Octahedral-based redox molecular sieve M-PKU-1: Isomorphous metal-substitution, catalytic oxidation of sec-alcohol and related catalytic mechanism

Article abstract of DOI:10.1016/j.jcat.2017.05.007

Octahedral-based redox molecular sieves M-PKU-1 (M[dbnd]Cr, Fe) were synthesized by isomorphous metal-substitution and used as catalysts for catalytic dehydrogenation of sec-alcohols using H₂O₂ as oxidant. Various characterizations, such as X-ray diffraction Rietveld refinement and XPS spectrum, confirmed that transition metals were embedded successfully inside the PKU-1 framework with a high level (about ～50 atom %) and presented in the valence state of +3. Molecular probe analyzes suggested that Cr sites catalyzed the quick formation of active OH radicals, however strongly suppressed the generation of superoxide O₂^- ions. Under the performed reaction conditions, 10%Cr-PKU-1 exhibited excellent catalytic performances (>99?% selectivity) and kept favorable recyclable stability. A hypothetical mechanism was proposed, which involved a Cr³⁺-Cr²⁺-Cr³⁺ circle when the oxidative dehydrogenation reaction happened. Furthermore, qualitative and quantitative analyzes were performed to illustrate the stepwise by-products generated in the probable pathway due to the over-oxidization, but the selectivity to the two proposed pathways seemed to be in equal portions and didn't have any obvious preference. Obviously, our preliminary results still merit further exploration; we believe however, they would provide helpful information to better understand the structure-activity relationship and the key function of Cr-PKU-1 in the catalytic activation of H₂O₂.

Full text of DOI:10.1016/j.jcat.2017.05.007

Journal of Catalysis 352 (2017) 130–141
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Octahedral-based redox molecular sieve M-PKU-1: Isomorphous
metal-substitution, catalytic oxidation of sec-alcohol and related
catalytic mechanism
⇑
⇑
Weilu Wang, Shixiang Hu, Lingjie Li, Wenliang Gao , Rihong Cong, Tao Yang
College of Chemistry and Chemical Engineering, Chongqing University, Chongqing 400044, People’s Republic of China
a r t i c l e i n f o
a b s t r a c t
Article history:
Octahedral-based redox molecular sieves M-PKU-1 (M@Cr, Fe) were synthesized by isomorphous metal-
substitution and used as catalysts for catalytic dehydrogenation of sec-alcohols using H as oxidant.
Various characterizations, such as X-ray diffraction Rietveld refinement and XPS spectrum, confirmed
that transition metals were embedded successfully inside the PKU-1 framework with a high level (about
Received 1 March 2017
Revised 9 May 2017
Accepted 14 May 2017
2 2
O
ꢀ
50 atom %) and presented in the valence state of +3. Molecular probe analyzes suggested that Cr sites
Å
catalyzed the quick formation of active OH radicals, however strongly suppressed the generation of
Keywords:
Å
À
superoxide O ions. Under the performed reaction conditions, 10%Cr-PKU-1 exhibited excellent catalytic
2
Aluminoborates
performances (>99 % selectivity) and kept favorable recyclable stability. A hypothetical mechanism was
Octahedral-based framework
Redox molecular sieve
Sec-alcohol oxidative dehydrogenation
H O activation
2 2
Mechanism
proposed, which involved a Cr3 -Cr -Cr circle when the oxidative dehydrogenation reaction happened.
Furthermore, qualitative and quantitative analyzes were performed to illustrate the stepwise by-
products generated in the probable pathway due to the over-oxidization, but the selectivity to the two
proposed pathways seemed to be in equal portions and didn’t have any obvious preference. Obviously,
our preliminary results still merit further exploration; we believe however, they would provide helpful
information to better understand the structure-activity relationship and the key function of Cr-PKU-1
+
2+
3+
Over-oxidation
2 2
in the catalytic activation of H O .
Ó 2017 Elsevier Inc. All rights reserved.
1
. Introduction
Liquid-phase catalytic oxidation is widely used in the manufac-
in 1983 and has been extensively investigated over the past few
decades [5]. This microporous material with redox sites was
proved to efficiently catalyze a wide variety of oxidative transfor-
ture industry of both bulk and fine chemicals, and the traditional
process usually involves stoichiometric inorganic oxidants, thus
being unfavorable due to the environmental protection [1].
Employing homogeneous catalysts, i.e. soluble metal salts or com-
plex molecules, are conventional and usually efficient, however
suffer from the disadvantages in difficult recovery, low recycling
ability and easy deactivation [2]. On the other hand, heterogeneous
catalytic systems using redox molecular sieves are easy to handle
these issues, by incorporating metal ions into the framework of
zeolites or zeolite-like materials [3,4]. The local confinement of
the redox active sites in their channels/cavities can provide a
microenvironment analogy with enzymes and therefore poten-
tially endow the catalyst with unique activity and selectivity.
A typical example is TS-1 (a titanium-substituted silica-based
molecular sieve), which was first synthesized by Taramaso et al.
2 2 2 2
mations with H O , RO H, or O under relatively mild conditions
[6–11]. Nevertheless, the titanium concentration in TS-1 is quite
low (the maximum atomic ratio of Si/Ti is about 2.5 %), because
of the incompatible ionic radii and different coordination numbers
4
+
4+
between Ti and Si . In fact, transition metals such as titanium
prefer octahedral coordination, and along this line, octahedral-
based molecular sieves (OMS) are an alternative and promising
system to accommodate transition metal ions as redox sites. Till
now, Todorokite-type OMS is the typical and successful case, in
which the rock salt blocks of manganese oxides are interlinked
forming one-dimensional tunnels via rutile-type connection
[12,13]. Indeed, Todorokite-type catalysts have already exhibited
good performances in some oxidation and hydration reactions
[14–18].
On the other hand, there is another interesting system of
octahedral-based molecular sieves (denoted as PKU-n), which
consist of cost-effective elements such as aluminum and
boron [19–24]. Among them, PKU-1 was the most extensively
⇑
Corresponding authors.
E-mail addresses: gaowl@cqu.edu.cn (W. Gao), taoyang@cqu.edu.cn (T. Yang).
http://dx.doi.org/10.1016/j.jcat.2017.05.007
021-9517/Ó 2017 Elsevier Inc. All rights reserved.
0

W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
131
investigated both in structure and in catalytic properties, [25–29]
beneficial from its easy preparation, high thermal stability, sub-
organic molecules through multistep oxidative reactions. We also
performed qualitative and quantitative analyzes on over-oxidized
products in order to better understand the function of Cr-PKU-1
3+
stantial specific acidic sites, and most importantly, Al in PKU-1
3
+
3+
3+
can be replaced by Cr , Fe or Ga in a very high level [19]. Single
crystal X-ray diffraction revealed that PKU-1 is consisted of AlO
octahedra by edge-sharing exclusively, affording 18-ring channels
along the c-axis. Borates groups, in the forms of BO(OH) , BO (OH)
and B (OH), attach to AlO -framework by sharing vertex oxygen
atoms. The BO(OH) and BO (OH) groups reside within the 18-ring
channels and narrow these channels along the [001] direction
Fig. 1a), while B (OH) groups attach to the walls of 10-ring
channels and almost block the channels along the [100] direction
in H
OH (from H
2
O
2
activation, the results indicate a gentle generation rate of
Å
6
2
O
2
) would lead to the high selectivity of ketone, but
Å
a higher OH concentration could lead to diverse by-products such
as adipic acid or cyclic lactone. In general, this study is a prelimi-
2
2
2
O
4
6
2 2
nary study on the controllable and selective activation of H O cat-
2
2
alyzed by Cr-PKU-1, and considering the significant capacity to
accommodate desired metals, M-PKU-1 catalysts really holds high
potentials, for example, on catalyzing other redox reactions, prob-
ably with additional loading of cocatalysts.
(
2 4
O
(
Fig. 1b).
In this study, we synthesized metal-incorporated PKU-1
denoted as M-PKU-1, M@Cr or Fe) and Cr-PKU-1 was proved to
(
2
. Experimental section
be quite efficient to catalyze the oxidation of sec-alcohols into
the corresponding ketone compounds with a high selectivity
2.1. Syntheses of Cr-PKU-1, Fe-PKU-1 or Fe-Cr-PKU-1 catalyst
(
2 2
99 % using H O as the oxidant), which is considered as one of
the important transformations because of the wide utility in
pharmaceutical syntheses and fine chemical engineering [30–32].
Obviously, homogeneous catalysis on this reaction hold several
drawbacks in reactor corrosion and catalyst recovery et al.
The syntheses of M-PKU-1 were carried out in a closed Teflon
autoclave by the boric acid flux method as described in the litera-
ture [19]. Typically, Cr(NO O (0.5 mmol), Al(NO Á9H
Á9H
4.5 mmol) and concentrated HNO (0.3 mL) were first loaded to
a 25 mL Teflon container and slowly heated until a homogeneous
solution was formed. Then, large amount of boric acid
100 mmol), functioned as both reaction medium and reactant,
3
)
3
2
3
)
3
2
O
(
3
[
33–36] therefore a green oxidation system using recyclable
heterogeneous catalysts is desired [37–40].
a
The systematic survey on the catalysis conditions of M-PKU-1
suggests that 10%Cr-PKU-1 is the most efficient catalyst for a vari-
ety of sec-alcohols. The transformation from cyclohexanol to cyclo-
hexanone (abbreviated as CHOL and CYC for clarity, hereinafter)
was selected as the model reaction to understand the catalytic
mechanism. The results indicate the presence of Cr3 in the frame-
(
was charged into the autoclave. Next, the autoclave was sealed
and heated at 513 K for 120 h. After the reaction, the autoclave
was naturally cooled to room temperature and the resultant solid
was washed with warm water (343 K) in order to completely
remove residual boric acid or other soluble components. The
as-synthesized Cr-PKU-1 powder is greenish.
+
work of PKU-1 promotes the activation of H
2
O
2
into ÁOH radicals
Å
À
exclusively (prohibiting the formation of O
2
ions in the mean-
In the literature, the molecular formula of PKU-1 was identified
Å
time). An appropriate concentration of OH radical is the key to
the high selectivity (99 %) to CYC under mild conditions, and the
oxidation process was speculated accordingly, during which a
to be HAl
with 10 atom % substitution, the exact formula is HCr0.3Al2.7
OH) . For simplicity, these as-synthesized Cr-PKU-1 samples were
3 6 4
B O12(OH) [19]. Therefore, as for Cr-PKU-1, for example
6 12
B O
(
4
3
+
2+
3+
Cr -Cr -Cr cycle was involved.
denoted sequentially as 0%, 6%, 10%, 20%, 30%, 40% or 50%Cr-PKU-1,
according to the increase x.
As proved in the following, a low-level Cr in PKU-1 tends to pro-
vide lower concentration of OH radical and thus results into a long
The syntheses of Fe-PKU-1 and Fe-Cr-PKU-1 samples were sim-
3
+
ilar with those of Cr-PKU-1. Fe(NO)
source. According to the initial loading amount of transition
metals, the as-obtained Fe-PKU-1 samples were marked as 0%,
3
Á9H
2
O was used as the Fe3+
reduction period, while the over-high concentration of Cr in PKU-
Å
1
(i.e. 30 atom %) generated excessive OH radicals and side-
reactions/over-oxidation were therefore unavoidable. For example,
CHOL was degraded successively into some low-molecular-weight
5
%, 10%, 20% or 30%Fe-PKU-1, while bimetal-substituted PKU-1
Fig. 1. Octahedral-based framework of PKU-1 stabilized with three types of borate hydroxyl groups along 18-ring (a) and 10-ring (b) windows; red octahedron, AlO
octahedron, CrO ; cyan, yellow and grey spheres represent O, B and H atoms, respectively.
6
; green
6

1
32
W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
samples were denoted as 0%Fe-5%Cr, 5%Fe-5%Cr, 10%Fe-5%Cr or
rate of 80 mL/min), the samples were heated from room tempera-
ture to 378 K, keeping at 378 K for 10 min to remove the moisture
content, and then further heated up to 1173 K. Besides, all samples
were kept at 10 ±1 mg for avoiding heat transfer limitations.
2
0%Fe-5%Cr-PKU-1, respectively.
2.2. Catalyst characterizations
2
.2.7 Photoluminescence (PL) spectroscopy was used to detect
2
.2.1 Powder X-ray diffraction data were collected at room tem-
perature on a PANalytical X’pert powder diffractometer equipped
with a PIXcel 1D detector under Cu K radiation (k = 1.5406 Å).
OH radicals formed over Cr-PKU-1 catalyst using terephthalic acid
(TA) as a probe molecule. In a typical run, 40 mg of 10%Cr-PKU-1
was dispersed in a mixed aqueous solution containing 3 mmol
a
The operating voltage and current were 40 kV and 40 mA, respec-
tively. Le Bail fitting was performed using the TOPAS software
package [41]. In order to in-situ observe the phase transformation
under higher temperature, a high-temperature powder XRD exper-
iments were carried out on a Bruker D8 Advance diffractometer
3 2 2
TA, 20 mmol CHOL, 20 mL CH CN, and 40 mmol H O . After stirring
for a specified interval of time, the suspension was withdrawn and
then centrifuged to remove the powder. The upper and transparent
solution was used for the PL measurement with an excitation irra-
diation at 328 nm on a Hitachi F4600 fluorescence spectrometer.
PMT voltage was fixed to be 700 V, and the width of excitation
and emission slit were both set to 2.5 nm.
(
Cu Ka radiation, k = 1.5406 Å) with a computer-controlled fur-
nace. The sample was loaded on a platinum strip and heated from
room temperature to 1300 K in air at a heating rate of 5 K/min, and
the XRD data were collected after the temperature was stabilized
for 1 h at each chosen temperature.
2.2.8 UV–Vis absorption spectra were used to detect superoxide
Å
À
2
ion ( O ) formed over Cr-PKU-1 catalyst using nitroblue tetra-
zolium (NBT) as a probe molecule. In a typical run, 40 mg of 10%
Cr-PKU-1 was dispersed in a mixed aqueous solution containing
3 mmol NBT, 20 mmol CHOL, 20 mL CH CN, and 40 mmol H O .
3 2 2
After stirring for a specified interval of time, the suspension was
withdrawn and then centrifuged to remove the catalyst. The upper
solution was diluted for 20 times with DMSO and then used for
UV–Vis absorption measurement in a UV–Vis spectrometer (PUXI
TU-1810, China).
2
.2.2 X-ray photoelectron spectra (XPS) were acquired with UK
Kratos Axis Ultra spectrometer with Al K X-ray source operated at
5 kV and 15 mA. Kinetic energies of photoelectrons were mea-
a
1
sured using a hemispherical electron analyzer working at the con-
stant pass energy (40 eV). The pressure in the chamber was less
À9
than 5.0 Â 10 Torr. The XPS data were collected in increments
of 0.1 eV with dwell time of 500 ms. Electron binding energies
b
were calibrated against the C 1 s emission at E = 284.6 eV to cor-
rect the contact potential differences between the sample and
the spectrometer.
2.3. Catalytic dehydrogenation
2
.2.3 Cyclic voltammetry (CV) was performed to identify the
All organic solvents were purified by distillation method, other
substrates such as sec-alcohol were used as received. Catalytic
dehydrogenation of sec-alcohols was performed in a 50 mL
single-neck round-bottom flask immersed into a silicon oil bath
and stirred with a magnetic stirrer. In a standard run, 20 mmol
redox behavior of Cr-PKU-1. To prepare the testing electrode, a vis-
cous slurry containing 74 wt % Cr-PKU-1, 18.5 wt % acetylene
black, and 7.5 wt % PTFE was mixed and pressed onto a nickel foam
current collector. After that, the electrode was dried in vacuum and
maintained at 333 K for 24 h in prior to measurements. Electro-
chemical tests were examined on an electrochemical workstation
3
CHOL, 20 mL CH CN, 40 mg 10%Cr-PKU-1 catalyst and 40 mmol
2 2
H O
(30 wt %) were mixed in the flask. After heated at 353 K for
(
Autolab Pgstat302N, Switzerland). The working solution was pre-
pared by mixing CH CN (80 mL), dimethyl sulphoxide (DMSO)
0.5 mL) and 0.1 M KCl solution (10 mL). In a three-electrode cell,
8 h, reaction mixture was extracted with a syringe to transfer to
a centrifuge tube, and subsequently centrifuged to remove the
solid catalyst. Finally, the obtained transparent solution was ana-
lyzed qualitatively or quantitatively with chromatography-mass
spectrometry (GC–MS), gas chromatography (GC) and high perfor-
mance liquid chromatography (HPLC), respectively.
3
(
Pt foil, Ag/AgCl and the above loaded nickel foam were used as
the counter, reference and working electrodes, respectively. Cyclic
voltammetry curves were obtained in the potential range of
À1.0 ꢀ 0 V vs. Ag/AgCl by varying the scan rate from 5 to
GC–MS was performed on an Agilent 7890N gas chromatograph
coupled with a capillary column (DB-5: length, 30 m; inner diam-
eter, 0.25 mm) and mass spectrum system (Agilent 5975C). Helium
was used as both the carrier (1 mL/min) and make-up gas (40 mL/
min). GC–MS spectra were identified by comparing the m/z value of
each peak with the NIST Mass Spectral database.
GC was used to determine the concentration of the substrates
and products on a Techcomp GC 7800 gas chromatograph
equipped with a flame ionization detector (FID). A capillary column
(CP-43: length, 25 m; inner diameter, 0.53 mm; film thickness,
À1
1
00 mVÁs
.
2
.2.4 Sample morphology was examined with a field emission
scanning electron microscopy (JEOL, JSM-7800F) equipped with
an energy dispersive spectrometer (EDS) analyzer. Samples were
prepared by dispersing dry powder on double-sided conductive
adhesive tape followed by coating with a gold film. The recorded
SEM image shown in this work was at a magnification of 1300Â.
2
.2.5 UV–Vis diffuse reflectance spectrum (UV–Vis DRS) was
collected using a Shimadzu UV-3100 spectrometer equipped with
an integrating sphere and using BaSO as a reflection standard.
4
1.0 lm) was chosen to separate the mixed components at the fol-
Spectra were recorded in the range of 190–1000 nm in 0.5 nm
steps with a scan speed of 60 nm/min and time constant of 2 s.
In the visible range, a constant slit width of 2 nm was used for
the monochromator, and the reflected radiation was detected with
a photon multiplier detector. The wet pastes were placed on the
window of the integrating sphere for testing.
lowing condition: 423 K for capillary column, 473 K for injector
port, and 523 K for FID detector.
An Agilent 1260 infinity HPLC system coupled with an auto-
sampler, quaternary pump, and ultraviolent detector was used to
determine the concentration of adipic acid and glutaric acid in
the study. The HPLC was performed on a COSMOSIL 5C18-PAQ col-
umn (length, 250 mm; inner diameter, 4.6 mm) and the column
was held at 298 K. The mobile phase was a combination of solvent
A (20 mM potassium phosphate dibasic in water adjusted pH to 3.0
with phosphoric acid) and solvent B (methanol). The gradient anal-
ysis was performed using 84 % solvent A and 16 % B solvent. The
2
.2.6 Combined thermogravimetric (TG) and differential scan-
ning calorimeter (DSC) analyzes were performed on a Mettler-
Toledo TGA/DSC1 instrument at a heating rate of 10 K/min from
room temperature to 1173 K. Its temperature precision was
±0.5 K and microbalance sensitivity was less than ±0.1 lg. Blank
experiments without samples were carried out to obtain the base-
lines to calibrate the experiments with samples at different heating
rates. Under an inner atmosphere of high purity nitrogen (a flow
injection volume was 20 lL. For all gradient segments, the elution
flow rate was 1.2 mL/min, and the detection wavelength was set at
510 nm.

W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
133
3
. Results and discussion
identify their chemical compositions and valence states for
Cr-PKU-1. The detailed discussions about these characterizations
were placed in the Figs. S6–S9, SI.
3.1. Isomorphous substitution by transition metals
Fig. 2 gives the XRD patterns of the as-synthesized Cr-PKU-1
samples and the Le Bail fitting for the pattern of 10%Cr-PKU-1, as
a representative. XRD patterns for all Cr-containing PKU-1 samples
3
3
.2. Catalytic performance
.2.1. Optimizations of reaction conditions
In this study, Cr-PKU-1 catalysts were applied in the liquid-
(
the right inset of Fig. 2) are almost the same with that of the par-
ent PKU-1, suggesting that Cr incorporations do not change their
host structures. In addition, Le Bail fitting to XRD pattern of 10%
Cr-PKU-1 gives a good convergence, and the refined cell parame-
phase oxidation of several selected sec-alcohols with 30 wt %
as oxidant (see Fig. 3). Under the given conditions, 10%Cr-
2 2
H O
PKU-1 was demonstrated to have a high selectivity for all sec-
alcohols to ketones. For example, when CHOL was used as reaction
substrate, the corresponding ketone (CYC) was obtained with a
high selectivity (>99 %) along with the highest turnover frequency
3
ters are a = 22.06 Å, c = 7.06 Å and V = 2978 Å . The refined unit cell
volumes give a linear dependence on the incorporated Cr content,
and the up-limit amount of incorporated Cr is in the high level and
reaches up to ꢀ50 atom %. The expansion of unit cell volumes coin-
cides with the fact that Cr3 ion has a larger radius (0.615 Å, coor-
dination number (CN) = 6) than Al3+ (0.535 Å, CN = 6) [42].
In one sense, Cr-PKU-1 can be regarded as a solid solution
(
TOF). The cyclic sec-alcohols were easily converted to the corre-
+
sponding carbonyl compounds comparing to acyclic ones. Maybe,
cyclic structure is beneficial to stabilize the corresponding fatty
ketones and to lower its activation energy in dehydrogenation pro-
cess. Owing to the highest TOF in CHOL-to-CYC, CHOL was chosen
as the model reaction substrate in the following investigations.
Reaction medium is a key factor for catalytic reactions both in
homogeneous and in heterogeneous systems. An appropriate sol-
vent can effectively reduce the surface Gibbs energy and minimize
side reactions. Therefore, 7 polar solvents, including acetonitrile
formed by the substitution of Al3 with Cr . However, under the
present synthetic conditions, only partial solid solutions can be
prepared, namely that purely Cr-substituted PKU-1 cannot be pre-
pared. Nevertheless, the significant changes of cell parameters def-
initely confirmed the successful incorporation of Cr3 into the PKU-
+
3+
+
1
framework (UV–vis spectra also give another evidence for the
successful Cr-incorporation, which was shown in Fig. S1 in the
Supporting information, SI). In some tetrahedral-based molecular
sieves, their unit cell parameters generally have no obvious change
due to the extra-framework insertion or low-concentration-type
incorporation in the framework, so the framework substitution
by transition metal was often doubtful and cannot be conclusively
drawn [43,44].
Besides Cr-PKU-1 catalysts, Fe-PKU-1 and Fe-Cr-PKU-1 were
also prepared, whose powder XRD patterns as well as refined cell
volumes were provided in Figs. S2–S5, SI. The XRD analyzes indi-
cate these as-prepared catalysts are successful and phase-pure
(
CH
3
CN), tetrahydrofuran (THF), dimethyl formamide (DMF), ethyl
O)
alcohol (EtOH), methyl alcohol (MeOH), acetone and water (H
2
were chosen as reaction media due to their miscibility with reac-
tants, while nonpolar solvents such as toluene, hexane and dichlor-
omethane were not the appropriate candidates because of the
formation of two immiscible phases. As shown in Fig. 4, nearly
1
00 % selectivity to CYC was achieved in all cases, however, the
reaction rates differed a lot. Experimentally, the maximum TOF
was obtained when CH CN was chosen as the solvent.
3
The influence of temperature on the reaction rate and CYC
selectivity was also investigated in the range of 298–354 K. As
shown in Fig. 5, the reaction rate was enhanced significantly with
increasing temperatures, which means this dehydrogenation reac-
tion was temperature-sensitive. In the meantime, the selectivity to
CYC almost remained in a very high level (ca. >95 %) in the whole
range of reaction temperature. The above results indicated a high
temperature is beneficial to the selective oxidation of CHOL. There-
fore, in the current study, the reaction mixtures were heated at
their reflux temperature (about 354 K) in order to achieve a better
reaction rate and CYC selectivity simultaneously.
(
with no impurity); the refined unit cell volume from Le Bail fitting
also give a linear expansion along with the increasing doped metal
cations.
The physicochemical properties of the as-synthesized M-PKU-1
catalysts were also characterized by many measurements. For
example, as for Cr-PKU-1 catalysts, SEM was employed to observe
their morphology (and homogeneity), in-situ high temperature
XRD patterns and TGA-DSC analysis were performed to monitor
their thermal behaviors, and XPS spectrum was collected to
Fig. 2. Le Bail fitting of powder XRD for 10%Cr-PKU-1, blue circles s represents the observed data and red solid line is the calculated pattern; the marks below are the
expected reflections. The inset is the XRD patterns of Cr-PKU-1 samples with different Cr contents.

1
34
W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
For an oxidation reaction, an appropriate equivalence of oxidant
will generally ensure a high yield of the product, while an exces-
sive amount of oxidant tends to induce an over-oxidization, which
means reduced selectivity to the targeted product. According to the
stoichiometric equation, the dehydrogenation of one mole of CHOL
requires one mole of H O . Herein, the influence of H O /substrate
2 2 2 2
molar ratio (varying from 0.25/1 to 4/1) was investigated as shown
in Fig. 6. Apparently, the reaction rate increased significantly with
the increase in H
to CYC remained high (>99 %) until H
2
2
O
2
/CHOL molar ratio; differently, the selectivity
/CHOL molar ratio reached
/1 and decreased thereafter. Despite of promoting turnover fre-
2 2
O
2 2
quency, the larger stoichiometry of H O will also result into a dee-
per oxidation, which will be discussed in later sections. Here, the
optimal ratio of H /CHOL (about 2/1) was higher than the stoi-
chiometry, because some of H would decompose into O which
is chemically inert to oxidize CHOL.
2 2
O
2
O
2
2
Fig. 3. The reaction rate of 6 different sec-alcohol substrates and the selectivity to
corresponding ketones over 10%Cr-PKU-1. Reaction conditions: 20 mmol sec-
alcohol substrates, 40 mg catalyst, 20 mL CH CN, 2 equivalent H O oxidant, and
3 2 2
refluxed at 354 K for 18 h. Noting: Turnover frequency (TOF) is calculated with the
division of the moles of consumed substrate by the moles of used catalyst and
reaction time.
3
.2.2. Screening of metal-incorporated PKU-1 and their recyclability
In this study, three types of catalysts, i.e. Cr-PKU-1, Fe-PKU-1
and Fe-Cr-PKU-1, were synthesized. Dehydrogenation reactions
indicated the type and amount of incorporated transition-metals
played a crucial role on the reaction rate of CHOL and CYC selectiv-
ity. Briefly, Cr-substituted PKU-1 samples were more active than
Fe-substituted ones; on the contrary, for bimetal-substituted cata-
lysts Fe-Cr-PKU-1, the incorporated iron ions are detrimental to
CHOL transformation and CYC selectivity.
First, the catalytic performance of Cr-PKU-1 samples with 5 dif-
ferent Cr contents was examined, and the reaction rate of CHOL
and CYC selectivity was presented in Fig. 7. Chromium-free PKU-
1
catalyst did not show any obvious catalytic activity, while Cr ions
were evidenced to act as active sites and played an essential role in
catalytic dehydrogenation of CHOL. With the increase of Cr/(Cr
+
Al) atom ratio from 0 to 30 atom %, the TOF was enhanced dra-
À1 À1
matically from 0.25 to 18.25 molSubÁmolCatÁh . The selectivity to
CYC, however, was not linearly dependent on the Cr content. The
high CYC selectivity (ca. 99 %) was only achieved when Cr content
was below 10 atom %, beyond which, the excessive chromium
would greatly accelerate the side-reactions and reduce CYC selec-
tivity. As a consequence, 10%Cr-PKU-1 sample was selected as
the model catalyst in order to further discuss the catalytic
mechanism.
Fig. 4. Reaction rate of CHOL and CYC selectivity in the different solvents over 10%
Cr-PKU-1. Reaction conditions: 20 mmol CHOL, 40 mg catalyst, 20 mL solvent, 2
equivalent H
2
O
2
oxidant, and refluxed at the individual boil point for 8 h.
Iron ion sometimes was reported to act as an efficient active
metal for some redox reactions, such as the case of Fe/AlPO-5
[
45,46]. However, Fe-substituted PKU-1 samples in this study did
not give a positive catalytic capacity for the dehydrogenation of
Fig. 5. The reaction rate of CHOL and CYC selectivity under different reaction
temperatures over 10%Cr-PKU-1 catalyst. Reaction conditions: 20 mmol CHOL,
Fig. 6. Reaction rate of CHOL and CYC selectivity under the different stoichiometric
4
0 mg catalyst, 20 mL CH
3
CN, 2 equivalent H
2
O
2
oxidant, and heated at different
proportions of
20 mmol CHOL, 40 mg catalyst, 20 mL CH
H
2
O
2
/CHOL over 10%Cr-PKU-1 catalyst. Reaction conditions:
CN, and refluxed at 354 K for 8 h.
temperature for 8 h.
3

W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
135
stability. However, the further ICP analysis on reaction solution
indicated there was a slight release of Cr active metal into the solu-
tion, the metal loss was about 4.13 %. Though unavoidable some-
times, active metal leakage seems to be greatly influenced by the
performed reaction conditions, especially the type and concentra-
tion of the used oxidant. For example, when the oxidant TBHP with
weak oxidizing power was used to take place of H
2 2
O , the released
amount of active metal Cr was trace and can be ignorable.
3
3
.3. Discussions about reaction mechanism
.3.1. Effect of Cr sites on the formation of active intermediate species
As mentioned above, chromium ions in Cr-PKU-1 were assumed
to function as active sites to quickly active H O into oxygenated-
2 2
intermediate species during the sec-alcohols oxidation. In order to
confirm this statement, the following experiments were designed.
2 2 3
First, a certain amount of H O (30 wt %) and CH CN were trans-
ferred into a test tube; next, a trace of Cr-PKU-1 powder was added
into the above solution; then, the test tube was semi-sealed by Par-
Fig. 7. The influence of Cr amount incorporated in PKU-1 on reaction rate of CHOL
and CYC selectivity. Reaction conditions: 20 mmol CHOL, 40 mg catalyst, 20 mL
CH CN, 2 equivalent H O oxidant, and refluxed at 354 K for 8 h.
3 2 2
afilm; finally, the residual amount of H
determined by redox titration using KMnO
A photograph was taken to monitor O
played in Fig. 10a, where bubbles were observed and some of them
attached to the wall of the test tube. In fact, the gas was identified
2
O
2
after 4 h activation was
as standard solution.
bubbling rate and dis-
4
CHOL. As shown in Fig. 8 (left side), all Fe-PKU-1 samples exhibited
poor catalytic activities, and the maximum TOF was only 1.08
molSubÁmolCatÁh when using 30 atom % Fe-PKU-1 as the catalyst,
2
À1 À1
which was much smaller than the ones of Cr-substituted PKU-1
catalysts. Leithall et al. reported that bimetal-substituted molecu-
lar sieves VTiAlPO-5 have exhibited a superior performance than
the monometallic analogues in the epoxidation of cyclohexene, in
which V and Ti ions were proved to give a cooperative and syner-
gistic effect [47]. In our present study, Fe-Cr-PKU-1 catalysts with
different Fe/Cr molar ratio were successively synthesized, and their
catalytic performances in CHOL dehydrogenation were assessed,
but were rather poor. In other words, the introduction of iron ions
had a negative impact on CHOL transformation and CYC selectivity
as O
2
and should be originated from the catalytic decomposition of
H O over Cr-PKU-1. This means Cr-PKU-1 can significantly acti-
2 2
vate the peroxide bond (AOAOA) of H
2
O
2
and thus give an even
Å
more active oxygen-inclusive intermediate species, either OH rad-
Å
À
2
icals or superoxide ions O . If these highly-active species were not
completely consumed by the reaction substrates (like in this
experiment), they would therefore prefer to deactivate quickly
and combine with each other to form O . As for the reaction of
2
CHOL dehydrogenation, these highly-active oxidative species are
supposed to efficiently capture the hydrogen from CHOL, and lar-
gely accelerate dehydrogenation process.
(
see right side of Fig. 8). When increasing the amounts of iron ions,
their catalytic activity decreased dramatically. It can be drawn that
iron ions are unfavorable to liquid phase dehydrogenation for sec-
alcohol.
The powder catalysts can be easily separated from the reaction
solution by simple filtration followed by washing with acetone and
water. After that, the reusability of 10%Cr-PKU-1 catalyst was eval-
uated. As shown in Fig. 9, 10%Cr-PKU-1 maintained its original
framework structure and did not exhibit any obvious loss of
catalytic activity after 5 runs, evidencing the excellent framework
Quantitative analyzes on the residual H
redox titration using KMnO as oxidant, where the amount of con-
sumed KMnO was shown in Fig. 10b. Apparently, the O evolution
2 2
O were performed by
4
4
2
3+
rate shows a positive dependence on the concentration of Cr . In
other words, a higher chromium concentration in Cr-PKU-1 would
lead to a better catalytic efficiency due to the quicker decomposi-
Fig. 9. Recyclability of 10%Cr-PKU-1 catalyst for CHOL dehydrogenation evidenced
by XRD pattern after 5 runs and the catalytic performances for each run. Because
the powdered catalyst unavoidably has a small loss after each run, the amounts of
reactants and solvent will be reduced accordingly in the identical proportion with
above operations.
Fig. 8. Reaction rate of CHOL and CYC selectivity over Fe-substituted PKU-1 and Fe-
substituted 5%Cr-PKU-1 with different Fe contents. Reaction conditions: 20 mmol
3 2 2
CHOL, 40 mg catalyst, 20 mL CH CN, 2 equivalent H O oxidant, and refluxed at
3
54 K for 8 h.

1
36
W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
Fig. 10. O
2
bubbling rate monitored by catalytic decomposition of H
2 2 4
O over the Cr-PKU-1 catalysts (a) and the consumed amount of KMnO (b), respectively. Reaction
conditions: 20 mg Cr-PKU-1 catalyst, 1.5 mL H
2
O
2
, 2.5 mL CH CN, and heated at 313 K for 4 h.
3
tion of H
2
O
2
into active intermediate. In this sense, it is under-
high conversion rate. It is a convincing proof to confirm the
enhanced production of OH radicals by Cr-PKU-1 catalysts. In
Å
standable about the positive correlation between the reaction rate
of CHOL and Cr content in Cr-PKU-1, while excessive chromium
sites are detrimental to CYC selectivity, as is previously shown in
Fig. 7.
order to further demonstrate this structure-activity relationship,
we also employed fluorescence molecular probe to monitor the
formation of ÁOH radicals in the presence of different type of M-
Å
PKU-1 catalysts. Fig. 12 gives the amount of OH radicals resulted
from the catalytic decomposition of H
PKU-1 catalyst, and shows OH generation rate was greatly domi-
2
O
2
by Cr-PKU-1 or Fe-
3
.3.2. Detection of the involved critical radicals
Å
2 2
H O
is popularly believed to decompose into hydroxyl radicals
3
+
(ÅOH) and/or superoxide ions ( O
species in redox reactions [48,49]. In the following, it is important
to figure out the real active intermediate species when H
activated by Cr-PKU-1.
Å
À
nated by the type and amount of active metal sites. Cr ions effi-
2
), which thereafter act as critical
Å
ciently promote the formation of OH radicals, while iron ions
Å
exhibit a poor capacity to catalytically convert H
2
O
2
into OH rad-
2 2
O was
icals. This result explains why Cr-PKU-1 catalysts have much better
catalytic activity than Fe-PKU-1 catalysts. Fig. 12 also indicates that
the fluorescence intensity of TAOH has a rapid increase along with
increase in the amounts of incorporated amounts transition metal.
That is to say, a higher concentration of Cr ions will have a stron-
ger capacity to accelerate the generation of ÁOH radicals, thus can
greatly improve the catalytic conversion of CHOL.
First, the so-called terephthalic acid (TA) fluorescent method
was performed to in-situ monitor the formation of ÁOH radicals
[
50]. In detail, the hydroxyl radicals can be trapped by TA to form
3
+
a
highly fluorescent compound, 2-hydroxyterephthalic acid
(
TAOH), whose maximum fluorescence emission is centered about
at 440 nm and the emission intensity is proportional to the amount
of the produced TAOH in the liquid phase [51]. If without any cat-
alyst, the PL was too weak to be detected, which means the amount
Accordingly, we believe Cr-PKU-1 holds a great potential
because it has a controllable capability of H
2
O
2
activation, and
Å
Å
OH is the exclusive product. In our study, the model reaction
of OH radicals produced under the catalyst-free conditions was
was the dehydrogenation of sec-alcohols, and the compatible con-
negligible. Once 10%Cr-PKU-1 was added, a broad absorption band
centered at 440 nm was observed and the intensity gradually grew
with time increasing from 1 to 4 h (see Fig. 11a).
On the other hand, NBT is a widely used superoxide ion indica-
tor for the formation of purple formazan with the absorbance at
k = 560 nm [52,53]. Here, NBT was used as the probe molecule to
3
+
centration of Cr was estimated to be 10 % to achieve both high
conversion rate and selectivity (>99 %). Apparently, the allowed-
3
+
concentration of Cr in the octahedral-based framework of PKU-
is much higher than the case of TS-1. Although the further
1
3
+
increase of the Cr concentration in Cr-PKU-1 will give a deeper
oxidation in the dehydrogenation of CYC, the Cr-PKU-1 catalyst
with high concentration might provide a great potential in other
catalytic oxidation systems.
Å
À
2
in-situ detect another possible reactive species O . The blank
experiment was performed without adding any catalyst, and the
concentration of formazan also has a continuous increase (see
Fig. 11b); however, when 10%Cr-PKU-1 is added into the test solu-
tion, the concentration of formazan was far below the ones gener-
ated in the blank experiment.
3.3.3. Discussions about induction period
The catalytic CHOL dehydrogenation was largely dominated by
the concentrations of hydroxyl group radicals, and a critical con-
centration seemed required to push forward this dehydrogenation
reaction. Table 1 gives the time dependence of CHOL conversion
and CYC selectivity over the Cr-PKU-1 with different Cr contents.
A pre-activation stage for CHOL dehydrogenation was needed,
and the higher the chromium content, the shorter the activation
time. For instance, in the presence of 6%Cr-PKU-1 catalyst, CHOL
conversion was negligible within 3 h; for 10%Cr-PKU-1 catalyst,
the induction period was shortened to about 1 h, and CHOL conver-
sion only reached 4 % after 2 h; if 20%Cr-PKU-1 was chosen, the
induction period of about 1 h was still required, but the CHOL con-
version could reach up to 25 % after 2 h. In the meantime, the
It is quite interesting to observe the opposite tendency of the
Å
Å À
2
concentration variation of OH and O
Cr sites catalyzed the quick formation of active OH radicals from
radicals. In other words,
Å
H
2
O
2
, however strongly suppressed the generation of superoxide
Å
À
Å
O
2
ions. Therefore, we reasonably considered OH radicals as the
critical active species to catalyze the dehydrogenation of sec-
alcohols. It is also very important that Cr-PKU-1 suppressed the
generation of superoxide O
ble side reactions caused by O
As evidenced in Figs. 7 and 8, the type or concentration of incor-
porated transition-metals played a vital role on the conversion rate
and the selectivity. First, a high concentration of Cr leads to a
Å
À
2
ions, which will eliminate the possi-
Å
À
2
.
3
+

W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
137
Å
Å
À
Å
Fig. 11. Formation of OH radicals and O radicals ion monitored by molecular probe method: (a) OH radicals evidenced by the production of the fluorescent TAOH and (b)
2
Å
À
O
2 2 3
ion evidenced by UV-vis absorption spectrum of the production of formazan. Operational condition: TA or NBT (3 mmol), CHOL (20 mmol), H O (40 mmol) and CH CN
2
(
20 mL). The blue and red lines represent the time-dependent spectra measured with and without 10%Cr-PKU-1 catalyst, respectively.
Å
Fig. 12. Formation of OH radicals catalyzed by Cr-PKU-1 or Fe-PKU-1 with different transition metal contents evidenced by the production of the fluorescent TAOH after 2 h
2 2 3
reaction time. Operational condition: TA (3 mmol), CHOL (20 mmol), H O (40 mmol) and CH CN (20 mL).
Table 1
Time dependence of CHOL conversion and CYC selectivity over the Cr-PKU-1 with different Cr contents.
Time (h)
6% Cr-PKU-1
Conv. (%)
10% Cr-PKU-1
Conv. (%)
20% Cr-PKU-1
Conv. (%)
Sel. (%)
Sel. (%)
Sel. (%)
1
2
3
4
5
6
7
8
<1
<1
<1
4
99
99
99
99
99
99
99
99
<1
4
8
17
21
27
40
46
99
99
99
99
99
99
99
99
<1
25
40
47
48
50
54
58
99
82
79
77
76
75
74
73
6
11
15
18
3 2 2
Reaction conditions: 20 mmol CHOL, 40 mg catalyst, 20 mL CH CN, 2 equivalent H O oxidant, and refluxed at 354 K.
instantaneous reaction rate also decayed very rapidly because of
3.3.4. Redox couples evaluated by electrochemical measurement
Å
3+
the quick transformation of oxidant H
2
O
2
into active OH radicals
As discussed above, Cr ions were proved to play a key role in
catalyzed by the higher-chromium catalyst. In other words, the
observation of induction period suggests that CHOL dehydrogena-
tion reaction will be activated, only when the concentration of OH
2 2
the catalytic transformation of H O into active radicals, therefore
the identification of Cr functioned in this process was much
attractive. In some transition-metal-catalyzed organic reactions
3
+
Å
radicals reaches the minimum threshold.
2 2
with using H O as oxidant, the higher-valence-state transition

1
38
W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
2+
3
+
metal ion was generally discovered to be reduced to a lower one.
For example, Fe species were commonly identified in the Fenton
or Fenton-like reaction system (Fe -H O system), among which, a
2 2
Fe -Fe -Fe cycle was established accompanied with the gener-
ation of ÁOH radicals according to the so-called Fenton-like mech-
anism [54–56]. Similarly, some kind of chromium-containing ion
couple (probably the coupled Cr /Cr ) is highly possible to be
involved in the Cr -H
coupled Cr /Cr (À0.63 V vs Ag/AgCl). The shift toward a more
negative potential should attribute to the difference in coordina-
tion modes between solid and solvated environments [64]. There
was no detectable redox peak when scanning from 0 to 1 V. Thus,
from another perspective, the absence of higher E1/2 implied that
2
+
3
+
3
+
2+
3+
3
+
6+
3+
5+
the redox circulation for coupled Cr /Cr or Cr /Cr was unlikely
to proceed [65].
3+
2+
3
+
2
O
2
reaction system, which was hereinafter
Some researchers suggested that an extremely stable frame-
work is often unfavorable to configuration transitions of active
metal cations, and the resistance against redox process seems to
highly depend on the porous structure and the resultant frame-
work with strong stress, which often results in a local distortion
or even a lower structural asymmetry [66,67]. In fact, the crystal
structure of Cr-PKU-1 possesses three independent metal sites in
the space group of R3, and one of them is located in a quite dis-
torted octahedral environment [28]. From the structural point of
identified by cyclic voltammetry (CV) measurement.
Cyclic voltammetry is in fact an useful technique in characteriz-
ing the local environment in the case of transition-metal-bearing
porous solids. For example, Cartro-Martins et al. applied the CV
measurements to prove that only the Ti4+ ions in the framework
underwent electron-transfer processes, not the extra-framework
Ti ions [57,58]. Venkatathri and his co-workers investigated sev-
eral vanadium-containing molecular sieves and found that two dis-
tinct sets of redox couples were present in the CV curves, which
were speculated to be V /V in two different coordination envi-
ronments, namely a distorted tetrahedral and a square pyramidal
symmetry [59–62]. Borges et al. studied the electrochemical prop-
erties of a cobalt-substituted aluminophosphate (CoIST-2) and pro-
vided a clear evidence for the presence of at least two types of Co
species [63].
4
+
6 6
view, the distorted AlO /CrO octahedra seem to afford the static
5
+
4+
3+
stress in the main framework, for this particular reason, Cr ions
are reasonably speculated to easily undergo a catalytic circulation
for a given redox reaction. In the next step, the real functionality of
3
+
Cr and the mechanism are validated by catalytic oxidative dehy-
drogenation of sec-alcohol shown in the following sections.
2
+
The CV curves in Fig. 13 exhibit a single oxidation and reduction
process for all Cr-PKU-1 samples (10%, 30% and 50%), and their rel-
ative peak potentials (E1/2) located at ꢀÀ0.48 V vs Ag/AgCl refer-
ence electrode, in which E1/2 is the average value of the oxidation
potentials (around À0.64 V) and reduction ones (around
À0.32 V). The peak separation and similar current intensity ratio
indicate that the redox process is quasi-reversible and repro-
ducible for Cr species within the PKU-1 framework. Since the CV
peaks are the indications of the valence alternation of Cr ions, it
is as expected that the relevant current intensities become more
and more significant with the increase in Cr concentration. For
comparison, no such redox behavior is observed in the
chromium-free PKU-1 sample at the same experimental conditions
3
.3.5. Reaction mechanism
Based on above observations as well as those in previous stud-
ies [68–70], a possible reaction mechanism is proposed in Fig. 14.
3+
First, the sixfold coordination Cr (1) easily reacts with H
2
O
2
to
form Cr AOOH (2) accompanied with the production of H . Then,
undergoes a hemolytic cleavage of its O-OH bond to generate
2+
+
2
2+
Å
Å
Cr AO (3) and OH radicals. Next, 3 can capture a hydrogen atom
from organic substrate CHOL to form Cr AOH (4), along with the
formation of C H11O (5). Finally, 4 reacts with H back to 1 and
6
2+
Å
+
simultaneously generates a water molecule; Meanwhile, 5 com-
bine with ÁOH radicals (which was generated in the previous step)
2
to generate CYC and H O.
(
see the inset of Fig. 13).
As previously indicated by XPS analyzes (Fig. S10, SI), chromium
3
3
.4. Reaction pathways under the over-oxidative conditions
species in the Cr-PKU-1 are proved to be Cr3 . Accordingly, these
paired anodic and cathodic peaks are supposed to result from the
reversible reaction between Cr and another type of Cr couple,
reasonably Cr
potentials. As shown in Fig. 13, the observed E1/2 is about
À0.48 V (vs Ag/AgCl), and is reasonably close to the one of the
+
.4.1. Reaction pathway for over-oxidation of CHOL
As described in previous sections, CYC selectivity was depen-
3+
2
+
2 2
dent on reaction conditions, especially the amount of H O or
Cr atom % in the Cr-PKU-1 catalyst. For instance, when the
equivalent amount of H /CHOL was raised from 2 to 3, the CYC
ions, based on the related standard redox
3
+
2 2
O
Fig. 14. The proposed reaction mechanism for catalytic dehydrogenation over Cr-
Fig. 13. Cyclic voltammetry curves of Cr-PKU-1 samples with different Cr contents.
PKU-1 catalyst.

W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
139
selectivity would decrease about 50 % because of the over-
oxidation (see Fig. 6), and consequently a large amount of by-
products such as dicarboxylic acids were generated. Interestingly,
the identification of these emerging intermediates could provide
solid evidences to recognize the degradation pathways of CHOL.
In the present case, most intermediates were identified using
GC–MS measurement, whose molecular structure was included
in Fig. S10, SI. For clarify, the degradation of CHOL can be reason-
ably classified into 3 continuous routes: Route I (Fig. 15a), II and
III (Fig. 15b).
oxidized into o, m, or p-cyclohexanediol, respectively, through a
hydroxylation reaction, subsequently the dehydrogenation process
to corresponding 2, 3, or 4-hydroxycyclohexanone occurs. While in
Route III, the hydroxycyclohexanone can be produced in the
sequence of dehydrogenation and hydroxylation reactions. Eventu-
ally, two competitive reactions, i.e. dehydrogenation or dehydra-
tion steps, are followed by the production of cyclohexanedione or
cyclohexenone.
3.4.2. Discussion about the selectivity for each pathway
In route I, CHOL was degraded successively into some smaller
organic molecules through multistep oxidative reactions, such as
catalytic dehydrogenation, Baeyer-Villiger oxidation and oxidative
ring-opening reaction. In detail, CHOL was first oxidized to CYC
through a dehydrogenation process. In an over-oxidative environ-
ment, the so-called Baeyer-Villiger oxidation will occur, and a cyc-
As might be expected, these parallel over-oxidation routes are
not equivalent and will be largely influenced by different reaction
conditions. Probably the concentration of intermediates will give a
hint about the most favorable route. As already proved, both the
3
+
Cr concentration and the oxidant equivalence have an important
influence on the CYC selectivity. Rationally, three batches with dif-
ferent catalytic conditions were performed, and the concentrations
lic lactone called e-carprolactone was subsequently produced. In
this way, these over-oxidative procedures would continue to con-
secutively produce other intermediates, such as adipic acid and d-
valerolactone. Formic acid molecules, however, were produced as
aliphatic dicarboxylic acids are transformed into cyclic lactone.
The oxidation process in routes II and III contains two same
reactions: dehydrogenation and hydroxylation reactions, the only
difference between these two routes is the sequence of these two
reactions (Fig. 15b). For example in Route II, CHOL is initially
(
selectivity) of main intermediates were also determined using GC
or HPLC.
As shown in Table 2, the selectivity to main intermediates was
determined, and especially the summed selectivity for each route
was emphasized. First in Batch 1, when the high chromium catalyst
(
4
30%Cr-PKU-1) was used, CHOL conversion was enhanced from
6 % to 62 %, however, the CYC selectivity was dramatically
reduced from 99 % to 46 % compared to the 10%Cr-PKU-1 model
Fig. 15. The supposed reaction pathway for over-oxidation of CHOL, the main product and by-products of CHOL (1) oxidation is listed as below: CYC (2);
e-carprolactone (3);
adipic acid (4); d-valerolactone (5); glutaric acid (6);
c-butyolactone (7); o, m or p-cyclohexanediol (8); 2, 3 or 4-hydroxycyclohexanone (9); 2-cyclohexen-1-one (10) and p-
cyclohexanedione (11).

1
40
W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
Table 2
The selectivity of some by-products emerged in the proposed reaction routes for the over-oxidation of CHOL under the different reaction conditions.
Compounds No.
Batch 1 (62 %)a
Selectivity (%)
Batch 2 (80 %)b
Selectivity (%)
Batch 3 (70 %)c
Selectivity (%)
2
46
38
43
Route I
3
4
Trace
12
Trace
18
Trace
17
5
6
0.1
14
0.9
11
0.2
9
7
Sum
8
9
0.2
26.3
8
0.2
30.1
10
0.6
26.8
10
Route II and III
14
18
18
1
1
Sum
0
1
0.6
1.0
23.6
0.6
2.0
30.6
0.6
1.0
29.6
Reaction conditions for each batch:
a
2
2
2
0 mmol CHOL, 40 mg 30%Cr-PKU-1 catalyst, 20 mL CH
0 mmol CHOL, 40 mg 10%Cr-PKU-1 catalyst, 20 mL CH
0 mmol CHOL, 40 mg 30%Cr-PKU-1 catalyst, 20 mL CH
3
CN, 2 equivalent H
2
O
2
, refluxed at 354 K for 8 h.
O , refluxed at 354 K for 12 h.
b
c
3
CN, 4.5 equivalent H
CN, 2.5 equivalent H
2 2
3
2
O
2
, refluxed at 354 K for 12 h. The value in parentheses is the CHOL conversion for each
batch.
catalyst (see Fig. 7). The over-oxidation selectivity through Route I
and II and III is 26.3 % and 23.6 %, respectively. In general, a higher
Cr concentration in Cr-PKU-1 did not show a significant prefer-
ence to boost either Routes I or II and III, which seems to be really
equivalent.
incorporated into the framework sites, and the linear expansions
of cell unit volume indicate M-PKU-1 have a higher upper limit
3
+
3+
for doping transition metal cations (50 atom % for Cr ). XPS, cyc-
lic voltammetry and crystal structure analyzes suggested Cr3+ ions
were much likely to locate in a highly-distorted octahedral envi-
ronment and thereby can easily and quickly forward the redox
circles when the oxidative reaction happened.
2 2 2 2
In Batch 2, a higher equivalent H O (H O /CYOL = 4.5) was
added into the reaction system. The results indicate that CHOL con-
version further increased from 46 % to 80 %, in the meantime, the
CYC selectivity declined to an even smaller value of 38% comparing
Under the performed conditions, Cr-PKU-1 samples show much
superior catalytic performance to Fe-incorporated ones (99 %
selectivity and ꢀ15 times higher in TOF for 10 atom % substitu-
tion). The isolated Cr sites are proved to function as active sites
to the optimized ratio of H
either Route I or II and III exhibited a further increase (30.1 and
0.6 %, respectively) compared to Batch 1. That is to say, the com-
2 2
O /CYOL = 2. The overall selectivity to
3
and be responsible to activate rapidly H
2 2
O to release the active
petition between Routes I and II and III was well matched and not
OH radical. The fluorescence probe method indicates the presence
dominated by any of the above two factors (Cr3 concentration or
+
of Cr in the framework of PKU-1 promotes the activation of H
into OH radicals exclusively and prohibits the formation of O
3+
O
2
2
Å
Å
À
equivalence of H
In Batch 3, H
2
O
2
).
(in the equivalence of 2.5) was added with a
2
2
O
2
ions in the meantime, and the amount of ÁOH radicals was greatly
dominated by the type and amount of active metal sites. A low-
level Cr in PKU-1 (6%Cr-PKU-1) tends to provide lower concentra-
slow rate and in a successive manner, which has been proved to
affect the selectivity of main products [8,71]. In the current case,
the CHOL conversion and CYC selectivity were 70 % and 43 %,
respectively, but the previously competitive Routes I and II and
III became somewhat selective. In other words, the selectivity to
Route I substantially unchanged comparing to Batch 1, while the
ones toward Routes II and III elevated from 23.6 % to 29.6 %. This
Å
tion of OH radical and thus results into a long reduction period; on
the contrary, an excessive amount of Cr active centers (20%Cr-PKU-
Å
1) will bring an over-high concentration of OH (at least ten times,
compared to 6%Cr-PKU-1 catalyst) and thus greatly shorten induc-
tion period, thereafter however produce a variety of by-products
such as adipic acid or hydroxycyclohexanone. A hypothetical
mechanism was also proposed to demonstrate the functions of Cr
is a hint that the successive adding of H
2
O
2
provided a sustainable
Å
but a low concentration of ÁOH radicals, and a low but instanta-
neous concentration of active species is beneficial to hydroxylation
reactions involved in Routes II and III, while a higher one is in favor
of oxidative degradation involved in Route I.
3
+
2+
3+
catalytic sites, in which a Cr -Cr -Cr circle was involved and
3
+
6+
3+
very different with the Cr -Cr -Cr circle presented in the
tetrahedral-based redox molecular sieves. The stepwise by-
products were analyzed by GC–MS and HPLC to demonstrate the
probable reaction route under over-oxidation conditions, and three
parallel pathways containing dehydrogenation, hydroxylation,
Baeyer-Villiger oxidation or ring-opening reactions, were specu-
lated to clarify the progressive oxidation of CHOL under the over-
oxidation environment. Although this study is a preliminary study
3+
On the whole, an excessive concentration of Cr
in the
Cr-PKU-1 catalyst definitely leads to over-oxidative conditions for
deep-oxidized by-products through two completely different path-
ways, which was only probably controlled by the instantaneous
concentration of oxidative species in solution. The particularity of
Cr-PKU-1 is to catalyze the selective decomposition of H
2
O
2
into
Å
OH and interestingly in a controllable rate. For the sake of a high
selectivity from sec-alcohols to ketones, the most efficient catalyst
is 10%Cr-PKU-1 under the performed conditions, while the further
improvement to develop possible bi-catalyst systems is proposed.
2 2
on the controllable and selective activation of H O catalyzed by
Cr-PKU-1, and considering their significant capacity to accommo-
date desired metals with high level, M-PKU-1 catalysts really holds
high potentials, for example, on catalyzing other redox reactions,
probably with additional loading of cocatalysts.
4
. Conclusions
Acknowledgment
As a novel octahedral-based redox molecular sieve, M-PKU-1
M@Cr or Fe) has been facilely synthesized by the boric acid
reflux method. X-ray Diffraction Rietveld refinement method
has evidenced that transition metals have been successfully
This work was financially supported by the National Natural
Science Foundation of China (Grants nos. 91222106 and
21671028) and Natural Science Foundation of Chongqing (Grants
nos. 2014jcyjA50036 and 2016jcyjA0291).
(

W. Wang et al. / Journal of Catalysis 352 (2017) 130–141
141
[
25] W.L. Gao, Y. Jing, J. Yang, Z.Y. Zhou, D.F. Yang, J.L. Sun, J.H. Lin, R.H. Cong, T.
Yang, Inorg. Chem. 53 (2014) 2364–2366.
Appendix A. Supplementary material
[
26] G.J. Wang, W.L. Wang, F.H. Zhang, W.L. Gao, R.H. Cong, T. Yang, Catal, Sci.
Technol. 6 (2016) 5992–6001.
In supporting information, the supplementary materials in
Figs. S1–S10 have been provided: UV-Vis absorption spectra of
Cr O and Cr-PKU-1 with different Cr content (Fig. S1); Powder
2 3
[27] W.L. Wang, Y. Wang, B. Wu, R.H. Cong, W.L. Gao, B. Qin, T. Yang, Catal.
Commun. 58 (2015) 174–178.
[
28] Q.Q. Li, C. Lin, Z.Y. Zhou, J. Ju, G.B. Li, J.H. Lin, W.L. Gao, R.H. Cong, T. Yang, Inorg.
Chem. 53 (2014) 5600–5608.
XRD patterns of Fe-PKU-1 with different Fe content (Fig. S2); The
dependence of unit cell volumes for Fe-PKU-1 samples on the dif-
ferent concentration of Fe atoms (Fig. S3); Powder XRD patterns of
Fe-5%Cr-PKU-1 with different Fe content (Fig. S4), The dependence
of unit cell volumes for Fe-5%Cr-PKU-1 samples on the different
concentration of Fe atoms (Fig. S5), SEM images and EDS analysis
of 20%Cr-PKU-1 sample (Fig. S6), In-situ high temperature XRD pat-
terns for 10%Cr-PKU-1 sample (Fig. S7), TG-DSC analysis of 10%Cr-
PKU-1 sample (Fig. S8), XPS survey spectrum of Cr-PKU-1 (Fig. S9)
and the emerging intermediates determined by GC-MS measure-
ment (Fig. S10). This material is available free of charge via the
Internet at http://www.elsevier.com. Supplementary data associ-
ated with this article can be found, in the online version, at
http://dx.doi.org/10.1016/j.jcat.2017.05.007.
[29] W.L. Gao, T. Yang, Y.X. Wang, G.B. Li, F.H. Liao, J.H. Lin, J. Phys. Chem. B 109
2005) 22775–22779.
[
(
30] R.A. Sheldon, J. Mol. Catal. A: Chem. 107 (1996) 75–83.
31] T. Mallat, A. Baiker, Chem. Rev. 104 (2004) 3037–3058.
[
[32] R.A. Sheldon, I.W.C.E. Arends, G.-J.T. Brink, A. Dijksman, Acc. Chem. Res. 35
2002) 774–781.
33] C.J. Weiss, P. Das, D.L. Miller, M.L. Helm, A.M. Appel, ACS Catal. 4 (2014) 2951–
958.
[34] M. Uyanik, M. Akakura, K. Ishihara, J. Am. Chem. Soc. 131 (2009) 251–262.
(
[
2
[
[
[
35] M.B. Lauber, S.S. Stahl, ACS Catal. 3 (2013) 2612–2616.
36] B.A. Steinhoff, A.E. King, S.S. Stahl, J. Org. Chem. 71 (2006) 1861–1868.
37] Y.-C. Son, V.D. Makwana, A.R. Howell, S.L. Suib, Angew. Chem. Int. Ed. 40
(2001) 4280–4283.
[
[
38] T.L. Lu, Z.T. Du, J.X. Liu, H. Ma, J. Xu, Green Chem. 15 (2013) 2215–2221.
39] U. Mandi, M. Pramanik, A. Roy, N. Salam, A. Bhaumik, S.M. Islam, RSC Adv. 4
(
2014) 15431–15440.
[40] R. Kumar, S. Sithambarama, S.L. Suib, J. Catal. 262 (2009) 304–313.
[
[
[
41] TOPAS V: TOPAS, V4.1-beta. Bruker AXS, Karlsruhe, Germany, 2004.
42] R. Shannon, Acta Crystallogr. A 32 (1976) 751–767.
43] H.J. Sung, C. Jong-S, K.H. Young, G. Jean-M, K.C. Anthony, J. Phys. Chem. B 109
References
(
2005) 845–850.
[
1] Z. Guo, B. Liu, Q.H. Zhang, W.P. Deng, Y. Wang, Y.H. Yang, Chem. Soc. Rev. 43
2014) 3480–3524.
2] M.E. Davis, Acc. Chem. Res. 26 (1993) 111–115.
[
[
44] L. Canesson, A. Tuel, Zeolites 18 (1997) 260–268.
45] D. Goldfarb, M. Bernardo, K.G. Strohmaier, D.E.W. Vaughan, H. Thomann, J. Am.
Chem. Soc. 116 (1994) 6344–6353.
46] Y. Wang, Q.H. Zhang, T. Shishido, K. Takehira, J. Catal. 209 (2002) 186–196.
47] R.M. Leithall, V.N. Shetti, S. Maurelli, M. Chiesa, E. Gianotti, R. Raja, J. Am.
Chem. Soc. 135 (2013) 2915–2918.
48] Y. Zhang, N. Zhang, Z.-R. Tang, Y.-J. Xu, Chem. Sci. 4 (2013) 1820–1824.
49] K.-C. Yang, J.-H. Zheng, Y.-L. Chen, K.-C. Lee, E.-C. Cho, RSC Adv. 6 (2016)
(
[
[
3] R.A. Sheldon, M. Wallau, I.W.C.E. Arends, U. Schuchardt, Acc. Chem. Res. 31
[
[
(
1998) 485–493.
[
[
[
[
4] M. Hartmann, L. Kevan, Chem. Rev. 99 (1999) 635–663.
5] J. Taramasso, G. Perego, B. Natari, U. S. Patent 4,410,501, 1983.
6] N. Wilde, M. Pelz, S.G. Gebhardt, R. Glaser, Green Chem. 17 (2015) 3378–3389.
7] S.T. Du, F. Li, Q.M. Sun, N. Wang, M.J. Jia, J.H. Yu, Chem. Commun. 52 (2016)
[
[
5
3025–53033.
3
368–3371.
8] F. Song, Y.M. Liu, H.H. Wu, M.Y. He, P. Wu, T. Tatsumi, J. Catal. 237 (2006) 359–
67.
9] S. Ekambaram, S.C. Sevov, Angew. Chem. Int. Ed. 38 (1999) 372–375.
[
50] F. Raziq, Y. Qu, X.L. Zhang, M. Humayun, J. Wu, A. Zada, H.T. Yu, X.J. Sun, L.Q.
Jing, J. Phys. Chem. C 120 (2016) 98–107.
[
3
[
[
[
[
51] J. Hu, J. Men, Y.Y. Liu, H. Huang, T.F. Jiao, RSC Adv. 5 (2015) 54028–54036.
52] H. Goto, Y. Hanada, T. Ohno, M. Matsumura, J. Catal. 225 (2004) 223–229.
53] Y. Wang, K. Deng, L. Zhang, J. Phys. Chem. C 115 (2011) 14300–14308.
54] F. Gong, L. Wang, D.W. Li, F.Y. Zhou, Y.Y. Yao, W.Y. Lu, S.Q. Huang, W.X. Chen,
Chem. Eng. J. 267 (2015) 102–110.
[
[
[
10] L.H. Chen, X.Y. Li, G. Tian, Y. Liu, J.C. Rooke, G.S. Zhu, S.L. Qiu, X.Y. Yang, B.L. Su,
Angew. Chem. Int. Ed. 50 (2011) 11156–11161.
11] W.G. Cheng, X.S. Wang, G. Li, X.W. Guo, S.J. Zhang, J. Catal. 255 (2008) 343–
3
46.
[
55] J. Sha, E.J. Zheng, W.J. Zhou, A. Liebens, M. Pera-Titus, J. Catal. 337 (2016) 199–
[
[
12] C.L. Lin, R.K. Chiang, K.H. Lii, J. Am. Chem. Soc. 131 (2009) 2068–2069.
13] Y.F. Shen, R.P. Zerger, R.N. DeGuzman, S.L. Suib, L. McCurdy, D.I. Potter, C.L.
O’Young, Science 260 (1993) 511–515.
14] S.L. Suib, Acc. Chem. Res. 41 (2008) 479–487.
15] L. Espinal, W. Wong-Ng, J.A. Kaduk, A.J. Allen, C.R. Snyder, C. Chiu, D.W.
Siderius, L. Li, E. Cockayne, A.E. Espinal, S.L. Suib, J. Am. Chem. Soc. 134 (2012)
944–7951.
16] X. Meng, J.Q. Zhang, B.H. Chen, Z.Q. Jing, P.Q. Zhao, Catal, Sci. Technol. 6 (2016)
90–896.
17] A.M. EI-Sawy, C.K. Kingondu, C.H. Kuo, D.A. Kriz, C.J. Guild, Y.T. Meng, S.J.
Frueh, S. Dharmarathna, S.N. Ehrlich, S.L. Suib, Chem. Mater. 26 (2014) 5752–
2
07.
[
[
56] S. Hwang, S.G. Huling, S. Ko, Chemosphere 78 (2010) 563–568.
57] S. de Cartro-Martins, S. Khouzanpi, A. Tuel, Y.B. Taarit, N.E. Mur, A. Sellami, J.
Electroanal. Chem. 350 (1993) 15–28.
58] S. de Cartro-Martins, A. Tuel, Y.B. Taarit, Zeolites 14 (1994) 130–136.
59] M. Chatterjee, D. Bhattacharya, N. Venkatathri, S. Sivasanker, Catal. Lett. 35
[
[
[
[
7
(
1995) 313–326.
[
[
[
60] N. Venkatathri, M.P. Vinod, K. Vijayamohanan, S. Sivasanker, J. Chem. Soc.,
Faraday Trans. 92 (1996) 473–478.
61] N. Venkatathri, Appl. Catal. A 242 (2003) 393–401.
62] N. Venkatathri, Appl. Catal. A 310 (2006) 31–39.
63] C. Borges, M.F. Ribeiro, C. Henriques, J.P. Lourencüo, D.M. Murphy, A. Louati, Z.
Gabelica, J. Phys. Chem. B 108 (2004) 8344–8354.
64] R. Ganesan, B. Viswanathan, J. Mol. Catal. A: Chem. 223 (2004) 21–29.
65] T. Selvam, M.P. Vinod, Appl. Catal. A: Gen. 134 (1996) L197–L201.
66] A.A. Verberckmoes, B.M. Weckhuysen, R.A. Schoonheydt, Micropor. Mesopor.
Mater. 22 (1998) 165–178.
8
[
[
[
5
760.
[
[
[
[
18] M. Ousmane, G. Perrussel, Z. Yan, J.M. Clacens, F. De Campo, M. Peta-Titus, J.
Catal. 309 (2014) 439–452.
[
[
[
19] J. Ju, J.H. Lin, G.B. Li, T. Yang, H.M. Li, F.H. Liao, C.K. Loong, L.P. You, Angew.
Chem. Int. Ed. 42 (2003) 5607–5610.
20] T. Yang, A. Bartoszewicz, J. Ju, J.L. Sun, Z. Liu, X.D. Zhou, Y.X. Wang, G.B. Li, F.H.
Liao, B. Martin-Matute, J.H. Lin, Angew. Chem. Int. Ed. 50 (2011) 12555–12558.
21] H. Chen, J. Ju, Q.P. Meng, J. Su, C. Lin, Z.Y. Zhou, G.B. Li, W.L. Wang, W.L. Gao, C.
M. Zeng, C. Tang, J.H. Lin, T. Yang, J.L. Sun, J. Am. Chem. Soc. 137 (2015) 7047–
050.
22] J. Ju, T. Yang, G.B. Li, F.H. Liao, Y.X. Wang, L.P. You, J.H. Lin, Chem. Eur. J. 10
2004) 3901–3906.
23] T. Yang, J. Ju, G.B. Li, F.H. Liao, X.D. Zhou, F. Deng, L. Chen, Y.X. Wang, J.H. Lin,
Inorg. Chem. 46 (2007) 4772–4774.
[
[
[
[
[
67] P.A. Barret, G. Sankar, C.R.A. Catlow, J.M. Thomas, J. Phys. Chem. 100 (1996)
8
977–8985.
68] W.C.E. Arends, R.A. Sheldon, Ma. Wallau, U. Schuchardt, Angew. Chem Int. Ed.
6 (1997) 1144–1163.
7
3
[
[
[
69] R.A. Sheldon, J.K. Kochi, Metal-Catalyzed Oxidations of Organic Compounds,
Academic Press, New York, 1981.
70] Y.J. Wang, X. Wen, C.Y. Rong, S.P. Tang, W.F. Wu, C. Zhang, Y.C. Liu, Z.H. Fu, J.
Mol. Catal. A Chem. 411 (2016) 103–109.
(
24] W.L. Gao, Y.X. Wang, G.B. Li, L.P. You, J.H. Lin, Inorg. Chem. 47 (2008) 7080–
71] J.H. Ding, P. Wu, Appl. Catal. A: Gen. 488 (2014) 86–95.
7
082.