Vous êtes sur la page 1sur 3


Controllable graphene N-doping with ammonia plasma

Yung-Chang Lin, Chih-Yueh Lin, and Po-Wen Chiua
Department of Electrical Engineering, National Tsing Hua University, Hsinchu 30013, Taiwan

Received 6 December 2009; accepted 16 February 2010; published online 1 April 2010
Here we show that gas-phase doping by means of NH3 plasma exposure is a highly flexible and
manufacturable process for graphene electronics. The nitrogen-containing radicals can readily form
covalent bonds with the carbon lattice and keep stable in the postannealing for damage restoration.
The amount of charge transfer can be fine tuned by controlling the exposure time and monitored by
the systematic shift in the Raman G mode and the Gds Vg curves in transport measurements. The
maximum doping level can reach 1.5 1013 cm2. 2010 American Institute of Physics.
Tremendous potential for technological applications has
thrust graphene, a single atomic layer of hexagonal carbon
network, into one of the hottest areas of research activity.
This has been fueled by recent experimental breakthroughs
in diverse application areas, ranging from gas sensors,1 to
structural reinforcing agents,2 to transparent conducting
devices,3 and to nanoelectronic devices.4 For the latter case,
it has been shown that carrier mobility in graphene fieldeffect devices can reach 200,000 cm2 / V s,5 opening up the
possibility of building electronic architectures in a size approaching a molecule or a cluster of atoms.
Graphene nanoribbon is a narrow-gap semiconductor
with holes as the majority carrier.6 To construct complex
electronics with logic operation, it is desired to have both
p- and n-type conductions and the ability to control the
carrier density in an active channel. The classical approach
used to obtain electron-doped graphite is via alkali metal
intercalation.7 This approach preserves the carbon network
from being severely damaged but the rapid degradation upon
exposure to air sets the fundamental constriction for its practical applications. Therefore, graphene functionalization by
various inorganic or organic substances, which provides airstable dopant and permanently alters the electronic structure
of graphene sheets, has been of emerging importance in this
respect. Adsorption of various organic molecules has revealed the extreme sensitivity of the electrical properties of
graphene to molecule species.8 However, the precise control
over the adsorption coverage and the sustainability at high
processing temperature remain challenged for the noncovalent functionalization.
Here we describe a simple technique which makes use of
ammonia plasma for adding a variety of atomic and molecular radicals H, N, NH, and NH2 to graphene. It provides not
only a better control over the doping density but also the
compatible processing flow with state-of-the-art silicon technologies. Importantly, the resulting functional groups on
graphene are stable in air and sustain high temperature treatments. The excess defects caused by the exposure of plasma
can be greatly reduced by postannealing. It was found that
the nitrogen functionalities keep stable in graphene after
postannealing up to the tested 800 C.
Mechanical exfoliated graphene as well as bilayer
graphene grown by chemical vapor deposition were both

Electronic mail: pwchiu@ee.nthu.edu.tw.


used in the current study. It has been shown that bilayer

graphene grown on transition metals recovers the characteristic Dirac cones of isolated monolayer graphene due to the
effective screening of the inner graphene layer from the
metal d-states and, as a result, bilayer graphene on transition
metals behaves like isolated monolayer graphene.9 To avoid
the mesoscopic variation in charge density in graphene,10,11
gold markers were made in the vicinity of the graphene studied so that the evolution of charge density upon doping on
the same places can be clearly monitored. The NH3 plasma at
a dose of 3 1014 1 / cm2 was applied to graphene in a stepwise fashion. Between two steps, a Raman spectrum at 633
nm excitation was taken. It reveals the functionalization status on the same position. X-ray photoelectron spectroscopy
XPS was used to provide insight into the chemical bonding
of graphene with nitrogen. To do this, the Ni film with overlying plasma-exposed graphene was first transferred to polymethyl methacrylate PMMA with the aid of hydrochloric
acid. The etching process isolated graphene sheets on
PMMA, which were then transferred to a silicon chip with
the residual PMMA removed by acetone and thermal annealing.
Figures 1a and 1b display the XPS C 1s and N 1s
spectra of graphene treated with NH3 plasma, respectively.
The survey spectrum in Fig. 1a inset indicates the presence
of carbon, nitrogen, and oxygen on the samples. The oxygen
signal can be related to either residual PMMA or to air contamination due to ex situ analysis. For peak fittings, the C 1s
and N 1s background spectra are removed by using Shirley
algorithm. In Fig. 1a, the main C 1s peak, centered at
284.7 eV, shows a clear shift to higher binding energy with
respect to the position of the pure sp2 CC bonding 284.3
eV, indicating the incorporation of n-type functional groups
as a result of plasma exposure.12,13 This shift is also in agreement with increased structural disorder due to the disruptions
in the sp2 carbon framework from the incorporation of
nitrogen.14 The other two peaks locate at 285.8 and 287.1 eV
can be identified as different CN bonding structures, with
sp2- and sp3-hybridized carbon, respectively.12 The analysis
of the nitrogen chemical environment is shown in Fig. 1b
where the asymmetric N 1s spectrum is deconvoluted using
three Gaussian peaks. The peak at 398.3 eV accounts for
most of the nitrogen in graphene and can be assigned to
pyridinelike N, i.e., N has two carbon neighbors in a hexagonal ring.15 This bonding environment can occur at the edge

96, 133110-1

2010 American Institute of Physics

Downloaded 14 Sep 2010 to Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions


Appl. Phys. Lett. 96, 133110 2010

Lin, Lin, and Chiu




Intensity (a.u.)

Intensity (a.u.)


600 500 400 300 200

Binding Energy (eV)









Intensity (a.u.)

Intensity (a.u.)






EF (c)















Binding Energy (eV)
1550 1600 1650




1550 1600 1650

Raman shift (cm-1)



FIG. 2. Color online a Raman spectra of NH3 plasma-treated graphene

on Ni. The spectra from the top to the bottom correspond to 0, 3, 6, and 9
min exposure. The G peaks were scaled to have similar intensity. b and c
show the evolution of G peak upon plasma exposure for graphene with
initial Fermi level lying in conduction band and valance band, respectively.
The dashed lines indicate the G peak position of pristine graphene.



FIG. 1. Color online XPS core level spectra of a C 1s and b N 1s of

graphene treated with NH3 plasma. c Schematic of postulated nitrogenation on graphene. The inset in a is the XPS survey spectrum.

of graphene or if a carbon vacancy is created next to the

nitrogen inside the carbon network, as illustrated in Fig. 1c.
The absence of the third coordination of the nitrogen is
associated with the defects into the graphene sheet and
reflected in the pronounced D peak in the corresponding
Raman spectra discussed below. Upon HN3 plasma exposure,
covalent functionalization with amine groups, fitted at
399.0 eV, can occur at the edge or defect sites.16 Substitution
of C atoms by NH radicals in hexagonal rings has been
shown to cause a more effective n-doping in graphene17 than
similar reaction in pentagonal rings pyrrolelike N observed
at 400.2 eV.18 The specific assignment of the small peak at
401.4 eV is, however, difficult. It can be considered to quaternarylike N or graphitelike N, i.e., threefold coordinated
sp2 N in the hexagonal rings of graphene. This graphitic
form of nitrogen also contributes to the n-doping.
Figure 2 shows the evolution of micro-Raman spectra of
bilayer graphene on Ni under NH3 plasma exposure. For
pristine graphene, the D peak is barely seen, indicative of the
absence of significant number of defects. As graphene is exposed to NH3 plasma, the D peak intensity rises rapidly and
saturates after 3 min exposure. With the increase in exposure
time, the 2D peak, the second order of D peak, monotonically decreases in intensity. The further reduction in 2D intensity after the saturation of D peak is relevant to the excess
charges from dopant and the change in the I2D / IG intensity

ratio provides a simple measure of the doping level. Referring to the electrostatic gating of graphene on Si substrate,19
we obtain a doping level of 1 1013 cm2 for the change
in the I2D / IG intensity ratio from 1.7 to 0.7. The chemical
doping-induced suppression of the 2D intensity is consistent
with the Raman spectra obtained in electrostatically gated
graphene.19,20 The charge transfer to graphene is also evident
in the shift in G peak frequency. Depending on the initial
position of the Fermi level with respect to the Dirac point in
the outer layer, stiffening or softening of the G peak can
occur upon NH3 plasma exposure.20 Figures 2b and 2c
show the systematic G peak upshift and downshift, respectively. The maximum G peak shift is about 5 cm1, corresponding to a doping level of 8 1012 cm2.19,20 This is in
accord with the doping level estimated by the reduction in
I2D / IG intensity ratio. Also note that the doping is inhomogeneous over the graphene flakes. The edge exhibits the
maximum shift in G peak, while the central areas are relatively inert.
The incorporation of nitrogen-containing molecules into
graphene is found to be quite stable even at elevated temperatures where the damaged hexagonal lattice can be partially repaired. Figure 3 compares the Raman spectra of exfoliated graphene before and after annealing up to 800 C.
On one hand, the D peak induced by plasma is significantly
reduced after annealing, indicative of the restoration of damaged lattice. On the other hand, the G peak can be shifted
even to higher frequency upon annealing, showing the robust
covalent bonds between carbon and nitrogen and the enhanced electron donation to graphene. The doping level can

Downloaded 14 Sep 2010 to Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions


Appl. Phys. Lett. 96, 133110 2010

Lin, Lin, and Chiu


Intensity (a.u.)

plasma 2 min



Raman shift (cm-1)

FIG. 3. Color online Raman spectra of the same exfoliated graphene on Si

substrate, with postannealing at indicated temperatures for 5 min.

reach 1.5 1013 cm2 after applying the drive-in annealing.

Charge transport in doped graphene exhibits asymmetry
in electron and hole conduction. This is attributed to the
asymmetric metal-induced broadening of density of electronic states in graphene electrodes that misaligns to the
channel neutrality points.8 The suppression of electron conduction occurs as the channel is hole-doped, and vice versa
for the case of electron doping. In Fig. 4, we show the evolution of the Gds Vg curves of the same exfoliated graphene
on Si substrate upon NH3 plasma exposure. The device is
doped stepwise from p- to n-type. The corresponding Raman
spectra for each doping state are displayed in Fig. 4d for
comparison. Prior to doping, graphene shows p-type
behavior.1 The Gds Vg curve exhibits explicitly the reduction in electron conduction for the gate voltage greater than
VDirac where the minimum conductivity occurs. Upon NH3
plasma exposure, the minimum conductivity successively
shifts to more negative gate voltages. As shown in Fig. 4b,
the VDirac meets the zero gate voltage after 1 min exposure.
The overall conductance decreases due to the creation of
excess defects in the channel, as revealed by the increase in
the D peak intensity in the corresponding Raman spectrum in
Fig. 4d. The electron and hole conduction are essentially




G (e2/h)

-10 0


10 20 30 40

Intensity (a.u.)


1 min

-20 -10
1.5 (c)



1 min



2 min
2 min

-40 -20

0 20
Vg (V)


1400 1600 2600


Raman shift (cm-1)

FIG. 4. Color online ac Gds Vg curves of the same exfoliated

graphene at different doping states. Measurements were carried out at 10 K.
The dashed lines indicate the gate voltage at the charge neutrality point. d
Raman spectra correspond to the graphene at different doping states shown
in ac.

symmetric, except for a discernible shoulder developed in

the positive gate voltage. In addition, the conductance at the
charge neutrality point is temperature-independent not
shown. Increasing the exposure time to 2 min Fig. 4c
moves the Gds Vg curve to more negative gate voltages,
indicative of the n-doping on graphene. The pronounced suppression of the hole conduction can be clearly seen. At Vg
0, the associated shoulder in electron conduction turns into
a plateau reproducible at different temperature measurements. It should be noted that the physisorbed NH3 molecules on the carbon surface do not directly dope graphene in
the absence of water vapor.21 The Gds Vg curve measured in
vacuum is thus unaffected even at the presence of partial
physisorption of NH3.
In conclusion, this paper presents a facile electron doping of graphene by ammonia plasma. Unlike the classical
approach in which graphene is doped during growth, the
postdoping offers the controllability over dopant concentration and silicon compatible processing technologies. The
shift in the Raman G peak as well as the reduction in I2D / IG
intensity ratio are sensitive to the change in charge density in
graphene and provide an easy mean to calculate the doping
Work carried out under the auspices of NTHU booster
project, TSMC JDP project under Contract No. NTHU0905,
and Taiwan National Science Council under contract No.
NSC 97-2112-M-007-016-MY3.

F. Schedin, A. K. Geim, S. V. Morozov, E. W. Hill, P. Blake, M. I.

Katsnelson, and K. S. Novoselov, Nature Mater. 6, 652 2007.
S. Stankovich, D. A. Dikin, G. H. B. Dommett, K. M. Kohlhaas, E. J.
Zimney, E. A. Stach, R. D. Piner, S. T. Nguyen, and R. S. Ruoff, Nature
London 442, 282 2006.
X. Li, G. Zhang, X. Bai, X. Sun, X. Wang, E. Wang, and H. Dai, Nat.
Nanotechnol. 3, 538 2008.
Y.-M. Lin, K. A. Jenkins, A. Valdes-Garcia, J. P. Small, D. B. Farmer, and
P. Avouris, Nano Lett. 9, 422 2009.
S. V. Morozov, K. S. Novoselov, M. I. Katsnelson, F. Schedin, D. C. Elias,
J. A. Jaszczak, and A. K. Geim, Phys. Rev. Lett. 100, 016602 2008.
X. Wang, Y. Ouyang, X. Li, H. Wang, J. Guo, and H. Dai, Phys. Rev. Lett.
100, 206803 2008.
M. S. Dresselhaus and G. Dresselhaus, Adv. Phys. 51, 1 2002.
D. B. Farmer, R. Golizadeh-Mojarad, V. Perebeinos, Y.-M. Lin, G. S.
Tulevski, J. C. Tsang, and P. Avouris, Nano Lett. 9, 388 2009.
P. Sutter, M. S. Hybertsen, J. T. Sadowski, and E. Sutter, Nano Lett. 9,
2654 2009.
C. Casiraghi, S. Pisana, K. S. Novoselov, A. K. Geim, and A. C. Ferrari,
Appl. Phys. Lett. 91, 233108 2007.
J. Martin, M. Akerman, G. Ulbricht, T. Lohmann, J. H. Smet, K. von
Klitzing, and A. Yacoby, Nat. Phys. 4, 144 2007.
C. Morant, J. Andrey, P. Prieto, D. Mendiola, J. M. Sanz, and E. Elizalde,
Phys. Status Solidi A 203, 1069 2006.
S. Maldonado, S. Morin, and K. J. Stevenson, Carbon 44, 1429 2006.
S. Maldonado and K. J. Stevenson, J. Phys. Chem. B 109, 4707 2005.
N. Hellgren, J. Guob, Y. Luoc, C. Sthed, A. Aguid, S. Kashtanov, J.
Nordgren, H. gren, and J.-E. Sundgren, Thin Solid Films 471, 19
A. Felten, C. Bittencourt, J. J. Pireaux, G. Van Lier, and J. C. Charlier, J.
Appl. Phys. 98, 074308 2005.
X. Wang, X. Li, L. Zhang, Y. Yoon, P. K. Weber, H. Wang, J. Guo, and H.
Dai, Science 324, 768 2009.
J. Robertson and C. A. Davis, Diamond Relat. Mater. 4, 441 1995.
A. Das, S. Pisana, B. Chakraborty, S. Piscanec, S. K. Saha, U. V. Waghmare, K. S. Novoselov, H. R. Krishnamurthy, A. K. Geim, A. C. Ferrari,
and A. K. Sood, Nat. Nanotechnol. 3, 210 2008.
J. Yan, Y. Zhang, P. Kim, and A. Pinczuk, Phys. Rev. Lett. 98, 166802
C. W. Bauschlicher, Jr. and A. Ricca, Phys. Rev. B 70, 115409 2004.

Downloaded 14 Sep 2010 to Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions