Vol. 36, No. 1
Journal of Semiconductors
January 2015
Redistribution of carbon atoms in Pt substrate for high quality monolayer graphene
synthesis
Li Yinying(李银英), Wu Xiaoming(伍晓明), Wu Huaqiang(吴华强)Ž , and Qian He(钱鹤)
Institute of Microelectronics, Tsinghua University, Beijing 100084, China
Abstract: The two-dimensional material graphene shows its extraordinary potential in many application fields.
As the most effective method to synthesize large-area monolayer graphene, chemical vapor deposition has been
well developed; however, it still faces the challenge of a high occurrence of multilayer graphene, which causes the
small effective area of monolayer graphene. This phenomenon limits its applications in which only a big size of
monolayer graphene is needed. In this paper, by introducing a redistribution stage after the decomposition of carbon
source gas to redistribute the carbon atoms dissolved in Pt foils, the number of multilayer flakes on the monolayer
graphene decreases. The mean area of monolayer graphene can be extended to about 16 000 m2 , which is eight
times larger than that of the graphene grown without the redistribution stage. A Raman spectrograph is used to
demonstrate the high quality of the monolayer graphene grown by the improved process.
Key words: graphene synthesis; platinum; redistribution stage; monolayer grapheme
DOI: 10.1088/1674-4926/36/1/013005
EEACC: 2520
1. Introduction
Due to its fascinating electrical and mechanical properties, graphene has attracted increasing attention since being
mechanically exfoliated from bulk graphite. Many graphene
production techniques have been developed, such as the mechanical exfoliation of highly oriented pyrolytic graphiteŒ1 ,
epitaxial growth on SiCŒ2; 3 , graphite oxide reductionŒ4 , CVD
growth on metal substratesŒ5 11 etc. Among these, CVD
growth using Cu, Ni or Pt as the substrate is the foremost way
to synthesize large-area graphene with relatively acceptable
uniformity. Many high performance graphene FETs and novel
carbon-based circuits are reported using CVD grown monolayer grapheneŒ12 16 . However, there are always many flakes
of multilayer or few-layer graphene (2–4 layers)Œ17 scattered
on a large part of the CVD grown monolayer graphene on
Cu, Ni and Pt substratesŒ5; 6; 9; 18 25 . The random existence
of the multilayer flakes strongly influences the performance
of graphene FETs and circuits. This paper discusses the CVD
growth mechanism of monolayer graphene on a Pt substrate
and presents an improved process to reduce the density and
size of the multilayer and few-layer graphene flakes.
Platinum films are used as the substrate in this work for
several reasons. (1) The relatively low carbon solubility of Pt
around the growing temperature (0.07–0.08wt%, 1000 ıC)
is beneficial to monolayer graphene growthŒ26 . (2) The high
melting point (1768 ıC) and low thermal expansion coefficient (8.8 m/(mK)) of Pt reduce surface agglomeration during graphene synthesis and provide a smooth graphene morphology, which suppresses the formation of pentagonal and/or
heptagonal arrangements of carbon atoms and is in favor of
subsequent graphene transfer to other substrates. (3) Unlike
Cu or Ni, Pt is hard to oxidize because of its inertness, which
may also reduce the surface roughness of graphene. (4) After annealing over 600 ıC, the (111) orientation dominates in
Pt heteromorphism due to the minimization of the surface energy of the face-centered cubic metalŒ27 29 , which induces the
weak bonding interaction between the graphene and the Pt substrateŒ30 32 . These properties make Pt a very suitable material
to synthesize uniform large-area graphene films.
The basic mechanism of graphene growth on the Pt substrate is based on carbon segregation or precipitationŒ10; 11; 31 .
Curve I in Figure 1(a) shows the temperature profile of the
graphene synthetic process. Normally, the process consists of
four stages: (1) the temperature rising in the H2 ambient, (2)
the annealing stage in the H2 ambient to clean the substrate
surface, (3) the decomposition stage where carbon atoms decomposed from carbon sources such as methaneŒ9; 22; 25; 33 or
ethyleneŒ31; 34 etc. dissolve into the Pt substrate till carbon saturation, and (4) the cooling or graphene forming stage. Carbon
solubility in platinum decreases as the temperature is decreasing, which leads to a carbon segregation or precipitation process. Then carbon atoms gather together to form continuous
graphene films on the Pt surface. However, the carbon atoms
dissolved in the polycrystalline platinum might distribute nonuniformly, which might be caused by the size and crystal orientation of Pt crystalline grains, or by defects and impurity in
and between the crystalline grains. Taking the cooling condition into accountŒ35 , the decomposition stage plays the most
important role in the formation of the final graphene morphology. Because the cooling stage is very short, graphene grown
on a Pt surface in the cooling stage tend to maintain the distribution characteristics of the dissolved carbon atoms in Pt foils
in the decomposition stage. This is the reason why there are
many flakes of multilayer or few-layer graphene scattered on
the CVD grown monolayer graphene. In this paper, we introduce a redistribution stage before cooling to allow the carbon
* Project supported by the National Natural Science Foundation of China (No. 61377106).
† Corresponding author. Email: [email protected]
Received 5 June 2014, revised manuscript received 29 July 2014
013005-1
© 2015 Chinese Institute of Electronics
J. Semicond. 2015, 36(1)
Li Yinying et al.
Figure 1. (a) Curves I and II are the temperature profiles for graphene synthesis on polycrystalline Pt foils without and with a redistribution
stage. (b) Schematic diagram of the two different graphene growth processes.
Exp.
S1
S2
S3
S4
S0
Process temp
(ıC)
1060
1060
1060
1060
1060
H2 flow
rate (sccm)
794
794
794
794
794
Table 1. Graphene growth parameters.
CH4 flow
Decomposition
Ar flow
rate (sccm)
time (min)
rate (sccm)
6
8
20
6
8
20
6
8
20
6
8
20
6
40
0
Redistribution
time (min)
2
20
40
60
0
Cooling rate
About 200 ıC/min
at the first three
minutes
S0 is the process without a redistribution stage for comparison. Considering the intense etching reaction of H2 to graphene in the cooling
stage, a longer decomposition stage (40 min here) is essential to get a full coverage of graphene by the method without a redistribution stage.
atoms to diffuse in the Pt substrate. Then, obtaining a graphene
film with fewer multilayer graphene flakes is promising. An
improved experimental process, as shown in Figure 1(b) II,
was designed and carried out. In the redistribution stage, H2
and CH4 are purged and argon is introduced to prevent further
decomposition of CH4 and the etching effect of hydrogen during the cooling stage.
2. Experiments
We used polycrystalline platinum (50 m thick, fineness
of 99.95%, Alfa Aeser company) as the substrate. The CVD
system is described in another literature from our groupŒ35 .
Firstly, the Pt foils were soaked in acetone and cleaned by ultrasound cleaner for five minutes to remove the organic impurities on the surface. After being dried, the foils were loaded
into the reactor chamber, which is made of a quartz tube. Before the temperature rising stage, the chamber was pumped
into a high vacuum, usually < 0.1 Pa, then piped in argon to
1 atm. Then, hydrogen was introduced. During the decomposition process, methane was introduced into the chamber as the
carbon source gas. For the process without the redistribution
stage, the chamber was filled with H2 and CH4 till the sample was cooled down and graphene was synthesized. For the
process with the redistribution stage, H2 and CH4 were purged
and argon gas was used to fill the chamber to maintain an inert
ambient for the following redistribution stage. In the cooling
stage, the quartz tube chamber was pulled off and cooled down
to room temperature. Four comparative trials were carried out
and the process parameters are listed in Table 1. Scanning electron microscopy (QUANTA FEG 450) was used to observe
the graphene on the platinum. A bubbling method was used
for graphene transfer. The Pt/graphene was firstly spin coated
with polymethyl methacrylate (PMMA) followed by 150 ıC
baking for 30 min, then the Pt/graphene/PMMA clamped by
the cathode side was soaked into a 0.15 mol/L NaOH solution and the current was tuned at 0.1–0.2 A to separate the
graphene/PMMA layer from the Pt substrate. After being transferred to a Si/SiO2 substrate and removing the PMMA by acetone, optical microscopy (Olympus MX5) and Raman spectroscopy (Horiba Jobin Yvon HR 800) were used to characterize the quality of the graphene.
3. Results and discussion
3.1. Experimental results
Figure 2 shows typical SEM images of graphene on polycrystalline platinum synthesized by the process without a carbon redistribution stage. According to the Raman spectra of
the target spots marked by the circle in Figures 2(a) and 2(b),
it is indicated that the very big multilayer graphene or graphite
and the very dense few-layer graphene scattered on the monolayer grapheneŒ36 reduces the available size of pure monolayer
graphene.
Figure 3 shows the SEM images of the graphene grown
in different process conditions. As discussed above, graphene
grown by the process without the carbon redistribution stage
has big multilayer zones and very dense few-layer flakes (Figure 3(a’)). As presented in Figures 3(b) to 3(e), the quantity
and size of the multilayer and few-layer graphene flakes change
significantly. In order to quantify the experimental results, four
parameters are introduced to discuss the results (to simplify
013005-2
J. Semicond. 2015, 36(1)
Li Yinying et al.
Figure 2. Typical SEM images of graphene on platinum synthesized by the process without the redistribution stage. (a) Multilayer graphene
flakes. (b) Dense distribution of few-layer graphene. The scale bars are 50 m and 5 m in a and b respectively. (c) Raman spectra of the target
spots marked by the circle in SEM image (a) and (b), after being transferred to 90 nm SiO2 /Si.
Figure 3. SEM images of graphene grown on polycrystalline platinum foils. (a) Graphene synthesized without the redistribution stage. Graphene
synthesized by the improved method with (b) 2 min, (c) 20 min, (d) 40 min, and (e) 60 min of the carbon atom redistribution stage. (a’)–(e’)
Images with larger magnification corresponding to (a)–(e), respectively. The scale bars in (a)–(e) are 200 m, and 50 m in (a’)–(e’).
the calculation, we call both multilayer and few-layer graphene
flakes multilayer graphene hereinafter): (1) Nmulti , the amount
of multilayer flakes in a given area (600 600 m2 in this
paper when the SEM image is with 500 magnification), (2)
Aavemulti ; the average area of multilayer graphene flakes, (3)
Rcov , the coverage ratio of graphene on the surface of the
Pt substrate, and (4) Aavemono , the average area of monolayer
graphene, which is obtained from Equation (1).
Aavemono D
Atotal
Aavemulti Nmulti
;
Nregion
(1)
where Nmulti , Aavemulti and Aavemono are defined as above, Atotal
is the given area (600 600 m2 in this paper) and Nregion is
the maximum amount of triangle regions on the area divided by
multilayer flakes. It can be derived from Euler’s formula (2),
Nregion D e
v C 2;
(2)
where v is the number of vertices in the area (here is the amount
of multilayer flakes Nmulti /, and e D 3(v 2) is the maximum
amount of line segments mutually disjoint to each other connected from the vertices v Œ37 .
This parameter Aavemono implies the average intact area that
can be used for device or circuit fabrication without any multilayer flakes. The relationship between the four parameters and
013005-3
J. Semicond. 2015, 36(1)
Li Yinying et al.
Figure 4. (a) Relationship between parameter Nmulti and Aavemulti and the duration of the redistribution stage. (b) Relationship between parameter
Rcov and Aavemono and the duration of the redistribution stage.
Figure 5. (a) Optical image of graphene transferred on 90 nm SiO2 /Si substrate. (b) Raman spectrum at the place labeled by the circle in (a). (c)
FWHM of 2D and G peaks for 18 random sampling points of monolayer graphene.
redistribution duration is shown in Figure 4.
3.2. Characterization of results
For the monolayer graphene synthesized by the process
with the redistribution stage, Raman spectroscopy was used
to evaluate its quality. Typically, the Raman spectrum of the
graphene sample synthesized with a redistribution time of
60 min is shown in Figure 5(b). The G peak is at 1584 cm 1
and the 2D peak is centered at 2681 cm 1 . Additionally, the
intensity ratio of the G to the 2D band is about 0.26, which
is smaller than 0.5. The 2D band has a sharp and symmetrical
Lorentz profile. No apparent D peak at 1350 cm 1 is observed.
These properties indicate a high crystalline quality and the integrity of the SP2 structure in the grapheneŒ38;39 . 18 random
points on the monolayer graphene were sampled to take Raman spectra. The full width half maximum (FWHM) of the 2D
and G band of the spectra is shown in Figure 5(c). The distribution ranges of FWHM of the 2D and G band are 28–35 cm 1
and 10–18 cm 1 respectively, which prove the high uniformity
and high quality of the graphene grown on Pt substrates by the
improved method.
3.3. Analysis of the results
Based on the experimental results which are shown in Figures 3 and 4, the introduced redistribution stage strongly affects the quantity and size of few-layer and multilayer graphene
flakes on the CVD synthesized monolayer graphene. Basically, the longer the redistribution stage lasts, the lower the
density of the multilayer graphene flakes becomes. The multilayer graphene flakes almost totally vanish on the monolayer
graphene grown in the process condition of 60-min redistribution. However, the average size of the multilayer and few-layer
graphene flakes increases with the increase of the redistribution
stage duration if this stage lasts less than 60 min. Fortunately,
the net size of monolayer graphene, reflected by parameter
Aavemono , increases with the redistribution stage duration. For
the graphene sample grown with the 60 min redistribution duration, parameter Aavemono is about eight times larger than that
of graphene grown without the redistribution stage. This result
implies that the redistribution stage effectively smoothes the
carbon atom concentration in the Pt substrate, even though the
few-layer graphene flakes do not completely disappear. When
the redistribution stage is very long (e.g. 60 min), the coverage
of graphene on the Pt substrate is no longer 100%, as shown
in Figure 4(b). The possible reason is that hydrogen atoms also
dissolve into the Pt substrate during the CH4 decomposition
stage besides carbon atoms. In the redistribution stage, part of
the carbon and hydrogen atoms are released from the Pt substrate together to form alkane gases, which are vented. Then,
fewer carbon atoms are left in the Pt substrate when the redistribution stage is much longer. This might cause some areas
on the Pt substrate to lack carbon atoms and the coverage of
graphene decreases. Considering the disagreeable issue of incomplete graphene coverage, and the fact that a longer redistri-
013005-4
J. Semicond. 2015, 36(1)
Li Yinying et al.
bution stage makes a much smaller Nmulti , Aavemulti and a larger
Aavemono , the optimal duration of the redistribution stage should
be between 40 and 60 min for our study.
4. Conclusions
In summary, this paper presents an improved CVD process
for graphene growth on a Pt substrate by introducing a redistribution stage to decrease the quantity of few-layer and multilayer flakes on monolayer graphene. In the redistribution stage,
the concentrated carbon atoms dissolved in the Pt substrate diffuse again. When the redistribution stage is between 40 and
60 min, the multilayer zones almost totally vanish and only a
tiny amount of few-layer graphene remains. The mean area of
the monolayer graphene is enlarged by about eight times compared to the graphene grown without the redistribution stage.
Additionally, Raman spectrograph measurement results show
that the quality of graphene grown by the improved process is
similar to that of the micromechanically cleaved graphene.
Acknowledgements
The authors appreciate the support from the Tsinghua
Nanofabrication Technology Center.
References
[1] Novoselov K S, Geim A K, Morozov S V, et al. Electric field
effect in atomically thin carbon films. Science, 2004, 306(5696):
666
[2] Hupalo M, Conrad E H, Tringides M C. Growth mechanism for
epitaxial graphene on vicinal 6H-SiC (0001) surfaces: a scanning
tunneling microscopy study. Phys Rev B, 2009, 80(4): 041401
[3] Emtsev K V, Bostwick A, Horn K, et al. Towards wafer-size
graphene layers by atmospheric pressure graphitization of silicon
carbide. Nature Mater, 2009, 8(3): 203
[4] Eda G, Chhowalla M. Chemically derived graphene oxide: towards large-area thin-film electronics and optoelectronics. Adv
Mater, 2010, 22(22): 2392
[5] Kim K S, Zhao Y, Jang H, et al. Large-scale pattern growth
of graphene films for stretchable transparent electrodes. Nature,
2009, 457(7230): 706
[6] Gao L, Guest J R, Guisinger N P. Epitaxial graphene on Cu (111).
Nano Lett, 2010, 10(9): 3512
[7] Wu Y, Chou H, Ji H, et al. Growth mechanism and controlled
synthesis of AB-stacked bilayer graphene on Cu–Ni alloy foils.
Acs Nano, 2012, 6(9): 7731
[8] Jacobberger R M, Arnold M S. Graphene growth dynamics on
epitaxial copper thin films. Chem Mater, 2013, 25(6): 871
[9] Li X, Cai W, An J, et al. Large-area synthesis of high-quality
and uniform graphene films on copper foils. Science, 2009,
324(5932): 1312
[10] Kang B J, Mun J H, Hwang C Y, et al. Monolayer graphene
growth on sputtered thin film platinum. J Appl Phys, 2009,
106(10): 104309
[11] Gao J H, Sagisaka K, Kitahara M, et al. Graphene growth on a Pt
(111) substrate by surface segregation and precipitation. Nanotechnology, 2012, 23(5): 055704
[12] Wu Y, Lin Y, Bol A A, et al. High-frequency, scaled graphene
transistors on diamond-like carbon. Nature, 2011, 472(7341): 74
[13] Han S J, Jenkins K A, Garcia A V, et al. High-frequency graphene
voltage amplifier. Nano Lett, 2011, 11(9): 3690
[14] El-Kady M F, Strong V, Dubin S, et al. Laser scribing of highperformance and flexible graphene-based electrochemical capacitors. Science, 2012, 335(6074): 1326
[15] Chen S Y, Ho P H, Shiue R J, et al. Transport/magnetotransport
of high-performance graphene transistors on organic moleculefunctionalized substrates. Nano Lett, 2012, 12(2): 964
[16] Schall D, Otto M, Neumaier D, et al. Integrated ring oscillators
based on high-performance graphene inverters. Scientific Reports, 2013: 3
[17] Yao Y, Li Z, Lin Z, et al. Controlled growth of multilayer, fewlayer, and single-layer graphene on metal substrates. The Journal
of Physical Chemistry C, 2011, 115(13): 5232
[18] Bae S, Kim H, Lee Y, et al. Roll-to-roll production of 30-inch
graphene films for transparent electrodes. Nature Nanotechnol,
2010, 5(8): 574
[19] Yamada T, Ishihara M, Kim J, et al. A roll-to-roll microwave
plasma chemical vapor deposition process for the production of
294 mm width graphene films at low temperature. Carbon, 2012,
50(7): 2615
[20] Delamoreanu A, Rabot C, Vallee C, et al. Wafer scale catalytic
growth of graphene on nickel by solid carbon source. Carbon,
2014, 66: 48
[21] Gao T, Xie S, Gao Y, et al. Growth and atomic-scale characterizations of graphene on multifaceted textured Pt foils prepared by
chemical vapor deposition. ACS Nano, 2011, 5(11): 9194
[22] Li X, Magnuson C W, Venugopal A, et al. Graphene films with
large domain size by a two-step chemical vapor deposition process. Nano Lett, 2010, 10(11): 4328
[23] Gao L, Ren W, Zhao J, et al. Efficient growth of high-quality
graphene films on Cu foils by ambient pressure chemical vapor
deposition. Appl Phys Lett, 2010, 97(18): 183109
[24] Ping J, Fuhrer M S. Carbon impurities on graphene synthesized
by chemical vapor deposition on platinum. arXiv preprint arXiv:
1304.5123, 2013
[25] Li X, Magnuson C W, Venugopal A, et al. Large-area graphene
single crystals grown by low-pressure chemical vapor deposition
of methane on copper. J Ame Chem Soc, 2011, 133(9): 2816
[26] Siller R H, Oates W A, McLellan R B. The solubility of carbon
in palladium and platinum. Journal of the Less Common Metals,
1968, 16: 71
[27] Liu C L, Cohen J M, Adams J B, et al. EAM study of surface selfdiffusion of single adatoms of fcc metals Ni, Cu, Al, Ag, Au, Pd,
and Pt. Surf Sci, 1991, 253(1): 334
[28] Kawamura M, Mashima T, Abe Y, et al. Formation of ultra-thin
continuous Pt and Al films by RF sputtering. Thin Solid Films,
2000, 377: 537
[29] Wang J, Zhao D G, Sun Y P, et al. Thermal annealing behaviour
of Pt on n-GaN Schottky contacts. J Phys D: Appl Phys, 2003,
36(8): 1018
[30] Giovannetti G, Khomyakov P A, Brocks G, et al. Doping
graphene with metal contacts. Phys Rev Lett, 2008, 101(2):
026803
[31] Sutter P, Sadowski J T, Sutter E. Graphene on Pt (111): growth
and substrate interaction. Phys Rev B, 2009, 80(24): 245411
[32] Gao M, Pan Y, Huang L, et al. Epitaxial growth and structural
property of graphene on Pt (111). Appl Phys Lett, 2011, 98(3):
033101
[33] Gao L, Ren W, Xu H, et al. Repeated growth and bubbling transfer of graphene with millimetre-size single-crystal grains using
platinum. Nature Commun, 2012, 3: 699
[34] De Parga A L V, Calleja F, Borca B, et al. Periodically rippled graphene: growth and spatially resolved electronic structure.
013005-5
J. Semicond. 2015, 36(1)
Li Yinying et al.
Phys Rev Lett, 2008, 100(5): 056807
[35] Xiao K, Wu H, Lv H, et al. The study of the effects of cooling conditions on high quality graphene growth by the APCVD method.
Nanoscale, 2013, 5(12): 5524
[36] Graf D, Molitor F, Ensslin K, et al. Spatially resolved Raman
spectroscopy of single-and few-layer graphene. Nano Lett, 2007,
7(2): 238
[37] Rosen K. Discrete mathematics and its applications, 6th ed. Beijing: China Machine Press, 2008
[38] Malard L M, Pimenta M A, Dresselhaus G, et al. Raman spectroscopy in graphene. Physics Reports, 2009, 473(5): 51
[39] Lee D S, Riedl C, Krauss B, et al. Raman spectra of epitaxial
graphene on SiC and of epitaxial graphene transferred to SiO2 .
Nano Lett, 2008, 8(12): 4320
013005-6