Heteroatom-induced band-reconstruction of metal vanadates for photocatalytic cyclohexane oxidation towards KA-oil selectivity

Article abstract of DOI:10.1016/j.apcata.2019.02.015

Structure-function correlation constitutes a central theme in the selective photocatalytic oxidation of cyclohexane with molecular oxygen, as a result of precise tailoring band-structure towards KA-oil selectivity. Herein, we comparatively and systematically investigate the correlation among optical properties, electronic structure, and photocatalytic performance of two d°-vanadate photocatalysts (BiVO₄ and Cu₃V₂O₈) using theoretical calculations and experimental techniques. Results show that BiVO₄ promotes the selective oxidation pathway under visible-light irradiation with higher efficiency (TOF = 7.5 h^?1, based on the molar amount of V sites) and chemoselectivity (cyclohexanone/cyclohexanol molar ratio = 3.0) than the Cu₃V₂O₈. Meanwhile, the photocatalytic performance of BiVO₄ demonstrates high photocatalytic selectivity towards KA-oil (95%), while Cu₃V₂O₈ seems to facilitate the accommodation of chlorine promoters, leading to very different chlorocyclohexane selectivity (46%). Such the distinct photocatalytic performances of BiVO₄ and Cu₃V₂O₈ can be ascribed to their unique electronic structures to facilitate the e^??h⁺ charge carriers and generation of reactive species during photoirradiation. The Bi³⁺ 6p states of BiVO₄ are predominant components of the conduction band minimum (CBM) and hybridize with Bi³⁺ 6 s, and O^2? 2p and V⁵⁺ 3d orbitals at the valence band maximum (VBM); whereas the Cu²⁺ 3d and V⁵⁺ 3d orbitals contribute to the CBM of Cu₃V₂O₈, and Cu²⁺ 3d, O^2? 2p and V⁵⁺ 3d orbitals at the VBM. So BiVO₄ with s ? p matrix elements is more effective than the Cu₃V₂O₈ in producing ^[rad]OH, h⁺ and H₂O₂, which can rapidly oxidize the C ? Cl band cleavage for chlorocyclohexane intermediate to regenerate newly Cl^[rad] radicals in the next cycle reactions. Heteroatom-tuning of photocatalytic performance is further supported by universal evidences from experimental results by replacing Bi or Cu of the above vanadates with Ag, La, Sr, Fe or Ti atom and the study of the electrons, holes and radicals mediated reaction pathway by scavengers, respectively.

Full text of DOI:10.1016/j.apcata.2019.02.015

Accepted Manuscript
Title: Heteroatom-induced band-reconstruction of metal
vanadates for photocatalytic cyclohexane oxidation towards
KA-oil selectivity
Authors: Liping Xiang, Jingjing Fan, Wenzhou Zhong, Liqiu
Mao, Kuiyi You, Dulin Yin
PII:
S0926-860X(19)30071-7
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2019.02.015
APCATA 16985
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
30 August 2018
30 January 2019
9 February 2019
Please cite this article as: Xiang L, Fan J, Zhong W, Mao L, You K, Yin
D, Heteroatom-induced band-reconstruction of metal vanadates for photocatalytic
cyclohexane oxidation towards KA-oil selectivity, Applied Catalysis A, General (2019),
https://doi.org/10.1016/j.apcata.2019.02.015
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.

Heteroatom-induced band-reconstruction of metal
vanadates for photocatalytic cyclohexane oxidation
towards KA-oil selectivity
a
a
,a
a
b
a
Liping Xiang , Jingjing Fan , Wenzhou Zhong* , Liqiu Mao , Kuiyi You , Dulin Yin
a
National & Local United Engineering Laboratory for New Petrochemical Materials & Fine Utilization of
Resources, Key Laboratory of Chemical Biology Traditional Chinese Medicine Reserach Ministry of Education,
College of Chemistry and Chemical Engineering, Hunan Normal University, Changsha 410081, P. R.
b
College of Chemical Engineering, Xiangtan University, Xiangtan 411105, PR China
*Corresponding authors: Wenzhou Zhong
Fax: +86-731-88872531; Tel.: +86-731-88872576
E-mail:zwenz79@163.com (W. Zhong)
Graphical Abstract
1

Highlights
0





► d -Vanadate shows superior visible-light responsiveness.
► Heteroatom-tuning vanadates can alter the competitive reaction pathways.
► Electronic structure of vanadates interface to various reaction intermediates change.
-
1
(2.8 h-1).
► BiVO
4
promotes the oxidation with the higher TOF (7.5 h ) than Cu
3
V
2 8
O
► A different KA-oil selectivity is obtained for BiVO
4
(95%) and Cu
3
V
2 8
O
(54%).
2

Abstract
Structure-function correlation constitutes a central theme in the selective photocatalytic
oxidation of cyclohexane with molecular oxygen, as a result of precise tailoring band-structure
towards KA-oil selectivity. Herein, we comparatively and systematically investigate the correlation
0
among optical properties, electronic structure, and photocatalytic performance of two d -vanadate
photocatalysts (BiVO
Results show that BiVO
with higher efficiency (TOF=7.5 h , based on the molar amount of V sites) and chemoselectivity
cyclohexanone/cyclohexanol molar ratio =3.0) than the Cu . Meanwhile, the photocatalytic
performance of BiVO demonstrates high photocatalytic selectivity towards KA-oil (95%), while
Cu seems to facilitate the accommodation of chlorine promoters, leading to very different
chlorocyclohexane selectivity (46%). Such the distinct photocatalytic performances of BiVO and
can be ascribed to their unique electronic structures to facilitate the e −h charge carriers
and generation of reactive species during photoirradiation. The Bi3+ 6p states of BiVO
are
predominant components of the conduction band minimum (CBM) and hybridize with Bi 6s, and
4
and Cu
3
V
2
O
8
) using theoretical calculations and experimental techniques.
4
promotes the selective oxidation pathway under visible-light irradiation
-1
(
3
V
2
O
8
4
3
V
2
O
8
4
−
+
Cu
3
V
2
O
8
4
3
+
2
−
5+
2+
5+
O 2p and V 3d orbitals at the valence band maximum (VBM); whereas the Cu 3d and V 3d
2
+
2−
5+
orbitals contribute to the CBM of Cu
BiVO
3
V
2
O
8
, and Cu 3d, O 2p and V 3d orbitals at the VBM. So
+
4
with s−p matrix elements is more effective than the Cu
3
V
2
O
8
in producing •OH, h and H
2
2
O ,
which can rapidly oxidize the C−Cl band cleavage for chlorocyclohexane intermediate to regenerate
newly Cl• radicals in the next cycle reactions. Heteroatom-tuning of photocatalytic performance is
further supported by universal evidences from experimental results by replacing Bi or Cu of the
above vanadates with Ag, La, Sr, Fe or Ti atom and the study of the electrons, holes and radicals
mediated reaction pathway by scavengers, respectively.
Keywords: Cyclohexane; Photocatalytic oxidation; Visible light; Structure-function correlation;
Electronic structure
3

1
. Introduction
Direct insertion of molecule oxygen in aliphatic C−H bond is one of the most challenging and
promising subjects from industrial and synthesis standpoints due to the high level of activation
required [1, 2]. Catalytic and selective C−H bond functionalization of cyclohexane to KA-oil (a
mixture of cyclohexanone and cyclohexanol) is a commercial procedure in the chemical industry as
the KA-oil is an irreplaceable intermediate in the manufacture of nylon-6 and nylon-6, 6 polymers
-1
[
3]. Owing to the intrinsic inertness of C−H bonds (dissociation energy=439 kJmol ), modern
industrial oxidation of cyclohexane to KA-oil via the conventional thermocatalytic process has to be
conducted under drastic conditions (1.5~3 MPa, 130~160 °C) in the presence of soluble based-cobalt
or manganese catalysts, which leads to high energy consumption and a large amount of undesirable
byproducts [4]. More importantly, maintaining a higher KA-oil selectivity at moderate under
elevated heating conditions to high conversion rates is very difficult in view of the susceptibility of
cyclohexane and KA-oil via the C−C bond cleavage to be completely oxidized to CO
2
and H O in
2
the presence of oxygen. With these facts, given the demand for a green and sustainable
transformation with fewer or even zero byproducts, the construction of an environmentally benign
thermocatalytic-free oxidation process under mild conditions is highly desirable.
Photocatalytic reactions have received enormous attention not only for its potential applications
in environmental treatment [5, 6], but also for organic synthesis of various valuable chemicals [7, 8].
In recent year, considerable efforts on energy shortage have been devoted to the design of
photocatalytic oxidation systems alternative to the existing on the C−H bond thermo-activation of
cyclohexane, with the aim of high efficiency of cyclohexane oxidation. Particularly, getting a high
KA-oil production (e.g., less total oxidation products, CO ) under ambient conditions has been
2
intensively investigated with various photocatalysts such as semiconductor oxides [9], copper(II)
chloride [10], transition metal complexes [11] and NaY zeolite [12]. Among the various
photocatalytic oxidations, bulk TiO with wide band gap energy (3.0–3.2 eV) constitutes a
2
paradigmatic example of catalysts for the oxidative C−H activation of cyclohexane under UV-light
irradiation [13]. However, undesired subsequently overoxidation or mineralization of cyclohexanone
to CO
2
is essentially correlated with the photocatalytic decomposition over TiO due to the
2
−
or O −
[14, 15]. To obtain an
production of highly oxidizing radicals such as surface lattice •O
2
3
enhanced photocatalytic selectivity of TiO
2
, increasing concerns on promoting the formed
−
cyclohexanone desorption and inhibiting highly oxidizing •O
2
production have focused on the
4

modification of the surface or bulk properties of TiO
the Ag-substituted TiO exhibited a cyclohexane conversion of 9% with a 63% cyclohexanone
selectivity in chloroform medium under UV-light irradiation [16]. Mul et al. showed that the
UV-light-driven photooxidation of cyclohexane with surface silylated-TiO led a 95% selectively
conversion <1%) to cyclohexanone [17]. Among them, it is also reasonable to expect that the
formation of heterojunction structures between a narrow band gap materials and TiO can efficiently
extend the photosensitivity of TiO into the visible region. Ide et al. developed a FeO/TiO -catalyzed
visible-light-induced photocatalytic oxidation of cyclohexane into KA-oil (conversion <1%; KA-oil
selectively >99.9%) under a CO atmosphere [18]. Shiraishi et al. designed TiO /rGO hybrid
catalyst, which enhances the activity (conversion <0.5%; 83% KA-oil selectivity) to about twice that
of bare TiO [15]. Han and co-workers found that V /TiO exhibited the high photocatalytic
activity (18% conversion) with a single selectivity to KA-oil at 2.0 MPa oxygen pressures [19]. All
these reports suggest that rational designing of hetero-atom to TiO -containing materials probably
2
materials. The groups of Madras reported that
2
2
(
2
2
2
2
2
2
2
O
5
2
2
produces a multiple heterojunction structure with modified photocatalytic activity. However, the
photocatalytic activities with high partial oxidation selectivity in the visible-light region are
insufficient under ambient conditions. Meanwhile, the relationship between electronic structure and
photocatalytic performance towards KA-oil selectivity has been rarely studied, although over 20
photocatalysts have been discovered for cyclohexane photo-oxidation.
From the viewpoint of electronic state strategy, the transition metal V in +5 valence state
4
+
possesses the same electronic configuration of Ti , but its 3d orbital is usually located below the
analogous d orbitals of the Ti4+ in the energy spectrum and lowers the CBM to a more positive
position [20]. Accordingly, V5+ should be considered to be one of potential core metal ions for
visible-light photocatalyst. Although some literatures have reported that V5+ deposited on the
different supports such as TS-1 [21], TiO
2
[19], C
3
4
N [22] and molybdophosphoric acids [23] can
enhance the photocatalytic activity under visible-light irradiation, no photocatalysts based on core
metal vanadium have been found for cyclohexane photo-oxidation. Especially, critical issue is the
oxidizing ability of photogenerated hot holes within vanadium, which is thought to be much milder
than that of hot holes in the valence band for most semiconductor photocatalysts [24], giving a
possibility for the application in selective C−H bond activation. At present, some vanadates based on
core metal vanadium such as BiVO
4
, InVO
4
, AgVO
4
and Zn
3
V
2
8
O have been developed successfully
as visible-light photocatalysts for the photocatalytic degradation of dye molecules [25], water
5

splitting [26] and CO conversion [27]. In addition, these vanadates also allows a large change in the
2
crystal morphology and new band structure, because atomic arrangement is expected to form and
contribute to some new properties. Rational design and heteroatom tuning of electronic structure of
photocatalysts, therefore, have become one of the most effective strategies to control their
photocatalytic activity, which is also evident in several photocatalytic processes. For example, Fan
and his co-workers investigated the effect of three crystalline phases of BiVO
4
on the degradation of
had the highest photocatalytic activity [28].
, Zn ) heterostructure photocatalysts
methyl blue aqueous solution, and monoclinic BiVO
Cao et al. investigated TiO /vanadates (Sr10
for photoreduction of CO into CH
4
2
V
6
O25, Ni
3
V
2
O
8
2
V
2
O
7
2
4
under visible light irradiation, and revealed that TiO
2
/Sr10
V
6
O
25
could perform as an excellent photocatalytic material [29]. On the basis of the aforementioned
considerations, so this family of vanadates provides opportunities to design new efficient
photocatalysts for controlling cyclohexane oxidation towards KA-oil selectivity. Further, it will be
clear whether the heteroatom-induced band-reconstruction would alter the competitive reaction
pathways towards KA oil? If yes, what are its implications?
0
In the present study, we have chosen two d -vanadate photocatalysts (BiVO
4
and Cu
3
V
2
8
O ) by
taking advantage of strong heteroatom-vanadate interactions to selectively modify the structure of
photocatalyst. The special interactions between heteroatom and vanadate can introduce large
electronic perturbations in electronic structure and induce obvious variations in crystal structure,
which directly affects the catalytic performance. Cyclohexane selective oxidation on
heteroatom-tuning vanadate catalysts is systematically evaluated under visible-light irradiation.
When the heteroatoms are tuned, the electronic structures of vanadates interface to various reaction
intermediates change. This response enables the significant tuning of the photo-catalysts selectivity,
whereas the overall photo-catalytic activity is much less affected. A combination of comparative
experiments and electronic band structure calculations for the vanadate photocatalysts can provide
insights into the structure-function relationships: how the catalytic performance of vanadate catalysts
for cyclohexane oxidation can be fine-tuned.
2
. Experimental
2
.1. Catalyst preparation
The monoclinic BiVO was prepared via a simple precipitation method. In a typical procedure,
4
5
mmol Bi(NO
3
)
3
·5H
2
O was dissolved in 30 mL HNO
3
(1.5 M), and 5 mmol NH
4
VO was added to
3
6

5
mL of 15% NH
3
·H
2
O. Then, the above solutions were mixed together under stirring, and the pH
·H O. The resulting suspension was stirred for 3 h at room
value was adjusted to 7 by 15% NH
3
2
temperature. The yellow precipitate was collected and washed with ethanol and deionized water
three times each, and dried at 80 °C in air for 6 h.
For the synthesis of Cu
3
V
2
O
8
, 5 mmol Cu(NO
solution (0.1 M), which was prepared by dissolving NH
solution under vigorous stirring and the
3
)
2
·3H O was dissolved in 50 mL deionized water.
2
Then, a 50 mL NH
4
VO
3
4
VO into deionized
3
o
water at 80 C for 5 min, was slowly added to the Cu(NO
)
3 2
mixture was stirred for about 2 h. The precursor was collected and washed with ethanol and
deionized water three times each, and dried at 80 °C in air for 12 h. The final product was obtained
o
by calcination at 500 C in air for 2 h. In addition, TiVO
4
, SrVO
4
, CrVO
4
, LaVO
4
, CoVO
4
, AlVO
4
and FeVO
For the sake of comparison, single-metal oxides (Bi
calcination of the corresponding salts (Bi(NO ·5H O, NH
4
were prepared by a similar procedure.
2
O
3
, V
2
O
5
and CuO) were prepared by
o
3
)
3
2
4
VO
4
and Cu(NO
3
)
2
·3H
2
O) at 500 C for
8
h.
2
.2. Characterization techniques
The crystalline phase of the samples was analyzed with a Bruker D8 Advance X-ray
diffractometer using CuKα radiation (λ=1.542 Å). Transmission electron microscopy (TEM)
images and high-resolution transmission electron microscopy (HR-TEM) images were obtained by
using a JEOL-2010 TEM with an acceleration voltage of 200 kV. Raman spectra of samples were
analyzed with a Renishaw in Via Raman Microscope system with a 514.5 nm laser source and a
−
1
spectral resolution of 0.3 cm . The optical properties of the samples were analyzed by diffuse
reflectance UV-vis spectra using a Varian-Cary 5000 spectrometer with BaSO as a reference
4
material. Surface composition and surface states of samples were determined by X-ray
photoelectron spectroscopy (XPS) using an Imaging Photoelectron Spectrometer (Axis Ultra,
Kratos Analytical Ltd.) with Al Ka radiation (1486.7 eV). Binding energies were corrected with
reference to the C 1s peak of adventitious carbon (binding energy = 284.8 eV); fit XPS software
was used for curve fitting. Photoluminescence measurement was performed using a JASCO
FP-6500 type fluorescence spectrophotometer with an excitation wavelength of 290 nm light at
room temperature.
2
.3. DFT calculations
All our calculations were performed based on the plane-wave pseudopotential method using the
7

CASTEP code. The generalized gradient approximation (GGA) with Perdew-Burke-Ernzerhof (PBE)
functional was adopted for the exchange-correlation potential [30]. Hubbard-like potential U was
employed to better treat exchange and correlation associated with the localized d-states of the Cu
and V cations for Cu
Monkhorst−Pack k-points grid of 2×4×4 and 2×2×4 was setted for sampling in the first Brillouin
zone during geometry optimizations of BiVO and Cu , respectively. The atomic coordinates
3
V
2
O
8
[31] The plane-wave energy cutoff was set to 400 eV. A
4
3
V
2
O
8
and the shape of the cell were allowed to fully relax, until the residual forces on all atoms, the
stresses on the cell, and the convergence tolerance for self-consistent energy were less than 0.01
-5
eV/Å, 0.05 GPa, 0.001 Å, and 10 eV, respectively. The electronic structure calculation was
performed using the optimized geometric structure.
2
.4. Photocatalytic reactions
Photocatalytic selective oxidation of cyclohexane into KA-oil was performed using a
self-designed photochemical reactor equipped with a water-cooling jacket, which was connected
with a oxygen gas burette to monitor the gas uptake in our previously published work (Fig S1) [21].
The photoreactor consisted of a jacketed quartz tube of 3.8 cm i.d. and 20 cm length and an inside
reactor of 1.5 cm i.d. and 19 cm length. A 20 W tungsten-bromine lamp (light intensity, 275
2
mW•cm ) with an UV light filter (Osram brand) was immersed in a mixture of acetonitrile (5 mL),
cyclohexane (5 mmol), concentrated HCl (0.1 mL) and 50 mg photocatalyst. After being filled with
the desired amount of oxygen (1atm), the reaction suspension was vigorously stirred to eliminate the
influence of mass transfer. Owing to the heating effect of light irradiation illuminated, the
o
temperature of reaction solution could maintain at 28–30 C. After illumination reaction, the mixture
was centrifuged to completely remove the catalyst particles. The liquid products were quantitatively
analyzed by an Agilent 6890N gas chromatograph equipped with flame ionization detector and a
DB-17 polysiloxane capillary column (30 m × 0.32 mm × 0.50 μm) using n-hexanol as an internal
standard. The gas products were analyzed by gas chromatography equipped with a TCD detector. All
products were satisfactorily identified by comparing the MS spectra with those of the authentic
samples. All raw materials and products were established carbon mass balances. All the experiments
were repeated at least 4 times in parallel to obtain an average value.
3
. Results and discussion
3
.1. Crystal structure and morphology
The crystal structure and phase compositions of the materials are investigated by XRD analysis.
8

As shown in Fig. 1, the peaks of BiVO
6.71°, 47.31°, 50.32° and 53.25° are attributed to the diffractions of the (121), (040), (200), (002),
211), (051), (240), (042), (202) and (161) crystal planes of monoclinic BiVO (JCPDS No.
0-014-0688) [32], respectively. Interesting, Cu material structured by strong Cu -vanadate
interactions has a similar phase by comparison with BiVO . The peaks at 2θ values of 15.0°, 18.7°,
6.8°, 27.3°, 30.8°, 32.2°, 36.1°, 37.0°, 37.7°, 38.6°, 45.2° and 58.5° are distinctly indexed to the
100), (-110), (021), (111), (-211), (012), (-212), (130), (-131), (102), (040) and (-330) crystal planes
of Cu with a monoclinic phase (JCPDS No. 26-0567) [33], respectively. Compared with the
XRD patterns of the CuO (JCPDS No. 48-1548), Bi (JCPDS No. 27-0050) and V (JCPDS No.
1-1246), no other peaks for impurity are observed for BiVO and Cu , which confirms the
high purity of the resultant metal vanadates in phase composition. Further, Fig. 6 (A and B ) shows
schematic crystal structures of BiVO and Cu . The crystal structure of Cu consists of the
V−O tetrahedra and Cu−O octahedra, while the monoclinic phase BiVO is composed of the V−O
4
located at 28.82°, 30.55°, 34.50°, 35.22°, 39.78°, 42.47°,
4
(
4
2
+
0
3
V
2
O
8
4
2
(
3
V
2
O
8
2
O
3
2
O
5
4
4
3
V
2
O
8
0
0
4
3
V
2
O
8
3
V
2
O
8
4
tetrahedra and Bi−O polyhedra. We think that the Bi3+ cation in the polyhedral field with the
electronic states is obviously different from the Cu2+ cation in the octahedral field, which should
impose significant effects on the electronic structure and photocatalytic activities of the studied
metal vanadates.
The crystal structures of BiVO
Fig. 2). Evidently, BiVO exhibits a worm-like morphology with an average diameter of 50 nm.
The obvious contrast between the dark edge and the pale center confirms that the BiVO contains
hollow spheres, and the average diameter of hollow spheres is approximately 30-90 nm. Interesting,
TEM image of Cu clearly shows that the individual sphere of Cu is also porous hollow
4
and Cu
3
V
2
O
8
are investigated by TEM and HR-TEM analysis
(
4
4
3
V
2
O
8
3
V
2
O
8
sphere with an average diameter of 30-150 nm, but the wall is comprised of loosely packed
aggregates of irregular flakiness. Obviously, the morphology of these metal vanadates is different
from that of the corresponding the single-metal oxides (Bi
In addition, the corresponding HR-TEM image of BiVO
nm and 0.475 nm match those of the (110) and (121) planes of monoclinic BiVO
lattice distance of 0.363 nm is found in Cu , attributed to the (-211) plane of Cu
2
O
3
, V
2
O and CuO), as shown in Fig. S2.
5
4
reveals that the lattice distances of 0.309
4
[34]. A different
. These
3
V
2
O
8
3
V
2
O
8
lattice distances can be clearly seen in the corresponding HR-TEM image, indicating that the BiVO
and Cu possess a good crystallization, respectively.
.2. Surface properties and band gap energy
4
3
V
2
O
8
3
9

The incorporation of heteroatom into the vanadates solid has, therefore, important effects on the
local structure surrounding the core V atoms. The structural information of BiVO and Cu
with the single-metal oxides (Bi , V and CuO) from the Raman spectra is shown in Fig. 3.
Compared with the corresponding single-metal oxides (Fig. 3B), the vibrational characteristic
4
3
V
2
8
O ,
2
O
3
2
O
5
−
1
Raman peaks at 126, 209, 247, 326, 363, 758 and 825 cm are the clear signature of the typical
−
1
vibrations of monoclinic BiVO
external modes (rotation/ translation) of BiVO
asymmetric (B bending) and symmetric (A
bending) deformation modes of the VO 3−
due to a strong interaction between Bi and VO
4
(Fig. 3A) [34-36]. The peaks at 126 and 209 cm correspond to the
−
1
4
. The peaks at 326 and 363 cm are ascribed to the
g
g
4
tetrahedron
3
+
3−
−1
4
, respectively. The intense peak at 824 cm is
attributed to a symmetric V−O stretching mode with A
g
symmetry, along with a weak shoulder at
about 758 cm assigned to an antisymmetric V−O stretching mode with B symmetry. In addition,
the peak at 247 cm−1 relates to the Bi–O symmetric stretching in a deformation mode of BiO
−
1
g
8
−
1
dodecahedron and at 209 cm corresponds to the vibration of the crystal lattice (external modes),
2
+
respectively. In contrast, Raman peak pattern of Cu
3
V
2
O
8
by strong Cu -vanadate interactions,
however, is obviously different from that of BiVO
4
sample. The peak intensities of deformation
VO 3
−
modes are greatly decreased, which can be explained on the basis of the packing of the crystal
4
structure. (Fig. 3A) [36]. Furthermore, a positive shift in the vibrational mode of V−O has been
observed, which suggests that the V−O bond length of Cu is disturbed by the formation of the
V−O−Cu linkages. This is also indicative of the presence of the bending of the same structural motif
or a breaking of the total symmetry of the VO 3−
structure [37, 38], which is confirmed by the above
theoretical studies. Such a similar phenomenon was earlier predicted by Colón et al., on Cu-doped
TiO catalysts [39]. Therefore, it is clear from the figure that the Raman study provides the valuable
3
V
2
O
8
4
2
information on the structural variations by incorporation of heteroatoms into vanadates. However, it
is unclear the electronic structure difference of the as-prepared composites and in what state the
heteroatom remains into the vanadates lattice. Further, we conducted XPS test to analyze the
changes of valence state of elements and binding energy.
For comparison, the XPS spectra of the BiVO
surface survey XPS spectrum given in Fig. 4A reveals the presence of Bi, V, O and C elements in
BiVO sample and Cu, V, O and C elements in Cu sample, confirming that other metals and
impurities are not present. The appearance of C is atmospheric in origin. The high resolution XPS
spectrum of Bi for BiVO shows the binding energies of Bi 4f located at 164.6 and 159.0 eV
4
and Cu
3
V
2
O
8
are exposed in Fig. 4. The full
4
3
V
2
O
8
4
1
0

corresponding to the Bi 4f7/2 and Bi 4f5/2 (Fig. 4B), respectively, which can be assigned to Bi3+
valence state of the composites [40]. Meanwhile, the curve of V 2p displays two strong peaks
centering at about 516.0 and 524.1 eV, which is attributed to the binding energy of V 2p3/2 and V
2
p1/2 respectively, indicating that V species exist in +5 valence state for BiVO
41]. Interesting, a binding energy shift of V 2p3/2 and V 2p1/2 for Cu toward higher binding
energy at 516.7 and 524.6 eV takes place compared with BiVO (due to the different
electronegativities of Bi and Cu ), indicating that the electron cloud density is reduced. This may
4
samples (Fig. 4D)
[
3
V
2
O
8
4
3+
2+
2+
5+
be ascribed to the strong electrostatic attraction of Cu , influencing the local chemical state of V
species. That is to say, to achieve local charge balance in the Cu lattice, some of the V ions
3
V
2
O
8
may acquire higher oxidation state by releasing electrons, which is consistent with the structural
variations in spectroscopic (Raman) studies. In addition, as shown in Fig. 4C, the binding energy of
the Cu 2p3/2 (934.33 eV) and Cu 2p1/2 (955.1 eV) peaks, together with the strong satellite peaks in
the range of 940−945 eV [42], indicates that Cu2+ is the dominant valence state for Cu
V
2
3
8
O . A
+
weak component (3%) at 931.59 eV, associated with the presence of Cu , is observed after
deconvolution of Cu 2p3/2 peak [42]. Therefore, one more convincing follow up is the gap states in
+
the forbidden band, which is formed due to the introduction of Cu states. Thus, electrons present in
the gap states of forbidden band can be excited to the conduction band by means of a wide range of
energy. This facilitates the elongation of light absorption into the visible region, which is further
confirmed by the following UV-vis spectra. From the observation in O 1s spectra (Fig. 4E), the
resolved O 1s peaks located at 529.1 and 531.2 eV are assigned to O–V and O–Bi bonds of BiVO
43], while peaks at 529.2 and 531.5 eV are ascribed to O–V and O–Cu entities of Cu
respectively. Obviously, the peak of O–V slightly shifted toward a higher binding energy value of
about 0.10 eV for Cu , further suggesting that there is strengthening of O–V bond with a
4
[
3
V
2
8
O ,
3
V
2
O
8
modified bonding state. As a result, the differences in the intensities and peak position of the peaks
in the Raman and XPS spectra clearly demonstrate the variations in the local crystal structure of the
vanadates samples, as the packing of the structure related to the interactive forces between the
3
−
heteroatom and VO
to further understand the change of band structure after heteroatom-doping, we determined the
position of VBM. As shown in Fig. 4F, the VBM value of BiVO and Cu is 2.71 and 1.65 eV,
respectively, which is in line with the following calculated value of VBM by the empirical formula
. This suggests that the presence of heteroatom causes band-reconstruction of metal
4
cations is directly related to the electronic transitions in the materials. In order
4
3
V
2
O
8
EVB = ECB + E
g
11

vanadates and its influence on KA-oil selectivity.
To understand the variation of the optical bandgap upon the change in the substitution of metal
atom in the vanadates, the UV-vis absorption spectra of BiVO
oxides (Bi , V and CuO) are analyzed using UV-vis spectroscopy, as shown in Fig. 5A.
Comparison of all the BiVO and Cu photocatalysts spectra with those of referent single-metal
oxides, both of BiVO and Cu generally exhibit a strong absorption from UV-light to
visible-light region, implying the possibility of visible-light photocatalytic activity over these
samples. Compared with the absorption edge of BiVO , Cu has a red shift from 500 to 540 nm,
suggesting that the Cu can absorb longer wavelength light. Moreover, an apparent absorption
hump with an extrapolated absorption edge at about 750 nm is appeared for Cu , which is
associated with the Cu d–d orbital intraband transition in the octahedral field. In addition, band-gap
energy is calculated by the formula αhν = A(hν − E and A are the Planck
)n/2, where h, α, ν, E
4
and Cu
3
V
2
O
8
, with the single-metal
2
O
3
2
O
5
4
3
V
2
O
8
4
3
V
2
O
8
4
3
V
2
O
8
3
V
2
O
8
3
V
2
O
8
2
+
g
g
constant, the absorption coefficient, light frequency, band gap and proportionality constant,
respectively [44]. Among them, the n value depends on the type of electronic transition: n = 2 for a
direct transition; n = 0.5 for indirect transition. Based on the following theory calculation results (Fig.
6
), the electronic transitions of BiVO
the E of BiVO
eV, whereas that of Cu
about 1.83 eV (Figs. 5B and 5C). Further, the conduction band (CB) for vanadates can be also
4
and Cu
3
V
2
O
8
are direct and indirect, respectively. As a result,
2
g
4
can be calculated from the plot of (αhv) versus (hv) and founded to be about 2.34
3
V
2
O
can be estimated from the plot of (αhv)1/2 versus (hv) and founded to be
8
e
e
calculated by the equation: ECB = χ − E − 0.5E
g
, in which ECB, χ, E and E are the CB potential, the
g
geometric mean of the absolute electronegativities of the constituent atoms (the absolute
electronegativity of an individual atom is the arithmetic mean of the atomic electron affinity and the
first ionization energy [45]), the free electron energy on the hydrogen scale (~4.5 eV), and the
vanadate band gap energy [46-48]. Additionally, the value of EVB can be obtained by EVB = ECB + E
Thus, the CBs for BiVO and Cu are calculated to be 0.37 and −0.18 eV, and the VBs for
BiVO and Cu are calculated to be 2.71 and 1.65 eV, respectively.
.3 DFT simulated electronic and optical properties
Aiming at exploring the impact of heteroatom-tuning on the crystal structure and electronic
structure of metal vanadates, the full PBE calculations are performed for the unique light absorption
and photocatalysis properties presented by the BiVO and Cu catalyst. As shown in Table 1,
the optimized cell constants of BiVO and Cu reproduce very well the experimental data [49,
g
.
4
3
V
2
O
8
4
3
V
2
O
8
3
4
3
V
2
O
8
4
3
V O
2 8
1
2

5
0], and this variation of lattice parameters may be associated with the different ionic radii of the
2
+
3+
Cu and Bi cations. The average V–O bond length calculated is 1.740Å in BiVO
Cu , and less V–O bond length will lead to form a more dispersive band. The Bi–O bond length
in BiVO is obviously larger than the Cu–O bond length in Cu , which makes the crystal field
of Cu stronger than that of BiVO . Figs. 6A and 6B show the calculated electronic band
structures for the BiVO and Cu , respectively. BiVO has an indirect and minimum energy gap
4
and 1.980 Å in
3
V
2
O
8
4
3
V
2
O
8
3
V
2
O
8
4
4
3
V
2
O
8
4
of 2.22 eV, along with two direct and only slightly larger energy gaps (2.28 eV and 2.29 eV) at the A
point (-0.5, 0, 0) and Z point (0, -0.5, 0.5). Usually, the direct band gap is favorable to recombine
photogenerated electrons and holes in the semiconductors, which indicates that these two narrow and
direct energy gaps for BiVO
4
should mainly contribute to the high photocatalytic activity in the
visible light region [51]. In comparison, the indirect energy gap is observed for Cu
3
V
2
8
O , which is
consistent with the experimental and previous theoretical results [52]. However, it is noted that all of
the calculated band gaps are generally smaller than the corresponding experimental values due to the
well-known LDA error, which is only expected to provide a rather approximate description of the
band gaps and excited [53].
Fig. 7 shows the total and site-projected density of states (DOS). Obviously, the CBM of
is mainly contributed by the Bi3+ 6p and V 3d states, whereas the CBM of Cu
5+
V
2
BiVO
4
3
8
O is
5
+
2+
composed primarily of the V 3d and Cu 3d states. Similarly, the lowest unoccupied molecular
is predominantly composed of Bi3+ 6p, V5+ 3d and O 2p orbitals.
2–
orbital (LUMO) of BiVO
4
However, the LUMO of Cu
3
V
2
O
mainly derives from the hybridization of V5+ 3d and Cu2+ 3d
8
orbitals. As we known, it will be useful to have more p-character in the CBM to avoid recombination
in photocatalytic reactions, when the VBM mainly consists of oxygen p-orbitals [54]. Thus, Bi3+
contributes the p-character of BiVO
4
in the stability of the electrons in the CBM by reduction of
recombination rates of excited charges. The partial density of states of individual elements in BiVO
4
show that the contribution of the Bi3+ 6s lone pair plays an important role in the construction of
VBM with a dispersive band structure (Fig. 7A), which contributes in decreasing the effective mass
of electrons and improves the photocatalytic activity of photocatalysts [55]. Further, the contribution
3
+
2+
of the Bi 6p orbitals to the CBM of BiVO
4
is about 10%, whereas Cu 6p accounts for only 0.4%
processes a strong covalent feature of the Bi−O and
of the orbital formation, indicating that BiVO
4
V−O bonds. This shows that more electrons should transfer from Bi (or V) to adjacent oxygen atoms
rather than being shared between Bi (or V) and oxygen atoms, facilitating the charge separation of
1
3

photogenerated carriers and thus benefiting the photocatalytic performance [56]. On the other hand,
the VBM structures of photocatalysts are also quite different (Figs. 7A and 7B). The VBM of BiVO
4
is mainly composed of O 2p, Bi 6s and V5+ 3d orbitals, whereas the VBM of Cu
2
–
3+
V
2
O
8
is
3
2
–
2+
5+
3+
2–
dominated by O 2p, Cu 3d and V 3d character. As a result, Bi 6s mixed with O 2p orbitals
2
+
2–
dominate the HOMO of BiVO
Cu . So the potential for valence band top of BiVO
required for the photocatalytic oxidation of organic compounds in the presence of oxygen. Based on
the PDOS, the unoccupied Cu 3d states contribute to the lowest VB bottom of Cu , resulting in
that the visible-light absorption of Cu with the narrowest band gap is the strongest (which is
confirmed by the XPS and UV-vis spectra analysis).
4
; whereas Cu 3d mixed with O 2p orbitals dominate the HOMO of
3
V
2
O
8
4
is different from that of Cu
3
V
2
8
O , which is
3
V
2
O
8
3
V
2
O
8
3
.4 Photocatalytic oxidation of cyclohexane
Photocatalytic activity of the as-synthesized vanadates are evaluated to establish how the extent
of heteroatom-tuning for vanadates can exert an influence on the photooxidation of cyclohexane into
KA-oils in the presence of concentrated HCl as an additive under visible light irradiation, as shown
in Table 2. No product is detected in the control experiments without catalyst or light irradiation,
which indicates that the cyclohexane oxidation is truly driven by a photocatalytic process (entry 1).
Similar visible light stimulation of pure V
photocatalytic activity for cyclohexane oxidation in CH
electron-hole recombination is the dominant process under these experimental conditions (entry 2).
In contrast, as shown by entries 5−6, BiVO and Cu structured by heteroatom-vanadate
2
O
5
as a core metal element results in a considerable
3
CN, which is due to the fact that fast
4
3
V
2
O
8
interactions remarkably enhance visible light photocatalytic activity for the KA-oil formation,
indicating that the modification of the crystal and electronic structure in the metal vanadate is
responsible for the increasing visible light absorption performance. To compare the intrinsic activity
of vanadium species in the different metal vanadate, turnover frequency (TOF) toward selective
formation of KA oils is calculated for cyclohexane conversion based on V sites on catalyst. BiVO
4
-1
catalyst successfully and repeatedly displays a high TON value of 7.5 h , giving 15.5% cyclohexane
conversion and 95.0% KA-oil selectivity (cyclohexanone/cyclohexanol molar ratio =3.0). Compared
-1
with BiVO
4
, Cu
3
V
2
O catalyst shows a relative low TON value of 2.8 h , giving 9.1% cyclohexane
8
conversion and 54.2% KA-oil selectivity (cyclohexanone/cyclohexanol molar ratio =1.3). The
decrease in KA-oil selectivity may be mainly due to less further photo-oxidation of the formed
chlorocyclohexane intermediate to the re-formation of KA-oils via secondary reactions, since the
1
4

BiVO
4
catalyst is different from Cu
3
V
2
8
O in generation of interactive intermediates (such as
electrons, holes and •OH) for the cleavage of C−Cl bond of chlorocyclohexane to re-form radical Cl•
species. Actually, the reactive Cl• as highly effective species to capture H atom of cyclohexane in
the photocatalysis cycling of BiVO
4
and Cu
3
V
2
8
O , is further confirmed by compared to that without
the additive HCl (entries 7 and 8). Moreover, the additive concentrated HBr instead of concentrated
HCl is investigated as a possible reference (entry 9). No photo-catalysis efficiency, however, is
obtained under similar conditions, revealing that the Cl• generated in entries 5 and 6 can undergo
reversible chlorine-atom transfer with carbon-centered radicals. This is likely because the Br atom
formed is no easy to capture H atom of cyclohexane due to its low activity and the cleavage of C−Br
bond is also difficult in the existence of high Br free radicals [57]. These catalytic activities of metal
vanadates by the substitution of Bi3+ ions for Cu2+ demonstrate not only the close correlations
between photocatalytic abilities and generation of interactive intermediates and charge carriers.
To understand why the substituting of metal element into the framework of vanadates is able to
enhance its photocatalytic activity significantly, other metal vanadates such as TiVO
CrVO , LaVO , CoVO , AlVO and FeVO are also prepared and submitted to the oxidation reaction
Table 2). Among different metal vanadates, the substituting of Co , Fe and Al into the vanadate
4
, SrVO ,
4
4
4
4
4
4
2
+
3+
3+
(
framework is also an effective method to improve visible photocatalytic activity on oxidation of
cyclohexane into KA-oils (entries 14−16), and a high KA-oil selectivity of 93.1% can be obtained
for AlVO
4
. In contrast, TiVO
4
, SrVO
4
, CrVO
4
and LaVO show much less photocatalytic activity
4
(
entries 10−13). Interesting, TiVO
4
exhibits the highest cyclohexanol/cyclohexanone molar ratio
value of 4.1 among all metal vanadates, giving a 77.6% cyclohexanol selectivity. These results show
that the catalytic activities of metal vanadates appear to be closely related to its heteroatom-induced
band-reconstruction, as above discussed in the differences of BiVO
4
and Cu
3
V
2
8
O . This further
indicates that the selective cyclohexane oxidation towards KA-oils selectivity can be tuned using
heteroatom-tuning electronic structure.
Perhaps the compelling evidence regarding the singular nature of the BiVO
catalyst can be obtained from the different behavior of the product evolution in the cyclohexane
photo-oxidation. To study this, the product distributions measured with the BiVO and Cu
catalyst at different time intervals is shown in Fig. 8. A similar low conversion of cyclohexane is
observed for BiVO and Cu after irradiation for 0.5 h, and then the level of cyclohexane
conversion increases linearly with an increasing irradiation time from 0.5 to 8 h. A possible
4
and Cu
3
V
2
O
8
4
3
V O
2 8
4
3
V
2
O
8
1
5

explanation for this phenomenon is a similar radicalic-type autoxidation process with the BiVO
4
or
Cu as catalyst. Interesting, for the BiVO catalyst the cyclohexane conversion rate initially is
3
V
2
O
8
4
the higher under visible-light irradiation, implying a higher utilization efficiency of visible-light.
However, the evolution of the rates of formation of cyclohexanone, cyclohexanol and
chlorocyclohexane are remarkably different during the cyclohexane photo-oxidation process. For the
BiVO
4
and Cu
3
V
2
O
8
catalyst, the increased cyclohexanone formation rate and decreased
cyclohexanol formation rate are appeared for prolonged reaction times, indicating that the products
are formed in parallel and cyclohexanol formed initially can be transformed to cyclohexanone via a
sequential oxidation process as the reaction proceeds. As a comparison, the formation rate and
ketone/alcohol molar ratio, however, exceed that of Cu
the chlorocyclohexane formation rate initially increases, reaching a maximum (16.9%) after
irradiation for 1 h and smoothly decreasing to 6.7% (3 h); whereas the Cu catalyst initially
3
V
2
O
8
used. In addition, for BiVO catalyst,
4
3
V
2
O
8
shows a slight decrease for the chlorocyclohexane formation rate, but after 2 h the formation rate
always remains “frozen” in that state, independent of the irradiation time. The analysis of
dependences of the formation rates of KA-oil and chlorocyclohexane confirms that KA-oil is formed
by reversible chlorine-atom transfer process. To find out the other reaction intermediates, we took
aliquots of the sample at regular intervals and analyzed them by iodometric titration. No peroxides
such as hydroperoxide and cyclohexyl hydroperoxide are detected, indicating that the peroxide
intermediates can be quickly decomposed into the active radicals by catalysts under irradiation
conditions.
3
.5 Photocatalytic reactive species
Molecular oxygen plays a major role in photocatalytic selective oxidations elicited by
photocatalyst, because molecular oxygen is involved in the generation of reactive species (such as h+,
−, Cl• and •OH) and driving catalytic reactions. Electronic structures and optical properties of the
photocatalysts, therefore, will have significant effects on the photocatalytic activity by tailoring
reactive species production. As a result, trapping experiments for both BiVO and Cu are
•O
2
4
3
V
2
O
8
conducted to trace the different active species in cyclohexane photocatalysis. The scavengers (10
mmol/L) triethanolamine (TEOA), benzoquinone (BQ), dimethylbutadiene (DMBD) and
+
−), halogen (Cl•) and
isopropanol (IPA) are used to trap holes (h ), superoxide anion radicals (•O
2
hydroxyl radicals (•OH), respectively [58-60]. As shown in Table 3, introducing TEOA into the
photocatalytic system causes the BiVO photocatalyst to quickly deactivate, confirming that
4
1
6

+
photogenerated h is the mainly active species in BiVO
4
sample during the selective oxidation of
has less opportunity to take part in
+
cyclohexane. However, scavenger of h for Cu
3
V
2
O
8
photocatalytic oxidation process (entry 1). This is attributed to the fact that the lone pair of electrons
of N element of TEOA can combine with the empty d-orbitals of the transition metal Bi element of
photocatalysts to form strong covalent bonds. Differently, when IPA is added into the photocatalytic
reaction solution, cyclohexane conversion decreases from 15.5% to 5.8% and 9.1% to 7.7% for
BiVO
selective oxidation reaction and BiVO
compared to Cu . By contrast, no obvious change is observed on the photo-oxidation of
4
and Cu
3
V
2
O
8
, respectively (entry 2); therefore, it is concluded that •OH takes part in the
4
photocatalyzes more production of hydroxyl radical
3
V
2
O
8
cyclohexane with the introduction of p-benzoquinone into catalytic system (entry 3), indicating that
−
the less •O
2
radicals produce for BiVO
4
and Cu
3
V
2
8
O under visible light irradiation. In addition, a
low photoconversion of cyclohexane into KA-oil is also obtained with the addition of DMBD into
the photocatalytic reaction solution (entry 4), revealing that the formed Cl• radicals from Cl−
oxidation are expected to efficiently initiate the same chain process by abstraction of hydrogen from
cyclohexane under our conditions.
−
As is well known, the photogenerated electrons (e ) are another key reactive species in
−
−
photocatalytic reaction because e can produce •O
2
and •OH. Therefore, we used the AgNO
3
(1 mM)
−
to trap e generating from BiVO
4
and Cu
3
V
2
O
8
photocatalysts, because the separated electrons are
used up in making metallic silver. Interesting, the higher cyclohexane photoconversion is observed
−
over BiVO
BiVO
take part in reaction process. In this case, the •OH radicals generated in BiVO
are not from photogenerated electrons, but from photogenerated holes. Further, the remarkably
different photocatalytic performance of the BiVO photocatalyst, compared with that of Cu
4
under visible light irradiation (Table 3, entry 5), implying that the scavenger of e for
4
has less opportunity to join recombination of electron−hole pairs and produces more holes to
4
photocatalyst system
4
3
V
2
O
8
,
can be also ascribed to the different charge carriers trapping, migration and separation efficiency.
Photoluminescence emission spectroscopy (PL) is carried out to investigate the transfer and fate of
photoinduced carriers: the lower PL intensity generally represents the higher separation capacity of
charge carriers in semiconductor-based systems, resulting in a higher photocatalytic activity [61].
Obviously, a strong emission peak at about 520 nm is observed for the Cu
), demonstrating that the electrons and holes over the Cu can facilely recombine. By contrast,
the relatively low PL intensity of BiVO implies that BiVO can migrate the photogenerated
3
V
2
O
8
photocatalyst (Fig.
9
3
V
2
O
8
4
4
1
7

electrons and holes more effectively, with inhibiting the recombination of charge carriers. The
discovery by the PL spectra is an evidence for the previous Raman spectroscopy study and crystal
3–
structure analysis in that the activity is tuned by the V–O bond length of VO
4
for metal vanadates.
tetrahedron in BiVO is 1.74Å, and it is
catalyst. Based on the results of Shannon et al., the electron
3
–
We have proved that the average V–O bond length of VO
4
4
obviously shorter than that of Cu
3
V
2
O
8
delocalization of the metal-oxygen bond in oxides easily occurs in the reduced metal-oxygen bond
length in comparison to metal oxides [62]. Therefore, increased electron delocalization of the short
V–O bond for BiVO
4
will increase effective transfer of charge carriers, and then lead to the
enhanced photocatalytic activity.
3
.6 Photocatalytic reaction mechanism
The above results provide the direct evidence for that (i) BiVO
4
and Cu
3
V
2
O
8
distinctly differ in
+
photo-generation of reactive species. BiVO
and Cl•, and (ii) processes a higher separation capacity of charge carriers than Cu
observations are well predicted by the energy band structures of BiVO and Cu
4
can photocatalyze more production of •OH, h , H
2
O
2
3
V
2
O
8
. These
4
3
V
2
O
8
. Under
–
irradiation of light, BiVO
4
and Cu
3
V
2
O
8
are excited simultaneously and the electrons (e ) in the VB
+
can be excited to the CB with the generation of the same amount of holes (h ) on VB in both BiVO
4
5
+
and Cu
3
V
2
O
8
. The normal photoexcitation in Cu
V
3 2
O is from the hybridization state of V 3d and
8
2
+
2+
2–
3+
Cu 3d orbitals to the hybridization state of Cu 3d and O 2p orbitals. However, Bi 6s states are
2
–
5+
3+
predominant components of the bottom of the CB and hybridize with O 2p and V 3d, and Bi 3p
5
+
and V 3d orbitals at the bottom of the VB for the BiVO
big, which contributes in reducing of recombination rates of excited charges with achieving the
better separation of the photogenerated electron-hole pairs for the BiVO . The separation capacity of
charge carriers is well consistent with its remarkably low PL intensity (Fig. 9). In addition, the CB
level of BiVO (0.37 eV/NHE) and Cu (−0.18 eV/NHE) is more positive than the redox
potential of O (−0.33 eV/NHE) [63], the electrons located on the CB of BiVO
/•O −
cannot reduce O
to yield O −
to product H via a two-electron process [O
and Cu , which can be attributed to the fact that the CB levels of BiVO
negative compared to the redox potential of O /H (0.68 eV/NHE) and H
24]. In contrast, for BiVO with a strong covalent feature of the Bi−O and V−O bonds, the more
electrons can be transfer by Bi→O or (V→O) rather than being shared, and the formed intensified
4
. The values of s−p matrix elements are
4
4
3
V
2
O
8
2
2
4
and BiVO
4
2
4
−
−). The reduction of O
2
2
via single-electron transfer routes (O
2
+e = •O
2
−
+
2
O
2
2
+2e +2H =H
2
O
2
(aq)], however, can occur on BiVO
and Cu
/•OH (0.71 eV/NHE)
3
V
2
O
8
4
3
V
2
O are more
8
2
O
2 2
2
O
2
[
4
1
8

local electric field at the BiVO surface can promote the transfer of trapped hot electrons to adsorbed
4
−
O
2
and Cl , therefore benefiting more production of •OH and Cl•. Furthermore, based on the
+
scavenger reactivity in trapping experiments (Table 3), one can conclude that h and •OH are the
major active species in the photocatalytic oxidation of cyclohexane by using BiVO catalysts. At the
same time, the more photo-induced holes can also oxidize H O to form •OH radicals, because the
VB level (+2.71 eV/NHE) of BiVO is also sufficiently positive than the •OH/H
4
2
4
2
O potential (+2.27
−
eV/NHE) [64]. In addition, these photo-induced holes can also oxidize cyclohexane and Cl to form
cyclohexyl and Cl• radicals, which are expected to initiate the same chain process by abstraction of
hydrogen from cyclohexane. The addition of DMBD as scavengers into the reaction solution
provides evidence that Cl• radicals are involved in the photocatalytic oxidation process (Table, entry
4
). Therefore, these results imply that the •OH, Cl• and cyclohexyl radicals are the mainly active
species for the selective oxidation of cyclohexane to KA-oil. In contrast, based on the obtained
active species trapping experiments, the photocatalytic reaction might involve more electrons as
reactive species for the production of the above radicals through the oxidation pathway by using
Cu
3
V
2
O
8
catalyst.
According to the experimental and theoretical investigation, some promising photooxidation
processes and mechanisms are proposed for the different photocatalytic activities, as shown in Fig.
0 and Fig. S3. Visible light promotes photoexcitation of metal vanadate to produce the VB holes
1
+
−
(
h ) and CB electrons (e ) (Eq. (1)). Then the CB electrons are captured by water molecule in the
acidic solution to product H
hydroxyl radicals (•OH) by electron trapped in the CB (only for BiVO
the formed holes can react with H
2
O
2
via a two-electron process, and H
2
O
2
can be further reduced to
photocatalyst). In addition,
4
2
O molecules to produce •OH radicals; these •OH radicals further
react with water to yield hydroxyl radical-water complexes (Eq. (2)). Meanwhile, cyclohexane is
oxidized with the VB holes to cyclohexyl radical. This related to the large positive redox potential of
VB holes and OH• radicals as opposed to the mild driving force for the chain initiation process with
the production of cyclohexyl radical. Molecular oxygen cannot reduce to yield O
−, which is not
2
effective in directly reacting with the C−H bond of cyclohexane, but can be captured by cyclohexyl
radical to form a free cyclohexyl-peroxy radical (Eq. (4)). Subsequently, cyclohexyl peroxy radical
couples by two cyclohexyl peroxy radicals to product KA-oil and O , or is reduced to cyclohexanone
2
by CB electrons; additionally, this radical can also react with vicinal cyclohexane to form another
cyclohexyl radical for the next cycle reactions and cyclohexyl hydroperoxide. The latter (CyOOH)
1
9

mostly rearranges to cyclohexanone and water under photocatalytic conditions (Eq. (5)). In addition,
another minor pathway via •OH radicals coupling with cyclohexyl radicals to product a
cyclohexanol may also exist and all formed cyclohexanol may be subsequently oxidized to
cyclohexanone by the VB holes or •OH radicals (Eq. (6)), resulting a high cyclohexanone/
cyclohexanol molar ratio. It is confirmed that a small quantity of ionic chlorine exists in the acidic
solution, showing a synergistic factor in accelerating the photocatalytic reaction, since the same
experiment performed with addition of other Brønsted acids (such as HBr and H
2
SO
4
) and CCl
4
cannot obtain similar effects under photocatalytic conditions. This is possible for chlorine radicals
from hydrochloric acid to be formed by the VB holes under visible light irradiation in that Cl•
−
radicals are less energetic than OH• radicals (Eq. (3)) [65]. In addition, Cl may be oxidized with the
VB holes to yield •OH. Thus the Cl• radicals can effectively attack the C−H bond of cyclohexane by
a chain transfer oxidation that is faster than the corresponding reaction with OH• radicals (Eq. (4)).
However, it is also necessary for there to be present enough holes or OH• radicals for the oxidative
cleavage of C−Cl band of chlorocyclohexane intermediate to regenerate newly Cl• radicals for the
next cycle reactions (Eq. (7)), resulting in good stability. As a result, BiVO
4
processes a higher
separation capacity of charge carriers than Cu , which leads to more production of
3
V
2
O
8
photogenerated holes and OH• radicals under the same photo-radiation, resulting in a higher
cyclohexane conversion and KA-oil selectivity simultaneously. This is due to the evidence that holes
and HO• have more oxidizing power to convert to the complete oxidizing product cyclohexanone.
This is not surprising, as high selectivity for chlorocyclohexane is usually observed in Cu
3
V O
2 8
photocatalysts, which is further confirmed by the fact that the various reactive species after adding
−
the scavengers are quite different under our experimental conditions. Interesting, no •O
2
is formed
−
via the reduction of O
2
by the photo-formed CB e on BiVO
4
and Cu
3
V
2
8
O , which is the key species
for the cyclohexanone decomposition to CO
2
via the cleavage of C–C bond. Thus, the formed
cyclohexanone is not subsequently decomposed to CO
production of KA-oil.
2
by photocatalysis, leading to selective
4
. Conclusion
In summary, metal vanadates have been proven to be a novel series of visible-light-driven
photocatalysts for the efficient photooxidation of cyclohexane to KA-oil for the first time, along with
the intermediate is obtained as the total oxidation product. Excellent partial oxidation KA-oil
-1
selectivity (95%) and efficiency (TOF=7.5 h ) are achieved with 15.5% conversion using a BiVO
4
2
0

in oxygen under ambient conditions (room temperature and 1atm pressure). In more detail, BiVO
4
3
+
2+
and Cu
3
V
2
O vanadates, as tunable optical properties by incorporating metal Bi or Cu ions, are
8
compared systematically for their abilities to generate reactive species and charge carriers toward
selective oxidation pathways during photoexcitation. The formation of reactive species is found to
highly depend on the crystal structure and surface electronic state of BiVO
4
and Cu
3
V
2
O
8
photocatalysts by experimental and theoretical studies. The construction of VBM and CBM
3
+
positions by hybridization of Bi 6s and 6p orbitals facilitates the direct trapping and mobility of hot
electrons from the VBM to CBM for BiVO , promoting the generation of hydroxyl radical and H
4
2
O
2
reactive species toward selective formation of KA oils. These atomic-level findings will provide
valuable insights for understanding the photocatalytic mechanisms of the heteroatom-tuning
vanadates for associated reactive species with their different trends in photocatalytic activation of
aliphatic C−H bond.
Acknowledgments
This work was supported by the Natural Science Foundation of China (Grant No.21576078 and
2
1878074) and the Natural Science Foundation of Hunan Province (Grant No.2016JJ2081) and
Innovation Platform Open Fund of Hunan College (16K052).
References
[
1] A. Henríquez, H.D. Mansilla, A.M. Martínez-de la Cruz, J. Freer, D. Contreras, Appl. Catal. B:
Environ. 206 (2017) 252–262.
2] G. de Ruiter, K.M. Carsch, S. Gul, R. Chatterjee, N.B. Thompson, M.K. Takase, J. Yano, T.
Agapie, Angew. Chem. Int. Ed. 56 (2017) 4772–4776.
3] U. Schuchardt, D. Cardoso, R. Sercheli, R. Pereira, R.S. da Cruz, M.C. Guerreiro, D. Mandelli,
E.V. Spinacé, E.L. Pires, Appl. Catal. A Gen. 211 (2001) 1–17.
[
[
[
[
[
4] A.A. Alshehri, A.M. Alhanash, M. Eissa, M.S. Hamdy, Appl. Catal. A Gen. 554 (2018) 71–79.
5] L. Suhadolnik, A. Pohar, B. Likozar, M. Čeh, Chem. Eng. J. 303 (2016) 292–301.
6] A.A. Memon, A.A. Arbab, S.A. Patil, N. Mengal, K.C. Sun, I.A. Sahito, S.H. Jeong, H.S. Kim,
Appl. Catal. A Gen. 566 (2018) 87–95.
2
1

[
[
[
[
(
[
7] J. Kou, C. Lu, J. Wang, Y. Chen, Z. Xu, R.S. Varma, Chem. Rev. 117 (2017) 1445−1514.
8] V. D.B.C. Dasireddy, B. Likozar, J. Taiwan Inst. Chem. E. 82 (2018) 331–341.
9] A. Maldotti, A. Molinari, R. Amadelli, Chem. Rev. 102 (2002) 3811–3836.
10] K. Takaki, J. Yamamoto, K. Komeyama, T. Kawabata, K. Takehira, Bull. Chem. Soc. Jpn. 77
2004) 2251–2255.
11] A. Maldotti, A. Molinari, P. Bergamini, R. Amadelli, P. Battioni, D. Mansuy, J. Mol. Catal. A
1
13 (1996) 147–157.
[
[
(
[
[
12] H. Sun, F. Blatter, H. Frei, J. Am. Chem. Soc. 118 (1996) 6873–6879.
13] M.N. Lyulyukin, P.A. Kolinko, D.S. Selishchev, D.V. Kozlov, Appl. Catal. B: Environ. 220
2018) 386–396.
14] P. Du, J.A. Moulijn, G. Mul, J. Catal. 238 (2006) 342–352.
15] Y. Shiraishi, S. Shiota, H. Hirakawa, S. Tanaka, S. Ichikawa, T. Hirai, ACS Catal. 7 (2017)
2
93−300.
[
[
[
[
[
[
16] R. Vinu, G. Madras, Appl. Catal. A Gen. 366 (2009) 130–140.
17] A.R. Almeida, J.T. Carneiro, J.A. Moulijn, G. Mul, J. Catal. 273 (2010) 116–124.
18] Y. Ide, H. Hattori, S. Ogo, M. Sadakane, T. Sano, Green Chem. 14 (2012) 1264–1267.
19] D. Yang, T. Wu, C. Chen, W. Guo, H. Liu, B. Han, Green Chem. 19 (2017) 311–318.
20] M. Oshikiri, M. Boero, J. Ye, Z. Zou, G. Kido, J. Chem. Phys. 117 (2002) 7313–7318.
21] W. Zhong, T. Qiao, J. Dai, L. Mao, Q. Xu, G. Zou, X. Liu, D. Yin, F. Zhao, J. Catal. 330 (2015)
2
08–221.
[
(
[
22] S. Verma, R.B. Nasir Baig, M.N. Nadagouda, R.S. Varma, ACS Sustainable Chem. Eng. 4
2016) 2333–2336.
23] J. She, Z. Fu, J. Li, B. Zeng, S. Tang, W. Wu, H. Zhao, D. Yin, S.R. Kirk, Appl. Catal. B:
Environ. 182 (2016) 392–404.
24] S. Samanta, S. Khilari, D. Pradhan, R. Srivastava, ACS Sustainable Chem. Eng. 5 (2017)
562–2577.
25] A. Malathi, J. Madhavan, M. Ashokkumar, P. Arunachalam, Appl. Catal. A Gen. 555 (2018)
7–74.
[
2
[
4
[
[
26] D. Ke, T. Peng, L. Ma, P. Cai, P. Jiang, Appl. Catal. A Gen. 350 (2008) 111–117.
27] S. Gao, B. Gu, X. Jiao, Y. Sun, X. Zu, F. Yang, W. Zhu, J. Am. Chem. Soc. 139 (2017)
3
438–3445.
2
2

[
28] H.M. Fan, T.F. Jiang, H.Y. Li, D.J. Wang, L.L. Wang, J.L. Zhai, D.Q. He, P. Wang, T.F. Xie, J.
Phys. Chem. C 116 (2012) 2425–2430.
[
[
[
[
[
29] Y. Yan, Y. Yu, D. Wu, Y. Yang, Y. Cao, Nanoscale, 8 (2016) 949–958.
30] G. Kresse, D. Joubert, Phys. Rev. B 59 (1999) 1758–1775.
31] M. Yashima, R.O. Suzuki, Phys. Rev. B 79 (2009) 125201–125206.
32] X. Xu, S. Kou, X. Guo, X. Li, X. Ma, H. Mao, J. Phys. Chem. C 121 (2017) 16257–16265.
33] M. Li, Y. Gao, N. Chen, X. Meng, C. Wang, Y. Zhang, D. Zhang, Y. Wei, F. Du, G. Chen,
Chem. Eur. J. 22 (2016) 1–9.
[
[
[
34] C. Yin, S. Zhu, Z. Chen, W. Zhang, J. Gu, D. Zhang, J. Mater. Chem. A 1 (2013) 8367–8378.
35] J. Yu, A. Kudo, Adv. Funct. Mater. 16 (2006) 2163−2169.
36] S.M. Thalluri, C.M. Suarez, M. Hussain, S. Hernandez, A. Virga, G. Saracco, N. Russo, Ind.
Eng. Chem. Res. 52 (2013) 17414−17418.
[
(
[
(
[
[
37] T. Kawada, A. Ikematsu, T. Tajiri, S. Takeshima, M. Machida, Int. J. Hydrogen Energ. 40
2015) 10726–10733.
38] K. Bhattacharyya, S. Varma, A. K. Tripathi, S.R. Bharadwaj, A.K. Tyagi, J. Phys. Chem. C 112
2008) 19102–19112.
39] J.M. Valero, S. Obregón, G. Colón, ACS Catal. 4 (2014) 3320−3329.
40] J. Su, X.-X. Zou, G.-D. Li, X. Wei, C. Yan, Y.-N. Wang, J. Zhao, L.-J. Zhou, J.-S. Chen, J.
Phys. Chem. C 115 (2011) 8064–8071.
[
[
41] M. Balamurugan, G. Yun, K.-S. Ahn, S.H. Kang, J. Phys. Chem. C 121 (2017) 7625–7634.
42] L. Zhou, Q. Yan, J. Yu, R.J.R. Jones, N. Becerra-Stasiewicz, S.K. Suram, A. Shinde, D.
Guevarra, J.B. Neaton, K.A. Persson, J.M. Gregoire, Phys. Chem. Chem. Phys. 18 (2016)
349–9352.
9
[
[
[
43] P.C.J. Graat, M.A.J. Somers, Appl. Surf. Sci. 100 (1996) 36–40.
44] L. Zhang, J. Zhou, C. Zhang, J. Mater. Chem. A 2 (2014) 14903–14907.
45] Q. Li, H. Meng, P. Zhou, Y.Q. Zheng, J. Wang, J.G. Yu, J.R. Gong, ACS Catal. 3 (2013)
8
82–889.
[
[
[
[
46] M. Butler, D. Ginley, J. Electrochem. Soc. 125 (1978) 228–232.
47] R.G. Pearson, Inorg. Chem. 27 (1988) 734–740.
48] Y.I. Kim, S.J. Atherton, E.S. Brigham, T.E. Mallouk, J. Phys. Chem. 97 (1993) 11802–11810.
49] X.-L. Wei, H. Zhang, G.-C. Guo, X.-B. Li, W.-M. Lau, L.-M. Liu, J. Mater. Chem. A 2 (2014)
2
3

2
101–2109.
[
[
[
50] P.J. Coing-Boyat, Acta Cryst. B38 (1982) 1546–1548.
51] Z. Zhao, Z. Li, Z. Zou, Phys. Chem. Chem. Phys. 13 (2011) 4746–4753.
52] L. Zhou, Q. Yan, A. Shinde, D. Guevarra, P.F. Newhouse, N. Becerra-Stasiewicz, S.M.
Chatman, J.A. Haber, J.B. Neaton, J.M. Gregoire, Adv. Energy Mater. 5 (2015) 1500968-1500980.
[
[
53] R.W. Godby, M. Schl¨uter, L.J. Sham, Phys. Rev. Lett. 56 (1986) 2415–2418.
54] S. Murugesan, M.N. Huda, Y. Yan, M.M. Al-Jassim, V. (Ravi) Subramanian, J. Phys. Chem. C
1
14 (2010) 10598–10605.
[
[
55] A. Kudo, Y. Miseki, Chem. Soc. Rev. 38 (2009) 253–278.
56] Z. Chen, P. Sun, B. Fan, Q. Liu, Z. Zhang, X. Fang, Appl. Catal. B: Environ. 170–171 (2015)
1
0–16.
57] J.T. Wang, B.S. Zhang, Y.M. Wang, Q.M. Hu, Organic Chemistry, 2nd ed., Nankai University
Press, Tianjing, 2002 (Chapter 2).
58] G. Li, K.H. Wong, X. Zhang, C. Hu, J.C. Yu, R.C. Chan, P.K. Wong, Chemosphere, 76 (2009)
185–1191.
[
[
1
[
[
(
[
[
59] J. Cao, B. Xu, H. Lin, B. Luo, S. Chen, Chem. Eng. J. 185–186 (2012) 91–99.
60] U.A. Joshi, J.R. Darwent, H.H.P. Yiu, M.J. Rosseinsky, J. Chem. Technol. Biotechnol. 86
2011) 1018−1023.
61] R. Georgekutty, M.K. Seery, S.C. Pillai, J. Phys. Chem. C 112 (2008) 13563–13570.
62] R. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 32 (1976)
7
51–767.
[
[
[
63] J. Kim, C.W. Lee, W. Choi, Environ. Sci. Technol. 44 (2010) 6849–6854.
64] H.F. Cheng, B.B. Huang, Y. Dai, X.Y. Qin, X.Y. Zhang, Langmuir 26 (2010) 6618–6624.
65] M. Lewandowski, D.F. Ollis, J. Catal. 217 (2003) 38–46.
2
4

Table 1. Calculated lattice parameters and previously reported experimental structure data for the BiVO and
4
3
Cu V
2
O
8
samples.
Theoretical calculation values
BiVO Cu
Experimental values
Crystal structure
d
e
4
3
V
2
O
8
BiVO
4
3
Cu V
2
O
8
a/ Å
b/ Å
c/ Å
α
5.194
5.090
11.697
90.00
90.00
90.39
1.740
2.459
5.161
5.431
6.534
69.07
88.36
68.21
1.980
1.940
5.197
5.096
11.702
90.00
90.00
90.40
1.745
2.441
5.196
5.355
6.505
69.22
88.69
68.08
-
Lattice parameters
β
γ
a
D
D
V–O
M–O
/Å
/Å
b
-
a
b
The D V–O represents the average bond length of V–O; the D M–O represents the average bond length of Bi–O and
c
d
e
Cu–O for BiVO
4
and Cu
3
V
2
O , respectively; the D O–O represents the average bond length of O–O; [36]; [37].
8
2
5

Table 2 Visible light-driven oxidation of cyclohexane with molecular oxygen catalyzed by the different catalysts a.
Selectivity (%)c
Conversion
-
1 d
Entry
Catalyst
TOF (h )
(
%)b
KA-oil
1
2
Blank
0
-
-
-
-
-
V
2
O
5
3.8
0.1
74.9
100
48.6
95.0
54.2
100
100
100
96.5
75.2
80.1
73.7
85.1
93.1
87.9
25.1
0
20.6
49.6
37.0
24.0
23.1
27.4
36.5
-
54.3
50.4
11.6
71.0
31.1
72.6
63.5
100
0.5
-
3
Bi
2
O
3
4
CuO
0.6
51.4
4.9
45.8
-
-
5
BiVO
4
15.5
9.1
7.5
2.8
0.2
0.03
0.13
1.6
3.3
2.3
3.5
3.6
2.7
3.3
6
Cu
BiVO
Cu
BiVO
TiVO
SrVO
3
V
2
O
8
8
7e
8e
9f
4
0.3
3
V
2
O
0.1
-
4
0.2
-
1
0
1
4
6.4
3.5
24.8
16.9
26.3
14.9
6.9
12.1
77.6
18.4
19.9
23.3
23.5
22.8
25.1
18.9
56.8
63.2
50.4
61.6
70.3
62.9
1
4
9.0
1
1
1
1
1
2
3
4
5
6
CrVO
4
4
9.1
LaVO
9.1
CoVO
4
12.0
12.7
13.0
AlVO
FeVO
4
4
a
All reactions were done with 0.0050 g of catalyst, 4 mmol cyclohexane, O
C, time (8 h), concentrated HCl (0.1 ml) as an additive, tungsten-bromine lamp (20W) as visible light source.
2
(1 atm), 5 mL of acetonitrile, at 28-30
◦
b
Conversion (%) based on substrate = {1−[(concentration of substrate left after reaction) × (initial concentration
c
of substrate)−1]} × 100; Product selectivity = content of this product/(adding cyclohexane amount (mmol)-the
amount of cyclohexane recovered (mmol)) × 100%.
d
Turnover frequency (TOF): number of moles of cyclohexane converted per the molar amount of V sites and hour
of reaction time.
e
Without adding concentrated HCl as an additive in photo-oxidation system.
f
Using concentrated HBr as an acidic additive instead of concentrated HCl.
2
6

Table 3 Reactive species trapping experiments of BiVO
BiVO
4
and Cu
3
V
2
O
8
samples.
4
3
Cu V
2
O
8
Sele. (%)
Sele. (%)
Entry
Scavenger
Conv.
%)
Conv.
(%)
(
1
2
3
4
5
TEOA
IPA
3.0
5.8
12.1
0
0
41.6
16.9
32.6
-
58.1
79.7
62.8
-
8.6
7.7
9.0
0.5
6.1
35.8
23.1
19.7
10.8
43.5
17.8
41.1
59.3
53.7
6.0
3.4
4.6
-
21.1
35.5
50.5
35.7
BQ
DMBD
AgNO
3
13.6
5.0
19.5
75.5
46.5
2
7

Figure captions:
Fig. 1. XRD patterns of BiVO
4 3
, Cu V
2
O
8 2
, Bi O
3
, V
2
O and CuO samples.
5
Fig. 2. TEM images of BiVO
4
(A
0
, A
1
and A
2 3 2
) and Cu V O
8
samples (B
0
, B
1
and B ).
2
Fig. 3. Visible Raman spectra of (A) for BiVO
4
and Cu
3
V
2
O
8
samples and (B) for Bi
2
O
3
, V
2
5
O and CuO samples.
Fig. 4. XPS spectra of (A) survey, (B) Bi 4f, (C) Cu 2p, (D) V 2p, (E) O 1s and (F) valence band of BiVO
Cu samples
4
and
3
V
2
O
8
Fig. 5. UV-Vis spectra of (A) for BiVO
4 3
, Cu V
2
O
8 2
, Bi O
3
, V
2
O
5
and CuO samples; Tauc plots of BiVO (B) for
4
direct band gap and Cu for indirect band gap.
3
V
O
2 8
Fig. 6. Crystal structures of BiVO
4
(A
0
3
) and Cu V
2
O
8
(B
0
) samples; band structures of BiVO
4
(A) and Cu
3
V
2
O (B)
8
samples.
Fig. 7. Total density of state of BiVO
Cu (B) samples.
4
(A
0
) and Cu
3
V
O
2 8
(B
0
) samples; partial density of state of BiVO (A) and
4
3
V
2
O
8
Fig. 8. Effect of the irradiation time on the product evolution for BiVO
4
(A) and Cu
3
V
2
8
O (B) samples. The error
bars represent the standard deviations of three independent measurements of the same sample.
Fig. 9. Photoluminescence spectra with the excitation wavelength of 290 nm.
Fig. 10. Photooxidation processes of cyclohexane with O
2
vapor over the BiVO
4
and Cu
3
V
2
O samples.
8
2
8

Fig. 1.
2
9