about
210 °C. After that reaction or inter-diffusion, fracture
mode changes from flex-to-ACF to the ACF itself. This
reaction or inter-diffusion of the ACF is proposed to be
responsible for higher adhesion strength at higher
bonding temperature.
Acknowledgements
The authors would like to acknowledge the financial
support of the Strategic Research Grant of City University of Hong Kong (project no. 70001387, Reliability
Study of Anisotropic Conductive Joints––Effect of
Thermally Induced Warpage) and the Hong Kong SAR
Government––Innovation and Technology Fund (ITF),
University–Industry Collaboration Programme (project
no. UIT/22).
References
[1] Xu S. Evaluating Thermal and Mechanical Properties of
Electrically Conductive Adhesives for Electronic Applications. PhD Dissertation. Engineering Science and Mechanics Department. Virginia Polytechnic Institute and State
University, 2002. p. 168–9.
[2] Kim HK, Shi FG. Electrical reliability of electrically
conductive adhesive joints: dependence on curing condition and current density. Microelectron J 2001;32(4):315–
21.
[3] Dudek R et al. Reliability investigations on conductive
adhesive joints with emphasis on the mechanics of the
514
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
M.A. Uddin et al. / Microelectronics Reliability 44 (2004) 505–514
conduction mechanism. IEEE Trans––CPMT-A 2000;
23(3):462–9.
Liu J et al. Surface characteristics, reliability and failure
mechanisms of tin, copper and gold metallizations. IEEE
Trans––CPMT-A 1997;20(1):21–30.
Vona SA et al. Surface mount conductive adhesives with
superior impact resistance. Proc Fourth Advanced Packaging Materials Conference, Chateau Elan, Braselton,
Georgia, 1998. p. 261–7.
Zwolinski M et al. Electrically conductive adhesives for
surface mount solder replacement. IEEE Trans––CPMT-C
1996;19(4):241–50.
Fu Y, Willander M, Liu J. Statistic of electric conductance
through anisotropically conductive adhesive. IEEE Trans
Comp Packag Technol 2000;24(2):250–5.
Liu J et al. Overview of conductive adhesive joining
technology in electronic packaging applications. Proceedings of the 3rd Adhesive Joining and Coating Technology
in Electronics Manufacturing Conference, Binghamton,
NY, September 1998. p. 1–18.
Uddin MA, Alam MO, Chan YC, Chan HP. Plasma
cleaning of the flex substrate for flip chip bonding with
anisotropic conductive adhesive film (ACF). J Electron
Mater, in press.
Chan YC, Uddin MA, Alam MO, Chan HP. Curing
kinetics of anisotropic conductive adhesive film. J Electron
Mater 2003;32(3):131–6.
Chan YC, Luk DY. Effects of bonding parameters on the
reliability performance of anisotropic conductive adhesive
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
interconnects for flip-chip-on-flex packages assembly I.
Different bonding temperature. Microelectron Reliab 2002;
42(8):1185–94.
Connell G, Zenner RLD, Gerber JA. Conductive adhesive flip-chip bonding for bumped and unbumped die. In:
Proceedings of the 47th Electronic Components and
Technology Conference, 1997. p. 274–8.
Yarlagadda PKDV, Stoynov LA, Kim Ill-Soo. An investigation into the performance of solar cured adhesive
bonded joints. J Mater Process Technol 2001;113(1–3):
160–6.
Clayton AM. Epoxy resins: chemistry and technology, 2nd
ed. New York: M. Dekker; 1988. p. 603–718.
Yun HK, Cho K, Kim JK, Park CE, Sim SM, Oh SY, Park
JM. Effects of plasma treatment of polyimide on the
adhesion strength of epoxy resin/polyimide joints. J Adhes
Sci Technol 1997;11(1):95–104.
Ghosh MK, Mittal KL. Polyimides: Fundamental and
application. Marcel Dekker Inc; 1996. p. 430–3.
Stearns TH. Flexible printed circuitry. In: Electronic packaging and interconnection series. New York: McGraw-Hill;
1996. p. 86–91.
Chuks NO, Samjid HM, David CW, David JW. Conduction mechanism in anisotropic conducting adhesive assembly. IEEE Trans Comp, Packaging, Manufact Technol––
Part A 1998;21(2):235–42.
Yim MJ, Paik KW. The contact resistance and reliability
of anisotropically conductive film (ACF). Adv Packag,
IEEE Trans Adv Packag 1999;22(2):166–73.