You are on page 1of 8

Microelectronics Reliability 52 (2012) 595602

Contents lists available at SciVerse ScienceDirect

Microelectronics Reliability
journal homepage: www.elsevier.com/locate/microrel

The effects of functionalized graphene nanosheets on the thermal and


mechanical properties of epoxy composites for anisotropic conductive adhesives
(ACAs)
Jiwon Kim a, Byung-seung Yim b, Jong-min Kim b, Jooheon Kim a,
a
b

School of Chemical Engineering & Material Science, Chung-Ang University, Seoul 156-756, Republic of Korea
School of Mechanical Engineering, Chung-Ang University, Seoul 156-756, Republic of Korea

a r t i c l e

i n f o

Article history:
Received 11 July 2011
Received in revised form 11 October 2011
Accepted 1 November 2011
Available online 25 November 2011

a b s t r a c t
Functionalized graphene/epoxy composites were prepared using the epoxy resin diglycidyl ether of
bisphenol A. Graphene oxide (GO) and Al(OH)3-coated graphene (Al-GO) llers were fabricated using
the Hummers method and a simple solgel method, with aluminum isopropoxide as the aluminum precursor. X-ray photoelectron spectroscopy veried the successful formation of functional groups onto the
GO and Al-GO. The dispersion of functionalized graphene llers showed an even distribution within the
epoxy resins. A dynamic mechanical analysis was used to investigate the changes in the mechanical properties of the epoxy composites, which included neat epoxy and epoxy with various concentrations of
graphene-based llers. The storage modulus and tan d graphs illustrate the enhancement achieved by
increasing the amount of ller. The composite with 3 wt.% GO had the highest storage modulus and glass
transition temperature. The thermal conductivities of the composites with graphene-based llers were
enhanced compared to those without llers. The 3 wt.% GO/epoxy composite had the highest thermal
conductivity, which was nearly twice that of the neat epoxy resin.
2011 Elsevier Ltd. All rights reserved.

1. Introduction
Microelectronic requirements are motivating the development
of smaller, higher density, and lower cost solutions. Polymer-based
anisotropic conductive adhesives (ACAs) have drawn attention as
an environmentally friendly solution for lead-free interconnects
due to increased environmental awareness. In addition to their
environmental advantages, these materials also require few processing steps and have mild processing conditions and ne pitch
interconnect capability due to the availability of small-sized conductive llers [15]. However, the application of polymer-based
ACAs in electronic devices could be limited by their low thermal
conductivities because of the increasing need for rapid heat dissipation [68]. Thus, in the quest to fabricate electronic materials
that easily dissipate heat, polymers with thermally conductive llers have emerged as a cost-effective way to address thermal management issues.
Epoxy polymers have been widely used in engineering components because of their outstanding mechanical and thermal properties and processability. In ACAs, the use of epoxy resins has
been the state of the art because of their benecial properties such
as low shrinkage, high adhesion and resistance to thermal and
mechanical shocks. Furthermore, they also have good resistance
Corresponding author. Tel.: +82 2 820 5763; fax: +82 2 824 3495.
E-mail address: jooheonkim@cau.ac.kr (J. Kim).
0026-2714/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.microrel.2011.11.002

to moisture, solvents and chemical attacks [9,10]. However, the


thermal conductivity of the epoxy matrix is very low and is therefore not suitable for the thermal design loads in ACAs.
In recent years, nanollers employing carbon-based reinforcement materials have been dominated by carbon nanotubes (CNTs)
which produce high performance composites with enhanced properties [1114]. However, the use of CNTs in composites has been
limited by the high cost, as well as challenges in processing and dispersion. Compared to CNTs, graphene continues to attract considerable attention because of its outstanding properties. Graphene, a
two-dimensional material, shows promise as a nanoller material
in polymer composites due to its extremely high aspect ratio, unique graphitized planar structure, and low manufacturing cost. Theoretical and experimental studies of individual graphene sheets
show they may possess a high Youngs modulus (1 TPa), high thermal conductivity (5000 W/m K) and large surface area (2600 m2/
g) [15,16]. However, in spite of their desirable properties, the use of
graphene-based materials has been restricted to heat management
applications because of graphenes great electrical conductivity
(6000 S/cm) [17], which can result in malfunctions at electrode
junctions, such as short circuiting in ultra ne pitch packages
[18]. Therefore, a graphene-based composite with insulating materials has the potential to be a usable material for heat management.
In this study, chemically modied graphene nanosheets were
prepared to improve the thermal conductivity and electrical
insulation properties of epoxy resins for ACAs. Two types of

596

J. Kim et al. / Microelectronics Reliability 52 (2012) 595602

functionalized graphene sheets were fabricated: graphene oxide


and Al(OH)3-coated graphene. These were embedded in bisphenol
A resin, a widely used epoxide adhesive. The inuence of functionalized graphene sheets was evaluated in terms of the increase in
the mechanical and thermal conductive properties of epoxy composite systems. Furthermore, the functionalized graphene/epoxy
composites were analyzed to conrm the effects of functionalized
graphene on the microstructure and interconnections in electronic
devices that use low melting point alloy (LMPA) solder as a conductive material between the quad at package (QFP) leads and
the substrate.
2. Experimental
2.1. Materials
Graphite with an average particle size of less than 20 lm was
obtained from SigmaAldrich. Sodium nitrate (NaNO3), potassium
permanganate (KMnO4) and aluminum isopropoxide (AIP, Al(O-iPr)3) were purchased from SigmaAldrich. Sulfuric acid (H2SO4,
95.0%), hydrochloric acid (HCl, 35.037.0%), hydrogen peroxide
(H2O2, 34.5%), methanol (CH3OH, 99.5%), and acetone (99.5%) were
purchased from Samchun Chemical.
The composite in this study consisted of a polymer binder, a curing agent, an LMPA solder, and other minor organic addictives. To
obtain the base polymer binders, diglycidyl ether of bisphenol A
(DGEBA; Kukdo Chemical; epoxy equivalent weight (E.E.W.) =
186.8 g/eq) was dried in a vacuum oven for 2 h to remove the residual solvent. 4,40 -Diaminodiphenylmethane (DDM), prepared by TCI
Korea, was used as a curing agent. The LMPA solder ball (Sn-58Bi;
Tm = 412 K (139 C); 45 lm) was obtained from Senju Industrial
Company. 3-Butanoic acid (Aldrich, 97%) was used as a reductant
to remove the oxide layer from the surfaces of the LMPA and the
conductive pad.
2.2. Functionalization of graphene sheets
Fig. 1 shows the synthetic procedure [19,20]. Graphene oxide
(GO) was synthesized using the modied Hummers method [19].
In a typical synthesis, 1 g of graphite powder is added to a mixture
of 46 mL concentrated sulfuric acid and 1 g of NaNO3 and stirred in
an ice bath for 4 h. After the homogeneous dispersion of the graphite mixture, KMnO4 (6 g) was slowly added to the solution over 1 h
at 273 K (0 C) in an ice bath. The mixture was then stirred for 6 h
at 308 K (35 C) in an oil bath. After the reaction, the mixture was
distilled with 92 mL de-ionized water (D.I. water) and heated at
368 K (95 C). After 20 min, the mixture was poured into 200 mL
of D.I. water with 20% H2O2. The mixture was washed and puried
with HCl and D.I. water several times and dried in a vacuum oven
at 323 K (50 C). The collected graphite oxide powder was

dispersed in D.I. water using ultrasonication, leading to dispersed


GO sheets in solution, and then dried in a vacuum oven at 323 K
(50 C) for 48 h.
The hydroxyl groups on the GO sheets provide the hydrogen for
bonding via the hydrolysis of AIP [20]. In order to form hydrogen
bonds between the hydroxide groups of GO and the hydroxide
groups of the precursor, AIP (0.5 g) was mixed with 500 mL of an
H2O/methanol mixture (4:1) at 363 K (90 C) for 30 min. The reactions led to the successful hydrolysis of AIP, which contains hydroxyl groups. A suspension of graphene cake in a methanol and D.I.
water mixture was then added to the AIP solution. Hydrolysis
and condensation of the mixture were carried out at 333 K
(60 C) for 6 h, and the liquid mixture in the solution was then
evaporated using a rotary evaporator and washed with D.I. water
to remove the residual and unreacted AIP. At the end of the procedure, the nal products were collected via ltration and dried under vacuum at 373 K (100 C) for 12 h. AIP can form Al(OH)3 via
hydrolysis and can be easily converted to Al2O3 through heat
treatment.
2.3. Preparation of graphene/epoxy composites
GO and aluminum covered graphene (Al-GO) were dispersed
separately in acetone and stirred for 6 h; this dispersion was then
mixed with DGEBA to give concentrations of 1, 3, or 5 wt.% GO or
Al-GO as compared to the weight of the epoxy resin. Above
5 wt.%, the mixture was viscous and difcult to process due to
the high surface energies of GO and Al-GO. The mixture was sonicated for 30 min and stirred for 12 h for homogenization. In order
to evaporate the acetone, the mixture was placed in a convection
oven at 343 K (70 C) for 6 h, and the residual solvent was eliminated in a vacuum oven at room temperature for 12 h. Epoxy resins
were mixed with stoichiometric amounts of curing agent (DGEBA:DDM = 1:0.4) at 373 K (100 C) for 15 min. Then, 3-butenoic
acid was added to the resins to eliminate the oxide layer on LMPA
solder that is used to create a conductive path in the hybrid interconnection test. The organic carboxylic acid can realize good coalescence and wetting of the LMPA solders. Bubbles in the mixture
were removed by placing the mixture in a vacuum oven for
30 min at room temperature.
2.4. Measurements
2.4.1. X-ray photoelectron spectroscopy
Surface functionalizations of GO and Al-GO were qualitatively
analyzed using X-ray photoelectron spectroscopy (XPS). XPS was
performed using a VG-Microtech ESCA2000 spectrometer equipped
with a hemispherical electron analyzer and an Mg Ka (hm = 1.25
36 keV) X-ray source.

Fig. 1. Preparation of modied Al(OH)3-coated GO composite.

J. Kim et al. / Microelectronics Reliability 52 (2012) 595602

597

2.4.2. Dynamic mechanical analysis


The dynamic mechanical analyses (DMAs) of both pure epoxy and
functionalized graphene/epoxy composites were performed on a
model TTDMA dynamic mechanical analyzer (Triton Technology,
UK) to determine their thermomechanical properties, such as storage modulus E0 , loss modulus E00 , damping factor (tan d) and glass
transition temperature (Tg). The experiments were carried out on cubic samples (20 mm  10 mm  3 mm) in the single cantilever
mode. A frequency of 1 Hz (corresponding to a strain rate of 0.05%
per second) with a temperature ramp of 5 K/min and a scanning temperature range from 298 K (25 C) to 460 K (187 C) were employed.
The glass transition temperature, Tg, was determined from the peak
of the tan d curve. At least three tests were carried out for each case.
2.4.3. Laser ash apparatus
Thermal conductivity measurements were determined using a
Netzsch 457 laser ash analysis (LFA) instrument. All measurements were taken at room temperature (298 K, 25 C) with a laser
voltage power of 1538 V and a laser transmission lter of 100%.
Samples were cut into 10  10  1 mm bars. A total of 510 shots
were taken per sample set.
2.4.4. Field emission scanning electron microscopy
The morphologies of the functionalized graphene/LMPA/epoxy
composites were investigated using a eld emission scanning electron microscope (FE-SEM, SIGMA, Carl Zeiss) at a 3.0 kV accelerating voltage. All the samples were cut using liquid nitrogen prior to
coating with platinum (Pt). The 5 wt.% graphene ller load level
was chosen as the representative loading level at which to compare the microstructures of GO and Al-GO in LMPA/epoxy
composites.
2.4.5. Hybrid interconnection test
The adhesive joining process of the ACAs was the same as reported elsewhere [2123]. The morphology of the conductive path
and the mechanical strength analysis were evaluated through the
hybrid interconnection test. The size of the QFP was
14  14  2.7 mm3, and it had a 1.0 mm lead pitch. The printed circuit board (PCB) was 32  32  1.0 mm3 in size, and the PCB pattern
was plated with 18 lm-thick Cu. The substrate test board was
cleaned with acetone for 1 min and then washed with D.I. water
and dried using an air jet. LMPA/graphene/epoxy composites were
prepared with 1 wt.% of functionalized graphene and a volume fraction of LMPA solders of 0.4. The composites were selectively applied
to the QFP lead. After completion of the QFP and substrate alignment, the QFP was mounted onto the PCB electrode pads using a ip
chip bonder (LAMBDA: FINETECH Co.). Then, the test assembly was
heated according to the temperature prole determined from the
DSC analysis test. After the adhesive bonding was complete, the
morphologies of the conduction paths formed between the QFP
leads and the substrate were observed using an optical microscope.
A 45 pull test (JIS Z 3198-6) was performed on the QFP lead to assess
the mechanical properties of the hybrid interconnections. Specically, the QFP lead was pulled upward at a speed of 6 mm/min, with
a total of 22 pulls tested for each composite.
3. Results and discussion
3.1. Characterization of graphene sheets
XPS analyses of GO and Al-GO were performed to determine
their chemical structures and the results are presented in Figs. 2
and 3. XPS can quantify the types of atoms in the samples and
can also identify the types of chemical bonds. Wide scan spectra
in the range 01100 eV identify the surface elements present and

Fig. 2. X-ray photoelectron spectra survey scans of graphene: (a) graphite, (b)
graphene oxide and (c) Al(OH)3-fuctionalized graphene.

perform a quantitative analysis, as shown in Fig. 2. In the spectra


of graphite specimens, GO and Al-GO, C 1s and O 1s signals appeared at 284 eV and 532 eV. The Al 2p signal appeared at 75 eV
in the spectrum of Al-GO in Fig. 2c. As shown in Table 1, the initial
oxygen concentration on graphite was only small, 2.71%, but after
oxidation, it drastically increased to 64.49% along with a reduction
in carbon concentration. This phenomenon was mainly attributed
to the generation of oxygenated functional groups, such as epoxide,
hydroxyl, carbonyl and carboxyl groups [24]. Furthermore, the
atomic concentrations of C 1s and O 1s were 35.51% and 64.49%
in GO but were 30.08% and 66.50% in Al-GO. After the reaction with
AIP, the ratio of O:C gradually increased with the appearance of a
low peak of Al 2p at a concentration of 3.42% in Al-GO. This result
indicates that a functional group containing oxygen atoms is introduced onto the surface of Al-GO after the reaction of GO with AIP,
indicating a newly created species containing Al atoms.
The detailed chemical bonding of the fabricated graphene sheets
in this study was conrmed through the high resolution spectra of C
1s and O 1s based on Gaussian spectral deconvolution. In Fig. 3,
these peaks provide clear evidence that GO and Al-GO were chemically modied. Carbonyl, epoxy/hydroxyl and carboxylate signatures were detected in the XPS spectrum of the GO powder using
a narrow C 1s scan, as shown in Fig. 3b. Small and large peaks were
found at 284.5 eV (C@C/CAC), 286.5 eV (CAO), 287.9 eV (C@O), and
289.3 eV (O@CAO) [25], suggesting successful manufacture of GO
layers according to the Hummers method. The C 1s spectrum for
Al-GO (Fig. 3c) is a similar spectrum to that of GO; however, a
new peak was observed at 283.6 eV (CAOAAl), conrming the presence of an aluminum-based material on Al-GO [26]. For GO, the XPS
peaks of O 1s in Fig. 3e are reasonably decomposed into four Gaussian peaks with binding energies of 530.6 (C@O/COOH), 532.0
(AOH), 532.6 (CAO), and 534.7 eV (H2O) [27,28]. After the solgel
reaction, the Al-GO spectrum shows several peaks at 530.6, 531.8,
532.6, and 535.1 eV, observed at similar binding energies to those
of GO, along with two new peaks. These new peaks at 531.1
(AlOOH) and 533.1 eV (Al(OH)3) also veried the presence of
Al(OH)3 on the Al-GO [20,29]. According to the XPS analysis, GO
was completely fabricated via the Hummers method, and Al(OH)3
successfully covered the surface of GO using the solgel method.
3.2. Thermomechanical properties
A DMA is often used to study the viscoelastic properties of polymers under stress and increased temperature. Fig. 4 shows the

598

J. Kim et al. / Microelectronics Reliability 52 (2012) 595602

Fig. 3. C 1s X-ray photoelectron spectra of (a) graphite, (b) graphene oxide and (c) Al(OH)3-fuctionalized graphene and O 1s spectra of (d) graphite, (e) graphene oxide and (f)
Al(OH)3-fuctionalized graphene.

variations in storage modulus and tan d as a function of temperature from below the glassy state temperature range to the rubbery
plateau of pure epoxy and its composites with GO and Al-GO. As
shown in Fig. 4, the storage modulus and Tg of the composites were
increased by the addition of GO and Al-GO llers.
The DMA results showed that the adhesives with modied
graphene llers had a signicantly higher storage modulus at room

temperature and a considerably higher rigidity, as seen in Fig. 4.


The storage modulus plots for the GO/epoxy and Al-GO/epoxy
composites are presented in Fig. 4a. The initial elastic moduli of
the GO/epoxy and Al-GO/epoxy composites were higher than that
of the pure epoxy resin, and the rubbery modulus steadily increased with increasing ller concentration for both systems. At
298 K (25 C), the GO/epoxy composites with 1 and 3 wt.% GO

J. Kim et al. / Microelectronics Reliability 52 (2012) 595602


Table 1
Atomic concentrations of functionalized graphene sheets based on the XPS analysis.

Graphite
GO
Al-GO

C (%)

O (%)

Al (%)

O/C ratio

97.29
35.51
30.08

2.71
64.49
66.50

3.42

0.028
1.82
2.21

599

In this study, the Tg of pure epoxy polymer was about 344 K


(71 C). With the addition of 1 wt.% Al-GO and GO, the Tg values increased to 355.5 K (82.5 C) and 362 K (89 C), respectively. When
the ller content was 3 wt.%, the Tg values reached 373.5 K
(100.5 C) in the case of Al-GO and 376 K (103 C) for GO. In general, the increase in Tg in any polymeric system is associated with
a restriction in molecular motion, a reduction in free volume and a
higher degree of crosslinking. In this case, a restriction of the
molecular motion and reduction of the free volume resulted in restricted polymer chain mobility.

3.3. Thermal conductivity


The graphene-based materials provide an efcient thermal conductivity enhancement compared to those of other carbon materials, including single-wall carbon nanotubes (SWNTs), because of
the high aspect ratio, dimensionality, rigidity of the graphene layer,
and the thermal interface resistance between the graphene and the
polymer matrix [32].
Fig. 5 compares the thermal conductivities of GO/epoxy and AlGO/epoxy composites prepared with 0, 1 and 3 wt.% llers at room
temperature. The thermal conductivity of the pure DGEBA epoxy
resin was around 0.188 W/m K, and, as shown in Fig. 5, the thermal
conductivities of the composites increased with the addition of GO
and Al-GO llers. The thermal conductivity of pure epoxy was
0.188 W/m K, which increased to 0.226 and 0.358 W/m K with
the addition of 1 and 3 wt.% GO, respectively. Al-GO/epoxy composites also had increased thermal conductivities compared to
pure epoxy: 0.223 and 0.254 W/m K at 1 and 3 wt.%, respectively.
At a 3 wt.% loading, the GO/epoxy composite showed an enhanced
thermal conductivity that was almost twice as great as that of the
neat epoxy resin. The increases in the thermal conductivity values
due to the addition of llers in this work were much higher than
those reported in the literature for other electrically insulated ller/epoxy composite systems [33].
Thermal conductivity is affected by the carbon nanoller structure quality within the matrix, loading, dispersion and the thermal
resistance of the interface between the nanoller and the polymer
matrix [3436]. In Fig. 5, the GO/epoxy composites demonstrated a
better thermal ller performance than did the Al-GO/epoxy composites. The proportion of Al(OH)3 in Al-GO can signicantly decrease the total thermal conductivity of the Al-GO/epoxy
composite because Al(OH)3 has a lower thermal conductivity value
than that of the GO sheet.

Fig. 4. Dynamic mechanical properties of neat epoxy composites and of epoxy


composites with GO and Al-GO.

showed about 31.4% and 56.4% higher storage moduli than that of
pure epoxy resin (5892 MPa), respectively. The storage moduli of
Al-GO/epoxy composites also increased by 34.5% (1 wt.%) and
36.4% (3 wt.%) over that of pure epoxy. As the temperature increased, both pure epoxy and functionalized graphene/epoxy composites experienced a gradual decrease in storage modulus,
followed by a sudden decrease at the glass transition temperature
(Tg). The decrease in the modulus value is related to the material
transition from a glassy state to a rubbery state.
The tan d plots for the pure epoxy and the chemically treated GO
and Al-GO are presented in Fig. 4b, and the peaks illustrate the Tg of
each polymer. In several cases, the dispersion, the particle size and
the surface modication of the llers play important roles in the
changes in Tg [30,31]. As shown in Fig. 4b, the Tg of the epoxy composites increased gradually as the ller concentration increased.

Fig. 5. Thermal conductivities of epoxy composites of neat epoxy, graphene oxide


and Al(OH)3-functionalized graphene.

600

J. Kim et al. / Microelectronics Reliability 52 (2012) 595602

Fig. 6. FE-SEM images of functionalized graphene/LMPA/epoxy composites: (a) GO at low magnication, (b) GO at high magnication, (c) Al-GO at low magnication and (d)
Al-GO at high magnication.

3.4. Microstructures of the composites


The microstructure of 5 wt.% functionalized graphene/epoxy
composites with 30 vol.% of LMPA solders was investigated using
FE-SEM to compare the dispersion and compatibility of the composites. In order to prevent the melting of LMPA solders, composites were cured at 403 K (130 C) for 18 h.
The fractured surface exhibited dispersion and compatibility of
LMPA and functionalized graphene in the epoxy resins, as shown in
Fig. 6. For each specimen, two types of micrographs are presented,
one at a relatively low magnication (Fig. 6a and c) and the other at
a higher magnication (Fig. 6b and d). As shown in Fig. 6a and c,
the LMPA solders maintain a spherical shape because the curing
temperature of the functionalized graphene/LMPA/epoxy composites was lower than the melting temperature of LMPA solder. The
GO/LMPA/epoxy and Al-GO/LMPA/epoxy composite images
showed that the distribution of functionalized graphene llers
was uniform without segregation after curing, as shown in
Fig. 6a and c. The good dispersion of functionalized graphenes in
epoxy could be attributed to the intermolecular interactions between the epoxy matrix and the graphene llers, such as intermolecular hydrogen bonding between the hydroxyl group in the
epoxy and the hydroxyl group in the ller surface [37,38]. The
images also revealed that the GO and Al-GO randomly dispersed

as a 3D network through the polymer matrix rather than simply


aligning parallel to the surface of the sample. At higher magnication, however, it was observed that the Al-GO had a slightly poorer
dispersion with a tendency to restack compared to that of GO. The
oxygen and hydroxyl groups on GO were more suited for forming
an intimate llermatrix interaction than those on Al-GO, which
consisted almost completely of hydroxyl groups.
3.5. Hybrid interconnection process
Fig. 7 shows the morphology of the conduction path between
the QFP lead and the substrate formed through the melting-coalescence-wetting behavior of the LMPA solders. The bonding image
consisted of two parts: the conduction path between the Cu pattern and the QFP lead made by the melted LMPA solder ball and
the polymer around the lead, which joined the lead and substrate.
Fig. 7ac shows the cross-sections of LMPA soldered joints with
raw epoxy polymers (LMPA/epoxy composite) and those blended
with 1 wt.% GO or Al-GO (GO/LMPA/epoxy and Al-GO/LMPA/epoxy
composites), respectively. Most of the LMPA solders were wetted
and developed conduction paths in all systems, with the exception
of several non-wetted LMPA solder balls around the QFP leads.
ACAs provide not only electrical interconnections, but also
mechanical support. The pull test was performed to evaluate the

Fig. 7. OM images of the conduction path with 40 vol.% LMPA solders for (a) pure epoxy, (b) GO/DGEBA and (c) Al-GO/DGEBA.

J. Kim et al. / Microelectronics Reliability 52 (2012) 595602

601

Furthermore, the pull force and the wetting tests showed no significant changes. These results may be attributed to the functional
groups on the GO and Al-GO surfaces, which determine the
improvement in compatibility between graphene-based llers
and epoxy resins.

Acknowledgement
This research was supported by the Seoul R&BD program (No.
PA090933).

References

Fig. 8. Pull force test of the QFP leads of the LMPA/DGEBA composites with 1 wt.%
of GO and Al-GO.

effect of functionalized graphene sheets on the mechanical pull


force of interconnected QFPs. Fig. 8 shows the pull test results of
GO/LMPA/epoxy and Al-GO/LMPA/epoxy composites with 1 wt.%
functionalized graphene llers compared to that of LMPA/epoxy
resin. The pull force results of the solder joints indicated that the
LMPA/epoxy composites containing GO and Al-GO had lower values compared to that of the neat LMPA/epoxy composite. The pull
force in the neat LMPA/epoxy composite was 18.4 N, while those of
the GO and Al-GO blended composites were slightly decreased to
17.7 N and 18.2 N respectively. These results are nearly equal to
those of the neat LMPA/epoxy resin. As was stated above, however,
hydrogen bonds between epoxy and functionalized graphene llers can improve the toughness of epoxy composites and increase
compatibility between epoxy resin and llers [37,39]. Therefore,
functionalized graphene sheets can be easily dispersed in epoxy
resins and improve the mechanical properties of functionalized
graphene embedded composites. This effect leads to the similar
pull test values of composites compared to neat LMPA/epoxy resin.

4. Conclusions
In the present work, the effects of functionalized graphene
sheets within epoxy composites on the thermal and mechanical
properties, morphology and pull test of a hybrid interconnection
were investigated. Graphene oxide (GO) and Al(OH)3-functionalized graphene nanosheets (Al-GO) were fabricated using Hummers method and a simple solgel method with aluminum
isopropoxide. Using ake graphite as a starting material, strong
chemical oxidants were used for the synthesis of GO, and the hydroxyl groups of GO and AIP were reacted via a simple solgel
method.
The XPS analysis conrmed that GO and Al-GO were successfully fabricated. The storage modulus and tan d increased with
the amount of ller. Compared to neat epoxy, the DMA results indicated 56.5% and 36.4% improvements in the storage modulus for
composites with 3 wt.% GO and Al-GO at 298 K (25 C), respectively. The thermal conductivities of composites with llers were
also higher than that of the neat epoxy, and the thermal conductivity values improved according to ller concentration. The 3% GO/
epoxy composite had the highest thermal conductivity, about
twice that of the neat epoxy resin. The dispersion of functionalized
graphenes in the epoxy resin was good, and coalescence of the
added LMPA solders of the GO and Al-GO assembly demonstrated
similar properties to those in the neat epoxy assembly.

[1] Lin YC, Zhong J. A review of the inuencing factors on anisotropic conductive
adhesives joining technology in electrical applications. J Mater Sci
2008;43:307293.
[2] Eom YS, Jang KS, Moon JT, Nam JD. Electrical interconnection with a smart ACA
composed of uxing polymer and solder powder. ETRI J 2010;32:41421.
[3] Jagt JC, Beris PJM, Lijten GFCM. Electrically conductive adhesives: a prospective
alternative for SMD soldering? IEEE Trans Compon Package Manuf Technol B
1995;18:2928.
[4] Lau JH. Flip chip technologies. New York: McGraw-Hill; 1995.
[5] Minges ML. Electronics materials handbook packaging. ASM International;
1989.
[6] He Y, Moreira BE, Overson A, Nakamura SH, Bider C, Briscoe JF. Thermal
characterization of an epoxy-based underll material for ip chip packaging.
Thermochim Acta 2000;357358:18.
[7] Viswanath R, Wakharkar V, Watwe A, Lebonheur V. Thermal performance
challenges from silicon to systems. Intel Technol J 2000;Q3:116.
[8] Xu Y, Luo X, Chung DDL. Lithium doped polyethylene-glycol-based thermal
interface pastes for high thermal contact conductance. J Electron Packag
2002;124:18891.
[9] Kapur P, McVittie JP, Saraswat KC. Technology and reliability constrained
future copper interconnects. I. Resistance modeling. IEEE Trans Electron Dev
2002;49:5907.
[10] Obreja VVN. On the reliability of power silicon rectier diodes above the
maximum permissible operation junction temperature. IEEE Int Symp Ind
Electron ISIE06 2006:540835.
[11] Zhu J, Kim JD, Peng H, Margrave JL, Khabashesku VN, Barrera EV. Improving the
dispersion and integration of single-walled carbon nanotubes in epoxy
composites through functionalization. Nano Lett 2003;3:110713.
[12] Geng Y, Liu MY, Li J, Shi XM, Kim JK. Effects of surfactant treatment on
mechanical and electrical properties of CNT/epoxy nanocomposites. Compos
Part A 2008;39:187683.
[13] Flahaut E, Peigney A, Laurent Ch, Marlie Ch, Chastel F, Rousset A, et al. Carbon
nanotubemetaloxide
nanocomposites:
microstructure
electrical
conductivity and mechanical properties. Acta Mater 2008;48:380312.
[14] Ma PC, Kim JK, Tang BZ. Functionalization of carbon nanotubes using a silane
coupling agent. Carbon 2006;44:32328.
[15] Lee C, Wei X, Kysar JW, Hone J. Measurement of the elastic properties and
intrinsic strength of monolayer graphene. Science 2008;321:3858.
[16] Balandin AA, Ghosh S, Bao W, Calizo I, Teweldebrhan D, Miao F, et al. Superior
thermal conductivity of single-layer graphene. Nano Lett 2008;8:9027.
[17] Du X, Skachko I, Barker A, Andrei EY. Approaching ballistic transport in
suspended graphene. Nat Nanotechnol 2008;3:4915.
[18] Yim MJ, Paik KW. Recent advances on anisotropic conductive adhesives (ACAs)
for at panel displays and semiconductor packaging applications. Inter J Adhes
Adhes 2006;26:30413.
[19] Hummers WS, Offeman RE. Preparation of graphitic oxide. J Am Soc
1958;80:1339.
[20] Hernadi K, Couteau E, Seo JW, Forro L. Al(OH)3/multiwalled carbon nanotube
composite: homogeneous coverage of Al(OH)3 on carbon nanotube surfaces.
Langmuir 2003;19:70269.
[21] Kim HM, Kim JM, Kim JH. Effects of novel carboxylic acid-based reductants on
the wetting characteristics of anisotropic conductive adhesive with low
melting point alloy ller. Microelectron Reliab 2010;50:25865.
[22] Kwon YM, Yim BS, Kim JM, Kim JH. Dispersion, hybrid interconnection and
heat dissipation properties of functionalized carbon nanotubes in epoxy
composites for electrically conductive adhesives (ECAs). Microelectron Reliab
2011;51:8128.
[23] Yim BS, Kim JM, Jeon SH, Lee SH, Kim JH, Han JG, et al. Hybrid Interconnection
process using solderable ICAs (isotropic conductive adhesives) with lowmelting-point alloy llers. Mater Trans 2009;50:264955.
[24] Geng Y, Wang SJ, Kim JK. Preparation of graphite nanoplatelets and graphene
sheets. J Colloid Inter Sci 2009;336:5928.
[25] Stankovich S, Dikin DA, Piner RD, Kohlhaas KA, Kleinhammers A, Jia Y, et al.
Synthesis of graphene-based nanosheets via chemical reduction of exfoliated
graphite oxide. Carbon 2007;45:155865.
[26] Martin JM, Vovelle L, Bou M, Le Mogne TH. Chemistry of the interface between
aluminium and polyethyleneterephthalate by XPS. Appl Surf Sci 1991;47:149.

602

J. Kim et al. / Microelectronics Reliability 52 (2012) 595602

[27] Yang DX, Velamakanni A, Bozoklu G, Park SJ, Stoller M, Piner RD, et al.
Exfoliation of graphite oxide in propylene carbonate and thermal reduction of
the resulting graphene oxide platelets. Carbon 2009;47:14552.
[28] Costa D, Marcus P, Yang WP. Resistance to pitting and chemical composition of
passive lms of a Fe17% Cr alloy in chloride-containing acid solution. J
Electrochem Soc 1994;141:266976.
[29] Rotole JA, Sherwood PMA. Valence band X-ray photoelectron spectroscopic
studies to distinguish between oxidized aluminum species. J Vac Sci Technol A
1999;17:10916.
[30] Ash BJ, Schadler LS, Siegel RW. Glass transition behavior of alumina/
polymethylmethacrylate nanocomposites. Mater Lett 2002;55:837.
[31] Xiong M, Cu G, You B, Wu L. Preparation and characterization of poly(styrene
butylacrylate) latex/nano-ZnO nanocomposites. J Appl Polym Sci
2003;90:1923.
[32] Yu A, Ramesh P, Itkis ME, Bekyarova E, Haddon RC. Graphite
nanoplateletepoxy composite thermal interface materials. J Phys Chem C
2007;111:75659.
[33] Kochetov R, Korobko AV, Andrithsch T, Morshuis PHF, Picken SJ, Smit JJ.
Modelling of the thermal conductivity in polymer nanocomposites and the

[34]

[35]
[36]

[37]
[38]

[39]

impact of the interface between ller and matrix. J Phys D: Appl Phys
2011;44:395401.
Yang SY, Ma CCM, Teng CC, Huang YW, Liao SH, Huang YL, et al. Effect of
functionalized carbon nanotubes on the thermal conductivity of epoxy
composites. Carbon 2010;48(3):592603.
Lin C, Chung DDL. Graphite nanoplatelet pastes vs. carbon black pastes as
thermal interface materials. Carbon 2009;47:295305.
Biercuk MJ, Llaguno MC, Radosavljevic M, Hyun JK, Johnson AT, Fischer JE.
Carbon nanotube composites for thermal management. Appl Phys Lett
2002;80:27679.
Park SJ, Jin FL, Lee JR. Synthesis and thermal properties of epoxidized vegetable
oil. Macromol Rapid Commun 2004;25:7247.
Medhekar NV, Ramasubramaniam A, Ruoff RS, Shenoy VB. Hydrogen bond
networks in graphene oxide composite paper: structure and mechanical
properties. ACS NANO 2010;4:23006.
Park SJ, Jin FL, Lee C. Preparation and physical properties of hollow glass
microspheres-reinforced epoxy matrix resins. Mater Sci Eng A
2005;402:33540.

You might also like