Enhancement of the Photocatalytic Activity of Carbon Nitrides by Complex Templating

Article abstract of DOI:10.1002/chem.201500579

Carbon nitrides with the increased surface areas (≈90 m² g^-1) were prepared by the heat-treatment of melamine-[Al(NO₃)₃] 9 H₂O complexes. Here, [Al(NO₃)₃] 9 H₂O was successfully used as a green template and a hydrogen-bonding partner of melamine, simultaneously providing a distinct porosity and defined particle morphology for the resulting carbon nitrides. The influence of the solvent and precursors ratio on the stoichiometry of the complexes, as well as the morphology, optical properties, and photocatalytic activity of resulting products for Rhodamine B degradation and sacrificial hydrogen evolution were investigated. We also studied the temperature-induced transformations of the starting complex to access the reaction mechanism. The developed synthesis procedure results in a significantly increased photocatalytic activity (up to 20-fold) compared with that of the ground reference material. Green template for C₃N₄ preparation: [Al(NO₃)₃] 9 H₂O was successfully used as a hydrogen-bonding partner of melamine, and subsequently as a green template , simultaneously providing a distinct porosity and defined particle morphology to the resulting carbon nitrides obtained through a heat treatment of the preformed complexes. The developed synthesis procedure resulted in 20-fold increase of the photocatalytic activity compared with that of the ground, graphitic carbon nitride reference (see figure).

Full text of DOI:10.1002/chem.201500579

DOI: 10.1002/chem.201500579
Full Paper
&
Photochemistry
Enhancement of the Photocatalytic Activity of Carbon Nitrides by
Complex Templating
[
a]
Zu Peng Chen, Markus Antonietti, and Dariya Dontsova*
Abstract: Carbon nitrides with the increased surface areas
well as the morphology, optical properties, and photocata-
lytic activity of resulting products for Rhodamine B degrada-
tion and sacrificial hydrogen evolution were investigated.
We also studied the temperature-induced transformations of
the starting complex to access the reaction mechanism. The
developed synthesis procedure results in a significantly in-
creased photocatalytic activity (up to 20-fold) compared
with that of the ground reference material.
2
À1
(
ꢀ90 m g ) were prepared by the heat-treatment of mela-
mine–[Al(NO ) ]·9H O complexes. Here, [Al(NO ) ]·9H O was
3
3
2
3 3
2
successfully used as a “green template” and a hydrogen-
bonding partner of melamine, simultaneously providing
a distinct porosity and defined particle morphology for the
resulting carbon nitrides. The influence of the solvent and
precursors ratio on the stoichiometry of the complexes, as
Introduction
volves hazardous etchants, HF or NH HF , which are fatal to
4
2
humans and the environment. Nanoporous carbon nitrides
prepared by the “soft-template approach”, which uses some
Visible-light responsive photocatalysis is emerging as a sustain-
able and convincing solution to counteract energetic and envi-
ronmental problems. The search for suitable semiconductor
photocatalysts for use in solar energy conversion and solar
light-driven degradation of organic pollutants has been the
[
5]
[6]
surfactants and self-organized biopolymers as the “green”
templates, do not finally have the structure of the idealized
C N (with C/N molar ratio of 0.75), but are enriched in carbon
3
4
resulting from the template decomposition. On the other
[1]
focus of research for the past 40 years. A weakly ordered
graphitic carbon nitride (simplified as g-C N ), which is known
hand, aluminium nitrate nonahydrate ([Al(NO ) ]·9H O), also
3
3
2
known as a strong oxidizing and nitrating agent, was recently
used by Zou’s group as a “green” template for the synthesis of
3
4
as the most stable allotrope in the carbon nitride family, has
stood out as a promising candidate among various photocata-
lysts. Recently, g-C N has received tremendous research inter-
[
7]
porous carbons. Meanwhile, it also exhibits the merits of low
decomposition temperature and easy removal of the remain-
ing Al species.
3
4
est because of its low cost, simple preparation procedure, and,
in particular, its outstanding catalytic performance in sacrificial
water splitting, heterogeneous catalysis, photodegradation of
[
8]
[9]
Recently, the groups of Thomas and Shalom proposed
a new and facile method that was based on supramolecular
preorganization of monomers to synthesize carbon nitrides
with increased surface areas and improved photoactivity for
hydrogen evolution reaction (HER) and RhB degradation. In
[2]
organic pollutants, oxidation reactions, and so on.
Bulk carbon nitride prepared by thermal condensation of ni-
trogen-rich organic monomers (e.g., melamine or dicyandia-
mide (DCDA)) shows only modest catalytic activity and low
this procedure, cyanuric acid and melamine first form a supra-
[
2d]
[10]
quantum efficiency because of its low surface area and poor
electronic properties. To improve its catalytic activity, various
modifications of the carbon nitride preparation procedure
have been developed, including the hard templating method,
heteroatom doping, copolymerization, and chemical functiona-
molecular complex with a defined morphology,
which is
then heated up to give carbon nitride. However, this specific
synthetic method is so far limited to complex formation be-
tween organic monomers (mostly substituted melamines and
cyanuric acid derivatives), and it is desirable to expand the
method to other classes of compounds.
[
2c,3]
lization.
Previously, mesoporous silica and silica nanoparti-
cles have been utilized as “hard templates” to synthesize mes-
Motivated by these findings, we successfully developed
a new route to nanoporous carbon nitrides starting from pre-
formed melamine–[Al(NO ) ]·9H O complexes. The products
[4]
oporous carbon nitride (mpg-C N ) featuring a high overall
3
4
surface area. Unfortunately, the removal of the templates in-
3
3
2
exhibit greatly enhanced photoactivity in the reactions of pol-
[
a] Z. P. Chen, Prof. Dr. M. Antonietti, Dr. D. Dontsova
Max-Planck Institute of Colloids and Interfaces
Department of Colloid Chemistry, Research Campus Golm
lutant degradation and H evolution under visible-light irradia-
2
tion. The synthetic strategy illustrated here may potentially
open a new avenue to develop more efficient porous carbon
nitrides for visible light-driven catalytic applications.
1
4476 Potsdam (Germany)
E-mail: dariya.dontsova@mpikg.mpg.de
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500579.
Chem. Eur. J. 2015, 21, 10805 – 10811
10805
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Results and Discussion
Complex composition
The complexes between melamine and [Al(NO ) ]·9H O were
3
3
2
obtained by the addition of melamine to a solution of
Al(NO ) ]·9H O in ethanol or water, by using different molar
[
3
3
2
ratios of the precursors (1:2, 2:3, 1:1, 2:1, 3:1, 6:1) and subse-
quent agitation. In ethanol, complexes appeared as milky sus-
pensions, whereas in water we observed gels or gel-like sus-
pensions. Within the complexes, amino groups of melamine
are likely bound to water molecules of [Al(NO ) ]·9H O through
3
3
2
an extended network of hydrogen bonds, thus at a ratio of 3:1
the formation of the stoichiometric complex could be expect-
ed. However, the real composition of complexes was found to
be different from the expected one, as analyzed by EDX and
reflected in Table 1. Complexes prepared in ethanol (2:1 and
Figure 1. XRD patterns and FTIR spectra of MAxy-complexes prepared in a,
c) water and b, d) ethanol and of precursors. Here, x and y indicates the
3
:1 ratios, Table 1, entries 9 and 10) always contained some un-
planned excess of melamine (as evidenced by the increased C
content). When an excess of melamine was intentionally taken
molar ratio between melamine and [Al(NO
complex.
3 3 2
) ]·9H O used to prepare the
(
6:1 ratio, Table 1, entry 11), the content of water molecules in
the complex was lowered to close to 6. Complexes prepared in
water (Table 1, entries 2–6) were also characterized by an
excess of melamine and contained additional water molecules
improved by heating up the gels to 60 or 808C (as demon-
strated by entries 7 and 8 compared with entry 6 (Table 1)),
which is essential for photocatalytic activity of resulting carbon
nitrides as will be shown below.
(
more than 9), as deduced from low Al/O ratios. This off-stoi-
chiometry is explained by rather quick formation of the ex-
tended supramolecular H-bonded network that results in in-
creased viscosity of the reaction mixture, thus, leading to hin-
drance of diffusion of the reactant molecules, drop of complex
formation rate, and entrapment of water molecules. Both mel-
amine and water content in water-derived complexes can be
The crystalline structures of the preformed melamine–[Al(-
NO ) ]·xH O complexes in water and ethanol are reflected in
3
3
2
powder X-ray diffraction (XRD) patterns (Figure 1a and b, re-
spectively). In all the cases, the presence of the three diffrac-
tion peaks at 2V=12.3, 17.6, and 19.48 confirms the successful
formation of the new aggregate,
which is long-range ordered. Be-
[
a]
sides, there is a number of char-
acteristic peaks in the region of
Table 1. Elemental composition of complexes prepared under different conditions.
Entry Complex
C [wt.%] N [wt.%] Al [wt.%] O [wt.%] C/N ratio Al/O ratio
2
V=26–318, probably due to
1
Theoretical stoichiometry for
/[Al(NO ]·9H O M =753.5
14.35
39.04
3.58
38.22
0.37
0.094
monoclinic crystal structure of
3
C
3
N
6
H
6
3
)
3
2
w
[
Al(NO ) ]·9H O containing octa-
3
3
2
2
3
4
5
6
7
8
9
1
MA12-H
MA23-H
2
O
O
18.47
19.93
19.14
19.57
20.27
20.36
18.06
19.62
21.60
21.99
16.82
20.84
18.28
51.91
53.12
52.77
54.17
54.47
51.37
49.53
48.93
48.34
49.72
43.81
47.31
49.74
1.02
1.02
0.72
0.42
0.94
1.11
2.56
2.55
2.34
2.76
3.36
5.56
4.56
25.59
25.94
27.36
25.84
24.32
27.16
29.84
28.90
27.72
25.55
36.02
26.29
24.35
0.36
0.38
0.36
0.36
0.37
0.40
0.36
0.40
0.45
0.44
0.38
0.44
0.37
0.040
0.039
0.026
0.016
0.039
0.041
0.086
0.088
0.084
0.108
0.093
0.211
0.187
3
+
[11]
hedral
[Al(H O) ]
units.
2
6
2
MA11-H
MA21-H
MA31-H
MA31-H
MA31-H
2
O
Among those, the peak at 2V=
27.58 is the most intense one
and probably characterizes the
lamellar stacking motif in the
complex. Although the stoichi-
ometry and composition of com-
plexes prepared under different
conditions varied significantly
(Table 1), all PXRD patterns look
quite similar, and the presence
of crystalline melamine is evi-
dent only in the case in which
the excess of melamine was
high from the beginning (at 6:1
ratio in ethanol). This suggests
that in the remainder of cases
melamine was probably surface-
bound onto the formed particles
2
O
O
2
2
2
O-608C
O-808C
MA21-EtOH
MA31-EtOH
MA61-EtOH
MA21-EtOH-top
0
1
1
1
1
1
2
3
4
MA21-EtOH-top-2708C
Theoretical stoichiometry for
/[Al(NO ] M =591.4
MA61-EtOH-top
MA61-EtOH-top-2708C
3
C
3
N
6
H
6
3
)
3
w
1
1
1
1
1
2
2
5
6
7
8
9
0
1
22.04
19.52
47.15
38.19
3.69
10.02
7.19
27.13
32.47
76.77
74.74
71.90
67.60
65.04
0.43
0.51
0.136
0.309
0.094
0.112
0.140
0.187
0.337
[Al(NO
[Al(NO
[Al(NO
[Al(NO
3
)
3
)
3
)
3
)
3
3
3
3
]·9H
]·6H
]·3H
]
2
2
2
O
O
O
8.40
10.10
12.67
21.94
[Al(OH)
2
(NO
3
)]
[
[
a] The complexes are abbreviated as MAxy-S, in which x and y represent the ratio between melamine and
Al(NO ]·9H O, whereas S indicates the solvent used.
3
)
3
2
Chem. Eur. J. 2015, 21, 10805 – 10811
www.chemeurj.org
10806
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
of the 3:1 stoichiometric complex, whose structure is reflected
in XRD patterns.
Further evidence for the complex formation was obtained
from the results of FTIR-ATR investigations (Figure 1c and d).
The FTIR spectra of the complexes prepared at different condi-
tions look very similar to each other and reveal several distinct
features compared with those of the starting compounds.
There are two main absorption bands that have similar intensi-
ties: 1) the band of the bending vibrations of NÀH bonds,
Figure 3. PXRD patterns of starting complexes and the intermediates ob-
tained by heating these starting complexes to 2708C: a) both 3:1 and 6:1
patterns; b) enlarged 3:1 patterns.
À1
À1
which is shifted to 1685 cm (1640 cm in melamine), con-
firming the formation of hydrogen bonds between the nona-
[8]
hydrate and the NH groups of melamine; 2) the band at
2
À1
À
1
368 cm assigned to the stretching vibrations in [NO ] ,
3
which confirms the incorporation of [Al(NO ) ]·9H O in the
highest photocatalytic activity (further referred to as 6:1). Both
consisted of the smallest complex particles and were collected
from the top of the corresponding suspensions.
3
3
2
[
12]
complex structure. The disappearance of the second band of
the NH stretch confirms the transformation of the free primary
amino-groups of melamine to the bound ones of the complex.
The morphology of the prepared complexes is preferentially
rod-like, though an excess of [Al(NO ) ]·9H O in ethanol results
Upon heating to 2708C, the Al/O ratio in the 3:1 complex
changes from 0.093 to 0.211, which is attributed to the loss of
all crystal water molecules. The composition of the complex
becomes close to the expected one for 3C N H /[Al(NO ) ]
3
3
2
in a granular, open-structure morphology, as illustrated by rep-
resentative scanning electron microscopy (SEM) images
3
6
6
3 3
(Table 1, entries 13 and 14). The loss of nine water molecules
would imply a weight loss of ꢀ21.5%, which is consistent with
the weight loss observed by TGA after first two decomposition
steps at 115.5 and 194.58C (the Supporting Information,
Figure S2B). If compared with the stability of pure
(
Figure 2 and the Supporting Information, Figure S1).
[
Al(NO ) ]·9H O, which first undergoes partial dehydration, and
3 3 2
then simultaneous dehydration and denitrification before T
[
13b,c]
reaches 2008C
(the Supporting Information, Figure S2A),
the stability of [Al(NO ) ]·9H O confined in the supramolecular
3
3
2
complex appears to be greatly enhanced. Upon heating to
708C, a 6:1 sample, having an initial Al/O ratio of 0.136 (corre-
sponding to [Al(NO ) ]·xH O with x=3–4), yields the intermedi-
2
3
3
2
ate with the Al/O ratio of 0.309 (Table 1, entries 15 and 16).
This ratio is close to the Al/O ratio in [Al(OH) (NO )] (0.337) and
2
3
implies that here both dehydration and partial denitrification
of the starting [Al(NO ) ]·xH O have occurred. We can conclude
that decomposition mechanism depends on the composition
of the preformed complex.
3
3
2
Figure 2. Representative SEM images of MA complexes prepared by using
different solvents and reagent ratios.
Partial decomposition of [Al(NO ) ]·xH O upon heating, lucki-
3
3
2
ly, does not lead to the loss of crystallinity as follows from
comparison of the corresponding XRD patterns (Figure 3a and
b). On the contrary, the crystalline structure becomes more de-
fined upon heating as observed by the intensification of some
originally present peaks, especially those at 2V=27.5 and 318,
suggesting the improvement of layer stacking and growth of
crystallites in the c-direction. Interestingly, this tendency is
much more pronounced in the 6:1 complex than in the 3:1
one, which might be responsible for the higher photocatalytic
activity of the 6:1 product (the Supporting Information,
Table S1, entries 17 and 19). The morphological changes of the
3:1 and 6:1 complexes upon heating to 2708C (the Supporting
Information, Figure S3) imply densification of the structure
through a certain sintering of particles though maintaining
their original shape that partially supports a topotactic transi-
tion mechanism.
Discussion of the condensation mechanism
To obtain nanoporous carbon nitrides, the preformed com-
plexes were heated at 5508C under a nitrogen flow. During
the heat-treatment [Al(NO ) ]·xH O undergoes dehydration and
3
3
2
[
13]
denitrification, and finally turns into an amorphous alumini-
um hydroxide-oxide species, which can be easily removed
after the synthesis by an acidic wash.
We attempted to understand the mechanism of complex
transformation upon heat-treatment by using TGA analysis
(
Figure S2, the Supporting Information). In addition, chemical
and structural characterization of the intermediates was ach-
ieved by heating the selected starting complexes to 2708C,
just before the onset of melamine condensation temperature
(
Figure 3 and Table 1 (entries 12, 13, 15, and 16), and the Sup-
porting Information, Figure S3). The complexes selected for
these investigations were the ones with 3:1 stoichiometry (fur-
ther referred to as 3:1) and the ones that finally showed the
Chem. Eur. J. 2015, 21, 10805 – 10811
www.chemeurj.org
10807
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Products
the tertiary N atoms in the center of the tri-s-triazine ring NC3
and to partially hydrogenated N atoms of ÀNH (x=1, 2)
x
During further heating of the 3:1 complex to 5508C, denitrifi-
groups arising from incomplete condensation, respectively.
These results are in full agreement with the literature data re-
cation of [Al(NO ) ] and condensation of melamine to polymer-
3
3
[
2d]
ic carbon nitride occur. The resulting C N products are denot-
ported for the graphitic carbon nitride
and thus indicate
3
4
ed in the following as MAxy-S-C N , in which x and y represent
that the chemical structure of the synthesized products is built
of tri-s-triazine units. The O1s signal shown in the Supporting
Information (Figure S6D) reveals two contributions assigned to
the surface-absorbed water (533.1 eV) and surface hydroxyl
groups (531.7 eV) originating from rather oxidative synthesis
path as well as aqueous work up of the products. As it was
mentioned above, there is no residual aluminium in the prod-
ucts (see Al 2p binding energy region in the Supporting Infor-
mation, Figure S6E), demonstrating that the template was suc-
cessfully removed after the synthesis.
3
4
the taken ratio between melamine and [Al(NO ) ]·9H O, and S
3
3
2
indicates the solvent. PXRD patterns and FTIR spectra of the
final products are similar to those of the reference bulk carbon
nitride prepared from melamine, indicating the successful for-
mation of polymeric carbon nitride network. In the case of
bulk g-C N , typical reflections at 27.3 (002) and 12.88 (100) in
3
4
the PXRD pattern correspond to graphite-like interlayer stack-
[
2d]
ing and in-plane structural repeating motifs, respectively
the Supporting Information, Figure S4A and B). In MAxy-C N4
(
3
products, both peaks are much broader and weaker than in
the reference material, which suggests a decreased structural
order, smaller crystallite size, and more distorted graphite-like
All products are characterized by a particle-like morphology
(the Supporting Information, Figure S7), particles having rather
small grain size, random shapes and size distribution, and
being perforated as the result of gas evolution during the de-
composition of [Al(NO ) ]·9H O, condensation of melamine,
[
14]
packing. The weakening of layer stacking motif in products,
which was well-pronounced in intermediates (Figure 3), is
probably caused by rapid and violent denitrification of
3
3
2
and etching of the remaining aluminium species that provides
macroporosity to most of prepared materials. The study of the
nitrogen sorption properties of MAxy-C N (N adsorption–de-
[
Al(NO ) ]. The FTIR spectra of MAxy-C N , displayed in the Sup-
3 3 3 4
porting Information (Figure S4C and D) are identical to those
3
4
2
À1
of g-C N . The vibrations between 1200 and 1650 cm are as-
sorption isotherms are shown in the Supporting Information,
Figure S8) shows that all the products possess higher specific
surface areas than the reference, ground g-C N (13–90 vs.
3
4
sociated with the stretching vibrations of aromatic CN hetero-
À1
cycles, and the strong band at 808 cm is attributed to the
3
4
2
À1
deformation vibrations of the tri-s-triazine ring. We did not ob-
serve significant differences between water and ethanol-de-
rived products, except for a slightly better structural order in
water-derived carbon nitrides, which is due to the lower initial
content of the template, leading to less distortion caused by
template decomposition.
9.5 m g , see the Supporting Information, Table S2). This is ac-
companied by the development of some micro-, meso- and
macroporosity and a 26-fold increase of the total pore volume,
which is expected to be beneficial for the photocatalytic activi-
ty of the prepared materials. Typically, ethanol-derived prod-
ucts have slightly higher surface areas than water-derived ones
(see the Supporting Information, Table S2, entries 11, 12, and
14 vs. entries 3 and 4), which are due to the higher initial tem-
plate content. These values can be further improved by either
increasing template content in the starting complex (the Sup-
porting Information, Table S2, entries 8 and 9) or directly heat-
ing the obtained gel to 5508C (Table S2, entry 10 and Fig-
Elemental analysis and EDX spectroscopy were employed to
determine the composition of the resulting products. Both
methods showed C/N molar ratio similar to the one of bulk
C N , indicating the formation of polymeric carbon nitride
3
4
structure from the complexes (the Supporting Information,
Table S1 and Figure S5). The absence of the aluminium species
in nanoporous MAxy-C N was confirmed by a missing Al 2p
2
À1
ure S9). Expectedly, the highest surface areas (50–90 m g ,
the Supporting Information, Table S2, entries 17–19) are ob-
served for the smallest complex particles collected from top of
suspensions. The porous character of the products is demon-
strated in Figure 4 by representative TEM images.
3
4
peak in the high-resolution X-ray photoelectron spectrum (the
Supporting Information, Figure S6E).
X-ray photoelectron spectroscopy (XPS) provided more infor-
mation about the chemical composition and the valence states
in the products (the Supporting Information, Figure S6). Fig-
Compared with surface areas (s.a.) of carbon nitrides pre-
pared by using SiO2 templates (e.g., mpg-C N , s.a.
ure S6A, showing the survey spectrum of the MA23-C N4
3
3
4
2
À1
sample, indicates that C, N, and some O atoms are present in
the material. The C1s spectrum (the Supporting Information,
ꢀ200 m g ), the surface areas obtained by using
Figure S6B) reveals two main contributions at 288.2 and
2
2
84.8 eV, which correspond to sp -carbon atoms in tri-s-triazine
rings (CN ) and the adventitious carbon, respectively. The small
3
contribution at 286.2 eV is assigned to the hydroxylated sur-
face carbon atoms as corroborated by O 1s signal at 531.7 eV
(
see below). The N 1s signal (the Supporting Information, Fig-
ure S6C) is deconvoluted into three peaks having binding en-
ergies of 398.6, 399.7, and 400.8 eV. The strongest peak at
2
3
98.6 eV is assigned to sp -nitrogen atoms in tri-s-triazine ring
3 4
Figure 4. TEM images of a) MA21-EtOH-top-C N , and of a directly heated
(
CÀN=C), whereas the peaks at 399.7 and 400.8 originate from
gel b) MA31-H O-C
2
3 4
N .
Chem. Eur. J. 2015, 21, 10805 – 10811
www.chemeurj.org
10808
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
2
À1
[
Al(NO ) ]·9H O (<90 m g ) are lower. On one hand, this is
3 3 2
because the amount of “green” template which is possible to
introduce into the initial complex is limited by complex stoichi-
ometry. On the other hand, Al has much higher affinity to O
than to N (in case for Si, the binding preference is not so dra-
matic), so it may be partially demixed from the structure
during the synthesis. Still, [Al(NO ) ]·9H O provides the advan-
3
3
2
tages of low-cost and easy template removal after the synthe-
sis (using 1m HCl) compared with SiO NPs (using 4m NH HF ).
2
4
2
The influence of the ratio of the precursors on the intrinsic
semiconductor properties of the resulting nanoporous carbon
nitrides, such as their optical absorption and emission as well
as the electronic structure, was studied by using UV/Visible ab-
sorption, steady-state photoluminescence (PL), and electron
paramagnetic resonance (EPR) spectroscopies. The appearance
of the optical absorption spectra of the products, shown in
Figure 5a and b, is typical for semiconductor materials and
originates from the electron transitions from the valence band
[2a,15]
(
VB) to the conduction band (CB).
the reference bulk C N , the absorption edge of MAxy-C N
4
Compared with that of
3
4
3
products is blueshifted. This obvious hypsochromic shift is pre-
sumably due to the decreased conjugation length in the mate-
[9,16]
rials prepared from MA complexes
or might be due to
a slightly altered stacking of C N layers as in an H-aggregate.
3
4
The optical band gaps of MAxy-EtOH-C N products were esti-
3
4
mated from the corresponding Tauc plots as shown in the
insets of the Figure 5a and b, and lie between 2.85 and
2
.93 eV, whereas the value for the reference C N is 2.74 eV.
3 4
Similar structure-induced changes of the electronic properties
are observed in the PL spectra of products (Figure 5c and d).
In agreement with the optical absorption spectra, we observed
a slight blueshift of the fluorescence emission peak. The inten-
sities of the emission signals of all MAxy-C N are lower than
3
4
the one of the reference C N , which is probably because of an
3
4
increased degree of the disorder and improved separation of
photo-generated charge-carriers in MAxy-C N . The extent of
3
4
the fluorescence quenching is typically higher in water-derived
products, but does not seem to depend on the ratio of mela-
mine to [Al(NO ) ]·9H O; rather it is because of different mor-
3
3
2
phologies, surface states, band gap narrowing effects, and de-
[17]
fects within the product structures.
The extended covalent carbon nitride network of the prod-
ucts was further characterized by using room-temperature EPR
spectroscopy. In the corresponding spectrum (the Supporting
Information, Figure S10), only one single Lorentzian line, which
yields a g value of 2.0034, is observed for both MA23-EtOH-
C N and the reference g-C N because of the delocalized elec-
Figure 5. UV/Visible absorption spectra (a, b) and room temperature PL
spectra (lex =350 nm) (c, d) of carbon nitrides derived from water (a, c) and
ethanol (b, d). Insets of (a, b) show the corresponding Tauc plots. Naming
order of products matches the order of plotted curves indicated by dashed
arrows.
3
4
3
4
[18]
trons in the CB of the materials. The EPR signal intensity ob-
served for MA23-C N is slightly increased compared with that
3
4
of the reference bulk-C N indicating an improved charge-sep-
3
4
aration in the former semiconductor.
spite the decreased structural order when compared with that
of the reference g-C N , the activities of all complex-derived
Photocatalytic activity
3
4
The photocatalytic performance of MAxy-C N products was
carbon nitrides (MAxy-C N ) in this reaction are much higher
3 4
3
4
first evaluated in the reaction of Rhodamine B (RhB) degrada-
tion under illumination with blue (465 nm) or white light. De-
than the activity of bulk-C N (see Figure 6 and the Supporting
3 4
Information, Table S2).
10809 ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2015, 21, 10805 – 10811
www.chemeurj.org

Full Paper
can be explained by the large number of suitable surface
active sites of the photocatalysts, which are created during
synthesis and template removal.
À3
À1
However, the highest rate constant, k=81.7”10 min
(
which is ꢀ22 times higher than that of ground g-C N , k=
3
4
À3
À1
3
.68”10 min , and ꢀ3.8 times higher than that of carbon
nitride prepared from melamine–cyanuric acid complex, k=
À3
À1
2
1.7”10 min ), was measured for MA61-EtOH-top-C N
3 4
2
À1
product with a surface area of only 52 m g , prepared from
the precursor containing both melamine and melamine-[Al(-
NO ) ]·xH O complex. A tentative explanation is that there is
3
3
2
a heterojunction between two carbon nitride phases, one of
which originates from melamine and other one from the mela-
mine–[Al(NO ) ]·xH O complex, in-situ created during the syn-
3
3
2
thesis, and a gradient structural order of the intermediates
Figure 3a) and products in this case, which improves separa-
(
tion of charge carriers.
The photocatalytic behavior of selected MAxy-C N products
3
4
was further studied in water reduction reaction upon visible-
light irradiation. A 50 W white LED array was used as a light
source, triethanolamine (TEOA) as a sacrificial hole scavenger
and in-situ deposited Pt as a co-catalyst, and the results are
shown in Figure S13 (the Supporting Information). Also, in case
Figure 6. Dependences of the relative RhB concentration on irradiation time
by using a) MAxy-H O-C , b) MAxy-EtOH-C , and ground reference g-
as photocatalysts. Dashed arrows indicate the naming order of prod-
2
À1
2
3
N
4
3
N
4
of HER, MA61-EtOH-bulk-C N product (s.a. ꢀ20 m g ) expos-
3
4
C
3
N
4
À1
es the highest activity by producing 2.62 mmol(H )h , illustrat-
2
ucts.
ing the benefits of this composed material for charge separa-
tion. No catalyst deactivation was observed during 12 h of re-
action time. The activity of g-C N in HER is (under the speci-
3
4
The change of the absorption spectrum as a function of the
irradiation time caused by decrease of the RhB concentration
in the solution is shown in the Supporting Information (Fig-
ure S11) for the case in which MA31-EtOH-C N is used as a cat-
fied conditions) below the detection limit of the set-up used
À1
(<0.25 mmol(H )h ). In general, better photocatalytic per-
2
formance of MAxy-C N compared with that of g-C N can be
3
4
3
4
ascribed to higher surface areas, but especially to a favorable
microstructure of the products that improves the number and
3
4
alyst. A pronounced blueshift of the most intense absorption
peak from 554 to 497 nm is observed, which is attributed to
the stepwise N-de-ethylation of RhB in the course of the reac-
[
19,20]
the nature of the surface active sites,
as well as a reduced
recombination of photogenerated charge carriers due to short-
er diffusion paths and better charge separation.
[
19]
tion.
The intensity of the main RhB absorption peak at
5
54 nm, which is proportional to the dye concentrations, was
MAxy-C N photocatalysts were stable during RhB degrada-
3
4
used to calculate the rate constants k. On one hand, RhB deg-
radation rates tend to increase with increasing surface areas of
photocatalysts employed. Thus, products derived from ethanol
show higher k values than those of water-derived ones (Fig-
ure 6b). The activities of water-derived photocatalysts can be
improved by increasing their surface areas by using the ap-
proaches discussed above, that is, by improving the stoichiom-
etry of the starting complex (the Supporting Information, Fig-
ure S12A) or by directly heating the gels (Figure S12B). The
second factor that might be responsible for the high photoca-
talytic activity of complex-derived products is their small grain
size. This was assisted by deposition of melamine onto com-
plex particles (despite an excess of melamine in starting com-
plexes, no crystalline-phase melamine was observed in PXRD
patterns, except for the 6:1 complex) and leads to shortening
of the diffusion paths of photo-generated charge-carriers,
tion and hydrogen evolution experiments: Neither structural
nor chemical changes can be detected in MA61-C N after 2 h
3
4
of RhB photodegradation or 12 h of H evolution reactions (the
2
Supporting Information, Figure S14). Therefore, we can con-
clude that all the products obtained from the preformed MAxy
complexes show improved activity in RhB degradation and the
H evolution reaction compared with that of bulk g-C N .
2
3
4
Conclusion
A new complex-templating synthetic approach for preparation
of functional carbon nitrides was developed in which we com-
bined the benefits of using a “green” template and supra-
molecular assembly of the precursors. In this way, the supra-
molecular assembly method was expanded to inorganic and
soluble materials, which allowed us to obtain higher final
yields of carbon nitrides. Despite the lack of crystallinity, the
products obtained show significantly improved photocatalytic
[
9]
which decreases the probability of their recombination. Over-
all, the photocatalytic activity of the surface layer of products
substantially overcompensates the photocatalytic inactivity of
the bulk of products. Such a high activity of the surface layer
activity in RhB degradation and visible light-driven H evolu-
2
tion reactions compared with that of the reference g-C N . The
3
4
Chem. Eur. J. 2015, 21, 10805 – 10811
www.chemeurj.org
10810
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
factors responsible for such a high activity are increased sur-
face area and pore volume, as well as improved charge separa-
tion achieved by smaller grain size of photocatalysts and in-
situ created heterojunction between the melamine- and com-
plex-derived carbon nitride phases. Potentially, this approach
might be further expanded to other inorganic precursors such
as boric acid or layered hydroxides to introduce boron into
C N structure or to improve the crystallinity of the final prod-
Ed. 2012, 51, 68–89; d) A. Thomas, A. Fischer, F. Goettmann, M. Anto-
nietti, J.-O. Muller, R. Schlogl, J. M. Carlsson, J. Mater. Chem. 2008, 18,
4
1
893–4908; e) J. Liu, M. Antonietti, Energ. Environ. Sci. 2013, 6, 1486–
493; f) J. Xu, T. J. K. Brenner, L. Chabanne, D. Neher, M. Antonietti, M.
Shalom, J. Am. Chem. Soc. 2014, 136, 13486–13489.
[3] X. L. Wang, W. Q. Fang, H. F. Wang, H. Zhang, H. Zhao, Y. Yao, H. G.
Yang, J. Mater. Chem. A 2013, 1, 14089–14096.
4] a) M. Groenewolt, M. Antonietti, Adv. Mater. 2005, 17, 1789–1792; b) F.
Goettmann, A. Fischer, M. Antonietti, A. Thomas, Angew. Chem. 2006,
118, 4579–4583; Angew. Chem. Int. Ed. 2006, 45, 4467–4471; c) X.
Wang, K. Maeda, X. Chen, K. Takanabe, K. Domen, Y. Hou, X. Fu, M. An-
tonietti, J. Am. Chem. Soc. 2009, 131, 1680–1681.
[
3
4
ucts, respectively.
[
[
[
5] a) Y. Wang, X. Wang, M. Antonietti, Y. Zhang, ChemSusChem 2010, 3,
4
35–439; b) H. Yan, Chem. Commun. 2012, 48, 3430–3432.
6] Y. Zhang, Z. Schnepp, J. Cao, S. Ouyang, Y. Li, J. Ye, S. Liu, Sci. Rep. 2013,
, 2163, 1–5.
7] W. Xia, B. Qiu, D. Xia, R. Zou, Sci. Rep. 2013, 3, 1935, 1–7.
Experimental Section
3
Synthesis procedure
Melamine (800 mg, 6.34 mmol) was dispersed in ethanol or water
[8] a) Y.-S. Jun, E. Z. Lee, X. Wang, W. H. Hong, G. D. Stucky, A. Thomas, Adv.
Funct. Mater. 2013, 23, 3661–3667; b) Y.-S. Jun, J. Park, S. U. Lee, A.
Thomas, W. H. Hong, G. D. Stucky, Angew. Chem. 2013, 125, 11289–
(
[
50 mL) containing an appropriate amount of the dissolved
Al(NO ) ]·9H O, for example, in the case of the MA31 complex, an
3
3
2
11293; Angew. Chem. Int. Ed. 2013, 52, 11083–11087.
amount of [Al(NO ) ]·9H O (2.32 mmol) corresponding to 10%
3
3
2
[9] M. Shalom, S. Inal, C. Fettkenhauer, D. Neher, M. Antonietti, J. Am.
excess to stoichiometric ratio was used. The mixture was agitated
for 1–2 days in the case of water or 1–2 weeks in the case of etha-
nol, to ensure the complex formation, then the complex was sepa-
rated by using centrifugation and dried at 508C under vacuum.
Chem. Soc. 2013, 135, 7118–7121.
[
10] a) J. P. Mathias, E. E. Simanek, J. A. Zerkowski, C. T. Seto, G. M. White-
sides, J. Am. Chem. Soc. 1994, 116, 4316–4325; b) C. T. Seto, G. M.
Whitesides, J. Am. Chem. Soc. 1993, 115, 905–916.
The obtained white powder was placed in the crucible, heated to
[11] K. Hermansson, Acta Crystallogr. Sect. C 1983, 39, 925–930.
[12] D. J. Goebbert, E. Garand, T. Wende, R. Bergmann, G. Meijer, K. R. Asmis,
D. M. Neumark, J. Phys. Chem. A 2009, 113, 7584–7592.
À1
5
5
508C at a heating rate of 2.38Cmin and maintained for 4 h at
508C under a flow of nitrogen (15 Lmin ). After cooling down,
À1
[
13] a) E. El-Shereafy, M. Abousekkina, A. Mashaly, M. El-Ashry, J. Radioanal.
Nucl. Chem. 1998, 237, 183–186; b) B. Pacewska, M. Keshr, Thermochim.
Acta 2002, 385, 73–80; c) P. Melnikov, V. A. Nascimento, I. V. Arkhangel-
sky, L. Z. Zanoni Consolo, J. Therm. Anal. Calorim. 2013, 111, 543–548.
14] a) G. Zhang, J. Zhang, M. Zhang, X. Wang, J. Mater. Chem. 2012, 22,
the resulting product was washed thoroughly with 1m HCl solu-
tion for 15 h, then with deionized water until neutral pH, and final-
ly dried under vacuum at 508C overnight. Reference bulk C N was
3
4
À1
prepared by heating melamine at 5508C (heating rate 2.38Cmin )
for 4 h under a flow of nitrogen (15 mLmin ); the resulting prod-
[
À1
8083–8091; b) J. Zhang, M. Zhang, C. Yang, X. Wang, Adv. Mater. 2014,
uct was ground thoroughly.
26, 4121–4126.
[
[
[
15] K. Maeda, X. Wang, Y. Nishihara, D. Lu, M. Antonietti, K. Domen, J. Phys.
Chem. C 2009, 113, 4940–4947.
16] X.-H. Li, J. Zhang, X. Chen, A. Fischer, A. Thomas, M. Antonietti, X.
Wang, Chem. Mater. 2011, 23, 4344–4348.
17] a) M. Shalom, S. Gimenez, F. Schipper, I. Herraiz-Cardona, J. Bisquert, M.
Antonietti, Angew. Chem. 2014, 126, 3728–3732; Angew. Chem. Int. Ed.
Description of materials and characterization techniques used as
well the details of photocatalytic experiments can be found in the
Supporting Information.
Acknowledgements
2014, 53, 3654–3658; b) Z. P. Chen, J. Xing, H. B. Jiang, H. G. Yang,
Chem. Eur. J. 2013, 19, 4123–4127; c) J. Zhang, M. Zhang, G. Zhang, X.
Wang, ACS Catal. 2012, 2, 940–948; d) J. Zhang, G. Zhang, X. Chen, S.
Lin, L. Mçhlmann, G. Doł e˛ ga, G. Lipner, M. Antonietti, S. Blechert, X.
Wang, Angew. Chem. 2012, 124, 3237–3241; Angew. Chem. Int. Ed.
Dr. Jian Liu is acknowledged for help with the XPS measure-
ment.
2012, 51, 3183–3187.
[
[
[
18] M. Tabbal, T. Christidis, S. Isber, P. MØrel, M. A. El Khakani, M. Chaker, A.
Amassian, L. Martinu, J. Appl. Phys. 2005, 98, 044310.
19] T. Takizawa, T. Watanabe, K. Honda, J. Phys. Chem. C 1978, 82, 1391–
Keywords: carbon nitride · green chemistry · hydrogen
bonds · photochemistry · template synthesis
1
396.
20] J. Nowotny, T. Bak, M. K. Nowotny, L. R. Sheppard, J. Phys. Chem. B 2006,
10, 18492–18495.
[
1] a) X. Chen, S. Shen, L. Guo, S. S. Mao, Chem. Rev. 2010, 110, 6503–6570;
b) C. Chen, W. Ma, J. Zhao, Chem. Soc. Rev. 2010, 39, 4206–4219; c) A.
Kudo, Y. Miseki, Chem. Soc. Rev. 2009, 38, 253–278.
1
[
2] a) X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K.
Domen, M. Antonietti, Nat. Mater. 2009, 8, 76–80; b) X. Wang, S. Ble-
chert, M. Antonietti, ACS Catal. 2012, 2, 1596–1606; c) Y. Wang, X.
Wang, M. Antonietti, Angew. Chem. 2012, 124, 70–92; Angew. Chem. Int.
Received: February 11, 2015
Published online on June 17, 2015
Chem. Eur. J. 2015, 21, 10805 – 10811
www.chemeurj.org
10811
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim