Cr/ZSM-5 for ethane dehydrogenation: Enhanced catalytic activity through surface silanol

Article abstract of DOI:10.1016/j.apcata.2016.12.025

Dehydrogenation of ethane to ethylene in the presence of CO₂was investigated over CrO_xcatalysts supported on ZSM-5 with different crystallite sizes, including nanometer, submicron and micrometer. These catalysts behave differently in the dehydrogenation reaction, but no relationship between the activity and crystallite size of ZSM-5 was found. XPS, UV-vis, laser Raman and HR-TEM studies show that the better activity comes from better CrO_xdispersion on ZSM-5 surface. The characterization results of ZSM-5 supports by N₂adsorption, SEM, XRD, XRF,²⁷Al and²⁹Si MAS NMR reveal that the difference of the dispersion is associated with their surface properties, not with the textural ones. Silanols on the surface play an important role in dispersing the CrO_xactive species. Silanol nests are more helpful in efficient CrO_xdispersion than terminal silanols, which subsequently result in higher dehydrogenation activity.

Full text of DOI:10.1016/j.apcata.2016.12.025

Accepted Manuscript
Title: Cr/ZSM-5 for ethane dehydrogenation: enhanced
catalytic activity through surface silanol
Author: <ce:author id="aut0005"
author-id="S0926860X16306299-
dc47cf4472a236c2869f8214d90e0471"> Yanhu
Cheng<ce:author id="aut0010"
author-id="S0926860X16306299-
a63bc1e197de018727d9541dce09e76f"> Changxi
Miao<ce:author id="aut0015"
author-id="S0926860X16306299-
3cd11af28adcd8f86b5dc7dfd92efa5c"> Weiming
Hua<ce:author id="aut0020"
author-id="S0926860X16306299-
acfdbcbef47c2dc6987ed291f3f9c27c"> Yinghong
Yue<ce:author id="aut0025"
author-id="S0926860X16306299-
b548be9757cb14b661d461e00a1c6f52"> Zi
Gao
PII:
S0926-860X(16)30629-9
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2016.12.025
APCATA 16110
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
2-11-2016
26-12-2016
30-12-2016
Please cite thisarticle as: YanhuCheng, Changxi Miao, WeimingHua, YinghongYue, Zi
Gao, Cr/ZSM-5 for ethane dehydrogenation: enhanced catalytic activity through surface
silanol, Applied Catalysis A, General http://dx.doi.org/10.1016/j.apcata.2016.12.025
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.

Paper submitted for publication in Applied Catalysis A: General
Cr/ZSM-5 for ethane dehydrogenation: enhanced catalytic
activity through surface silanol
Yanhu Cheng a, Changxi Miao b, Weiming Hua a, Yinghong Yue a,*, Zi Gao a
a Department of Chemistry, Shanghai Key Laboratory of Molecular Catalysis and Innovative
Materials, Fudan University, Shanghai 200433, PR China
b Shanghai Research Institute of Petrochemical Technology, Shanghai 201208, PR China
* Corresponding author. Tel.: +86 21 65642409; Fax: +86 21 65641740.
E-mail address: yhyue@fudan.edu.cn (Y. Yue)

Graphical abstract

Highlights
 Surface silanol rather than surface area of ZSM-5 is crucial for
CrOx dispersion.
 Silanol nests are more helpful in efficient CrOx dispersion than
terminal silanols.
 Better CrOx dispersion results in better activity in ethane
dehydrogenation.

ABSTRACT

Dehydrogenation of ethane to ethylene in the presence of CO2 was investigated over
CrOx catalysts supported on ZSM-5 with different crystallite sizes, including
nanometer, submicron and micrometer. These catalysts behave differently in the
dehydrogenation reaction, but no relationship between the activity and crystallite size
of ZSM-5 was found. XPS, UV-vis, laser Raman and HR-TEM studies show that the
better activity comes from better CrOx dispersion on ZSM-5 surface. The
27
characterization results of ZSM-5 supports by N2 adsorption, SEM, XRD, XRF, Al
29
and Si MAS NMR reveal that the difference of the dispersion is associated with
their surface properties, not with the textural ones. Silanols on the surface play an
important role in dispersing the CrOx active species. Silanol nests are more helpful in
efficient CrOx dispersion than terminal silanols, which subsequently result in higher
dehydrogenation activity.
Keywords: Ethane dehydrogenation; Cr oxide; Dispersion; ZSM-5; Crystallite size;
Silanol nests

1. Introduction
Ethylene is an important raw material in the petrochemical industry, which is
traditionally obtained from the pyrolysis of gaseous alkanes and liquid petroleum
products. Due to the rapidly growing demand for ethylene and the depletion of
petroleum resource, much effort has been dedicated to new routes of ethylene

production from alternative feedstocks, such as natural gas and shale gas.
Dehydrogenation of ethane into ethylene (DHE) is a promising way. Oxidation
dehydrogenation of ethane with O2 (ODHE) has been studied extensively for the past
decades with advantages of higher equilibrium conversion and lower energy
consumption [1-3]. However, application of this technology is still limited by low
ethylene selectivity because of the over-oxidation of ethane to carbon dioxide.
The utilization of carbon dioxide as a mild oxidant for the selective
dehydrogenation of ethane to ethylene (CO2-DHE) has drawn intense attention due to
the energy and environmental concerns [4-8]. Cr-based catalysts in particular showed
excellent activity for the dehydrogenation of ethane with CO2, which were usually
prepared by dispersing Cr species on supports with high surface areas [9-11]. These
supports included inorganic oxides like SiO2, ZrO2, Al2O3, TiO2 and porous materials
such as SBA-15, MCM-41, MSU-x, TS-1 [6-11, 12-14]. ZSM-5 has been widely used
in industrial processes either as a catalyst or support, owing to its three-dimensional
microporous structure, high surface area, and high thermal/hydrothermal stability.
Recently ZSM-5 supported metal oxide catalysts were investigated with regard to
ethane or propane dehydrogenation in the presence of CO2 [6-8]. H-form ZSM-5 with
a Si/Al ratio over 1900 was found to be preferred as the support for chromium oxides
in the dehydrogenation of ethane [6]. The amount of Al incorporated in the framework
of ZSM-5 was thought to have an impact on the valence state of chromium species.
Cr6+ became the dominant state and the cycle between Cr6+ and Cr3+ in the presence
of CO2 operated well when Si/Al ratio is above 1900. Our previous work showed that

Na-form of the ZSM-5 support favored CrOx dispersion [8] in comparison to HZSM-5,
thus, better activity was obtained over Cr/NaZSM-5 in the CO2-DHE reaction though
their selectivities as well as stabilities are almost the same. The influence of acid
characteristics on catalytic performance is not obvious since the dehydrogenation over
Cr-based catalysts is thought to proceeds through a redox mechanism. Cr loadings, Cr
precursors and the way of Cr introduction also affected the dispersion and the state of
the Cr species on the support, resulting in different catalytic behaviors [11, 15-19].
More recently, Cr supported on Na-type ZSM-5 having a smaller crystal size (ca. 400
nm) was found to be more effective when applied to the dehydrogenation of propane
in the presence of CO2, although the exact reason was not clear [7]. The above
catalyst also exhibited superior activity in the oxidative dehydrogenation of ethane
with CO2, and the promotional effect of CO2 on the dehydrogenation reaction was
quite evident over this catalyst [8].
In this contribution, CrOx supported catalysts on NaZSM-5 with different crystallite
sizes (including nanometer, submicron and micrometer) were prepared and
characterized by SEM, N2 adsorption, IR, XRF, 27Al MAS NMR and 29Si MAS NMR.
The nature of the supported chromium species was characterized by HR-TEM, TPR,
XPS, UV-vis and laser-Raman. The catalytic performance of these catalysts for
dehydrogenation of ethane to ethylene in the presence of CO2 was compared and
discussed in relation to their textural and physicochemical properties as well as the
supported CrOx species.

2. Experimental
2.1. Catalyst preparation
Submicron ZSM-5 zeolites with different sizes were prepared following the
procedure in the literature [20]. Typically, NaAlO2 was dissolved in an aqueous
tetrapropylammonium hydroxide (TPAOH) solution, and then tetraethylorthosilicate
(TEOS) was added. The resulting mixture was stirred for 6 h at room temperature,
followed by heating at 50 °C with stirring to evaporate the ethanol resulting from the
reaction. A clear gel was obtained with the molar composition 120 SiO2 : 0.375 Al2O3 :
120x TPAOH : 3600 H2O (x = 0.15, 0.3 and 0.4). This gel was transferred to an
autoclave and crystallized by heating at 170 °C for 2 d. ZSM-5 zeolite with
micrometer size was synthesized according to another literature [21]: Na2SiO3·9H2O,
NaAlO2 and tetrapropylammonium bromide (TPABr) were dissolved in water. Diluted
H2SO4 aqueous solution was then added slowly under stirring until the pH value is
near 9.0. The gel with molar composition of 120 SiO2 : 0.375 Al2O3 : 48 TPABr :
19200 H2O was crystallized by heating at 170 °C for 2 d. To prepare nanosize ZSM-5
zeolite, TPAOH was first added to the mixture of TEOS and water, then stirred at
80 °C for 24 h. NaAlO2 aqueous solution was dropped into the above mixture to reach
the molar composition of 120 SiO2 : 0.375 Al2O3 : 30 TPAOH : 960 H2O. After stirred
for another 1 h, the gel was transferred to an autoclave and crystallized for 24 h at
170 °C. All obtained as-synthesized samples were dried at 110 °C overnight and then
calcined in air at 600 °C for 6 h to remove the template. They were turned into

Na-form by three consecutive ion exchanges with 1 M NaNO3 aqueous solution for a
total of 12 h. The sub-microsize, microsize and nanosize ZSM-5 zeolites are labelled
as ZSM-5-S-x, ZSM-5-L and ZSM-5-N, respectively. (x represents the molar ratio of
TAPOH/SiO2).
(NH4)2SiF6 treated ZSM-5-S-0.4 sample was prepared as follow: ZSM-5-S-0.4 was
first turned into the NH4 form by ion-exchange with 1 M NH4NO3 aqueous solution
for 12 h, and then treated with a solution of (NH4)2SiF6 (AHFS) in order to insert Si
atoms into the framework defects [22]. After washing, it was turned into the Na-form
again by ion-exchange with 1 M NaNO3 and the final sample was named
ZSM-5-S-0.4-Si.
The supported chromium oxide catalysts were prepared by impregnating an
aqueous solution of Cr(NO3)3·9H2O on Na-form ZSM-5 using an incipient wetness
method. The impregnated samples were dried at 110 °C overnight and calcined in air
at 650 °C for 6 h. The obtained catalysts are denoted as Cr/ZSM-5-S-x, Cr/ZSM-5-L
and Cr/ZSM-5-N, respectively. The loading amount of chromium species in all
catalysts was 3 wt %.
2.2. Catalyst characterization
X-ray diffraction (XRD) patterns were recorded on a Persee XD-2 X-ray
diffractometer using nickel-filtered Cu Kα radiation at 40 kV and 30 mA. The BET
surface areas and pore volumes of the catalysts were analyzed by N2 adsorption at
-196 °C using a Micromeritics ASAP 2000 instrument. Bulk Si/Al ratios of the

prepared ZSM-5 samples were measured by X-ray fluorescence (XRF) on a
Bruker-AXS S4 Explorer instrument. Scanning electron microscopy (SEM) was
recorded digitally on a FEI Nova NanoSem 450 microscope operating at 40 kV. 2-D
atomic mapping were recorded on an FEI Tecnai G2 F20 S-TWIN instrument.
Laser Raman spectra were recorded on a Horiba JY XPloRA spectrometer using the
532-nm radiation of air cooled solid state laser as an excitation source. The laser
power was 25 mW. The spectra were obtained at room temperature under ambient
conditions with acquire time of 30 s and accumulation number of 8. The spectral
resolution was 2 cm−1. Diffuse reflectance UV–vis spectra (DRS) were collected on a
Shimadzu UV-2450 spectrometer equipped with an integrating sphere attachment.
X-ray photoelectron spectroscopy (XPS) measurements were carried out on a Perkin–
Elmer PHI 5000C spectrometer with Mg Kα radiation as the excitation source. All
binding energy values were referenced to the C1s peak at 284.6 eV.
27Al and 29Si MAS NMR measurements were performed on a Bruker DSX
27
spectrometer. For Al MAS NMR spectra a resonance frequency of 104.3 MHz, a
recycle delay of 0.5 s, short 0.8 s pulses, a spectral width of 62.5 kHz and a spin rate
29
of 12 kHz were applied, whereas for Si MAS NMR spectra a resonance frequency
of 79.5 MHz, a recycle delay of 3 ms, 3.5 s pulses, a spectral width of 32.05 kHz and
3+
a spin rate of 4 kHz were applied. External Al(H2O)6 and tetramethylsilane were
used as references.
DRIFTS data were recorded on an FTIR (Nicolet 6700) equipped with an MCT
detector cooled by liquid N2. The DRIFTS cell was fitted with CaF2 windows and a

heating cartridge. Samples were activated in the He atmosphere at 450 °C for 4 h, and
cooled to 300 °C under flowing He (50 ml/min), and treated in flowing N2 (15 ml/min)
for 20 min before the scanning. Spectra were collected at 300 °C in order to exclude
the impact of water on the samples under flowing He, with a resolution of 4 cm−1 and
accumulation of 32 scans.
Temperature-programmed reduction (TPR) profiles were obtained on a
Micromeritics AutoChem II apparatus loaded with 100 mg of catalyst. The
experiments were carried out in 10 vol% H2/Ar flowing at 30 ml/min, with a ramping
rate of 10 °C/min. H2 consumption was monitored using a thermal conductivity
detector.
2.3. Catalyst testing
Catalytic tests for ethane dehydrogenation in the presence of CO2 were performed
at 650 °C in a fixed-bed flow microreactor at atmospheric pressure. The catalyst load
was 200 mg, and it was pretreated at 650 °C for 2 h in nitrogen flow prior to the
reaction. The gas reactant contained 3 vol% ethane, 15 vol% carbon dioxide and
balancing nitrogen. The total flow rate of the gas reactant is 30 ml/min. The
hydrocarbon reaction products were analyzed using an on-line GC equipped with a
6-m packed column of Porapak Q and a FID. The gas products were analyzed on-line
by another GC equipped with a TCD and a carbon molecular sieve 601 column. The
reaction data in the work were reproducible with a precision of less than 5%.

3. Results and Discussion
3.1. Textural characterization
Five ZSM-5 supports with different crystallite sizes have been prepared and their
XRD patterns are shown in Fig. 1. All the samples exhibit well crystallized MFI
structure, with the typical characteristic reflections of 2θ = 8.0, 8.9, 23.1, 23.4, 24.0.
No peaks corresponding to the chromium oxide are found after the CrOx supported
(not shown in the text), indicating the chromium oxide is well dispersed on all these
ZSM-5 supports. SEM images of the ZSM-5 supports are presented in Fig. 2. All
samples are well crystallized without any presence of amorphous materials and
exhibit a uniform crystallite size distribution. However, the average crystallite sizes
are quite different, from 80 nm to 15 μm, which are in the range of nanometer,
submicron meter and micrometer, respectively.
The textural properties of the ZSM-5 supports are quite similar, which are listed in
Table 1. The specific surface areas (Stotal) increase a little with the decrease of
crystallite size as well as the increase of external surface (Sexternal). The Si/Al ratios of
the ZSM-5 supports are also close to each other, as confirmed by XRF measurement.
3.2. State of Cr species
27Al MAS NMR spectra were acquired to characterize the local coordination
environment of the aluminium atoms in these five zeolites. There is only one intense
line at 55 ppm assigned to the ZSM-5 framework aluminium atoms in tetrahedral

coordination appeared in the spectra, as shown in Fig. S1. No discernable signal at 0
ppm, attributed to extra-framework aluminium atoms in octahedral coordination, is
observed, indicating that all the Al atoms in the gels have been incorporated in the
framework of MFI structure.
XPS was employed to investigate the oxidation states of Cr species, and the valence
of Cr was thought to play an important role in the Cr-based catalysts in the
dehydrogenation reaction [12-16]. The deconvoluted spectra in the Cr (2p) region are
illustrated in Fig. S2. Two peaks at around 576.5 and 579.5 eV associated to Cr3+ and
Cr6+ can be observed and the quantitative data after fitting were listed in Table 2. All
the catalysts are found to have similar binding energy (BE) values. However, the
Cr6+/Cr3+ ratio is quite different. The catalyst taking ZSM-5-S-0.4 as the support had
the highest Cr6+/Cr3+ ratio, while the lowest Cr6+/Cr3+ ratio was observed on the
ZSM-5-L supported one.
The above results were confirmed by H2-TPR tests. The results are presented in Fig.
3 and Table 2. There is one reduction peak around at 350~370 °C, with a shoulder at
250~270 °C, that can be attributed to the reduction of Cr6+ to Cr3+ (except for
Cr/ZSM-5-L). Another small peak at 469 °C emerged on the spectrum of Cr/ZSM-5-L,
which can also be corresponded to the reduction of Cr6+, caused by the chromium
species in a hidden position in the support matrix [19, 23]. H2 consumption decreases
in the order of Cr/ZSM-5-S-0.4 > Cr/ZSM-5-S-0.15 > Cr/ZSM-5-S-0.3 >
Cr/ZSM-5-N > Cr/ZSM-5-L, indicating the amount of redox Cr(VI) species decreases
following the above sequence, which is in accordance with XPS results.

Additional information on the state of Cr species can be obtained from UV–vis
diffuse reflectance measurements, and the results are illustrated in Fig. 4. Two peaks
centered at 272 and 370 nm are observed for all the catalysts, which can be assigned
to O2- → Cr6+ charge transfer transition of monochromate species in tetrahedral
coordination [12, 14-17, 24]. The peak at around 468 nm is also associated with Cr6+,
which is not so obvious for Cr/ZSM-5-S-0.4 catalyst. It is believed that the intensity
of 468 nm peak increases with the degree of the oligomerization in the chromium
species. The fourth peak at 605 nm was found on all the catalysts except the
Cr/ZSM-5-S-0.4 one, which resulted from the typical octahedral symmetry transition
of tiny crystalline chromium oxides [12, 14-17, 24]. It can also be seen from Fig. 4
that the intensity of the peaks at 468 and 605 nm have the same sequence of
Cr/ZSM-5-L
>
Cr/ZSM-5-N
>
Cr/ZSM-5-S-0.3
>
Cr/ZSM-5-S-0.15
>
Cr/ZSM-5-S-0.4, suggesting that Cr species polymerization tended to increase in the
order of Cr/ZSM-5-L > Cr/ZSM-5-N > Cr/ZSM-5-S-0.3 > Cr/ZSM-5-S-0.15 >
Cr/ZSM-5-S-0.4. In other words, worse dispersion of chromium was observed on
ZSM-5-L and ZSM-5-N supports as compared with ZSM-5-S ones, and the best
dispersion of chromium was obtained on the support ZSM-5-S-0.4.
Laser Raman spectra were acquired to identify the molecular nature of the Cr
species dispersed on the supports, and the results were depicted in Fig. 5. A band at
550 cm–1, assigned to crystalline Cr2O3, appeared in all the spectra except that of
Cr/ZSM-5-S-0.4 [10, 15, 23-26]. The intensity of this band decreased in the order of
Cr/ZSM-5-L
>
Cr/ZSM-5-N
>
Cr/ZSM-5-S-0.3
>
Cr/ZSM-5-S-0.15
>

Cr/ZSM-5-S-0.4. Another two intense bands at approximately 980 and 1000 cm–1
were present in the spectra of each of the catalysts, which can be ascribed to isolated
and polymeric chromium oxide species, respectively [10, 15, 23-26]. But their
intensities showed a contrary order. Meanwhile, a weak and broad band at 810 cm–1,
attributed to the bending mode of the Cr-O-Cr linkage of polymeric chromates, also
appeared. It can be concluded that the dispersed Cr(VI) species on the catalyst
surfaces were gradually transformed from monochromates and polymeric chromates
into crystalline Cr2O3 in the order of Cr/ZSM-5-S-0.4, Cr/ZSM-5-S-0.15,
Cr/ZSM-5-S-0.3, Cr/ZSM-5-N, Cr/ZSM-5-L, which was in accordance with the UV–
vis results.
An analytical HR-TEM with large angel detection was used for the 2-D atomic
mapping, which helped to confirm the level of aggregation of chromium species in the
samples (shown in Fig. 6). It was clear that chromium oxide particles with large size
aggregated closely in Cr/ZSM-5-L catalyst, while no aggregated form but the
scattered distribution of chromium species were observed in Cr/ZSM-5-S-0.4. The
situation for the other two Cr/ZSM-5-S samples and Cr/ZSM-5-N was in the middle.
The 2-D mapping results again confirm the differences of molecular nature of
chromium species on the five ZSM-5 supports with different crystallite sizes, which
was obtained from XPS, UV-vis and laser Raman measurements.
3.3. Surface characters of support
Fig. 7 depicted the 29Si MAS NMR spectra of the ZSM-5 supports. Only one broad

resonance peak was observed for all the samples, which can be deconvoluted into
multiple peaks by Gaussian fitting, indicating the existence of Si atoms with different
chemical environments in the framework. According to the previous studies, peaks
with chemical shift below -110 ppm was usually assigned to siloxane Q4[(OSi)4]
while peaks at around -106 ppm and -102 ppm was assigned to siloxane
Q3[(AlO)Si(OSi)3] and Q3[HOSi(OSi)3], respectively [27-30]. Chemical shift above
-100 ppm was not common to be observed in high siliceous zeolite materials. As can
be seen from the spectra, all samples showed a detailed correspondence in the number,
shifts and intensities of the peaks, indicating they have identical structures. However,
the proportion of the Q3[HOSi(OSi)3] among the five sites was quite different, which
decreased in the order of ZSM-5-N (4.1%) > ZSM-5-S-0.4 (3.6%) > ZSM-5-S-0.15
(2.6%) > ZSM-5-S-0.3 (2.1%) > ZSM-5-L (1.0%), showing that the amount of the
silanols on surface changed with the crystallite sizes of the support. However, no
obvious relationship between the amount of silanols and crystalline sizes is observed.
A further investigation was performed by infrared spectroscopy in order to get a
deep understanding of hydroxyl groups on ZSM-5 surface, which are displayed in Fig.
8 by comparing the IR spectra of the samples in the hydroxyl stretching region (3800
~ 3200 cm–1). More detailed information can be seen from the deconvoluted spectra
shown in Fig. S3. Three distinct peaks at 3740 cm–1, 3720 cm–1 and 3500 cm–1 were
presented in each of the supports. The narrow peak at 3740 cm–1 was involved with
the terminal silanol groups usually on the exterior of the crystal, while the peak
centered at about 3720 cm–1 was assigned to the vicinal silanols and the broad peak

around 3500 cm–1 was in relation to the silanol nests [31-36]. The band at 3610 cm–1
associated with O–H vibrations of strongly acidic bridging Si-OH-Al groups in
zeolites was not observed in any of the spectra, indicating a thorough exchange of
protons by Na cations during the ion exchange process. The relative intensities of
peaks of terminal silanol groups decreased in the order of ZSM-5-N > ZSM-5-S-0.3 >
ZSM-5-S-0.4 > ZSM-5-S-0.15 > ZSM-5-L, contrary to that of the crystallite size,
which was in consistent with the literature results that zeolite with larger crystal size
had smaller external surface areas and weaker band at 3740 cm–1 [31, 35]. The
intensity of the peak at 3720 cm–1 and 3500 cm–1 had a different sequence, i.e.
ZSM-5-S-0.4 > ZSM-5-S-0.15 > ZSM-5-S-0.3 > ZSM-5-L > ZSM-5-N, meaning that
ZSM-5-S-0.4 possessed largest amount of vicinal and nest silanols. This can be
confirmed by the ratio of A3500/A3740 listed in Table 1. It should be reminded that
ZSM-5-L showed a much higher ratio of A3500/A3740 because its amount of terminal
silanol was so low. In a word, the total amount of the silanols as well as silanol
distribution on these ZSM-5 supports is clearly different from each other although
they have the same topology and similar texture properties.
3.4. Catalytic performance
The prepared Cr/ZSM-5 catalysts were evaluated during ethane dehydrogenation in
the presence of CO2, and the results are shown in Table 3. The major product formed
in the reaction is ethylene, and the minor products are methane and propylene. Mass
transfer and heat transfer limitations can be excluded according to the Weisz-Prater

analysis and Mears analysis (shown in the supporting information) [37-38]. It is
interesting to note that these Cr/ZSM-5 catalysts behave quite differently in the
reaction and the catalyst with the smallest crystallite size do not show the best activity.
The initial conversion of ethane on the catalysts decreased in the order of
Cr/ZSM-5-S-0.4 > Cr/ZSM-5-S-0.15 > Cr/ZSM-5-S-0.3 > Cr/ZSM-5-N >
Cr/ZSM-5-L, which has no obvious relationship with the crystallite size of ZSM-5.
For all the catalysts, the ethane conversion dropped gradually with reaction time, but
the decreasing rates were also quite different. Cr/ZSM-5-S samples were more stable
than Cr/ZSM-5-N and Cr/ZSM-5-L. Cr/ZSM-5-S-0.4 was the most stable one, only
6.4% of its initial activity was lost during the first 6 h, while for Cr/ZSM-5-N and
Cr/ZSM-5-L, the dropping data were 26.2% and 34.8%, respectively, during the same
period. In general, ZSM-5-S samples served better as the supports for chromium
oxide than ZSM-5-L and ZSM-5-N when they were applied to the catalytic
dehydrogenation of ethane reaction in the presence of CO2. Particularly,
Cr/ZSM-5-S-0.4 exhibited the highest activity and the best stability.
Cr-based catalysts have been documented as one of the most promising catalysts
for light alkane dehydrogenation reaction due to their high catalytic efficiency either
in the absence or in the presence of CO2. Two types of coordinatively unsaturated
Cr(III) species are generally considered as the active sites, one is the Cr(III) ions
formed from the reduction of Cr(VI) and the other is dispersed Cr(III) ions on fresh
catalysts. The former is thought to be more active. As a result, the amount of reducible
Cr(VI) on the calcined samples is crucial for the dehydrogenation reaction [10, 13, 24,

39-41]. H2 consumption allows one to estimate the amount of redox Cr(VI) species. It
can be seen from TPR results that the amount of reducible Cr(VI) on fresh catalysts
has the same sequence as that of dehydrogenation activity, confirming that redox
Cr(VI) makes quite an important contribution to this reaction. Because of the same Cr
loadings, Cr6+/Cr3+ ratio obtained by XPS (Table 2) can also prove the above
conclusion that the catalyst with the largest amount of surface Cr(VI) species (i.e.
highest Cr6+/Cr3+ ratio) exhibits the highest dehydrogenation activity.
The higher amount of reducible Cr(VI) species on Cr/ZSM-5 catalyst comes from
better CrOx dispersion on its surface. In other words, a high dispersion of CrOx
species, which can avoid the formation of less active crystalline chromium oxide
(Cr2O3), is needed for the high dehydrogenation activity. As evidenced by UV-vis in
Fig. 4, Cr species polymerization tends to increase in the order of Cr/ZSM-5-L >
Cr/ZSM-5-N > Cr/ZSM-5-S-0.3 > Cr/ZSM-5-S-0.15 > Cr/ZSM-5-S-0.4, transforming
the monochromate and polychromate species (Cr(VI)) to the tiny crystalline Cr2O3
(Cr(III)). This is why the intensity of the peaks at 605 nm corresponding to Cr(III)
species in Cr2O3 has the order of Cr/ZSM-5-L > Cr/ZSM-5-N > Cr/ZSM-5-S-0.3 >
Cr/ZSM-5-S-0.15 > Cr/ZSM-5-S-0.4. This transformation would reduce the amount
of active Cr(VI) species on the surface. The same conclusion can be obtained from the
1
laser Raman spectra in Fig. 5. Peak at 550 cm– assigned to crystalline Cr2O3 was very
sharp for Cr/ZSM-5-L catalyst while it was weak for Cr/ZSM-5-N, Cr/ZSM-5-S-0.3
and Cr/ZSM-5-S-0.15. Meanwhile, no obvious peak corresponding to crystalline
Cr2O3 could be observed on Cr/ZSM-5-S-0.4 catalyst, meaning dispersion increases in

the order of Cr/ZSM-5-S-0.4 > Cr/ZSM-5-S-0.15 > Cr/ZSM-5-S-0.3 > Cr/ZSM-5-N >
Cr/ZSM-5-L. The differences of chromium oxide dispersion over Cr/ZSM-5 can also
be proved by the 2-D atomic mapping shown in Fig. 6.
The above results are a bit surprising since they are quite different from the
previous observation that better dispersion is often obtained on supports of smaller
crystallite size, which has relative higher specific surface area. All the supports used
here have similar structure and composition. Their surface areas are in the order of
ZSM-5-N > ZSM-5-S-0.15 > ZSM-5-S-0.4 > ZSM-5-S-0.3 > ZSM-5-L, which is not
consistent with the dispersion abilities. Cavani once reported that the nature of the
support, mainly the surface hydroxyl chemistry, was found to affect the nature of the
degree of oligomerization of chromium species, as well as the ability to stabilize Cr6+
at high temperature [42]. Weckhuysen held the view that the dispersion of Cr
depended on the availability of the reactive surface hydroxyl groups since chromium
species were anchored to the surface of SiO2 by esterification with the hydroxyl
groups of the support [15]. Quite similar dispersing state of CrOx was obtained on
SBA-15 materials with different surface areas but the same amount of silanols in our
recent studies [41]. Thus, similar phenomenon may occur during dispersing the CrOx
29
over ZSM-5 supports. Si MAS NMR spectra of five ZSM-5 supports (depicted in
Fig. 7) show that the proportion of the Q3[HOSi(OSi)3] decrease in the sequence of
ZSM-5-N >ZSM-5-S-0.4 > ZSM-5-S-0.15 > ZSM-5-S-0.3 > ZSM-5-L, which is in
the same order of the decreasing dispersion ability for CrOx except ZSM-5-N. The
intensity of Q3 peaks decreased after the chromium species loading, and a new peak at

29
about -107 ppm appeared. In Dzwigaj’s work, a new peak observed in the Si NMR
spectrum of β zeolites with the loading of transition metal V species was taken as an
evidence for the incorporation of V species into the framework and forming the Si
(3Si, 1V) sites [43]. Hamdy also found a new peak in their 29Si NMR spectrum when
Cr was introduced to TUD-1 zeolite, which was thought to arise from the formation of
the sites Si (3Si, 1Cr) [44]. Thus, it seems reasonable that the new peak at around
-107 ppm can be attributed to the formation of Si-O-Cr bond by the reaction between
the supported Cr species and the silanol groups on the surface. As a result, surface
silanol play an important role in CrOx dispersion.
More detailed results about the surface silanol groups can be obtained by IR
measurements. As depicted in Fig. 8, there exist three kinds of silanol groups on the
surface, including terminal, vicinal and nest ones. Terminal hydroxyl was once
thought to be responsible for the interaction between chromium species and ZSM-5
support [15, 45]. On the contrary, Zecchina et al believed that nests of hydroxyl
groups inside silicalite molecular sieves can react with chromic acid to give anchored
chromates [44-47]. Dzwigaj et al did a lot of work on creating vacant-T sites (nests of
silanols) by post-synthesis method and believed that such vacant-T sites were much
helpful in the efficient dispersion of transition metals such as V、Cr、Mo on β zeolite
[43, 48-50]. In our present work, both of terminal and nest silanols are thought to be
responsible for the dispersion, since the intensities of both IR bands decreased after
the modification with CrOx (see Cr/ZSM-5-S-0.4 in Fig. 8). Besides, better dispersion
was obtained on ZSM-5-N supports as compared with ZSM-5-L, though the former is

short of nest silanols. On the other hand, dispersion of chromium on ZSM-5-N
support was worse than that on ZSM-5-S supports despite of its higher amount of
silanols. Considering their difference in silanol distributions, silanol nests might be
more helpful in the efficient dispersion of chromium species.
To confirm the crucial role of nest silanol on the dispersion of supported chromium
oxide, (NH4)2SiF6 treated ZSM-5-S-0.4 sample was prepared. As shown in Table 1,
the textural properties changed little after the modification. However, IR
measurements (ZSM-5-S-0.4-Si in Fig. 8) show that the intensity of the broad peak at
3500 cm–1 decreased obviously after modification while the peak at 3740 cm–1
increased a little, indicating a large number of silanol nests were consumed by the
insertion of Si atoms into the zeolite framework. The dispersion of CrOx species on
ZSM-5-S-0.4 became worse after this treatment, which could be demonstrated by the
fact that crystalline Cr2O3 were observed in the Raman spectrum at 550 cm–1 (see
Cr/ZSM-5-S-0.4-Si in Fig. 5). The dehydrogenation activity also decreased greatly
because of the worse dispersion, and the initial activity drops to only one-third of the
data before the treatment. This confirms that nest silanols play more important role in
the CrOx dispersion and finally determine the dehydrogenation activity.
4. Conclusions
ZSM-5 with different crystallite sizes were synthesized and characterized. The
catalytic dehydrogenation of ethane to ethylene in the presence of CO2 was carried
out over ZSM-5 supported CrOx catalysts. The results showed that chromium oxide

supported on submicron ZSM-5-S-0.4 exhibited the best activity and stability. The
excellent catalytic behaviour comes from the improved dispersion of CrOx species on
the surface of ZSM-5-S-0.4, resulting in a large quantity of the reducible Cr(VI)
species that act as active species in the ethane dehydrogenation reaction. Silanol
groups play an important role in the dispersion of the supported CrOx species.
Terminal and nest silanol groups were both responsible for the introduction of
chromates onto the ZSM-5 supports, whereas the latter is more efficient and helpful.
An active zeolite supported catalyst can be designed after understanding the
interaction between the zeolites and the species supported.
Acknowledgements
This work was supported by the National Natural Science Foundation of China
(21273043) and the Science & Technology Commission of Shanghai Municipality
(13DZ2275200).
Reference
[1] M. Cherian, M.S. Rao, W.T. Yang, J.M. Jehng, A.M. Hirt, G. Deo, Appl. Catal. A.
233 (2002) 21–33.
[2] J. Janas, J. Gurgul, R.P. Socha, J. Kowalska, K. Nowinska, T. Shishido, M. Che,
S. Dzwigaj, J. Phys. Chem. C 113 (2009) 13273–13281.
[3] B. Grzybowska, J. Słoczynski, R. Grabowski, L. Keromnes, K. Wcisło, T.

Bobinska, Appl. Catal. A: Gen 209 (2001) 279–289.
[4] K. Takehira, Y. Ohishi, T. Shishido, T. Kawabata, K. Takaki, Q.H. Zhang, Y.
Wang, J. Catal. 224 (2004) 404–416.
[5] X.N. Ye, Y.H. Yue, C.X. Miao, Z.K. Xie, W.M. Hua, Z. Gao, Green Chem. 7
(2005) 524–528.
[6] N. Mimura, M. Okamoto, H. Yamashita, S. Ted Oyama, K. Murata, J. Phys.
Chem. B 110 (2006) 21764-21770.
[7] F. Zhang, R.X. Wu, Y.H. Yue, W.M. Yang, S.Y. Gu, C.X. Miao, W.M. Hua, Z.
Gao, Microporous Mesoporous Mater. 145 (2011) 194–199.
[8] Y.H. Cheng, F. Zhang, Y. Zhang, C.X. Miao, W.M. Hua, Y.H. Yue, Z. Gao, Chin.
J. Catal. 36 (2015), 1242–1248.
[9] T.V. M. Rao, E. M. Zahidi, A. Sayari, J. Mol. Catal. A: Chem 301 (2009) 159–
165.
[10] Y. Wang, Y. Ohishi, T. Shishido, Q.H. Zhang, W. Yang, Q. Guo, H.L. Wan, K.
Takehira, J. Catal. 220 (2003) 347–357.
[11] S.B. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki, Appl. Catal. A:
Gen 196 (2000) 1–8.
[12] T. Shishido, K. Shimamura, K. Teramura, T. Tanaka, Catal. Today 185 (2012)
151–156.
[13] J. Baek, H.J. Yun, D. Yun, Y. Choi, J. Yi, ACS Catal. 2 (2012) 1893–1903.
[14] P. Michorczyk, J. Ogonowski, K. Zenczak, J. Mol. Catal. A: Chem. 349 (2011)
1–12.

[15] B.M. Weckhuysen, I.E. Wachs, R.A. Schoonheydt, Chem. Rev. 96 (1996) 3327–
3349.
[16] M.S. Kumar, N. Hammer, M. Rønning, A. Holmen, D. Chen, J.C. Walmsley, G.
Øye, J. Catal. 261 (2009) 116–128.
[17] M. Kocon, P. Michorczyk, J. Ogonowski, Catal. Lett. 101 (2005) 53–57.
[18] L.C. Liu, H.Q. Li, Y. Zhang, Catal. Today, 115 (2006) 235–241.
[19] F. Ayari, M. Mhamdi, T. Hammedi, J. Álvarez-Rodríguez, A.R. Guerrero-Ruiz,
G. Delahay, A. Ghorbel, Appl. Catal. A: Gen. 439–440 (2012) 88–100.
[20] A. Zatta, M.R. Rabaioli, P. Radici, R. Aiello, F. Crea, P. Frontiera, G. Giuffrida,
P. Frontera, WO Patent 048188 (2006).
[21] Y. Wang, L.F. Guo, Y. Ling, Y.M. Liu, X.H. Li, H.H. Wu, P, Wu, Appl. Catal. A:
Gen. 379 (2010) 45–53.
[22] S. Schallmoser, T. Lkuno, M.F. Wagenhofer, R. Kolvenbach, G.L. Haller, M.
Sanchez-Sanchez, J.A. Lercher, J. Catal. 316 (2014) 93–102.
[23] R.L. Puurunen, B.M. Weckhuysen, J. Catal. 210 (2002) 418–430.
[24] T.V.M. Rao, G. Deo, J.M. Jehng, I.E. Wachs, Langmuir, 20 (2004) 7159–7165.
[25] B.M. Weckhuysen, J.M. Jehng, I.E. Wachs . J. Phys. Chem. B 104 (2000) 7382–
7387.
[26] F. Ayari, M. Mhamdi, J. Álvarez-Rodríguez, A.R. Guerrero Ruiz, G. Delahay, A.
Ghorbel, Appl. Catal. B: Environmental 134–135 (2013) 367–380.
[27] M. Hunger, D. Freude, H. Pfeifer, W. Schwieger, Chem. Phys. Lett. 167 (1990)
21–26.

[28] J.M. Chezeau, L. Delmotte, J.L. Guth, Z. Gabelica, Zeolites, 11 (1991) 598–
606.
[29] M. Inui, T. Ikeda, T. Suzuki, K. Sugita, F. Mizukami, Bull. Chem. Soc. Jpn. 82
(2009) 1160–1169.
[30] C.A. Fyfe, G.C. Gobbi, G.J. Kennedy, J. Phys. Chem. 88 (1984) 3248–3253.
[31] T. Armaroli, L.J. Simon, M. Digne, T. Montanari, M. Bevilacqua, V. Valtchev, J.
Patarin, G. Busca, Appl. Catal. A. 306 (2006) 78–84.
[32] G.L. Qin, L. Zheng, Y.M. Xie, C.C. Wu, J. Catal. 95 (1985) 609–612.
[33] R.M. Dessau, K.D. Schmitt, G.T. Kerr, G.L. Woolery, L.B. Alemany, J. Catal.
104 (1987) 484–489.
[34] R.M. Dessau, K.D. Schmitt, G.T. Kerr, G.L. Woolery, L.B. Alemany, J. Catal.
109 (1988) 472–473.
[35] T. Aamaroli, M. Trombetta, A.G. Alejandre, J.R. Solis, G. Busca, Phys. Chem.
Chem. Phys. 2 (2000) 3341–3348.
[36] K. Barbera, F. Bonino, S. Bordiga, T.V.W. Janssens, P. Beato, J.Catal. 280 (2011)
196–205.
[37] D.E. Mears, Ind. Eng. Chem. Process Des. Dev., 10 (1971) 541.
[38] S.T. Oyama, X.M. Zhang, J.Q. Lu, Y.F. Gu and T. Fujitani, J. Catal. 257 (2008)
1–4.
[39] A. Kytokivi, J. Jacobs, A. Hakuli, J. Merilainen, H.H. Brongersma, J.Catal. 162
(1996) 190–197.
[40] M. Cherian, M. S. Rao, W.T. Yang, J.M. Jehng, A. M. Hirt, G. Deo, Appl. Catal.

A. 233 (2002) 21–33.
[41] Y.H. Cheng, L.B. Zhou, J.X. Xu, C.X. Miao, W.M. Hua, Y.H. Yue, Z. Gao,
Micropor. Mesopor. Mater. 234 (2016) 370–376.
[42] Cavani F, Koutyrev M, Trifiro F, Bartolini A, Ghisletti D, Iezzi R, Santucci A,
Del Piero G. J. Catal. 158 (1996) 236-250.
[43] S. Dzwigaj, M.J. Peltre, P. Massiani, A. Davidson, M. Che, T. Senb, S.
Sivasanke, Chem. Commun. (1998) 87–88.
[44] M.S. Hamdy, G. Mul, Appl. Catal. B: Environ. 174-175 (2015) 413-420.
[45] J. Gao, Y. Zheng, Y. Tang, J.M. Jehng, R. Grybos, J. Handzlik, I.E. Wachs, S.G.
Podkolzin, ACS Catal. 5 (2015) 3078–3092.
[46] A. Zecchina, G. Spoto, G. Ghiotti, E. Garrone, J. Mol. Catal. 86 (1994) 423–
436.
[47] G. Spoto, S. Bordiga, E. Garrone, G. Ghiotti, A. Zecchina, J. Mol. Catal. 74
(1992) 175–184.
[48] M. Trejda, M. Ziolek, Y. Millot, K. Chalupka, M. Che, S. Dzwigaj, J. Catal. 281
(2011) 169-176.
[49] S. Dzwigaj, T. Shishido, J. Phy. Chem. C 112 (2008) 5803-5809.
[50] R. Baran, F. Averseng, Y. Millot, T. Onfroy, S. Casale, S. Dzwigaj, J. Phy. Chem.
C 118 (2014) 4143-4150.

Table 1
Textural properties of ZSM-5 supports.
e
Samples
Si/Ala Si/Alb Stotal
m2/g
Sexteral
m2/g
167
96
V
Crystal sizesc A3500/A3740
cm3/g nm
ZSM-5-N
160
160
160
160
160
-
181
175
168
163
160
412
387
370
388
367
336
0.28
0.26
0.22
0.24
0.20
0.19
80
3
ZSM-5-S-0.15
ZSM-5-S-0.3
ZSM-5-S-0.4
ZSM-5-L
1000
300
19
8
100
103
56
400
30
700
-
15000
d
d
ZSM-5-S-0.4-Si
-
90
-
a from the synthesis gel
b from XRF
c from SEM
d not detected
e A3500 and A3740 referred to the peak area of silanol nests (3500 cm1) and terminal silanol (3740
cm1) after deconvolution, respectively.

Table 2
XPS and TPR data for various ZSM-5 supported catalysts.
Samples
Binding energy (eV)
Cr6+/Cr3+
atomic ratio mmol/g
H2 Consumption
Cr6+
Cr3+
Cr/ZSM-5-N
579.4
579.7
579.3
579.5
579.4
576.6
576.5
576.5
576.5
577.1
0.66
0.87
0.75
1.15
0.51
0.11
0.31
0.20
0.39
0.07
Cr/ZSM-5-S-0.15
Cr/ZSM-5-S-0.3
Cr/ZSM-5-S-0.4
Cr/ZSM-5-L

Table 3
Reaction data for ZSM-5 supported catalysts in the DHE in the presence of CO2.a
Samples
Conversion
(%)b
Selectivity(%)b
C2H4
Yield
(%)b
Carbon
balance
CH4
C3H6
Cr/ZSM-5-N
33.2(24.5)
56.3(48.9)
44.2(39.1)
65.5(61.3)
11.5(7.5)
8.9(4.9)
91.1(95.1) 0.0(0.0)
30.2(23.3) 96
46.5(42.7) 95
39.5(36.1) 97
49.4(48.4) 86
Cr/ZSM-5-S-0.15
Cr/ZSM-5-S-0.3
Cr/ZSM-5-S-0.4
Cr/ZSM-5-L
17.1(12.4) 82.6(87.4) 0.2(0.2)
10.4(7.4) 89.3(92.4) 0.2(0.1)
24.3(20.9) 75.4(78.9) 0.3(0.2)
3.7(4.7)
6.9(4.8)
96.3(95.3) 0.0(0.0)
93.1(95.2) 0.0(0.0)
11.1(7.2)
99
Cr/ZSM-5-S-0.4-Si 22.7(18.1)
21.2(17.2) 90
a Reaction conditions: 200 mg catalyst, 650 oC, 3% C2H6/15% CO2 in N2, flow rate 30 ml/min.
b The values out and in the brackets are obtained at 10 min and 6 h, respectively.