Bifunctional catalytic system effective for oxidative dehydrogenation of 1-butene and n -butane using Pd-based intermetallic compounds

Article abstract of DOI:10.1021/cs500920p

The catalytic performance of Pd-based intermetallic compounds supported on silica (Pd_mM_n/SiO₂: M = Bi, Fe, Ge, In, Sn, Zn) was investigated in oxidative dehydrogenation of 1-butene and n-butane. A remarkable increase in selectivity (and also yield) of dehydrogenation products (1,3-butadiene and/or 1-butene) was obtained when PdIn, PdBi, or Pd₃Fe replaced monometallic Pd. Temperature-programmed reduction and X-ray diffraction (XRD) studies on the catalysts with various treatments (i.e., the fresh, spent, and calcined) suggested that redox of the second metal was involved in the catalysis. XRD, X-ray photoelectron spectra, and transmission electron microscopy-energy dispersive X-ray spectroscopy analyses using the spent PdIn/SiO₂ catalyst revealed the formation of a core-shell structure consisting of a PdIn₁δ core and Pd-In₂O₃ composite shell formed by oxidative decomposition of PdIn during the reaction. A mechanistic study suggested that the presence of In atoms, adjacent to Pd atoms, effectively inhibits the undesired combustion by capturing O₂ as lattice oxygen. The incorporated lattice oxygen reacts with hydrogen atoms derived from the hydrocarbon by C-H activation over Pd sites, resulting in the formation of water and oxygen vacancy and/or the parent intermetallic phase. We thus reveal that a combination of C-H activation by Pd and the redox of the second metal provides a unique and effective dehydrogenation system.

Full text of DOI:10.1021/cs500920p

Research Article
pubs.acs.org/acscatalysis
Bifunctional Catalytic System Effective for Oxidative
Dehydrogenation of 1‑Butene and n‑Butane Using Pd-Based
Intermetallic Compounds
Shinya Furukawa,† Masahiro Endo,‡ and Takayuki Komatsu*,‡
‡
†Department of Chemistry, Department of Chemistry and Materials Science, Tokyo Institute of Technology, 2-12-1-E1-10
Ookayama, Meguro-ku, Tokyo 152-8551, Japan
S
* Supporting Information
ABSTRACT: The catalytic performance of Pd-based intermetallic
compounds supported on silica (PdmMn/SiO2: M = Bi, Fe, Ge, In, Sn,
Zn) was investigated in oxidative dehydrogenation of 1-butene and n-
butane. A remarkable increase in selectivity (and also yield) of dehydrogen-
ation products (1,3-butadiene and/or 1-butene) was obtained when PdIn,
PdBi, or Pd3Fe replaced monometallic Pd. Temperature-programmed
reduction and X-ray diffraction (XRD) studies on the catalysts with various
treatments (i.e., the fresh, spent, and calcined) suggested that redox of the
second metal was involved in the catalysis. XRD, X-ray photoelectron
spectra, and transmission electron microscopy−energy dispersive X-ray
spectroscopy analyses using the spent PdIn/SiO2 catalyst revealed the
formation of a core−shell structure consisting of a PdIn1−δ core and Pd−In2O3 composite shell formed by oxidative
decomposition of PdIn during the reaction. A mechanistic study suggested that the presence of In atoms, adjacent to Pd atoms,
effectively inhibits the undesired combustion by capturing O2 as lattice oxygen. The incorporated lattice oxygen reacts with
hydrogen atoms derived from the hydrocarbon by C−H activation over Pd sites, resulting in the formation of water and oxygen
vacancy and/or the parent intermetallic phase. We thus reveal that a combination of C−H activation by Pd and the redox of the
second metal provides a unique and effective dehydrogenation system.
KEYWORDS: oxidative dehydrogenation, butadiene, intermetallic compound, bifunctional, 1-butene
and known to be effective for the oxidative dehydrogenation of
1-butene to 1,3-butadiene.4−7 On the other hand, very few
researchers have studied oxidative dehydrogenation of hydro-
carbons using noble metal catalysts,12,13 despite possessing
greater C−H activation abilities than transition metal oxides,
which is extremely important in the dehydrogenation process.
This may be a product of the highly active properties of noble
metals causing undesired combustion of hydrocarbons into
COx, significantly decreasing selectivity. In this context, the
development of an effective catalyst based on the use of a noble
metal for oxidative dehydrogenation would open up a new field
of catalytic dehydrogenation. It is an ideal approach to modify
the nature of the metallic catalysts so that combustion is
suppressed without lowering their C−H activation ability. One
can modify the nature of metallic catalysts by changing catalyst
support, adding cocatalysts, and by the formation of an alloy or
intermetallic phase. We previously reported that some
intermetallic catalysts exhibited superior catalytic performances
for the nonoxidative dehydrogenation of hydrocarbons than
monometallic catalysts, for example, Ni3Sn/SiO2 and Ni3Sn2
INTRODUCTION
■
1,3-Butadiene is an important feedstock of butadiene polymer,
the demand for which is expanding.1 In general, it is produced
by dehydrogenation of 1-butene.2 The dehydrogenation
process is divided into two categories: (1) nonoxidative and
(2) oxidative dehydrogenation,3 as described in the following
equations.
C4H8 → C4H6 + H2
ΔH° = 110kJ·mol−1
(1)
C4H8 + 1/2 O2 → C4H6 + H2O ΔH° = −132kJ·mol−1
(2)
The nonoxidative process is highly endothermic, requiring
high temperatures to obtain a reasonable conversion. In
addition, the deposition of coke should be suppressed in
order to maintain the catalytic performance. In this context,
oxidative dehydrogenation possesses attractive advantages. The
reaction is exothermic with no limitation by equilibrium;
therefore, it is essentially possible to obtain a complete
conversion at low temperatures. Moreover, deposited coke
can be removed by oxidation into COx.
Oxidative dehydrogenation of hydrocarbons has been widely
investigated using metal oxide catalysts.4−11 Among a number
of oxide catalysts, bismuth molybdates are widely investigated
Received: June 28, 2014
Revised: September 1, 2014
Published: September 2, 2014
© 2014 American Chemical Society
3533
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
showed higher selectivities than Ni/SiO2 in the dehydrogen-
ation of cyclohexane to benzene.14 Pt3Sn/SiO2 exhibited not
only higher selectivity but also greater catalytic activity than Pt/
SiO2 in n-butane dehydrogenation to butenes.15 Although this
Pt-based catalyst showed high catalytic performance in
dehydrogenation of n-butane, the high oxidizing ability of Pt
may result in a larger amount of combustion to COx in the
oxidative dehydrogenation condition. Therefore, we focused on
Pd-based intermetallic compounds as candidates of effective
catalysts for oxidative dehydrogenation of 1-butene.
In this study, we investigated the catalytic properties of
various Pd-based intermetallic compounds supported on SiO2
(PdmMn/SiO2: M = Bi, Fe, Ge, In, Sn, Zn) in the oxidative
dehydrogenation of 1-butene, for the formation of 1,3-
butadiene. Dehydrogenation of n-butane to 1-butene and 1,3-
butadiene was also examined using the Pd-based catalysts. A
mechanistic study and detailed characterizations were per-
formed to understand the catalysis of intermetallic compounds.
Herein, we report the development of catalysts superior to
monometallic Pd and unique bifunctional catalysis played by Pd
and the second metal.
Characterization. The crystal structure of prepared
intermetallic compounds was examined by powder X-ray
diffraction (XRD) with a Rigaku RINT2400 using an X-ray
Cu Kα source. Estimation of the crystallite size was performed
for the most intense peak, based on the Scherrer equation with
a 0.9 Scherrer constant. Transmission electron microscopy
(TEM) was conducted on JEOL JEM-2010F at the accelerating
voltage of 200 kV. Quantitative elemental analysis of a
microscopic region was also performed using energy dispersive
X-ray spectroscopy (EDX) equipped with TEM (EDAX,
Genesis). In order to prepare the TEM specimen, all samples
were sonicated in tetrachloromethane and dispersed on a
copper grid supported by an ultrathin carbon film. X-ray
photoelectron spectra (XPS) of fresh and spent catalysts were
recorded using an ULVAC PHI 5000 VersaProbe spectrometer.
Fresh catalysts were pressed into a pellet and placed in a quartz
reactor, where it was reduced under flowing hydrogen (60 mL·
min−1) at 600 °C for 1 h. Spectra were obtained with an Al Kα
X-ray source using C 1s (248.8 eV) for the calibration of
binding energy. For spent catalysts, boron nitride (BN, 99%;
Wako) was physically mixed into the sample with no reduction
pretreatment performed. Energy calibration was conducted
with N 1s (398.1 eV) of BN to avoid the contribution of
formed coke to the calibration with C 1s. The reduction
behavior of the second metal M of PdmMn/SiO2 was examined
by temperature-programmed reduction (TPR). Prior to
measurement, the prepared catalyst was calcined at 400 °C
for 4 h in air. Under flowing H2(5%)/Ar, the temperature of the
sample bed was raised from room temperature to 800 °C at a
heating rate of 10 °C·min−1, and the consumption of hydrogen
was continuously measured by a thermal conductivity detector
(TCD). The amount of coke formed on the catalysts during the
reaction was determined by thermogravimetry (TG) using an
SII TG/DTA 7200. The sample (16 mg) was heated from
room temperature to 800 °C at a ramping rate of 10 °C·min−1,
under a flow of air. The amount of exposed Pd on the catalyst
was estimated by CO pulse adsorption. Prior to measurement,
the catalyst was reduced under flowing hydrogen (60 mL·
min−1) at 400 °C for 0.5 h, followed by He treatment (60 mL·
min−1) at 400 °C for 1 h to remove adsorbed hydrogen. After
cooling to room temperature, the catalyst bed was cooled to
−80 °C (dry ice with acetone). A specific amount of CO(5%)/
He pulse was introduced to the catalyst bed and passed
(unadsorbed) CO was measured by a TCD. The reduction
treatment was omitted for the spent catalyst.
Dehydrogenation of Hydrocarbons. Oxidative dehydro-
genation of 1-butene was conducted in a fixed bed flow system
with a quartz reactor of 8 mm internal diameter. The dead
volume of the reactor was minimized by casting a quartz tube
(6 mm outside diameter) to reduce the contribution of
hydrocarbon thermal reaction. The catalyst (50 mg) was
reduced under flowing hydrogen (99.9999%, 60 mL·min−1;
Nippon Sanso) at 400 °C for 1 h prior to the reaction. The
reactor was purged with He, followed by feeding of the reaction
mixture: 1-butene (>98%, 15 mL·min−1; Takachiho Kogyo), O2
(15 mL·min−1), and He (90 mL·min−1) at 400 °C. The
produced hydrocarbons, including CH4, were analyzed using a
flamed-ionization-detector gas chromatograph (FID−GC,
Shimadzu GC-14B) with a column of CP-Silica PLOT
(VARIAN, 0.32 mm × 30 m). COx (CO and CO2), O2, and
CH4 were analyzed using an on line TCD−GC (Shimadzu GC-
8A) with a column of active carbon (GL science). CH4 was
analyzed using both FID−GC and TCD−GC to integrate the
EXPERIMENTAL SECTION
■
Catalyst Preparation. Monometallic Pd/SiO2 was pre-
pared by pore-filling impregnation. An aqueous solution of
Pd(NO3)2 (4.8 × 10−2 g·mL−1, 0.64 mL, N.E. ChemCat) was
diluted with ion-exchanged water and added to silica gel (5.0 g,
Cariact G-6, Fuji Silysia Co., SBET = 673 m2·g−1) previously
dried at 130 °C and cooled in air to room temperature. The
quantity of palladium solution was calculated to fill the silica gel
pores, achieving a palladium loading of 3 wt %. The mixture
was sealed with plastic film overnight at room temperature. It
was then dried over a boiling water bath with stirring, followed
by drying at 130 °C for 1 h, and subsequently calcined in the air
at 400 °C for 1 h. After the calcination, the catalyst was reduced
under flowing hydrogen (99.9995%, 60 mL·min−1; Taiyo
Nippon Sanso) at 400 °C for 2 h. The catalyst was cooled to
room temperature with flowing helium and kept in a desiccator.
Pd/In2O3 was prepared in a similar way using In2O3 (99%,
Wako) as a support. Pd-based intermetallic compound catalysts
(PdmMn/SiO2; M = Ge, In, Sn, Zn) were prepared by
successive impregnation with Pd/SiO2. In the case of Pd−Zn
(1:1), an aqueous solution of Zn(NO3)2·6H2O (99%, Kanto)
was added to Pd/SiO2 so that the atomic ratio of Pd/Zn was
adjusted to 1. The mixture was sealed using plastic film and left
overnight at room temperature. It was then dried over a boiling
water bath with stirring and reduced under flowing hydrogen
(99.9995%, Taiyo Nippon Sanso) at 800 °C 2 h. Other
intermetallic catalysts were prepared in a similar way using
specific second metal precursors, solvents, and reduction
temperatures: PdIn, In(acetylacetonate)3 (95%, Kanto), nitric
acid (60%, Kanto), 400 °C; Pd2Ge, tetrabutylgermane (98%,
Aldrich), n-hexane (, 99%, Kanto), 700 °C; Pd3Sn2, SnCl2·
2H2O (98%, Kanto), H2O/MeOH (1:1), 600 °C. PdBi/SiO2
was prepared by coimpregnation in a mixed nitric acid solution
of Pd(NO3)3 and Bi(NO3)3 and reduced under H2 at 500 °C
for 1 h. Pd3Fe/SiO2 was also prepared by coimpregnation in a
similar manner to PdBi/SiO2 using Fe(acetylacetonate)3 (99%,
Soekawa) at an 800 °C reduction temperature. In2O3/SiO2 (In:
3 wt %) was prepared by impregnation of a nitric acid solution
of In(acetylacetonate)3 in a pore-filling fashion. The impreg-
nated catalyst was then calcined in the air at 400 °C for 1 h.
3534
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
amounts of hydrocarbons and COx. Oxidative dehydrogenation
of n-butane (>99%, Takachiho Kogyo) to 1,3-butadiene was
also conducted in a similar fashion, at a reduction and reaction
temperature of 500 °C with 30 mg of catalyst. The reaction
using 18O2 was conducted in a glass circulation system
equipped with a vacuum line and a Quadrapole Mass
Spectrometer (Spectra International, MICROVISION). Pd/
SiO2 (50 mg) was reduced under flowing hydrogen (60 mL·
min−1) at 400 °C for 0.5 h, followed by evacuation. A mixture
of 1-butene (1.2 kPa) and 18O2 (>99 atom %, 0.46 kPa;
ISOTEC) was introduced into the reactor at 400 °C. In the
case of PdIn/SiO2 (200 mg), a similar procedure was
employed, except that the catalyst was oxidized under 2.0 kPa
of 18O2 at 400 °C after the reduction treatment. The H2O/CO2
molar ratio was determined as a product of the corresponding
mass intensity ratio and the correlation coefficient of the
calibration curve (0.68).
RESULTS AND DISCUSSION
Figure 1 shows XRD patterns of the prepared catalysts. Pd/
SiO2 displayed two broad peaks assigned to 111 and 200
■
Figure 2. TEM images of (a) Pd/SiO2 and (b) PdIn/SiO2.
intermetallic compounds were larger than that of monometallic
Pd, and differed from each other. No obvious correlation was
observed between crystallite size and preparation procedure
(successive or coimpregnation), reduction temperature, and
Pd/M ratio. Interestingly, PdIn was much larger in size (50
nm) than other intermetallic compounds, despite having with
the lowest reduction temperature (400 °C). This may be due to
the low melting point of In (157 °C); the lower melting point
increases its mobility on the SiO2 surface and promotes
aggregation. In the XRD pattern of PdIn, a broad feature
overlapped with the intense sharp peak was observed. Peak
deconvolution with two Lorenzian curves well-reproduced the
experimental data (Supporting Information, Figure S1). The
broader component gave a 5 nm estimation of the crystallite
size. The presence of small and large particles was confirmed by
TEM, as shown in Figure 2b. The estimated crystallite sizes
were consistent with the actual particle size observed in the
TEM image.
The catalytic properties of the prepared intermetallic
compounds were investigated in the oxidative dehydrogenation
of 1-butene to 1,3-butadiene, as shown in Figure 3. In all cases,
isomerization into 2-butene and combustion into COx also
proceeded, as side or main reactions. As anticipated,
combustion predominantly proceeded over Pd/SiO2; therefore,
the 1,3-butadiene yield was very low. In contrast, intermetallic
catalysts, with the exception of Pd3Sn2, showed higher 1,3-
butadiene selectivities and yields than Pd/SiO2, with com-
parable 1-butene conversion. Pd3Fe had the highest 1,3-
butadiene yield, followed by PdIn and PdBi. The catalytic
properties of these catalysts were also tested in the oxidative
dehydrogenation of n-butane (Figure S2). Dehydrogenation
products (1-butene and 1,3-butadiene) were hardly generated
over Pd/SiO2. The selectivity of the combustion products
(COx: 97 C-%) was higher than in 1-butene dehydrogenation.
On the other hand, dehydrogenation occurred to some extent
over intermetallic compounds. PdIn showed much higher
selectivity to dehydrogenation products (20 C-%) than
monometallic Pd (1.8 C-%). Although the selectivity was not
sufficiently high, emphasis should be placed on the remarkable
increase in selectivity using the intermetallic catalyst. The order
Figure 1. XRD patterns of Pd-based intermetallic compounds
supported on silica. Figures in parentheses indicate crystallite size.
diffractions of metallic Pd at 40.1° and 45.2°, respectively. The
crystallite size of Pd particles was estimated as 6 nm using
Scherrer’s equation with the 111 diffraction. The crystallite size
was roughly consistent with the particle size (5−9 nm)
observed in the TEM image, as shown in Figure 2a. For the
Pd−Bi (1:1) catalyst, the desired PdBi intermetallic phase was
predominantly observed, indicating the formation of the PdBi
phase as the main species. A small broad feature assigned to the
Pd3Bi phase was also observed.16 The XRD pattern of the Pd−
Fe (3:1) catalyst agreed with that of the intermetallic Pd3Fe
phase. The appearance of the 110 diffraction at 33.1° strongly
supported the formation of a Pd3Fe intermetallic phase (space
group: Pm−3m) but not of a Pd0.75Fe0.25 alloy phase (space
group: Fm−3m). In general, the 110 diffraction of the fcc
structure (Fm−3m) is not observed by distinction rule due to
its high symmetry. For other bimetallic systems, the desired
intermetallic compounds (Pd2Ge, PdIn, Pd3Sn2, PdZn) were
formed in a single-phase. The estimated crystallite sizes of the
3535
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
to a decrease in the PdIn lattice constant by an increase in Pd/
In ratio. These results strongly indicate that the formation of
In2O3 from PdIn resulted in the formation of PdIn1−δ. The
actual composition ratio can be estimated as Pd0.55In0.45 (δ =
0.18) by shifting the peak position. It is also likely some Pd
formed small monometallic clusters unobservable with XRD.
Calcination of PdIn/SiO2 under air caused a disappearance of
the PdIn intermetallic phase and a formation of In2O3, Pd, and
PdO, showing complete decomposition of the PdIn phase (c).
The detection of metallic Pd by XRD suggests that the small Pd
clusters grew up to be observable with the deep oxidation of
PdIn. We also performed oxidative dehydrogenation of 1-
butene using the calcined PdIn/SiO2 catalyst. Surprisingly, the
spent catalyst showed a similar XRD pattern to (b): peaks
assigned to In2O3 and PdIn1−δ were observed (d). This result
indicates that a part of In2O3 was reduced by 1-butene and
PdIn intermetallic phase was regenerated. The reduction of
In2O3 by 1-butene was confirmed in dehydrogenation of 1-
butene without O2 using the calcined PdIn/SiO2; In2O3
completely disappeared and PdIn was regenerated (e). In this
case, 1-butene can be regarded as a reducing agent (or
hydrogen donor) for In2O3, in other words, dehydrogenation of
1-butene is induced by the lattice oxygen of In2O3. The
pronounced reducibility of In2O3 can be found in the
literature;18,19 Penner et al. reported that In2O3 was reduced
not only by H2 but even by CO. Based on these results, redox
of In and dehydrogenation by lattice oxygen are significantly
involved in the catalytic cycle of the Pd−In system.
Figure 3. Catalytic performances of Pd-based catalysts in the oxidative
dehydrogenation of 1-butene.
of catalytic performances (yield of dehydrogenation products)
was similar to that in 1-butene dehydrogenation (1,3-butadiene
yield). In both dehydrogenation reactions, conversions of O2 in
the case of most catalysts were ca. 100%.
To understand the mechanism of the remarkable increase in
the catalytic performances, we subsequently performed detailed
characterizations using PdIn/SiO2 that showed high catalytic
performances in both dehydrogenation reactions. Figure 4
In this context, it is important to investigate the effect of the
oxide species reducibility on catalytic performance. We
therefore performed a TPR study using calcined intermetallic
catalysts and Pd/SiO2, as shown in Figure 5. The TPR profile
Figure 4. XRD patterns of a PdIn/SiO2 catalyst after treatment at 400
°C with various conditions: (a) H2 reduction, (b) 1-butene/O2/He
(reaction) after (a), (c) oxidation in the air, (d) 1-butene/O2/He
(reaction) after (c), and (e) 1-butene/He after (c).
Figure 5. TPR profiles of Pd-based catalysts calcined at 400 °C. Filled
triangles indicate the position of the most intense peak assigned to the
reduction of the second metal oxide.
shows XRD patterns of PdIn/SiO2 catalyst pretreated under
various conditions. PdIn intermetallic phase was solely
observed after reduction pretreatment (a). PdIn/SiO2 catalyst
spent in oxidative dehydrogenation of 1-butene showed small
peaks assigned to In2O3 with the 111 diffraction of remaining
PdIn (b), indicating some In in PdIn was oxidized by O2 during
the reaction. Formation of In2O3 from PdIn should be
associated with the formation of Pd. However, no Pd particle,
observable by XRD, was detected. Interestingly, the peak
position of the 111 diffraction of PdIn was shifted slightly
higher in angle. The PdIn (space group: Pm−3m) intermetallic
phase is classified in the Berthollide-type intermetallic
compound, having a small solubility range (e.g., 50−57 Pd
atom % for PdIn).17 Therefore, the peak shift can be attributed
of the calcined Pd/SiO2 did not reveal any positive
consumption peak, only a negative peak. It has been reported
that PdO is reduced by H2, even at room temperature.20−24
Therefore, Pd should be in a metallic state before the
temperature becomes elevated. The observation of a negative
peak in the TPR profile of Pd is well-known and attributed to
the release of H2 from β-PdHx species.20−25 Similar negative
peaks were also observed with the calcined intermetallic
compounds. In the case of Pd-based bimetallic sys-
tems,21,23,26,27 the formation of PdHx species tends to be
inhibited, relative to monometallic Pd. Therefore, the
appearance of the negative peaks supports the decomposition
3536
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
of the intermetallic phase into Pd and the second metal oxide.
For intermetallic compounds, several reduction peaks were
observed. We classified these into two categories: lower (100−
250 °C) and higher (>300 °C) temperature. The lower group
can be assigned to the reduction of PdO,21,23,28,29 with a
relatively strong interaction with the second metal oxide.21,29
The higher category is assigned to the second metal
oxide.22,28,29 The reduction temperatures of the lower group
were not so different from each other, whereas those of the
higher category differed significantly depending on the second
metal. Figure 6 shows the relation between the product yield in
Figure 7. Catalytic performances in 1-butene dehydrogenation over
various In2O3-containing materials.
required. Therefore, we conducted FT-IR, XPS and TEM−
EDX analyses on the PdIn/SiO2 catalysts spent in oxidative
dehydrogenation of 1-butene. Figure 8 shows the FT-IR spectra
Figure 6. Relation between yield of dehydrogenation products in the
oxidative dehydrogenation of 1-butene or n-butane over Pd-based
intermetallic catalysts and reduction peak temperature of the second
metal in TPR profiles.
oxidative dehydrogenation of n-butane or 1-butene and the
temperature of the most intense reduction peak (filled
triangles) of the second metal oxide in TPR profiles. The
lower reduction temperature gave a higher product yield in
both reactions. These trends were prominent below reduction
temperatures of 400 or 500 °C, which was the reaction
temperature of the dehydrogenation of 1-butene or n-butane,
respectively. These results strongly indicate that the reduction
of the second metal oxide is the key step for the bimetallic
catalysis and, hence, supports involvement of the redox of the
second metal.
To understand the interplay of Pd and In in the catalysis,
several control experiments were tested using the following
catalysts: Pd/SiO2, PdIn/SiO2, In2O3/SiO2, Pd/In2O3, and Pd/
SiO2 + In2O3 (physical mixture). Figure 7 shows their catalytic
performances in the oxidative dehydrogenation of 1-butene.
In2O3/SiO2 had much lower conversion and selectivity than
Pd/SiO2 and PdIn/SiO2, suggesting that Pd is necessary to
activate the C−H bonds for dehydrogenation. In this case, 2-
butene was generated as a main product with almost 1:1 cis/
trans ratio (data not shown), indicating that isomerization of 1-
butene proceeded over Brønsted acid sites on In2O3.30 Pd/
In2O3 exhibited higher conversion and selectivity than Pd/SiO2
and a comparable 1,3-butadiene yield to that of PdIn/SiO2.
This result suggests that a combination of Pd and In2O3 is
important to obtained high catalytic performance. Physical
mixture of Pd/SiO2 and bulk In2O3, however, did not enhance
selectivity compared to Pd/SiO2. On the basis of these results,
we concluded that a key point for selective dehydrogenation is
that Pd and In2O3 are adjacent in the atomic order.
Figure 8. FT-IR spectra of CO adsorbed on fresh Pd/SiO2 and PdIn/
SiO2. Data for PdIn/SiO2 spent in oxidative dehydrogenation of 1-
butene (for 2 and 5 min) and oxidized by O2 at 400 °C for 1 h were
also shown.
of CO adsorbed on fresh and spent catalysts. CO adsorbed on
fresh Pd/SiO2 exhibited peaks assigned to linearly adsorbed CO
on Pd at 2088 cm−1 and bridged CO at around 1950 cm−1.31
For fresh PdIn/SiO2, no bridged species were observed, which
corresponds to the disappearance of Pd ensembles by the
formation of the intermetallic compound. The peak position of
linear CO on PdIn/SiO2 was lower in energy than that on Pd/
SiO2 (2065 cm−1), indicating that Pd atoms in PdIn were
negatively charged compared with those in monometallic Pd.
This result is consistent with a similar previously reported FT-
IR study.32 After the reaction, however, the peak position
shifted to that of Pd/SiO2. This position was very close to that
of PdIn/SiO2 oxidized by O2, which decomposed into Pd and
To discuss the nature of the reaction site and mechanism in
detail, multiple characterization of the catalyst structure is
3537
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
In2O3. These results support the formation of monometallic Pd
at the catalyst surface under the reaction condition. Moreover,
the absence of bridged CO on the spent PdIn/SiO2 indicates
that the generated Pd was highly dispersed or isolated. The
slight deviation of the peak position between pure Pd/SiO2 and
the spent or oxidized PdIn/SiO2 may be due to a difference in
the surrounding species of Pd, either SiO2 or In2O3. Figure 9
PdIn phase and remained as In2O3. The presence of a small
amount of In2O3 corresponds to a slight surplus of Pd species
due to the 1:1 Pd:In loading ratio. Indeed, a trace amount of
monometallic Pd was observed in the Pd 3d XPS as a small
shoulder. The composition of In2O3 increased, reaching almost
100% as the reaction proceeded. These results support the
decomposition of PdIn into Pd and In2O3. Although PdIn1−δ
was mainly observed by XRD, the XPS results did not indicate
the presence of intermetallic Pd species at the catalyst surface
after 25 min. According to these results, a core−shell structure,
in which the PdIn1−δ core is covered with a In2O3 shell
containing Pd clusters, is likely to be formed in the reaction.
This is consistent with PdIn being oxidized by O2 from outside
of the particle. Comparing the results of FT-IR and XPS, the
change in the catalyst structure observed in the FT-IR study
completed at a much earlier stage (5 min) than that in the XPS
study (25 min). This deviation can be attributed to the fact that
XPS includes information about the near-surface region,
whereas FT-IR study using CO as a probe molecule only
provides the information on the outmost layer. These are also
consistent with the oxidation of PdIn from the outside. Figure
10 shows TEM images of nanoparticles on the fresh and the
Figure 9. (a) Pd 3d and (b) In 3d XPS of PdIn/SiO2 before and after
oxidative dehydrogenation of 1-butene.
shows (a) Pd 3d and (b) In 3d XPS for fresh and the spent
catalysts. Pd/SiO2 indicated 3d5/2 and 3d3/2 emissions at 335.5
and 340.8 eV, respectively, consistent with the reported values
of metallic Pd supported on SiO2.33 The peak positions of Pd
3d5/2 and 3d3/2 for the fresh PdIn/SiO2 (336.3 eV for 3d5/2 and
341.5 eV for 3d3/2) were higher in energy than those of
monometallic Pd/SiO2. The higher binding energy with PdIn/
SiO2 is seemingly contradictory to the negatively charged Pd as
mentioned in the FT-IR study. Similar results have also been
reported for PdGa27,34 and PdZn.35−37 In these cases, valence
band XPS and DFT calculations revealed that the valence band
of the intermetallic compound was occupied to a higher degree
than that of monometallic Pd, making Pd atoms negatively
charged.27,34 The higher binding energy in the Pd 3d core level
was interpreted as a change in the screening of the core holes
by the increased occupancy of the valence band.37 As the
reaction proceeded, the peak feature gradually changed into
that of the monometallic Pd. After 25 min, the peaks agreed
well with that of Pd/SiO2. The In 3d XPS of flesh PdIn/SiO2
are deconvoluted into two components: metallic In in PdIn
(444.1 and 451.6 eV)38 and In3+ (444.8 and 452.4 eV).38,39 The
presence of the In3+ species in fresh PdIn/SiO2 indicates that a
small fraction of In did not participate in the formation of the
Figure 10. TEM images of PdIn/SiO2 catalyst: (a) fresh and (b) spent
in the dehydrogenation of 1-butene. Circles indicate where EDX
analysis was performed.
spent PdIn/SiO2 catalysts. Chemical analyses by EDX were also
conducted at the edge and center of the nanoparticles (Nos. 1−
6 designated with circles). The atomic ratio of Pd, In, and O at
each point is listed in Table 1. Fresh catalyst had Pd/In ratios
close to 1:1, regardless of the position (Nos. 1−3), supporting
homogeneous formation of the PdIn intermetallic compound.
However, In contents were slightly lower than Pd, with the
differences close to O content. This is consistent with the
observation of In2O3 at the catalyst surface by XPS. After the
reaction, the structure of the nanoparticle drastically changed,
as shown Figure 9b; the particle indeed had a core−shell
structure consisting of a dark core and relatively bright shell.
The EDX analysis on the particle center (No. 5) revealed a
composition ratio with high Pd content: Pd (55) > In (28) > O
(17). Assuming that Pd atoms form Pd0.55In0.45 and
3538
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
Table 1. EDX Analysis of Atomic Composition of Fresh and
Spent PdIn/SiO2 Catalysts
Scheme 1. (a) Change in Catalyst Structure during the
Reaction and (b) Proposed Reaction Mechanism of the
Oxidative Dehydrogenation of 1-Butene to 1,3-Butadiene
b
atomic composition (%)
a
over PdIn/SiO2
a
catalyst
fresh
site no.
Pd
In
O
1
2
3
4
5
6
50.7(2)
48.2(5)
47.3(1)
20.3(1)
54.8(2)
38.2(5)
43.7(3)
42.2(8)
42.5(2)
50.6(1)
27.7(4)
31.5(8)
6.3(13)
9.6(32)
10.2(7)
29.3(1)
17.5(9)
spent
30.3(14)
a
b
Site numbers correspond to those designated in Figure 9. Figure in
parentheses indicates an error in the first decimal place.
monometallic Pd, we can estimate the composition ratio of Pd/
Pd0.55In0.45/In2O3 as 6.1:3.6:1.0. A considerable amount of Pd
seems to be in the monometallic phase, consistent with XPS
results. On the other hand, the edge regions were mainly
composed of In and O (Nos. 4 and 6). Therefore, formation of
a core−shell structure consisting of the PdIn1−δ core and the
Pd−In2O3 shell was demonstrated. The relatively high Pd
content at the center can be attributed to the presence of
noncrystalline Pd clusters at the surface or near-surface region.
Because EDX analysis is performed by acquiring transmitted X-
rays along with the zone axis of the observation, a small
contribution of the near surface region should be included in
the output. Based on the Pd content comparable with the In
content at the shell (No. 6) and the substantial width of the
shell, the observed high Pd content at the center (No. 5) is
compatible with the core−shell structure. Figure S3 shows a
high-resolution TEM image of the core−shell particle. At the
shell region, lattice fringes (4.12 Å) corresponding to In2O3
(211) planes (4.13 Å) were clearly observed. In contrast, at the
core region, lattice fringes with 2.32 Å were observed
overlapped with those of In2O3, which is consistent with the
d spacing of PdIn (110) planes (2.30 Å). However, this spacing
is also similar to that of Pd (111) (2.25 Å). Considering a few
percent error, an additional support is required for assignment
to the PdIn phase. In this context, the fact that the core is
composed of a large crystallite reinforces the assignment, as the
presence of crystalline PdIn has been demonstrated by XRD
analysis.
a
Reaction mechanism: (1) oxidative decomposition of PdIn to Pd and
In2O3, (2) dehydrogenation of 1-butene to 1,3-butadiene at the
interface, (3) reoxidation of In by O2, (4) over-reduction, and (5)
combustion by O2.
dehydrogenation of the hydrocarbon occurs at the interface
through C−H activation via Pd and hydrogen consumption by
the lattice oxygen of In2O3 to form water (2). This step can be
regarded as partial reduction of In2O3 by the hydrocarbon. The
oxygen vacancy is filled by reoxidation with O2 (3). Over-
reduction of In2O3 to metallic In may regenerate the parent
PdIn phase (4). O2 is also used for combustion of hydrocarbon
over Pd sites (6). Therefore, the selectivity for dehydrogenation
strongly depends on whether O2 is taken to form lattice oxygen
of In2O3 or is directly used for combustion. In other words,
incorporation of O2 into the lattice of In2O3 decreases its
opportunity to be used in combustion. In the case of the Pd/
In2O3 catalyst, a similar chemical process may occur at the
interface of Pd particles and In2O3 supports. Indeed, it has been
reported that a PdIn phase is formed by reduction treatment
with Pd/In2O3 and that the subsequent oxidation treatment
causes decomposition of the PdIn phase to Pd and In2O3.40
However, one should consider the ratio of Pd−In2O3 interfacial
sites and monometallic Pd sites, which has a determining
influence on selectivity for dehydrogenation. In the case of
PdIn, the adjacency of Pd and In, on an atomic scale, provides
highly dispersed Pd sites and maximizes the number of Pd−
In2O3 sites. In contrast, in the case of Pd/In2O3, the
corresponding process can occur only at the interfacial site.
This difference may be the reason for the 2-fold deviation in
selectivity.
Scheme 1a illustrates the change in the catalyst structure
during the reaction. In atoms of PdIn are oxidized into In2O3
from the surface by O2. Generation of In2O3 from PdIn forms
small monometallic Pd clusters that are unobservable by XRD.
The remaining intermetallic phase core is covered with a In2O3
layer shell containing small Pd clusters and the In content of
the core is decreased slightly by the transformation to PdIn1−δ
.
According to the results observed above, a similar structure was
formed under dehydrogenation of 1-butene regardless of
whether the reduced or calcined catalyst was used, and the
catalyst seemed to reach a steady-state structure. In this case,
both catalysts showed comparable 1,3-butadiene yields even
after 5 min of TOS (Table S1). This is consistent with the
result of the FT-IR study that the catalyst structure at the
surface reaches the steady-state at the initial stage of the
reaction.
On this experimental basis, we proposed the following
reaction mechanism for the oxidative dehydrogenation of
hydrocarbons over PdIn/SiO2 (Scheme 1b). Initially, surface
PdIn is oxidized into Pd and In2O3 (1). Pd and In2O3 are
placed adjacent to each other during this process. Oxidative
To confirm that the Pd−In2O3 composite indeed shows a
higher selectivity to oxidative dehydrogenation than Pd, we
evaluated H2O/CO2 ratios in the reactions (Scheme 2). For
example, oxidative dehydrogenation of one 1-butene molecule
yields equimolar amounts of 1,3-butadiene and H2O, whereas
combustion of 1-butene gives 4 CO2 and 4 H2O (Scheme 2a).
Therefore, the H2O/CO2 ratio reflects the selectivity to
3539
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
Scheme 2. (a) Products in the Oxidative Dehydrogenation
and Combustion of 1-Butene and [H218O]/[C18O2] Ratios
Obtained in the Presence of (b) 1-Butene, 18O2, and Pd/
SiO2 and (c) 1-Butene and Pd−In218O3/SiO2
dehydrogenation or combustion. To evaluate the ratio in the
reaction with O2 in the absence of In2O3, 1-butene and 18O2
were introduced on Pd/SiO2 at 400 °C. 18O2 was used to
distinguish CO2 (C16O2, MW: 44) from C3H8 (MW: 44),
fragmented from C4H8. The ratio, 1.4, was close to unity
(Scheme 2b), indicating that combustion was dominant. On
the other hand, the reaction over Pd−In218O3 composite
(PdIn/SiO2 calcined under 18O2) gave a much higher H2O/
CO2 ratio (4.5, Scheme 2c). These results demonstrated that
combustion was indeed suppressed in the presence of In2O3.
We subsequently performed a kinetic study in oxidative
dehydrogenation of 1-butene over Pd/SiO2 and PdIn/SiO2.41
The dependences of the formation rate of 1,3-butadiene on
partial pressure of 1-butene (PC H ) and O2 (PO ) were as
Figure 11. Time course of 1,3-butadiene yield (red circles) and coke
amount (black diamonds) in oxidative dehydrogenation of 1-butene
over PdIn/SiO2 and the relation between them (blue triangles).
slight deactivation occurred, it should be noted that this
catalytic system possesses resistance to coke accumulation. The
saturation of coke amount during the reaction is probably due
to a steady state being reached between coke formation and
combustion. We also measured the change in the amount of
exposed Pd by CO adsorption among the spent catalysts
(Figure S5). The amount of exposed Pd was halved in the
initial stage of the reaction (∼5 min), followed by a moderate
decline (5−135 min). The initial drop corresponds to the
change in the catalyst surface structure from PdIn to a Pd−
In2O3 composite. The latter decrease can be attributed to coke
deposition. In contrast, turnover frequency did not so vary
during the reaction (5−90 min). These results support that the
decrease in 1,3-butadiene yield is due to coke deposition.
On the basis of the insights obtained in this study, the unique
catalytic property of the Pd-based intermetallic compounds
observed in this study appears to correlate with the reducibility
of the second metal element. As mentioned above, the high
reducibility of In2O3 may contribute to the catalytic property.
In addition to the reducibility, the formation enthalpy (ΔHf) of
the intermetallic compound (i.e., the easiness of compound
formation) may be involved. In general, the more negative ΔHf,
the more negative the corresponding ΔGf, promoting
compound formation. For example, PdBi has a moderately
negative ΔHf (−27 kJmol−1),43 whereas PdIn has a
considerably negative ΔHf (−61 kJmol−1).44 Moreover, PdIn
possesses a certain solubility range, which would make the
transferring of In atoms less constrained. Thus, the
combination of these specific properties of PdIn may induce
the greatest catalytic performance. However, there is still room
to improve the selectivity of this catalytic system. Based on the
obtained insights, it is important that Pd atoms are highly
dispersed or isolated by In2O3 during the reaction. Therefore, it
is a promising way to reduce the Pd/In ratio below 1 so that Pd
atoms become less aggregated. Alternatively, modification of
the catalyst support may be an effective way to inhibit
aggregation of Pd and/or In2O3. Because SiO2 has a relatively
weak metal−support interaction,45 the loaded species tend to
aggregate easily. It is better to employ an appropriate support
having a substantially strong metal−support interaction, such as
γ-Al2O3,45 so that Pd and In2O3 remain well dispersed. In this
catalytic cycle, involvement of O2 is essential for the formation
4
8
2
follows (eqs 3 and 4):
rC H = kPC H 0.97PO 0.38 (Pd/SiO2)
(3)
(4)
4
6
4
8
2
rC H = kPC H 0.17PO
(PdIn/SiO2)
1.04
4
6
4
8
2
For Pd/SiO2, a first-order relationship with PC4H8 was observed,
whereas the dependence on PO2 was low. This result suggests
that adsorption of 1-butene on Pd surface is the rate-
determining step and the catalyst surface was dominantly
covered with O2. The excess amount of O2 on the catalyst
surface compared to 1-butene seems to easily result in
undesired combustion. This is also consistent with the results
observed in Figure 3. On the other hand, PdIn/SiO2 exhibited
opposite features (i.e., a first-order relationship with PO2 and
almost zero-order dependence on PC4H8). Hence, 1-butene
saturated the catalyst surface, and adsorption of O2 (or
oxidation of In to In2O3) was the rate-determining step for
PdIn/SiO2a situation favorable for the dehydrogenation
process, in preference to combustion.
Finally, the stability of the catalyst during the oxidative
dehydrogenation of 1-butene was tested. Figure 11 (red circles)
shows the time course of 1,3-butadiene yield during the
reaction. The product yield decreased slightly with time on
stream, almost plateauing around 135 min. The amount of coke
deposited on the catalyst, determined by TG-DTA analysis (see
Supporting Information Figure S4 for the results of TG-DTA),
was also plotted versus time on stream (Figure 10, black
diamonds). The amount of coke increased at the initial stage of
the reaction, becoming saturated around 135 min.42 A negative
linear correlation was observed between butadiene yield and
the amount of coke (Figure 10, blue triangles), suggesting a
diminished catalytic performance attributable to a decrease in
the number of active sites due to coke deposition. Although a
3540
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
(6) Soares, A. P. V.; Dimitrov, L. D.; de Oliveira, M. C. R. A.; Hilaire,
L.; Portela, M. F.; Grasselli, R. K. Appl. Catal., A 2003, 253, 191−200.
(7) Batist, Ph. A.; van de Moesdijk, C. G. M.; Matsuura, I.; Schuit, G.
C. A. J. Catal. 1971, 20, 40−57.
of ideal reaction sites. Therefore, a similar catalytic cycle may
work in some other reaction systems involving O2, such as
alcohol oxidation, oxidative steam reforming of methanol, and
ammonia-deNOx.
(8) Krishnan, V. V.; Suib, S. L. J. Catal. 1999, 184, 305−315.
(9) Liaw, B. J.; Cheng, D. S.; Yang, B. L. J. Catal. 1989, 118, 312−
326.
CONCLUSION
■
In this study, catalytic performances of Pd-based intermetallic
compounds (PdmMn/SiO2: M = Bi, Fe, Ge, In, Sn, Zn) in the
oxidative dehydrogenation of 1-butene and n-butane were
investigated. A remarkable increase in selectivity (and also
yield) to dehydrogenation products (1,3-butadiene and/or 1-
butene) was obtained when monometallic Pd was replaced by
PdIn, PdBi, or Pd3Fe. Undesired combustion into COx
proceeded over monometallic Pd, significantly decreasing
selectivity. The presence of the second metal (In, Bi, or Fe)
adjacent to Pd effectively inhibits undesired combustion, likely
by capturing O2 as a lattice oxygen. The incorporated lattice
oxygen reacts with hydrogen atoms derived from the
hydrocarbon by C−H activation over Pd sites, resulting in
the formation of water and oxygen vacancy and/or the parent
intermetallic phase. Thus, a combination of C−H activation by
Pd and redox of the second metal provides a unique and
effective dehydrogenation process, compared to monometallic
Pd or the second metal oxide alone. Under the reaction
conditions tested, the catalyst surface decomposes into Pd and
the second metal oxide (Pd−MOx nanocomposite), forming a
core−shell structure with a parent intermetallic core and a
composite shell as a steady state. The insights obtained in this
study will open the way to the development of a more effective
dehydrogenation system based on bifunctional catalysis.
(10) Herniman, H. J.; Pyke, D. R.; Reid, R. J. Catal. 1979, 58, 68−73.
(11) Gibson, M. A.; Hightower, J. W. J. Catal. 1976, 41, 420−430.
(12) She, Y.; Han, J.; Ma, Y. H. Catal. Today 2001, 67, 43−53.
(13) Fujimoto, K.; Kunugi, T. Ind. Eng. Chem. Prod. Rd. 1981, 20,
319−323.
(14) Onda, A.; Komatsu, T.; Yashima, T. J. Catal. 2004, 221, 378−
385.
(15) Furukawa, S.; Tamura, A.; Ozawa, K.; Komatsu, T. Appl. Catal.,
A 2014, 469, 300−305.
(16) Under the reaction atmosphere, any intermetallic species at the
catalyst surface will be oxidized into Pd metal and the second metal
oxide phases. Therefore, the presence of a small amount of the
impurity phase does not seem to strongly affect the reaction outcome.
(17) Okamoto, H. Desk Handbook: Phase Diagrams for Binary Alloys;
ASM International: Materials Park, OH, 2000.
(18) Bielz, T.; Lorenz, H.; Amann, P.; Klotzer, B.; Penner, S. J. Phys.
Chem. C 2011, 115, 6622−6628.
(19) Bielz, T.; Lorenz, H.; Jochum, W.; Kaindl, R.; Klauser, F.;
Klotzer, B.; Penner, S. J. Phys. Chem. C 2010, 114, 9022−9029.
(20) Bond, G. C.; Tripathi, J. B. P. J. Chem. Soc., Faraday Trans. 1
1976, 72, 933−941.
(21) Benedetti, A.; Fagherazzi, G.; Pinna, F.; Rampazzo, G.; Selva,
M.; Strukul, G. Catal. Lett. 1991, 10, 215−223.
(22) Iwasa, N.; Mayanagi, T.; Ogawa, N.; Sakata, K.; Takezawa, N.
Catal. Lett. 1998, 54, 119−123.
(23) Batista, J.; Pintar, A.; Mandrino, D.; Jenko, M.; Martin, V. Appl.
Catal., A 2001, 206, 113−124.
ASSOCIATED CONTENT
* Supporting Information
(24) Hong, C. T.; Yeh, C. T.; Yu, F. H. Appl. Catal. 1989, 48, 385−
396.
■
S
(25) Fagherazzi, G.; Benedetti, A.; Polizzi, S.; Dimario, A.; Pinna, F.;
Additional catalytic data, peak deconvolution of XRD pattern,
TG−DTA, amount of exposed Pd. This material is available
free of charge via the Internet at http://pubs.acs.org.
Signoretto, M. Catal. Lett. 1995, 32, 293−303.
(26) Ziemecki, S. B.; Michel, J. B.; Jones, G. A. React. Solid 1986, 2,
187−202.
(27) Kovnir, K.; Armbruster, M.; Teschner, D.; Venkov, T. V.;
Szentmiklosi, L.; Jentoft, F. C.; Knop-Gericke, A.; Grin, Y.; Schlogl, R.
Surf. Sci. 2009, 603, 1784−1792.
(28) Iwasa, N.; Ogawa, N.; Masuda, S.; Takezawa, N. Bull. Chem. Soc.
Jpn. 1998, 71, 1451−1455.
AUTHOR INFORMATION
Corresponding Author
*E-mail: komatsu.t.ad@m.titech.ac.jp. Fax: +81-3-5734-2758.
Tel.: +81-3-5734-3532.
■
(29) Pinna, F.; Selva, M.; Signoretto, M.; Strukul, G.; Boccuzzi, F.;
Benedetti, A.; Canton, P.; Fagherazzi, G. J. Catal. 1994, 150, 356−367.
(30) Dapaah, J. K. A.; Uemichi, Y.; Ayame, A.; Matsuhashi, H.;
Sugioka, M. Appl. Catal., A 1999, 187, 107−113.
(31) Kim, W. J.; Shin, E. W.; Kang, J. H.; Moon, S. H. Appl. Catal., A
2003, 251, 305−313.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by JSPS KAKENHI Grant No.
23360353. We thank Center for Advanced Materials Analysis
Tokyo Institute of Technology for the aid of TEM−EDX
analysis.
(32) Hirano, T.; Kazahaya, Y.; Nakamura, A.; Miyao, T.; Naito, S.
Catal. Lett. 2007, 117, 73−78.
(33) Huang, L.; Wang, Z.; Ang, T. P.; Tan, J.; Wong, P. K. Catal. Lett.
2006, 112, 219−225.
(34) Kovnir, K.; Armbruster, M.; Teschner, D.; Venkov, T. V.;
Jentoft, F. C.; Knop-Gericke, A.; Grin, Y.; Schlogl, R. Sci. Technol. Adv.
Mater. 2007, 8, 420−427.
(35) Armbruster, M.; Behrens, M.; Fottinger, K.; Friedrich, M.;
Gaudry, E.; Matam, S. K.; Sharma, H. R. Catal. Rev. 2013, 55, 289−
367.
(36) Iwasa, N.; Takezawa, N. Top. Catal. 2003, 22, 215−224.
(37) Friedrich, M.; Teschner, D.; Knop-Gericke, A.; Armbruster, M.
J. Catal. 2012, 285, 41−47.
(38) Marchesini, F. A.; Irusta, S.; Querini, C.; Miro, E. Appl. Catal., A
2008, 348, 60−70.
(39) Skala, T.; Veltruska, K.; Moroseac, M.; Matolinova, I.;
Korotchenkov, G.; Matolin, V. Appl. Surf. Sci. 2003, 205, 196−205.
REFERENCES
■
(1) Passmann, W. Ind. Eng. Chem. 1970, 62, 48−51.
(2) Dumez, F. J.; Froment, G. F. Ind. Eng. Chem. Proc. Dd 1976, 15,
291−301.
(3) Hakonsen, S. F.; Holmen, A. In Handbook of Heterogeneous
̊
Catalysis; Ertl, G., Knozinger, H., Schuth, F., Weitkamp, J., Eds.; Wiley-
VCH Verlag GmbH & Co. KGaA: Weinheim, 2008; Chapter 14; pp
3384−3386.
(4) Park, J. H.; Noh, H.; Park, J. W.; Row, K. H.; Jung, K. D.; Shin, C.
H. Res. Chem. Intermed. 2011, 37, 1125−1134.
(5) Jung, J. C.; Lee, H.; Kim, H.; Chung, Y. M.; Kim, T. J.; Lee, S. J.;
Oh, S. H.; Kim, Y. S.; Song, I. K. J. Mol. Catal. A: Chem. 2007, 271,
261−265.
3541
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542

ACS Catalysis
Research Article
(40) Lorenz, H.; Turner, S.; Lebedev, O. I.; Van Tendeloo, G.;
Klotzer, B.; Rameshan, C.; Pfaller, K.; Penner, S. Appl. Catal., A 2010,
374, 180−188.
(41) Sampling for the kinetic study was performed at 5 min of TOS,
in which time activity already reached a steady state as mentioned
above. Therefore, the initial change in the catalytic structure was
excluded from the kinetic study.
(42) As the coke amount (e.g., 0.35 mmol at 135 min) corresponds
to only 0.1% of the integrated reactant flow (330 mmol for 135 min),
its contribution to the carbon balance can be ignored.
(43) Vrest’al, J.; Pinkas, J.; Watson, A.; Scott, A.; Houserova, J.;
Kroupa, A. CALPHAD: Comput. Coupling Phase Diagrams Thermo-
chem. 2006, 30, 14−17.
(44) Jordan, J. L.; Deevi, S. C. Intermetallics 2003, 11, 507−528.
(45) Furukawa, S.; Ozawa, K.; Komatsu, T. RSC Adv. 2013, 3,
23269−23277.
3542
dx.doi.org/10.1021/cs500920p | ACS Catal. 2014, 4, 3533−3542