Effect of potassium addition on bimetallic PtSn supported θ-Al 2O3 catalyst for n-butane dehydrogenation to olefins

Article abstract of DOI:10.1016/j.cattod.2013.10.070

PtSn/θ-Al₂O₃ catalysts with different amount of potassium (0.4, 0.7, 0.95, 1.2 and 1.45 wt.%) were prepared by an impregnation method, and their catalytic activity in n-butane dehydrogenation was investigated at 823 K, an atmospheric pressure and a GHSV of 18,000 mL(g _cat h)^-1. The compositions listed in order of n-C ₄⁼ yields at 823 K were as follows: K_0.95(PtSn) _1.5 > (PtSn)_1.5 > K_0.4(PtSn) _1.5 > K_0.7(PtSn)_1.5 > K _1.2(PtSn)_1.5 > K_1.45(PtSn)_1.5 > K_0.9(Pt)_1.5. The K_0.9(Pt)_1.5 and K_0.95(Sn)_1.5 catalyst severely deactivated in n-butane dehydrogenation. The (PtSn)_1.5 (without K) catalyst showed the highest n-butane conversion, while K_0.95(PtSn)_1.5 did the highest n-C₄⁼ yield. The small amount of potassium on bimetallic PtSn/θ-Al₂O₃ catalyst improved n-C ₄⁼ selectivity, but slightly decreased n-butane conversion, resulting in the increase of n-C₄⁼ yield. The effect of potassium was caused by blocking the acid sites of Pt catalyst. The TPR and HAADF STEM-EDS study suggested the reduction procedure of the Pt, Sn and K species. However, the higher loaded potassium (1.2 and 1.45 wt.%) doped (PtSn)_1.5 catalysts were rather highly deactivated because the sizes of Pt particles were increased by weakening the interaction between Pt and Sn. The n-C₄⁼ selectivity of the (PtSn)_1.5 catalyst increased with respect to the reaction, while that of the potassium doped catalysts maintained the high n-C₄⁼ selectivity from the beginning of the reaction. Also, different alkali metals (Ca, Na and Li) were tested for the comparison with K. The potassium doped catalyst showed the highest n-C₄⁼ yield among the other alkali metals for n-butane dehydrogenation.

Full text of DOI:10.1016/j.cattod.2013.10.070

Catalysis Today 232 (2014) 40–52
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Effect of potassium addition on bimetallic PtSn supported ␪-Al2O3
catalyst for n-butane dehydrogenation to olefins
Bhari Mallanna Nagarajaa, Heon Junga,b, Dae Ryook Yangb, Kwang-Deog Junga,∗
a Clean Energy Research Centre, Korea Institute of Science and Technology, P.O. Box 131 Cheongryang, Seoul 136-791, Republic of Korea
b Department of Chemical & Biological Engineering, Korea University, Anam-dong, Seongbuk-gu, Seoul 136-701, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 5 July 2013
Received in revised form 15 October 2013
Accepted 27 October 2013
Available online 4 December 2013
PtSn/␪–Al2O3 catalysts with different amount of potassium (0.4, 0.7, 0.95, 1.2 and 1.45 wt.%)
were prepared by an impregnation method, and their catalytic activity in n-butane
dehydrogenation was investigated at 823 K, an atmospheric pressure and
a
GHSV of
18,000 mL(gcat h)−1
.
The compositions listed in order of n-C4 yields at 823 K were as follows:
=
K
0.95(PtSn)1.5 > (PtSn)1.5 > K0.4(PtSn)1.5 > K0.7(PtSn)1.5 > K1.2(PtSn)1.5 > K1.45(PtSn)1.5 > K0.9(Pt)1.5
.
The
K
0.9(Pt)1.5 and K0.95(Sn)1.5 catalyst severely deactivated in n-butane dehydrogenation. The (PtSn)1.5
Keywords:
n-Butane dehydrogenation
n-Butenes
KPtSn/␪-Al2O3 catalyst
Effect of alkali metal
PtSn alloy formation.
=
(without K) catalyst showed the highest n-butane conversion, while K0.95(PtSn)1.5 did the highest n-C4
yield. The small amount of potassium on bimetallic PtSn/␪-Al2O3 catalyst improved n-C4= selectivity, but
=
slightly decreased n-butane conversion, resulting in the increase of n-C4 yield. The effect of potassium
was caused by blocking the acid sites of Pt catalyst. The TPR and HAADF STEM-EDS study suggested
the reduction procedure of the Pt, Sn and K species. However, the higher loaded potassium (1.2 and
1.45 wt.%) doped (PtSn)1.5 catalysts were rather highly deactivated because the sizes of Pt particles were
=
increased by weakening the interaction between Pt and Sn. The n-C4 selectivity of the (PtSn)1.5 catalyst
increased with respect to the reaction, while that of the potassium doped catalysts maintained the high
=
n-C4 selectivity from the beginning of the reaction. Also, different alkali metals (Ca, Na and Li) were
=
tested for the comparison with K. The potassium doped catalyst showed the highest n-C4 yield among
the other alkali metals for n-butane dehydrogenation.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
for isobutane dehydrogenation by reducing the size of surface Pt
ensembles [10]. Siri et al. studied the modifying effects of alkaline
The dehydrogenation of butane (n- and iso) into olefins is a sig-
nificant process used for the industrial application such as plastics,
synthetic rubbers, automotive fuel components, and petrochemi-
cal products [1–4]. Butenes are used for the production of methyl
tertiary butyl ether and methacrylates as an additive to gasoline
to improve the octane number [5]. In particular, n-butenes from
the dehydrogenation of butane can be used as a feedstock for the
production of 1,4-butadiene. Many authors reported on PtSn cat-
alysts with various supports [3,4,6–9] and alkali promoted (K, Li,
Mg, and La) [5,10–18] PtSn catalysts for isobutane dehydrogena-
tion reaction. Llorca et al. reported the magnesia supported PtSn
catalysts with different Pt content for isobutane dehydrogenation
reaction. The conversion and selectivity of Pt catalyst was increased
with the addition of Sn, which was explained by the dilution and
electron effect on Pt by Sn [9]. Cortright and Dumesic reported that
the addition of potassium to Pt/Sn/SiO2 increased the selectivity
metals on the acid properties of the support and on the behavior
of PtSn/␥-Al2O3 catalytic systems for isobutane dehydrogenation
[11]. They reported that the isopropyl alcohol transformation rate
decreased drastically with the addition of alkaline metals according
to the sequence of Li < Na < K. The addition of K improved the selec-
tivity to butenes as well as the stability of the catalyst. Zhang et al.
observed that the suitable addition of potassium and tin resulted
in an increase in the platinum dispersion and a decrease in the
catalyst acidity, which was due to the synergistic effect between
potassium and tin [12]. The synergistic effect could prevent coke
formation that covered the active metal and facilitate the transfer
of the deposited carbon from the metal to the support. Tasbihi et al.
reported the effect of potassium and lithium addition on catalytic
performance of Pt–Sn/Al2O3 catalyst for isobutane dehydrogena-
tion [14]. The potassium doped catalyst showed better activity as
compared with lithium one, which was mainly related to their
structural characteristics and electronic properties. The improved
selectivity to isobutene was suggested by reducing acidity of the
support. Rodriguez et al. reported that the incorporation of Sn
and Li also improved the activity and selectivity for light-paraffin
∗
Corresponding author. Tel.: +82 2 9585218; fax: +82 2 9585219.
E-mail address: jkdcat@kist.re.kr (K.-D. Jung).
0920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2013.10.070

B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
41
dehydrogenation by inhibiting isomerization and cracking reaction
[15]. Zhang and co-workers reported that the presence of Mg or
La in the PtSnK/Al2O3 catalysts increased the activity and selectiv-
ity of butenes for isobutane dehydrogenation reaction [17,18]. The
addition of Mg or La stabilized the oxidation states of Sn species,
increasing the platinum dispersion and decreasing carbon depo-
sition. The addition of alkali metals to oxide catalysts has been
found to increase selectivity to olefins in isobutane dehydrogena-
tion reactions. The alkali metals such as Li, Na and K could improved
the catalytic activity, selectivity and stability for isobutene dehy-
drogenation [11,12]. The important phenomenon was attributed
to a geometric modification of the metallic phase by the alkali
metals, which could effectively inhibited the acidity on the supports
[11,14,19,20].
supplied by Hyosung Company, South Korea). The ␪-Al2O3 support
was added into the beaker with Pt and Sn solution. Ethanol was
removed by evaporation on a hot plate, and the residue was dried
at 393 K overnight and calcined at 773 K for 4 h in air. The calcined
␪-alumina supported PtSn catalyst was designated as (PtSn)1.5
.
Samples of the support material containing different amounts
of potassium were prepared by adding the appropriate quanti-
ties of aqueous solutions of potassium hydroxide (Fluka, Purity = ≥
86%) to (PtSn)1.5, so that the final resultant materials contained
0.4, 0.7, 0.95 1.2 and 1.45 wt.% of K; after drying the samples at
393 K overnight the catalysts were calcined in air at 773 K for
4 h. The samples were designated as K0.4(PtSn)1.5, K0.7(PtSn)1.5
,
K0.95(PtSn)1.5, K1.2(PtSn)1.5 and K1.45(PtSn)1.5 and mono metallic
catalysts were designated as K0.9(Pt)1.5 and K0.95(Sn)1.5. The potas-
sium, Pt and Sn contents of the samples (Table 1) were determined
by an atomic absorption spectroscopy. The other alkali metals such
as 1.0 wt.% Na, Ca and Li supported (PtSn)1.5 catalysts were also pre-
pared by an impregnation method. The samples were designated
as Ca0.95(PtSn)1.5, Na0.95(PtSn)1.5 and Li0.95(PtSn)1.5. NaOH (Junsei
Chemicals, Co. Ltd., Purity: 97%), Ca(NO3)2.4H2O (Kanto chemicals
Co. Inc., Purity:98.5%) and LiNO3 (Junsei Chemicals, Co. Ltd., Purity:
98%) were used for the impregnation.
A few researchers have studied n-butane dehydrogenation reac-
tion using PtSn catalyst with different supports [21–25]. Bocanegra
et al. reported bimetallic PtSn and trimetallic InPtSn catalysts with
different supports (␥-Al2O3, ZnAl2O4 and MgAl2O4) for n-butane
dehydrogenation. The addition of In and/or Sn on Pt sites induced
geometric effects resulted from a dilution of the Pt surface, lead-
ing to the formation of alloys and/or intermetallic compounds to
increase the catalytic activity [21–23]. Ballarini et al. have stud-
ied on the n-butane dehydrogenation reaction with PtSn and PtGe
supported on ␥-Al2O3 deposited on sphere shaped ␣-Al2O3 by a
wash coating method [22,23]. In their report, it was observed that
the catalytic activity was improved by the addition of Sn on 0.5Pt
and Ge on ␥-Al2O3/␣-Al2O3 support. The increases of the catalytic
activity mainly depended on the composition of Sn and Pt, which
exhibited low electronic interaction in probable alloys [24,25].
Recently, we studied the selective and stable bimetallic (PtSn)x.x
and Pt1.5Snx.x supported ␪-Al2O3 catalysts for dehydrogenation
of n-butane to n-butenes [26]. 1.5 wt.% (PtSn)/␪-Al2O3 catalyst
2.2. Characterizations of catalysts
The XRD patterns of the prepared PtSn/␪-Al2O3 catalysts were
recorded on a diffractometer (M/S, Shimadzu Instruments, Japan)
˚
with Ni-filtered CuK␣ (␭ = 1.5418 A) as a radiation source. The
operating voltage was 40 kV and the current was 30 mA with
a scanning rate of 2␪ was 2◦ min−1. The BET surface areas and
N2 adsorption-desorption measurements were performed at 77 K
using an automated gas sorption system (Belsorp II mini, BEL Japan,
Inc.,). The Barrett–Joyner–Halanda (BJH) method was used for the
pore size distribution. The TPR experiments were carried out using
a temperature program analyzer (BELCAT, BEL Japan, Inc.).
The NH3-TPD was performed to determine the acidic prop-
erties of the catalysts; the 0.1 g of calcined sample was loaded
into a U-tube quartz reactor. Prior to the NH3 adsorption, the
sample was pre-treated at 773 K for 1 h with the flow of helium
(30 mL min−1) in order to remove the moisture, physically adsorbed
water and other impurities. The sample was cooled down to 373 K
in a stream of helium. After cooling, the ammonia (5%NH3/He,
flow: 30 mL min−1) was introduced into the reactor for 1 h. The
weakly adsorbed ammonia was removed at 373 K for 1 h. Then,
the temperature was ramped from 373 to 1073 K with a heating
rate of 10 K min−1 under the flow of helium (30 mL min−1). The
NH3 desorption profile was exhibited in a thermal conductivity
detector.
For TPR studies, 0.1 g of a calcined sample was placed between
quartz wool in a U-type quartz reactor. The sample was thermally
treated under an Ar stream at 673 K for 2 h to remove physically
adsorbed water and other impurities. The catalysts were cooled
down to the room temperature under pure Ar gas. After the pre-
treatment, the catalysts were heated at 10 K min−1 from room
temperature up to 1073 K in 5% H2/Ar stream with a flow rate
of 30 mL min−1. The chemisorptions of CO were performed with a
pulse chemisorption mode (BELCAT, BEL Japan, Inc.). Prior to mea-
surements, 0.050 g of a sample was thermally treated under a He
stream at 773 K for 50 min to remove physically adsorbed water
and other impurities. The sample was cooled down to room tem-
perature and heated to 823 K with a heating rate of 10 K min−1 using
pure H2 at a flow rate of 50 mL min−1. The sample was then reduced
in a pure H2 flow at 823 K for 2 h. After the reduction at 823 K,
the sample was purged with He gas at the same temperature for
1 h. After cooling to 323 K, the 10% CO/He gas was introduced for
the CO chemisorption. CO loop gas was used for each pulse, and
=
showed highest n-C4 yield and the catalyst was stable among
the other samples for n-butane dehydrogenation reaction. So far,
there were little reports on potassium doped PtSn/␪-Al2O3 cata-
lyst for n-butane dehydrogenation reaction. In the present work,
the different wt.% of potassium doped on 1.5 wt.% (PtSn)/␪-Al2O3
catalysts were studied for n-butane dehydrogenation reaction. The
=
0.95 wt.% K doped catalyst showed maximum n-C4 yield among
the other potassium loaded samples. XRD patterns and TEM images
confirmed the presence of PtSn alloy on the prepared bimetallic
PtSn catalysts. The Pt, Sn and K particles were clearly observed
at the same position, which was confirmed by HAADF STEM and
corresponding EDS mapping. The n-butane dehydrogenation was
conducted at different temperatures (773–873 K), and the activity
of the catalysts was explained by the characteristics of the cata-
lysts in CO chemisorption, XRD and TPR analysis. We were also
studied on the different alkali metals (K, Ca, Na and Li) doped
(PtSn)1.5/␪-Al2O3 catalyst for n-butane dehydrogenation reaction.
=
The potassium doped catalyst showed the better n-C4 yield and
=
n-C4 selectivity among the other catalysts.
2. Experimental
2.1. Preparation of the catalyst
The equal weight percentage of Pt and Sn (1.5 wt.%) was sup-
ported on ␪-Al2O3 by co-impregnation method, using hydrogen
hexachloroplatinate(IV) hydrate (H2PtCl6. nH2O, n = 5.8, Kojima
Chemicals Co., Ltd., Japan, purity = 99%) and tin(II) chloride dihy-
drate (SnCl2. 2H2O, Sigma-Aldrich, St. Louis, USA, purity = 98%) salt
as precursors. Weighed amount of Pt and Sn salt was placed in
a 50 ml beaker and dissolved in a required quantity of ethanol.
Sphere-shaped ␪-Al2O3 support (∼3 mm) was prepared by cal-
cining the spherical ␥-Al2O3 at 1273 K for 6 h (␥-Al2O3 support was

42
B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
Table 1
Physicochemical properties of mono and bimetallic 1.5 wt.% (PtSn)/␪-Al2O3 catalysts with different potassium loadings.
Catalysts
From ICP (wt.%)
BET surface area (m2 gm−1
)
Pore volume of calcined
Average pore diameter of
calcined catalysts (nm)
catalysts (cm3 g−1
)
Pt
Sn
K
Calcined
Reduced
Used
(PtSn)1.5
K0.4(PtSn)1.5
1.3
1.3
1.3
1.3
1.3
1.3
1.35
–
1.35
1.35
1.35
1.35
1.35
1.35
–
–
86
81
80
80
81
81
84
83
82
79
78
79
79
78
84
82
84
78
76
78
81
76
84
80
0.327
0.302
0.301
0.303
0.298
0.297
0.326
0.323
15.1
15.3
15.3
15.3
14.6
14.7
15.6
15.5
0.40
0.70
0.95
1.2
1.45
0.90
0.95
K
K
K
0.7(PtSn)1.5
0.95(PtSn)1.5
1.2(PtSn)1.5
K1.45(PtSn)1.5
K0.9(Pt)1.5
K0.95(Sn)1.5
1.35
the pulse injections were repeated until saturation. The amount
of CO was measured using a thermal conductivity detector. The Pt
dispersion of each catalyst was then calculated from the amount
of CO adsorbed, taking the stoichiometry factor (SF) as one for
3. Results and discussion
3.1. Catalysts characterization
CO/Ptatm
:
3.1.1. Physical properties of the catalyst
Figs. 1 and 2 show the calcined and reduced XRD patterns of
(PtSn)1.5 catalyst with different amount of K loadings. No XRD
peaks from Pt/PtO or Sn/SnO2 or K phases were not observed in
the calcined catalysts (Fig. 1). Only ␪-Al2O3 phases appeared at
2␪-values of 31.05◦, 33.05◦, 37.15◦, 39.52◦, 45.10◦, 48.00◦, 60.20◦
and 67.58◦ (ICDD file no.: 47-1771), it is because the metal com-
ponents were highly dispersed on the support and/or too small to
be detected by XRD techniques. The 2␪ values of Pt were at 39.76◦,
46.24◦ and 67.45◦ (ICDD file no.: 04-0802), and those of Sn were at
15.57◦, 16.35◦, 22.10◦, 27.64◦ and 28.99◦ ((ICDD file no.: 19-1365).
No characteristic peaks of Pt/PtOx and Sn/SnOx were observed in the
XRD patterns of the reduced potassium doped (PtSn)1.5 catalysts,
too (Fig. 2). The reduced mono- and bimetallic potassium doped
(PtSn)1.5 catalysts showed ␪-Al2O3 (ICDD file no.: 47-1771) and
PtSn (ICDD file no.: 25-0614) phases with the 2␪-values at 25.08◦,
30.07◦ and 41.88◦ with hexagonal structure, indicating the forma-
tion of PtSn alloy during the reduction. Additionally, no diffraction
peaks assigned to the crystalline potassium were detected, sug-
gesting that potassium species were highly dispersed on ␪-Al2O3
support.
Dispersion (%) = 100 × Vs × SF × MW/(SW × Fn × 22414)
where Vs is the cumulative volume of adsorbed CO (cm3 at STP),
MW is the molecular weight of Pt metal (g mol−1), SW is the weight
of sample and Fn is the Pt fraction in relation to the total catalyst
sample weight.
Transmission electron microscopy (TEM) is a technique to cal-
culate the Pt particle size by using CM30 microscope (Philips)
with accelerating voltage 50–300 kV using LaB6 source. The high
resolution TEM images of the Pt and Sn nanoparticles were charac-
terized by using FEI (Technai F20 G2, The Netherland) microscopy
working in a STEM mode (Accelerating voltage: 50–200 kV, Image
resolution: < 0.23 nm, Electro probe size: < 0.3 nm, Magnification:
25–1,030,000×, Specimen double tilting: 40◦/ 20◦). The cata-
lyst sample was dispersed in ethanol solution and the suspension
was carried out using ultrasonic bath for 30 min. The resulting
solution was deposited on copper grid coated with a carbon film,
and the alcohol was evaporated. Finally, the sample containing
the copper grid was used for the TEM and HRTEM/HAADF STEM
analysis. CHNS analysis was carried out using a Fisions EA1108
instrument for the determination of the amount of accumulated
carbon on the surface of the catalyst after the reaction. The Pt, Sn
and K were determined by atomic absorption apectroscopy (AAS)
and inductively coupled plasma-atomic emission spectrometer
(ICP-AES).
Table 1 shows the physical properties such as BET surface area
and pore size distribution. The loading amount (Pt, Sn and K) of
the prepared catalysts were analyzed using AAS/ICP analysis. The
BET surface area, pore volume and average pore diameter of ␪-
Al2O3 were 98 m2 g−1, 0.361 cm2 g−1 and 14.6 nm, respectively. The
surface areas of the calcined, reduced and used (PtSn)1.5 catalyst
2.3. Activity measurement
The prepared catalysts were first crushed and sieved. The final
particle sizes were between 80–140 mesh before the activity test.
The n-butane dehydrogenation was carried out at 823 K at atmo-
spheric pressure using 0.1 g of the catalyst after the H2 reduction.
The internal diameter of a fixed bed tubular quartz reactor was
10 mm, and the catalyst bed length between the quartz wool
layers in the reactor was approximately 3–4 mm. Before each
experiment, the sample was reduced in a pure hydrogen (99.99%)
flow (30 mL min−1). The temperature during the reduction was
increased slowly from room temperature to 823 K (10 K min−1) and
maintained at 823 K for 2 h. After the reduction, H2:N2:n-C4H10
feed gases with a molar ratio of 1.0:1.0:1.0 to were introduced
into the catalytic reactor for the reaction, and the total gas flow
rate was kept constant at 30 mL min−1 (GHSV: 18,000 mL/gcat/h).
The products and reactants were analyzed with an on-line GC (FID
detector, Series M 600D, South Korea) equipped with a capillary col-
umn (GS-Alumina, Agilent Technologies, USA, i.d.: 0.53 mm, length:
50 m).
Fig. 1. XRD profiles of the calcined PtSn/␪-Al2O3(with and without K) catalysts. (a)
␪-Al2O3, (b) K0.95(Sn)1.5, (c) K0.9(Pt)1.5, (d) (PtSn)1.5, (e) K0.4(PtSn)1.5, (f) K0.7(PtSn)1.5
(g) K0.95(PtSn)1.5, (h) K1.2(PtSn)1.5 and (i) K1.45(PtSn)1.5
,
.

B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
43
Fig. 2. XRD profiles of the reduced PtSn/␪-Al2O3(with and without K) catalysts; (a) ␪-Al2O3, (b) K0.95(Sn)1.5, (c) K0.9(Pt)1.5, (d) (PtSn)1.5, (e) K0.4(PtSn)1.5, (f) K0.7(PtSn)1.5, (g)
K0.95(PtSn)1.5, (h) K1.2(PtSn)1.5 and (i) K1.45(PtSn)1.5
.
samples were 86, 82 and 84 m2 g−1, respectively. The surface areas
of the samples containing potassium were about 79–81 m2 g−1. The
BET surface area of the (PtSn)1.5 catalyst was not so much changed
with an increase in potassium content. The pore volume (327 to
297 cm3 g−1) and pore size (15.1 to 14.7 nm) were slightly varied
with the increase in potassium content.
temperature, hydrogen consumption and percentage of the reduc-
tion of potassium doped mono- and bimetallic catalysts. All the
TPR profiles were de-convoluted using Lorentzian–Gaussian func-
tions. The mono metallic K0.9(Pt)1.5 (Fig. 4a) catalyst showed the
low temperature reduction peak at 478 K with a small shoulder
at 400 K, which was assigned as the reduction of the Pt species in
the weak interaction with the support. The other peak at 521 and
633 K was assigned to the reduction of the Pt species in the strong
interaction with the support [23]. The reduction peaks at 714 and
823 K can be attributed to oxychlorinated Pt species. [24,29,30].
A small reduction peak for a finely dispersed Sn species at 530 K
appeared with K0.95(Sn)1.5 catalyst (Fig. 4b). The broad reduction
peak at 712 and 752 K could be due to the reduction of Sn4+/Sn2+
and/or Sn2+/Sn0 species, respectively [17,31]. The broad peak at
886 K can be assigned to the reduction of Sn2+ to Sn0 in a high
interaction with the support [5,32].
It is observed that the (PtSn)1.5 without K (Fig. 4c) showed the
reduction peaks at 476, 548, 677, and 946 K. The new reduction
peak at 476 K (␣) can be assigned to the co-reduction of the Pt
species with Sn species, which is mainly related to the Pt species in
the weak interaction with the support. The peak at 548 (␤) can be
assigned to the co-reduction of the Pt species with Sn species, which
is related mainly to the Pt species in the strong interaction with
the support. The reduction peak at 677 (␥) can be assigned mainly
to the reduction of isolated Sn4+/Sn2+ and/or Sn2+/Sn0 species. The
reduction peak at 946 K (␦) can be assigned mainly to the reduction
of Sn2+/Sn0 species with the strong interaction with the support,
respectively. The K0.4(PtSn)1.5 catalyst (Fig. 4d) showed the reduc-
tion peaks at 520, 579, 753 and 787 K. The reduction peaks at
520 K (␣) and 579 K (␤) can be assigned to the reduction of the
Pt with Sn and K species [27]. The reduction peaks at 753 and
787 K (␥) can be assigned mainly to the reduction of Sn4+/Sn2+
and/or Sn2+/Sn0 species, respectively. The K0.7(PtSn)1.5 (Fig. 4e) cat-
alyst showed ␣, ␤, ␥, and ␦ peak at 515, 611, 786 and 809 K. The
K0.95(PtSn)1.5 (Fig. 4f) catalyst also showed ␣, ␤, ␥, and ␦ peak at
503, 688 and 830 K. The similar trend was observed in K1.2(PtSn)1.5
and K1.45(PtSn)1.5 (Fig. 4g and h) catalyst. The reduction peak at low
temperature (␣ and ␤) shifted towards higher temperature with
the addition K. The reduction peak area of Sn species at ␦ increased
and the reduction temperature decreased with the addition of K,
indicating that the addition of potassium promoted the reduction
of Sn species [14]. The hydrogen consumption increased with the
increase of potassium content because the reduction amount of
Sn species increased with the addition of potassium (Table 3). The
experimental result indicates that the excess addition of K weak-
ened the interaction between Sn and the support, which caused
low catalytic activity and stability. The addition of K could also give
an effect on the interaction between Pt and Sn as Perak et al. [33]
3.1.2. NH3-TPD studies
Fig. 3 shows the NH3 TPD profiles of ␪-Al2O3, (PtSn)1.5 and
1.0 wt.% of K, Ca, Na and Li supported (PtSn)1.5 catalysts. The
catalytic activity of n-butane dehydrogenation reaction was
mainly related to the acidity of the catalysts. All the NH3 TPD
profiles were de-convoluted using Lorentzian–Gaussian functions.
Table 2 shows the NH3 uptakes at different temperatures for the
prepared samples. All the samples exhibited three peaks. The peak
(I) at 464–480 K is assigned as weak acidic sites, the peak (II) at
504–522 K and 557–587 K can be attributed to medium acidic
sites, and the final peak (III) at 613–692 K can be considered as
strong acidic sites [5]. The peaks at 760–973 K were assigned as
desorption of the strongly adsorbed water or as de-hydroxylation
(Fig. 3). For the confirmation of the peaks at 760–973 K, tem-
perature programmed desorption with ␪-Al2O3 was performed
without NH3 adsorption. The helium gas was introduced into
the reactor with the alumina support at 373 K for 1 h. Then, the
temperature was ramped from 323 to 1073 K with a heating
rate of 10 K min−1 under the flow of helium (30 mL min−1). The
water peak was clearly observed at 760–973 K and the peak
profile was shown in Fig. 3(a1). The (PtSn)1.5 catalysts showed
the total NH3 uptake of 0.526 mmol (g cat)−1, which is low as
compared with the support ␪-Al2O3 (0.664 mmol (g cat)−1). The
result indicates that the addition of Pt and Sn on ␪-Al2O3 decreased
the acidic sites of the support [18]. The NH3 TPD results of alkali
metals such as K, Ca, Na and Li doped (PtSn)1.5 were also stud-
ied, and the results were also shown in Table 2. The addition
of K to (PtSn)1.5 decreased the total NH3 uptake (0.315 mmol
(g cat)−1). The order of the total NH3 uptakes were as follows: ␪-
Al2O3 > (PtSn)1.5 > Li0.95(PtSn)1.5 > Na0.95(PtSn)1.5 > Ca0.95(PtSn)1.5
>
K0.95(PtSn)1.5. Zhang et al. and Bariås et al. reported that the
addition of K or Sn reduced the acidic sites of the support due
to its basic character [27,28]. The experimental results clearly
showed that the addition of K to (PtSn)1.5 suppressed the acidity,
minimizing isomerization products in n-butane dehydrogenation
reaction as shown in Table 7.
3.1.3. Temperature programmed reduction studies
Fig. 4 shows the TPR profiles of the K0.9(Pt)1.5, K0.95(Sn)1.5
,
(PtSn)1.5 (without K) and different wt.% of potassium doped
bimetallic (PtSn)1.5 catalysts and Table 3 shows the reduction

Table 2
NH3-TPD results of ␪-Al2O3, (PtSn)1.5 and different alkali metals supported on (PtSn)1.5/␪-Al2O3 catalysts.
Catalysts
Weak acidic sites (I)
Medium acidic sites (II)
Strong acidic sites (III)
Total NH3
uptake/mmol (g cat)−1
Temperature (K);
Percentage (%)
NH3 uptake/mmol
(g cat)−1
Temperature (K);
Percentage (%)
NH3 uptake/mmol
(g cat)−1
Temperature (K);
Percentage (%)
NH3 uptake/mmol
(g cat)−1
␪-Al2O3
(PtSn)1.5
K0.95(PtSn)1.5
Ca0.95(PtSn)1.5
Na0.95(PtSn)1.5
Li0.95(PtSn)1.5
472; (9)
464; (9)
470; (8)
475; (13)
480; (9)
474; (7)
0.060
0.047
0.025
0.058
0.042
0.034
512; (20), 575; (22)
504; (22), 560; (31)
507; (18), 564; (35)
513; (26), 561; (21)
522; (18), 587; (38)
508; (21), 557; (28)
0.133, 0.146
0.116, 0.163
0.057, 0.110
0.116, 0.093
0.085, 0.179
0.102, 0.135
629; (8), 667; (41)
652; (38)
652; (39)
613; (22), 683; (18)
692; (35)
667; (44)
0.053, 0.272
0.200
0.123
0.098, 0.080
0.165
0.213
0.664
0.526
0.315
0.445
0.471
0.484

Table 3
H2-TPR profiles of mono and bimetallic 1.5 wt.% (PtSn)/␪-Al2O3 with different potassium loadings.
Catalysts
␣ Peak
␤ Peak
␥ Peak
␦ Peak
Temperature (K);
Percentage (%)
H2 consumption/mmol
H2 (g cat)−1
Temperature (K);
Percentage (%)
H2 consumption/mmol
H2 (g cat)−1
Temperature (K);
Percentage (%)
H2 consumption/mmol
H2 (g cat)−1
Temperature (K);
Percentage (%)
H2 consumption/mmol
H2 (g cat)−1
(PtSn)1.5
K0.4(PtSn)1.5
476; (29)
520; (21)
515; (35)
503; (26)
529; (35)
508; (24)
400; (4), 478; (37)
530; (6)
0.050
0.044
0.079
0.062
0.089
0.065
0.0038, 0.035
0.008
548; (22)
579; (36)
611; (31)
569; (38)
584; (25)
559; (10)
521; (25), 633; (11)
–
0.040
0.075
0.070
0.091
0.064
0.027
0.023, 0.010
–
677; (43)
753; (28), 787; (15)
786; (23)
688 (9)
698; (15), 787; (25)
655; (54)
0.080
0.059, 0.031
0.052
0.022
0.038, 0.064
0.146
946; (6)
–
809; (11)
830; (27)
–
882; (12)
823; (19)
886; (3)
0.010
–
0.025
0.065
–
0.032
0.018
0.004
K
0.7(PtSn)1.5
K0.95(PtSn)1.5
1.2(PtSn)1.5
K
K1.45(PtSn)1.5
K0.9(Pt)1.5
K0.95(Sn)1.5
714; (4)
712; (73), 752; (18)
0.0038
0.103, 0.025

46
B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
Table 4
CO chemisorptions and amount of coke deposited on the catalyst after n-butane dehydrogenation reaction at 823 K for 5 h.
Catalysts/wt.%
Amount of CO adsorbed
in/cm3 STP (g cat)−1
Pt dispersion/%
Pt surface area/m2 (g cat)−1
Average particle
size (nm)
Coke amount (wt.%)c
COa
TEMb
(PtSn)1.5
K0.4(PtSn)1.5
0.112
0.100
0.096
0.090
0.051
0.033
0.082
7.5
6.7
6.4
6.1
3.4
2.2
5.3
0.241
0.216
0.206
0.194
0.110
0.072
0.176
15.1
16.8
17.6
18.7
33.0
50.3
21.4
7.1
8.6
9.8
10.5
16.4
23.5
15.4
0.8
0.5
0.4
0.3
1.2
1.8
2.9
K
K
K
0.7(PtSn)1.5
0.95(PtSn)1.5
1.2(PtSn)1.5
K1.45(PtSn)1.5
K0.9(Pt)1.5
a
Calculated from CO chemisorptions.
Mean particle size of the metal catalyst from TEM images using at least 30 visible particles.
From elemental analysis.
b
c
acidic sites of Pt particles can be inhibited by the interaction with
potassium.
suggested that the addition of Sn and K to Pt could lead to a direct
interaction between Pt with Sn and K. To investigate the K effect on
active Pt metal, CO chemisorption study was performed.
3.1.6. Transmission electron microscopy (TEM) studies
The Fig. 6 shows TEM images and particle size distribution
of the reduced Kx.x/(PtSn)1.5 catalysts. The TEM images exhibited
3.1.4. CO chemisorption study
Table 4 shows the Pt surface areas (MSA), Pt dispersions (D) and
particle sizes (PS) of the (PtSn)1.5 and potassium doped (PtSn)1.5
catalysts by CO chemisorption. The mono metallic K0.9(Pt)1.5 cata-
lyst exhibited a low Pt dispersion. The bimetallic (PtSn)1.5 catalyst
showed the highest Pt dispersion and Pt surface area. The Pt dis-
persion of (PtSn)1.5 catalyst slightly decreased with an increase in
amount of potassium. However, the Pt metal dispersion abruptly
decreased when potassium was added to the (PtSn)1.5 catalyst more
than 1.2 wt.%. The potassium addition below 1.0 wt.% rarely cov-
ered the exposed Pt surface, indicating the presence of the chemical
interaction with Pt, while that above 1.0 wt.% significantly covered
the Pt surface resulting in low Pt dispersion. As previously dis-
cussed, the high Pt dispersion of the (PtSn)1.5 can be ascribed due
to the prevention of Pt migration by the PtSn alloy formation at low
reduction temperature [26].
that the Pt particle sizes of the reduced K0.9(Pt)1.5, K0.4(PtSn)1.5
,
K0.7(PtSn)1.5, K0.95(PtSn)1.5, K1.2(PtSn)1.5 and K1.45(PtSn)1.5 cata-
lysts were about 15.4, 8.6, 9.8, 10.5, 16.4 and 23.5 nm, respectively.
The Pt particles size of the (PtSn)1.5 catalyst without potassium was
the smallest among the prepared catalysts, which is corresponded
with the CO chemisorption results. The Pt particle size increased
and the Pt metal dispersion decreased with an increase in potas-
sium content on the (PtSn)1.5 catalyst (Table 4). The average Pt
particle sizes by TEM analysis were approximately half of those
from the chemisorption data (Table 4). A conventional CO/Ptatm of
1.0 was used for the Pt particle size calculation. In the supported
Pt/Al2O3 catalysts, the CO/H ratio for a Pt atom was estimated as
0.83–0.96 [34]. In the well-dispersed catalysts, a CO/H ratio of 0.7
was reported for an H/Pt ratio of 1.0, which was due to the CO
bridged bonding of 50% on Pt [35]. Here, the Pt metal particle size
on the bimetallic catalysts by TEM analysis will be equivalent to
that by CO chemisorption, if the CO/Pt ratio is 0.5.
3.1.5. HAADF STEM and corresponding EDX mapping
The XRD pattern did not show any metal or metal oxides (Pt/PtOx
or Sn/SnOx) or potassium peaks in the calcined catalysts, and only
peaks of ␪-Al2O3 phase were observed. To identify the structures of
the Pt, Sn, K and PtSn particles on ␪-Al2O3, the K0.95(PtSn)1.5 sample
was characterized using a FEI microscopy in STEM mode. Fig. 5(a)
and (b) show typical high-angle annular dark-field scanning TEM
(HAADF STEM) images and the corresponding EDS maps of the cal-
cined and reduced K0.95(PtSn)1.5 catalysts, respectively. Fig. 5(a)
shows the HAADF STEM-EDS compositional maps of the calcined
Fig. 7 shows the HRTEM micrograph of the calcined and reduced
0.95(PtSn)1.5 catalyst, indicating the presence of individual parti-
K
cles with clear lattice fringes. The fast Fourier transform (FFT) was
applied to calculate the lattice spacing of the crystallite. The FFT
image (Fig. 7(a)) of the calcined K0.95(PtSn)1.5 catalyst exhibited
the lattice spacing of 0.236 nm corresponding to the (1 1 1) reflec-
tion planes of the Pt, respectively. The FFT images of the reduced
K0.95(PtSn)1.5 catalyst (Fig. 7(b)) showed a spacing of 0.218 nm, cor-
responding to the (1 0 2) planes of the PtSn alloy.
K0.95(PtSn)1.5 catalyst. The distribution of the Pt, Sn and K particles
was clearly visible. The Pt particles were scattered on the support,
while the Sn and K particles were uniformly well distributed as
shown in the EDS mapping. Fig. 5(b) also shows the HAADF STEM-
EDS compositional maps of the reduced K0.95(PtSn)1.5 catalyst. The
distributions of Pt, Sn and K particles were also visible. The Pt, Sn
and K particles on the reduced K0.95(PtSn)1.5 catalyst were observed
at the same position. The number of Pt particles of the calcined
K0.95(PtSn)1.5 sample decreased after the reduction, indicating the
Pt particles were sintered. However, the highly dispersed Sn and
K particles on the calcined K0.95(PtSn)1.5 sample were migrated to
the Pt particles during the reduction. From HAADF STEM and cor-
responding EDX mapping, it is clear that Pt, Sn and K are on the
same position after the reduction. The experimental results sug-
gest the PtSn alloy can form during the reduction step and the
major of potassium are not on the support, but on the PtSn par-
ticles. Therefore, it is concluded that the main potassium effect on
the PtSn catalysts are closely related to the interaction with Pt. The
3.2. Catalytic activity
3.2.1. Effect of potassium on 1.5 wt.% PtSn supported ꢀ-Al2O3
catalysts
Fig. 8 shows the conversion of n-butane with respect to the
reaction time on the (PtSn)1.5 catalyst, the potassium doped (Pt)1.5
(Sn)1.5, and the potassium doped (PtSn)1.5 catalysts. Table 5 shows
the initial and final (Xi and Xf, measured at 30 to 300 min. of the
reaction time) conversion, n-C4 selectivity and n-C4 yields of the
tested catalysts. The reaction was carried out at 823 K for 5 h in n-
butane dehydrogenation. In the product distribution, 1,3-butadiene
was produced by the dehydrogenation of n-butenes during n-
butane dehydrogenation. The selectivity to 1,3-butandiene was
limited by the thermodynamic equilibrium in the reaction condi-
tion. The products of C1–C3 included methane, ethane, propane and
propylene, which could be produced on the hydrocracking sites on
the Pt metal species. The cracking products were the highest on
,
=
=

B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
47
Fig. 5. The high angle annular dark-field scanning TEM (HAADF STEM) images and the corresponding EDS maps of the (a) calcined and (b) reduced 0.95 wt.% potassium
doped (PtSn)1.5 catalyst.
Table 5
Activity and deactivation parameter of 1.5 wt.% (PtSn)/␪-Al2O3 catalysts (with and without K) for n-butane dehydrogenation reaction at 823 K and a GHSV of 18,000 mL/gcat/h.
Catalysts (wt.%)
Conversion (%)
Deactivation
Selectivity (Si/Sf) (%)
Yield of n-C4=(Yf) (%)
parameter* (ꢁX)
=
Initial (Xi)
Final (Xf)
C1–C3
n-C4
i-C4
1,3-Butadiene
1-Butene
2-Butene
(PtSn)1.5
46.7
41.9
40.4
44.1
33.5
13.2
24.1
0.95
43.6
39.5
38.4
42.8
27.3
9.5
6.63
5.72
4.95
2.95
18.51
28.03
31.12
–
1.7/1.7
1.1/1.2
1.5/1.5
1.6/1.5
1.4/3.1
3.8/5.4
6.5/7.5
55.4/56.9
30.9/34.5
37.5/37.2
37.6/37.2
36.1/37.1
37.0/33.1
31.0/28.0
33.7/32.1
2.8/2.6
45.8/51.9
55.8/56.0
55.1/55.7
56.4/55.8
55.3/55.7
56.2/56.0
51.3/52.0
5.9/5.3
18.3/8.0
1.8/1.6
1.8/1.6
1.6/1.3
1.5/3.0
4.4/5.9
5.4/5.0
35.9/35.2
3.3/3.9
3.8/4.0
4.0/4.0
4.3/4.3
4.8/5.1
4.6/4.7
3.0/3.4
0.0/0.0
37.7
36.8
35.7
39.8
24.2
8.0
K0.4(PtSn)1.5
K0.7(PtSn)1.5
K0.95(PtSn)1.5
K1.2(PtSn)1.5
K1.45(PtSn)1.5
K0.9(Pt)1.5
16.6
0.95
14.0
0.07
K0.95(Sn)1.5
*: ꢁX = (Xi −Xf)/Xi × 100.

48
B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
Fig. 6. TEM images of the reduced PtSn/␪-Al2O3(with and without) catalysts. (a,a1) (PtSn)1.5, (b,b1) K0.4(PtSn)1.5, (c,c1) K0.7(PtSn)1.5, (d,d1) K0.95(PtSn)1.5, (e,e1) K1.2(PtSn)1.5
(f,f1) K1.45(PtSn)1.5 and (g,g1) K0.9(Pt)1.5
,
.
the K0.9(Pt)1.5 catalyst. The cracking sites on the Pt catalyst was
little influenced by the potassium addition, but was blocked by
the Sn addition as shown in Table 5. The selectivity to C1–C3 of
(PtSn)1.5 catalyst was not so much changed with the addition of
potassium, but rather increased with the potassium addition more
than 1.5 wt.%.
Isobutane was a main isomerization product in n-butane dehy-
drogenation. The isomerization is due to the acid sites of the
catalysts. The acid sites can be present on both Pt sites and support.
It is clear that the potassium addition suppressed the isomerization
sites of the (PtSn)1.5 catalyst as shown in Table 5. The selectivity to
the isobutane decreased with a potassium addition up to 1 wt.%,
but rather increased with the further potassium addition. If the
potassium suppressed the acid sites on the support, the selectiv-
ity to isobutene should have maintained at low level even at the
potassium addition more than 1.5 wt%. As shown in Fig. 4, most
of potassium particles are not on the support, but on Pt particles,
indicating that the potassium suppressed the acidic sites on the

B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
49
Fig. 7. HRTEM micrographs and corresponding FFTs of the (a) K0.95(PtSn)1.5-calcined and (b) K0.95(PtSn)1.5-reduced 0.95 wt.% potassium doped PtSn/␪-Al2O3 catalysts.
Pt particles. It has been known that the acid sites for isomeriza-
tion were related with chlorine presence on the Pt/alumina catalyst
[36–38]. In our reaction temperature, oxychlorinated platinum was
reduced during the reduction as shown in Fig. 4. It was shown that
the hydro-dechlorination can induce the strong acid sites on the
catalysts [37] which were titrated with potassium, resulting in low
isobutene yield.
The initial conversions of (PtSn)1.5, K0.4(PtSn)1.5, K0.7(PtSn)1.5
,
K0.95(PtSn)1.5
,
K1.2(PtSn)1.5
K1.45(PtSn)1.5
,
K0.9(Pt)1.5
and
K
0.95(Sn)1.5 catalysts were 46.7, 41.9, 40.4, 44.1, 33.5 13.2,
24.1 and 0.95%, respectively. The conversion of each catalyst
after 5 h decreased to 43.6, 39.5, 38.4, 42.8, 27.3, 9.5, 16.6
and 0.95%, respectively. The catalysts listed in order of the
n-C4 conversion for 5 h reaction at 823 K were as follows:
Fig. 8. Conversion vs time on stream on (PtSn)1.5 catalyst (with and without K) with
different amount of potassium for n-butane dehydrogenation reaction at 823 K and
a GHSV of 18,000 mL/gcat/h.
Fig. 9. Stability test on (PtSn)1.5 and K0.95(PtSn)1.5 catalyst at 823 K for 15 h for the
n-butane dehydrogenation reaction at 823 K and a GHSV of 18,000 mL/gcat/h.

50
B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
The n-C4= selectivity of the (PtSn)1.5 catalyst was 76.7/86.4% for
Si/Sf, while that of the K0.95(PtSn)1.5 catalyst was 92.5/92.9%. The
=
high n-C4 selectivity of the potassium doped (PtSn)1.5 catalysts
resulted from inhibiting the isomerization sites on the (PtSn)1.5
catalyst as discussed previously, because the selectivity to C1–C3
was not so much changed with the potassium addition. The
potassium contained PtSn alloy particles restrict the sites for
cracking, hydrogenolysis, and isomerization, resulting in high
=
n-C4 selectivity and resistance to the coke formation. Finally, the
=
catalysts listed in the order of the n-C4 yield were as follows:
K0.95(PtSn)1.5 > (PtSn)1.5 > K0.7(PtSn)1.5 > K0.4(PtSn)1.5 > K1.2(PtSn)1.5
> K0.9(Pt)1.5 > K1.45(PtSn)1.5
.
3.2.2. Effect of reaction temperature and stability test on
K0.95(PtSn)1.5
catalyst
Fig. 10. Conversion vs time on stream on K0.95(PtSn)1.5 catalyst with different
temperatures for the n-butane dehydrogenation reaction at 823 K and a GHSV of
18,000 mL/gcat/h.
=
The K0.95(PtSn)1.5 catalyst showed the highest n-C4 yield
among the prepared catalysts. Fig. 9 shows the butane conversion
of K0.95(PtSn)1.5 catalyst with respect to the time on stream in the
range of 773–873 K, and Table 6 shows the deactivation parameter,
butane conversion, n-C4 selectivity, and n-C4 yield. The deac-
tivation parameter increased with an increase in the reaction
temperature. The n-butane conversion rapidly deactivated at 873 K.
The cracking products of C1–C3 also increased with an increase in
=
=
(PtSn)1.5 > K0.95(PtSn)1.5 > K0.4(PtSn)1.5 > K0.7(PtSn)1.5 > K1.2(PtSn)1.5
> K0.9(Pt)1.5 > K1.45(PtSn)1.5. The conversion is closely related to the
Pt dispersion except K0.95(PtSn)1.5. The K0.95(PtSn)1.5 showed the
optimized composition for the potassium addition on the (PtSn)1.5
catalyst. Veldurthi et al. reports that the improvement of catalytic
activity was mainly related to the Pt dispersion, Pt surface area
and particle size of the catalysts [39]. However, the catalysts listed
in the order of the stability from the deactivation parameters were
as follows: K0.95(PtSn)1.5 > K0.7(PtSn)1.5 > K0.4(PtSn)1.5 > (PtSn)1.5
> K1.2(PtSn)1.5 > K1.45(PtSn)1.5 > K0.9(Pt)1.5 > K0.95(Sn)1.5. The deac-
tivation parameters of the prepared catalysts are shown in Table 5
(ꢁX = 100 × (Xi−Xf)/Xi, where Xi is the initial and Xf is the final
conversion, respectively). The monometallic K0.9(Pt)1.5 showed
very high deactivation parameter and the catalyst was strongly
deactivated with time on steam was expected. K0.9(Pt)1.5 catalyst
showed very low Pt dispersion (5.3%) and Pt surface area (0.176 m2
(g cat) −1), and a large Pt particles size (>15.4 nm). The coke
formation after the reaction was 2.9 wt.%, which was the highest
among the prepared catalysts. The deactivation parameters are
closely related to the coke amount after 5 h reaction. The coke
formation on the catalyst involves several processes such as
dehydrogenation condensation, polymerization, and cyclization
of hydrocarbon species on the catalyst surface. It was known that
the olefinic precursors on the metal active sites migrated to the
support, which were converted to polyolefinic or polyaromatic
types of coke through condensation and polymerization reaction
[2,3,15,17,40]. The relationship between the particle size of the
active metal and the amount of carbon deposited on the surface of
the support was reported [13,41,42].
=
the reaction temperature, resulting in the low n-C4 selectivity.
The higher the reaction temperature, the higher the 1,3-butadiene
yield in the thermodynamic point. The reaction at 873 K showed the
=
highest n-C4 yield after 5 h reaction, but the catalyst was strongly
deactivated as compared with other temperatures for n-butane
dehydrogenation reaction.
Fig. 10 shows the catalytic activity of the (PtSn)1.5 and
K0.95(PtSn)1.5 catalyst at 823 K for 15 h. The (PtSn)1.5 catalyst
was deactivated with time on stream more rapidly than the
K0.95(PtSn)1.5 catalyst. After 15 h, the conversion of n-butane on
K
0.95(PtSn)1.5 catalyst was higher than that on (PtSn)1.5. The initial
=
n-C4 selectivity of the (PtSn)1.5 catalyst were 77%, which steadily
increased up to 90% after 15 h. The amount of coke was measured
to be about 3.3 wt.% after 15 h. The acid sites of the (PtSn)1.5 cata-
lyst were blocked by the coke formation during the reaction, which
=
resulted in the increase of n-C4 selectivity. On the other hand,
=
the n-C4 selectivity of the K0.95(PtSn)1.5 catalyst (ca. 93%) was not
so much changed during the reaction for 15 h, because the potas-
sium blocked the acid sites in advance. The amount of coke on the
K0.95(PtSn)1.5 catalyst was 0.6 wt.% after 15 h reaction.
3.2.3. Effect of alkali (K, Ca, Na and Li) doped 1.5 wt.% (PtSn)
supported ꢀ-Al2O3 catalysts
Fig. 11 shows the conversion of n-butane with respect
to the reaction time on the 0.95 wt.% of alkali (K, Ca, Na
and Li) doped (PtSn)1.5 and the reaction was carried out at
823 K for 5 h catalysts for n-butane dehydrogenation. The ini-
tial conversion of K0.95(PtSn)1.5, Ca0.95(PtSn)1.5, Na0.95(PtSn)1.5 and
Li0.95(PtSn)1.5 catalyst were 44.1%, 37.6%, 8.6% and 5.0%. The
conversion of each catalyst after 5 h decreased to 42.8, 33.8,
7.1 and 4.9%, respectively. The conversion of the K0.95(PtSn)1.5
catalyst was not only the highest among the prepared cat-
alysts, but also the most stable. Both the conversion of
n-butane and the stability of the catalysts were in the order of
K0.95(PtSn)1.5 > Ca0.95(PtSn)1.5 > Na0.95(PtSn)1.5 > Li0.95(PtSn)1.5
The deactivation parameters of 0.95 wt.% of alkali (K, Ca,
Na and Li) doped (PtSn)1.5 catalysts are shown in Table
(ꢁX = 100 × (Xi−Xf)/Xi, where Xi is the initial and Xf is the final
conversion, respectively). The Na0.95(PtSn)1.5 catalyst showed very
high deactivation rate among the other catalysts. The Li0.95(PtSn)1.5
catalyst showed very low deactivation parameter, but the initial
and final conversions were very low among the other alkali doped
The coke deposition on the (PtSn)1.5 catalyst was lower than
K0.9(Pt)1.5, while that on the (PtSn)1.5 catalyst decreased with an
increase in the potassium amount up to 1 wt.%. The coke can form
on the cracking and isomerization sites on the Pt particles. The coke
deposition on Pt catalyst occurred mainly on the cracking sites on
the Pt particles and the cracking sites were inhibited by the Sn addi-
tion. The potassium could inhibit the acid sites for isomerization
on the Pt particles, resulting in decreasing the carbon amount on
the (PtSn)1.5 catalyst further as shown in Table 4. The stability of
(PtSn)1.5 catalyst also increased with an increase in the potassium
amount up to 1.0 wt.%. The high stability of (PtSn)1.5 catalyst with
low potassium content is ascribed due to the suppression of the acid
sites on the (PtSn)1.5 catalyst, while the low stability of (PtSn)1.5 cat-
alyst with high potassium content can be due to large Pt particle
sizes as shown in Table 4. Additionally, the high selectivity to C1–C3
of K1.45(PtSn)1.5 indicates that PtSn interaction was weakened by
the potassium addition of high content.
.
7

B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
51
Table 6
Activity and deactivation parameter of K0.95(PtSn)1.5 catalyst with different temperatures for n-butane dehydrogenation reaction at 823 K and a GHSV of 18,000 mL/gcat/h.
Temperature (K)
Conversion (%)
Deactivation
Selectivity (Si/Sf) (%)
Yield of n-C4=(Yf) (%)
parameter (ꢁX)
=
Initial (Xi)
Final (Xf)
C1–C3
n-C4
i-C4
1,3-Butadiene
1-Butene
2-Butene
773
823
873
19.6
44.1
62.0
18.5
42.8
52.5
5.61
2.95
15.32
0.2/0.2
1.6/1.5
5.8/6.8
40.4/39.8
36.1/37.1
32.4/31.9
55.8/56.1
56.4/55.8
51.5/51.1
1.5/1.7
1.6/1.3
1.0/0.9
2.1/2.2
4.3/4.3
9.3/9.3
17.7
39.8
43.6
Table 7
Activity (conversion, selectivity and yield) and deactivation parameter of (PtSn)1.5 catalyst with different alkali metals for n-butane dehydrogenation reaction at 823 K and a
GHSV of 18,000 mL/gcat/h.
Catalyst (wt.%)
Conversion (%)
Deactivation
Selectivity (Si/Sf) (%)
Yield of n-C4=(Yf) (%)
parameter (ꢁX)
=
Initial (Xi)
Final (Xf)
C1–C3
n-C4
i-C4
1,3-butadiene
1-Butene
2-Butene
K0.95(PtSn)1.5
Ca0.95(PtSn)1.5
Na0.95(PtSn)1.5
Li0.95(PtSn)1.5
44.1
37.6
8.6
42.8
33.8
7.1
2.95
10.10
17.44
2.00
1.6/1.5
1.5/1.8
5.4/7.1
36.1/37.1
37.1/36.6
28.5/25.7
23.9/23.2
56.4/55.8
55.7/55.7
55.4/55.1
50.5/50.5
1.6/1.3
1.5/1.7
5.7/7.0
4.3/4.3
4.2/4.2
5.0/5.1
4.9/5.0
39.8
31.2
5.7
5.0
4.9
10.0/10.7
10.7/10.6
3.6
sion. It has been known that the coke formed on the cracking and
isomerization sites of the Pt catalysts. The formation of C1–C3 can
be produced on the cracking sites, while that of i-C4 can be done
on the isomerization sites. K0.9(Pt)1.5 catalyst showed the C1–C3
selectivity of 6.5%, the i-C4 selectivity of 5.4% and the coke amount
of 2.9 wt.%. The Sn added (PtSn)1.5 catalyst showed the low C1–C3
selectivity (1.7%) and coke amount (0.8 wt%), but the high i-C4 selec-
=
tivity (18.3%), resulting in the low n-C4 selectivity (76.7%). When
potassium was added to the (PtSn)1.5 catalyst upto 0.95 wt%, both
the i-C4 selectivity and the coke amount were minimized to be
1.5%–1.6% and 0.5%–0.3 wt.%, respectively. The further potassium
addition rather increased the C1–C3 selectivity, the i-C4 selectivity
and the coke amount. From these experimental results, Sn addition
suppressed the cracking sites on the Pt catalysts, while potassium
addition blocked the isomerization sites (acid sites) on the Pt cata-
lysts. A lot of potassium addition to the (PtSn)1.5 rather weakened
the interactions among Pt, Sn and K, resulting in the increase of the
C1–C3 selectivity as well as of the i-C4 selectivity, resulting in the
highest coke amount.
Fig. 11. Conversion vs time on stream on 0.95 wt.% alkali doped (PtSn)1.5 catalyst
for n-butane dehydrogenation reaction at 823 K and a GHSV of 18,000 mL/gcat/h.
Other alkali metals such as Ca, Na, and Li were tested for blocking
the acid sites on the (PtSn)1.5 catalysts in n-butane dehydrogena-
tion. The potassium added catalyst was still the best one, because
the potassium strongly suppressed the acidic sites as compared
with other alkali metals as shown in NH3-TPD studies. From these
experimental results, the K0.95(PtSn)1.5 not only showed the high-
est n-C4= yield in n-butane dehydrogenation, but also the highest
stability.
=
catalyst. The Ca0.95(PtSn)1.5 catalyst showed better C4 yield than
Na and Li doped catalyst, but this catalyst had low activity and high
deactivation rate than K0.95(PtSn)1.5 catalyst. The K0.95(PtSn)1.5 cat-
alyst showed the best catalytic activity among the other alkali
doped catalysts.
4. Conclusions
PtSn/␪-Al2O3 catalysts with different amount of potassium (0.4,
0.7, 0.95, 1.2 and 1.45 wt.%) were prepared by an impregnation
method, and their catalytic activity in n-butane dehydrogenation
was investigated at 823 K, an atmospheric pressure and a GHSV
of 18,000 mL(gcat h)−1 XRD patterns and TEM images showed the
PtSn alloy formation after the reduction of the (PtSn)1.5 based cat-
alysts. It was observed by HAADF STEM and corresponding EDX
mapping that the components of Sn and K were observed on the
same position with Pt particles on the ␪-Al2O3 support after the
H2 reduction, indicating that both Sn and K modified the charac-
ters of Pt particles. The particle sizes from the Pt dispersion by CO
chemisorption were consistent with those by TEM analysis. It can
be concluded from the activity test of the prepared catalysts that
the conversion of n-butane was mostly related to the Pt disper-
Acknowledgements
This work was supported by the basic research project of KIST
and by Converging Research Center Program through the National
Research Foundation of Korea (NRF) funded by the Ministry of Edu-
cation, Science and Technology (2011K000660).
References
[1] S.D. Jackson, S. Rugmini, J. Catal. 251 (2007) 59–68.
[2] S.T. Korhonen, S.M.K. Airaksinen, M.A. Ban˜ares, A.O.I. Krause, Appl. Catal., A 333
(2007) 30–41.
[3] S.M. Stagg, C.A. Querini, W.E. Alvarez, D.E. Resasco, J. Catal. 168 (1997) 75–94.
[4] S.J. Miller, U.S. Patent 4 727 216, 1986.

52
B.M. Nagaraja et al. / Catalysis Today 232 (2014) 40–52
[5] Y. Zhang, Y. Zhou, X. Sheng, L. Wan, Y. Li, Y. Xiao, B. Yu, Z. Zeng, Fuel Process.
[22] A.D. Ballarini, P. Zgolicz, I.M.J. Vilella, S.R. de Miguel, A.A. Castro, O.A. Scelza,
Appl. Catal., A 381 (2010) 83–91.
Technol. 104 (2012) 23–30.
[6] R.D. Cortright, J.A. Dumesic, J. Catal. 148 (1994) 771–778.
[7] G.J. Siri, J.M. Ramallo-López, M.L. Casella, J.L.G. Fierro, F.G. Requejo, O.A. Ferretti,
Appl. Catal., A 278 (2005) 239–249.
[8] M. Ohta, Y. Ikeda, A. Igarashi, Appl. Catal., A 266 (2004) 229–233.
[9] J. Llorca, N. Homs, J. León, J. Sales, J.L.G. Fierro, P.R. Piscina, Appl. Catal., A 189
(1999) 77–86.
[23] A.D. Ballarini, C.G. Ricci, S.R. de Miguel, O.A. Scelza, Catal. Today 133–135 (2008)
28–34.
[24] G.D. Angel, A. Bonilla, Y. Pena, J. Navarrete, J.L.G. Fierro, D.R. Acosta, J. Catal. 219
(2003) 63–73.
[25] Y.W. Zhang, Y.M. Zhou, H. Liu, Y. Wang, Y. Xu, P.C. Wu, Appl. Catal., A 333 (2007)
202–210.
[10] R.D. Cortright, J.A. Dumesic, J. Catal. 157 (1995) 576–583.
[11] G.J. Siri, G.R. Bertolini, M.L. Casella, O.A. Ferretti, Mater. Lett. 59 (2005)
2319–2324.
[12] Y. Zhang, Y. Zhou, L. Wan, M. Xue, Y.Z. Duan, X. Liu, J. Natur. Gas Chem. 20 (2011)
639–646.
[13] J.M. Hill, R.D. Cortright, J.A. Dumesic, Appl. Catal., A 168 (1998) 9–21.
[14] M. Tasbihi, F. Feyzi, M.A. Amlashi, A.Z. Abdullah, A.R. Mohamed, Fuel Process.
Technol. 88 (2007) 883–889.
[26] B.M. Nagaraja, C.-H. Shin, K.-D. Jung, Appl. Catal., A 467 (2013) 211–223.
[27] S. Zhang, Y. Zhou, Y. Zhang, L. Huang, Catal. Lett. 135 (2010) 76–82.
[28] O.A. Bariås, A. Holmen, E.A. Blekkan, Catal. Today 24 (1995) 361–364.
[29] F.M. Dautzenberg, H.B.M. Wolters, J. Catal. 51 (1978) 26–39.
[30] G. Lietz, H. Lieske, H. Spindler, W. Hanke, J. Völter, J. Catal. 81 (1983) 17–25.
[31] D.L. Hoang, S.A.-F. Farrage, J. Radnik, M.-M. Pohl, M. Schneider, H. Lieske, A.
Martin, Appl. Catal., A 333 (2007) 67–77.
[32] R. Burch, J. Catal. 71 (1981) 348–359.
[15] D. Rodriguez, J. Sanchez, G. Arteaga, J. Mol. Catal. A: Chem. 228 (2005)
309–317.
[33] Y. Perak, F. Ribeiro, G. Somorjai, J. Catal. 178 (1998) 66–75.
[34] J. Freel, J. Catal. 25 (1972) 149–160.
[16] M.L. Casella, G.J. Siri, G.F. Santori, O.A. Ferretti, M. Ramírez de Agudelo, Langmuir
13 (2000) 5639–5643.
[17] Y. Zhang, Y. Zhou, L. Wan, M. Xue, Y.Z. Duan, X. Liu, Fuel Process. Technol. 92
(2011) 1632–1638.
[18] L.H. Wan, Y.M. Zhou, Y.W. Zhang, Y.Z. Duan, X. Liu, M.W. Xue, Ind. Eng. Chem.
Res. 50 (2011) 4280–4285.
[35] H.L. Gruber, J. Phys. Chem. 66 (1962) 48–54.
[36] E. Garbowski, J.-P. Candy, M. Primet, J. Chem. Soc. Faraday Trans. I 79 (1983)
835–844.
[37] P.-M. Bernard, M. Primet, J. Chem. Faraday Trans. 86 (3) (1990) 567–570.
[38] P.G. Menon, R.P. De Pauw, G.F. Froment, Ind. Eng. Chem. Prod. Res. Dev. 18 (2)
(1979) 110–116.
[19] S.A. Bocanegra, A.G. Ruiz, S.R. de Miguel, O.A. Scelza, Appl. Catal., A 277 (2004)
11–22.
[39] S. Veldurthi, C.H. Shin, O.S. Joo, K.D. Jung, Catal. Today 185 (2012)
88–93.
[20] S.A. Bocanegra, A.A. Castro, O.A. Scelza, S.R. de Miguel, Appl. Catal., A 333 (2007)
49–56.
[40] E. Merlen, P. Beccat, J.C. Bertolini, P. Delichère, N. Zanier, B. Didillon, J. Catal.
159 (1996) 178–188.
[21] S.A. Bocanegra, S.R. Miguel, I. Borbath, J.L. Margitfalvi, O.A. Scelza, J. Mol. Catal.
A: Chem. 301 (2009) 52–60.
[41] F.M. Brinkmeyer, D.F. Rohr, U.S. Patent 4 866 211, 1987.
[42] T. Imai, C.W. Hung, U.S. Patent 4 430 517, 1983.