22nd International Symposium on Plasma Chemistry
July 5-10, 2015; Antwerp, Belgium
Non-destructive method for doping of graphene using the flowing afterglow of
microwave plasmas at reduced pressure
L. Vandsburger, G. Robig and L. Stafford
Département de Physique, Université de Montréal, Montréal, Québec, Canada
Abstract: CVD-grown graphene films have been n-doped by downstream plasmaafterglow treatment. In the flowing afterglow of a microwave N 2 plasma, metastable
species and atomic nitrogen are present in significant densities. XPS and Raman are
combined with molecular dynamics simulations to develop a mechanism based on defect
localized N-incorporation by a three-body introduction reaction.
Keywords: graphene, plasma-doping, CVD, nudged-elastic band
1. Introduction
Graphene is leading the charge in the drive to develop
the next generation of electronics and optoelectronics.
Such applications, however, depend on the availability of
versatile post-growth processing techniques that can adapt
it for implementation in technologies where band gap is
necessary, especially in the domain of large-area films
grown by chemical vapour deposition (CVD).
The n-type doping of graphene by nitrogen atoms has
been addressed as a means either for increasing electron
density in graphene films or for inducing a band gap by
direct incorporation or carbon replacement [1]. Nitrogen
doping is generally realized during growth by various
physical and chemical methods [2, 3]. Conventional postgrowth methods, which offer more versatility than
variation of the growth conditions, are based either on wet
chemical or thermal annealing processes [4, 5] or direct
implantation of atoms [6-8]. In this work, we explore the
potential of the flowing afterglow of microwave plasmas
at reduced pressure for non-destructive doping of
CV-grown graphene films [9].
2. Methods
Graphene samples were synthesized by CVD on 25 µm
copper foils commercially available from Alfa Aesar
(item No. 13382). The foils were first chemically treated
with acetic acid and annealed for 30 minutes at 1030 °C
under a flow of 12 sccm (standard cubic centimeters per
minute) hydrogen gas in order to prepare the copper
surface. The temperature was maintained at 1030 °C
during a 30 minute CVD growth phase where a gas flow
of 4 sccm methane was introduced [10].
The late-afterglow approach was described in detail in a
previous paper [11]; it is explained visually in Fig. 1. In
Fig. 1a, a surface-wave plasma is produced in a fused
silica discharge tube by a surfatron [12], operating at
2.45 GHz. The power injected into the plasma (incident
less reflected power) was approximately 30 W, yielding a
surface-wave plasma column (zone 1 in Fig. 1a) of about
3 cm. The microwave plasma expands into a stainless
steel processing chamber by a flow of ultra-high-purity
O-20-1
N2 gas (99.999%). For all experiments reported here, the
system operated at 10 torr and a flowrate of 250 sccm.
The chamber reached a base pressure of 7×10-9 torr before
gas injection, ensuring a minimal presence of impurities.
Fig. 1. Treatment of graphene films in surface wave
plasma late afterglows. (a) Schematic of the plasma
system, where 1-4 show different zone positions of the
afterglow and 5 is the position of the graphene sample in
the downstream flow. The asterisk (*) denotes the
surfatron plasma generator. (b) Shows a graphical
diagram of the sample surface with nitrogen moieties of
interest and incident beam of plasma species, which is
shown in a photograph in (c). (d) The densities of plasma
species plotted against parametrized time, with the values
for the test system explicitly marked.
Beside the main plasma zone sustained by microwave,
two other bright regions appear in the flowing afterglow
of nominally pure N 2 plasmas: the early afterglow and the
late afterglow (see Fig. 1) [13]. Both afterglow regions
are linked to the N 2 vibration-vibration pumping
mechanism, which creates highly energetic N 2 vibrational
states that are pushed downstream by high gas flows.
These highly energetic N 2 states can collide to form
N 2 (A) and N 2 (a’) metastables (at 6 eV and 8.1 eV above
ground, respectively) and subsequent ion-electron pairs
1
by associative ionization reactions [14]. The number
densities of plasma-generated N atoms, N 2 (A)
metastables, and positive ions in the downstream region
was determined in a previous paper by optical emission
spectroscopy and Langmuir probe measurements. The
results are summarized in Fig. 1d as a function of the
parametrized time after the surface-wave discharge in the
flowing afterglow. The rise in all quantities at 10 ms
corresponds to the early afterglow [15], while longer
times can be ascribed to the late afterglow. At the
experimental conditions investigated here (dashed line in
Fig. 1d), the grounded substrate holder where graphene
samples are placed is exposed to the late afterglow.
Positive ion number densities in this region are less than
107 cm-3, much lower than the values observed in the
main plasma region (>1011 cm-3), as well as in other types
of low-pressure plasma reactors (typically between 1010
and 1011 cm-3). Furthermore, since these ions are not
accelerated towards the grounded substrate, ion
bombardment effects can be ignored. As shown in [16],
the heavy species temperature is close to 300 K, which
further eliminates flux-induced heating of the sample.
For nitrogen to be useful for n-doping, N-atoms must be
introduced in pyrrolic (5 member ring), pyridinic
(6 member ring), or quarternary (trinary planar) bonded
positions in order to participate in the hybridized orbital
structure and thus to alter the large-scale electronic
structure of graphene films [17]. The basic morphology
of these three groups is presented in Fig. 1b. The
de-excitation of long-lived N 2 (A) metastable species by
collision with the surface represents a significant source
of energy, at least within a very short time frame
(de-excitation time on the order of 1-10 ns), with internal
energy 6 eV above the ground state level N 2 (X) [18].
3. Results and discussion
Fig. 2 presents the XPS data collected from a
representative set of graphene samples treated for
0-5 minutes. Fig. 2 (a) is a high-resolution scan of the
N1S peak collected from a sample treated for 30 seconds.
The deconvolution focused on identifying aromatic
inclusions at 398.7 eV (pyridinic N), 400.1 eV (pyrrolic
N), and 401.0 eV (quaternary N) [19]. It is important to
note that no N1S peak was observed in untreated samples.
The de-convolution of the N1S peaks shows that nitrogen
was incorporated into both pyrrolic and pyridinic
moieties, with no composition attributed to quaternarybonded N atoms at 401.0 eV.
Considering that
incorporation of energetic N atoms typically produces
quaternary-bonded N atoms, these findings indicate that N
incorporation into aromatic configurations occurs via a
different mechanism. Surface chemisorbed amines and
amides also appear in the N1S as very broad peaks
(FWHM 1.8 eV), typically centered at 399.5 eV [20-22].
These species are known to be present, considering the
disparity between the values given for total
N composition and the sum of the individual deconvoluted peaks, as shown in Fig. 3.
Nitrogen
2
concentration increased during treatment, as shown in
both Fig 2 (a versus b) and in the steady increase in the
C-N signal (285.6 eV) in the C1S spectrum (Fig. 3).
Fig. 2. XPS of graphene films: (a) N1S spectrum from
high resolution scan of graphene after 0.5 min treatment
showing aromatic N bonding, (b) N1S spectrum of
graphene on Cu foil after 3 minute treatment. (c) C1S
spectra of graphene films for 0, 3, and 5 min treatments.
Fig. 3. XPS analysis of composition development with
treatment duration. The change in signal intensity from
C-N bonding in the C1S spectrum and total nitrogen
concentration indicates a development of N bonded in
non-aromatic groups.
Further analysis of the XPS data, presented in Fig. 3,
shows that the N1S signals arising from aromatic N cease
to increase soon after the first treatment. This indicates
that, over the range of experimental conditions
investigated, the dose of plasma-generated species is
much larger than the capacity of the graphene to accept
aromatic inclusions. N atom bonding, however, continues
to increase linearly with treatment time, as shown in
Fig. 3. Considering the constant surface flux of N-atoms
during late afterglow treatment, these findings strongly
indicate that incorporation of atomic N into aromatic
O-20-1
bonding is preceded by surface N adsorption and that the
graphene surface is not fully covered by adsorbed
N atoms. Published reports indicate that N adatoms are
mobile on unsaturated graphene films [23]. Therefore,
although adsorption does not necessarily occur at defect
sites, adsorbed N species can migrate to defect sites for
rapid, defect-guided, N incorporation. In this framework,
the density of as-grown defects limits the extent of
aromatic inclusions, as seen in the XPS data.
Raman data support the finding that nitrogen atoms
are introduced steadily during late afterglow treatment.
Raman spectra for three samples are shown in Fig. 4.
While as-grown samples show the typical 2D and G peaks
resulting from large area sp2 bonding, treated samples
show the development of the D peak, attributed to out-ofBy studying the
plane sp3-type bonding [24].
development of the average D/G ratio as shown in Fig. 4c,
it can be seen that D peak development increases linearly
with treatment time. Considering that the increase in
aromatic-bonded N found by XPS ceases early in that
period, it is concluded that nitrogen adatoms are attaching
on the graphene surface, by converting sp2 bonded
graphitic carbon into out-of-plane bonded ternary carbon.
The steady development of additional peaks, specifically
the D’ and the D+D’ peaks [25], is consistent with sp3
hybridization and corresponds with what has been
reported for nitrogenation of graphene. N heteroatoms
accentuate the Raman signals originating from point
defects [30]. In particular, the development of D’ peaks
have been shown to result from defect activation that
results from nitrogenation [26].
Fig. 4. Raman resonance spectra of graphene films at
514 nm (a) normalized spectra taken from graphene films
transferred onto Si by PMMA transfer. Peaks labeled by
‡ and * are artifacts of the transfer process arising from
residual PMMA and the Si/SiO 2 substrate, respectively.
(b) Raman spectra of graphene films taken directly on
copper foil. The spectra are offset to emphasize that the
D peak is induced after plasma treatment. (c) The steady
increase of the D:G ratio indicates a near constant rate of
surface reactions, corresponding to the steady increase in
C-N bonding signals in XPS data. (d) Peak shifts of both
the G and 2D peaks from transferred graphene samples as
a function of treatment time (min).
Raman data also indicate that in transferred samples,
Fig. 3a, the 2D peak was red-shifted up to 1.43 cm-1 after
treatment. A red-shift was also measured in the G peak
O-20-1
(shift up to 1.15 cm-1 after 5 minutes). The trends, shown
in Fig 3d, demonstrate that the shift is not progressive,
establishing a stable shift after the first treatment, and is
similar in magnitude for both G and 2D peaks. A redshift in G and 2D peak positions has been found to occur
in graphitic nanomaterials after N-doping by aromatic
group formation [27], which supports the findings of XPS
that nitrogen afterglow treatment results in nitrogen
incorporation, and that the incorporation saturates very
rapidly during treatment.
The defect-localized formation of aromatic N groups
in graphene was examined by computational simulations
using density functional theory (DFT) and the nudged
elastic band (NEB) method, similar to published analysis
of ammonia-based N graphitization reactions [28].
Systems based on single and double-vacancies in
graphene were built and were subsequently used to
examine the transition state energies of the aromatic N
incorporation reaction.
The premise of the model
approach was to evaluate whether an N adatom could be
introduced into a graphene point defect with an activation
energy of less than 6 eV. This threshold value of 6 eV
corresponds to the de-excitation energy of N 2 (A)
metastable species present in the late afterglow region.
For single-vacancy systems, a graphene cell was created
comprising 61 carbon atoms and a single nitrogen atom.
This was used to examine the formation of pyridinic N
heterocycles in graphene point defects. The cell was
larger for divacancy systems, containing 86 carbon atoms
and a single nitrogen atom. The divacancy system was
used to test incorporation of N adatoms into pyridinic and
pyrolic moieties. Both systems were saturated with
hydrogen atoms at the edges. DFT calculations were used
both to optimize the molecular images making up the
endpoints of the NEB chain-of-states, and to calculate the
energy of each image in the chain. Design of the endpoint
configurations was based on hypothetical molecular
geometries for N adatoms to translate into the heterocycle
bonded positions. These images were optimized, with
component atoms being free to relax, until the energy
change per optimization step dropped below 10-10 eV.
Simulations reveal that aromatic incorporation of
adsorbed N atoms in both single and double point
vacancies occurs by passing through a transition state
with energy less than 6 eV. The reaction paths are shown
in Fig. 5a-c. In Fig. 5a, a single vacancy forms a
pyridinic N inclusion, with a relative transition state
energy of 3.9 eV when starting with adsorbed adsorbed N,
while the reverse reaction has a relative energy barrier of
9.2 eV. The same effect is seen in divacancy systems,
where pyridinic inclusions (Fig. 5b) are formed at a lower
relative energy barrier of 1.3 eV, and are removed only
after introduction of at least 8.2 eV. Pyrrolic inclusions
(Fig. 5c), which outnumbered pyridinic inclusions in XPS
measurements, are formed in divacancy defects with a
slightly higher energy barrier of 5.8 eV, with a reverse
reaction requiring 9.9 eV. The lower energy of pyrrolic N
in graphene provides corroborating evidence to explain its
3
higher concentration in XPS measurements.
excitation mechanism.
The prevalence of pyrrolic
inclusions can be attributed to the proximity of the energy
of the transition state in the aromatic introduction reaction
to the 6 eV provided by the de-excitation of N 2 (A) and to
the lower overall energy of the pyrrolic configuration,
leading to increased overall stability.
4. Acknowledgments
The authors would like to acknowledge the vital
technical expertise of Dr. Josianne Lefebvre for her help
with XPS analysis, and Dr. Samir Elhouatik for his help
with access to the Raman facilities. The authors also
gratefully recognize the scientific and technical input of
Minh Nguyen, Dr. Pierre Lévesque, and Prof. Richard
Martel. The Canadian Natural Science and Engineering
Research Council provided funding for this work.
Fig. 5. (a) NEB chain of states model for the introduction
of adsorbed N into a pyridinic inclusion in a single point
vacancy in graphene. The energies of the states are
defined relative to the energy of the initial state, a single
adsorbed N atom siting in an minimum energy position
over the graphene lattice. (b) NEB chain of states model
for the introduction of adsorbed N into a pyridinic
inclusion in a divacancy point defect in graphene. (c)
NEB chain of states model for the introduction of
adsorbed N into a pyrrolic inclusion in a divacancy point
defect in graphene.
The analysis described above provides further
evidence that a single de-excitation collision between an
excited N 2 (A) metastable state could provide sufficient
energy to promote a nitrogen adatom into the transition
state and allow it to subsequently form an aromatic state.
The simulation further shows that the reverse reaction
cannot be stimulated in the same way, because the
activation energy for the reverse reaction exceeds the
de-excitation energy released during the N 2 (A) third-body
4
5. References
[1] Wang B, Pantelides ST. Controllable healing of
defects and nitrogen doping of graphene by CO and
NO molecules. Phys Rev B 2011;83:245403.
doi:10.1103/PhysRevB.83.245403.
[2] Gao H, Song L, Guo W, Huang L, Yang D, Wang F,
et al. A simple method to synthesize continuous
large area nitrogen-doped graphene. Carbon N Y
2012;50:4476–82.
doi:10.1016/j.carbon.2012.05.026.
[3] Ito Y, Christodoulou C, Nardi MV, Koch N,
Sachdev H, Müllen K. Chemical vapor deposition of
N-doped graphene and carbon films: the role of
precursors and gas phase. ACS Nano 2014;8:3337–
46. doi:10.1021/nn405661b.
[4] Wang X, Li X, Zhang L, Yoon Y, Weber PK, Wang
H, et al. N-doping of graphene through
electrothermal reactions with ammonia. Science
2009;324:768–71. doi:10.1126/science.1170335.
[5] Xu J, Jang SK, Lee J, Song YJ, Lee S. The
Preparation of BN-Doped Atomic Layer Graphene
via Plasma Treatment and Thermal Annealing. J
Phys
Chem
C
2014;118:22268–73.
doi:10.1021/jp504773h.
[6] Zhao W, Höfert O, Gotterbarm K, Zhu JF, Papp C,
Steinrück H-P. Production of Nitrogen-Doped
Graphene by Low-Energy Nitrogen Implantation. J
Phys
Chem
C
2012;116:5062–6.
doi:10.1021/jp209927m.
[7] Tang Y-B, Yin L-C, Yang Y, Bo X-H, Cao Y-L,
Wang H-E, et al. Tunable band gaps and p-type
transport properties of boron-doped graphenes by
controllable ion doping using reactive microwave
plasma.
ACS
Nano
2012;6:1970–8.
doi:10.1021/nn3005262.
[8] Lin Y-P, Ksari Y, Prakash J, Giovanelli L,
Valmalette J-C, Themlin J-M. Nitrogen-doping
processes of graphene by a versatile plasma-based
method.
Carbon
N
Y
2014;73:216–24.
doi:10.1016/j.carbon.2014.02.057.
[9] U.S. Provisional Patent Application: Stafford L,
O-20-1
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
Vandsburger L, Leonelli R, Afonso-Ferreira J.
Methods for atom incorporation into materials using
a plasma afterglow, USPTO No. 62/028,789
Yu V, Whiteway E, Maassen J, Hilke M. Raman
spectroscopy of the internal strain of a graphene
layer grown on copper tuned by chemical vapor
deposition. Phys Rev B 2011;84:205407.
doi:10.1103/PhysRevB.84.205407.
Afonso Ferreira J, Stafford L, Leonelli R, Ricard A.
Electrical characterization of the flowing afterglow
of N2 and N2/O2 microwave plasmas at reduced
pressure. J Appl Phys 2014;115:163303.
doi:10.1063/1.4872468.
Moisan M, Zakrzewski Z, Pantel R. The theory and
characteristics of an efficient surface wave launcher
(surfatron) producing long plasma columns. J Phys
D Appl Phys 1979;12:219–37. doi:10.1088/00223727/12/2/008.
Guerra V, Tatarova E, Dias FM, Ferreira CM. On
the self-consistent modeling of a traveling wave
sustained nitrogen discharge. J Appl Phys
2002;91:2648. doi:10.1063/1.1446229.
Boudam MK, Saoudi B, Moisan M, Ricard A.
Characterization of the flowing afterglows of an N 2
–O 2 reduced-pressure discharge: setting the
operating conditions to achieve a dominant late
afterglow and correlating the NO β UV intensity
variation with the N and O atom densities. J Phys D
Appl Phys 2007;40:1694–711. doi:10.1088/00223727/40/6/019.
Kang N, Lee M, Ricard A, Oh S. Effect of
controlled O2 impurities on N2 afterglows of RF
discharges. Curr Appl Phys 2012;12:1448–53.
doi:10.1016/j.cap.2012.04.009.
Wang Y, Shao Y, Matson DW, Li J, Lin Y.
Nitrogen-doped graphene and its application in
electrochemical
biosensing.
ACS
Nano
2010;4:1790–8. doi:10.1021/nn100315s.
Ferreira JA, Nguyen HPT, Mi Z, Leonelli R,
Stafford L. Improvement of the emission properties
from InGaN/GaN dot-in-a-wire nanostructures after
treatment in the flowing afterglow of a microwave
N2 plasma. Nanotechnology 2014;25:435606.
doi:10.1088/0957-4484/25/43/435606.
Kumar A, Ganguly A, Papakonstantinou P. Thermal
stability study of nitrogen functionalities in a
graphene network. J Phys Condens Matter
2012;24:235503.
doi:10.1088/09538984/24/23/235503.
Fally F, Doneux C, Riga J, Verbist JJ.
Quantification of the functional groups present at
the surface of plasma polymers deposited from
propylamine, allylamine, and propargylamine. J
Appl
Polym
Sci
1995;56:597–614.
doi:10.1002/app.1995.070560509.
Burg P, Fydrych P, Cagniant D, Nanse G, Bimer J,
Jankowska A. The characterization of nitrogenenriched activated carbons by IR, XPS and LSER
O-20-1
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
methods. Carbon N Y 2002;40:1521–31.
doi:10.1016/S0008-6223(02)00004-0.
Nakada K, Ishii A. Migration of adatom adsorption
on graphene using DFT calculation. Solid State
Commun
2011;151:13–6.
doi:10.1016/j.ssc.2010.10.036.
Dresselhaus MS, Dresselhaus G, Saito R, Jorio A.
Raman spectroscopy of carbon nanotubes. Phys Rep
2005;409:47–99.
doi:10.1016/j.physrep.2004.10.006.
Bernard S, Whiteway E, Yu V, Austing DG, Hilke
M. Probing the experimental phonon dispersion of
graphene using ^{12}C and ^{13}C isotopes. Phys
Rev
B
2012;86:085409.
doi:10.1103/PhysRevB.86.085409.
Sharifi T, Nitze F, Barzegar HR, Tai C-W,
Mazurkiewicz M, Malolepszy A, et al. Nitrogen
doped multi walled carbon nanotubes produced by
CVD-correlating XPS and Raman spectroscopy for
the study of nitrogen inclusion. Carbon N Y
2012;50:3535–41.
doi:10.1016/j.carbon.2012.03.022.
Luo Z, Lim S, Tian Z, Shang J, Lai L, MacDonald
B, et al. Pyridinic N doped graphene: synthesis,
electronic structure, and electrocatalytic property. J
Mater
Chem
2011;21:8038.
doi:10.1039/c1jm10845j.
Wei D, Liu Y, Wang Y, Zhang H, Huang L, Yu G.
Synthesis of N-doped graphene by chemical vapor
deposition and its electrical properties. Nano Lett
2009;9:1752–8. doi:10.1021/nl803279t.
Wang B, Tsetseris L, Pantelides ST. Introduction of
nitrogen with controllable configuration into
graphene via vacancies and edges. J Mater Chem A
2013;1:14927. doi:10.1039/c3ta13610h.
Hanwell MD, Curtis DE, Lonie DC, Vandermeersch
T, Zurek E, Hutchison GR. Avogadro: an advanced
semantic chemical editor, visualization, and analysis
platform.
J
Cheminform
2012;4:17.
doi:10.1186/1758-2946-4-17.
5