Dehydrogenation of propane over Pt/KL catalyst: Investigating the role of L-zeolite structure on catalyst performance using isotope labeling

Article abstract of DOI:10.1016/j.apcata.2010.10.022

Dehydrogenation of propane using an equimolar mixture of propane-d ₀ and propane-d₈ was investigated over 1%Pt/KL and 1%Pt/SiO₂ at atmospheric pressure and different reaction temperatures. A normal kinetic isotope effect exists (k_H/k _D = 1.4-1.5) when the reaction is conducted on Pt/KL at different temperatures (400, 500, and 600 °C), suggesting that C-H bond activation is involved in the kinetically relevant steps. Furtheremore, there is hardly any H-D exchange in the recovered propane during dehydrogenation at temperatures above 400 °C, suggesting that adsorption and subsequent dehydrogenation of propane are essentially irreversible and that C-H bond activation is the rate determining step. Unlike the case of hexane aromatization, the unique structure of L-zeolite does not help in controlling the entry of propane molecules into the lobes of the L-zeolite containing the active sites; hence, bimolecular reactions do occur, leading to coke formation and catalyst deactivation.

Full text of DOI:10.1016/j.apcata.2010.10.022

Applied Catalysis A: General 390 (2010) 264–270
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Dehydrogenation of propane over Pt/KL catalyst: Investigating the role of
L-zeolite structure on catalyst performance using isotope labeling
Khalid G. Azzam, Gary Jacobs, Wilson D. Shafer, Burtron H. Davis∗
Center for Applied Energy Research, University of Kentucky, 2540 Research Park Drive, Lexington, KY 40511, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 11 June 2010
Received in revised form 13 October 2010
Accepted 20 October 2010
Available online 27 October 2010
Dehydrogenation of propane using an equimolar mixture of propane-d0 and propane-d8 was investigated
over 1%Pt/KL and 1%Pt/SiO2 at atmospheric pressure and different reaction temperatures. A normal kinetic
isotope effect exists (kH/kD = 1.4–1.5) when the reaction is conducted on Pt/KL at different temperatures
(400, 500, and 600 ◦C), suggesting that C–H bond activation is involved in the kinetically relevant steps.
Furtheremore, there is hardly any H–D exchange in the recovered propane during dehydrogenation at
temperatures above 400 ◦C, suggesting that adsorption and subsequent dehydrogenation of propane are
essentially irreversible and that C–H bond activation is the rate determining step. Unlike the case of
hexane aromatization, the unique structure of L-zeolite does not help in controlling the entry of propane
molecules into the lobes of the L-zeolite containing the active sites; hence, bimolecular reactions do
occur, leading to coke formation and catalyst deactivation.
Keywords:
Dehydrogenation
Propane
C3H8
C3D8
© 2010 Elsevier B.V. All rights reserved.
Pt/KL
L-zeolite
Isotope effect
Diffusion
H–D exchange
1. Introduction
to C–H in alkanes, isotopic switching can be used to determine if
a normal kinetic isotope effect (NKIE) exists which would suggest
Dehydrogenation of propane to propylene is an important reac-
tion, since propylene is a vital feedstock in the petrochemical
industry. The majority of propylene is produced as a by-product
from processes such as steam cracking of naphtha and fluid cat-
alytic cracking (FCC) [1]. However, due to increased global demand,
several new plants for the dehydrogenation of propane to propy-
lene have been built worldwide [1]. Dehydrogenation of propane
is a highly endothermic reaction favored by high temperatures and
low hydrogen partial pressures. Typically, Pt–Sn supported on alu-
mina or Cr2O3 supported on alumina or zirconia are the commercial
catalysts for this reaction [1,2]. A variety of Pt catalysts, often pro-
moted with other electropositive metals and supported on various
carriers (e.g., SiO2 [3], ZSM5 [4,5], L-zeolite [6], SBA-15 [2,5,7], Y-
zeolite [6], MOR [6], Beta [6,8], and Mg(Al)O [9]), were investigated
for the dehydrogenation of propane to propylene reaction in an
effort to find an optimum catalyst.
C–H bond breaking as the rate limiting step among the elemen-
tary steps. Furthermore, the degree of H–D scrambling in the
recovered-unreacted alkanes [10–13] provides important informa-
tion regarding the extent of reaction of the alkane reactant with
the surface hydrogen pool. Jackson et al. [14] studied the dehydro-
genation of propane using a C3H8/C3D8 mixture on chromia and
found an NKIE of 1.85; thus, C–H bond activation was identified
to be the rate determining step. Biscardi and Iglesia [15] investi-
gated the elementary reaction steps of propane dehydrogenation
over H-ZSM5 and Ga/H-ZSM5 catalysts using an equimolar mix-
ture of C3H8/C3D8. At a temperature of 500 ◦C, an NKIE of 1.03
and 1.2 for H-ZSM5 and Ga/H-ZSM5, respectively, were reported
and significant H–D cross-exchange between C3H8 and C3D8 in the
recovered propane was observed. Therefore, they [15] concluded
that the C–H bond activation step does not limit the overall dehy-
drogenation rate and H–D exchange is strongly influenced by the
Bronsted acid sites on those catalysts. Recently, we [16] have found
using isotope labeling that the unique structure of L-zeolite plays
a significant role in controlling the entry of hexane molecules into
the wider lobes of L-zeolite containing the Pt active sites during
the dehydrocyclization reaction; in that case, the catalyst was pre-
pared by chemical vapor deposition (CVD) of Pt(AcAc)2 to ensure
that ultra-small Pt clusters were formed, situated inside the uni-
Over Pt, alkane dehydrogenation requires elementary steps
including activation of the C–H bond and elimination of the hydro-
gen formed. Due to the lower reactivity of the C–D bond relative
∗
Corresponding author. Tel.: +1 859 257 0253; fax: +1 859 257 0302.
E-mail address: davis@caer.uky.edu (B.H. Davis).
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.10.022

K.G. Azzam et al. / Applied Catalysis A: General 390 (2010) 264–270
265
resolution of 4 to give a data spacing of 1.928 cm−1. The cata-
lyst (∼40 mg) was first activated at 450, 500 or 600 ◦C in 75 sccm
33%H2/He for 1 h, purged with He for 15 min, and cooled to 40 ◦C
in 50 sccm of He flow. To probe the Pt0 clusters, the catalyst was
subjected to 50 sccm of 2%CO/He at 40 ◦C for 30 min. The catalyst
was then purged in 50 sccm He for 20 min. Scans were recorded
during CO adsorption and following the He purge. A 1%Pt/SiO2 cat-
alyst, prepared by incipient wetness impregnation, served as the
reference sample.
Temperature programmed oxidation (TPO) of the used catalysts
was conducted using a stainless steel fixed-bed tubular reactor
(0.38 in, i.d.) system in which the outlet stream was directed to
a sampling valve on the gas chromatograph (SRI 8610C). The GC
includes two columns (6ꢀ silica gel packed and 3ꢀ molecular sieve
packed) and two detectors (FID and TCD). To boost the sensitivity of
the CO and CO2 signals, the GC incorporates a methanizer, such that
these products can be analyzed by FID. The catalysts were heated in
a 1% O2/He stream from room temperature to 750 ◦C (heating rate
1 ◦C/min−1). The reactor effluent was analyzed every 7 min.
axial channels and with a majority nestled within the wider lobes.
Because of this unique property, the catalyst exhibited unparalleled
stability for hexane aromatization, as bimolecular reactions leading
to coking were avoided [15,16]. Since Pt supported on L-zeolite is
reported to be a promising catalyst for the dehydrogenation of alka-
nes to alkenes [6,17,18], it is important to investigate if the unique
structure of L-zeolite can play a role in those Pt-catalyzed reac-
tions as well. Using propane as the reactant, the aim of this study
is to determine if an NKIE exists and to investigate the extent to
which intracrystalline diffusion plays a role in the dehydrogenation
of lighter alkanes; this is accomplished by carrying out experiments
using an equimolar ratio of C3H8 and C3D8 in the feed.
2. Experimental
2.1. Catalyst preparation
A sample of potassium exchanged zeolite LTL (Süd-Chemie) was
calcined in flowing air (100 cm3/min) at 400 ◦C for 4 h. After cool-
ing to ambient temperature, the calcined zeolite LTL powder was
transferred to a glove bag filled with inert gas and carefully mixed
with platinum acetylacetonate (Alfa Aesar); the material was then
loaded into a sublimation tube, which was sealed at one end. The
other end was attached to a high vacuum system, equipped with an
oil diffusion pump and capable of achieving a vacuum of <10−6 Torr.
The sublimation tube was held at room temperature for 4 h, and
then the temperature was slowly ramped (1 ◦C/min) to 70, 80, 90,
and then 100 ◦C, holding at each temperature for 1 h. Finally, the
temperature was ramped to 130 ◦C, held for 15 min, and cooled to
room temperature. Following the CVD procedure, the sample was
light yellow in color. To decompose the Pt(AcAc)2 compound, the
catalyst was calcined at 350 ◦C in flowing air for 4 h. The final cata-
lyst was light gray in color. The loading was 1%Pt by weight. Various
sublimation methods for obtaining highly dispersed Pt/KL catalysts
can be found in the literature [16–22].
The Pt/SiO2 catalyst was prepared using the incipient wetness
impregnation (IWI) method. Silica (The PQ Corporation, CS-2133)
was used as the catalyst support. In this method, IWI of the sup-
port was performed with the Pt salt, tetraamineplatinum (II) nitrate
(Alfa, Stock #88960, Lot #A29T027). The loading solution volume
was determined based on the average pore volume of silica used.
After impregnation, the sample was dried for 6 h at 110 ◦C and sub-
sequently calcined in stagnant air at 350 ◦C (ramp of 3 ◦C/min) for
4 h.
2.3. Activity test
Catalytic tests for propane dehydrogenation were carried out
using a stainless steel fixed-bed tubular reactor (0.38 in, i.d.) under
steady state conditions. Typically, the catalysts were pressed into
pellets, crushed, and sieved to yield grains 355–600 ␮m in diam-
eter. A 400 mg catalyst sample was packed between two layers of
quartz wool. The temperature of the catalyst bed was monitored
using an internal thermocouple (Fe–Cr) and maintained by a tem-
perature controller (ꢀ Omega CN 3251-R). Propane-d0 (Scott Gross,
33%/He balance) and propane-d8 (Cambridge Isotope Laboratories,
98% purity) were controlled by Brooks mass flow controllers to yield
a mixture of 25% propane-d0/25% propane-d8/balance He. Prior to
catalytic testing, the catalyst was reduced in H2 at the reaction tem-
perature (i.e., 500 or 600 ◦C for 15 min with a temperature ramp of
4 ◦C/min). For the experiments conducted at 300 and 400 ◦C, the
catalyst was reduced in H2 at 450 ◦C to ensure complete reduction
of platinum to Pt0. Reaction testing was conducted at a WHSV of
2.6 h−1 at atmospheric pressure and temperatures of 300, 400, 500,
and 600 ◦C. A purge-valve was used to automatically direct sam-
ples to the GC for analysis. The feed and products were analyzed
using an Agilent GC (7890A) which includes an HP-PLOT/Al2O3 ‘S’
deactivated capillary column to achieve product separation. Prod-
ucts were detected using a flame ionization detector (FID). A linear
40 min temperature ramp from 35 to 180 ◦C provided the means for
adequate peak separation in the GC column. Due to the good sep-
aration of propane-d0 and propane-d8 in the column (see Fig. 1),
the conversions of both reactants were calculated using the FID
data. The products were collected in sample bags and analyzed
using an Agilent GC (6890)–MS (5973N) to investigate the extent
of H-labeled hydrogen incorporation into the deuterated propane
reactant, and likewise, the extent of D incorporation into the H-
labeled propane reactant.
2.2. Catalyst characterization
Hydrogen chemisorption by a temperature programmed des-
orption (TPD) method was carried out using a Zeton-Altamira
AMI-200 unit. The catalyst was slowly heated (1 ◦C/min) to either
450, 500 or 600 ◦C in 30 cm3/min of a 33%H2/Ar mixture, and held
at 450, 500 or 600 ◦C for 1 h. The catalyst was then cooled in flowing
hydrogen to 80 ◦C, and purged with 25 cm3/min of flowing argon
to remove weakly bound hydrogen. The catalyst was then heated
in 25 cm3/min of flowing argon to 450, 500 or 600 ◦C to desorb the
hydrogen, and the evolved hydrogen was monitored using a ther-
mal conductivity detector (TCD). To calibrate the signal, 10 pulses
of standard amounts of hydrogen were passed into a flowing argon
stream.
3. Results and discussion
The amounts of hydrogen evolved from the 1%Pt/KL CVD cat-
alysts during sequential TPD were 51.5, 50.0, and 45.7 ␮mol/g,
respectively, for the reduction temperatures of 450, 500, and 600 ◦C
employed. These values correspond closely to a H/Pt ratio of ∼2.0
(Table 1). Virtually identical results, obtained using the volumet-
ric chemisorption approach, were reported previously [19,21], and
indicate that the Pt clusters in the catalyst used in this work are
extremely well-dispersed. Although H/Pt ratios above 1.0 are com-
mon in the literature [19,23,24] for Pt/KL catalysts, the CVD method
Diffuse reflectance infrared Fourier transform spectroscopy
(DRIFTS) spectra were recorded using
a Nicolet Nexus 870
spectrometer equipped with a DTGS-TEC detector. A Thermo
Spectra-Tech cell capable of high pressure/high temperature oper-
ation and fitted with ZnSe windows served as the reaction chamber
for in situ CO adsorption measurements. Scans were taken at a

266
K.G. Azzam et al. / Applied Catalysis A: General 390 (2010) 264–270
ages occur, a large fraction of Pt clusters become entrapped and
the catalyst deactivates. However, the greater spatial separations
of tiny Pt clusters within the pores of CVD catalysts prevent this
from happening, and an initial drop is not observed [27]. Another
testimony to this is that, even after high temperature treatment at
600 ◦C followed by H2-TPD, though a slight decrease is observed
compared to the TPD experiments carried out after 450 and 500 ◦C
treatments, a very high dispersion is retained. Recall that the lobes
are separated by narrow apertures such that if the Pt is well sepa-
rated along the lengths of the channels (i.e., by employing the CVD
method), then this geometry of the channel structure provides a
physical activation energy barrier that inhibits particle migration
and agglomeration.
One method to verify the likelihood that the Pt0 clusters are
mainly within the zeolite channels is to use CO as a probe and follow
the evolution of platinum carbonyl species by DRIFTS. These plat-
inum carbonyl species are formed by the disruption of only highly
dispersed Pt clusters by CO; they are stabilized by a ligand effect,
proposed to be due to the interaction between electron-rich zeo-
lite O atoms, located inside the zeolite channels, and the Ptx(CO)y
species [28,29]. Fig. 2A shows the Pt-carbonyl species evolved
over the 1%Pt/KL CVD catalyst (solid line) and CO adsorption on a
1%Pt/SiO2 reference catalyst (dashed line). Prior to recording the
spectra, both catalysts were reduced in H2 at 600 ◦C, in contact
with CO for 30 min at 40 ◦C, and finally purged with He for 20 min.
As can be seen, the majority of bands for 1%Pt/KL CVD are posi-
tioned at wave numbers much lower than the linear ꢁ(CO) band
observed for the typical 1%Pt/SiO2 catalyst. The low wave number
bands appear at discrete positions (bands at 2051, 2035, 2011, 1982,
and 1962 cm−1), suggesting that they likely correspond to Ptx(CO)y
species stabilized by the L-zeolite support. Although the DRIFTS
results of adsorbed CO indicate that a high fraction of Pt resides
inside the channels (Pt carbonyl bands ranging from 2051 cm−1
to 1962 cm−1), there are also two shoulders present at 2084 and
2068 cm−1 for CO adsorbed on larger Pt metal crystallites. These
particles are likely large enough to block the narrow window of the
L-zeolite and therefore completely cut off reactant access to a frac-
tion of the channels. The DRIFTS spectra for the catalysts reduced in
H2 at 600, 500, and 450 ◦C were fitted using a least squares fitting
routine with Gaussian peaks (Fig. 2B–D, respectively). Two Gaus-
sian peaks were used to define the ꢁ(CO) of linearly bound CO on
larger Pt crystallites (2068 cm−1 +) and five Gaussian peaks were
used to define the ꢁ(CO) bands of Ptx(CO)y species stabilized by the
KL zeolite (<2068 cm−1), as depicted in Fig. 2B–D. For the cases of
450 and 500 ◦C reduction, approximately 20% of the linear stretch-
ing bands correspond to those assigned to CO adsorbed on larger
Pt crystallites; the fraction is somewhat lower (∼16%) for the case
of 600 ◦C reduction due to the agglomeration of some of the larger
particles—the bands at lower wave numbers being largely retained.
This is consistent with a decrease in the BET surface area and pore
volume by a fraction of ∼1/3 (see Table 1) observed after Pt addition
and calcination relative to KL zeolite alone. Nevertheless, DRIFTS
results demonstrate that the majority of Pt is small enough, and
in sufficent contact with the KL zeolite channels, to be converted
to KL support-stabilized Pt carbonyl species upon CO adsorption
(80–85% of the linear ꢁ(CO) bands). Therefore, from the hydrogen
chemisorption and DRIFTS results, one can conclude that the major-
ity of Pt clusters are likely ultra-well-dispersed and located within
the channels of the L-zeolite.
Fig. 1. FID gas chromatogram showing the efficiency of separating the propane-d0
and propane-d8 mixture using the HP-PLOT/Al2O3 ‘S’ deactivated capillary column,
(A) before reaction and (B) after reaction.
consistently yields among the highest dispersions. In the case of
Pt/SiO2, 27.1 ␮mol/g was evolved during the TPD measurments (see
Table 1) corrosponding to a H/Pt ratio of unity. The catalyst has high
Pt dispersion with an average Pt particle size of about 1 nm. Despite
the high dispersion of Pt particles in the case of Pt/SiO2, they are
still larger in size compared with those of the Pt/KL catalyst. There-
fore, in the case of Pt/KL, not only are the particles considerably
smaller than those reported for other methods (e.g., impregnation),
but they generally exhibit greater spatial separations within the
zeolite pores [19]. Indeed, conventional catalysts (e.g., prepared by
incipient wetness impregnation or ion exchange, where particles
are often clustered near the pore mouth [25]) typically exhibit an
initial drop in conversion that is not observed with CVD catalysts.
In a rigorous HR-TEM investigation, Treacy [26] found that after
particle growth reaches a point where two or more channel block-
Table 1
Physical and chemical characteristics of the catalysts.
Catalyst
SBET (m2/g)
Pore volume
(m3/g)
H/Pta
–
KLb
291
190
350
0.143
0.100
0.788
–
–
1 wt% Pt/KLc
1 wt% Pt/SiO2
2.01d 1.95e 1.78f
The relative conversions of propane-d0 and propane-d8 and the
NKIE (kH/kD) versus time on-stream (TOS) during the dehydrogena-
tion reaction of propane over the 1%Pt/KL at 600 ◦C, are shown in
Table 2. It is obvious that the conversion of undeuterated propane
is higher than that of deuterated propane over the testing inter-
val. The NKIE (kH/kD) is greater than unity, with an average value of
about 1.5 for all measured conversions over the testing interval. It is
c
–
–
1.06f
a
Measured by H2 – TPD.
Calcined at 400 ◦C.
b
c
d
e
f
Calcined at 350 ◦C.
Reduced in H2 at 450 ◦C.
Reduced in H2 at 500 ◦C.
Reduced in H2 at 600 ◦C.

K.G. Azzam et al. / Applied Catalysis A: General 390 (2010) 264–270
267
Fig. 2. (A) DRIFTS spectra for CO adsorption over 1%Pt/KL (solid line) and 1%Pt/SiO2 (dashed line) at 40 ◦C. The catalysts were reduced at 600 ◦C in 75 ccm of 33%H2/He, cooled
to 40 ◦C in 50 ccm He, treated with 50 ccm of 2%CO/He for 30 min, and finally purged with He for 20 min. Note the height of the ꢁ(CO) band for Pt/SiO2 was divided by 3.8 to
normalize. Gaussian fitting of ꢁ(CO) for linearly bound CO (dashed, two components) and ꢁ(CO) in Pt-carbonyls (dash-dotted, five components) for (B) 600 ◦C, (C) 500 ◦C and
(D) 450 ◦C after similar treatments to (A).
well established in the literature [11,30,31] that a reaction exhibits
a significant decrease in conversion when switching from a C–H
label to the C–D label if the reaction involves C–H bond breaking in
the rate limiting step due to the NKIE arising from the difference in
bond energy. Therefore, the NKIE of 1.5 suggests the involvement
of C–H bond activation in kinetically relevant steps. The degree of
H–D exchange in the recovered propane when the reaction was
conducted at 600 ◦C is shown in Fig. 3. There is hardly any H–D
exchange between propane-d0 and propane-d8 or between the
gaseous hydrocarbons and H2/D2 at the testing conditions sug-
gesting that C–H bond activation is irreversible and hence all the
propane chemically adsorbed on the Pt active sites completed the
reaction to final products. When conducting the reaction at 500 ◦C
(Table 3 and Fig. 4) or 400 ◦C (Table 4 and Fig. 5), the NKIE also
exists with an average value of about 1.4 for both temperatures
and with no significant H–D exchange in the recovered propane
reactants. This indicates that C–H bond activation is the rate deter-
mining step for the dehydrogenation of propane over Pt/KL catalyst
in this range of temperatures (400–600 ◦C).
Similarly, the NKIE exists when the propane dehydrogenation
reaction, using an equimolar propane-d0 and propane-d8 mix-
ture, was carried out on Pt/SiO2 (Table 5) and no significant H–D
Table 2
Activity and selectivity results of dehydrogenation of propane over 1 wt% Pt/KL at 600 ◦C, 1 atm, and 2.6 h−1 WHSV using an equimolar propane-d0 and propane-d8
(25%C3H8/25%C3D8/balance He) mixture.
TOS (min)
XC
(C%)
XC
(C%)
D
8
kH/kD
SPropene (%)
SMethane (%)
SEthene (%)
SEthane (%)
SBenzene (%)
H
3
8
3
45
80
115
150
185
220
28.2
24.3
21.6
20.2
16.6
15.9
18.9
16.0
14.2
13.3
10.9
11.2
1.49
1.52
1.52
1.52
1.53
1.42
41
39
37
38
38
38
49
52
54
54
52
50
6
5
5
5
6
7
4
4
4
3
4
5
0.4
0.3
0.3
0.3
0.2
0.2

268
K.G. Azzam et al. / Applied Catalysis A: General 390 (2010) 264–270
Fig. 4. Deuterium distribution in the recovered propane (equimolar feed of 25%
propane-d0/25% propane-d8/balance He) during the dehydrogenation reaction over
Fig. 3. Deuterium distribution in the recovered propane (equimolar feed of 25%
propane-d0/25% propane-d8/balance He) during the dehydrogenation reaction over
1%Pt/KL at 500 ◦C, 1 atm, and WHSV of 2.6 h−1
.
1%Pt/KL at 600 ◦C, 1 atm, and WHSV of 2.6 h−1
.
exchange in the recovered propane was observed (Fig. 6). The
NKIE (kH/kD) is almost constant with an average value of ∼2 and
the catalyst deactivates with TOS. Comparing the performance of
Pt/SiO2 and Pt/KL catalysts during the dehydrogenation of equimo-
lar propane-d0 and propane-d8 at 600 ◦C, both catalysts displayed
an NKIE and severe catalyst deactivation with TOS, suggesting that
the unique structure of KL-zeolite hardly influences catalyst per-
formance during the propane dehydrogenation reaction.
may explain the high deactivation rate as a function of TOS. Indeed,
significant deactivation of the catalysts with TOS was observed at
500 and 600 ◦C. For instance, at 600 ◦C, the XC
and XC
dropped
H
D
3
8
3
from 28.2% to 18.9% at a TOS of 45 min to 15.9% and 11.28%, respec-
tively, at a TOS of 220 min. To investigate the role of bimolecular
reactions on catalyst deactivation, the analysis of carbon deposits
by TPO was carried out for the used catalysts after 220 min of TOS
under propane dehydrogenation at 600, 500, and 400 ◦C (Fig. 7). An
intense CO2 peak was observed from the TPO result of the spent
catalyst for the reaction at 600 ◦C compared with ones conducted
at 500 or 400 ◦C. The large amount of carbon deposited on the cat-
alyst when the reaction was conducted at 600 ◦C relative to the
500 and 400 ◦C conditions leads to severe catalyst deactivation and
may explain the lower activity of the catalyst at 600 ◦C. It seems that
within the first 45 min in which the first sample was analyzed, some
of the L-zeolite channels were rapidly blocked by carbon deposits
and hence a fraction of the Pt active sites were lost during reac-
tion testing at 600 ◦C. Furthermore, the carbon deposits formed by
carrying out the reaction at 600 ◦C seem to be different (i.e., more
reactive) compared to the carbon formed when the reaction was
The existence of NKIE at 400, 500, and 600 ◦C and the lack of
H–D exchange in the recovered propane reactants at these temper-
atures suggest that (i) C–H bond activation is the rate determining
step and (ii) entry of propane through the narrow windows of the
L-zeolite to access the active Pt0 sites (i.e., located on the Pt0 crys-
tallites residing within the wider lobes of the uni-axial channels of
the L-zeolite) is relatively fast with respect to the surface reaction
rate. This is in contrast with what we observed in the case of hex-
ane dehydrocyclization over the same catalyst [16]. Since the entry
rate of propane molecules through the narrow windows to access
the active Pt0 sites is faster than propane activation on the same
sites, then the probability for bimolecular reactions increases; this
Table 3
Activity and selectivity results for the dehydrogenation of propane over 1 wt% Pt/KL at 500 ◦C, 1 atm, and 2.6 h−1 WHSV using an equimolar propane-d0 and propane-d8
(25%C3H8/25%C3D8/balance He) mixture.
TOS (min)
XC
(C%)
XC
(C%)
D
8
kH/kD
SPropene (%)
SMethane (%)
SEthene (%)
SEthane (%)
SBenzene (%)
H
3
8
3
45
80
115
150
185
220
31.5
31.0
27.6
26.2
24.5
22.0
23.7
22.7
19.2
18.1
17.5
16.5
1.33
1.37
1.44
1.44
1.40
1.33
40
41
43
44
45
45
48
48
47
46
45
45
3
3
3
3
3
4
9
7
7
6
6
6
0.3
0.3
0.3
0.3
0.3
0.3
Table 4
Activity and selectivity results for the dehydrogenation of propane over 1 wt% Pt/KL at 400 ◦C, 1 atm, and 2.6 h−1 WHSV using an equimolar propane-d0 and propane-d8
(25%C3H8/25%C3D8/balance He) mixture.
TOS (min)
XC
(C%)
XC
(C%)
D
8
kH/kD
SPropene (%)
SMethane (%)
SEthene (%)
SEthane (%)
SBenzene (%)
H
3
8
3
45
80
115
150
185
220
6.0
6.5
7.0
6.6
6.7
6.7
4.4
4.5
5.1
4.7
4.7
4.8
1.36
1.43
1.36
1.41
1.42
1.40
52
55
57
59
60
61
37
35
33
32
31
31
1
2
2
2
2
2
9
8
8
7
7
7
0.3
0.0
0.0
0.0
0.0
0.0

K.G. Azzam et al. / Applied Catalysis A: General 390 (2010) 264–270
269
Fig. 7. Temperature programmed oxidation profiles (i.e., CO2 – GC-FID areas) as a
function of temperature for the used 1%Pt/KL catalysts using 1%O2/He at 50 ml/min.
The used catalysts were tested for the dehydrogenation of propane at 400, 500, or
600 ◦C prior to carrying out TPO.
Fig. 5. Deuterium distribution in the recovered propane (equimolar feed of 25%
propane-d0/25% propane-d8/balance He) during the dehydrogenation reaction over
1%Pt/KL at 400 ◦C, 1 atm, and WHSV of 2.6 h−1
.
There have been attempts to address the coking issue. The group
of Dumesic [32,33] studied the dehydrogenation of isobutane reac-
tion over Pt, PtSn and PtSnK catalysts supported on either silica or
K-L-zeolite. The authors found that the PtSn/KL catalysts had signif-
icantly higher rates of isobutene production relative to PtSnK/silica.
The authors suggested that Sn and K addition to Pt decreased cok-
ing by decreasing the ensemble size of Pt surface atoms, which
in turn was suggested to inhibit the formation of coke precursors,
the highly dehydrogenated surface species. Another possibility was
that Sn facilitated transfer of coke to the support and away from
active Pt sites. The high activity was suggested to be due to a sta-
bilization of the molecularly adsorbed isobutane molecule by the
zeolite, and by the assistance of K in facilitating isobutane dissoci-
ation.
When the dehydrogenation reaction was conducted at 300 ◦C,
no activity was observed. However, the H–D exchange between
propane-d0 and propane-d8 (Fig. 8) was more pronounced com-
pared with the H–D exchanges when the reaction was conducted
at 400, 500, or 600 ◦C. The occurrence of H–D exchange at 300 ◦C
and the lack of exchange at 400, 500, or 600 ◦C indicates a change
from reversible adsorption at lower temperature to irreversible
adsorption and subsequent reaction at temperatures higher than
400 ◦C. This can be explained by considering the energy diagram
presented in Scheme 1. At low temperature, adsorption of propane
is reversible, as the large activation energy barrier for double bond
formation prevents further reaction. However, single bond activa-
tion and exchange with the surface hydrogen pool does occur at low
temperature, and is a well-established process [34,35]. Chrysosto-
mou et al. [36] investigated the thermal activation of propyl groups
on Pt(1 1 1) and in one experiment, carried out TPD-MS of 1 and
2-propyl groups (5 L) adsorbed on a Pt(1 1 1) surface pretreated
with 20 L of deuterium. At −23 ◦C, they observed a TPD feature
for both propyl intermediates, with all fragments up to C2D5+,
indicating multiple H–D exchanges in individual molecules, up to
Fig. 6. Deuterium distribution in the recovered propane (equimolar feed of 25%
propane-d0/25% propane-d8/balance He) during the dehydrogenation reaction over
1%Pt/SiO2 at 600 ◦C, 1 atm, and WHSV of 2.6 h−1
.
conducted at 500 and 400 ◦C. We will not go into the details as to
why this is the case, because it is outside the scope of this paper.
Nevertheless, it is clear that carbon deposits play a significant role
in catalyst deactivation due to the occurrence of bimolecular reac-
tions on Pt sites (i.e., arising from the simultaneous contact of more
than one propane molecule over Pt sites). In contrast to the case of
hexane [16], the zeolite microstructure did not sufficiently control
the entry and diffusion of propane molecules to the Pt sites relative
to the Pt-catalyzed dehydrogenation of propane surface reaction.
Table 5
Activity and selectivity results for the dehydrogenation of propane over 1 wt% Pt/SiO2 at 600 ◦C, 1 atm, and 2.6 h−1 WHSV using an equimolar propane-d0 and propane-d8
(25%C3H8/25%C3D8/balance He) mixture.
TOS (min)
XC
(C%)
XC
(C%)
D
8
kH/kD
SPropene (%)
SMethane (%)
SEthene (%)
SEthane (%)
SBenzene (%)
H
3
8
3
45
80
115
150
185
220
15.4
11.2
7.6
7.1
7.0
7.8
5.6
3.2
3.2
3.5
3.7
1.98
2.01
2.36
2.24
1.97
1.89
66
64
59
55
53
51
20
21
24
25
26
27
10
13
16
18
20
21
2
1
1
1
1
1
0.5
0.9
0.3
0.3
0.0
0.0
7.1

270
K.G. Azzam et al. / Applied Catalysis A: General 390 (2010) 264–270
400 ◦C (Tables 2–4), a selectivity to propene in the range of 40–60%
was obtained at the tested tempertatures. The high selectivity of
methane combined with the low selectivities of ethane, ethene, and
benzene indicate that the dehydrogenation reaction is combined
with cracking, ␤-scission, and dehydrocyclization (with very low
activity) reactions.
4. Conclusions
Hydrogen chemisorption and DRIFTS revealed that the majority
of Pt clusters are ultra-small and well dispersed within the channels
of L-zeolite when the catalyst was prepared using the CVD method.
The existence of an NKIE (1.4–1.5) and the lack of H–D exchange
for the conversion of an equimolar mixture of propane-d0 and
propane-d8 over 1%Pt/KL catalyst at temperatures of 400, 500, and
600 ◦C and 1 atm, suggests that the adsorption of propane on Pt sites
is irreversible and that C–H bond activation is the rate determining
step for the dehydrogenation reaction of propane at these condi-
tions. In contrast to the case of hexane aromatization, the unique
structure of L-zeolite did not sufficiently inhibit bimolecular reac-
tions to prevent catalyst deactivation for propane dehydrogenation.
At 300 ◦C, the H–D exchange between propane-d0 and propane-d8
was more pronounced, suggesting that the adsorption of propane
is switched from irreversible to reversible at lower temperature.
Fig. 8. Deuterium distribution in the recovered propane (equimolar feed of 25%
propane-d0/25% propane-d8/balance He) during the dehydrogenation reaction over
1%Pt/KL at 300 ◦C, 1 atm, and WHSV of 2.6 h−1
.
and including propane-d8. This led the authors to conclude that a
cyclic adsorbed propyl-propene-propyl surface reaction took place,
with the 1-propyl species responsible for exchanging hydrogens at
the central carbon atom, and the 2-propyl species responsible for
exchanging end carbons.
References
[1] C.H. Bartholomew, R.J. Farrauto, Fundamentals of Industrial Catalytic Processes,
John Wiley & Sons, New Jersey, 2006, pp. 534–554.
At higher temperatures it is proposed that the activation energy
barrier for dehydrogenation is shifted lower; in this case, exchange
with the surface pool of hydrogen ceases, as the adsorbed propane
species rapidly dehydrogenates. At this point, adjacent adsorbed
propene molecules can react to form coke.
It is well established that the rate of isotope exchange rel-
ative to the rate of dehydrogenation reaction decreases and, as
a consequence, causes the degree of exchange to increase as
the temperature decreases [10]. For instance, Shi and Davis [10]
reported that the H–D exchange between methylcyclohexane-d14
and octane-d0 achieved almost equilibrium over Pt/SiO2 catalyst at
300 ◦C but not at 482 ◦C. Although H–D exchange occurs between
propane-d0 and propane-d8 (see Fig. 8), it is far from equilib-
rium, indicating some influence of the structure of the L-zeolite
on restricting propane molecules to Pt sites at lower temperature
(300 ◦C).
[2] N.R. Shiju, V.V. Guliants, Appl. Catal. A 356 (2009) 1–17.
[3] O.A. Bariås, A. Holmen, E.A. Blekkan, Catal. Today 24 (1995) 361–364.
[4] T. Waku, J.A. Biscardi, E. Iglesia, J. Catal. 222 (2004) 481–492.
[5] M.S. Kumar, A. Holmen, D. Chen, Micropor. Mesopor. Mater. 126 (2009)
152–158.
[6] P.L. de Cola, R. Gläser, J. Weitkamp, Appl. Catal. A 306 (2006) 85–97.
[7] M.S. Kumar, D. Chen, J.C. Walmsley, A. Holmen, Catal. Commun. 9 (2008)
747–750.
[8] P. Andy, M.E. Davis, Ind. Eng. Chem. Res. 43 (2004) 2922–2928.
[9] D. Akporiaye, S.F. Jensen, U. Olsbye, F. Rohr, E. Rytter, M. Rønnekleiv, A.I. Spjelka-
vik, Ind. Eng. Chem. Res. 40 (2001) 4741–4748.
[10] B. Shi, B.H. Davis, J. Catal. 147 (1994) 38–47.
[11] B. Shi, B.H. Davis, Stud. Surf. Sci. Catal. 101 (1996) 1145–1154.
[12] B. Shi, B.H. Davis, J. Catal. 168 (1997) 129–132.
[13] J.A. Bicardi, E. Iglesia, Catal. Today 31 (1996) 207–231.
[14] S.D. Jackson, J. Grenfell, I.M. Matheson, G. Webb, Stud. Surf. Sci. Catal. 122 (1999)
149–155.
[15] J.A. Biscardi, E. Iglesia, J. Catal. 182 (1999) 117–128.
[16] K.G. Azzam, G. Jacobs, W. Shafer, B.H. Davis, J. Catal. 270 (2010) 242–248.
[17] J.M. Hill, R.D. Cortright, J.A. Dumesic, Appl. Catal. A 168 (1998) 9–21.
[18] R.D. Cortright, J.A. Dumesic, Appl. Catal. A 129 (1995) 101–115.
[19] G. Jacobs, W.E. Alvarez, D.E. Resasco, Appl. Catal. A: Gen. 206 (2001) 267–282.
[20] G. Jacobs, F. Ghadiali, A. Pisanu, A. Borgna, W.E. Alvarez, D.E. Resasco, Appl.
Catal. A 188 (1999) 79–98.
Finally, comparing the selectivities of the products during the
dehydrogenation of propane over the Pt/KL catalyst at 600, 500, and
[21] M. Ballatreccia, R. Zanoni, C. Dossi, R. Psaro, S. Recchia, G. Vlaic, J. Chem. Soc.
Faraday Trans. 91 (1995) 2045–2049.
[22] E. Mielczarski, S.B. Hong, M.E. Davis, J. Catal. 134 (1992), 349–358 and 359–369.
[23] F. Ghadiali, G. Jacobs, A. Pisanu, A. Borgna, W.E. Alvarez, D.E. Resasco, Stud. Surf.
Sci. Catal. 130 (2000) 2537–2542.
[24] G. Larsen, G.L. Haller, Catal. Lett. 3 (1989) 103–110.
[25] G. Jacobs, F.A. Ghadiali, A.M. Pisanu, A. Borgna, W.E. Alvarez, D.E. Resasco, Appl.
Catal. 188 (1999) 79.
[26] M.M.J. Treacy, Micropor. Mesopor. Mater. 28 (1999) 271.
[27] G. Jacobs, F.A. Ghadiali, C. Padro, A. Borgna, W.E. Alvarez, D.E. Resasco, J. Catal.
191 (2000) 116.
[28] A. Yu Stakheev, E.S. Shpiro, N.I. Jaeger, G. Schulz-Ekloff, Catal. Lett. 34 (1995)
293–303.
[29] A. Yu Stakheev, E.S. Shpiro, N.I. Jaeger, G. Schulz-Ekloff, Stud. Surf. Sci. Catal. 98
(1995) 108–109.
[30] H.A. Dabbagh, T.S. Huang, B. Franzus, B.H. Davis, Tetrahedron 47 (1991)
949–960.
[31] A. Ozaki, Isotopic Studies of Heterogeneous Catalysis, Academic Press, New
York, NY, 1977.
[32] J.M. Hill, R.D. Cortright, J.A. Dumesic, Appl. Catal. A: Gen. 168 (1998) 9.
[33] R.D. Cortright, J.A. Dumesic, Appl. Catal. A: Gen. 129 (1995) 101.
[34] A. Farkas, Trans. Faraday Soc. 36 (1940) 522.
[35] J.A. Oliver, C. Kemball, R. Brown, E.W. Jamieson, J. Chem. Soc., Faraday Trans 1:
Phys. Chem. In Condensed Phases 81 (1985) 1871.
[36] D. Chrysostomou, C. French, F. Zaera, Catal. Lett. 69 (2000) 117.
Scheme 1.