Facile one-pot synthesized ordered mesoporous Mg-SBA-15 supported PtSn catalysts for propane dehydrogenation

Article abstract of DOI:10.1016/j.apcata.2016.12.026

A series of high-surface area and well-ordered mesoporous Mg-incorporated SBA-15 (xMg-SBA-15, x?=?0, 1, 2, 4 and 6) was prepared using a facile one-pot hydrothermal synthesis method. The resulting materials were used to disperse Pt, Sn particles and further studied the catalytic performance for propane dehydrogenation. The physicochemical properties of the supports and catalysts were characterized by XRD, N₂adsorption-desorption, FT-IR, TEM, NH₃-TPD, XPS, H₂-TPD and TG-DTA techniques. Compared with the sample synthesized by the impregnation method, the xMg-SBA-15 materials showed a better Mg species distribution and higher specific surface area. It was found that incorporation of appropriate amounts of Mg promoted the dispersion of metallic nanoparticles and facilitated the transfer of the carbon deposits from the active metal to the support. Such a modification of support not only enhanced the interaction of the Sn-support, but also stabilized the oxidation state of Sn species, leading to the improvement of catalytic activity and stability. The significant increases in conversion and selectivity were obtained on the modified catalysts with appropriate content of Mg. The best catalytic performances were obtained on PtSn/2Mg-SBA-15 catalyst with 2?wt.% of magnesium, showing a higher than 98% selectivity toward propene at a 38.0% conversion of propane even after 6?h reaction.

Full text of DOI:10.1016/j.apcata.2016.12.026

Accepted Manuscript
Title: Facile one-pot synthesized ordered mesoporous
Mg-SBA-15 supported PtSn catalysts for propane
dehydrogenation
Author: <ce:author id="aut0005"
author-id="S0926860X16306305-
6
969984c0b08071abd30819afb605e63"> Bing Li<ce:author
id="aut0010" author-id="S0926860X16306305-
9c0a403870028c5fa6a19b821da21b2"> Zhenxin
0
Xu<ce:author id="aut0015"
author-id="S0926860X16306305-
db72363ea7fe04e7971fe9cc178cc62e"> Fangli
Jing<ce:author id="aut0020"
author-id="S0926860X16306305-
ea6430820f42aa85ae09a1de8ea94962"> Shizhong
Luo<ce:author id="aut0025"
author-id="S0926860X16306305-
ff5643b2ead50279d71d1cfcf10dc4b3"> Wei
Chu
PII:
S0926-860X(16)30630-5
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2016.12.026
APCATA 16111
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
18-8-2016
29-12-2016
31-12-2016
Please cite this article as: Bing Li, Zhenxin Xu, Fangli Jing, Shizhong Luo,
Wei Chu, Facile one-pot synthesized ordered mesoporous Mg-SBA-15 supported
PtSn catalysts for propane dehydrogenation, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2016.12.026
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.

Facile one-pot synthesized ordered mesoporous Mg-SBA-15
supported PtSn catalysts for propane dehydrogenation
a
b
a
a,
a,
Bing Li , Zhenxin Xu , Fangli Jing , Shizhong Luo , Wei Chu
a. Department of Chemical Engineering, Sichuan University, Chengdu 610065,
Sichuan, China.
b. Institut de Chimie et Procédés pour l'Energie, l'Environnement et la Santé
(
ICPEES), UMR 7515 CNRS-Université de Strasbourg (UdS), 25, rue Becquerel,
6
7087 Strasbourg Cedex 08, France.
*
Corresponding authors:
Shizhong Luo: luosz@scu.edu.cn
Wei Chu: chuwei1965@ scu.edu.cn
Tel.: +86 28-85403836; Fax: +86 28-85461108.
1

Graphical Abstract
2

Highlights
◆
A series of high-surface area and well-ordered mesoporous Mg-incorporated
SBA-15 (Mg-SBA-15) were prepared using a one-pot hydrothermal synthesis method.
Incorporation of magnesium modified the structural characteristics of the metallic
phase and strengthened the interactions of platinum, tin species and support.
The existence of Mg stabilized the oxidation states of Sn and facilitated the transfer
of the carbon deposits from the active metal to the support.
The significant improvements in conversion and selectivity were obtained on the
◆
◆
◆
catalyst incorporated with appropriate content of magnesium.
3

Abstract:
A series of high-surface area and well-ordered mesoporous Mg-incorporated
SBA-15 (xMg-SBA-15, x=0, 1, 2, 4 and 6) was prepared using a facile one-pot
hydrothermal synthesis method. The resulting materials were used to disperse Pt, Sn
particles and further studied the catalytic performance for propane dehydrogenation.
The physicochemical properties of the supports and catalysts were characterized by
XRD, N adsorption-desorption, FT-IR, TEM, NH -TPD, XPS, H -TPD and TG-DTA
2
3
2
techniques. Compared with the sample synthesized by the impregnation method, the
xMg-SBA-15 materials showed a better Mg species distribution and higher specific
surface area. It was found that incorporation of appropriate amounts of Mg promoted
the dispersion of metallic nanoparticles and facilitated the transfer of the carbon
deposits from the active metal to the support. Such a modification of support not only
enhanced the interaction of the Sn-support, but also stabilized the oxidation state of
Sn species, leading to the improvement of catalytic activity and stability. The
significant increases in conversion and selectivity were obtained on the modified
catalysts with appropriate content of Mg. The best catalytic performances were
obtained on PtSn/2Mg-SBA-15 catalyst with 2 wt.% of magnesium, showing a higher
than 98% selectivity toward propene at a 38.0% conversion of propane even after 6 h
reaction.
Keywords: Propane dehydrogenation; one-pot hydrothermal synthesis;
Mg-incorporated SBA-15; PtSn catalyst; structure-activity relationships;
4

1
. Introduction
As a kind of important industry material, propene has been widely implemented
in industrial processes for the production of polypropene, polyacrylonitrile, acrolein,
and acrylic acid etc. At present, the effective routes for producing propene mainly
contain steam cracking and fluidized catalytic cracking of naphtha.[1, 2] Since the
traditional production modes are unable to satisfy the increasing requirement for
propene, propane dehydrogenation as one of the most promising approach has gained
much attention.[3, 4] Dehydrogenation of low-cost hydrocarbons to corresponding
value-added olefins makes this route more economic dominance.[5] Propane
dehydrogenation is a highly endothermic and equilibrium limited reaction, meaning
higher temperature and lower pressure conditions are needed to accomplish a
considerable propene yield.[6] Unfortunately, these demanding conditions facilitate
side reactions such as cracking and coke deposition, resulting in the deactivation of
catalysts. Therefore, the development of new types of catalysts with high efficiency
under these severe reaction conditions is an important subject in this field.
Pt-based catalyst has been broadly applied in alkane dehydrogenation reaction,
owing to its excellent ability in activating C-H bonds and conserving simultaneously
the C-C bonds.[7, 8] However, carbon deposition and metallic sintering are still the
crucial issues in affecting the long term catalytic stability. Coke deposits derive from
polymerization of propene or carbon intermediates, which usually occurs at acidic
sites located on the support and metal nanoparticles.[9] To our best knowledge,
supported PtSn catalyst displays good catalytic performances in propane
dehydrogenation process and attracts widely attention. Generally, the properties of
PtSn catalysts closely related to the interaction between Pt and Sn, the chemical state
of Sn and the features of support.[10, 11] The specific functions of tin were expressed
as geometric and electronic modification of platinum phase.[12] It was proposed that
Sn promoter decreased the size of surface Pt ensembles, therefore restraining the deep
dehydrogenation reactions. On the other hand, the electronic effect of Sn can either
n+
favor the charge transfer from Sn to Pt or create a disparate electronic environment
5

in Pt-Sn alloys.[13] With regard to support, it plays a substantial role for the
dispersion of Pt, Sn active sites. PtSn catalysts supported on various materials such as
Al O , SiO , ZSM-5, and SAPO-34 have been extensively studied for
2
3
2
dehydrogenation of propane.[14-16] Recently, the SBA-15 shows its potential
application in catalysis as well as in propane dehydrogenation thanks to the well
defined and ordered mesoporous structure. Such unique features make SBA-15 very
well suited as host for the confinement and stabilization of Pt nanoparticles, which is
beneficial to obtaining higher dispersion of Pt nanoparticles and preventing particles
from agglomeration.[17] Therefore, SBA-15 supported PtSn catalysts may be a
promising prospect for the dehydrogenation of propane.
The acidity on the PtSn catalyst hinders the desorption of olefins product and
extend the retention time on catalyst surface, leading to the further conversion and
carbon deposition due to secondary dehydrogenation and isomerization reactions.[18]
The secondary dehydrogenation reaction must be suppressed to improve the
selectivity of the desired product propene. It is well established that improved propene
yield can be achieved by adding alkaline or alkaline earth metals to the catalyst
formulation for propane dehydrogenation.[19] A. Bliek et al.[20] reported that the
existence of alkali-earth metals in Pt-Sn catalyst could not only adapt the acid sites,
but also embellish the interaction between the support and the active sites. The
addition of Mg on PtSn/γ-Al O catalysts was described by Shi et al.[21], showing
2
3
that Mg promoter could improve the distribution of Pt particles and strengthen the
interaction among Pt, Sn, and support, leading to an enhanced catalytic performance.
6

Scheme 1. - Schematic synthesis procedures for the Mg-incorporated/supported
SBA-15 mesoporous materials
The introduction of alkaline metal and alkaline earth metal in the published
literature was mostly loaded through the post-synthesis procedure.[22] This process
needed two stages calcination to get the final product, which could cause a
non-uniform dispersion of Mg component on the support and a loss in the efficiency
of the Mg promoter. Moreover, the additive promoter might clog the mesoporous
channel thus decrease the surface area and pore size of SBA-15.[23] In the present
paper, we prepared the Mg-incorporated SBA-15 with different magnesium loadings
by one-pot method (Scheme 1) and used them as supports to disperse Pt, Sn particles.
The catalytic activity of these catalysts was tested by propane dehydrogenation.
Mg/SBA-15 sample was synthesized using the traditional impregnation method as the
referential sample. The differences between these samples on structure-activity
relationships were identified and discussed. This can provide us with important
information to identify the influence of magnesium promoter on the physicochemical
characteristics and catalytic properties of PtSn/SBA-15 in the propane
dehydrogenation.
2
. Experimental
7

2
.1. Catalyst preparation
The silica mesoporous support SBA-15 and Mg-incorporated SBA-15 were
synthesized using a triblock copolymer, Pluronic P123 (EO PO EO , Aldrich) as a
2
0
70
20
template according to the procedure reported in the literature.[24] The silica source
was the tetraethyl orthosilicate (TEOS, 98%). In a typical synthesis, 2 g of P123 was
first dissolved in 60 ml of 2M HNO and 15 ml of deionized water to form a clear
3
solution at room temperature. Then a calculated amount of magnesium acetate
(
Mg(CH COO) ) was added to the previous solution under stirring. After 3 h, 4.25 g
3 2
of TEOS was added dropwise to the solution and continuously stirred for 24 h at 40
o
C. The resulting suspension was placed into a Teflon-lined autoclave to undergo
o
hydrothermal treatment at 100 C for 24 h. Finally, the mixture was evaporated under
o
o
o
8
0 C with stirring and dried at 80 C for 12 h, then calcined at 550 C for 6 h to
remove the surfactant and release the pores. The received samples were denoted as
xMg-SBA-15 (the nominal mass fraction x = 0, 1, 2, 4 and 6) hereinafter. Another
sample was synthesized by the incipient-wetness impregnation method and denoted as
xMg/SBA-15-IM. First, the calcined SBA-15 powder was introduced into a calculated
amount of Mg(CH COO) aqueous solutions. After impregnation for 12 h, the final
3
2
o
mixture was evaporated under 80 C and then dried and calcined by the same
procedure as previously mentioned.
PtSn catalysts were prepared by incipient-wetness impregnation method.
SnCl ·2H O and H PtCl ·6H O precursors were dissolved in deionized water to form
2
2
2
6
2
a solution and the solution was impregnated with the support powder. Then, the
mixture was sonicated for 30 min and then under static condition for 12 h. After that,
o
o
the solution was dried at 80 C for 12 h and calcined at 500 C for 4 h. The nominal
compositions of these catalysts were 1 wt.% Pt and 2 wt.% Sn.
2
.2. Catalyst characterization
Low-angle XRD patterns were collected in the 2θ range of 0.5-10° with a step
size of 0.02° using an EMPYREAN diffractometer. Wide-angle XRD patterns were
obtained by an X-ray diffraction apparatus (Philips X’ pert PRO) using Cu Kα (45 kV,
5
0 mA) radiation.
8

o
The nitrogen adsorption-desorption isotherms were measured at -196 C on an
automated surface area & pore size analyzer (Quadrasorb SI apparatus). Before each
o
measurement, the samples were outgassed under vacuum at 300 C for 3 h.
The Fourier transform infrared (FT-IR) spectra were recorded on a Bruker Tensor
2
7 Fourier transform spectrometer at room temperature. The spectra were collected in
-1
the range of 4000-400 cm and KBr was used as background.
The X-ray photoelectron spectra (XPS) were achieved on the XSAM800
spectrometer with an Al Kα (hν = 1486.6 eV) X-ray source and the binding energies
were amended by using C (1 s) at 284.6 eV as an internal standard.
The temperature-programmed desorption of ammonia (NH -TPD) measurements
3
were implemented by a Micromeritics AUTOChem II2920 instrument to analyze the
acidity of the catalysts. The sample (100 mg) was pretreated at 250 °C for 2 h under a
helium flow (30 ml/min) before being cooled to 50 °C and saturated with adsorbed
ammonia. During desorption, a thermal conductivity detector (TCD) was employed to
record the TPD profiles from 50 to 900 °C with a heating rate of 10 °C /min.
The temperature-programmed desorption of hydrogen (H -TPD) was measured
2
over a Quantachrome Autosorb-1 instrument equipped with a TCD detector. The
catalyst (100 mg) was first reduced in situ in a 10 % H /N mixture gas flow of 30
2
2
ml/min, with a 10 °C/min ramping rate from 30 °C to 500 °C and maintained at that
temperature for 1 h. After that, the catalyst was cooled to 30 °C and subjected to
H -saturation in a stream of 10 % H /N . Then, the sample was purged with nitrogen
2
2
2
for 3 h to remove H in the gas phase, and then was heated from 30 °C to 800 °C with
2
a ramping rate of 10 °C /min.
Transmission electron micro-scope (TEM) images were taken on JEOL JEM
2
010 electron micro-scope at an acceleration voltage of 120.0 kV. The samples were
ultrasonic treatment and well suspended in ethanol. Drops of the suspension were
employed and well dispersed on a copper grid.
The thermogravimetric analysis was performed with a TGA Q500
thermogravimetric analyzer. The samples were heated under air atmosphere from
room temperature to 750 °C at a heating rate of 10 °C /min.
9

2
.3. Catalytic performance
The propane dehydrogenation reaction was performed in a fix-bed quartz tubular
reactor with an inner diameter of 8 mm. The catalyst (0.2 g) was placed into the center
o
of the reactor and reduced in H at 500 C for 1 h before catalytic evaluation. The
2
o
reaction conditions were as follows: 580 C of reaction temperature, atmospheric
pressure, C H /Ar= 7:3 (molar ratio) and the propane weight hourly space velocity
3
8
−1
(
WHSV) is 8.25 h . The reaction products were analyzed on-line using a GC-3000B
model gas chromatography equipped with a thermal conductivity detector (TCD). The
conversion of propane and the selectivity for propene were defined using following
equations:
F
 F
C H ,in
C H ,out
3 8
3
8
Propane conversion(%) 
Propene selectivity(%) 
100%
100%
(
1)
2)
F
C H ,in
3
8
F
 F
C H ,out
C H ,in
3 6
3
6
(
F
 F
C H ,in
C H ,out
3 8
3
8
Where Fi,in/out is the flow rate of each component in the feed or effluent.
3
. Results and discussion
3
.1. Characterization of Mg-modified SBA-15 supports
10

Fig. 1 - (a) Low-angle and (b) wide-angle XRD patterns of different supports: (A)
SBA-15, (B) 1Mg-SBA-15, (C) 2Mg-SBA-15, (D) 4Mg-SBA-15, (E) 6Mg-SBA-15,
and (F) 2Mg/SBA-15-IM
The low-angle XRD patterns of SBA-15 and Mg-modified SBA-15 samples are
shown in Fig. 1(a). The XRD pattern of SBA-15 showed a sharp peak at about 0.92°
together with two additional weak peaks, which were indexed to (100), (110), and
(200) reflections, respectively. The XRD profile verified the formation of typical
hexagonally-structured SBA-15 silica with highly ordered mesoporous channels as
reported by Zhao et al.[24] All of the patterns of xMg-SBA-15 exhibited these three
peaks with high intensity, indicating that the hexagonal mesostructure was intact even
if magnesium was introduced. Wei et al.[25] succeeded in synthesizing the
2
+
well-ordered structure MgO/SBA-15 composites since Mg ions were known to
11

interact strongly with the non-ionic surfactant P123. The intensity of diffraction peaks
showed gradual decrease with the introduction of Mg component, suggesting that
there was a deterioration of the ordered pore structures.[26] The calculated unit cell
parameters (a ) corresponding to the d spacings in the XRD pattern are listed in
0
100
Table 1. It could be observed that the unit cell parameter a of SBA-15 was 10.7 nm
0
while that was 10.7-10.9 nm for the magnesium-doped samples, indicating the
symmetry of pore structure did not significantly alter.
The wide-angle XRD patterns for these materials are depicted in Fig. 1(b). All of
the patterns showed a broad peak in 2θ region of 20-30°, corresponding to the typical
reflection of amorphous silica.[27] Except the typical patterns of SBA-15, the peaks at
2
θ of 42.9° and 62.3° were observed in 2Mg/SBA-15-IM sample, which indexed to
reflections of magnesium oxide (PDF Card No. 45-0946).[28] The absence of these
XRD diffractions indicated a better dispersion of magnesium species for the
xMg-SBA-15 samples synthesized by one-pot method.
Table 1 Textural properties of SBA-15 and Mg-modified SBA-15 materials
a
2
-1
b
3
-1
c
3
-1
d
e
f
Sample
SBA-15
SBET (m g ) V (cm g ) V
(cm g ) D (nm) d100 (nm) a (nm)
t
mic
p
0
571.5
504.4
420.5
385.5
381.0
0.78
0.69
0.65
0.70
0.67
0.61
0.10
0.10
0.08
0.05
0.04
0.10
5.4
5.5
6.1
6.6
6.6
5.4
9.3
9.4
9.3
9.4
9.4
9.4
10.7
10.9
10.7
10.9
10.9
10.9
1
2
4
6
Mg-SBA-15
Mg-SBA-15
Mg-SBA-15
Mg-SBA-15
2
Mg/SBA-15-IM 343.3
BET specific areas.
a
b
c
d
e
f
The total pore volumes were obtained at P/P = 0.99.
0
The microporous volume was calculated using the t-plot method.
Average pore diameter calculated using the BJH method.
d100 was the lattice spacing obtained from low-angle XRD.
a was the hexagonal unit-cell parameter calculated using the equation: a =2d / .
0
3
0
100
12

Fig. 2 - (a) N adsorption-desorption isotherms and (b) BJH pore size
2
distribution of different samples: (A) SBA-15, (B) 1Mg-SBA-15, (C) 2Mg-SBA-15,
(D) 4Mg-SBA-15, (E) 6Mg-SBA-15, and (F) 2Mg/SBA-15-IM
The textural properties of the samples were investigated by nitrogen
adsorption-desorption and the corresponding isotherms and pore size distributions are
shown in Fig. 2. All the samples exhibited typical isotherm of type IV with hysteresis
loops of type H1 according to the IUPAC classification, indicating the characteristic
of mesoporous materials.[29] It should be noted that the isotherms of Mg-modified
SBA-15 samples were very similar to that of SBA-15 in shape, suggesting that these
materials retained the initial mesoporous texture. The pore size distributions of the
support materials (Fig. 2(b)) displayed the pore size with high uniformity and mainly
distributed in the range of 4-7 nm. The results showed that all of the support materials
13

exhibited highly ordered mesoporous of the cylindrical channels with a narrow pore
size distribution, which was in good consistent with the results obtained by low-angle
XRD.
The specific surface area, pore volume and average pore diameter of SBA-15
and Mg-modified SBA-15 materials are summarized in Table 1. The surface areas
and total pore volumes of xMg-SBA-15 mesoporous materials decreased gradually
with the increasing amount of magnesium in the supports. Magnesium species
interacted with silanol groups and formed the Si-O-Mg monolayer, the residual
magnesium species covered on the top and converted to the smooth MgO layer after
x
calcination, thus decreasing both the surface area and pore volume.[30] Davidson et
al.[31] reported that the silica walls structure of SBA-15 materials was more complex
and showed a “corona” region with microporous around the cylindrical organic
aggregates due to the partial occlusion of the PEO chains in the silica matrix.
Magnesium species incorporated in SBA-15 would partly fill the “microporous
corona”. Therefore, the average pore diameter increased while the micropore volume
decreased when magnesium amount changed from 1wt.% to 6wt.%. Meanwhile,
2
Mg/SBA-15-IM possessed the lowest surface area and pore volume value, indicating
that the one-pot method was better able to maintain the mesoporous structure of silica
materials.
Fig. 3 - FT-IR spectra of samples: (A) SBA-15, (B) 1Mg-SBA-15, (C)
14

2
Mg-SBA-15, (D) 4Mg-SBA-15, (E) 6Mg-SBA-15
Infrared spectroscopy was used to characterize the surface property of the
as-made materials, the spectra were collected in Fig. 3. It could be clearly seen that
the framework vibrations of xMg-SBA-15 were characteristically similar to those of
-
1
SBA-15 structure. The symmetric stretching mode showed at 804 cm , while the
-1
band at 463 cm was assigned to the Si-O-Si bending mode.[32] The band in the
-1
range of 1050-1100cm was related to characteristic of Si-O-Si vibrational stretching,
and the peak in the samples was somewhat red-shifted as the amount of incorporated
magnesium increased, which indicated the interaction between magnesium and
silicon.[33]
Fig.4 - TEM images of 2Mg-SBA-15: (a) taken along the [001] directions and (b)
taken along the [110] directions; (c and d) HAADF-STEM image of 2Mg-SBA-15
and EDS elemental mapping images of O, Si, and Mg
To investigate the morphology of Mg-incorporated SBA-15 and the distribution
of magnesium species, TEM analyses were carried out. 2Mg-SBA-15 sample (Fig.
4
(a, b)) exhibited highly ordered hexagonal structure with narrow pore diameter size
distribution, which was in accordance with the XRD and N adsorption-desorption
2
15

results. There was no MgO particles observed on the 2Mg-SBA-15 material surface or
pores, indicating the well dispersion of magnesium on the support. The homogeneous
dispersion of Mg species in SBA-15 led to the less-blocked mesopores, consequently
these samples prepared by in situ incorporating method displayed the higher surface
area, pore diameter and pore volume than the impregnated sample. Distribution of Mg
on the support was confirmed by means of HAADF-STEM combined with EDS
elemental mapping analysis. From the elemental mapping images, it was clear that O,
Si and Mg were evenly distributed in mesoporous structure.
3
.2. Characterization of PtSn catalysts
Fig. 5 - (a) Low-angle and (b) wide-angle XRD patterns of reduced catalysts: (A)
PtSn/SBA-15, (B) PtSn/1Mg-SBA-15, (C) PtSn/2Mg-SBA-15, (D)
16

PtSn/4Mg-SBA-15, (E) PtSn/6Mg-SBA-15, and (F) PtSn/2Mg/SBA-15-IM
XRD patterns of reduced PtSn/xMg-SBA-15 and PtSn/2Mg/SBA-15-IM samples
are shown in Fig. 5. From low-angle XRD patterns, the features of all the samples
were rather similar, with three well-resolved diffraction peaks that indexed to (100),
(110) and (200) reflections, indicating the hexagonal mesostructure was intact even
after loading Pt and Sn. Wide-angle XRD patterns showed the typical peak of
amorphous silica at 2θ = 23.5°. The XRD reflections of 40.0°, 46.3°, 67.4° and 80.8°
at 2θ could be assigned to inter-planar spacing (111), (200), (220) and (311) of cubic
Pt metal structure, respectively.[34] There was no peak of Sn or SnO , probably due to
x
the highly-dispersed Sn species or the formation of amorphous SnO phases.
x
Fig. 6 - TEM images of the reduced samples: (a, b) PtSn/SBA-15,
(c)PtSn/2Mg-SBA-15, and (d) PtSn/2Mg/SBA-15-IM
To further investigate the morphology of the catalysts and the effect of Mg on the
modification of catalysts, the TEM experiments were implemented. The TEM images
of the corresponding samples (Fig. 6) showed that all catalysts remained the highly
ordered structure after Pt and Sn loading. For PtSn/SBA-15 sample, metallic particles
17

were irregularly distributed on the whole support and some particles were confined in
the porous channel appeared in Fig. 6(a), while some agglomerated particles
distributed on the external surface of the catalyst and the ends of mesopores were
partially blocked by the large particles. The movement of Pt particles within the pore
channels or from pores to the surface of support could occur under the reduction
conditions because of the weak interaction between Pt and support. The
agglomeration of metallic particles resulted in the low accessible metal surface area,
which in consequence suppressed the catalytic ability to adsorb the hydrogen at low
temperature.[35] Meanwhile, for the Mg-decorated samples, no large aggregates of
the metals on the external surface were observed, indicating a high distribution
throughout the pores of Pt and Sn metallic particles. The results suggested that the
modification of Mg could markedly influence metallic nanoparticles distribution. This
could be explained by that suitable content of magnesium additives could prevent the
agglomeration of dispersed Pt into discrete particles, thus an increase in platinum
dispersion was observed.[36] It was noted that the surfaces of PtSn/2Mg/SBA-15-IM
sample seemed to be covered by dark clouds, which could be caused by the unequally
dispersion of Mg species and be further responsible for the decline in both pore
volume and surface area. The results were in accord with the textural properties
characterized by nitrogen adsorption-desorption (Table 1).
Fig. 7 - NH -TPD profiles of different catalysts: (A) PtSn/SBA-15, (B)
3
18

PtSn/2Mg-SBA-15, and (C) PtSn/2Mg/SBA-15-IM
Fig. 7 shows the NH -TPD profiles of different catalysts. There were two
3
desorption peaks for each sample: the first desorption peak at about 60-230 °C and the
second one at above 230 °C. The first peak was attributed to the desorption of the
adsorbed NH on the weak acid sites, whereas the second peak was assigned to the
3
moderate acid sites.[37] In a comparison, a distinct decrease of ammonia desorption at
low temperature and the sharply decreased ammonia desorption peak at high
temperature were observed after loading of magnesium component. This phenomenon
showed that the catalyst acidity decreased with the addition of magnesium, which
could increase catalytic selectivity and stability.
19

Fig. 8 - (a) Pt 4f and (b) Sn 3d XPS spectra corresponding to the reduced
5
/2
catalysts: (A) PtSn/SBA-15, (B) PtSn/2Mg-SBA-15, (C) PtSn/6Mg-SBA-15, and
D) PtSn/2Mg/SBA-15-IM
(
X-ray photoelectron spectroscopy was conducted on the reduced
PtSn/xMg-SBA-15 catalysts to further study the electronic properties of the surface
elements. Fig. 8(a) shows the XPS spectra corresponding to the Pt 4f of the reduced
0
catalysts. The binding energies at 71.2 and 74.5 eV could be assigned to Pt , only
metallic Pt species presented on the catalysts. A slight shift to higher binding energy
was appeared as magnesium incorporated, which could be due to the strong
metal-support interaction (SMSI) between Pt and xMg-SBA-15 supports, it was
conducive to make Pt species with better resistance of sintering. Fig. 8(b) shows the
XPS spectra corresponding to the Sn 3d5/2 region. Three peaks were obtained at 485.9,
4
86.6 and 487.2 eV after deconvolution treatment, corresponding to different types of
tin species. The binding energy at 485.9 eV was associated with a reduced tin phase,
0
either in the metallic (Sn ) or in the alloy state.[38] While the second and third peak
2
+
4+
were attributed to different types of oxidized Sn i.e. Sn and/or Sn . It was well
2
+
4+
established that Sn and Sn could not be discriminated since their binding energies
were very close.[10]
Table 2 The binding energy and surface composition of catalysts derived from
XPS analysis
0
Tol
Binding energy (eV)
Sn 3d5/2
71.2 74.5 485.9 486.6 487.2 —
Sn /Sn Pt
Sn
Sample
a
b
b
Pt 4f
Mg 2p Si 2p O 1s (%)
103.3 532.5 14.1
wt.% wt.%
0.69 1.80
0.86 1.93
0.80 1.84
0.79 1.83
PtSn/SBA-15
PtSn/2Mg-SBA-15
PtSn/6Mg-SBA-15
71.4 74.6 486.0 486.7 487.5 50.5 103.0 532.3 6.5
71.6 75.0 486.0 486.7 487.4 50.6 102.9 532.2 9.0
PtSn2Mg/SBA-15-IM 71.3 74.5 486.0 486.7 487.2 50.4 103.1 532.3 8.1
a
Calculated using the peak area.
b
The mass percent of surface Pt and Sn elemental on the catalysts.
20

The binding energies of surface elements Pt 4f, Sn3d , Mg 2p, Si 2p and O 1s
5
/2
for various catalysts were summarized in Table 2. Mostly amount of the Sn species
0
were in the oxidized state. What's more, the content of Sn species in PtSn/SBA-15
catalyst (14.1%) was much more than that of the Mg-modified samples, indicating
2
+
4+
that more amounts of tin existed in a non-metallic state (Sn or Sn ) after
magnesium was added. However, an excess of Mg loading (6 wt.%) weakened the
interaction between Sn species and support, resulting in the increase of metallic tin
content. Thus, it could be concluded that a proper amount of Mg addition would
strengthen the Sn-support interaction, then inhibit the exhaustive reduction of tin
species and stabilize the oxidized tin species. It was found that the oxidative Sn
4
+
2+
species (Sn or Sn ) could modify the electronic and geometric properties of Pt
species, leading to the great reactivity and stability.[39] The mass percent of surface
Pt and Sn (Table 2) showed that incorporation of magnesium in SBA-15 enhanced
the dispersion of Pt and Sn species on the catalysts surface.[40] Meanwhile, much
stronger interactions of Pt, Sn and supports were observed over PtSn/2Mg-SBA-15 in
comparison with the impregnated sample. In addition, the binding energy of Mg 2p
over PtSn/2Mg/SBA-15-IM was smaller than that over PtSn/xMg-SBA-15, while it
was opposite for Si 2p. This might be caused by the formation of MgSiO over
3
PtSn/xMg-SBA-15 samples.[41]
21

o
Fig. 9 - H -TPD profiles of the catalysts after reduction at 500 C: (A)
2
PtSn/SBA-15, (B) PtSn/1Mg-SBA-15, (C) PtSn/2Mg-SBA-15, (D)
PtSn/4Mg-SBA-15, (E) PtSn/6Mg-SBA-15, and (F) PtSn/2Mg/SBA-15-IM
The H -TPD profiles of the reduced catalysts are displayed in Fig. 9, where two
2
hydrogen desorption peaks could be observed in the profiles. In general, the
o
desorption peak at low temperature (84-125 C) was attributed to the hydrogen on
o
metallic Pt, whereas the high temperature desorption peaks at 650 C could be
assigned to spillover hydrogen.[22, 42] The amount of H desorption for different
2
catalysts is shown in Table 3. In the case of magnesium promoted catalysts, the main
difference originated from the hydrogen desorption at low temperature, as the amount
of hydrogen was up to 94 μl/gcat. regardless of the magnesium content in the catalysts.
Meanwhile, the corresponding temperature moved towards the higher temperature,
revealing that the capacity of absorbing H became much stronger at low temperature.
2
It was worthy noting that the PtSn/SBA-15 sample showed the least amounts of
hydrogen adsorbed compared with that of the Mg-SBA-15 based catalysts, which was
probably caused by the diminished Pt particles because of sintering. Generally, the
main reason for the decrease in amounts of hydrogen on metallic Pt was the sintering
of particles.[43] According to the hydrogen desorption behaviour at low temperature,
one can conclude that Pt particles dispersed well on catalyst surface, which would
provide accessible metal surface area for the dehydrogenation of propane.[35]
Nevertheless, when the concentration of magnesium was excessive (6 wt.%), more
amounts of magnesium species might block the platinum particles, which led to the
decrease of the metal dispersion. [44, 45] These results suggested that the promoter
Mg could make a significant difference in metallic distribution. Moreover, the catalyst
with modified Mg species by the one-pot method possessed higher metal dispersion
than that by the impregnation method.
Table 3 The amount of different hydrogen desorption for the reduced catalysts
22

H uptake (μl/g )
2
cat.
Sample
o
o
8
4-125 C
650 C
38.5
38.9
38.4
40.4
38.9
38.2
Total
PtSn/SBA-15
13.5
36.6
53.2
93.8
81.4
44.2
52.0
75.5
91.6
134.2
120.3
82.4
PtSn/1Mg-SBA-15
PtSn/2Mg-SBA-15
PtSn/4Mg-SBA-15
PtSn/6Mg-SBA-15
PtSn/2Mg/SBA-15-IM
3
.3. Catalytic performance for propane dehydrogenation and coke analysis
23

Fig. 10 - (a) Propane conversion and (b) propene selectivity as functions of time
o
on stream for different catalysts (reaction conditions: T = 580 C, C H :Ar = 7:3
3
8
-1
(
molar ratio), 0.1 MPa, WHSV = 8.25 h ; m(cat) = 0.2 g)
Fig. 10 shows the catalytic performance of propane dehydrogenation over the
different catalysts. For PtSn/SBA-15 catalyst, the initial conversion of propane was
3
5.7% and then reduced to 27.2% after 6 h reaction. The catalytic activity and
stability were clearly enhanced when 1-4 wt.% Mg was introduced into the catalysts.
Nevertheless, the catalytic activity declined rapidly with the further increase of Mg,
suggesting that excess Mg content would suppress the propane dehydrogenation
24

reaction. The promotion effect of Mg on the catalytic performance could be related to
the better metallic particle distribution and stabilized effect of nonmetallic state Sn
species. The influence of preparation method was also dramatic, PtSn/2Mg-SBA-15
prepared by one-pot method exhibited better catalytic activity than
PtSn/2Mg/SBA-15-IM. At the same time, the PtSn/2Mg-SBA-15 catalyst showed the
most excellent catalytic activity with a maximum initial propane conversion of 43.0%
among all the samples.
Fig. 10(b) shows propene selectivity versus reaction time for different catalysts.
It could be observed that the propene selectivity kept stable for 6 h propane
dehydrogenation reaction. With the incorporation of Mg (1-4 wt.%), the propene
selectivity for other catalysts were markedly higher than that of PtSn/SBA-15 catalyst,
and stabilized at above 95.0%. The PtSn/2Mg-SBA-15 catalyst showed the highest
propene selectivity reached 99.2% among all catalysts. As for PtSn/SBA-15 sample,
some metal particles were agglomerated due to the weak interaction among Pt, Sn and
support (Fig. 6(a)). Hydrogenolysis reaction would be carried out easily since this
unwanted side reaction needed relatively large platinum ensemble, which resulted in a
low selectivity of propene.[46] However, the continuous addition of Mg (6 wt.%)
resulted in the decreased propene selectivity compared with that of the PtSn/SBA-15
catalyst.
It is known that platinum was the only active metal in PtSn/SBA-15 catalyst and
propene was formed only by dehydrogenation on the metallic sites, while the main
by-products (C1-C2 hydrocarbons) were mainly from propane cracking on the
support.[47] Therefore, the changes of acidity and Pt active sites were responsible for
the selectivity of propene. The addition of Mg promoted metal dispersion and
decreased the acidity of the catalyst, inhibiting the side reactions, so a better catalyst
performance was obtained. On the other hand, it was widely reported that two types of
Pt active sites existed in the PtSn catalyst, denoted as M sites and M sites.[48]
1
2
Platinum directly anchored on the support surface formed multiple Pt centers i.e. M1
sites, while Pt particles attached to the SnO surface with a “sandwich structure” were
x
assigned to the M sites. Generally, M sites were connected with the side reactions
2
1
25

such as hydrogenolysis and coking, whereas the M sites were responsible for the
2
dehydrogenation of propane. According to the XPS results, the presence of Mg
dramatically restrained the reduction of Sn species and stabilized the oxidation states
of Sn, which was favorable to form the M sites. Consequently, more M sites were
2
2
produced and better catalytic performance was realized.
Concerning the PtSn/6Mg-SBA-15 sample with an excess of Mg loading and
PtSn/2Mg/SBA-15-IM introducing of Mg by impregnation method, the decrease in
catalytic performance might be due to the changed balance metallic function and a
lower percentage of the SnO species. However, with the continuous addition of Mg
x
and the impregnation method both led to a deterioration of the ordered pore structure
and decreased the surface area of catalyst, which were damage to the catalytic
performance. Table S1 shows the carbon balance for different catalysts, and good
balances were obtained.
Table 4 Values of initial conversion (X ), final conversion (X ), deactivation parameter
o
f
(D) after 6 h reaction, and initial selectivities to olefins at 580 °C for different
catalysts
Sample
b
S (%)
C H C H
4
a
Xo (%)
Xf (%)
D
C H
CH4
2.35
1.33
0.97
3.41
5.49
2.36
3
6
2
6
2
PtSn/SBA-15
35.9
40.9
43.0
38.3
32.8
37.5
27.2
35.5
38.1
30.4
24.8
30.0
24
13
11
20
24
20
94.09
96.31
97.84
94.23
89.00
95.51
1.83
0.84
0.57
0.92
3.03
1.26
1.28
1.26
0.63
0.96
2.18
0.62
PtSn/1Mg-SBA-15
PtSn/2Mg-SBA-15
PtSn/4Mg-SBA-15
PtSn/6Mg-SBA-15
PtSn/2Mg/SBA-15-IM
a
The deactivation parameter (D) was defined as: D = 100×(X - X )/X , where X was
o
f
o
o
the initial conversion and X was the final conversion.[45]
f
b
The initial major by-product selectivity.
Table 4 shows the percentage of deactivation after 6 h on stream for the catalysts.
26

It could be observed that appropriate amount of Mg-incorporated in SBA-15 (1 wt.%
4 wt.%) was good to improve the catalytic stability, whereas an excess of Mg
~
loading (6 wt.%) played a negative effect on stability. The stability of the catalysts
decreased in the following sequence: PtSn/2Mg-SBA-15 > PtSn/1Mg-SBA-15 >
PtSn/4Mg-SBA-15 = PtSn/2Mg/SBA-15-IM > PtSn/6Mg-SBA-15 = PtSn/SBA-15.
The initial major by-product selectivity is shown in Table 4. The PtSn/2Mg-SBA-15
catalyst exhibited the lowest by-product selectivity. With the continuous increase of
magnesium from 1 to 4 wt.%, the selectivity of by-products significantly decreased,
suggesting that the side reactions were well restrained. These results were well in
accordance with propene selectivity in Fig. 10(b). Table S2 lists some related
literature data over the different Pt catalysts for propane dehydrogenation. Despite
differences in catalyst compositions and reaction conditions, the PtSn/2Mg-SBA-15
catalyst used in this study showed better catalytic performance than most catalysts.
27

Fig. 11 - TEM results for spent catalysts: (a, b) PtSn/SBA-15, (c,
d)PtSn/2Mg-SBA-15, (e, f) PtSn/2Mg/SBA-15-IM
The TEM images for the spent catalysts are shown in Fig. 11, all of the samples
maintained ordered meso-porous structure after 6 h reaction. For the spent
PtSn/SBA-15 catalyst, some porous channels were blocked by the deposited
carbonaceous species, resulting in the activity loss. The coke formed on catalyst
surface not only covered the active sites but also plugged up the channel where the
propene outflowed.[11] There was no obvious carbon species was observed on the
spent Mg-modified catalysts, and most of the metallic particles were confined within
the channels of the porous. The confinement effect from the channel of the SBA-15
28

molecular sieve would help to stabilize the metallic particles and improve the
sintering resistance, which could increase the catalytic stability during the propane
dehydrogenation reaction.[49]
Fig. 12 - TG-DTA profiles of spent catalysts: (A) PtSn/SBA-15, (B)
PtSn/2Mg-SBA-15, (C) PtSn/6-MgSBA-15, and (D) PtSn/2Mg/SBA-15-IM
As far as catalytic deactivation was concerned, coke deposition was an inevitable
issue. The deposited coke could be easily identified and quantified under the help of
TG analysis. As shown in Fig. 12, there were two successive weight loss regions on
the spent Mg-modified catalysts, representing two different coke deposits. The first
weight loss at around 250 °C was assigned to the coke deposited on the metal surface,
while the weight loss at high temperature (520 °C) corresponded to the coke located
on the support.[50] It was clear that for the spent PtSn/SBA-15 catalyst, most of
carbon deposits covered the active metal and only a small proportion of them located
the external surface of the carrier. In addition, it was important to note that the
combustion temperature for the coke on Mg-modified catalysts shifted to the lower
temperature, suggesting that the deposited carbon on the surface of metal became
more active. These phenomena suggested that the presence of Mg in the catalyst
promoted the migration of the coke deposits from the metal sites to the support.
29