of 12
S
upplementary
Information
Supplementary
Fig
ure
1
.
A schematic of the experimental setup used for graphene
fabrication.
Supplementary
Fig
ure
2
.
Emission spectrum
of the plasma:
Negative peaks
indicate an increase in particular species while positive peaks indicate a decrease.
Supplementary
Fig
ure
3
.
Response of the FID detector to methane in
nitrogen
:
The data are shown in units of PPM, and the line is a linear fit of the
FID integrated area.
Supplementary
Fig
ure 4.
XPS studies of PECVD
-
graphene on Cu:
(
a
) XPS spectrum of a
PECVD
-
graphene sample on Cu foil, showing no discernible signals at the
N 1s core level
binding energy (398.1eV)
. (
b
) Further investigation of two s
mall peaks at 407.5 eV (Peak A) and
395.5 eV (Peak B) are attributed to the Auger lines of underlying Cu, as manifested by the
consistent XPS spectral features for a bare Cu foil (black curve) with those of our graphene
sample on Cu (red curve). The shifti
ng in energy of the two peaks with changing the source of
XPS from Mg to Al (blue curve) further support the fact that the peaks are from the Auger lines
of Cu rather than the nitrogen 1s core level because the binding energy of nitrogen is
independent of
the x
-
ray energy.
a
41
0
405 400 395 390
Binding Energy (eV)
b
a
b
c
Cu foil
Cu
(100)
Cu
(111)
Supplementary Figure 5.
STM studies of the s
urface morphology of
PECVD
-
graphene
on different substrates:
From left to right, the height histograms of successively decreasing
areas for PECVD
-
graphene o
n
(
a
)
Cu foil;
(
b
)
Cu (100) single crystal; and
(
c
)
Cu (111)
single crystal.
The overall surface morphology for
PECVD
-
grown graphene
appears to be
much smoother than that of the 1000
C
thermal CVD
-
grown graphene at all length scales.
Supplementary
Figur
e
6
.
Raman spectral analysis of PECVD
-
graphene
on Cu
:
(a)
Histogram of the
D
/
G
intensity ratios
(
I
D
/
I
G
)
of the Raman map in Figure 2
b
, showing nearly
complete absence of the
D
-
band signal (which is associated with defects) in the PECVD
-
grown graphene.
(b)
Histogram of the 2D/
G
intensity ratios
(
I
2D
/
I
G
)
of the Raman map in
Figure 2
b
, showing the majority of the ratios exceeding 1, which is a figure of merit for the
quality of monolayer graphene.
(c) Histogram of the 2D
-
band linewidth (FWHM) of the
same PECVD
-
graphene on Cu taken several days after growth, showing a mean FWHM =
37.0 cm
1
. (d) Histogram of the 2D
-
band FWHM of the same sample as in (c) taken at 30
hours later while continuously exposed to atmosphere, showing a mean FWHM = 43.9 cm
1
.
(e) A comparative histogram of the 2D
-
band FWHM of a commercially purchased thermal
-
CVD gr
own graphene on Cu (Graphene Supermarket), showing a mean FWHM = 54.1 cm
1
.
Counts
D/G ratio
Counts
2
D/G ratio
a
b
c
e
d
(
I
D
/
I
G
)
(
I
2D
/
I
G
)
Supplementary
Figur
e 7.
Raman spectral analysis of PECVD
-
graphene
tran
sferred
from Cu to SiO
2
after one year of its growth:
(a)
Spatial map
of the
FWHM of the 2D
-
band over an area of (160
150)
m
2
, showing typical FWHM values < 30
cm
1
, which is a
figure of merit for monolayer graphene according to S. Lee
et al.
,
Nano Lett.
10,
4702
(2010)
.
(b)
Spatial map
of the 2D/
G
intensity ratios
(
I
2D
/
I
G
)
over the same sample area of as
in (a), showing
(
I
2D
/
I
G
)
> 2 over most of the sample
,
again consistent with the figure of
merit
(
I
2D
/
I
G
)
> 1 for
monolayer graphene.
(c) Histogram of the FWHM of the 2D
-
band in
(a), showing a mean FWHM = 28.8 cm
1
. (d) Histogram of the
2D/
G
intensity ratios
(
I
2D
/
I
G
)
in (b), s
howing a mean value of
(
I
2D
/
I
G
)
=
2.7.
Supplementary
Fig
ure
8
.
Comparison of
the Raman spectroscopy of PECVD
-
graphene and
thermal CVD
-
grown graphene on different substrates:
From left to right, Raman spectra of
PECVD
-
graphene,
1000
C
-
grown graphene, comparison of the 2D
-
band peaks, and comparison
of the zone
-
centre G
-
band peaks f
or samples grown on
(
a
)
Cu foils,
(
b
)
Cu (100) and
(
c
)
Cu
(111).
T
he Raman spectra
taken here
were collected with a Renishaw M1000 @ 514 nm
.
Supplementary
F
ig
ure
9
.
Simulations of the real
-
space Moiré patterns (
left panels
) and the
corresponding FT (
right panels
) for the 2D graphene honeycomb
lattice on
(
a
)
Cu (100)
square lattice
and
(
b
)
Cu (111) hexagonal lattice. The FT
Moiré pattern
for graphene on Cu
(100) is consistent with a honeycomb lattice at
= (12
2)
relative to the square lattice of Cu
(100), whereas that for graphene on Cu (111
) is consistent with a honeycomb lattice at
=
(
6
2)
relative to the hexagonal lattice of Cu (111)
.
The FT
Moiré pattern
s of
a,
and
b,
compare favorably with the FT spectra of Fig. 3
b
and Fig. 3
c
, respectively.
S
upplementary
Notes
:
1.
Analysis of
the
plasma
in
P
ECVD growth
The optical emission spectra (OES) of the plasma were measured using a fiber coupled
spectrometer (S2000, Ocean Optics, Inc) placed directly below the Evenson cavity. Absorption
spectra were taken prior to graphene deposition away from the
area of the copper sample. The
spectra were referenced to a pure hydrogen plasma,
i.e.
, before the addition of the flow of
methane. Shown in
Supplementary
F
igure
2 is a typical emission spectrum featuring CN (388
nm), CH (431 nm), and CC (516 nm),
1
where n
egative absorption peaks indicate an increase in a
component and positive peaks indicate a decrease.
Graphene growth was performed on a variety of copper substrates, including high purity
copper foil, common OFHC sheet, single crystal (100), and single c
rystal (111). Gas
chromatography was used to measure the amount of methane in the hydrogen stream during
graphene growth. The gas chromatograph (GC) was an HP 5890 II employing a flame ionization
detector (FID). The FID was calibrated using reference stand
ards (Mesa Gass, Inc) of 500, 1000,
2000, 4000, and 10000 ppm of CH
4
in nitrogen. The response (integrated peak area) of the FID
to methane was found to be linear and is shown in
Supplementary Figure
3
.
Based on our investigations, w
e believe that the m
ost likely plasma
species acting upon
Cu during the PECVD graphene growth
are atomic hydrogen and CN radicals. Atomic hydrogen
via a hydrogen plasma is known to be effective for removing native atmospheric derived species
including Cu
2
O, CuO, Cu(OH)
2
, and
CuCO
3
. It has been recently
reported that CN radicals can
be highly reactive towards removing
Cu from a semiconductor at room temperature.
2
Our
overall observation of the PECVD graphene growth was in agreement with this notion.
As an
example, we
increase
d the amount of nitrogen present in the system slightly and found that the
PECVD growth could occur at half the normal plasma power. Conversely, under excess methane
conditions the Cu substrate would not etch at even more than double the normal plasma powe
r.
The presence of both atomic hydrogen and CN
species
in the plasma could allow simultaneous
preparation of the copper surface and deposition of high quality graphene at reduced
temperatures. However, more experimental work will be necessary to fully unde
rstand the role of
each radical plays in the growth process even though reproducible
PECVD
-
growth without any
active heating has been made.
While the presence of trace nitrogen in the plasma was essential for successful PECVD
growth, we note that nitrogen
was not incorporated into our PECVD
-
graphene, as verified by the
studies of x
-
ray photoemission spectroscopy (XPS) in Supplementary figure 4a, where no
discernible
N 1s core level binding energy
at 398.1eV could be resolved for a sample of
monolayer graph
ene on Cu foil. We further note that the two small peaks at 407.5 eV (Peak A)
and 395.5 eV (Peak B) are associated with the Auger lines of the Cu substrate rather than the
nitrogen 1s core level in the graphene sample. This realization was achieved by comp
aring the
XPS spectrum of a bare Cu foil as the control sample. As demonstrated in
Supplementary Figure
4b, the XPS data obtained from the control sample (pure Cu foil) over the same energy range
(black curve) are essentially identical to that of our graph
ene sample (red curve). Moreover,
when the XPS source was changed from Mg to Al, Peak A and Peak B both disappeared,
indicating that they must be Auger lines of the underlying Cu substrate and that there was no
nitrogen present in our graphene sample, beca
use the nitrogen binding energy cannot be
dependent on the x
-
ray energy.
2.
S
ubstrate
influence
on the Raman spectra of graphene
While the
2D
-
band FWHM of graphene
may be used for
analyzing whether a graphene
sample consists of monolayer or multi
-
l
ayers,
possible changing properties of the underlying
substrate
could lead to complications in interpreting the Raman spectroscopic data. For instance,
the
2D
-
band FWHM of
our PECVD
-
grown graphene
on Cu immediately after growth was ~ 29
cm
1
, which increas
ed with time after growth and exposure to air, as exemplified in
Supplementary
F
igure 6c
-
d. Similarly broadened FWHM of the 2D
-
band (
Supplementary
F
igure
6e) was also observed in a commercial monolayer graphene sample on copper (from Graphene
Supermarket)
grown by the thermal CVD method. These findings are consistent with oxidation
of the underlying Cu substrate, as discussed by Yin
et al
.
3
The oxidation of the Cu substrate for our PECVD
-
graphene is likely the result of plasma
-
induced damages to the top
side of Cu during the graphene growth. On the other hand, the
quality of our PECVD
-
graphene remains intact despite the oxidation of the underlying Cu
substrate, which has been verified by Raman spectroscopic studies of one of the PECVD
-
graphene samples tr
ansferred from Cu to SiO
2
after approximately one year of its growth. As
demonstrated in
Supplementary
F
igure 7a
-
d, a large spatial map of the sample over an area of
(160
150)
m
2
reveals a mean FWHM = 28.8 cm
1
for the 2D
-
band and a mean 2D/
G
intensity
ratio
(
I
2D
/
I
G
)
= 2.7, both are consistent with predominantly monolayer graphene if we use either
the criterion FWHM < 30 cm
1
for the 2D
-
band as the figure of merit for monolaye
r graphene on
SiO
2
,
4
or the criterion
(
I
2D
/
I
G
) > 1 for monolayer graphene on various substrates.
5
,
6
,
7
,
8
We further
note the apparent correlation between the spatial maps of
Supplementary
F
igure
7a
-
b
, suggesting
the consistency for determining monolayer gra
phene by using either the criterion for the FWHM
of the 2D
-
band or that for the intensity ratio (
I
2D
/
I
G
).
S
upplementary
References
:
1.
Chatei,
H. Bougdira, J. Rémy, M. Alnot, P. Bruch, C. Krüger,
J.
K.
Effect of nitrogen
concentration on plas
ma reactivity and diamond growth in a H
2
-
CH
4
-
N
2
microwave
discharge
.
Diamond and Related Materials
6
,
107
-
119 (
1997
).
2.
For
patent information
, see the website
http://www.faqs.org/patents/app/2013001
7672
.
3.
Yin,
X.
et al
.
Evolution of the Raman spectrum of graphene grown on copper upon
oxidation of the substrate.
Nano Research
, 1
-
10 (2014); doi:10.1007/s12274
-
014
-
0521
-
0.
4.
Lee, S.
et al.
Wafer scale homogeneous bilayer graphene films by chemical vapor
de
position.
Nano Lett.
10,
4702 (2010).
5.
Ferrari, A. C.
et al.
Raman spectrum of graphene and graphene layers.
Phys. Rev. Lett.
97
, 187401 (2006).
6.
Wu, W.
et al
. Control of thickness uniformity and grain size in graphene films for
transparent conductive electr
odes.
Nanotechnology
23
, 035603 (2012).
7.
Dimiev, A.
et al.
Layer
-
by
-
layer removal of graphene for device patterning.
Science
331
,
1168 (2011).
8.
Zhou, H.
et al
. Chemical vapour deposition growth of large single crystals of monolayer
and bilayer graphene.
Na
ture Communications
4
, Article number: 3096.