Intimately coupled hybrid of graphitic carbon nitride nanoflakelets with reduced graphene oxide for supporting Pd nanoparticles: A stable nanocatalyst with high catalytic activity towards formic acid and methanol electrooxidation

Article abstract of DOI:10.1016/j.electacta.2016.03.169

A novel nitrogen-rich support material (CNNF-G) consisting of graphitic carbon nitride (g-C₃N₄) nanoflakelets (CNNF) and reduced graphene oxide (rGO) is designed and fabricated for loading Pd nanoparticles. Structural characterizatio

Full text of DOI:10.1016/j.electacta.2016.03.169

Accepted Manuscript
Title: Intimately coupled hybrid of graphitic carbon nitride
nanoflakelets with reduced graphene oxide for supporting Pd
nanoparticles: a stable nanocatalyst with high catalytic activity
towards formic acid and methanol electrooxidation
Author: Wenyao Zhang Qiushi Yao Xiaodong Wu Yongsheng
Fu Kaiming Deng Xin Wang
PII:
S0013-4686(16)30753-8
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.electacta.2016.03.169
EA 26999
To appear in:
Electrochimica Acta
Received date:
Revised date:
Accepted date:
16-1-2016
23-3-2016
28-3-2016
Please cite this article as: Wenyao Zhang, Qiushi Yao, Xiaodong Wu,
Yongsheng Fu, Kaiming Deng, Xin Wang, Intimately coupled hybrid of
graphitic carbon nitride nanoflakelets with reduced graphene oxide for
supporting Pd nanoparticles: a stable nanocatalyst with high catalytic activity
towards formic acid and methanol electrooxidation, Electrochimica Acta
http://dx.doi.org/10.1016/j.electacta.2016.03.169
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.

Intimately coupled hybrid of graphitic carbon nitride nanoflakelets with reduced graphene
oxide for supporting Pd nanoparticles: a stable nanocatalyst with high catalytic activity
towards formic acid and methanol electrooxidation
a
b
a
a,
b
a,
Wenyao Zhang , Qiushi Yao , Xiaodong Wu , Yongsheng Fu *, Kaiming Deng *, Xin Wang *
a
Key Laboratory for Soft Chemistry and Functional Materials of Ministry Education, Nanjing
University of Science and Technology, 210094, Nanjing, China.
b
Department of Physics, Nanjing University of Science and Technology, 210094, Nanjing,
China.
*
To whom correspondence should be addressed. Tel.: +86-25-8431-5943; fax: +86-25-8431-
5054. E-mail address: wangx@njust.edu.cn (X. Wang), fuyongsheng0925@163.com (Y. S. Fu),
kmdeng@njust.edu.cn (K. M. Deng)
1

Graphical abstract
Research Highlights
►
Intimately coupled hybrid of g-C
3
4
N nanoflakelets with rGO is fabricated via an in situ
chemical synthesis approach.
►
►
►
The detailed forming process of CNNF-G composite is carefully investigated.
The Pd-CNNF-G composite exhibits exceptional catalytic activity and durability.
The real fuel cells have been assembled to evaluate the performance of the electrocatalysts.
2

►
The effects of g-C
3
N nanoflakelets on Pd atom were investigated by DFT computations.
4
Abstract
A novel nitrogen-rich support material (CNNF-G) consisting of graphitic carbon nitride (g-
) nanoflakelets (CNNF) and reduced graphene oxide (rGO) is designed and fabricated for
loading Pd nanoparticles. Structural characterizations indicates that the CNNF is formed via
splitting decomposition of the g-C polymer on rGO at higher temperatures and the resulting
C
3
N
4
3
N
4
CNNF is intimately coupled to the rGO sheets. The CNNF can provide more exposed edge sites
and active nitrogen species for the high dispersion of Pd NPs. It is found that the Pd NPs with an
average diameter of 3.92 nm are uniformly dispersed on CNNF-G sheets. DFT computations
reveal that CNNF can trap Pd adatom and thus act as a Pd nucleation site at which Pd atoms tend
to accumulate to form Pd clusters. The Pd-CNNF-G nanocatalyst exhibits excellent
electrocatalytic activity for both formic acid and methanol oxidation reactions, including large
electrochemically active surface area (ECSA) values, significantly high forward peak current
densities, and reliable stability and durability, far outperforming the Pd-graphene, commercial
activated carbon-supported Pd catalyst or Pd-carbon nanotubes. Such a stable Pd/CNNF-G
nanocatalyst may bring new design opportunities for high-performance direct formic acid fuel
cell (DFAFC) and direct methanol fuel cell (DMFC) in the future.
Keywords: graphene, graphitic carbon nitride nanoflakelets, electrocatalysis, formic acid
oxidation, methanol oxidation
3

1
. Introduction
As the depletion of fossil fuels and the increasingly environmental pollution problems, a great
deal of research has been performed on how to obtain clean and renewable energy sources as
well as their devices [1]. In such an active field, fuel cells have been considered as one of the
most promising candidates in recent years because they can convert chemical energy from
hydrogen-rich fuels to electrical energy with little pollutant emission and high energy conversion
efficiency [2, 3]. Among the various fuel cell technologies, the direct formic acid fuel cell
(
DFAFC) and direct methanol fuel cell (DMFC) have attracted plenty of interest due to the
storage and transport of these liquid fuels is much easier and safer than that of hydrogen because
it does not need to be operated at low temperatures and high pressures [4, 5]. It is well known
that the nature of anode catalysts may heavily affect the performance of DFAFC or DMFC [6],
and then much more attention have been dedicated to develop new approaches for preparation of
electrochemical active metal NPs supported on advanced materials to improve the performance
[7-9].
Recently, Pd-based catalysts have aroused special interest owing to their lower cost, higher
activity and stronger tolerance to CO when compared with Pt-based catalysts [10, 11]. As has
been reported, the structure of Pd NPs, including the size distribution and morphology, can
greatly influence the catalytic performance for formic acid or methanol oxidation [12-14].
Inspired by this, tremendous efforts have been made to obtain highly dispersed Pd NPs with a
narrow size distribution supported on conductive carbon materials, such as carbon black [15],
carbon nanofibers [16], and carbon nanotubes [17] and so on. Apart from other carbon materials,
graphene sheets display a myriad of unique chemical and physical properties and can be used as
an ideal support material to load and disperse noble metal NPs for high performance
4

electrocatalytic or electrochemical devices [12, 18-20]. Moreover, based on intensive theoretical
calculations and detailed experiments it has been demonstrated that nitrogen doping is an
effective approach to tailor the electronic properties of graphene [21-23]. Graphitic carbon
nitride (g-C
3
N ) is an appealing class of nitrogen-rich carbon materials, which possesses not only
4
unique features including ideal two-dimensional structure, ultrahigh nitrogen content, large
aspect ratio, excellent mechanical resistance and anticorrosion ability in oxidative environments,
but also a promising potential in fields ranging from heterogeneous catalysis, energy conversion
and storage, to environmental applications [9, 24, 25]. Especially, because the electron accepting
ability of the nitrogen atoms can polarize the adjacent carbon atom and enhance the bonding
affinity with adsorbed OOH [26, 27], g-C
electrocatalytic activity for oxygen-reduction reaction (ORR) as a metal-free catalyst [28, 29].
On the other hand, we recently developed a covalently coupled hybrid of g-C -rGO for
supporting Pd nanoparticles [24]. It is found that incorporation of g-C into rGO plays an
3
N -based materials display an excellent
4
3
N
4
3
N
4
essential role in changing the nucleation and growth behaviours of Pd nanoparticles, thus
resulting in greatly enhanced electrochemical performances of the catalysts for fuel oxidation
reactions. Nevertheless, g-C
3
N
4
suffers from relatively low electrical conductivity due to its
sheets fully covered on the surface of rGO may
intrinsic porous microstructure, thus the g-C
3
N
4
disturb the electron transport pathways during the electrochemical process.
Inspired by this point, we design a facile approach to fabricate an intimately coupled hybrid of
g-C
3
N nanoflakelets with rGO (CNNF-G), which can expose much more edge sites and active
4
nitrogen species, facilitating the nucleation and growth of Pd nanocrystals as well as the high
dispersion of the resulting nanoparticles. Such unique features are not only favorable for
providing fast and barrier-free channels for rapid diffusion of the electrolyte, but also beneficial
5

for the removal of the incompletely oxidized carbonaceous intermediates from the catalyst
surface during the fuel oxidation process. As a consequence, the Pd-CNNF-G catalyst displays
an excellent electrocatalytic performance towards both formic acid and methanol
electrooxidation, including high electrocatalytic activity, unusual poison tolerance, and long-term
durability, all of which are superior to the Pd-graphene, Pd-CNT and commercial Pd-activated
carbon (Pd-AC) catalysts.
2
. Experimental section
2
.1 Preparation of Pd-CNNF-G catalyst
Graphite oxide was synthesized from powdered flake graphite by a modified
Hummers’ method [30]. Specifically, 100 mg GO was dispersed in 100 mL water by
ultrasonication. Then, melamine (200 mg) was added into the suspension and the reaction
mixture was heated to 80 °C for 2 h to form melamine chemically modified GO sheets
(
melamine-GO) via a covalent C-N bond. After the water of the mixture was removed
through freeze drying, the dried melamine-GO product was then transferred into a
crucible and first heated at 550 °C for 2 h under a nitrogen flow, resulting in the
polymerization of melamine to form g-C
of GO to yield the reduced GO (rGO). Then the sample was heated to the final calcination
temperature of 800 °C for 2 h. At such a high temperature, the g-C polymer
nanosheets can be split into many smaller nanoflakelets (CNNF) which are intimately
coupled with rGO to form the g-C nanoflakelets/rGO hybrid instead of decomposition
3
N
4
polymer nanosheets and the deoxygenation
3
N
4
3
N
4
due to the protection of graphene. For comparison, the same method was used to prepare
rGO without melamine.
6

Preparation of the Pd NPs-CNNF-G catalyst with 20 wt% Pd was similar to our
previous work [24]. Briefly, CNNF-G (10 mg) was ultrasonically dispersed in 100 mL of
ethylene glycol to form a uniform suspension. Then 0.025 mL of 0.73 M Pd(NO
3
) was
2
introduced to the suspension with magnetic stirring for 2 h. Subsequently, 0.425 mL of
hydrazine hydrate solution (50 wt% in water) was added into the mixture with stirring for
another 4 h under ambient conditions. The product, denoted as Pd-CNNF-G, was
centrifuged, washed with deionized water and freeze-dried. For comparison, Pd NPs
supported on rGO and CNT were prepared by the same procedure and were denoted as
Pd-rGO and Pd-CNT, respectively. The exact loadings of Pd in Pd-CNNF-G, Pd-g-C
3
4
N -
rGO, Pd-rGO and Pd-CNT samples were measured at 16.08, 17.9, 16.12 and 17.5 wt%,
respectively, which agree well with their theoretical metal loading (20.0 wt%). The
commercial activated carbon-supported palladium catalyst (Alfa Aesar 38308, Pd 20%,
denoted as Pd-AC) was also used without any further treatment.
2
.2 Characterization
Powder X-ray diffraction (XRD) analyses were performed using a Bruker D8 Advance
diffractometer with Cu Kα radiation (λ ≈ 1.54 Å). Fourier Transform-IR Spectroscopy (FT-IR)
was recorded on a Nicoletis10 instrument (Thermo Nicolet Corporation). Raman spectra of the
samples were collected with a LABRAM Aramis Raman Microprobe at wavelength of 532 nm.
Field emission scanning electron microscopy (FE-SEM) measurement and elemental mapping
were performed on a Hitachi S-3400N II microscope. Transmission electron microscopy (TEM)
and energy dispersive X-ray (EDX) spectroscopy were carried out on a JEOL JEM-2100
microscope. X-Ray photoelectron spectra (XPS) were recorded using an RBD upgraded PHI-
5000C ESCA system (Perkin Elmer) with Al Kα radiation (hν = 1486.6 eV). The Brunauer-
7

Emmett-Teller (BET) surface area was obtained by nitrogen sorption experiments conducting at
7 K using a Micromeritics TriStar II 3020 automated gas adsorption analyzer. The metal
7
loading of the catalysts was accurately determined with inductively coupled plasma atomic
emission spectrometry (ICP-AES, Varian Vista-AX). The TG analyses were carried out
on a TGA/SDTA851e thermogravimetric analyzer under a N atmosphere from 30 to
2
-
1
9
2
00 °C at a heating rate of 5 °C min .
.3 Electrochemical activity measurements
Electrocatalytic activities of various catalysts were measured in a conventional three-
electrode cell with a CHI 760E electrochemical workstation. A glassy carbon (GC) disk
3 mm in diameter) was used as working electrode, which was polished successively with
.0, 0.3 and 0.05 μm alumina slurries before each experiment. Catalysts were applied
onto the disk in the form of inks. The ink was a pure water-isopropyl alcohol-Nafion 117
(
1
-
1
(
10: 10: 1 volume ratio) solution with catalyst concentration approximately 2 mg mL . A
Pt wire and a saturated calomel electrode (SCE) were served as the counter electrode and
the reference electrode, respectively.
All electrochemical measurements were performed with the electrolytes of 0.5 M
HCOOH in 0.5 M H
2
SO
4
solutions and 1 M CH OH in 1 M NaOH solutions at 25 °C for
3
DFAFC and DMFC measurements, respectively. The solutions were deaerated by pure N
2
for at least 20 min prior to any measurements. To facilitate comparison of the catalytic
activity across samples, the Pd loading on the working electrode were determined by ICP-
AES measurements. All electrochemical data were normalized to the actual loading of Pd
and all the potentials were expressed with the reversible hydrogen electrode (RHE) by
8

⊖
calibrating the SCE to the RHE scale according to the Nernst equation: E = E -0.05916
*
pH.
2
.4 Membrane electrode assembly preparation
Nafion 117 (DuPont) was used as the proton exchange membranes and the pre-
treatment of the Nafion membrane was accomplished by successively treating it in a 5
wt.% H solution at 80 °C, distilled water at 80 °C, 8 wt.% H SO solution at 80 °C
2
O
2
2
4
and then in distilled water at 80 °C again, for 30 min each step. Prior to the fabrication of
hot-pressed Membrane electrode assemblies (MEA) and catalyst-coated membrane, a
carbon cloth was used as Gas diffusion layer for current collector and it also assists in
water management.
2
MEA with a 2.25 * 2.25 cm active cell area was fabricated using a ‘direct paint’
technique applied to the catalyst layer. The ‘catalyst inks’ were prepared by dispersing the
catalyst nanoparticles into appropriate amounts of ultrapure water, isopropyl alcohol and
5
wt.% Nafion ionomer solution. For all MEAs in this study, the cathode consisted of
-2
commercial Pt/C catalyst at a standard metal loading of 1 mg cm . The anode consisted
-2
of carbon supported Pd catalysts with a metal loading of 0.5 mg cm . A carbon cloth
diffusion layer (E-TeK) was placed on top of both the cathode and anode catalyst layers.
A single cell test fixture consisted of machined graphite flow fields with direct liquid
feeds and gold plated copper plates to avoid corrosion (Hephas Energy Co., ltd.). Hot-
pressing was conducted at 130 °C and 10 MPa for 120 s.
2
.5 Structure models and computational methods
Our calculations were based on DFT with generalized gradient approximation (GGA) [31] for
exchange correlation potential. Perdew-Burke-Ernzerholf (PBE) functional was used for the
9

GGA as implemented in Vienna Ab-initio Simulation Package (VASP) [32]. In order to simulate
the CNNF-G nanocomposite, a 3×3 graphene supercell was used to match the 1×1 tri-s-triazine
unit cell with a small 1.5% strain in graphene. The vacuum space in the third dimension (Z) is 15
Å, which is enough to separate the interaction between periodic images. The structures were
relaxed without any symmetry constraints with a cutoff energy of 500 eV, and the convergence
-
5
criteria of energy and force were set to 1×10 eV and 0.01 eV/Å, respectively. Reciprocal space
was represented by a Monkhorst-Pack special k-point scheme [33] with 3×3×1, and 5×5×1 grid
meshes for structure relaxation self-consistent field (SCF) computations, respectively.
3
. Result and discussion
3
.1 Structural characterization
The structures of the samples were characterized by means of Fourier transform
infrared spectroscopy (FT-IR), Raman spectroscopy and powder X-ray diffraction (XRD).
As shown in Fig. S1, in comparison with rGO, CNNF-G shows a typical breathing mode
-
1
of tri-s-triazine units at 807 cm as well as its stretching modes in the range of 1000-1600
-1
cm [34, 35]. With regard to the intense peak of tri-s-triazine units, it is speculated that
the g-C polymer nanosheets on rGO can be split into some smaller nanoflakelets,
which intimately interact with rGO to form the g-C nanoflakelets/rGO nanohybrid
CNNF-G) at higher temperatures.
Raman spectra (Fig. 2a) provide the further information on the structure of the resulting
smaller g-C nanoflakelets in CNNF-G. The G bands of the GO, rGO, and CNNF-G appeared
3
N
4
3
N
4
(
3
N
4
-1
at 1590.2, 1580.6 and 1582.5 cm , respectively. As compared with GO, the significant downshift
of the G band of rGO can be ascribed to the restoration of the π-conjugated structure during
pyrolysis [36]. While slight upshift of the G band of CNNF-G can be observed relative to rGO.
1
0

As will be seen in the XPS analysis, p-type doping is the dominant part in CNNF-G and
therefore its G band shifts to higher frequency as compared to rGO, which is consistent with the
phenomena of upshift in p-type doped and downshift in n-type CNTs and graphite [37-39]. Apart
from the shift of G band, the intensity ratio of D band to G band (I
feature that is often used to estimate the disordering degree of graphene-based materials. It is
found that the I /I ratios are 1.12, 1.24 and 1.40 for GO, rGO and CNNF-G, respectively. The
increase of I /I ratio is due to the effect of g-C nanoflakelets on the structural disorder and
the high baseline of g-C in the region of the D band.
Fig. 2b shows the X-ray diffraction (XRD) patterns of GO, Pd-rGO, Pd-CNNF-G and
g-C . The diffraction peak of GO at 10.5° can be indexed to (002) facets,
D
/I
G
) is another characteristic
D
G
D
G
3
N
4
3
N
4
3
N
4
corresponding to a layered structure with a basal spacing of 0.84 nm [40]. Upon the
formation of Pd-rGO or Pd-CNNF-G nanocompsites, the characteristic (002) peak shifts
to a higher angle as a result of the deoxygenation/reduction of GO during the pyrolysis
process [41]. Pd-CNNF-G shows a broad peak of (002) diffraction at 26.5°,
corresponding to a d-spacing of 3.40 Å calculated according to Bragg's equation, slightly
larger than that of Pd-rGO (3.36 Å). This indicates that the modification of CNNF on the
surface of rGO may cause the loosening of layers along the c-axis with fewer structural
defects. Besides, no typical diffraction peak of rGO (002) or g-C
3
4
N (002) is observable,
revealing that g-C nanoflakelets are evenly growing on both sides of rGO nanosheets
3
N
4
and. Furthermore, the intensive diffraction peaks at 39.8°, 46.2° and 67.7° are
corresponding to the (111), (200) and (220) planes of the face-centered-cubic (fcc) Pd
lattices (JCPDS no. 46-1043), respectively, suggesting a good crystallinity of Pd NPs in
the hybrid.
1
1

The morphologies of various catalysts were characterized by TEM. As shown in Fig.
a, CNNF-G has a two-dimensional structure with some wrinkles, which is similar to the
3
typical morphology of graphene. While upon the graphene sheets, the homogeneous
distribution of g-C nanoflakelets can be recognized only when observing carefully due
3
N
4
to the similar transparency of CNNF and graphene. To further confirm this, the
corresponding energy dispersive X-ray (EDX) spectrum was also carried out (Fig. S2),
and there are only four elements in the product, C, N, O, and Cu (the Cu signals come
from the copper grid), and the nitrogen content of the CNNF was calculated to be 4.5 at
%
. On the other hand, it is clearly seen from Fig. 3b that metallic Pd NPs are uniformly
dispersed on the surface of CNNF-G sheets while Fig. 3c reveals lattice fringes with
crystal plane distances of 0.223 nm, attributed to the spacing of the (111) planes in face-
centered cubic (fcc) Pd, respectively. Selected-area electron diffraction of the Pd-CNNF-
G displays a set of ring-like diffraction patterns (Fig. 3d), corresponding to the Pd (111),
(
200), (220) planes, which correlate well with the XRD results. Further inspection by
elemental mapping discloses that the Pd-CNNF-G architecture contains C, N and Pd as
the main components, and the elements are homogeneously distributed over the entire
area of the catalysts (Fig. 3e). This could be attribute to the presence of g-C
3
N
4
nanoflakelets on the surface of rGO, which possesses π-bonded planar C-N-C-layers
along with their incompletely condensed amino groups in the carrier, making it possible
to enhance the adhesion between Pd NPs and the support and effectively stabilize highly
dispersed Pd NPs and prevent their agglomeration [42]. In contrast, as presented in Fig.
S3, it can be seen that Pd NPs supported on rGO, CNT or AC have larger sizes with
aggregation to some extent, suffering from their relatively weak interactions of metal and
1
2

support. The statistical size distribution histograms obtained by measuring 200 randomly
chosen particles (Fig. 3c, Fig. S3) demonstrate that the average particle sizes are about 3.9
nm, 5.8 nm, 7.0 nm and 4.6 nm for Pd-CNNF-G, Pd-CNT, Pd-rGO and Pd-AC, further
suggesting that the CNNF-G facilitates the formation of smaller-sized Pd NPs with higher
dispersity. Moreover, the adsorption and desorption isotherms of nitrogen on the Pd-
CNNF-G or Pd-rGO hybrid exhibit a typical IUPAC type IV isotherm, implying the
existence of mesoporous structure (Fig. S4). Compared with Pd-rGO, the Pd-CNNF-G
2
-1
hybrid possesses a much higher specific surface area of 360.4 m g with an average pore
3
-1
diameter of 7.55 nm and a pore volume of 0.926 cm g , thereby not only offering rich
absorption sites but facilitating charge transfer across the electrode/electrolyte interface as
well.
X-ray photoelectron spectroscopy (XPS) was used to further characterize the elemental
composition and the valent state of the elements in Pd-CNNF-G and Pd-rGO. Fig. 4a presents
the survey spectrum of the XPS analysis. No peaks of other elements except C, O and Pd were
observed for Pd-rGO, by contrast, there was the co-existence of C, N, O and Pd in Pd-CNNF-G.
Percentages of the individual elements based on XPS analysis are listed in Table 1. The C1s XPS
2
spectrum of Pd-rGO can be deconvoluted into three components (Fig. 4b): the sp -hybridized C-
C (284.6 eV), C-OH (286.1 eV) and HO-C=O (289.0 eV) groups, respectively [43]. While the
C1s XPS spectrum of Pd-CNNF-G gives two additional peaks at 285.7 eV and 287.3 eV, which
can be assigned to C=N and C-N bonding configurations [24], revealing the existence of C-N
species in the Pd-CNNF-G nanocomposites. The broad and asymmetric N 1s spectrum can be
divided into three peaks at 398.3, 399.5 and 400.8 eV (Fig. 4c), corresponding to pyridinic N,
pyrrolic N and graphitic N, which are well consistent with the g-C
3
N sample [44, 45]. It should
4
1
3

be noted that the content of the pyridinic N together with pyrrolic N in the N-containing species
reaches as high as 62.4 % (Table S2) and such N atoms can serve as metal-anchoring sites
facilitating the nucleation and subsequent growth of metal nanoparticles [26]. Compared with the
g-C
3
N
4
nanosheets (Fig. S5), the increase of the graphitic N gives an indirect evidence of the
nanoflakelets and the covalent interaction between g-C and rGO,
formation of the g-C
3
N
4
N
3 4
which makes CNNF-G more stable during the electrocatalysis process. Moreover, theoretical
studies of N-doped carbon nanotubes indicate that the substitution of C atoms with graphitic N is
considered as n-type doping, whereas pyrrolic and pyridinic N may be taken as p-type doping
[
46-48]. Accordingly, p-type doping is dominant in the Pd-CNNF-G system. As shown in Fig.
d, the splitting pattern of Pd 4d band contains two pairs of doublets: the intensive doublet
335.8 and 341.0 eV) may be assigned to metallic Pd and another doublet (337.1 and 342.9 eV)
3
(
corresponds to the +2 oxidation state of Pd [42, 49]. Similar results were obtained for Pd-CNNF-
G, however, its Pd 3d binding energies were slightly shifted to lower values (Table S4).
According to the previous study, lower binding energies of supported Pd species are related to
stronger interaction between support and Pd NPs [50]. As will be seen later, the stronger
coupling between Pd NPs and CNNF-G gives rise to enhanced electrocatalytic performance
towards methanol and formic acid oxidation.
3
.2 Theoretical calculations
In order to better understand the effects of g-C
3
4
N nanoflakelets on Pd atom adsorption,
DFT computations were performed to investigate the adsorption energy and geometry of
Pd atom on graphene and CNNF-G. An adsorption location map and the optimized
geometry structures for Pd adsorption on CNNF-G are illustrated in Table. S6. As was
found for CNNF-G, Pd atoms preferentially bind at hollow sites with strongest
1
4

adsorption, Eads= 2.82 eV, roughly thrice that for Pd adsorption on graphene (Eads= 1.035
eV). Interesting, the g-C nanoflakelets’ effect on Pd adatoms is unique. Pd adatoms
3
N
4
moved significantly from their starting positions and relaxed toward the interlayer
between CNNF and rGO to forming a most stable structure (Fig. S8a-b). This well traps
Pd adatom, making it able to keep in contact with the substrate of rGO. Subsequently, it
could act as a Pd nucleation site at which Pd atoms will tend to accumulate and form Pd
clusters (Fig. S8c-d).
When the rGO sheet modified with CNNF, the entire CNNF-G are slightly charged with
respect to the graphene sheet, and upon Pd adsorption, it acts to redistribute charge. As Fig. 5
show, an intensive electron increase zone (yellow isosurface) can be observed from the
differential charge density. Combine with the overlap located at the Fermi energy in the PDOS
(
Fig. S9), it is suggestive of that the CNNF-G well acts to tether Pd adatom to the substrate, thus
reducing Pd cluster migration and agglomeration, which is attributed to the interaction of the d
2
orbitals (Pd) and the sp lone pair orbitals (N). Moreover, the electronic effect of formic acid or
methanol molecule on Pd-CNNF-G and Pd-rGO was also considered (Fig. S10).
3
.3 Electrochemical analysis
The electrochemically active surface area (ECSA) provides straightforward
information regarding the number of electrochemically active sites per unit mass of
catalyst [51]. Fig. 6 shows the CO stripping voltammograms for different electrodes in
-
1
0
.5 M H
2
SO solution at a scan rate of 50 mV s at 25 °C. A well-defined stripping peak
4
of COad at the potential of ca. 0.95 V can be observed for each catalyst. The ECSA values
are estimated supposing that the charge required to oxidize a full monolayer of CO on Pd
-2
is 420 mC cm [52]. As listed in Table 2, the Pd-CNNF-G catalyst exhibits an ECSA
1
5

2
-1
2
value as high as 91.2 m g , which is more competitive than the one of Pd-rGO (60.5 m
-
1
2
-1
2
-1
g ), Pd-AC (39.8 m g ), Pd-CNT (56.3 m g ), or recently reported state-of-the-art Pd-
2
-1
based nanomaterials such as formic acid-reduced Pd/graphene (56 m g ) [12], Pd/DNA-
2
-1
2
-1
modified graphene (82.2 m g ) [53], Pd/WC modified graphene (81.22 m g ) [54],
2
-1
oriented (111) Pd nanocrystals/graphene (29 m g ) [19], porous worm-like Pd nanotubes
2
-1
2
-1
(
17.7 m g ) [55], Pd/nitrogen-doped acetylene carbon black (82.7 m g ) [50], Pd-Au
2
-1
NPs/PDDA-functionalized graphene (58 m g ) [56], Pd-Cu bimetallic NPs/graphene
2
-1
2
-1
(
20.2 m g ) [57], porous Pt-Pd nanospheres/graphene (68.1 m g ) [58], Pd-Ni-P
2
-1
NPs/carbon black (62.9 m g ) [59]. It should be noted that the Pd-g-C
3
N
4
shows a low
2
-1
ECSA value of 6.9 m g due to the poor electrical conductivity of the g-C . In
3
N
4
addition, the electrocatalytic activities of as-obtained catalysts were also evaluated in an
alkaline medium. As illustrated in Fig. S11b, the potential limit of all CV curves is set at
1
.40 V, associating with the completion of formation of the first monolayer of the Pd (II)
oxide (Fig. S12), and a flat anodic (E ≈ 1.3 V) along with a strong cathodic (E ≈ 0.6 V)
peak are distinctly recognized, corresponding to the formation and reduction of Pd (II)
oxide, respectively [13, 60, 61] The ECSA values, calculated by the coulombic charge for
2
-1
reduction of Pd (II) oxide, follow an order of Pd-CNNF-G (89.4 m g ) ≈ Pd-g-C
3
4
N -rGO
2
-1
2
-1
2
-1
2
-1
(
90.6 m g ) > Pd-rGO (57.6 m g ) > Pd-CNT (54.3 m g ) > Pd-AC (39.6 m g ) > Pd-
2
-1
g-C
3
N (8.8 m g ), which is similar to that from CO stripping measurements in the acid
4
medium.
The Pd electrocatalytic behaviors for all Pd catalysts were characterized by cyclic
voltammogram measurement toward the formic acid oxidation reaction in acid medium
and the methanol oxidation reaction in alkaline medium. Fig. 7a depicts the CV responses
1
6

of the electrodes coated with the Pd-CNNF-G, Pd-g-C
3
N
4
-rGO, Pd-rGO, Pd-CNT, Pd-AC
or Pd-g-C catalyst for the electrochemical oxidation in 0.5 M H
3
N
4
2
SO + 0.5 M HCOOH
4
-
1
solution at 25 °C with a scan rate of 50 mV s . All the curves were calibrated for iR
losses (Fig. S13), and the ionic resistance (~20 Ω) from the solution was determined by
the EIS technique. It is well known that the electrooxidation of HCOOH typically follows
a dual-pathway mechanism [62, 63]:
(
1)
Direct dehydrogenation pathway
+
–
Pd + HCOOH → Pd + CO + 2H + 2e
2
(
2)
Indirect pathway
Pd + HCOOH → Pd-CO + H
O
2
+
–
Pd + H
2
O → Pd-OH + H + e
+
–
Pd-CO + Pd-OH → Pd + CO
Therefore, the current peak at ~ 0.4 V in the anodic scan may arise from the direct oxidation of
formic acid to CO , while the shoulder at ~ 0.8 V is related to the oxidation of both adsorbed CO
and HCOOH. Clearly, the Pd-CNNF-G catalyst exhibits an impressively high mass current
2
+ H + e
2
-1
-1
density of 1890 mA mg , which is more than 2 times as high as that of Pd-rGO (970 mA mg )
-1
-1
or Pd-CNT (830 mA mg ), and 3 times that of Pd-AC catalyst (515 mA mg ). Also, the mass
activity of Pd-CNNF-G is more competitive than those of recent state-of-the-art Pd-based
-1
nanostructures such as formic acid-reduced Pd/graphene (89.5 mA mg ) [12], Pd/DNA-modified
-1
-1
graphene (282.1 mA mg ) [53], Pd/WC modified graphene (360.5 mA mg ) [54], Pd/nitrogen-
-1
doped acetylene carbon black (1202 mA mg ) [50], Pd
3
Au
1
NPs/PDDA-functionalized graphene
580 mA mg ) [56]. Pd/ITO-CNTs (1365.4 mA mg ) [64], Ag S-hollow Pd nanocomposites
4.01 mA cm ) [65]. Furthermore, at a given oxidation current density (e.g. as indicated by the
-1
-1
(
(
2
-2
1
7

dashed line in Fig. S14a), the Pd-CNNF-G catalyst shows an obviously lower oxidation potential
than anyone of reference catalysts, implying that the formic acid electrooxidation occurred more
easily on Pd-CNNF-G [66]. The effects of the g-C
As shown in Fig. S15, the sample, obtained with the raw material's mass ratio of melamine to
GO of 1:2, reveals the optimal electrochemical activity. This is attributed that suitable g-C
3
N
4
load on the catalysts were also considered.
3
N
4
nanoflakelets can influence the spin density and charge distribution of neighboring carbon atoms
on the surface of graphene, which makes the CNNF-G support maintain acceptable conductivity
and considerable active sites, while further increasing the amount of g-C
may lead to a decrease in the electrocatalytic activity, which is due to the deteriorating electrical
conductivity derived from the intrinsic porous microstructure of the g-C
3
N in the composites
4
3 4
N .
In addition, Pd-CNNF-G also possesses a high catalytic activity in alkaline media (Fig.
b). In the presence of 1 M NaOH and 1 M CH OH, two irreversible current peaks can be
7
3
observed: one in the forward scan may be associated with the oxidation of freshly
chemisorbed species coming from methanol adsorption, while the reverse scan peak may
be assigned to the removal of the incompletely oxidized carbonaceous species formed
during the forward scan. The generally accepted mechanism for the electrooxidation of
methanol in alkaline medium, which involves the reaction of generation of formate ions,
can be summarized as follows [67-69]:
-
-
–
CH
3
OH + 5 OH → HCOO + 4 H O + 4e
2
methanol adsorbed on the electrode surface can undergo successive dehydrogenation.
However, the exact structure of the adsorbed intermediates is still unknown except for
(
CO)ads. The magnitude of the peak current in the forward scan is directly proportional to
electrocatalytic activity of Pd catalyst for methanol oxidation. As expected, Pd-CNNF-G
1
8

-
1
exhibits the highest peak current density of 1770 mA mg , which is much higher than
-
1
-1
those of Pd-g-C
3
N
4
-rGO (1550 mA mg ), Pd-rGO (835 mA mg ), Pd-CNT (725 mA
-
1
-1
67 mA mg-1
, or recent reported Pd-based
mg ), Pd-AC (535 mA mg ), Pd-g-C
3
N
4
(
)
-
1
nanostructures such as Pd-Cu bimetallic NPs/graphene (392.6 mA mg ) [57], Porous Pt-
-1
-
Pd nanospheres/graphene (705.8 mA mg ) [58], Pd-Ni-P NPs/carbon black (360 mA mg
1)
[59], Pd-SnO
/MWNTs (778.8 mA mg ) [70], Pd/graphene hydrogel on Ni foam (788
-1
2
-1
-1
mA mg ) [71], Pd
1
Ag
1
NPs/graphene (630 mA mg ) [72], Pd
1
Pt /graphite nanoplatelets
3
-
1
-1
(
385.2 mA mg ) [73], , Pd/CuO-TiO
2
(381 mA mg ) [74] (Table S7). Meanwhile, similar
to that of the formic acid oxidation, the onset potential of Pd-CNNF-G in alkaline
medium also suggests the significant enhancement in the kinetics of the methanol
oxidation reaction [14] (Fig. S14b).
Long-term stability is another important factor that affects the practical application of
electrocatalyst in fuel cell applications. It is known that some carboneous species, which
come from redox reactions on the surface of Pd NPs, readily cause the inactivation of
catalysts. Chronoamperometric measurement was used to appraise the electrochemical
durability of Pd-based catalysts. As shown in Fig. 7c, the polarization current of all the
catalysts decayed precipitously within the first 500 s due to the chemisorbed
carbonaceous species and then the hyperbolic decay slowly attained a pseudo-steady state
[
75]. This phenomenon is quite similar to that of the formic acid electrooxidation reaction
on the Pt-based electrocatalysts, suggesting the same poisoning mechanism of the
intermediate species coming from the formic acid oxidation on the electrocatalysts [76].
Upon the long-term operation, it can be seen from Fig. S16a that the Pd-CNNF-G
electrode retained 34.6 % of the initial current density after first 500 s, greatly surpassing
1
9

the electrodes based on Pd-g-C
3
4
N -rGO (30.4 %), Pd-rGO (26.4 %), Pd-CNT (17.2 %),
and Pd-AC (3 %). The larger retention of Pd-CNNF-G indicates the relatively lower
poisoning rate during the electrooxdiation reactions. Clearly, Pd-CNNF-G is able to
maintain the highest current and slowest decay rate for the whole duration of the scan,
exhibiting superior electrocatalytic activity and stability for formic acid oxidation of all
the investigated catalysts. It is worth mentioning that the Pd-C
decay after 500 s and retains 78.2 % for the whole time, suggesting that g-C
beneficial for improving the anti-poisoning property. Unfortunately, limited by the low
electrical conductivity, the g-C materials cannot be used alone in the field of
3
N
4
exhibits only 21.7 %
3
N
4
is
3
N
4
electrochemistry. Otherwise, the long-term stability for formic acid electrooxidation was
also considered at 0.25V, which is near the equilibrium potential (Fig. S17). As for
methanol oxidation, the variation of currents with time has a similar tendency (Fig. 7d,
Fig. S16b). Pd-CNNF-G shows the lowest current decay and retains the highest oxidation
current density over the testing time, which is also in good agreement with the results of
the CVs. Interestingly, when compared with the Pd-based catalyst supported on g-C
3
N
4
modified graphene, Pd-CNNF-G exhibits a slight improvement of current density for both
formic acid and methanol electrooxidation, but a great deal of enhancement of long-term
stability. This is due to the stronger interactions between the Pd NPs and the support,
which ensures the structure integrity maintenance during the harsh and long-term
electrochemical process.
To gain further insight into the oxidation of formic acid on the Pd-CNNF-G catalyst,
the kinetics of electron-transfer processes were studied by electrochemical impedance
spectroscopy (EIS). As shown in Fig. S18a, the variation trend of Nyquist impedance
2
0

polts at the Pd-CNNF-G electrode can be divided into three stages: (1) when potentials ≤
.3 V, the impedance spectra are located within the first quadrant; (2) as the potentials are
0
in the range of 0.4 V to 1.0 V, the negative impedance arcs can be observed in the second
quadrant; (3) at higher potentials (> 1.0 V), the impedance arcs return to the first
quadrant. According to the previously study in formic acid oxidation on the Pd-based
electrocatalysts [77-78], the negative faradaic impedance suggests the presence of an
inductive component. It can be considered that, as a co-catalyst, the g-C
3
N nanoflakelets
4
in the Pd-CNNF-G hybrids may promote the reaction because the oxophilic nature of the
N element can provide sites for the formation of chemisorbed hydroxyl species in this
potential range, which speeds up the oxidative removal of the adsorbed intermediate
poisoning species like COad [26, 27]. As compared to Pd-rGO (Fig. S18b), Pd-CNNF-G
possesses more negative onset potential and smaller impedance arcs, indicating
significantly facilitated for the removal of COad and faster charge-transfer kinetics for
formic acid oxidation, which are consistent with voltammetric and i-t response. To further
investigate reaction process, linear sweep voltammetry (LSV) was conducted on the
-1
catalyst. When the scan rates varied from 2 to 200 mV s , the peak current density (j )
p
1
/2
increased linearly with the square root of scan rate (v ) (Fig. S19a), indicating a
diffusion controlled behavior for electrooxidation of formic acid at the Pd-CNNF-G
electrode [79]. A good linear dependence of the oxidation peak potential (E ) versus
p
1
/2
log(v ) can be observed in Fig. S19b, which demonstrates that the oxidation of formic
acid was an irreversible electrode process. For a diffusion controlled and irreversible
electrode process on the Pd-CNNF-G catalyst, the value of the diffusion coefficient was
-7
2 -1
calculated to be 8.345 × 10 cm s [80].
2
1

Furthermore, the high-performance Pd-CNNF-G catalyst was used as anode material in
a real direct formic acid fuel cell. The corresponding steady-state polarization and power-
density curves of the studied catalysts are shown in Fig. 8a. Consistent with the results
from electrochemical measurements, the Pd-CNNF-G catalyst exhibits a higher power
-
2
density of 35 mW cm , which is about 1.5 times of that of Pd-rGO. It is also noted from
Fig. 8b that the discharge stability of the Pd-CNNF-G catalyst is largely improved
compared with the Pd-rGO catalyst at 0.35 V, which is consistent with those results
obtained from the CV and CA tests.
On the basis of the experimental results and theoretical analysis, the high catalytic activity and
durability of Pd-CNNF-G can be attributed to the specific characteristics of the unique
nanostructure of Pd-CNNF-G and the concerted effects of individual components. Firstly, the g-
C
3
4
N nanoflakelets with much more exposed edge sites on the surface of rGO, especially
2
+
pyridine-like N atoms, can be used as an excellent ligand for Pd ions, promoting the formation
of the highly dispersed Pd NPs. The bridging effect of g-C nanoflakelets facilitates the
3
N
4
intimate interaction among individual components of Pd-CNNF-G, which may ensure the
structure integrity maintenance, long-term electrochemical stability and durability. Secondly, the
N atoms at the exposed edge sites are also capable of activating their neighboring carbon atoms
and accelerating the formation of -OH by water dissociation, which can be a crucial factor for the
oxidative removal of the absorbed intermediate poisoning species on the Pd sites during catalytic
reactions [26]. Thirdly, the CNNF-G framework possesses abundant pore structures while
maintaining 2D nanostructure of rGO with acceptable conductivity and accessible internal
surface areas, providing a high rate of diffusion of the electrolyte and assisting the rapid removal
of the carbonaceous intermediates from the Pd sites in the oxidation process. Finally, the small-
2
2

sized and uniformly-dispersed Pd NPs on the CNNF-G offer more triple-phase boundaries at per
unit mass of Pd for the oxidation reaction of formic acid and methanol, improving the catalyst
utilization.
4
. Conclusions
In summary, we have demonstrated the successful fabrication of Pd NPs-decorated CNNF-G
nanocatalyst by a facile two-step approach: CNNF-G was obtained from splitting decomposition
of bulk g-C on rGO at higher temperatures, followed by the reaction of CNNF-G and Pd
3
N
4
nitrate to form Pd/CNNF-G nanocatalyst. The as-prepared Pd-CNNF-G exhibits an
unprecedented high and stable electrocatalytic activity for both formic acid and methanol
oxidation reactions, far outperforming the Pd-rGO, Pd-CNT or commercial Pd-AC catalysts. The
superior electrochemical performance of Pd-CNNF-G is attributed to the specific characteristics
of its unique nanostructure and the concerted effects of Pd NPs and CNNF-G, including
2
+
considerable edge sites and active nitrogen species for evenly dispersed decoration of Pd ions,
the strong adsorption energy of Pd adatoms on the CNNF-G resulting in a highly stable catalytic
system, and the abundant pore structures providing a high rate of diffusion of the electrolyte and
assisting the rapid removal of the carbonaceous intermediates from the Pd sites during the
reaction. Overall, such a stable CNNF-G-based nanocatalyst may bring new design opportunities
for high-performance fuel cells in the future.
Acknowledgements
This work was supported by NSF of China (No. 51572125), the Fundamental Research
Funds for the Central Universities (No. 30916014103) and the Zijin Intelligent program
of NUST (2014) and PAPD of Jiangsu.
2
3

References
[
1] D. Larcher, J. Tarascon, Towards greener and more sustainable batteries for electrical energy
storage. Nat. Chem., 7 (2015) 19-29.
[
2] F. Bonaccorso, L. Colombo, G. Yu, M. Stoller, V. Tozzini, A.C. Ferrari, R.S. Ruoff, V.
Pellegrini, Graphene, related two-dimensional crystals, and hybrid systems for energy
conversion and storage. Science, 347 (2015) 1246501.
[
3] B.C. Steele, A. Heinzel, Materials for fuel-cell technologies. Nature, 414 (2001) 345-352.
4] W.-Y. Yu, G.M. Mullen, D.W. Flaherty, C.B. Mullins, Selective Hydrogen Production from
Formic Acid Decomposition on Pd–Au Bimetallic Surfaces. J. Am. Chem. Soc., 136 (2014)
[
11070-11078.
[
5] M. Nielsen, E. Alberico, W. Baumann, H.-J. Drexler, H. Junge, S. Gladiali, M. Beller, Low-
temperature aqueous-phase methanol dehydrogenation to hydrogen and carbon dioxide.
Nature, 495 (2013) 85-89.
[
6] S. Uhm, H.J. Lee, Y. Kwon, J. Lee, A Stable and Cost-Effective Anode Catalyst Structure for
Formic Acid Fuel Cells. Angew. Chem. Int. Ed., 47 (2008) 10163-10166.
[
7] Z. Cui, M. Yang, F.J. DiSalvo, Mesoporous Ti0.5Cr0.5N Supported PdAg Nanoalloy as Highly
Active and Stable Catalysts for the Electro-oxidation of Formic Acid and Methanol. ACS
Nano, 8 (2014) 6106-6113.
[
8] M. Yang, R. Guarecuco, F.J. DiSalvo, Mesoporous chromium nitride as high performance
catalyst support for methanol electrooxidation. Chem. Mater., 25 (2013) 1783-1787.
9] H. Huang, S. Yang, R. Vajtai, X. Wang, P.M. Ajayan, Pt-Decorated 3D Architectures Built
from Graphene and Graphitic Carbon Nitride Nanosheets as Efficient Methanol Oxidation
[
2
4

Catalysts. Adv. Mater., 26 (2014) 5160-5165.
[
10] A.-L. Wang, H. Xu, J.-X. Feng, L.-X. Ding, Y.-X. Tong, G.-R. Li, Design of Pd/PANI/Pd
sandwich-structured nanotube array catalysts with special shape effects and synergistic
effects for ethanol electrooxidation. J. Am. Chem. Soc., 135 (2013) 10703-10709.
11] E. Antolini, Palladium in fuel cell catalysis. Energy Environ. Sci., 2 (2009) 915-931.
12] L.Y. Zhang, Z.L. Zhao, C.M. Li, Formic acid-reduced ultrasmall Pd nanocrystals on
graphene to provide superior electocatalytic activity and stability toward formic acid
oxidation. Nano Energy, 11 (2015) 71-77.
[
[
[
13] Y.X. Chen, A. Lavacchi, S.P. Chen, F. di Benedetto, M. Bevilacqua, C. Bianchini, P.
Fornasiero, M. Innocenti, M. Marelli, W. Oberhauser, Electrochemical milling and faceting:
Size reduction and catalytic activation of palladium nanoparticles. Angew. Chem. Int. Ed.,
51 (2012) 8500-8504.
[
14] L. Wang, Y. Nemoto, Y. Yamauchi, Direct synthesis of spatially-controlled Pt-on-Pd
bimetallic nanodendrites with superior electrocatalytic activity. J. Am. Chem. Soc., 133
(2011) 9674-9677.
[
15] V. Mazumder, S. Sun, Oleylamine-mediated synthesis of Pd nanoparticles for catalytic
formic acid oxidation. J. Am. Chem. Soc., 131 (2009) 4588-4589.
[
16] J. Huang, D. Wang, H. Hou, T. You, Electrospun Palladium Nanoparticle-Loaded Carbon
Nanofibers and Their Electrocatalytic Activities towards Hydrogen Peroxide and NADH.
Adv. Funct. Mater., 18 (2008) 441-448.
[
17] J. Zhang, Y. Cheng, S. Lu, L. Jia, P.K. Shen, S.P. Jiang, Significant promotion effect of
carbon nanotubes on the electrocatalytic activity of supported Pd NPs for ethanol oxidation
reaction of fuel cells: the role of inner tubes. Chem. Commun., 50 (2014) 13732-13734.
2
5

[
18] N.G. Sahoo, Y. Pan, L. Li, S.H. Chan, Graphene-Based Materials for Energy Conversion.
Adv. Mater., 24 (2012) 4203-4210.
[
19] Q.-S. Chen, Z.-N. Xu, S.-Y. Peng, Y.-M. Chen, D.-M. Lv, Z.-Q. Wang, J. Sun, G.-C. Guo,
One-step electrochemical synthesis of preferentially oriented (111) Pd nanocrystals
supported on graphene nanoplatelets for formic acid electrooxidation. J. Power Sources, 282
(2015) 471-478.
[
[
[
[
[
[
20] S. Zhang, Y. Shao, H. Liao, M. H. Engelhard, G. Yin, Y. Lin. Polyelectrolyte-induced
reduction of exfoliated graphite oxide: a facile route to synthesis of soluble graphene
nanosheets. ACS Nano, 5 (2011) 1785-1791.
21] L. Panchakarla, K. Subrahmanyam, S. Saha, A. Govindaraj, H. Krishnamurthy, U.
Waghmare, C. Rao, Synthesis, structure, and properties of boron-and nitrogen-doped
graphene. Adv. Mater., 21 (2009) 4726-4730.
22] J. Liang, Y. Jiao, M. Jaroniec, S. Z. Qiao, Sulfur and Nitrogen Dual-Doped Mesoporous
Graphene Electrocatalyst for Oxygen Reduction with Synergistically Enhanced
Performance. Angew. Chem. Int. Ed., 51 (2012) 11496-11500.
23] S. Wang, L. Zhang, Z. Xia, A. Roy, D.W. Chang, J.B. Baek, L. Dai, BCN Graphene as
Efficient Metal-Free Electrocatalyst for the Oxygen Reduction Reaction. Angew. Chem. Int.
Ed., 51 (2012) 4209-4212.
24] W. Zhang, H. Huang, F. Li, K. Deng, X. Wang, Palladium nanoparticles supported on
graphitic carbon nitride-modified reduced graphene oxide as highly efficient catalysts for
formic acid and methanol electrooxidation. J. Mater. Chem. A, 2 (2014) 19084-19094.
25] Y. Zheng, J. Liu, J. Liang, M. Jaroniec and S. Z. Qiao, Graphitic carbon nitride materials:
controllable synthesis and applications in fuel cells and photocatalysis. Energy Environ. Sci.,
2
6

5
(2012) 6717-6731.
[
26] Y. Zhou, K. Neyerlin, T.S. Olson, S. Pylypenko, J. Bult, H.N. Dinh, T. Gennett, Z. Shao, R.
O'Hayre, Enhancement of Pt and Pt-alloy fuel cell catalyst activity and durability via
nitrogen-modified carbon supports. Energy Environ. Sci., 3 (2010) 1437-1446.
[
27] G. Yang, Y. Li, R.K. Rana, J.-J. Zhu, Pt-Au/nitrogen-doped graphene nanocomposites for
enhanced electrochemical activities. J. Mater. Chem. A, 1 (2013) 1754-1762.
[28] Y. Zheng, Y. Jiao, J. Chen, J. Liu, J. Liang, A. Du, W. Zhang, Z. Zhu, S. C. Smith and M.
Jaroniec, Nanoporous graphitic-C
3
N @ carbon metal-free electrocatalysts for highly
4
efficient oxygen reduction. J. Am. Chem. Soc., 133(2011) 20116-20119.
[
29] J. Tian, R. Ning, Q. Liu, A. M. Asiri, A. O. Al-Youbi and X. Sun, Three-Dimensional
Porous Supramolecular Architecture from Ultrathin g-C
3
4
N Nanosheets and Reduced
Graphene Oxide: Solution Self-Assembly Construction and Application as a Highly
Efficient Metal-Free Electrocatalyst for Oxygen Reduction Reaction. ACS Appl.
Mater. Interfaces, 6 (2014) 1011-1017.
[
30] N. I. Kovtyukhova, P.J. Ollivier, B.R. Martin, T.E. Mallouk, S.A. Chizhik, E.V.
Buzaneva, A.D. Gorchinskiy, Layer-by-layer assembly of ultrathin composite films
from micron-sized graphite oxide sheets and polycations. Chem. Mater., 11(1999) 771-
7
78.
[
[
[
31] J.P. Perdew, K. Burke, M. Ernzerhof, Generalized gradient approximation made simple.
Phys. Rev. Lett., 77 (1996) 3865.
32] G. Kresse, J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations
using a plane-wave basis set. Phys. Rev. B, 54 (1996) 11169.
33] H. J. Monkhorst, J. D. Pack, Special points for Brillouin-zone integrations. Phys. Rev. B, 13
2
7

(
1976) 5188-5192.
34] Y. Fu, T. Huang, L. Zhang, et al. Ag/g-C
the degradation of dyes: a borohydride-generated superoxide radical approach[J]. Nanoscale,
(2015) 13723-13733.
[
[
[
3
N catalyst with superior catalytic performance for
4
7
35] M. Shalom, M. Guttentag, C. Fettkenhauer, S. Inal, D. Neher, A. Llobet and M. Antonietti,
In Situ Formation of Heterojunctions in Modified Graphitic Carbon Nitride: Synthesis and
Noble Metal Free Photocatalysis. Chem. Mater., 26(2014) 5812-5818.
36] K. S. Kim, Y.M. Um, J.-r. Jang, W.-S. Choe, P. J. Yoo, Highly sensitive reduced graphene
oxide impedance sensor harnessing π-stacking interaction mediated direct deposition of
protein probes. ACS Appl. Mater. Interfaces, 5(2013) 3591-3598.
[
[
[
[
37] M. S. Dresselhaus, G. Dresselhaus, R. Saito, A. Jorio, Raman spectroscopy of carbon
nanotubes. Phys. Rep., 2005, 409, 47-99.
38] D. Geng, Y. Chen, Y. Chen, Y. Li, R. Li, X. Sun, S. Ye, S. Knights, High oxygen-reduction
activity and durability of nitrogen-doped graphene. Energy Environ. Sci., 4 (2011) 760-764.
39] M. S. Dresselhaus, G. Dresselhaus, Intercalation compounds of graphite. Adv. Phys.,
51(2002) 1-186.
40] H.-K. Jeong, Y.P. Lee, R.J. Lahaye, M.-H. Park, K.H. An, I.J. Kim, C.-W. Yang, C.Y. Park,
R.S. Ruoff, Y.H. Lee, Evidence of graphitic AB stacking order of graphite oxides. J. Am.
Chem. Soc., 130(2008) 1362-1366.
[
41] B. Xiong, Y. Zhou, Y. Zhao, J. Wang, X. Chen, R. O’Hayre, Z. Shao, The use of nitrogen-
doped graphene supporting Pt nanoparticles as a catalyst for methanol electrocatalytic
oxidation. Carbon, 52 (2013) 181-192.
2
8

[
42] Y. Wang, J. Yao, H. Li, D. Su, M. Antonietti, Highly selective hydrogenation of phenol and
derivatives over a Pd@ carbon nitride catalyst in aqueous media. J. Am. Chem. Soc.,
133(2011) 2362-2365.
[
43] P. Kundu, C. Nethravathi, P.A. Deshpande, M. Rajamathi, G. Madras, N. Ravishankar,
Ultrafast microwave-assisted route to surfactant-free ultrafine Pt nanoparticles on graphene:
synergistic co-reduction mechanism and high catalytic activity. Chem. Mater., 23 (2011)
2772-2780.
[
44] Y. Cui, J. Zhang, G. Zhang, J. Huang, P. Liu, M. Antonietti, X. Wang, Synthesis of bulk and
nanoporous carbon nitride polymers from ammonium thiocyanate for photocatalytic
hydrogen evolution. J. Mater. Chem. 21(2011) 13032.
[
45] Y. Li, H. Zhang, P. Liu, D. Wang, Y. Li and H. Zhao, Cross-linked g-C
3
4
N /rGO
Nanocomposites with Tunable Band Structure and Enhanced Visible Light Photocatalytic
Activity. Small, 9(2013) 3336-3344.
[
[
[
46] H.S. Kang, S. Jeong, Nitrogen doping and chirality of carbon nanotubes. Phys. Rev. B, 70
(
2004) 233411.
47] C. Ewels, M. Glerup, Nitrogen doping in carbon nanotubes. J. Nanosci. Nanotechnol.,
(2005) 1345-1363.
5
48] E. Cruz-Silva, F. Lopez-Urias, E. Munoz-Sandoval, B.G. Sumpter, H. Terrones, J.-C.
Charlier, V. Meunier, M. Terrones, Electronic transport and mechanical properties of
phosphorus-and phosphorus− nitrogen-doped carbon nanotubes. ACS Nano, 3(2009) 1913-
1921.
[49] B. Hu, T. Wu, K. Ding, X. Zhou, T. Jiang, B. Han, Seeding growth of Pd/Au bimetallic
nanoparticles on highly cross-linked polymer microspheres with ionic liquid and solvent-
2
9