Hydrothermally Induced Oxygen Doping of Graphitic Carbon Nitride with a Highly Ordered Architecture and Enhanced Photocatalytic Activity

Article abstract of DOI:10.1002/cssc.201702278

As an amorphous or semicrystalline material, graphitic carbon nitride (g-C₃N₄) displays poor photocatalytic activity owing to rapid recombination of the photogenerated charge carriers, which is mainly caused by a high density of defects in the graphitic structure. In this work, a porous O-doped g-C₃N₄ (P-CNO) nanosheet with a highly ordered architecture is fabricated by introducing a novel hydrothermal treatment to the precursor before the final thermal condensation. The photocatalytic hydrogen evolution rate (HER) and HER per surface area of P-CNO are 13.9 and 1.7 times higher than that of bulk g-C₃N₄. The improved photocatalytic activity is ascribed to a synergistic effect of O doping, a porous sheet-like morphology, and increased crystallinity. This work also provides a new approach for the synthesis of other polymer-based photocatalysts with high crystallinity and excellent performance.

Full text of DOI:10.1002/cssc.201702278

Accepted Article
Title: Hydrothermally induced O-doping and porous structure of
graphitic carbon nitride with highly ordered architecture and
dramatically enhanced photocatalytic property
Authors: Chao Wang, Huiqing Fan, Xiaohu Ren, Jiangwei Ma, Jiawen
Fang, and Weijia Wang
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201702278
Link to VoR: http://dx.doi.org/10.1002/cssc.201702278
A Journal of
www.chemsuschem.org

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
Hydrothermally induced O-doping and porous structure of
graphitic carbon nitride with highly ordered architecture and
dramatically enhanced photocatalytic property
Chao Wang, Huiqing Fan*, Xiaohu Ren, Jiangwei Ma, Jiawen Fang, Weijia Wang
Abstract: As an amorphous or semi-crystalline material, graphitic
carbon nitride (g-C3N4) displays poor photocatalytic activity due to the
rapid recombination of photo generated charge carriers, which is
mainly caused by the high density of defects in the graphitic structure.
In this work, a porous O-doped g-C3N4 nanosheet (P-CNO) with highly
ordered architecture is fabricated by introducing a novel hydrothermal
treatment to the precursor before the final thermal condensation. The
photocatalytic hydrogen evolution rate (HER) and HER per surface
area of P-CNO are 13.9 times and 1.7 times higher than that of bulk
g-C3N4. The improved photocatalytic activity is ascribed to the
synergistic effect of O-doping, porous sheet-like morphology and
increased crystallinity. This work also provides a new approach for
synthesis of other polymer-based photocatalysts with high crystallinity
and excellent performances.
as it can greatly increase the specific surface area thus producing
abundant catalytic active sites [14-20]. Very recently, Li et al. prepared
carbon-rich g-C3N4 nanosheets via thermally treating bulk g-C3N4 in
air and N2 atmosphere [21]. The resulted g-C3N4 nanosheet exhibited
high specific surface area of 213.17 m2/g and could produce 79.2
μmol H2 in 2 hours under visible light irradiation (λ>400 nm), the
performance of which was 20 times higher than that of bulk g-C3N4.
Moreover, the carbon-rich g-C3N4 nanosheet showed great cycling
stability and displayed high hydrogen production capability even after
12 cycling runs. Zhou and co-workers reported a highly porous carbon
nitride with increased performance in removal of sulfamethazine
under visible light, which is a commonly used antibiotics in veterinary
industry, due to the large surface area arisen from the porous
structure [22]. Moreover, Rahman et al. designed a g-C3N4 nanosheet,
which not only possessed a large surface area, but also counteracted
the blue shift of optical absorption, a common issue that limited the
photon harvest of g-C3N4 nanosheets [23]. The as-synthesized g-C3N4
nanosheet displayed a narrowed band gap of 2.45 eV and achieved
an apparent quantum efficiency of 16% at 420 nm.
Introduction
Environmental and energetic issues have been drawing increasing
attention recent years due to the rapid development of industry. Using
semiconductor-based photocatalysts is believed to hold great
potential in producing clean and sustainable energy (H2) as well as
alleviating the serious environmental pollution [1]. Various metal oxide
photocatalysts, such as TiO2, ZnO and Bi2O3 have been developed to
achieve hydrogen evolution, degradation of organic pollutes, removal
of toxic gas and reduction of heavy metal ions in water [2-5]. However,
there are many drawbacks for these catalysts, including metal
containing, wide band gap, low absorption rate in visible light and
efficiency. Graphitic carbon nitride (g-C3N4) has attracted intensive
interests as a promising metal-free photocatalyst due to its moderate
band gap, proper band position, inexpensiveness, physical and
chemical stability [6,7]. However, bulk g-C3N4 still suffers from some
obstacles to achieve high photocatalytic activity, for instance, low
specific surface area, limited active sites, high charge carrier
recombination rate and low degree of crystallinity [8]. Several means
Researches show that doping with metal and non-metal atoms,
such as C, N, O, S, P, I, K, Na, Fe and Cu, also plays a vital role in
promoting the photocatalytic activity of g-C3N4 [24-33]. Usually, a dopant
energy level would be induced between the energy gap of g-C3N4 after
substitution of C and N with heteroatoms, which will broaden the light
responsive range and cause higher light efficiency as well as better
electrical conductivity. For instance, a calcination-refluxing method
was reported to synthesis Mn doped g-C3N4 for highly efficient
photocatalytic degradation of organic dyes and reduction of heavy
metal ions [34]. The doped g-C3N4 displayed extended light absorption
range and narrowed band gap, thus resulting in increased catalytic
properties under visible light.
Many doped g-C3N4 with porous structure or a sheet-like morphology
are also designed to integrate the preferred features mentioned above,
such as large surface area and good electrical conductivity [35-39]. She
et al. prepared porous ultrathin O-doped g-C3N4 nanosheets by
etching the thermally treated g-C3N4 with the mixture of nitric acid and
sulfuric acid [40]. The O-doped g-C3N4 nanosheets exhibited a high
surface area of 109.3 m2/g and improved separation efficiency of
photogenerated carriers, thus resulting in superior photocatalytic
performance in both hydrogen evolution and degradation of organic
dyes. All these modification methods mentioned above, however, are
based on thermally or chemically etching and distorting the intrinsic
structure of g-C3N4. Despite the large surface area or extended light
absorption capability achieved, this would unavoidably induce a large
amount of defects, which will serve as charge recombination sites and
greatly limit the photo efficiency. As an amorphous or semi-crystalline
semiconductor, g-C3N4 already has a poor crystallinity, let alone the
degree after these modifications. Previous research has proved that
more efficient charge separation can be obtained in polymers with
higher crystalline degree [41]. Zhou et al. recently synthesized a g-C3N4
have been developed to enhance the photocatalytic performance of
[10,11]
g-C3N4, such as elemental doping [9], design of co-catalysts
,
exfoliation of bulk g-C3N4 into 2D nanosheets [12], formation of
heterostructures with other materials [13], etc.
Formation of nanosheets and porous structure has been proved to
be a practical method to improve the photocatalytic property of g-C3N4,
State Key Laboratory of Solidification Processing, School of Materials
Science and Engineering, Northwestern Polytechnical University, Xi’an
710072, China
E-mail: hqfan@nwpu.edu.cn (*Corresponding author. Huiqing Fan)
Supporting information for this article is given via a link at the end of the
document.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
with modulated crystallinity by a post treatment, which is to heat the
mixture of g-C3N4, KCl and LiCl [42]. The g-C3N4 with high crystallinity
showed strong absorption capability and greatly enhanced
photocatalytic oxidation performance. However, the formation
mechanism of the high crystallinity was not clear. A good crystallinity
can encourage charge delocalization and promote the separation and
transfer kinetics of photo-generated charge carriers. If the low degree
of crystallinity of g-C3N4 can be overcome as well as keeping the
porous nanosheet structure and doping effects, the photocatalytic
performance of g-C3N4 would be promoted to a large extent. Herein,
we proposed a hydrothermally assisted two-step method to synthesis
porous O-doped carbon nitride nanosheets with highly ordered
architecture. The resulted sample showed a high crystallinity, a large
specific surface area, a porous structure as well as boosted
photocatalytic hydrogen production performance. The formation
mechanism of the enhanced degree of crystallinity and porous
structure was also studied. To the best of our knowledge, this is the
first carbon nitride nanosheet that simultaneously exhibits a porous
structure, O-doping and highly ordered architecture. It is anticipated
that this porous O-doped carbon nitride nanosheet photocatalyst can
be a promising candidate for practical solar energy conversion
applications.
Figure 1. (a) Schematic illustration for synthesis process of P-CNO (HT means
hydrothermal treatment), (b) XRD patterns of DCD and HTDCD and (c) FTIR
spectra of DCD, HTDCD and CN.
Results and Discussion
The porous O-doped carbon nitride (P-CNO) is synthesised by
heating the hydrothermally treated dicyandiamide (HTDCD) at 550 °C
for 2 h, as shown in Fig. 1a. Dicyandiamide (DCD) would aggregate
into new molecules when hydrothermally treated at 180 °C for 16 h.
The results are proved by XRD and FTIR spectra. The appearance of
new peaks in the XRD spectra of HTDCD is direct evidence for the
formation of new molecules. As displayed in the FTIR results in Fig.
1b, the broad peak at 2800-3600 cm-1 resulted from amino groups
becomes weaker and narrower for HT-DCD, indicating that some of
the amino groups were broken in the polymerization process after the
hydrothermal treatment of DCD. The FTIR peak of HTDCD located at
1882-2330 cm-1, which is ascribed to cyano groups, is also sharply
weakened and narrowed compared with DCD. It further indicates that
a large amount of cyano groups was destroyed and DCD polymerized
into new molecule in the hydrothermal process. The sharp peak at
808 cm-1 in CN associated to the tri-s-triazine units is not detected in
HT-DCD, demonstrating that the tri-s-triazine units were not formed
when DCD was hydrothermally treated at 180 °C. Interestingly, a
weak peak at 1082 cm-1 is emerged, which is resulted from hydroxide
radicals attached to sp2 C atoms [43]. It indicates that a little amount of
amino groups was oxidized into hydroxyls in the hydrothermal
treatment. Thus, oxygen could be induced in the network of graphitic
structure in the subsequent thermal condensation. The proposed
condensation route of P-CNO is shown in Fig. 2. Unlike CN, P-CNO
would not aggregate into macro bulks after the final thermal treatment,
but exists in the form of fluffy powder (see Fig. S1). Thus, the mass
loss (release of ammonia gas) in the thermal treatment process could
provide P-CNO with a porous structure. Therefore, we could easily
get a g-C3N4 simultaneously with O doping and a porous structure by
the introduction of a facile and green hydrothermal process.
Figure 2. Proposed condensation route of P-CNO.
The crystal structure of P-CNO and CN is characterized by powder X-
ray diffraction (XRD), as shown in Fig. 3a. Two dominated peaks are
detected in both CN and the P-CNO sample. It indicates that the
backbone graphitic structure is reserved in the P-CNO photocatalyst
and the pre-hydrothermal treatment process of DCD does not change
the crystal structure of the final thermal condensation products. The
(002) peak, which results from the interlayer of the stacking graphitic
structure, shifted from 27.5° to 27.7° for P-CNO. It indicates that P-
CNO displays a decreased interlayer distance about 0.002 nm
compared with CN. This could be originated from the strong
interaction between the graphitic layers of P-CNO. Interestingly, not
like other reported porous or doped g-C3N4 [4, 10, 12, 16, 44], the diffraction
peaks of P-CNO are much sharper and higher than the pristine bulk
CN. It indicates that the as-prepared P-CNO has more ordered
structure in both the in-plane heptazine framework and the stacking
of the conjugated aromatic layers. It could be originated from the ideal
arrangement of tri-s-triazine units with less dislocation in P-CNO.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
Thus, strong diffraction peaks are obtained for P-CNO even with less
stacking of graphitic layers and a porous structure. In addition, the full
width at half maximum (FWHM) of (002) diffraction peaks of CN and
P-CNO are calculated to be 1.84 and 1.10°, respectively, which further
confirms that P-CNO exhibits increased crystallinity. This suggests
that the pre-hydrothermal treatment can refine the graphitic structure
and increase the crystallinity of g-C3N4. As is known to all, defects in
the crystal structure of photocatalysts could serve as recombination
sites of the photoinduced carriers. The increased crystallinity, which
means fewer defects in the crystal structure of g-C3N4, would
suppress the rapid recombination of photoinduced charge carriers.
radicals connected to C in the C-N network. Furthermore, all
absorption peaks in the spectra of P-CNO are much sharper than that
of CN. Corresponding to the XRD analysis, it again verifies that the
tri-s-triazine units with more ordered in-plane structural packing motif
existed in P-CNO sample [46]. The room temperature electron
paramagnetic resonance (EPR) was further carried out to prove the
lower defects density in P-CNO. Due to the fact that orientation of
electron spins of the charge carrier is not conserved for a long time at
room temperature [19], EPR peak intensity for both CN and P-CNO is
rather weak. As shown in Fig. 3b, CN exhibits one single Lorentzian
line with a g value of 2.003, which is ascribed to lone pair electrons in
sp2 carbon in the aromatic rings of g-C3N4. However, nearly zero
signal is detected in P-CNO, indicating greatly quenched unpaired
electron density resulting from the low defects density [47]
.
Figure 4. XPS spectra of CN and P-CNO, (a) survey, (b) O 1s, (c) C 1s, (d) N
1s, (e) proposed O doping sites in P-CNO.
Figure 3. (a) XRD patterns, (b) Room temperature EPR spectra and (c) FTIR
spectra of CN and P-CNO.
The O-doping in P-CNO is further evidenced by the X-ray
photoelectron spectroscopy (XPS), as shown in Fig. 4. The survey
spectrum indicates the existence of C, N and O in CN and P-CNO,
except that the peak for O in P-CNO is higher than that in CN. The O
1s core level of CN at 532.2 eV is ascribed to the absorbed water.
Interestingly, two new peaks at 533.0 and 534.3 eV appear for P-CNO,
Fig. 3c shows the FTIR-spectra of CN and P-CNO. The IR
absorption bands of P-CNO are similar as that of CN. It indicates that
the original graphitic C-N network is largely reserved in the P-CNO
sample. The sharp peak at 808 cm-1 is assigned to the breathing mode
of the tri-s-triazine units. The peaks ranging from 950 to 1780 cm-1 are
attributed to the typical stretching modes of C-N heterocycles. A broad
absorption band observed in the range from 3500 to 2800 cm-1 is
originated from the stretching vibration of N-H bonds. Interestingly,
three additional peaks are observed for P-CNO, one located at 1280
which could be attributed to the hydroxide radical linked to carbon
[45]
atom and C-O-C in the heterocyclic ring of graphitic structure
.
According to the integral area of O 1s, the doping amount of O is
5.11%. The C 1s spectrum of CN could be deconvoluted into two
peaks located at 284.6 and 288.1 eV, which are the results of graphite
carbon and the sp2-hybridized carbon in the C=N-C aromatic ring,
respectively. Two additional peaks at 285.9 and 289.6 eV are found
in the C 1s of P-CNO, which are considered to be the C-O-C bond
and the hydroxide radicals connected to C atoms, respectively. The
peak indicating C=N-C aromatic ring of P-CNO displayed an obvious
shift of 0.7 eV compared with CN. The result could be ascribed to that
the substituted O in the graphitic structure of the C=N-C ring greatly
changed the chemical state of C element. The XPS analysis of O 1s
and C 1s further verifies that doping O atoms either take the place of
two-coordinated nitrogen atoms or exist as hydroxide radicals linked
to C atoms in the C=N-C ring, as shown in Fig. 4e. This is in good
accordance with the FTIR spectra. The N 1s spectrums of CN and P-
CNO are fitted into four peaks. The peaks located at 397.6, 398.9 and
cm-1 and two located at 1005 and 1075 cm-1, which are the results of
[43, 45]
C-O-C stretching and phenolic C-OH stretching, respectively
.
These additional peaks are direct evidence for O-doping in the crystal
structure of g-C3N4. DFT computation reported in another O-doped g-
C3N4 had proved that O atoms would preferentially substitute N atoms
in two-coordinated positions because of the extremely low formation
energy [20]. In addition, the electronegativity values of C, N and O are
2.5, 3.0 and 3.5, respectively. O is more intended to substitute the
atom whose electronegativity is closer to itself. Based on the reported
lectures and the above analysis, it suggests that oxygen has two
doping sites in the graphitic structure of C3N4, substitution of two-
coordinated N in the C-N heterocycles and existing as hydroxide
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
400.1 eV are the results of N bonded to carbon atoms in the tri-s-
triazine rings, the tertiary N and amino functional groups, respectively.
The peak at 402.6 eV is caused by the positive charge localization or
the charging effects in the cyano group and hetero cycles.
Interestingly, no shift of N 1s peaks is observed in P-CNO. However,
after calculating the integral area of each peak, the relative content of
various nitrogen species on the surface of P-CNO is found to be
completely different from that of CN (See Table S1). The decrease
amount of C=N-C could be explained that some two-coordinated
nitrogen atoms were replaced by O. Furthermore, the amount of N-H
bond increased from 9.8% to 32.4%, which indicates that a large
number of amino groups are detected in P-CNO. The primary and
secondary amino groups can easily form hydrogen bonds with water
molecules thus could enhance the dispersion capability of g-C3N4 in
water. As shown in the dispersibility test (see Fig. S3), the suspension
of P-CNO remained stable after 24 h, while precipitates could be
observed at the bottom of the sample cell only after 6 h for CN. It
indicates P-CNO can form homogeneous suspension with water thus
more photo-generated carriers could transport form the surface of
catalyst to water molecules to run photocatalytic reactions. The
abundant amino groups could also facilitate the strong interaction
between layers through hydrogen bonding interactions [48], thus
resulting in an ordered structure and decreased distance in the
stacking of conjugated aromatic layers, which is well agreed with the
XRD analysis. Elemental analysis (EA) is also carried out to study the
composition of the samples. The results are shown in Table S2. The
atomic ratio of N to C in P-CNO is 1.32, which is very closer to the
theoretical value of 1.33. It demonstrates that nearly ideal atomic
arrangement is achieved in P-CNO. However, the N/C atomic ratio
calculated in CN is much higher than 1.33, indicating that an
incompletely polymerized graphitic structure is formed in CN (See Fig.
S5).
increase the specific surface area and also provide abundant catalytic
active sites.
Figure 5. TEM images of (a) CN, (b) and (c) P-CNO, (d) SEM and (e) AFM
image of P-CNO, (f) height curve determined along the line between A-B and
C-D.
The porous structure and Brunauer-Emmett-Teller (BET) surface
area of the samples are studied using N2 adsorption-desorption
measurements at 77 K. Both CN and P-CNO both exhibit type IV
isotherms, which show a high adsorption capacity at high relative
pressure range (P/P0: 0.8-1), indicating the presence of large meso-
pores and macro-pores (also evidenced in TEM images). The
determined specific surface area of P-CNO is 48.96 m2/g, which is 8
times higher than that of CN. Pore distribution curve of P-CNO
showed two peaks at 3.8 nm representing small pores, and at 37 nm
representing large pores. The calculated total pore volume of CN and
P-CNO is 0.097 and 0.248 cm3/g. The large specific surface area and
pore volume enable P-CNO with a great number of mass and charge
transfusion channels and abundant catalytic active sites, which could
greatly facilitate the photocatalytic reaction.
The micro morphology and texture of CN and P-CNO are observed
by scanning electron microscopy (SEM) and transmission electron
microscopy (TEM). As observed in Fig. 5 and Fig. S4, CN is
composed by large irregular bulks with dense stacking layers and
smooth surfaces while P-CNO consists of nanosheets with a porous
structure. The enlarged TEM image of P-CNO shown in Fig. 5c clearly
demonstrates pores with diameter ranging from 10 to 120 nm. Further,
the edges of pores in P-CNO are very sharp and poignant (Fig. 5d
and S4d), indicating that the pores are formed in the polymerization
process caused by the shrinkage of the precursor, but not by etching
the completed g-C3N4 layers (In this case, the pores are often in the
shape of circles and the edges of pores are usually smooth [12, 49].).
Thus, few defects could be induced and a highly ordered architecture
is preserved in P-CNO. As shown in the insert of Fig. 5a and b, the
selected area electron diffraction (SAED) pattern of CN displays an
indistinct halation, while two clear diffraction rings indicating the (110)
and (002) crystal face are detected for P-CNO sample. This further
indicates that a higher crystallinity is obtained in the P-CNO sample,
corresponding to the previously discussed XRD, FTIR, XPS and EA
results. To measure the thickness of P-CNO nanosheet, atomic force
microscopy (AFM) was carried out. The results are shown in Fig. 5e
and f. The thickness of P-CNO is from about 3.1 to 4.1 nm, suggesting
that the sample is composed of about 9 to 14 atomic layers. The
ultrathin nanosheet morphology and porous structure could greatly
Figure 6. (a) N2 adsorption-desorption isotherms and (b) pore size distribution
curves of CN and P-CNO.
To understand the formation mechanism of porous structure and
ordered architecture in P-CNO, thermogravimetry (TG) is carried out
to study the condensation process of P-CNO. TG curves of DCD and
HTDCD are shown in Fig. 7. Not like that of DCD, the TG curve of
HTDCD displays two mass loss stages at 60 °C and 137 °C, which is
believed to be the release of structural water formed in the pre-
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
hydrothermal treatment of DCD. The fast pyrolysis of HTDCD could
be confined by the water molecules through hydrogen bonding. The
dehydration during polymerization process would induce a metastable
structure to form interaction between the incomplete dehydrated
HTDCD molecules. Therefore, the disordered and intense
condensation of HTDCD will be definitely hindered, which is clearly
depicted in the plot that HTDCD shows a delayed condensation
process, compared with DCD. Thus, the structural water in HTDCD is
the most essential reason for the ordered architecture of P-CNO.
Furthermore, DCD undergoes melting before the polymerization
the ultrathin porous structure, O-doping and increased
crystallinity of P-CNO are all favourable for the carrier
transportation, which will indirectly suppress the recombination
of photo-induced carriers. XPS valance band spectra indicates
the valance band potential of CN and P-CNO are 1.42 and 1.45
eV, respectively. Combining the results of Kubelka-Munk plots
and XPS valance band spectra of CN and P-CNO, the band
structure of CN and P-CNO is shown in Fig. 8d. The valance
band of P-CNO is more positive than that of CN, and its
conduction band is more negative, which is in good accordance
with the DRS analysis. This suggests that P-CNO possesses
stronger redox capabilities than CN, especially reduction ability,
which would greatly facilitate the photocatalytic hydrogen
evolution reaction.
reaction, which would cause disordered condensation and the tangled
[50]
CN chains
. However, HTDCD, after the pre-hydrothermal
treatment, will polymerize within a powder particle and does not melt
when condensate into P-CNO (See Fig. S1). This would also help
increasing the crystallinity of products. Plus, the abundant amino
groups in P-CNO could facilitate the strong interaction between the
graphitic layers through hydrogen bonding, which will result in an
ordered stacking at (002) crystal face. Interestingly, after full
condensation (T=550 °C), the remained products of DCD and HTDCD
are 53% and 30%, respectively. It indicates that compared with DCD,
HTDCD endures more mass loss and releases more NH3 in the
polymerization process, which will cause formation of more pores in
the structure of P-CNO. Therefore, an O-doped porous carbon nitride
with ordered architecture is obtained by the thermal condensation of
HTDCD.
Figure 8. (a) DRS spectra (insert is the corresponding Kubelka-Munk plots), (b)
PL spectra, (c) XPS valance band spectra and (d) schematic band structure of
CN and P-CNO.
Figure 7. TG curves of DCD and HTDCD.
UV-vis diffuse reflection spectrum (DRS) is used to study the
light absorption property of CN and P-CNO. As showed in Fig.
8a, the absorption edge of P-CNO displays a remarkable blue
shift compared with CN, which is resulted from the famous
Time-resolved fluorescence decay spectrum is used to further study
the charge carrier separation dynamics of CN and P-CNO. Compared
with CN, the fluorescence decay spectra of P-CNO is greatly
suppressed, indicating the photo-induced charge carriers in P-CNO
can be quickly captured by reactive substrates and are able to run
redox reactions. It also means more efficient separation of carriers,
which is ascribed to the porous sheet-like structure and highly ordered
crystalline structure and also is in correspondence with the PL spectra.
Bi-exponential decay function is used to fit the fluorescence decay
spectra and the results are shown in the insert of Fig. 9a. The average
carrier lifetime (τ) was calculated using Eq. (1):
quantum confinement effect caused by shifting the conduction
[11]
band and valence band edges along the opposite directions
.
However, the absorbance of P-CNO at wavelength longer than
460 nm is higher than that of CN. This could be attributed to the
[51, 52]
multiple reflection effect caused by the porous structure
.
The corresponding Kubelka-Munk plots (see insert of Fig. 8a)
demonstrate that the band gaps of CN and P-CNO are 2.71 and
2.85 eV, respectively. Fig. 8b shows the photoluminescence (PL)
emission spectra of CN and P-CNO. The suppressed emission
peak of P-CNO indicates that the recombination of photo-
induced carriers is largely restrained. There could be two
explanations for this effect. Firstly, the enlarged band gap, which
is further proved by the blue shift of emission peak, could reduce
the possibility of the recombination of charge carriers. Secondly,
n
2
∑
a τ
i
i=1
n
i
τ=
(1)
∑
a τ
i i
i=1
where τi is the lifetime and ai is the amplitude of the corresponding
component. The calculated fluorescence lifetimes for CN and P-CNO
are 5.43 and 3.46 ns, respectively. The quenching behavior of
emission intensity and carrier lifetime of P-CNO are solid evidence of
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
the drastically facilitated separation and transportation of charge
carriers.
visible light for 10 min, both CN and P-CNO showed increased signal
intensity. This process could be described using Eq. (2) [19]
,
[N-]+hν→[N]+e- (2)
It can be noticed that P-CNO displays much higher intensity than CN
under visible light, which indicates P-CNO can produce more
electrons, which would be used in the photocatalytic reaction to
reduce protons into hydrogen.
Figure 9. (a) Time-resolved fluorescence decay spectra (Insert shows the
fluorescence lifetimes and the percentages of photo-induced charge carriers)
and (b) calculated average lifetime of photo-induced carriers of CN and P-CNO.
Figure 11. (a) Hydrogen evolution curves, (b) Average and normalized
hydrogen evolution rate, (c) Cycling runs and (d) Wavelength dependence of
hydrogen evolution rate of CN and P-CNO.
Figure 10. (a) Transient photocurrent response and (b) EPR spectra (under
dark and visible light irradiation for 10 min) of CN and P-CNO.
The photocatalytic performance of the samples is evaluated by
hydrogen evolution under visible light (λ>400 nm) with 10 vol%
triethanolamine as sacrificial agent and 1% Pt as co-catalyst. As can
be seen in Fig. 11a, P-CNO displays remarkable enhancement in
photocatalytic hydrogen production compared with CN. The average
hydrogen evolution rate (HER) of P-CNO is 1748.6 μmol·h-1·g -1, which
is 13.9 times higher than that of CN. The normolized HER of P-CNO
is calculated to be 35.7 μmol·h-1·m -2, which is also far superior to that
of CN (21.1 μmol·h-1·m -2). It demonstrates that besides the increased
specific surface area cuased by the porous structure and ultralthin
sheet-like morphology, O-doping and the increased crystallinity,
which can inherently promote the charge seperation and migration in
the photoreaction process, are also the main reasons for the
enhanced photocatalytic property. The appearant quantum efficiency
(AQE) is calculated to understand the solar energy conversion
efficiency of P-CNO. The dependence of AQE on wavelength for P-
CNO is shown in Fig. S7. The AQE of P-CNO at 420 nm is calculated
to be 7.2%. A comparison of the photocatalytic performance in
hydrogen evolution of P-CNO with other recently reported carbon
nitrides is shown in Table S3. It is obvious that the photocatalytic
performance, especially the normolized HER is higher than that of the
most reported results. Cycling runs of photocatalytic tests suggests
that like CN, the hydrogen porduction performance of P-CNO also
shows no trend of deactivation even after 15 h. The photocatalytic
activity on wavelength tracks the characteristic absorption of P-CNO.
The photo-responsive property of CN and P-CNO is studied by the
transient photocurrent spectra and low-temperature EPR spectra. It
can be seen from Fig. 10a that the photocurrent density of P-CNO is
much higher than that of CN under visible light. It indicates that P-
CNO could generate more charge carriers under visible light
irradiation and exhibits higher electrical conductivity, which is proved
by the Mott-Schottky plots and electrochemical impedance spectra
(EIS) Nyquist plots (Fig S6). The charge carrier density in a
semiconductor can be determined by the reciprocal of slope in Mott-
Schottky plot [53]. The calculated charge density in P-CNO is 4.4 times
higher than that of CN. The enhanced electrical conductivity of P-CNO
is also evidenced by the smaller radius of Nyquist arc (Fig. S5 b),
suggesting stronger charge carrier transportation capability.
Interestingly, P-CNO also shows faster photo responsive rate than CN.
The current density of P-CNO can immediately reach the maximum at
the point when the light is on, and quickly return to baseline level when
the light is off. However, it takes about 10 seconds and 50 seconds
for the current density of CN to achieve maximum and return to the
level before irradiation, respectively. It is speculated that the fast
photo response of P-CNO is provided by its low density of lattice
defects, which in turn, means a high crystallinity. Low temperature (77
K) EPR is used to further prove the high photo response of P-CNO.
As depicted in Fig. 10b, CN and P-CNO both exhibited a single
Lorentzian line with quite weak intensity in dark. After irradiated under
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
It confirms that the hydrogen production is exclusively caused by the
photocatalytic process. In addition, an enhancement of photoactivity
is observed on the whole absorption spectrum for P-CNO. Based on
the above results, it demonstrates that P-CNO, with the cooperation
of porous structure, increased surface area, O-doping effect and
highly ordered architecture, displays a highly efficient photocatalytic
H2 evolution performance.
performance of P-CNO is resulted not only from the increased surface
area, but also the promoted transportation capability and efficient
separation of photo induced charge carriers caused by the O-doping
and highly ordered architecture. The present work demonstrates that
P-CNO catalyst holds great potential in application for photocatalysis
and other solar energy conversion systems.
Experimental Section
Catalysts Preparation
All reagents in the experiment were analytical grade and used
without further purification. In a typical synthesis, 1.5 g dicyandiamide
was added into 40 mL deionized water under magnetic stirring for 30
min to obtain a homogeneous solution. Then, the solution was sealed
in a 50 mL Teflon-lined autoclave and kept at 180 °C for 16 h. The
hydrothermally treated dicyandiamide was obtained by evaporating
water from the cooled solution at 100 °C. The porous O-doped carbon
nitride (P-CNO) was collected by heating the hydrothermally treated
dicyandiamide at 550 °C for 2 h with a ramping rate of 5 °C/min. For
comparison, Bulk carbon nitride (CN) was synthesized by directly
heating 4 g dicyandiamide at 550 °C for 4 h in air at a ramping rate of
5 °C/min.
Figure 12. Schematic illustrations for explanation of the enhanced
photocatalytic activity in P-CNO.
Characterization
The crystal structure of the samples was characterized by X-ray
diffraction (XRD; X'pert, Philips, Eindhoven, The Netherlands). The
chemical structure was measured with Fourier transform infrared
spectroscopy (FT-IR; TENSOR27, Bruker, Billerica, MA, USA). The
electron paramagnetic resonance (EPR; A300, Bruker, Billerica, MA,
USA) measurements were carried out on a Bruker A300 spectrometer.
The micromorphology was observed by field emission scanning
electron microscopy (FE-SEM; JSM-6701F, JEOL, Tokyo, Japan) and
high resolution transmission electron microscopy (HR-TEM, JEM-
3010, JEOL, Tokyo, Japan). The chemical states of the catalysts were
analysed by X-ray photoelectron spectroscopy (XPS, VG ESCALAB
220i-XL, Thermo Scientific, Waltham, MA, USA). The elemental
analysis is carried out on an elemental analyzer (Vario EL cub,
Elementar, Langenselbold, Germany). The Brunauer-Emmett-Teller
specific surface area was tested by nitrogen adsorption-desorption
(VSorb2800 P, Gold APP Corp., Beijing, China) at 77 K. The Diffuse
reflection spectra (DRS) of the as-synthesized carbon nitrides was
recorded by a spectrophotometer (UV3150; Shimadzu Corporation,
Kyoto, Japan). The room temperature fluorescence emission
spectrum and time-resolved fluorescence decay spectra were
recorded with a fluorescence spectrophotometer (FLS980, Edinburgh
Instruments, Edinburgh, Britain). Electrochemical impedance spectra
(EIS) were measured using an electrochemical workstation (CHI660E,
Chenhua, Shanghai, China). All electrochemical tests were carried
out using a three-electrode cell in which a Pt electrode and Ag/AgCl
electrode (3 M KCl) acted as the counter electrode and reference
electrode, respectively. The three electrodes were all immersed in 0.5
M Na2SO4 aqueous solution in the test.
Based on the above discussion, the pronounced enhancement in
photocatalytic water splitting of P-CNO can be possibly explained with
the following reasons as illustrated in Fig. 12. Firstly, the porous
sheet-like microstructure will greatly enhance the specific surface
area, thus resulting in abundant catalytically active sites, as well as
produce more channels for charge and mass transfer in the
photocatalytic reaction. Secondly, the O-doping in the graphitic
structure can increase the electronic conductivity of P-CNO, which in
turn, will promote the separation of charge carriers. Last but not the
least, the highly ordered architecture, which means less defects in the
lattice of P-CNO, will largely control the recombination and
dramatically enhance the efficiency of photo-induced charge carriers.
Therefore, for the first time that a highly ordered architecture as well
as porous structure and O-doping are achieved simultaneously in one
photocatalyst, P-CNO shows great potential in practically utilizing
semiconductor-based photocatalysts for energy and environmental
applications.
Conclusions
An O-doped porous g-C3N4 nanosheet with highly ordered
architecture was synthesized by a hydrothermally assisted two step
polymerization method. The O-doping effect and increased
crystallinity were caused by hydroxide radicals and structural water
introduced in the precursor during the hydrothermal treatment. A
porous structure was formed due to the melting-free thermal
condensation process. The photocatalytic hydrogen evolution rate of
P-CNO under visible light is 1748.6 μmol·h-1·g -1, which is 13.9 times
higher than that of CN. The greatly enhanced photocatalytic
Photocatalytic Test for Water Splitting
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
[13] X. She, J. Wu, H. Xu, J. Zhong, Y. Wang, Y. Song, K. Nie, Y. Liu, Y. Yang,
M. T. F. Rodrigues, R. Vajtai, J. Lou, D. Du, H. Li, P. M. Ajayan, Adv.
Energ. Mater. 2017, 7.
The photocatalytic activity of the samples was evaluated by
hydrogen production under visible light (λ>400 nm) irradiation.
50 mg catalyst was dispersed in 100 mL aqueous solution
containing 10 mL triethanolamine (TOEA). Pt (1 wt.%) was
photo-deposited on the catalysts using H2PtCl6·6H2O. Before
irradiation, the reaction system was evacuated 3 times to remove
air. The light source was a 300 W Xeon lamp (MAX-302, Asahi
Spectra, Torrance, CA, USA). A UV cut-off filter (λ>400 nm) was
applied to obtain visible light. The temperature of the reactant
system was maintained at 10 °C under external cooling
circulation. The evolved hydrogen was detected by a gas
chromatography device. To test the dependence of hydrogen
evolution rate and apparent quantum efficiency (AQE) on
wavelength, a serious of UV-vis filters (400 nm, 420 nm,450 nm,
500 nm, 550 nm) were applied. The light intensity is measured
by a irradiatometer (FZ-A, PEIFBNU, Beijing, China). AEQ is
calculated using Eq. (3):
[14] P. Niu, L. L. Zhang, G. Liu, H. M. Cheng, Adv. Funct. Mater. 2012, 22,
4763-4770.
[15] S. B. Yang, Y. J. Gong, J. S. Zhang, L. Zhan, L. L. Ma, Z. Y. Fang, R.
Vajtai, X. C. Wang, P. M. Ajayan, Adv. Mater. 2013, 25, 2452-2456.
[16] L. T. Ma, H. Q. Fan, J. Wang, Y. W. Zhao, H. L. Tian, G. Z. Dong, Appl.
Catal. B-Environ. 2016, 190, 93-102.
[17] Y. Li, R. Jin, Y. Xing, J. Li, S. Song, X. Liu, M. Li, R. Jin, Adv. Ener. Mater.
2016, 6.
[18] X. C. Wang, K. Maeda, X. F. Chen, K. Takanabe, K. Domen, Y. D. Hou,
X. Z. Fu, M. Antonietti, J. Am. Chem. Soc. 2009, 131, 1680-1681.
[19] H. J. Yan, Chem. Commun. 2012, 48, 3430-3432.
[20] A. Vinu, Adv. Funct. Mater. 2008, 18, 816-827.
[21] Y. F. Li, M. Yang, Y. Xing, X. C. Liu, Y. Yang, X. Wang, S. Y. Song, Small
2017, 13, 8.
[22] C. Zhou, C. Lai, D. Huang, G. Zeng, C. Zhang, M. Cheng, L. Hu, J. Wan,
W. Xiong, M. Wen, X. Wen, L. Qin, Appl. Catal. B-Environ. 2018, 220,
202-210.
2×(Number of hydrogen molecules)
[23] M. Rahman, K. Davey, S. Z. Qiao, Small 2017, 13.
[24] P. Zhang, X. Li, C. Shao, Y. Liu, J. Mater. Chem. A 2015, 3, 3281-3284.
[25] J. W. Fang, H. Q. Fan, M. M. Li, C. B. Long, J. Mater. Chem. A 2015, 3,
13819-13826.
AQE=
×100% (3)
Number of incident photons
[26] Z. F. Huang, J. Song, L. Pan, Z. Wang, X. Zhang, J. J. Zou, W. Mi, X.
Zhang, L. Wang, Nano Energy 2015, 12, 646-656.
Acknowledgements
[27] L. L. Feng, Y. Zou, C. Li, S. Gao, L. J. Zhou, Q. Sun, M. Fan, H. Wang,
D. Wang, G. D. Li, X. Zou, Int. J. Hydrogen Energ. 2014, 39, 15373-
15379.
This work is supported by the National Nature Science
Foundation (51672220), the 111 Program (B08040) of MOE, the
National Defense Science Foundation (32102060303), the Xi'an
[28] Y. P. Zhu, T. Z. Ren, Z. Y. Yuana, ACS Appl. Mater. Inter. 2015, 7,
16850-16856.
[29] Q. Han, C. Hu, F. Zhao, Z. Zhang, N. Chen, L. Qu, J. Mater. Chem. A
2015, 3, 4612-4619.
Science
and
Technology
Foundation
(CXY1706-5,
2017086CGRC049-XBGY005), the Shaanxi Provincial Science
Foundation (2017KW-018), and the NPU Gaofeng Project
(17GH020824) of China.
[30] S. Hu, F. Li, Z. Fan, F. Wang, Y. Zhao, Z. Lv, Dalton T. 2015, 44, 1084-
1092.
[31] J. Zhang, S. Hu, Y. Wang, Rsc Adv. 2014, 4, 62912-62919.
[32] Z. Li, C. Kong, G. Lu, J. Phys. Chem. C 2016, 120, 56-63.
[33] L. Jiang, X. Yuan, Y. Pan, J. Liang, G. Zeng, Z. Wu, H. Wang, Appl. Catal.
B-Environ. 2017, 217, 388-406.
Keywords: ordered architecture • doping • porous structure
• graphitic carbon nitride • photocatalysis
[34] J. C. Wang, C. X. Cui, Y. Li, L. Liu, Y.-P. Zhang, W. Shi, J. Hazard. Mater.
2017, 339, 43-53.
[35] G. Liu, P. Niu, C. Sun, S. C. Smith, Z. Chen, G. Q. Lu, H. M. Cheng, J.
Am. Chem. Soc. 2010, 132, 11642-11648.
[1]
[2]
A. Kudo, Y. Miseki, Chem. Soc. Rev. 2009, 38, 253-278.
P. Parnicka, P. Mazierski, T. Grzyb, Z. S. Wei, E. Kowalska, B. Ohtani,
W. Lisowski, T. Klimczuk, J. Nadolna, J. Catal. 2017, 353, 211-222.
X. N. Ren, Z. Y. Hu, J. Jin, L. Wu, C. Wang, J. Liu, F. Liu, M. Wu, Y. Li,
G. Van Tendeloo, B. L. Su, ACS Appl. Mater. Inter. 2017, 9, 29687-29698.
J. W. Fang, H. Q. Fan, G. Z. Dong, Mater. Lett. 2014, 120, 147-150.
L. Zhou, W. Z. Wang, H. L. Xu, S. M. Sun, M. Shang, Chem. Eur. J. 2009,
15, 1776-1782.
[36] Y. Deng, L. Tang, G. Zeng, Z. Zhu, M. Yan, Y. Zhou, J. Wang, Y. Liu, J.
Wang, Appl. Catal. B-Environ. 2017, 203, 343-354.
[37] N. Bao, X. Hu, Q. Zhang, X. Miao, X. Jie, S. Zhou, Appl. Surf. Sci. 2017,
403, 682-690.
[3]
[4]
[5]
[38] J. Hong, X. Xia, Y. Wang, R. Xu, J. Mater. Chem. 2012, 22, 15006-15012.
[39] S. Guo, Y. Zhu, Y. Yan, Y. Min, J. Fan, Q. Xu, Appl. Catal. B-Environ.
2016, 185, 315-321.
[6]
[7]
[8]
[9]
X. C. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson,
K. Domen, M. Antonietti, Nat. Mater. 2009, 8, 76-80.
J. W. Fang, H. Q. Fan, Z. Y. Zhu, L. B. Kong, L. T. Ma, J. Catal. 2016,
339, 93-101.
[40] X. She, L. Liu, H. Ji, Z. Mo, Y. Li, L. Huang, D. Du, H. Xu, H. Li, Appl. Catal.
B-Environ. 2016, 187, 144-153.
[41] N. Bansal, L. X. Reynolds, A. MacLachlan, T. Lutz, R. S. Ashraf, W.
Zhang, C. B. Nielsen, I. McCulloch, D. G. Rebois, T. Kirchartz, M. S. Hill,
K. C. Molloy, J. Nelson, S. A. Haque, Sci. Rep. 2013, 3.
[42] M. Zhou, P. Yang, R. Yuan, A. M. Asiri, M. Wakeel, X. Wang,
ChemSusChem 2017.
W. J. Ong, L. L. Tan, Y. H. Ng, S. T. Yong, S. P. Chai, Chem. Rev. 2016,
116, 7159-7329.
J. Ran, T. Y. Ma, G. Gao, X. W. Du, S. Z. Qiao, Energy Environ. Sci.
2015, 8, 3708-3717.
[43] Y. J. Zhou, L. X. Zhang, W. M. Huang, Q. L. Kong, X. Q. Fan, M. Wang,
J. L. Shi, Carbon 2016, 99, 111-117.
[10] J. Ran, G. Gao, F. T. Li, T. Y. Ma, A. Du, S. Z. Qiao, Nat. Commun. 2017,
8.
[44] H. Chen, J. H. Yao, P. X. Qiu, C. M. Xu, F. Jiang, X. Wang, Mater. Res.
Bull. 2017, 91, 42-48.
[11] J. Ran, B. Zhu, S. Z. Qiao, Angew. Chem. Int. Edit. 2017, 56, 10373-
10377.
[45] W. N. Xing, C. M. Li, G. Chen, Z. H. Han, Y. S. Zhou, Y. D. Hu, Q. Q.
Meng, Appl. Catal. B-Environ. 2017, 203, 65-71.
[12] L. Ma, H. Fan, M. Li, H. Tian, J. Fang, G. Dong, J. Mater. Chem. A 2015,
3, 22404-22412.
[46] X. Song, Q. Yang, X. Jiang, M. Yin, L. Zhou, Appl. Catal. B-Environ. 2017,
217, 322-330.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
[47] H. H. Ou, L. H. Lin, Y. Zheng, P. J. Yang, Y. X. Fang, X. C. Wang, Adv.
Mater. 2017, 29, 6.
[52] Q. Liang, Z. Li, Z. H. Huang, F. Kang, Q. H. Yang, Adv. Funct. Mater.
2015, 25, 6885-6892.
[48] M. Ayan-Varela, S. Villar-Rodil, J. I. Paredes, J. M. Munuera, A. Pagan,
A. A. Lozano-Perez, J. L. Cenis, A. Martinez-Aonso, J. M. D. Tascon,
ACS Appl. Mater. Inter. 2015, 7, 24032-24045.
[53] W. Che, W. Cheng, T. Yao, F. Tang, W. Liu, H. Su, Y. Huang, Q. Liu, J.
Liu, F. Hu, Z. Pan, Z. Sun, S. Wei, J. Am. Chem. Soc. 2017, 139, 3021-
3026.
[49] L. Shi, K. Chang, H. Zhang, X. Hai, L. Yang, T. Wang, J. Ye, Small 2016,
12, 4431-4439.
[54] L. H. Lin, H. H. Ou, Y. F. Zhang, X. C. Wang, ACS Catal. 2016, 6, 3921-
3931.
[50] G. Zhao, G. Liu, H. Pang, H. Liu, H. Zhang, K. Chang, X. Meng, X. Wang,
J. Ye, Small 2016, 12, 6160-6166.
[55] J. Li, B. Shen, Z. Hong, B. Lin, B. Gao, Y. Chen, Chem. Commun. 2012,
48, 12017-12019.
[51] Z. Wang, W. Guan, Y. Sun, F. Dong, Y. Zhou, W.-K. Ho, Nanoscale 2015,
7, 2471-2479.
[56] Z. Hong, B. Shen, Y. Chen, B. Lin, B. Gao, J. Mater. Chem. A 2013, 1,
11754-11761.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201702278
ChemSusChem
FULL PAPER
Table of Contents
FULL PAPER
A highly ordered architecture as well
as porous sheet-like structure and O-
doping are achieved simultaneously in
g-C3N4. The photocatalytic hydrogen
evolution rate (HER) and HER per
surface area are both greatly
enhanced.
Chao Wang, Huiqing Fan*, Xiaohu Ren,
Jiangwei Ma, Jiawen Fang, Weijia
Wang
Page No. – Page No.
Hydrothermally induced O-doping
and porous structure of graphitic
carbon nitride with highly ordered
architecture and dramatically
enhanced photocatalytic property
This article is protected by copyright. All rights reserved.