- Email: [email protected]
silicone composite with enhanced thermal conductivity and its application in heat dissipation of high-power light-emitting diodes
Accepted Manuscript Graphene nanosheet/silicone composite with enhanced thermal conductivity and its application in heat dissipation of high-power light-emitting diodes Haiyan Zhang, Yingxi Lin, Danfeng Zhang, Wenguang Wang, Yuxiong Xing, Jin Lin, Haoqun Hong, Chunhui Li PII:
S1567-1739(16)30275-9
DOI:
10.1016/j.cap.2016.10.004
Reference:
CAP 4339
To appear in:
Current Applied Physics
Received Date: 7 July 2016 Revised Date:
5 September 2016
Accepted Date: 4 October 2016
Please cite this article as: H. Zhang, Y. Lin, D. Zhang, W. Wang, Y. Xing, J. Lin, H. Hong, C. Li, Graphene nanosheet/silicone composite with enhanced thermal conductivity and its application in heat dissipation of high-power light-emitting diodes, Current Applied Physics (2016), doi: 10.1016/ j.cap.2016.10.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Graphene nanosheet/silicone composite with
2
enhanced thermal conductivity and its application in
3
heat dissipation of high-power light-emitting diodes
4
Haiyan Zhang,b,*, Yingxi Lin a, Danfeng Zhang a, Wenguang Wang a,b, Yuxiong Xing a,
5
Jin Lina, Haoqun Honga, Chunhui Lia
RI PT
1
6
a
7
China
8
b
9
University of Technology, Guangzhou, 510006, China
M AN U
SC
School of Materials and Energy, Guangdong University of Technology, Guangzhou, 510006, P. R.
Guangdong Provincial Key Laboratory of Functional Soft Condensed Matter, Guangdong
AC C
EP
TE D
10
*Corresponding author. E-mail address: [email protected] 1
ACCEPTED MANUSCRIPT
Abstract: Heat dissipation from light-emitting diodes (LEDs) has become a serious
2
problem because of the LEDs’ high luminosity and power. This problem could be
3
solved by improving the dissipation efficiency of each LED system component. A
4
microwave-reduced graphene nanosheet/silicone (GN/silicone) composite with a high
5
thermal conductivity and stability was prepared by mechanical blending. The thermal
6
conductivity of the composite reaches 2.7 W/(m K) with only 1.5 wt.% loading, and is
7
12 times higher than the pure silicone matrix. When used as a thermal interface
8
material between high-power LED chip module substrates and heat sinks, the thermal
9
conductive GN/silicone composite could decrease the temperature difference between
10
the substrate and shell. It could also improve the system heat transfer efficiency. The
11
temperature gap between the heat slug and the heat sink was less than 2°C with a 1.5
12
wt.% loading of GNs.
M AN U
SC
RI PT
1
14
TE D
13
KEYWORDS: Graphene nanosheet; thermal interface material; LED dissipation;
16
thermal conductivity.
AC C
17
EP
15
2
ACCEPTED MANUSCRIPT 1
1 Introduction Light-emitting diodes (LEDs) are two-lead semiconductor p–n junction light
3
sources that emit light when activated. Because they do not radiate at ultraviolet or
4
infrared wavelengths, they provide several advantages compared with other light
5
sources, such as a long service lifetime, a high luminaire efficiency and they are
6
environmentally friendly [1]. However, LEDs convert 75–85% of their electrical
7
power into heat, which reduces their efficiency and makes them less reliable. The
8
problem is more serious for high-power LEDs, which can be driven at currents from
9
hundreds of milliamperes to more than an ampere, compared with the tens of
SC
milliamperes for other LEDs [2].
M AN U
10
RI PT
2
Therefore, it is important to design an efficient thermal system within the LEDs.
12
One method to do so is to improve the heat transfer efficiency between the LED heat
13
slugs and sinks. Much work has been done to increase the efficiency by optimizing
14
the heat sink [3], adjusting the refrigerant in heat pipes [4] and reducing the heat
15
resistance between the LED chips and heat sinks [5]. Among these methods,
16
improving the heat transfer efficiency with thermal interface materials (TIMs)
17
between LEDs and heat sinks is seen as an inexpensive, simple way which can be
18
used in large-scale industrial production.
EP
AC C
19
TE D
11
In the practical application of high-power LEDs, LED chip modules are arranged in
20
arrays and are welded on a specialized substrate. When the substrates are clamped to
21
the heat sinks, microscopic surface undulations that form air gaps can hinder heat
22
transfer and decrease the heat transfer efficiency substantially [6]. Filling the air gap
23
with soft thermal conductive TIMs can facilitate heat transfer, because the thermal
24
conductivity of TIMs is usually two orders larger than air (~0.025 W/(m K) at room 3
ACCEPTED MANUSCRIPT 1
temperature) [7]. In general, widely used TIMs are made of polymer matrices with the addition of
3
high thermal-conductivity materials, such as metals (e.g., silver, copper), ceramics
4
(e.g., aluminium oxide, zinc oxide) or carbon materials (e.g., graphite, carbon
5
nanotubes). Metal particles can be oxidized so easily at high temperature that the heat
6
conducting property worsens. In addition, these thermal conductive particles require
7
high-volume fractions of filler to reach thermal conductivity of the composite from
8
~1–5 W/(m K) at room temperature, which may lead to poor composite mechanical
9
properties [8,9].
SC
RI PT
2
Carbon nanotubes (CNTs) have attracted interest as fillers for TIMs because of
11
their extremely high intrinsic thermal conductivity from ~3000–3500 W/(m K) at
12
room temperature [10]. However, their addition does not improve the thermal
13
conductivity of the matrix as expected or decrease with the addition of single-wall
14
CNTs [11]. It is believed that CNTs with an excellent thermal conductivity do not
15
couple well to the matrix material or contact surface. Therefore, there is optimism in
16
the search for alternative high thermal-conductivity fillers.
TE D
M AN U
10
In recent years, graphene has attracted considerable attention in various fields
18
because of its excellent mechanical, electric, and thermal performance [12–15].
19
Balandin et al. measured the thermal conductivity (λ) of single-layer graphene by
20
laser and found that the λ is much higher than most metals at 4800–5300 W/(m K)
21
[16]. This result is attributed to a long free path for phonon transport, which is by
22
ion-core vibration in a crystal lattice [17,18].
AC C
EP
17
23
High-quality and large-scale graphene production is required for use in TIMs.
24
Several methods have been developed to solve these problems, including epitaxial
4
ACCEPTED MANUSCRIPT 1
growth, chemical vapor deposition, and chemical exfoliation [19]. Of these methods,
2
chemical exfoliation of graphite oxide is seen as a simple and rapid way to obtain
3
graphene at a large scale [20–23]. In this paper, an industrial scalable, low-cost, and simple approach was used to
5
prepare graphene. This method produced graphene by exfoliating oxide graphite into
6
small, thin pieces by using microwave radiation [24,25]. Microwave-reduced
7
graphene nanosheets (GNs) were added to a silicone matrix to prepare
8
microwave-reduced GN/silicone composites, which was used as a thermal interface
9
material in the heat dissipation of high-power LEDs.
2. Experimental
11
2.1 Preparation of GNs
M AN U
10
SC
RI PT
4
As shown in Scheme 1, graphite oxide (GO) was prepared as follows. Firstly, 184
13
ml of concentrated H2SO4 was added to 4 g of NaNO3 in a 1000 ml conical flask at
14
0°C in a thermostatic water bath. The mixture was stirred continuously to form a
15
homogeneous solution. Then, 8 g of flake graphite and 24 g of KMnO4 were added
16
into the above solution, and the mixture was stirred at 0°C for 2 h. After that, the
17
temperature of the water in the thermostatic water bath was increased to 35°C and the
18
above mixed solution was kept stirring for 2 h. Afterwards, the thermostatic water
19
bath temperature was set to 75°C . Deionized water was added slowly into the conical
20
flask during the heating process. When the temperature reached 75°C, a certain
21
amount of H2O2 was added into the mixture to remove residual KMnO4. The conical
22
flask was removed and cooled naturally to room temperature. The resulting
AC C
EP
TE D
12
5
ACCEPTED MANUSCRIPT 1
golden-yellow mixed solution was washed with deionized water several times to pH 7.
2
Finally, graphite oxide was obtained after centrifugation and drying in a vacuum
3
drying oven at 60°C. The as-prepared GO was ground in a planetary ball mill for 10 h and then heated in
5
a microwave oven at 1000 W in a nitrogen atmosphere. When the temperature reached
6
300°C, the microwave oven was turned off. Black inflated GNs were obtained.
7
2.2 GN modification
SC
RI PT
4
To improve the compatibility of the GNs with the matrix, surface modification was
9
carried out by 3-aminopropyltriethoxysilane (KH-550, Kang Lan Biochemical Pte.,
10
Ltd., Shanghai, China) silane conpling agent addition. The as-prepared black powder
11
was added into a 500 ml solution that contained alcohol and 1.5 wt.% KH-550. The
12
mixed solution was sonicated for 30 min. The solution was filtrated and the
13
pre-treated GNs were dried at 120°C in a drying oven.
14
2.3 Preparation of GN/silicone composite
EP
TE D
M AN U
8
Silicone gel (Chun Chang Chemical Pte., Ltd., Shenzhen, China) consisted of
16
component A, a methyl vinyl silicone oil with the molecular structure as shown in
17
Scheme 2, and curing agent component B, a methyl hydrogenous silicone oil with the
18
molecular structure as shown in Scheme 3.
19 20
AC C
15
Polymerization was initiated by chloroplatinic acid addition. The reaction mechanism involved hydrogen addition to vinyl and is illustrated in Scheme 4.
6
ACCEPTED MANUSCRIPT Organic silicone gel was chosen as a polymeric matrix in this work for the
2
following reasons: (1) good chemical inertness, (2) wide service temperature range
3
from 5.5°C to 260°C, (3) outstanding electrical performance and weathering
4
resistance, (4) advantageous conformability, (5) ease of processability, (6) adjustable
5
hardness; (7) and low costs [26].
RI PT
1
Silicone A (10 g) was mixed with GNs homogeneously using an agate pestle in a 50
7
ml agate mortar. Silicone B (10 g) was added and blended uniformly by using
8
double-roll milling at room temperature for 15–20 min. The mixture was transferred
9
to a 15 cm diameter glass dish and placed in a vacuum oven for degassing. To
10
complete the curing reaction, curing at 150°C for 1 h and at 200°C for an additional 2
11
h was applied, after which the composite was peeled off easily from the glass dish. A
12
series of composites was prepared with different GN weight fractions of 0.1 wt.%, 0.3
13
wt.%, 0.5 wt.%, 1 wt.%, and 1.5 wt.%. A pristine silicone sample was produced for
14
comparison.
15
2.4 Characterization and measurements
M AN U
TE D
EP
AC C
16
SC
6
Transmission electron microscopy (TEM) was performed at an accelerating voltage
17
of 200 kV, using a JEOL JEM-2100. Micrographs were also obtained by using an
18
S-3400N-II scanning electron microscope (SEM).
19 20
Using CuKα radiation, X-ray diffraction (XRD) measurements were conducted at a scan rate of 8° min-1 in a D/Max-IV (Rigaku Co., Tokyo, Japan) at 40 kV and 30 mA.
7
ACCEPTED MANUSCRIPT Functional groups on the surface of GN were identified via Fourier transform
2
infrared (FTIR) spectroscopy (Thermo Scientific, Nicolet 6700). Thermogravimetric
3
analysis (TGA) was used to evaluate the thermal stabilities of the composite; this
4
analysis was performed in a nitrogen atmosphere and at 10°C min-1, using an SDT
5
2960 Simultaneous DSC-TGA (TA Instrument, USA). The thermal conductivities of
6
the composite were measured by using a thermal-conductivity meter LFA-440
7
(Netzsch, Germany), based on the laser flash method in which the front side of a
8
plane-parallel sample is heated by using a short light pulse and the temperature
9
increase on the rear surface was measured using an infrared detector.
10
3. Results and discussion
11
3.1 GN morphology and structure
M AN U
SC
RI PT
1
For the large-scale synthesis of graphene, graphite remains the most popular
13
precursor because of its higher yield and overall process cost. The general route for
14
graphene synthesis from graphite is oxidation to graphite oxide, followed by
15
exfoliation of the graphite oxide to graphene. Thermal exfoliation is the most
16
common method, and thermal degradation of the oxide functional groups results in an
17
evolution of substantial quantities of oxide and carbon dioxide, which can result in the
18
enhanced expansion of graphite oxide along the thickness direction. In general,
19
graphene is derived from GO, which can be reduced to graphene via direct thermal
20
treatment at elevated temperatures without using hazardous reducing agents. To
21
obtain completely exfoliated and thermal conductive graphene, “thermal shock” of
22
GO at temperatures of
AC C
EP
TE D
12
up to 1050°C was used. However, such a high temperatue 8
ACCEPTED MANUSCRIPT 1
lead to a low production efficiency and high cost, which wastes energy.
Therefore, microwave irradiation assisted synthesis is an ideal way to prepare
3
graphene on a large scale, and it is also an economic, environmentally friendly and
4
efficient synthesis technique. Compared with conventional heat treatment, microwave
5
irradiation can be used to synthesize graphene via superior “inert and instant heating”.
6
The GO volume can be reduced completely by the rapid, uniform and energy-efficient
7
microwave irradiation process.
SC
RI PT
2
SEM images show that the GNs consist of randomly aggregated, crumpled sheets
9
(Figure 1a–c). The folded sheet regions in Figure 1c had average widths of several
10
nanometers [27,28]. The nanosheet thickness as estimated from the edge of a single
11
GN (Figure 1a) is 1–5 nm, and it appears to be of a similar thickness compared with
12
Ahmad et al.’s report [28]. This result was confirmed by transmission electron
13
microscopy measurements. Figure 1d shows a large transparent, wrinkled
14
graphene-like area that is caused by a high specific surface. As reported in the
15
literature [29,30], this crimping results from thin walls that consist of a few layers of
16
graphene, and that corresponds to a 1–5 nm thickness.
TE D
EP
AC C
17
M AN U
8
The chemical structure of natural graphite (NG), GO, and GNs was determined via
18
XRD and FTIR spectroscopy. As shown in XRD (Figure 2), the strong NG diffraction
19
peak at 2θ = 26°, corresponds to the (002) reflection, which indicates a high degree
20
of graphitization of the NG. However, GNs shows a broad peak at 25°, and such a 9
ACCEPTED MANUSCRIPT broad curve is characteristic of amorphous material and indicates the successful
2
reduction of GO. According to the Scherrer equation [31], the interlayer gaps of NG,
3
GO and GN are 0.339 nm, 0.795 nm and 0.344 nm, respectively. These results
4
manifest that the interlayer gaps of GNs become larger than those of NG after
5
reduction, which suggests incomplete reduction and the existence of residual
6
functional groups. The incompletely reduced GNs can be dispersed better in a
7
polymer matrix with the help of polar functional groups, such as hydroxyl groups.
SC
RI PT
1
The FT-IR spectra of the GO and GNs are shown in Figure 3. The broad peak at
9
3400 cm-1 in the spectra is associated with OH groups. Significant peaks at 1382 cm-1
10
and 1216 cm-1 are attributed to carboxy C-O and epoxy C-O groups [32]. The peak at
11
1618 cm-1 arises from the existence of aromatic C=C. However, the peak at 1716 cm-1,
12
which is associated with a COOH group (C=O vibration), does not appear for the GNs.
13
It can be inferred that the GNs were partly reduced, so some functional groups of GO
14
remained. This result implies that GNs with functional groups are compatible with the
15
polymer matrix as indicated by the XRD results.
16
3.2 Dispersion of GNs in TIM
TE D
EP
AC C
17
M AN U
8
GNs have large specific surface areas, and a high surface energy, which causes
18
them to tend to aggregate. In double-roll milling, the surface between the materials
19
and the rollers was of a different velocity. This velocity initiated an internal shear
20
stress that can be controlled by adjusting the amount of materials and the distances 10
ACCEPTED MANUSCRIPT 1
between two rolls. By using KH-550, we could obtain a well-dispersed composite by mixing for 15–
3
20 min. As the SEM images show in Figure 4 (a–d), the GNs are dispersed in silicone
4
and are compatible with silicone. The GN remains as a sheet (red line) in the matrix
5
without any agglomeration. If the GNs are agglomerated, the agglomerate bulk could
6
be found
7
Shape remaining GNs were found even at a high loading of 1.5 wt.% (Figure d). It is
8
believed that the GNs disperse well in the matrix. The remaining sheet shape favors
9
the formation of conductivity chains because of the high aspect ratio of the GNs. The
10
silicone becomes black with increase in GNs (Figure 4e and f), and no obvious
11
bubbles, cracks or other defects are visible in the composite, which may hinder heat
12
transport in the composite. The composite size can be controlled by roll size. As a
13
result, we produced a composite of as large as 18 cm × 18 cm. An even larger
14
composite can be produced by using a suitable roll size, and large-scale production is
15
possible, which is significant in industrialization.
16
3.3 TIM thermal stability
EP
TE D
M AN U
SC
easily as sphere-like shape within the matrix as reported by Alireza [33].
AC C
17
RI PT
2
The DSC curves of a pure silicone and a GN/silicone composite with 1.0 wt.% GNs
18
is visible in Figure 5. For a sample of pure silicone, several strong endothermic peaks
19
appeared between 300°C and 600°C, which indicates that silicone decomposition
20
occurred, whereas the composite did not show significant endothermic peaks at the
21
same temperature. Therefore, composites could remain thermally stable at the same
11
ACCEPTED MANUSCRIPT 1 2
temperature at which pure silicone decomposed. The TGA results are shown in Figure 6. The composite possesses a good thermal stability compared with pure silicone. Firstly, there is no obvious mass loss (3–5 wt.%)
4
below 300°C for samples. The pure silicone (black line) began to lose mass at 300°C,
5
when the composites decomposing temperature were increasing gradually with fillers
6
addition increasing, which was 100°C higher than that of pure silicone at the loading
7
of 1.0 wt.% (~400°C). Secondly, the remainder of the composite at the loading of 1.0
8
wt.% is 30 wt.% higher than the silicone, which is also larger than the GN addition.
SC
RI PT
3
These results indicate that GNs in the composite can interconnect to form a
10
network frame, which serves as a frame and restricts silicone molecule movement.
11
The GN high specific surface area results in an increase in contact areas between the
12
GNs and the silicone. Therefore, the surface bond strength of these two materials can
13
be enhanced significantly. The interaction between GNs and silicone hinders
14
molecular segment movement in the composite [34].
15
3.4 TIM thermal conductivities
TE D
M AN U
9
The composite thermal conductivity increased with increase in filler content
17
(Figure 7). The error bars represent data scatter for several samples. The thermal
18
conductivity of pure silicone is 0.22 W/(m K). However, the composite could have a
19
thermal conductivity of 1.51 W/(m K) at a loading of 0.5 wt.% GNs. With the
20
addition of 1.5 wt.% GNs, the composite thermal conductivity reached 2.75 W/(m K),
21
which is 13 times that of pure silicone. This result is much higher than that of other
22
composites. In addition, GNs display better filler performance than for other kinds of
23
carbon materials, such as carbon nanotubes or GO [35]; this can be attributed to the
24
high quality of GNs prepared by microwave reduction, which damages the GNs less.
AC C
EP
16
12
ACCEPTED MANUSCRIPT The enhancement ratio varies approximately linearly with concentration. However,
2
when the filler exceeds 0.5 wt.%, the thermal conductivity increases rapidly, which
3
suggests that a heat transport chain has formed. This heat transport chain benefits
4
phonon transport, which is the main way to convey energy in the composite.
5
Therefore, with increase in fillers, the formed heat transport chain could increase the
6
thermal conductivity suddenly. GNs with a high aspect ratio, can form heat transport
7
chain more easily compared with other carbon materials [36,37].
SC
RI PT
1
Table 1 lists the thermal conductivities of several TIMs with various fillers,
9
including silver, CNTs, and GO. Table 1 shows that the thermal conductivity of the
10
reference TIMs is not enhanced obviously even for a high loading. In addition, such a
11
high loading can influence the machine performance of the TIMs. In this case, the
12
TIM with GNs is advantageous when applied in an electronic device.
13
3.5 Application in an LED
TE D
M AN U
8
As shown in Scheme 5, the TIMs prepared above are used in a high-power LED
15
system to improve the heat transport efficiency between the heat slug and sink. The
16
composite that is applied in a high-power LED improves heat dissipation, and reduces
17
the temperature of the LED chip module.
c
AC C
EP
14
18
The histogram in Figure 8(a) shows the temperature of the heat slug and sink in
19
high-power LEDs. Ti represent the mean temperature of the heat slug, and Tj is the
20
mean temperature of the heat sink, as shown in Scheme 5.
21
When no TIM exists, Ti and Tj were 55 and 35°C, respectively. The temperature
22
gap between Ti and Tj is 20°C, which confirms that heat transported with difficulty
23
from the LED module to a heat sink because of the poor thermal conductivity of air 13
ACCEPTED MANUSCRIPT 1
(~0.025 W/(m K)). However, this difference can be reduced with the thermal
2
conductive augmentation of TIMs. When the thermal conductivity of the TIM reached
3
2.7 W/(m K), the temperature difference (Ti = 37°C, Tj = 35°C) reached only 2°C. The result is more obvious from Figure 8(b), which shows the relationship between
5
thermal conductivity and temperature gap. With increasing TIM thermal conductivity,
6
the temperature of the heat dissipation base chip module (heat slug) Ti decreases
7
significantly. When the thermal conductivity of the TIMs is 1.5 W/(m K) (0.5 wt.%
8
loading), the temperature difference is less than 5°C. It can be inferred that the
9
temperature difference decreases and that the heat dissipation improves in a
SC
high-power LED with an increase in thermal conductivity.
M AN U
10
RI PT
4
Therefore, it can be inferred that the temperature difference between Ti and Tj
12
decreases with increasing TIM thermal conductivity, which suggests an enhancement
13
in heat dissipation. Therefore, using high thermal conductive TIMs can decrease the
14
LED chip temperature, which favors their lifetime and performance.
15
4. Conclusion
TE D
11
GNs with a suitable dispersibility were synthesized by a microwave-reduced
17
method, which has potential for application in industry. GN/silicone composites were
18
prepared by mechanical blending. Compared with pure silicone, the addition of GNs
19
delays the initial composite decomposition temperature, and results in an outstanding
20
stability (the initial decomposition temperature of the composites is 100°C higher than
21
that of pure silicone). The thermal conductivity of the TIMs increase with increase in
22
fillers. When the GN loading is 1.5 wt.%, the maximum thermal conductivity of the
23
composite reaches 2.7 W/(m K). Such high conductive materials are helpful when
24
used as TIMs in high power LED systems. The thermal conductivity of 1.5 W/(m K)
AC C
EP
16
14
ACCEPTED MANUSCRIPT 1
can reduce the temperature gap between the LED module and the heat sink to 5°C.
2
Acknowledgments This work is supported by the link project of the National Natural Science
4
Foundation of China and Guangdong Province (No. U1401246), by the National
5
Natural Science Foundation of China (Grant No. 51276044), by the Science and
6
Technology Program of Guangdong Province of China (Grant No. 2014B010106005,
7
2015B010135011, 2015A050502047), and by the Science and Technology Program
8
of Guangzhou City of China (Grant No. 201508030018), by Research Fund of Young
9
Teachers for the Doctoral Program of Higher Education of China (20134420120009),
10
and by Special Program for Public Interest Research and Capability Construction of
11
Guangdong (2014A010105047)
AC C
EP
TE D
12
M AN U
SC
RI PT
3
15
ACCEPTED MANUSCRIPT 1
Reference
2
[1]
B. L. Ahn, C. Y. Jang, S. B. Leigh, S. Yoo, H. Jeong, Effect of LED lighting on the cooling and heating loads in office buildings, Appl. Energ. 113 (2014)
4
1484–1489.
5
[2]
RI PT
3
E. D. Jung, Y. L. Lee, Development of a heat dissipating LED headlamp with silicone lens to replace halogen bulbs in used cars, Appl. Therm. Eng. 86 (2015)
7
143–150.
heat sinks with application to LED light bulbs, Int. J. Heat Mass Transf. 71
9
(2014) 496–502.
10
11
D. Jang, S. J. Park, S. J. Yook, K. S. Lee, The orientation effect for cylindrical
M AN U
[3]
[4]
J. Wang, Y. Cai, X. Zhao, C. Zhang, Thermal design and simulation of
TE D
8
SC
6
automotive headlamps using white LEDs, Microelectronics J. 45 (2014) 249–
13
255.
14
[5]
AC C
17
53 (2013) 428–434.
[6]
J. Due, A. J. Robinson, Reliability of thermal interface materials: A review,
Appl. Therm. Eng. 50 (2013) 455–463.
18
19
M. Maaspuro, A. Tuominen, Thermal analysis of LED spot lighting device operating in external natural or forced heat convection, Microelectron. Reliab.
15 16
EP
12
[7]
R. B. Montgomery, Viscosity and thermal conductivity of air and diffusivity of
16
ACCEPTED MANUSCRIPT water vapor in air, J. Meteorol. 4 (1947) 193–196.
1
2
[8]
L. C. Sim, S. R. Ramanan, H. Ismail, K. N. Seetharamu, T. J. Goh, Thermal characterization of Al2O3 and ZnO reinforced silicone rubber as thermal pads
4
for heat dissipation purposes, Thermochim. Acta. 430 (2005) 155–165.
5
[9]
RI PT
3
K. M. F. Shahil, A. A. Balandin, Thermal properties of graphene and multilayer graphene: Applications in thermal interface materials, Solid State
7
Commun. 152 (2012) 1331–1340.
SC
6
[10] E. Pop, D. Mann, Q. Wang, K. Goodson, H. Dai, Thermal conductance of an
9
individual single-wall carbon nanotube above room temperature, Nano Lett. 6
10
M AN U
8
(2006) 96–100.
[11] S. U. S. Choi, Z. G. Zhang, W. Yu, F. E. Lockwood, E. A. Grulke, Anomalous
12
thermal conductivity enhancement in nanotube suspensions, Appl. Phys. Lett.
13
79 (2001) 2252–2254.
EP
15
[12] S. Park, R. S. Ruoff, M. Engineering, Chemical methods for the production of graphenes, Nat. Nanotechnol. 4 (2009) 217–224.
AC C
14
TE D
11
16
[13] Y. Q. Bie, J. Horng, Z. Shi, L. Ju, Q. Zhou, A. Zettl, et al., Vibrational
17
spectroscopy at electrolyte/electrode interfaces with graphene gratings, Nat.
18
Commun. 6 (2015) 7593.
19
[14] A. Varykhalov, J. Sánchez-Barriga, D. Marchenko, P. Hlawenka, P. S. Mandal,
20
O. Rader, Tunable Fermi level and hedgehog spin texture in gapped graphene, 17
ACCEPTED MANUSCRIPT 1
Nat. Commun. 6 (2015) 7610. [15] G. Zhou, E. Paek, G. S. Hwang, A. Manthiram, Long-life Li/polysulphide
3
batteries with high sulphur loading enabled by lightweight three-dimensional
4
nitrogen/sulphur-codoped graphene sponge, Nat. Commun. 6 (2015) 7760.
RI PT
2
[16] A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, et al.,
6
Superior thermal conductivity of single-layer graphene, Nano Lett. 8 (2008)
7
902–907.
9
[17] A. A. Balandin, Thermal properties of graphene and nanostructured carbon
M AN U
8
SC
5
materials, Nat. Mater. 10 (2011) 569–581.
[18] S. Ghosh, W. Bao, D. L. Nika, S. Subrina, E. P. Pokatilov, C. N. Lau, et al.,
11
Dimensional crossover of thermal transport in few-layer graphene, Nat. Mater.
12
9 (2010) 555–558.
TE D
10
[19] S. Guo, S. Dong, Graphene nanosheet: synthesis, molecular engineering, thin
14
film, hybrids, and energy and analytical applications, Chem. Soc. Rev. 40
15
(2011) 2644–2672.
AC C
EP
13
16
[20] A. A. Green, M. C. Hersam, Solution phase production of graphene with
17
controlled thickness via density differentiation, Nano Lett. 9 (2009) 4031–
18
4036.
19
[21] M. Lotya, Y. Hernandez, P. J. King, R. J. Smith, V. Nicolosi, L. S. Karlsson, et
20
al., Liquid phase production of graphene by exfoliation of graphite in 18
ACCEPTED MANUSCRIPT 1
surfactant/water solutions, J. Am. Chem. Soc. 131 (2009) 3611–3620. [22] K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, et al.,
3
Large-scale pattern growth of graphene films for stretchable transparent
4
electrodes, Nature 457 (2009) 706–710.
RI PT
2
[23] X. Li, G. Zhang, X. Bai, X. Sun, X. Wang, E. Wang, et al., Highly conducting
6
graphene sheets and Langmuir-Blodgett films, Nat. Nanotechnol. 3 (2008)
7
538–542.
SC
5
[24] A. Amiri, G. Ahmadi, M. Shanbedi, M. Savari, S. N. Kazi, B. T. Chew,
9
Microwave-Assisted Synthesis of Highly-Crumpled, Few-Layered Graphene
10
and Nitrogen-Doped Graphene for Use as High-Performance Electrodes in
11
Capacitive Deionization, Sci. Rep. 5 (2015) 17503. [25] A.
V.
TE D
12
M AN U
8
Murugan,
T.
Muraliganth, of
Manthiram,
graphene
Rapid,
nanosheets
and
facile
13
microwave-solvothermal
their
14
polyaniline nanocomposites for energy storage, Chem. Mater. 21 (2009) 5004–
15
5006.
AC C
EP
synthesis
A.
16
[26] J. Lin, H. Zhang, H. Hong, H. Liu, X. Zhang, A thermally conductive
17
composite with a silica gel matrix and carbon-encapsulated copper
18
nanoparticles as filler, J. Electron. Mater. 43 (2014) 2759–2769.
19
[27] S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y.
20
Jia, et al., Synthesis of graphene-based nanosheets via chemical reduction of 19
ACCEPTED MANUSCRIPT 1
2
exfoliated graphite oxide, Carbon 45 (2007) 1558–1565. [28] I. Ahmad, M. Islam, H. S. Abdo, T. Subhani, K. A. Khalil, A. A. Almajid, et al., Toughening
4
nanosheet-reinforced alumina, Mater. Des. 88 (2015) 1234–1243.
6
and
mechanical
properties
of
graphene
[29] T. T. Baby, S. Ramaprabhu, Investigation of thermal and electrical conductivity of graphene based nanofluids, J. Appl. Phys. 108 (2010) 124308.
SC
5
mechanisms
RI PT
3
[30] G. Wang, B. Wang, J. Park, Y. Wang, B. Sun, J. Yao, Highly efficient and
8
large-scale synthesis of graphene by electrolytic exfoliation, Carbon 47 (2009)
9
3242–3246.
M AN U
7
[31] J. I. Langford, A. J. C. Wilson, Scherrer after sixty years: A survey and some
11
new results in the determination of crystallite size, J. Appl. Crystallogr. 11
12
(1978) 102–113.
TE D
10
[32] A. Kaniyoor, T. T. Baby, S. Ramaprabhu, Graphene synthesis via hydrogen
14
induced low temperature exfoliation of graphite oxide, J. Mater. Chem. 20
15
(2010) 8467–8469.
AC C
EP
13
16
[33] A. Ashori, S. Menbari, R. Bahrami, Mechanical and thermo-mechanical
17
properties of short carbon fiber reinforced polypropylene composites using
18
exfoliated graphene nanoplatelets coating, J. Ind. Eng. Chem. (2016) 6–11.
19
[34] X. Zhang, H. Zhang, J. Lin, L. Li, Q. Wu, Thermal conductivity and thermal
20
stability enhancement of ethylene propylene diene methylene with carbon 20
ACCEPTED MANUSCRIPT 1
nanotube, J. Reinf. Plast. Compos. 33 (2014) 767–774. [35] Y. A. Kim, T. Hayashi, M. Endo, Y. Gotoh, N. Wada, J. Seiyama, Fabrication
3
of aligned carbon nanotube-filled rubber composite, Scr. Mater. 54 (2006) 31–
4
35.
RI PT
2
[36] D.L. Nika, E.P. Pokatilov, A.S. Askerov, A.A. Balandin, Phonon thermal
6
conduction in graphene: Role of Umklapp and edge roughness scattering, Phys.
7
Rev. B 79 (2009) 155413.
9
[37] A. V. Savin, Y. S. Kivshar, B. Hu, Suppression of thermal conductivity in
M AN U
8
SC
5
graphene nanoribbons with rough edges, Phys. Rev. B 82 (2010) 2821–2829. [38] A. Yu, P. Ramesh, M. E. Itkis, E. Bekyarova, R. C. Haddon, Graphite
11
nanoplatelet-epoxy composite thermal interface materials, J. Phys. Chem. C
12
111 (2007) 7565–7569.
TE D
10
[39] P. Ding, S. Su, N. Song, S. Tang, Y. Liu, L. Shi, Highly thermal conductive
14
composites with polyamide-6 covalently-grafted graphene by an in situ
15
polymerization and thermal reduction process, Carbon 66 (2014) 576–584.
17
AC C
16
EP
13
21
ACCEPTED MANUSCRIPT Table 1. Thermal conductivity of TIMs with metal, carbon material fillers
Filler
Fraction
Thermal Conductivity
Reference
Epoxy
Sliver
5 vol.%
1.67 W/(m K)
Ref. [9]
Rubber
CNTs
5 wt.%
0.70 W/(m K)
Epoxy
Graphite
5 vol.%
1.45 W/(m K)
RI PT
Matrix
Ref. [35]
Ref. [38]
SC
1
M AN U
Nanosheets
polyamide-6
GO
10 wt.%
0.416 W/(m K)
Ref. [39]
Silicone
GNs
1.5 w.%
2.75 W/(m K)
This work
AC C
EP
TE D
2
3 4
Scheme 1. Visual image of graphene synthesis process.
5
22
ACCEPTED MANUSCRIPT CH3 H2 C
CH
Si
CH3 O
CH3
1
Si
CH3 O
CH3
Si
CH
CH3
3
H3C
Si
O
CH3
4
H
O
Si
CH3
CH3
Si
CH
CH2 +
H
M AN U
6
Si
Si
CH2
CH2
Si
TE D
Scheme 4. The reaction mechanism of polymerization.
AC C
EP
9
Si
CH3
Scheme 3. Molecular structure of methyl hydrogenous silicone oil.
5
8
H
SC
CH3
RI PT
Scheme 2. Molecular structure of methyl vinyl silicone oil.
2
7
CH2
10 11
Scheme 5. Thermal dissipated structure of high-power LED.
12
23
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
1
2
AC C
EP
4
Figure 1. The low (a) and high (b, c) magnifications of SEM and TEM (d) images of GNs.
TE D
3
5 6
Figure 2. Results of the XRD measurements performed on the NG, GO and GNs. 24
SC
RI PT
ACCEPTED MANUSCRIPT
1
Figure 3. FTIR of GO and GNs.
M AN U
2
4
AC C
EP
TE D
3
5
Figure 4. SEM images of the GNs/silicone grease composites with loading of (a) 0.3 wt.%; (b) 0.5
6
wt.%; (c) 1.0 wt.%; (d) 1.5 wt.% and (e, f) The optical photograph of the composites.
7
25
SC
RI PT
ACCEPTED MANUSCRIPT
1
Figure 5. DSC curve of silicone grease filled with 1.0wt.% GNs and neat silicone grease.
M AN U
2
4 5
AC C
EP
TE D
3
Figure 6. TGA of silicone grease filled with different loading of GNs.
6
26
SC
RI PT
ACCEPTED MANUSCRIPT
1
Figure 7. Thermal conductivity of silicone filled with different amount of GNs.
M AN U
2
5 6 7
Figure 8. (a) The measured mean temperatures of the heat slug (Ti) and heat sink (Tj); (b) The
AC C
4
EP
TE D
3
relationship of thermal conductivity of TIMs and temperature gap.
27
ACCEPTED MANUSCRIPT Highlights:
1. An microwave synthesis approach was used to scale graphene nanosheets. 2.
The thermal conductivity of TIM can be significantly enhanced in a low loading.
AC C
EP
TE D
M AN U
SC
RI PT
3. The as-prepared TIMs can efficiently decrease the temperature of LED chips.
S1567-1739(16)30275-9
DOI:
10.1016/j.cap.2016.10.004
Reference:
CAP 4339
To appear in:
Current Applied Physics
Received Date: 7 July 2016 Revised Date:
5 September 2016
Accepted Date: 4 October 2016
Please cite this article as: H. Zhang, Y. Lin, D. Zhang, W. Wang, Y. Xing, J. Lin, H. Hong, C. Li, Graphene nanosheet/silicone composite with enhanced thermal conductivity and its application in heat dissipation of high-power light-emitting diodes, Current Applied Physics (2016), doi: 10.1016/ j.cap.2016.10.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Graphene nanosheet/silicone composite with
2
enhanced thermal conductivity and its application in
3
heat dissipation of high-power light-emitting diodes
4
Haiyan Zhang,b,*, Yingxi Lin a, Danfeng Zhang a, Wenguang Wang a,b, Yuxiong Xing a,
5
Jin Lina, Haoqun Honga, Chunhui Lia
RI PT
1
6
a
7
China
8
b
9
University of Technology, Guangzhou, 510006, China
M AN U
SC
School of Materials and Energy, Guangdong University of Technology, Guangzhou, 510006, P. R.
Guangdong Provincial Key Laboratory of Functional Soft Condensed Matter, Guangdong
AC C
EP
TE D
10
*Corresponding author. E-mail address: [email protected] 1
ACCEPTED MANUSCRIPT
Abstract: Heat dissipation from light-emitting diodes (LEDs) has become a serious
2
problem because of the LEDs’ high luminosity and power. This problem could be
3
solved by improving the dissipation efficiency of each LED system component. A
4
microwave-reduced graphene nanosheet/silicone (GN/silicone) composite with a high
5
thermal conductivity and stability was prepared by mechanical blending. The thermal
6
conductivity of the composite reaches 2.7 W/(m K) with only 1.5 wt.% loading, and is
7
12 times higher than the pure silicone matrix. When used as a thermal interface
8
material between high-power LED chip module substrates and heat sinks, the thermal
9
conductive GN/silicone composite could decrease the temperature difference between
10
the substrate and shell. It could also improve the system heat transfer efficiency. The
11
temperature gap between the heat slug and the heat sink was less than 2°C with a 1.5
12
wt.% loading of GNs.
M AN U
SC
RI PT
1
14
TE D
13
KEYWORDS: Graphene nanosheet; thermal interface material; LED dissipation;
16
thermal conductivity.
AC C
17
EP
15
2
ACCEPTED MANUSCRIPT 1
1 Introduction Light-emitting diodes (LEDs) are two-lead semiconductor p–n junction light
3
sources that emit light when activated. Because they do not radiate at ultraviolet or
4
infrared wavelengths, they provide several advantages compared with other light
5
sources, such as a long service lifetime, a high luminaire efficiency and they are
6
environmentally friendly [1]. However, LEDs convert 75–85% of their electrical
7
power into heat, which reduces their efficiency and makes them less reliable. The
8
problem is more serious for high-power LEDs, which can be driven at currents from
9
hundreds of milliamperes to more than an ampere, compared with the tens of
SC
milliamperes for other LEDs [2].
M AN U
10
RI PT
2
Therefore, it is important to design an efficient thermal system within the LEDs.
12
One method to do so is to improve the heat transfer efficiency between the LED heat
13
slugs and sinks. Much work has been done to increase the efficiency by optimizing
14
the heat sink [3], adjusting the refrigerant in heat pipes [4] and reducing the heat
15
resistance between the LED chips and heat sinks [5]. Among these methods,
16
improving the heat transfer efficiency with thermal interface materials (TIMs)
17
between LEDs and heat sinks is seen as an inexpensive, simple way which can be
18
used in large-scale industrial production.
EP
AC C
19
TE D
11
In the practical application of high-power LEDs, LED chip modules are arranged in
20
arrays and are welded on a specialized substrate. When the substrates are clamped to
21
the heat sinks, microscopic surface undulations that form air gaps can hinder heat
22
transfer and decrease the heat transfer efficiency substantially [6]. Filling the air gap
23
with soft thermal conductive TIMs can facilitate heat transfer, because the thermal
24
conductivity of TIMs is usually two orders larger than air (~0.025 W/(m K) at room 3
ACCEPTED MANUSCRIPT 1
temperature) [7]. In general, widely used TIMs are made of polymer matrices with the addition of
3
high thermal-conductivity materials, such as metals (e.g., silver, copper), ceramics
4
(e.g., aluminium oxide, zinc oxide) or carbon materials (e.g., graphite, carbon
5
nanotubes). Metal particles can be oxidized so easily at high temperature that the heat
6
conducting property worsens. In addition, these thermal conductive particles require
7
high-volume fractions of filler to reach thermal conductivity of the composite from
8
~1–5 W/(m K) at room temperature, which may lead to poor composite mechanical
9
properties [8,9].
SC
RI PT
2
Carbon nanotubes (CNTs) have attracted interest as fillers for TIMs because of
11
their extremely high intrinsic thermal conductivity from ~3000–3500 W/(m K) at
12
room temperature [10]. However, their addition does not improve the thermal
13
conductivity of the matrix as expected or decrease with the addition of single-wall
14
CNTs [11]. It is believed that CNTs with an excellent thermal conductivity do not
15
couple well to the matrix material or contact surface. Therefore, there is optimism in
16
the search for alternative high thermal-conductivity fillers.
TE D
M AN U
10
In recent years, graphene has attracted considerable attention in various fields
18
because of its excellent mechanical, electric, and thermal performance [12–15].
19
Balandin et al. measured the thermal conductivity (λ) of single-layer graphene by
20
laser and found that the λ is much higher than most metals at 4800–5300 W/(m K)
21
[16]. This result is attributed to a long free path for phonon transport, which is by
22
ion-core vibration in a crystal lattice [17,18].
AC C
EP
17
23
High-quality and large-scale graphene production is required for use in TIMs.
24
Several methods have been developed to solve these problems, including epitaxial
4
ACCEPTED MANUSCRIPT 1
growth, chemical vapor deposition, and chemical exfoliation [19]. Of these methods,
2
chemical exfoliation of graphite oxide is seen as a simple and rapid way to obtain
3
graphene at a large scale [20–23]. In this paper, an industrial scalable, low-cost, and simple approach was used to
5
prepare graphene. This method produced graphene by exfoliating oxide graphite into
6
small, thin pieces by using microwave radiation [24,25]. Microwave-reduced
7
graphene nanosheets (GNs) were added to a silicone matrix to prepare
8
microwave-reduced GN/silicone composites, which was used as a thermal interface
9
material in the heat dissipation of high-power LEDs.
2. Experimental
11
2.1 Preparation of GNs
M AN U
10
SC
RI PT
4
As shown in Scheme 1, graphite oxide (GO) was prepared as follows. Firstly, 184
13
ml of concentrated H2SO4 was added to 4 g of NaNO3 in a 1000 ml conical flask at
14
0°C in a thermostatic water bath. The mixture was stirred continuously to form a
15
homogeneous solution. Then, 8 g of flake graphite and 24 g of KMnO4 were added
16
into the above solution, and the mixture was stirred at 0°C for 2 h. After that, the
17
temperature of the water in the thermostatic water bath was increased to 35°C and the
18
above mixed solution was kept stirring for 2 h. Afterwards, the thermostatic water
19
bath temperature was set to 75°C . Deionized water was added slowly into the conical
20
flask during the heating process. When the temperature reached 75°C, a certain
21
amount of H2O2 was added into the mixture to remove residual KMnO4. The conical
22
flask was removed and cooled naturally to room temperature. The resulting
AC C
EP
TE D
12
5
ACCEPTED MANUSCRIPT 1
golden-yellow mixed solution was washed with deionized water several times to pH 7.
2
Finally, graphite oxide was obtained after centrifugation and drying in a vacuum
3
drying oven at 60°C. The as-prepared GO was ground in a planetary ball mill for 10 h and then heated in
5
a microwave oven at 1000 W in a nitrogen atmosphere. When the temperature reached
6
300°C, the microwave oven was turned off. Black inflated GNs were obtained.
7
2.2 GN modification
SC
RI PT
4
To improve the compatibility of the GNs with the matrix, surface modification was
9
carried out by 3-aminopropyltriethoxysilane (KH-550, Kang Lan Biochemical Pte.,
10
Ltd., Shanghai, China) silane conpling agent addition. The as-prepared black powder
11
was added into a 500 ml solution that contained alcohol and 1.5 wt.% KH-550. The
12
mixed solution was sonicated for 30 min. The solution was filtrated and the
13
pre-treated GNs were dried at 120°C in a drying oven.
14
2.3 Preparation of GN/silicone composite
EP
TE D
M AN U
8
Silicone gel (Chun Chang Chemical Pte., Ltd., Shenzhen, China) consisted of
16
component A, a methyl vinyl silicone oil with the molecular structure as shown in
17
Scheme 2, and curing agent component B, a methyl hydrogenous silicone oil with the
18
molecular structure as shown in Scheme 3.
19 20
AC C
15
Polymerization was initiated by chloroplatinic acid addition. The reaction mechanism involved hydrogen addition to vinyl and is illustrated in Scheme 4.
6
ACCEPTED MANUSCRIPT Organic silicone gel was chosen as a polymeric matrix in this work for the
2
following reasons: (1) good chemical inertness, (2) wide service temperature range
3
from 5.5°C to 260°C, (3) outstanding electrical performance and weathering
4
resistance, (4) advantageous conformability, (5) ease of processability, (6) adjustable
5
hardness; (7) and low costs [26].
RI PT
1
Silicone A (10 g) was mixed with GNs homogeneously using an agate pestle in a 50
7
ml agate mortar. Silicone B (10 g) was added and blended uniformly by using
8
double-roll milling at room temperature for 15–20 min. The mixture was transferred
9
to a 15 cm diameter glass dish and placed in a vacuum oven for degassing. To
10
complete the curing reaction, curing at 150°C for 1 h and at 200°C for an additional 2
11
h was applied, after which the composite was peeled off easily from the glass dish. A
12
series of composites was prepared with different GN weight fractions of 0.1 wt.%, 0.3
13
wt.%, 0.5 wt.%, 1 wt.%, and 1.5 wt.%. A pristine silicone sample was produced for
14
comparison.
15
2.4 Characterization and measurements
M AN U
TE D
EP
AC C
16
SC
6
Transmission electron microscopy (TEM) was performed at an accelerating voltage
17
of 200 kV, using a JEOL JEM-2100. Micrographs were also obtained by using an
18
S-3400N-II scanning electron microscope (SEM).
19 20
Using CuKα radiation, X-ray diffraction (XRD) measurements were conducted at a scan rate of 8° min-1 in a D/Max-IV (Rigaku Co., Tokyo, Japan) at 40 kV and 30 mA.
7
ACCEPTED MANUSCRIPT Functional groups on the surface of GN were identified via Fourier transform
2
infrared (FTIR) spectroscopy (Thermo Scientific, Nicolet 6700). Thermogravimetric
3
analysis (TGA) was used to evaluate the thermal stabilities of the composite; this
4
analysis was performed in a nitrogen atmosphere and at 10°C min-1, using an SDT
5
2960 Simultaneous DSC-TGA (TA Instrument, USA). The thermal conductivities of
6
the composite were measured by using a thermal-conductivity meter LFA-440
7
(Netzsch, Germany), based on the laser flash method in which the front side of a
8
plane-parallel sample is heated by using a short light pulse and the temperature
9
increase on the rear surface was measured using an infrared detector.
10
3. Results and discussion
11
3.1 GN morphology and structure
M AN U
SC
RI PT
1
For the large-scale synthesis of graphene, graphite remains the most popular
13
precursor because of its higher yield and overall process cost. The general route for
14
graphene synthesis from graphite is oxidation to graphite oxide, followed by
15
exfoliation of the graphite oxide to graphene. Thermal exfoliation is the most
16
common method, and thermal degradation of the oxide functional groups results in an
17
evolution of substantial quantities of oxide and carbon dioxide, which can result in the
18
enhanced expansion of graphite oxide along the thickness direction. In general,
19
graphene is derived from GO, which can be reduced to graphene via direct thermal
20
treatment at elevated temperatures without using hazardous reducing agents. To
21
obtain completely exfoliated and thermal conductive graphene, “thermal shock” of
22
GO at temperatures of
AC C
EP
TE D
12
up to 1050°C was used. However, such a high temperatue 8
ACCEPTED MANUSCRIPT 1
lead to a low production efficiency and high cost, which wastes energy.
Therefore, microwave irradiation assisted synthesis is an ideal way to prepare
3
graphene on a large scale, and it is also an economic, environmentally friendly and
4
efficient synthesis technique. Compared with conventional heat treatment, microwave
5
irradiation can be used to synthesize graphene via superior “inert and instant heating”.
6
The GO volume can be reduced completely by the rapid, uniform and energy-efficient
7
microwave irradiation process.
SC
RI PT
2
SEM images show that the GNs consist of randomly aggregated, crumpled sheets
9
(Figure 1a–c). The folded sheet regions in Figure 1c had average widths of several
10
nanometers [27,28]. The nanosheet thickness as estimated from the edge of a single
11
GN (Figure 1a) is 1–5 nm, and it appears to be of a similar thickness compared with
12
Ahmad et al.’s report [28]. This result was confirmed by transmission electron
13
microscopy measurements. Figure 1d shows a large transparent, wrinkled
14
graphene-like area that is caused by a high specific surface. As reported in the
15
literature [29,30], this crimping results from thin walls that consist of a few layers of
16
graphene, and that corresponds to a 1–5 nm thickness.
TE D
EP
AC C
17
M AN U
8
The chemical structure of natural graphite (NG), GO, and GNs was determined via
18
XRD and FTIR spectroscopy. As shown in XRD (Figure 2), the strong NG diffraction
19
peak at 2θ = 26°, corresponds to the (002) reflection, which indicates a high degree
20
of graphitization of the NG. However, GNs shows a broad peak at 25°, and such a 9
ACCEPTED MANUSCRIPT broad curve is characteristic of amorphous material and indicates the successful
2
reduction of GO. According to the Scherrer equation [31], the interlayer gaps of NG,
3
GO and GN are 0.339 nm, 0.795 nm and 0.344 nm, respectively. These results
4
manifest that the interlayer gaps of GNs become larger than those of NG after
5
reduction, which suggests incomplete reduction and the existence of residual
6
functional groups. The incompletely reduced GNs can be dispersed better in a
7
polymer matrix with the help of polar functional groups, such as hydroxyl groups.
SC
RI PT
1
The FT-IR spectra of the GO and GNs are shown in Figure 3. The broad peak at
9
3400 cm-1 in the spectra is associated with OH groups. Significant peaks at 1382 cm-1
10
and 1216 cm-1 are attributed to carboxy C-O and epoxy C-O groups [32]. The peak at
11
1618 cm-1 arises from the existence of aromatic C=C. However, the peak at 1716 cm-1,
12
which is associated with a COOH group (C=O vibration), does not appear for the GNs.
13
It can be inferred that the GNs were partly reduced, so some functional groups of GO
14
remained. This result implies that GNs with functional groups are compatible with the
15
polymer matrix as indicated by the XRD results.
16
3.2 Dispersion of GNs in TIM
TE D
EP
AC C
17
M AN U
8
GNs have large specific surface areas, and a high surface energy, which causes
18
them to tend to aggregate. In double-roll milling, the surface between the materials
19
and the rollers was of a different velocity. This velocity initiated an internal shear
20
stress that can be controlled by adjusting the amount of materials and the distances 10
ACCEPTED MANUSCRIPT 1
between two rolls. By using KH-550, we could obtain a well-dispersed composite by mixing for 15–
3
20 min. As the SEM images show in Figure 4 (a–d), the GNs are dispersed in silicone
4
and are compatible with silicone. The GN remains as a sheet (red line) in the matrix
5
without any agglomeration. If the GNs are agglomerated, the agglomerate bulk could
6
be found
7
Shape remaining GNs were found even at a high loading of 1.5 wt.% (Figure d). It is
8
believed that the GNs disperse well in the matrix. The remaining sheet shape favors
9
the formation of conductivity chains because of the high aspect ratio of the GNs. The
10
silicone becomes black with increase in GNs (Figure 4e and f), and no obvious
11
bubbles, cracks or other defects are visible in the composite, which may hinder heat
12
transport in the composite. The composite size can be controlled by roll size. As a
13
result, we produced a composite of as large as 18 cm × 18 cm. An even larger
14
composite can be produced by using a suitable roll size, and large-scale production is
15
possible, which is significant in industrialization.
16
3.3 TIM thermal stability
EP
TE D
M AN U
SC
easily as sphere-like shape within the matrix as reported by Alireza [33].
AC C
17
RI PT
2
The DSC curves of a pure silicone and a GN/silicone composite with 1.0 wt.% GNs
18
is visible in Figure 5. For a sample of pure silicone, several strong endothermic peaks
19
appeared between 300°C and 600°C, which indicates that silicone decomposition
20
occurred, whereas the composite did not show significant endothermic peaks at the
21
same temperature. Therefore, composites could remain thermally stable at the same
11
ACCEPTED MANUSCRIPT 1 2
temperature at which pure silicone decomposed. The TGA results are shown in Figure 6. The composite possesses a good thermal stability compared with pure silicone. Firstly, there is no obvious mass loss (3–5 wt.%)
4
below 300°C for samples. The pure silicone (black line) began to lose mass at 300°C,
5
when the composites decomposing temperature were increasing gradually with fillers
6
addition increasing, which was 100°C higher than that of pure silicone at the loading
7
of 1.0 wt.% (~400°C). Secondly, the remainder of the composite at the loading of 1.0
8
wt.% is 30 wt.% higher than the silicone, which is also larger than the GN addition.
SC
RI PT
3
These results indicate that GNs in the composite can interconnect to form a
10
network frame, which serves as a frame and restricts silicone molecule movement.
11
The GN high specific surface area results in an increase in contact areas between the
12
GNs and the silicone. Therefore, the surface bond strength of these two materials can
13
be enhanced significantly. The interaction between GNs and silicone hinders
14
molecular segment movement in the composite [34].
15
3.4 TIM thermal conductivities
TE D
M AN U
9
The composite thermal conductivity increased with increase in filler content
17
(Figure 7). The error bars represent data scatter for several samples. The thermal
18
conductivity of pure silicone is 0.22 W/(m K). However, the composite could have a
19
thermal conductivity of 1.51 W/(m K) at a loading of 0.5 wt.% GNs. With the
20
addition of 1.5 wt.% GNs, the composite thermal conductivity reached 2.75 W/(m K),
21
which is 13 times that of pure silicone. This result is much higher than that of other
22
composites. In addition, GNs display better filler performance than for other kinds of
23
carbon materials, such as carbon nanotubes or GO [35]; this can be attributed to the
24
high quality of GNs prepared by microwave reduction, which damages the GNs less.
AC C
EP
16
12
ACCEPTED MANUSCRIPT The enhancement ratio varies approximately linearly with concentration. However,
2
when the filler exceeds 0.5 wt.%, the thermal conductivity increases rapidly, which
3
suggests that a heat transport chain has formed. This heat transport chain benefits
4
phonon transport, which is the main way to convey energy in the composite.
5
Therefore, with increase in fillers, the formed heat transport chain could increase the
6
thermal conductivity suddenly. GNs with a high aspect ratio, can form heat transport
7
chain more easily compared with other carbon materials [36,37].
SC
RI PT
1
Table 1 lists the thermal conductivities of several TIMs with various fillers,
9
including silver, CNTs, and GO. Table 1 shows that the thermal conductivity of the
10
reference TIMs is not enhanced obviously even for a high loading. In addition, such a
11
high loading can influence the machine performance of the TIMs. In this case, the
12
TIM with GNs is advantageous when applied in an electronic device.
13
3.5 Application in an LED
TE D
M AN U
8
As shown in Scheme 5, the TIMs prepared above are used in a high-power LED
15
system to improve the heat transport efficiency between the heat slug and sink. The
16
composite that is applied in a high-power LED improves heat dissipation, and reduces
17
the temperature of the LED chip module.
c
AC C
EP
14
18
The histogram in Figure 8(a) shows the temperature of the heat slug and sink in
19
high-power LEDs. Ti represent the mean temperature of the heat slug, and Tj is the
20
mean temperature of the heat sink, as shown in Scheme 5.
21
When no TIM exists, Ti and Tj were 55 and 35°C, respectively. The temperature
22
gap between Ti and Tj is 20°C, which confirms that heat transported with difficulty
23
from the LED module to a heat sink because of the poor thermal conductivity of air 13
ACCEPTED MANUSCRIPT 1
(~0.025 W/(m K)). However, this difference can be reduced with the thermal
2
conductive augmentation of TIMs. When the thermal conductivity of the TIM reached
3
2.7 W/(m K), the temperature difference (Ti = 37°C, Tj = 35°C) reached only 2°C. The result is more obvious from Figure 8(b), which shows the relationship between
5
thermal conductivity and temperature gap. With increasing TIM thermal conductivity,
6
the temperature of the heat dissipation base chip module (heat slug) Ti decreases
7
significantly. When the thermal conductivity of the TIMs is 1.5 W/(m K) (0.5 wt.%
8
loading), the temperature difference is less than 5°C. It can be inferred that the
9
temperature difference decreases and that the heat dissipation improves in a
SC
high-power LED with an increase in thermal conductivity.
M AN U
10
RI PT
4
Therefore, it can be inferred that the temperature difference between Ti and Tj
12
decreases with increasing TIM thermal conductivity, which suggests an enhancement
13
in heat dissipation. Therefore, using high thermal conductive TIMs can decrease the
14
LED chip temperature, which favors their lifetime and performance.
15
4. Conclusion
TE D
11
GNs with a suitable dispersibility were synthesized by a microwave-reduced
17
method, which has potential for application in industry. GN/silicone composites were
18
prepared by mechanical blending. Compared with pure silicone, the addition of GNs
19
delays the initial composite decomposition temperature, and results in an outstanding
20
stability (the initial decomposition temperature of the composites is 100°C higher than
21
that of pure silicone). The thermal conductivity of the TIMs increase with increase in
22
fillers. When the GN loading is 1.5 wt.%, the maximum thermal conductivity of the
23
composite reaches 2.7 W/(m K). Such high conductive materials are helpful when
24
used as TIMs in high power LED systems. The thermal conductivity of 1.5 W/(m K)
AC C
EP
16
14
ACCEPTED MANUSCRIPT 1
can reduce the temperature gap between the LED module and the heat sink to 5°C.
2
Acknowledgments This work is supported by the link project of the National Natural Science
4
Foundation of China and Guangdong Province (No. U1401246), by the National
5
Natural Science Foundation of China (Grant No. 51276044), by the Science and
6
Technology Program of Guangdong Province of China (Grant No. 2014B010106005,
7
2015B010135011, 2015A050502047), and by the Science and Technology Program
8
of Guangzhou City of China (Grant No. 201508030018), by Research Fund of Young
9
Teachers for the Doctoral Program of Higher Education of China (20134420120009),
10
and by Special Program for Public Interest Research and Capability Construction of
11
Guangdong (2014A010105047)
AC C
EP
TE D
12
M AN U
SC
RI PT
3
15
ACCEPTED MANUSCRIPT 1
Reference
2
[1]
B. L. Ahn, C. Y. Jang, S. B. Leigh, S. Yoo, H. Jeong, Effect of LED lighting on the cooling and heating loads in office buildings, Appl. Energ. 113 (2014)
4
1484–1489.
5
[2]
RI PT
3
E. D. Jung, Y. L. Lee, Development of a heat dissipating LED headlamp with silicone lens to replace halogen bulbs in used cars, Appl. Therm. Eng. 86 (2015)
7
143–150.
heat sinks with application to LED light bulbs, Int. J. Heat Mass Transf. 71
9
(2014) 496–502.
10
11
D. Jang, S. J. Park, S. J. Yook, K. S. Lee, The orientation effect for cylindrical
M AN U
[3]
[4]
J. Wang, Y. Cai, X. Zhao, C. Zhang, Thermal design and simulation of
TE D
8
SC
6
automotive headlamps using white LEDs, Microelectronics J. 45 (2014) 249–
13
255.
14
[5]
AC C
17
53 (2013) 428–434.
[6]
J. Due, A. J. Robinson, Reliability of thermal interface materials: A review,
Appl. Therm. Eng. 50 (2013) 455–463.
18
19
M. Maaspuro, A. Tuominen, Thermal analysis of LED spot lighting device operating in external natural or forced heat convection, Microelectron. Reliab.
15 16
EP
12
[7]
R. B. Montgomery, Viscosity and thermal conductivity of air and diffusivity of
16
ACCEPTED MANUSCRIPT water vapor in air, J. Meteorol. 4 (1947) 193–196.
1
2
[8]
L. C. Sim, S. R. Ramanan, H. Ismail, K. N. Seetharamu, T. J. Goh, Thermal characterization of Al2O3 and ZnO reinforced silicone rubber as thermal pads
4
for heat dissipation purposes, Thermochim. Acta. 430 (2005) 155–165.
5
[9]
RI PT
3
K. M. F. Shahil, A. A. Balandin, Thermal properties of graphene and multilayer graphene: Applications in thermal interface materials, Solid State
7
Commun. 152 (2012) 1331–1340.
SC
6
[10] E. Pop, D. Mann, Q. Wang, K. Goodson, H. Dai, Thermal conductance of an
9
individual single-wall carbon nanotube above room temperature, Nano Lett. 6
10
M AN U
8
(2006) 96–100.
[11] S. U. S. Choi, Z. G. Zhang, W. Yu, F. E. Lockwood, E. A. Grulke, Anomalous
12
thermal conductivity enhancement in nanotube suspensions, Appl. Phys. Lett.
13
79 (2001) 2252–2254.
EP
15
[12] S. Park, R. S. Ruoff, M. Engineering, Chemical methods for the production of graphenes, Nat. Nanotechnol. 4 (2009) 217–224.
AC C
14
TE D
11
16
[13] Y. Q. Bie, J. Horng, Z. Shi, L. Ju, Q. Zhou, A. Zettl, et al., Vibrational
17
spectroscopy at electrolyte/electrode interfaces with graphene gratings, Nat.
18
Commun. 6 (2015) 7593.
19
[14] A. Varykhalov, J. Sánchez-Barriga, D. Marchenko, P. Hlawenka, P. S. Mandal,
20
O. Rader, Tunable Fermi level and hedgehog spin texture in gapped graphene, 17
ACCEPTED MANUSCRIPT 1
Nat. Commun. 6 (2015) 7610. [15] G. Zhou, E. Paek, G. S. Hwang, A. Manthiram, Long-life Li/polysulphide
3
batteries with high sulphur loading enabled by lightweight three-dimensional
4
nitrogen/sulphur-codoped graphene sponge, Nat. Commun. 6 (2015) 7760.
RI PT
2
[16] A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, et al.,
6
Superior thermal conductivity of single-layer graphene, Nano Lett. 8 (2008)
7
902–907.
9
[17] A. A. Balandin, Thermal properties of graphene and nanostructured carbon
M AN U
8
SC
5
materials, Nat. Mater. 10 (2011) 569–581.
[18] S. Ghosh, W. Bao, D. L. Nika, S. Subrina, E. P. Pokatilov, C. N. Lau, et al.,
11
Dimensional crossover of thermal transport in few-layer graphene, Nat. Mater.
12
9 (2010) 555–558.
TE D
10
[19] S. Guo, S. Dong, Graphene nanosheet: synthesis, molecular engineering, thin
14
film, hybrids, and energy and analytical applications, Chem. Soc. Rev. 40
15
(2011) 2644–2672.
AC C
EP
13
16
[20] A. A. Green, M. C. Hersam, Solution phase production of graphene with
17
controlled thickness via density differentiation, Nano Lett. 9 (2009) 4031–
18
4036.
19
[21] M. Lotya, Y. Hernandez, P. J. King, R. J. Smith, V. Nicolosi, L. S. Karlsson, et
20
al., Liquid phase production of graphene by exfoliation of graphite in 18
ACCEPTED MANUSCRIPT 1
surfactant/water solutions, J. Am. Chem. Soc. 131 (2009) 3611–3620. [22] K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, et al.,
3
Large-scale pattern growth of graphene films for stretchable transparent
4
electrodes, Nature 457 (2009) 706–710.
RI PT
2
[23] X. Li, G. Zhang, X. Bai, X. Sun, X. Wang, E. Wang, et al., Highly conducting
6
graphene sheets and Langmuir-Blodgett films, Nat. Nanotechnol. 3 (2008)
7
538–542.
SC
5
[24] A. Amiri, G. Ahmadi, M. Shanbedi, M. Savari, S. N. Kazi, B. T. Chew,
9
Microwave-Assisted Synthesis of Highly-Crumpled, Few-Layered Graphene
10
and Nitrogen-Doped Graphene for Use as High-Performance Electrodes in
11
Capacitive Deionization, Sci. Rep. 5 (2015) 17503. [25] A.
V.
TE D
12
M AN U
8
Murugan,
T.
Muraliganth, of
Manthiram,
graphene
Rapid,
nanosheets
and
facile
13
microwave-solvothermal
their
14
polyaniline nanocomposites for energy storage, Chem. Mater. 21 (2009) 5004–
15
5006.
AC C
EP
synthesis
A.
16
[26] J. Lin, H. Zhang, H. Hong, H. Liu, X. Zhang, A thermally conductive
17
composite with a silica gel matrix and carbon-encapsulated copper
18
nanoparticles as filler, J. Electron. Mater. 43 (2014) 2759–2769.
19
[27] S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y.
20
Jia, et al., Synthesis of graphene-based nanosheets via chemical reduction of 19
ACCEPTED MANUSCRIPT 1
2
exfoliated graphite oxide, Carbon 45 (2007) 1558–1565. [28] I. Ahmad, M. Islam, H. S. Abdo, T. Subhani, K. A. Khalil, A. A. Almajid, et al., Toughening
4
nanosheet-reinforced alumina, Mater. Des. 88 (2015) 1234–1243.
6
and
mechanical
properties
of
graphene
[29] T. T. Baby, S. Ramaprabhu, Investigation of thermal and electrical conductivity of graphene based nanofluids, J. Appl. Phys. 108 (2010) 124308.
SC
5
mechanisms
RI PT
3
[30] G. Wang, B. Wang, J. Park, Y. Wang, B. Sun, J. Yao, Highly efficient and
8
large-scale synthesis of graphene by electrolytic exfoliation, Carbon 47 (2009)
9
3242–3246.
M AN U
7
[31] J. I. Langford, A. J. C. Wilson, Scherrer after sixty years: A survey and some
11
new results in the determination of crystallite size, J. Appl. Crystallogr. 11
12
(1978) 102–113.
TE D
10
[32] A. Kaniyoor, T. T. Baby, S. Ramaprabhu, Graphene synthesis via hydrogen
14
induced low temperature exfoliation of graphite oxide, J. Mater. Chem. 20
15
(2010) 8467–8469.
AC C
EP
13
16
[33] A. Ashori, S. Menbari, R. Bahrami, Mechanical and thermo-mechanical
17
properties of short carbon fiber reinforced polypropylene composites using
18
exfoliated graphene nanoplatelets coating, J. Ind. Eng. Chem. (2016) 6–11.
19
[34] X. Zhang, H. Zhang, J. Lin, L. Li, Q. Wu, Thermal conductivity and thermal
20
stability enhancement of ethylene propylene diene methylene with carbon 20
ACCEPTED MANUSCRIPT 1
nanotube, J. Reinf. Plast. Compos. 33 (2014) 767–774. [35] Y. A. Kim, T. Hayashi, M. Endo, Y. Gotoh, N. Wada, J. Seiyama, Fabrication
3
of aligned carbon nanotube-filled rubber composite, Scr. Mater. 54 (2006) 31–
4
35.
RI PT
2
[36] D.L. Nika, E.P. Pokatilov, A.S. Askerov, A.A. Balandin, Phonon thermal
6
conduction in graphene: Role of Umklapp and edge roughness scattering, Phys.
7
Rev. B 79 (2009) 155413.
9
[37] A. V. Savin, Y. S. Kivshar, B. Hu, Suppression of thermal conductivity in
M AN U
8
SC
5
graphene nanoribbons with rough edges, Phys. Rev. B 82 (2010) 2821–2829. [38] A. Yu, P. Ramesh, M. E. Itkis, E. Bekyarova, R. C. Haddon, Graphite
11
nanoplatelet-epoxy composite thermal interface materials, J. Phys. Chem. C
12
111 (2007) 7565–7569.
TE D
10
[39] P. Ding, S. Su, N. Song, S. Tang, Y. Liu, L. Shi, Highly thermal conductive
14
composites with polyamide-6 covalently-grafted graphene by an in situ
15
polymerization and thermal reduction process, Carbon 66 (2014) 576–584.
17
AC C
16
EP
13
21
ACCEPTED MANUSCRIPT Table 1. Thermal conductivity of TIMs with metal, carbon material fillers
Filler
Fraction
Thermal Conductivity
Reference
Epoxy
Sliver
5 vol.%
1.67 W/(m K)
Ref. [9]
Rubber
CNTs
5 wt.%
0.70 W/(m K)
Epoxy
Graphite
5 vol.%
1.45 W/(m K)
RI PT
Matrix
Ref. [35]
Ref. [38]
SC
1
M AN U
Nanosheets
polyamide-6
GO
10 wt.%
0.416 W/(m K)
Ref. [39]
Silicone
GNs
1.5 w.%
2.75 W/(m K)
This work
AC C
EP
TE D
2
3 4
Scheme 1. Visual image of graphene synthesis process.
5
22
ACCEPTED MANUSCRIPT CH3 H2 C
CH
Si
CH3 O
CH3
1
Si
CH3 O
CH3
Si
CH
CH3
3
H3C
Si
O
CH3
4
H
O
Si
CH3
CH3
Si
CH
CH2 +
H
M AN U
6
Si
Si
CH2
CH2
Si
TE D
Scheme 4. The reaction mechanism of polymerization.
AC C
EP
9
Si
CH3
Scheme 3. Molecular structure of methyl hydrogenous silicone oil.
5
8
H
SC
CH3
RI PT
Scheme 2. Molecular structure of methyl vinyl silicone oil.
2
7
CH2
10 11
Scheme 5. Thermal dissipated structure of high-power LED.
12
23
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
1
2
AC C
EP
4
Figure 1. The low (a) and high (b, c) magnifications of SEM and TEM (d) images of GNs.
TE D
3
5 6
Figure 2. Results of the XRD measurements performed on the NG, GO and GNs. 24
SC
RI PT
ACCEPTED MANUSCRIPT
1
Figure 3. FTIR of GO and GNs.
M AN U
2
4
AC C
EP
TE D
3
5
Figure 4. SEM images of the GNs/silicone grease composites with loading of (a) 0.3 wt.%; (b) 0.5
6
wt.%; (c) 1.0 wt.%; (d) 1.5 wt.% and (e, f) The optical photograph of the composites.
7
25
SC
RI PT
ACCEPTED MANUSCRIPT
1
Figure 5. DSC curve of silicone grease filled with 1.0wt.% GNs and neat silicone grease.
M AN U
2
4 5
AC C
EP
TE D
3
Figure 6. TGA of silicone grease filled with different loading of GNs.
6
26
SC
RI PT
ACCEPTED MANUSCRIPT
1
Figure 7. Thermal conductivity of silicone filled with different amount of GNs.
M AN U
2
5 6 7
Figure 8. (a) The measured mean temperatures of the heat slug (Ti) and heat sink (Tj); (b) The
AC C
4
EP
TE D
3
relationship of thermal conductivity of TIMs and temperature gap.
27
ACCEPTED MANUSCRIPT Highlights:
1. An microwave synthesis approach was used to scale graphene nanosheets. 2.
The thermal conductivity of TIM can be significantly enhanced in a low loading.
AC C
EP
TE D
M AN U
SC
RI PT
3. The as-prepared TIMs can efficiently decrease the temperature of LED chips.