Pt/TS-1 catalysts: Effect of the platinum loading method on the dehydrogenation of n-butane

Article abstract of DOI:10.1016/j.apcata.2021.118194

A series of catalysts in which platinum was supported on the crystalline pure titanium silicalite (TS-1) were prepared using two different loading methods, namely, an ethylene glycol (EG) reduction method and the conventional incipient wetness impregnation (IM) technique. Various characterization techniques were used to study the effect of the loading method on the physicochemical and morphological properties of the prepared catalysts. Also the effect of the platinum-loading method on the dehydrogenation of n-butane was investigated in a fixed-bed reactor. The results show that the EG method favors the formation of a more-concentrated Pt dispersion, which results in much better catalytic activities for the selective formation of butenes and butadiene (> 97 %). This phenomenon is interpreted by carrying out density functional theory (DFT) calculations with focus on the relationship between the coverage of n-butane on Pt surface and the activation barrier for the first C[sbnd]H bond cleavage.

Full text of DOI:10.1016/j.apcata.2021.118194

Applied Catalysis A, General 621 (2021) 118194
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Pt/TS-1 catalysts: Effect of the platinum loading method on the
dehydrogenation of n-butane
a,
b
,
c
a,
b
,
c
,
a,
c
c
a,b,c,
Mingyuan Shao
, Chaoquan Hu
*, Xuebing Xu , Yang Song , Qingshan Zhu
*
a
State Key Laboratory of Multiphase Complex Systems, Institute of Process Engineering, Chinese Academy of Sciences, Beijing, PR China
University of Chinese Academy of Sciences, Beijing, PR China
b
c
Nanjing IPE Institute of Green Manufacturing Industry, Nanjing, PR China
A R T I C L E I N F O
A B S T R A C T
Keywords:
A series of catalysts in which platinum was supported on the crystalline pure titanium silicalite (TS-1) were
prepared using two different loading methods, namely, an ethylene glycol (EG) reduction method and the
conventional incipient wetness impregnation (IM) technique. Various characterization techniques were used to
study the effect of the loading method on the physicochemical and morphological properties of the prepared
catalysts. Also the effect of the platinum-loading method on the dehydrogenation of n-butane was investigated in
a fixed-bed reactor. The results show that the EG method favors the formation of a more-concentrated Pt
dispersion, which results in much better catalytic activities for the selective formation of butenes and butadiene
N-butane
Direct dehydrogenation
Pre-reduction
DFT
Particle size
(
> 97 %). This phenomenon is interpreted by carrying out density functional theory (DFT) calculations with
focus on the relationship between the coverage of n-butane on Pt surface and the activation barrier for the first
H bond cleavage.
C
–
1
. Introduction
The catalytic dehydrogenation of n-butane is a highly important re-
support during heterogeneous catalysis [11,12]. In some cases, the
strong mutual interaction could stabilize the size of metal particles
during catalytic reactions. Simultaneously, the type or morphology of
the support is also a promising method for improving the catalytic
performances of Pt-based catalysts. For instance, in the dehydrogenation
process the use of structured mesoporous (silica foam [17], SBA-15 [8]
and MCM-41 [16]) silica can stabilize the metal nanoparticles through
the well-defined pores. In recent years, the TS-1 titanium silicalite mo-
lecular sieve has been considered to be one of the most promising sup-
ports for Pt particles due to its highly dispersed and isolated Ti [18,19].
TS-1 is very mechanically strong and has a high specific surface area and
pore structure suitable for loading active components. More impor-
tantly, its high thermal/hydrothermal stability [20] motivates us to
investigate TS-1 supported Pt particles towards n-butane
dehydrogenation.
action for production of butenes (n-butene, cis-2-butene, and trans-2-
butene) and butadiene, which are the raw materials of synthetic rubbers,
plastics, and other valuable chemical products [1,2]. There are two
types of dehydrogenation: (i) oxidative dehydrogenation [3,4] and (ii)
direct dehydrogenation [5–7]. Oxidative dehydrogenation requires an
oxidant, such as oxygen, carbon dioxide, or nitrous oxide, to be intro-
duced, which easily leads to a deeper n-butane oxidation; consequently,
selectivity for the target products is limited by side reactions [8–10]. By
comparison, direct dehydrogenation exhibits a competitive selectivity
for the target products, i.e. butenes and butadiene. However, this pro-
cess requires high temperatures to achieve sensible yields and hence
suffers catalyst deactivation. Since butane dehydrogenation is a struc-
ture sensitive reaction [11–13], controlling the particle size of Pt could
improve the activities of the catalysts at relatively low temperatures. In
this case, the deactivation could be inhibited.
Conventional preparation techniques based on wet impregnation
and the chemical reduction of metal precursors, typically by gas-phase
2
H , often do not provide adequate control of particle shape and size
Many kinds of Pt support have been tested for n-butane dehydroge-
[22]. Modified methods have been developed recently, and materials
that are more-active than commercial Pt-based catalysts have been
produced. In all cases, the connectivity between the Pt particles and the
surface of the support, and the dispersion of the Pt particles was found to
nation process, including MgO [1], SiO
Al [2,7,14]. Catalytic performance is known to be significantly
affected by the electronic interactions between the metal and the
2 2 2
[6], TiO [9], ZrO [15], and
2 3
O
*
Corresponding author at: State Key Laboratory of Multiphase Complex Systems, Institute of Process Engineering, Chinese Academy of Sciences, Beijing, PR China.
E-mail addresses: cqhu@ipe.ac.cn (C. Hu), qszhu@ipe.ac.cn (Q. Zhu).
https://doi.org/10.1016/j.apcata.2021.118194
Received 27 January 2021; Received in revised form 23 April 2021; Accepted 27 April 2021
Available online 30 April 2021
0
926-860X/© 2021 Elsevier B.V. All rights reserved.

M. Shao et al.
Applied Catalysis A, General 621 (2021) 118194
be important [23,24].
ICPOES730 apparatus. The amount of coke deposited on the catalyst has
been calculated from ICP analysis.
In this paper, we use two different preparation techniques to prepare
TS-1 supported Pt nanoparticles towards n-butane dehydrogenation.
Specifically, a series of TS-1 based catalysts was prepared by the
conversional wet impregnation method as well as the ethylene glycol
The catalyst surfaces were characterized by in-situ DRIFT under the
reaction conditions using the BRUKER TENSOR II spectrometer. The
catalyst was loaded into an in-situ cell fitted with ZnSe windows. Prior to
◦
(
EG) prereduction technique. The effect of the platinum-loading method
on the dehydrogenation of n-butane was investigated in a fixed-bed
reactor. N -adsorption/desorption, X-ray diffractometry (XRD), diffuse
any DRIFT experiment, the catalyst was reduced at 400 C under a flow
ꢀ 1
2
(20 mL min ) of 50 % H in Ar for 1 h, and then purged with pure Ar for
2
0.5 h at the same temperature to remove hydrogen adsorbed on the
catalyst. After the cell cooled down to room temperature, 100 ppm of CO
reflectance infrared Fourier-transform spectroscopy (DRIFTS), X-ray
photoelectron spectroscopy (XPS), and transmission electron micro-
scopy (TEM) were used to study the effects of the various preparation
methods on the physicochemical and morphological properties of the
prepared samples. The location of the platinum, the acidities of the
catalysts, and their catalytic performance toward the dehydrogenation
of n-butane were investigated. Furthermore, the relationship between
the particle size and the activity was studied by DFT calculations.
ꢀ 1
balanced with Ar was fed into the cell at a flow rate of 20 mL min , and
the IR spectra were collected after injecting the mixed gas for 0.5 h at
◦
50 C.
2.3. Dehydrogenation of n-butane
The dehydrogenation of n-butane to n-butenes and 1,3-butadiene
was carried out in a tubular quartz reactor (i.d.; 10 mm) at atmo-
spheric pressure. An aliquot of the catalyst (0.10 g) was placed at the
center of the reactor where the temperature was uniform. Butane
balanced with Ar was continuously introduced into the reactor during
2
. Catalyst synthesis
2
.1. Catalyst preparation
the dehydrogenation process. Before n-butane dehydrogenation, the IM
◦
The TS-1 support was synthesized according to the traditional hy-
and EG catalysts need to be reduced with H
2
atmosphere at at 400 C for
drothermal method with the required Si/Ti molar ratio [22]. First,
specific amounts of tetraethyl orthosilicate (TEOS) and tetrabutyl tita-
nate (TBOT) were dissolved in an aqueous solution of tetrapropy-
lammonium hydroxide (TPAOH). The mixture was stirred in ice water
1 h. The feed composition was fixed at n-butane: H
2
: Ar = 5: 1: 94 at a
ꢀ 1
total flow rate of 50 mL min . The reaction temperature was measured
using a thermocouple placed in the quartz reactor.
The reaction products were analyzed by a gas chromatography (GC)
with a thermal conductivity detector (TCD) and a flame ionization de-
◦
for 1 h and then heated to 70 C for 3 h, after which isopropanol (IPA)
was added and the mixture was stirred for 1 h. The molar composition of
tector (FID). A KB-Al
2
O
3
/Na
2
SO
4
column (50 m ×0.53 mm ×20
μ
m) and
the final gel was: 1 SiO
2
: 0.02 TiO
2
:4 IPA: 0.3 TPAOH: 35 H
2
O. Second,
a packed column (13X molecular sieve) were used for the GC. The
stream from the reactor was transported to the GC through a six-way
◦
the solution was subjected to hydrothermal crystallization at 170 C for
◦
4
d. After repeated washing and drying, the precursor was calcined in air
valve. The temperature of the GC oven was initially set at 80 C and
◦
◦
◦
at 550 C to produce the 100 nm TS-1 zeolite.
then increased to 170 C at 20 C/min. The chromatograph was used to
qualitatively determine the various products at different retention times
with an external standard used for accurate quantitative results. All
response factors of the target products and reactants were obtained
using calibration gases. The conversions of n-butane to butenes and 1,3-
butadiene and their selectivities were calculated on the basis of the
carbon balance as follows:
The IM-1 catalyst was prepared by the traditional impregnation
method, as reported in the literature [4,10]. Pt-TS-1-EG catalysts were
prepared by an ethylene glycol prereduction (EG) method. An aqueous
solution of
H
2
PtCl
6
⋅6H
2
O was added to a solution of poly-
vinylpyrrolidone (PVP) in ethylene glycol with stirring. The mixed so-
◦
lution was then heated at 140 C and for 2 h under reflux. The TS-1
support was then added to the mixture, ultrasonicated, and dried at
Conversion(%) = F
[
C4H10
in
]
ꢀ [FC4H10]out
× 100.
(1)
(2)
◦
◦
1
20 C for 12 h. The mixture was then calcined at 500 C for 4 h to obtain
[F
C4H10]in
the Pt-TS-1-EG catalyst. The size distributions of Pt particles were
investigated by varying the molar ratio of Pt to PVP. These catalysts are
referred to as “EG-1” (Pt:PVP = 1:20), “EG-2” (Pt:PVP = 1:50), and
ni[Fi]
out
Selectivity(%) =
× 100.
Σni × [Fi]
out
“
EG-3” (Pt:PVP = 1:70).
The turnover frequency (TOF) of n-butane in dehydrogenation re-
2
.2. Catalyst characterization
action was calculated as:
The crystalline phases of the TS-1-based catalysts were examined by
TOF = F
[
C4H10]in × conv. × MPt
(3)
m
cat × wPt × DPt
XRD (Rigaku, D-MAX2500-PC) at 50 kV and 100 mA using Cu-K radi-
ation (λ = 1.541 Å). The Brunauer-Emmett-Teller (BET) surface areas,
pore volumes, and average pore diameters of the catalysts were
where [FC4H10
]in is the molar flow rate of n-butane in the feed (mol/s);
conv. is the conversion of n-butane; MPt is the atomic weight of platinum
195.08 g/mol); mcat is the weight of catalyst (g) used; wPt is mass
2
measured by N -adsorption/desorption experiments using a BELSORP-
(
mini II (BEL Japan) instrument. The size distributions of the Pt parti-
cles on the support were determined by TEM using a JEOL 2100 in-
strument. The Pt particles were assumed to be spherical and the average
fraction of platinum and DPt is the dispersion of Pt particles. DPt is the
dispersion of Pt particles and was calculated by DPt = 1.13/Pt particle
size.
∑
3
particle size was calculated according to the equation: d =
n
i
d
i
/
∑
The activation energy (Ea) was calculated by the Arrhenius equation:
2
i
n
i
d
.
Ea
RT
XPS was performed on a Thermo Scientific ESCALab 250Xi spec-
k = a × exp(ꢀ
)
(4)
trometer with 200-W monochromatic Al K
α
radiation. The 500 m X-ray
μ
spot was used for small area XPS (SAXPS) analyses. The base pressure in
where a, R, and T are the pre-exponential factor, the universal gas
constant, and the absolute temperature, respectively. The activation
energy are researched with the conversions of n-butane lower than 10 %.
ꢀ 9
the analysis chamber was about 3 × 10 mbar. Typically, the hydro-
carbon C1s line at 284.8 eV from adventitious carbon was used for en-
ergy referencing. Thermogravimetric analysis (TGA) patterns were
recorded on a Perkim-Elmer Diamond TGA apparatus under N
2
atmo-
◦
sphere and a heating rate of 10 C/min. Inductively Coupled Plasma
Atomic Emission Spectrometer has been collected on
a Agilent
2

M. Shao et al.
Applied Catalysis A, General 621 (2021) 118194
2
.4. DFT calculations
3.1.2. X-ray diffractometry analysis
The XRD patterns of the catalysts are shown in Fig. 1. All samples
◦
◦
◦
◦
◦
The plane-wave density functional theory was carried out by using
exhibit clear five diffraction peaks at 8.0 , 8.9 , 23.1 , 24.0 , and 24.4 ,
which are the characteristic peaks of TS-1 with a mordenite framework
inversion (MFI) structure [21]. The sharp peaks indicate that the cata-
lysts contain the highly crystalline TS-1 phase. No obvious XRD peaks of
titanium were observed for all the catalysts, suggesting that titanium is
incorporated in the lattice or is highly dispersed in the form of very small
titania particles. In addition, the catalysts prepared by the ethylene
glycol did not show any characteristic platinum peaks. The absence of
the Pt reflections is most probably due to the presence of very small Pt
particles [22,23]. In contrast, the pattern of the IM-1 catalyst displays a
the Vienna ab initio simulation package (VASP) [1]. The electron-ion
interactions was investigated by the projector augmented wave (PAW)
method [2] and the exchange corrective interaction was described
through the generalized gradient approximation (GGA) with the
parametrization of Perde-Burke-Ernzerhof (PBE) [3]. The plane wave
basis energy cutoff was truncated at 400 eV. While performing the ge-
ometry optimization, the convergence criteria of energy and force were
ꢀ 5
used as 1 × 10 eV/atom and 0.02 eV/Å, respectively. For searching
the transition state, a force tolerance of 0.03 eV/Å and a total energy of
-
7
◦
1
× 10 eV/atom were converged to obtain the accurate forces.
peak at 40.1 , which is the characteristic peak of Pt (111). This indicates
Here, the increase in the super cell means a decrease in the coverage.
a poorly dispersed or rather large size of metallic Pt particles.
Grids of 5 × 5×1, 5 × 3 × 1, 3 × 3 × 1, 3 × 3 × 1 Monkhorst-Pack K
points were used for the calculations of surfaces (2ｘ2), (2ｘ3), (3ｘ4),
and (4ｘ4), respectively. To avoid the interaction between the n-butane
and the bottom of slab through periodic boundary, a 15 Å thickness
vacuum was employed in the z direction. Further, the bottom of the two
layers were fixed to simulate the bulk structure. The adsorption energy
between the adsorbed atom (molecule) and the surface was defined as
following:
3.1.3. Morphology
TEM was used to observe the size distribution, of the Pt particles on
the catalysts. As shown in Fig. 2a, IM-1 contains particles with an
average size of 5.62 nm. This figure shows poor dispersion of Pt species
for the biggest particle size is bigger than 15 nm. By comparison, the Pt
particle-size distributions on TS-1 prepared by ethylene glycol pre-
reduction are shown in Figs. 2b–d. The average size of Pt particles on
sample by impregnation method is much larger than those prepared by
the ethylene glycol prereduction method (1.81, 2.76, and 3.56 nm).
These EG samples show good dispersion of Pt particles, which reveals
uniform, nearly normally distributed sizes in the 1–4 nm range. These
TEM figures is consistent with the XRD results, revealing the absence of
large metal crystals in these samples [20]. Figs. 2b–d suggest that,
owning to the confinement effect of the preparation method, encapsu-
lation inhibits the metal from further agglomerating into larger particles
to a certain extent.
E
a
= Etotal ꢀ Esurface ꢀ Eadsorbate
(5)
where Etotal, Esurface, and Eadsorbate represent for the total energy of the
system, the surface, and the adsorbate, respectively.
In this work, the climbing-image nudge elastic band (CI-NEB) com-
bined with the dimer method were implemented to search the saddle
point. This has been proven to be one of the most efficient methods in
searching transition state [4–6]. Importantly, after obtaining the
so-called transition state, the harmonic vibrational frequency calcula-
tions had to be performed to ensure that it is the real transition state. The
energy barrier was described as follows
3
.1.4. Surface valences
Fig. 3 shows XPS Pt 4f spectra of IM-1 and EG-2. The Pt 4f XPS
E
a
= Ets ꢀ Ein
(6)
spectrum of IM-1 (Fig. 3a) shows peaks at 71.5 (Pt 4f7/2) and 74.6 eV (Pt
◦
4
f
5/2) that are assigned to metallic Pt (Pt ) [28]. Fig. 3b shows that EG-2
where Ets and Ein stand for the total energies of the transition state and
displays peaks at 72.1 and 75.5 eV which are more intense than the
peaks in Fig. 3a. The observed change in the peak position indicates that
the electronic environments of the Pt particles on the catalysts prepared
by the two methods are different. According to the literature, the XPS
data reveal that the Pt in the EG-2 catalyst is in an electron-deficient
state. This is attributable to strong interactions between the Pt species
and Ti ions, with the Ti ions withdrawing electrons from the Pt species
the initial state, respectively.
3
. Results and discussion
3
3
.1. Characterization
.1.1. Physicochemical properties of the catalysts
[
29,30]. As shown in Fig. 4, EG-2 and IM-1 catalysts contain Ti species in
The surface areas of the catalysts prepared by the two methods are
different chemical states; the binding energies of the Ti species in EG-2
are 464.8 and 459.1 eV, which are lower than those of IM-1 (465.5 and
listed in Table 1. The catalysts prepared by the ethylene glycol pre-
2
ꢀ 1
reduction method exhibited a close BET surface area (370 m g ); while
the catalyst prepared by the impregnation method exhibited a slightly
lower value. This could be due to the decrease in the total pore volume,
as shown in Table 1. It has been reported that the impregnation method
likely results in a poor distribution of Pt particles on the catalyst surfaces
and could block the pores of support [20]. In addition, the average pore
size for all the catalysts was similar. The contents of Pt (wt%) were
around 0.81 wt% as determined by the ICP analysis. Besides, the value in
IM-1 is slightly higher than that in EG samples.
4
59.6 eV).
Table 1
Physicochemical properties of the prepared catalysts.
Catalysts
Surface area
Average pore
(nm)
Pore volume
Pt
2
ꢀ 1
3
(
m
g
)
(cm ⋅gꢀ 1)
(wt%)
EG-1
EG-2
EG-3
IM-1
370
372
369
360
2.73
2.75
2.74
2.76
2.53
2.56
2.54
2.44
0.82
0.79
0.79
0.84
Fig. 1. XRD patterns of various Pt-TS-1 catalysts: (a) IM-1, (b) EG-1, (c) EG-2,
and (d) EG-3.
3

M. Shao et al.
Applied Catalysis A, General 621 (2021) 118194
Fig. 2. (Left) TEM images of: (a) IM-1, (b) EG-1, (c) EG-2, and (d) EG-3. (Right) the corresponding size distributions of Pt particles.
The DRIFT spectra for the adsorption of CO on reduced TS-1-based could be attributed to Ti-Pt-CO. [32] However, as shown in Fig. 5c, the
ꢀ
1
ꢀ 1
ꢀ 1
catalysts at room temperature (Fig. 5) show Pt-CO (2080 cm ), Pt-
peak at 2080 cm is absent and the peak at 2114 cm is quite strong
which is attributed to the good dispersion of Pt particles and their small
sizes [33].
ꢀ 1
ꢀ 1
CO-Pt (1850 cm ), and Ti-CO (2170 cm ) [31] peaks. Significantly,
these peaks, which correspond to adsorbed CO, are quite different. There
are mainly 4 bands of CO adsorbed on the Pt. CO bonded to Pt of IM
ꢀ 1
sample produces a peak at 2080 cm , while CO bridge bonding on Pt
3
3
.2. Catalytic performance
ꢀ 1
clusters exhibits a broad resonance at 1880 cm . And the peak at
ꢀ 1
2
170 cm
indicates the CO-Ti band strength. In EG samples, the
.2.1. Dehydrogenation of n-butane
ꢀ 1
absence of a broad band of 1800ꢀ 1950 cm , which is corresponding to
The activities of the catalyst for the dehydrogenation of n-butane
bridgebonded CO-Pt stretches, confirms a uniform dispersion of Pt. The
were evaluated using a fixed-bed reactor connected to a gas chromato-
-
1
peak at 2110 cm indicates a linear CO-Pt stretch: a single atom Pt site.
However, there is lack of enough characterization analysis to prove
whether Pt is single atom or not. Another possible cause is the strong
graph. The catalytic performance for n-butane dehydrogenation is
◦
shown in Fig. 6 as a function of the reaction temperature (300ꢀ 580 C).
By comparison, the catalysts prepared by the traditional impregnation
and the EG prereduction methods are researched by the same
-
1
interaction between Pt and Ti in the support, the peak at 2110 cm
4

M. Shao et al.
Applied Catalysis A, General 621 (2021) 118194
high temperature to occur. The conversion of n-butane increased from a
low temperature and then decreased as the temperature exceeded
◦
5
00 C, which is due to the formation of coke on the catalyst surfaces.
The selectivities for trans-2-butene and cis-2-butene are quite close at
about 37 %, and vary in a same way; i.e. decrease as the selectivity for
1
,3-butadiene increases. Fig. 6 shows the differences observed between
the catalysts prepared by the various methods. The selectivities of the EG
catalyst for dehydrogenated products (1-butene, trans-2-butene, cis-2-
butne, and 1,3-butadiene) are quite high (> 96 %) with few isomerized
and hydrogenolysis products (C1–C3) observed. However, TS-1-
supported Pt prepared by the EG prereduction was highly active,
while the catalyst prepared by the impregnation method was poorly
active. The conversions of n-butane on the catalysts prepared by the the
EG prereduction method are higher than that of the catalyst prepared by
the impregnation method. In addition, the selectivity of the IM catalyst
for the dehydrogenated products (1-butene, trans-2-butene, cis-2-butne,
and 1,3-butadiene) is about 88 % and the selectivity of C1–C3, which are
the hydrogenolysis products, is about 12 %.
Fig. 3. Pt 4f XPS spectra of selected TS-1 based catalysts.
3
.2.2. Stability testing the TS-1-based catalysts
The n-butane conversions of the variously prepared TS-1-based cat-
alysts are shown in Fig. 7. The n-butane conversions of these catalysts
slowly decrease with time on stream. The conversion of IM-1 decreased
more obviously with time than the catalysts prepared by the EG pre-
reduction method. The conversions of EG clearly higher than those of
the IM catalyst. Among these catalysts, EG-2 exhibited the best catalytic
activity and stability for conversion, with only a slightly decrease of
conversion observed after 10 h on stream. As shown in Fig. S3, EG-2
◦
sample shows the highest yield of all butenes at 500 C. Since the
trend for the selectivity of butenes on these samples are quite similar, the
trend of butenes yield are almost same with that of n-butane conversion.
Further research is required to obtain a better understanding of the
reason for the superior stability of the catalysts prepared by the EG
prereduction method.
3
.3. Density functional theory calculations
To gain a better understanding of the Pt particle size for tuning the
conversion of n-butane, density functional theory (DFT) calculations
Fig. 4. Ti 2p XPS spectra of selected TS-1 based catalysts.
were performed to investigate the reaction energy barrier for the first
C H bond cleavage (dehydrogenation) of n-butane. The nano-catalysts
–
with smaller Pt particle size show more active site in a certain area for
size effect. We connect the Pt particle size with different molecules on
the Pt surface. The adsorptions of n-butane on different surface cover-
ages are investigated by locating the molecules on p(2 × 2), p(2 × 3), p
(
3 × 3), p(3 × 4) surfaces, respectively. The catalyst with the smallest Pt
particle size means the most Pt (111) in a per unit area which is defined
as the molecules on p(3 × 4) surfaces. And the final snapshots of n-
butane adsorbed on the top slab are shown in Fig. 8. According to Eq.
(
5), it is calculated that the adsorption energies are ꢀ 0.08.ꢀ 0.10, ꢀ 0.09,
and ꢀ 0.09 eV, respectively. The energies indicate that the interactions
between n-butane and surfaces are physisorption and the strength has no
relationship with the surface coverage. The structure parameters of n-
butane are listed in Table 2. For n-butane on different coverages, all the
lengths of the C
.10 Å, respectively. It is found the structure of the n-butane on the Pt
–
C and C–H bonds in n-butane maintain at 1.52 and
1
(
111) has no difference with the isolated n-butane (C-C: 1.52–1.53 Å,
C–H: 1.10–1.11 Å), which confirms that the catalyst has no effect on the
structure of n-butane.
Usually, the formation process from n-butane to butyl suffers from
two steps. The first step is the dehydrogenation in the methyl or meth-
ylene group of the n-butane and the second step is the formation of butyl
via β–dehydrogenation. This was demonstrated in a number of studies
that the dehydrogenation first H atom is from the top methyl group of n-
butane. Thus in this calculation, we considered the transition state of
dehydrogenation in the methyl group. The configurations of transition
Fig. 5. DRIFT spectra for the adsorption of CO on reduced Pt-TS-1-based cat-
alysts at room temperature. (a) IM-1, (b) EG-1, (c)EG-2, and (d) EG-3.
experimental conditions. As shown in Fig. 6 the conversion of n-butane
was found to increase with increasing reaction temperature since n-
butane dehydrogenation is thermodynamically limited and requires a
5

M. Shao et al.
Applied Catalysis A, General 621 (2021) 118194
Fig. 6. Catalytic performance of the TS-1-based catalyst for the dehydrogenation of n-butane. (a) IM-1, (b) EG-1, (c) EG-2, and (d) EG-3. Reacrion condition: 0.10 g,
n-butane : H
2
: Ar = 5: 1: 94, 50 mL⋅minꢀ 1⋅.
3
.4. Discussion
As mentioned above, catalysts prepared by the various methods
show different catalytic activities. However, the intrinsic factors that
influence the dehydrogenation process are unclear. Several reviews
report that n-butane dehydrogenation is a structure-sensitive reaction
[
34,35]. Previous research showed that particle size and the distribu-
tions of active components affect catalytic activity. For example,
Natarajan [36] reported that smaller Pt particles result in a higher cat-
alytic activity for n-butane dehydrogenation. We used various methods
to characterize our catalysts with the aim of elucidating the reason for
their different catalytic activities.
As shown in Table 1, the catalysts prepared by the ethylene glycol
prereduction method have similar surface areas, average pore sizes, and
pore volumes. To compare with IM sample, the surface areas and pore
volumes of EG samples are slightly higher. This is possible because the
pores in IM sample were blocked by some Pt species. For the pore vol-
ume of IM sample is smaller than that of TS-1 support. This is possibly
due the pores of the catalysts prepared by the impregnation method
being blocked by some Pt species [34,35]. However the values are close,
which is in good agreement with the lack of obvious peaks corre-
sponding to Pt particles in the XRD patterns of these catalysts. In order to
further study the details of the Pt species, TEM images were acquired. As
shown in Fig. 2, the Pt particles in the catalysts prepared by the ethylene
glycol prereduction method were uniform in size and nearly normally
distributed in the 1–3 nm range. More than 90 % Pt particle of the EG
samples are between 1 and 3 nm while that of IM sample with a much
wide distribution of 1 and 9 nm. Bsides, Pt dispersion has been improved
via the pre-reduction method, it extremely increased from 35 % for IM-1
to 60 % for EG-2. Among these catalysts, EG-2 showed the smallest mean
Fig. 7. Catalytic stabilities of the TS-1-based catalyst during the dehydroge-
nation of n-butane. (a) IM-1, (b) EG-1, (c) EG-2, and (d) EG-3. Reacrion con-
◦
: Ar = 5: 1: 94, 50 mL⋅minꢀ 1.
dition: 500 C, 0.10 g, n-butane : H
2
states on different coverages are shown in Fig. 9. It is found that the
active C
–
H bonds fall in the range of 1.54–1.56 Å, which are elongated
from 1.1 Å in gaseous n-butane. Moreover, the C-Pt distances for tran-
sition state are maintained at 2.27 Å, which are obviously closer than the
ones for initial states. The phenomenon indicates that the formation of
C Pt bond weakens the C C bond in n-butane.
– –
6

M. Shao et al.
Applied Catalysis A, General 621 (2021) 118194
Fig. 8. Side view of the adsorption configurations: n-butane adsorbed at the top slab of (a) p(2 × 2), (b) p(2 × 3), (c) p(3 × 3) and (d) p(3 × 4) supercell. The first
layer of Pt: silver. C: brown. H: white.
Pt size (1.81 nm) with dispersed Pt particles that are quite concentrated,
Table 2
which is consistent with the n-butane-dehydrogenation results.
Structure parameters of n-butane on different coverages.
As mentioned above, catalysts with smaller Pt particles are more
p(2 × 2)
p(2 × 3)
p(3 × 3)
p(3 × 4)
selective for butenes and butadiene [37]. As shown in Fig. 10, the
trans-butene and cis-butene selectivities are similar and do not vary with
the particle size of Pt.
a
d
d
d
C-Pt (Å)
3.611
1.526
1.109
3.977
1.527
1.106
3.579
1.526
1.100
3.580
1.525
1.100
b
C-C (Å)
c
C-H (Å)
These catalysts were analysed by XPS (Figs. 3 and 4) in order to
determine other factors that influence the dehydrogenation process. The
Pt 4f peak of the EG catalyst is observed at higher energy because Ti
withdraws electrons form Pt. The IM-1 and EG-2 catalysts exhibit
different Pt 4f and Ti 2p XPS peaks that reveal strong interactions be-
tween Pt and Ti species in the latter catalyst. Clearly, the Pt species in the
two catalysts are different, with the Pt in EG-2 in an electron-deficient
state. This may be responsible for the superior catalytic activity of the
a
b
c
d
C-Pt presents the shortest C-Pt distance.
C-C presents the length of the C C bonds in n-butane.
C-H presents the length of the C H bonds in n-butane.
d
d
–
–
Fig. 9. Side view of the transition state configurations for n-butane dehydrogenation on the top slab of (a) p(2 × 2), (b) p(2 × 3), (c) p(3 × 3) and (d) p(3 × 4)
supercell. The first layer of Pt: silver. C: brown. H: white.
7

M. Shao et al.
Applied Catalysis A, General 621 (2021) 118194
researched by changing the reaction temperature. The EG-2 shows the
smallest reaction activation energy. It shows that catalyst with the
smaller Pt particle size has the smaller reaction activation energy. The
trend of reaction activation energy with Pt particle size are matched
with the catalytic activity. The affect of Pt particle size are researched by
changing the different n-butane coverage on the Pt surface. As
mentioned in the density functional theory calculations, the catalyst
with small Pt particle size means big n-butane coverage. Further the
energy barriers for the first step activation of n-butane on different
coverages are explored, which are shown in Fig. 11. It is seen that the
energy barrier decreases with increasing the supercell size. Thus, we
conclude that the energy barrier of dehydrogenation could be weakened
by increasing the coverage. The catalyst with the smallest particles
shows the lowest energy barrier of the dehydrogenation process. This
trend is consistent with the affect of particle size for the dehydrogena-
tion reaction activity energy.
Fig. 10. Correlations between Pt-particle size and n-butane-dehydrogenation
◦
selectivities at 500 C.
EG catalysts than that prepared by impregnation. However many reports
on PtSn alloy-based catalysts show that electron-rich Pt particles favor
the n-butane dehydrogenation reaction [38,39]. And some researchers
4. Conclusion
Pt-TS-1-based catalysts prepared using the ethylene glycol pre-
reduction method show quite good catalytic activities for n-butane
dehydrogenation. The result exhibit selectivities above 97 % for butenes
2
have reported that TiO -based catalysts exhibit good catalytic activities
for superior n-butane adsorption [40]. The superior activity of the EG
catalysts is attributed to the influence of Ti in the TS-1 structure. DRIFT
spectra for the adsorption of CO on reduced Pt-TS-1-based catalysts were
used to illustrate the state of Pt particles. As shown in Fig. 5, CO bridge
◦
and butadiene, and a conversion of about 20 % at 500 C over the EG-2
catalyst. In addition, this catalyst is very stable, with activity that was
barely observed to decrease during dehydrogenation for 10 h. The cat-
alysts were characterized to confirm that those prepared by the ethylene
glycol prereduction method are superior. Ethylene glycol prereduction
effectively reduces the average size of the Pt particles. And the distri-
bution of Pt particle size are in a narrow area. As shown by TEM, the Pt
particles are only 1.81 nm in size and are concentrated in the 1–2 nm
range. Pt dispersion has been improved via pre-reduction method.
Interestingly, Pt dispersion was extremely increased from 35 % for IM-1
to 60 % for EG-2. Strong metal-support interactions between the Pt
species and the Ti species in the TS-1 support were observed by XPS and
DRIFTS, which increase both catalytic activity and stability. The reac-
tion activation energy for dehydrogenation reaction on different cata-
lysts were decreased by decreasing the Pt particle size. The density
functional theory calculations show that the energy barrier of dehy-
drogenation could be weakened by increasing the coverage of n-butane,
which are belonged to the decreasing Pt particle size.
ꢀ
1
bonding on Pt clusters exhibits a broad resonance at 1880 cm , was
only observed in the spectrum of the IM catalyst [41]. By comparing the
ꢀ 1
strength of peaks at 2110 cm , EG-2 sample which has the smallest Pt
particle conforms the strongest strength.CO-atop adsorption is observed
-
1
at higher wavenumber (2110 cm ) for EG catalysts due to electron
donation from Pt to Ti [42]. EG-2 shows the best catalytic activity for
n-butane dehydrogenation may caused by the Pt which shows weakly
-
1
electropositivity. However some reports interpret the peak at 2110 cm
+
to correspond to CO linearly absorbed on Ptδ single-atom sites on metal
oxides [43]. However there is lack of enough characterization analysis
to prove whether Pt is single atom or not. This is also possibly due to the
smallest Pt particles leading to the strongest Pt/Ti interactions in the
TS-1 support among the EG catalysts [23,29]. The strong metal-support
interaction is helpful to improve the stability of catalysts [34]. It is worth
pointing out that the CO-adsorption behavior is very consistent with
how the initial activities of the catalysts depend on Pt particle size. The
catalyst with smaller Pt particle size exhibit higher n-butane conversion
and butenes selectivity.
CRediT authorship contribution statement
Although these values are a very rough approximation of the real
TOF values of the catalyst, all the TOFs of EG samples are bigger than
Mingyuan Shao: Methodology, Investigation, Formal analysis, Data
curation, Visualization, Writing - original draft, Writing - review &
editing. Chaoquan Hu: Funding acquisition, Project administration,
Investigation, Formal analysis, Writing - original draft. Xuebing Xu:
ꢀ 1
that of the IM-1 sample (428.4 h ). Note that EG-2 exhibited not only
ꢀ 1
the highest TOF (694.8 h ) and good durability but also the highest n-
butane conversion owing to the highest Pt dispersion. The trend of TOFs
for n-butane dehydrogenation is close to the trend of reaction activation
energy. The trend of reaction activation energy with Pt particle size are
matched by the catalytic activity. As shown in Table. 3, the reaction
activation energy for dehydrogenation reaction on different catalysts are
Table 3
Correlations between Pt-particle size, reaction activation energy of n-butane
dehydrogenation and TOF values.
Catalysts
IM-1
EG-1
EG-2
EG-3
a
ꢀ 1
TOF /h
428.4
98.64
5.62
568.8
81.07
2.73
694.8
74.25
1.81
507.6
93.33
3.56
b
ꢀ 1
Ea /kJ mol
Pt average
size/nm
a
◦
TOF values of catalysts at 500 C after 25 min of time on stream. Reaction
, 50 mL/min, 50 mg catalyst.
condition: 5 % butane, 1 % H
b
2
The activation energy are obtained by using the n-butane conversions lower
than 10 % under same conditions as that for TOF values.
Fig. 11. Energy barrier as a function of the supercell size.
8

M. Shao et al.
Applied Catalysis A, General 621 (2021) 118194
Methodology. Yang Song: Investigation, Validation. Qingshan Zhu:
Funding acquisition, Methodology, Project administration, Supervision,
Validation, Writing - review & editing.
[14] H. Seo, J. Lee, U. Hong, G. Park, Y. Yoo, J. Lee, H. Chang, I. Song, Catal. Commun.
4
7 (2014) 22–27.
15] G. Raju, B. Reddy, S. Park, J CO
16] B. Ajayi, B. Abussaud, R. Jermy, S. Khattaf, Prog. React. Kinet. Mec. 39 (2014)
[
[
2
Util. 5 (2014) 41–46.
3
41–353.
[
[
[
17] D. Shee, G. Deo, Catal Lett. 124 (2008) 340–351.
Declaration of Competing Interest
The authors report no declarations of interest.
Acknowledgments
18] Z. Li, W. Ma, Q. Zhong, Ind. Eng. Chem. Res. 58 (2019) 4010–4016.
19] J. Li, J. Li, Z. Zhao, X. Fan, J. Liu, Y. Wei, A. Duan, Z. Xie, Q. Liu, J. Catal. 352
(2017) 361–370.
[
[
20] W. Wannapakdee, T. Yutthalekha, P. Dugkhuntod, K. Rodponthukwaji,
A. Thivasasith, S. Nokbin, T. Witoon, S. Pengpanich, C. Wattanakit, Catalysts 9
(
2019) 174.
21] H. Chen, R. Zhang, H. Wang, W. Bao, Y. Wei, Appl. Catal. B Environ. 278 (2020),
19311.
1
The authors are grateful for the financial support from the National
Natural Science Foundation of China (22078328).
[
[
22] C. Brien, G. Jenness, H. Dong, D. Vlachos, I. Lee, J. Catal. 337 (2016) 122–132.
23] E. Ogel, M. Casapu, D. Doronkin, R. Popescu, H. Sto ¨r mer, C. Mechler, G. Marzun,
S. Barcikowski, M. Tu ¨r k, J. Grunwaldt, J. Phys. Chem. C. 123 (2019) 5433–5446.
24] A. Talati, M. Haghighi, F. Rahmani, RSC Adv. 6 (2016) 44195–44204.
28] X. Zhang, N. He, C. Liu, H. Guo, Catal Lett. 149 (2019) 974–984.
29] L. Cesar, C. Yang, Z. Lu, Y. Ren, G. Zhang, J. Miller, ACS Catal. 9 (2019)
5231–5244.
[
[
[
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2021.118194.
[30] L. Deng, H. Miura, T. Shishido, S. Hosokawa, K. Teramura, K. Tanaka,
ChemCatChem 6 (2014) 2680–2691.
[
[
31] Q. Zhang, S. Mo, Q. Li, Y. Sun, M. Zhang, P. Chen, M. Fu, L. Wu, L. Chen, D. Ye,
Catal. Sci. Technol. 9 (2019) 4538–4551.
References
32] J. Jones, H. Xiong, A. DeLaRiva, E. Peterson, H. Pham, S. Challa, G. Qi, S. Oh,
Science 353 (2016) 150–154.
[
1] J. Rischard, C. Antinori, L. Maier, O. Deutschmann, Appl. Catal. A Gen. 511 (2016)
3–30.
2
[33] D. Shee, G. Deo, Catal. Today 325 (2019) 25–32.
[
[
2] S. Furukawa, M. Endo, T. Komatsu, ACS Catal. 4 (2014) 3533–3542.
3] J. McGregor, Z. Huang, G. Shiko, L.F. Gladden, R.S. Stein, M. Duer, J.Z. Wu, P.
C. Stair, S. Rugmini, S.D. Jackson, Catal. Today 142 (2009) 143–151.
4] Y. Tian, X. Liu, M. Rood, Z. Yan, Appl. Catal. A Gen. 545 (2017) 1–9.
5] J. Camacho-Bunquin, S. Ferrandon, H. Sohn, J. Kropf, C. Yang, J. Wen, R. Hackler,
C. Liu, G. Celik, L. Marshall, C. Stair, M. Delferro, ACS Catal. 8 (2018)
[34] L. Deng, H. Miura, T. Ohkubo, T. Shishido, Z. Wang, S. Hosokawa, K. Teramura,
T. Tanaka, Catal. Sci. Technol. 9 (2019) 947–956.
[35] Y. Qiu, X. Li, Y. Zhang, C. Xie, S. Zhou, R. Wang, S. Luo, F. Jing, W. Chu, Ind. Eng.
Chem. Res. 58 (2019) 10804–10818.
[
[
[36] P. Natarajan, H. Khan, S. Yoon, K. Jung, J. Ind. Eng. Chem. 63 (2018) 380–390.
[37] T. Bauer, S. Maisel, D. Blaumeiser, J. Vecchietti, N. Taccardi, P. Wasserscheid,
P. Bonivardi, A. Gorling, J. Libuda, ACS Catal. 9 (2019) 2842–2853.
[38] B. Nagaraja, C. Shin, K. Junga, Appl. Catal. A Gen. 467 (2013) 211–223.
[39] A. Ballarini, P. Zgolicz, I. Vilella, S. Miguel, A. Castro, O. Scelza, Appl. Catal. A Gen.
381 (2010) 83–91.
1
0058–10063.
[
[
[
[
6] Q. Zhu, G. Wang, H. Zhang, X. Zhu, C. Li, Appl. Catal. A Gen. 566 (2018) 113–120.
7] B. Nagaraja, C. Shin, K. Jung, Appl. Catal. A Gen. 467 (2013) 211–223.
8] W. Liu, S. Lai, H. Dai, S. Wang, H. Sun, C. Au, Catal. Today 131 (2008) 450–456.
9] M. Shao, C. Hu, J. Sun, S. Xu, M. Xiang, Q. Zhu, Ind. Eng. Chem. Res. 591 (2020)
[40] D. Shee, G. Deo, J. Mol. Catal. A-Chem. 308 (2009) 46–55.
[41] Y. Pei, Z. Qi, T. Goh, L. Wang, R. Maligal-Ganesh, H. MacMurdo, S. Zhang, C. Xiao,
X. Li, F. Tao, D. Johnson, W. Huang, J. Catal. 356 (2017) 307–314.
[42] B. Yan, X. Yang, S. Yao, J. Wan, M. Myint, E. Gomez, Z. Xie, S. Kattel, W. Xu,
J. Chen, ACS Catal. 6 (2016) 7283–7292.
7
1–80.
10] V. Murgia, E. Torres, J. Gottifredi, E. Sham, Appl. Catal. A Gen. 312 (2006)
[
1
34–143.
[
[
11] M. Dixit, P. Kostetskyy, G. Mpourmpakis, ACS Catal. 8 (2018) 11570–11578.
12] J. Zhang, Y. Deng, X. Cai, Y. Chen, M. Peng, Z. Jia, Z. Jiang, P. Ren, S. Yao, J. Xie,
D. Xiao, X. Wen, N. Wang, H. Liu, D. Ma, ACS Catal. 9 (2019) 5998–6005.
[43] G. Sun, Z. Zhao, R. Mu, S. Zha, L. Li, S. Chen, K. Zang, J. Luo, Z. Li, S. Purdy,
A. Kropf, J. Miller, L. Zeng, J. Gong, Nature Commun. 9 (2018) 4454.
[
13] B. Vu, M. Song, I. Ahn, Y. Suh, D. Suh, W. Kim, H. Koh, Y. Choi, Appl. Catal. A Gen.
4
00 (2011) 25–33.
9