Pulse and IR study on the reaction pathways for the conversion of ethanol to propene over scandium-loaded indium oxide catalysts

Article abstract of DOI:10.1021/cs5006822

Potential reaction intermediates in the conversion of ethanol to propene, acetaldehyde, ethyl acetate, crotonaldehyde, acetic acid, acetone, and 2-propanol were introduced as pulses onto a scandium-loaded indium oxide catalyst. The product distributions were primarily measured as a function of the space velocity in the absence or presence of hydrogen and water. The FT-IR spectra of the surface adsorbates were also collected after ethanol adsorption and indicated the formation of ethoxide species, which were converted to acetate species over the catalyst. The proposed reaction route involved the dehydrogenation of ethanol to acetaldehyde, direct oxidation of acetaldehyde with water or a surface hydroxyl group to yield acetic acid, ketonization of acetic acid to acetone and carbon dioxide, and hydrogenation and subsequent dehydration of acetone to propene. The total reaction can be described as 2 CH₃CH₂OH → CH₂ - CHCH₃ + CO₂ + 3 H₂. A side reaction involving isobutene formation also occurred via the acetone intermediate.

Full text of DOI:10.1021/cs5006822

Research Article
pubs.acs.org/acscatalysis
Pulse and IR Study on the Reaction Pathways for the Conversion of
Ethanol to Propene over Scandium-Loaded Indium Oxide Catalysts
Masakazu Iwamoto,*,† Masashi Tanaka,† Shota Hirakawa,‡ Shota Mizuno,‡ and Mika Kurosawa‡
†Research and Development Initiative, Chuo University, 1-13-27 Kasuga, Bunkyo-ku, Tokyo 112-8551, Japan
‡Chemical Resources Laboratory, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8503, Japan
S
* Supporting Information
ABSTRACT: Potential reaction intermediates in the conversion
of ethanol to propene, acetaldehyde, ethyl acetate, crotonalde-
hyde, acetic acid, acetone, and 2-propanol were introduced as
pulses onto a scandium-loaded indium oxide catalyst. The
product distributions were primarily measured as a function of
the space velocity in the absence or presence of hydrogen and
water. The FT-IR spectra of the surface adsorbates were also
collected after ethanol adsorption and indicated the formation of
ethoxide species, which were converted to acetate species over
the catalyst. The proposed reaction route involved the
dehydrogenation of ethanol to acetaldehyde, direct oxidation of
acetaldehyde with water or a surface hydroxyl group to yield
acetic acid, ketonization of acetic acid to acetone and carbon
dioxide, and hydrogenation and subsequent dehydration of acetone to propene. The total reaction can be described as 2
CH3CH2OH → CH2CHCH3 + CO2 + 3 H2. A side reaction involving isobutene formation also occurred via the acetone
intermediate.
KEYWORDS: ethanol, propene, indium oxide, scandium, acetate, ketonization
INTRODUCTION
■
The catalytic conversion of ethanol to propene1−12 has very
recently been developed to use bioethanol as a starting material
for chemicals and reduce the carbon dioxide emission from
naphtha crackers. Among the active catalysts that have been
developed, Ni ion-loaded mesoporous silica MCM-41 (Ni-
M41),8−10 yttrium loaded cerium oxides (Y/CeO2),11 and
scandium-modified indium oxides (Sc/In2O3),12 Sc/In2O3
results in good propene yields and long durability. Therefore,
the elucidation of the reaction pathways on the catalyst would
be useful for understanding the catalytic process and improving
the activity.
The possible reaction pathways on the oxide catalysts for the
conversion of ethanol to propene are summarized in Figure 1,
where the notations of the compounds indicate that they might
be generated as intermediate products or surface intermediates.
The reaction pathways over zeolite catalysts3−5 were not
included in this figure. Oligomerization, aromatization, and
fission reactions on strong acid sites in zeolite pores are widely
known to result in the formation of ethene, propene, and
butenes due to shape selectivity. However, these uncontrolled
reactions that occur in the pores ultimately result in coke
formation and short catalyst life times. Therefore, the reaction
pathways over zeolites have been omitted here. There are two
ethanol reaction routes including dehydration to yield ethene
(Pathway 1) and dehydrogenation to afford acetaldehyde
(Pathway 6), that can occur over oxide catalysts.6−12 Our
Figure 1. Proposed reaction pathways for the conversion of ethanol to
propene over oxide catalysts.
recent paper on Ni-M419 revealed that ethene produced
through Pathways 1 and 2 or Pathways 6, 7, 8, and 18 was an
intermediate in the production of propene. Ethene was
dimerized to form 1-butene, which was isomerized to 2-
butenes. The resulting 2-butenes could react with another
ethene to produce propene via metathesis. The progress of the
metathesis reaction on Ni-M41 was confirmed in separate
Received: July 15, 2014
Revised: August 24, 2014
Published: August 25, 2014
© 2014 American Chemical Society
3463
dx.doi.org/10.1021/cs5006822 | ACS Catal. 2014, 4, 3463−3469

ACS Catalysis
Research Article
experiments.13 The acetic acid that is coproduced through
Pathways 6, 7, and 8 would react with ethanol or hydrogen to
form ethyl acetate (eq 1), acetaldehyde (eq 2), or ethanol (eq
3), as reported over several catalysts.6,7 Equation 1 is well-
known as the Fisher esterification. These equations suggested
the circulation of acetic acid on the Ni-M41 catalyst.
atomic % to that of In, whose amount was determined by an
induced coupled plasma analysis after the samples were
dissolved in HF solutions. The catalysts were calcined at
1073 K for 5 h, and the respective grain sizes were adjusted to
0.3−0.6 mm for use in catalytic runs. The catalyst was
designated as Sc3/In2O3, where “3” was the atomic % of loaded
scandium.
Ethanol and the other organic reagents were purchased
commercially from the Kanto Chemical Co. Japan, and used
without further purification.
CH3COOH + CH3CH2OH → CH3COOCH2CH3 + H2O
CH3COOH + H2 → CH3CHO + H2O
(1)
(2)
(3)
CH3COOH + 2H2 → CH3CH2OH + H2O
Apparatus and General Procedure for the Pulse
Experiments. As reported in the previous paper,12 the
coexistence of water or hydrogen in the continuous flow
experiments substantially improved the selectivity to propene
and the durability of the Sc/In2O3 catalysts. The cointroduction
of water or hydrogen with ethanol or other substrates as one
pulse is one method for determining the addition effect. For
example, in these experiments, the mixture of acetaldehyde and
water is introduced onto the catalyst dried upon heating in the
carrier gas flow. Therefore, it is possible that the product
distribution is different from that on a wet catalyst surface in
typical continuous flow systems. Similarly, the presence or
absence of hydrogen atoms preadsorbed on the catalyst might
greatly influence the dehydrogenation and hydrogenation
reactions. On the basis of these considerations, we employed
mixed flows of H2O/N2 or H2/N2 as the carrier gas in the pulse
experiments. However, the carrier gas in the reactor was also
employed as the carrier in the gas chromatograph system in the
typical pulse experiment apparatus. Therefore, this mixture
cannot be employed as the carrier gas. The carrier gas for the
gas chromatography system was segregated from the carrier gas
for the catalytic reaction in the current experimental apparatus
for the pulse experiments. A thermal conductivity detector
(TCD) was installed after the catalyst bed, and a sampling loop
consisting of a six-way valve was also attached after the TCD. In
the pulse experiments, the flow of products after substrate
injection was detected by the TCD, and the sampling loop
automatically picked up the portion of products according to
the TCD signal and sent it to the gas chromatography system.
Pulse experiments using various substrates were performed in
a quartz microreactor with an inner diameter of 7.6 mm at
atmospheric pressure. Prior to each reaction, 0.1−2.0 g of the
catalyst mounted in the reactor was treated at 673 K in N2 (50
mL/min) for 60 min to remove any residual gases in the system
followed by exposure to the pulses (a 2 μL aliquot of a liquid
sample). The products were analyzed with an online automatic
gas chromatograph (GC7100, J-Science, Japan) equipped with
four packed columns and a capillary column as follows:
activated charcoal, MS 5A, Porapak Q, and PEG-20M, and GS-
Q (30 m, Agilent, U.S.A.). The change in the catalytic activity
with the reaction time (durability) was analyzed by the
occasional injection of ethanol and the consistency of the
product distributions within the experimental errors. The
conversion and selectivity were calculated on the basis of the
carbon content in the products assuming a 100% carbon
balance.
On the Y/CeO2 catalyst, the simultaneous production of
ethene and propene was observed in the presence/absence of
water. The reactivity of ethyl acetate was studied using pulse
experiments, which indicated its decomposition through
Pathway 8.11 The produced ethene was observed without any
further reactions. However, acetic acid was converted to
acetone through ketonization (Pathways 9 and 10). Acetone
was converted to propene by cofed ethanol (Pathway 11),
which involved the reduction of acetone by ethanol (eq 4) and
the subsequent dehydration of 2-propanol to propene (eq 6).
Equation 4 is known as the Meerwein−Ponndorf−Verley
(MPV) reduction.14
CH3C(O)CH3 + CH3CH2OH → CH3CH(OH)CH3 + CH3CHO
(4)
CH3C(O)CH3 + H2 → CH3CH(OH)CH3
CH3CH(OH)CH3 → CH2 = CHCH3 + H2O
(5)
(6)
The reaction pathways 1−11 in Figure 1, in which Pathway
11 on Y/CeO2 was the MPV reduction, were already suggested
on Ni-M41 or Y/CeO2. In contrast to the reactions on these
catalysts, the amount of ethene produced on the Sc/In2O3
catalysts was very small, and isobutene was one of the major
products.12 In addition, the production ratios of acetone/
propene/isobutene, as well as the durability of the catalyst, were
dependent on the partial pressures of water and hydrogen.
These findings indicated that the reaction pathways on Sc/
In2O3 are different from those over Ni-M41 and Y/CeO2 and
should be studied separately.
Initially, we measured the contact time dependence of the
product distribution in a continuous flow system but did not
obtain significant conclusions due to complicated product
distributions. Therefore, a pulse technique was applied to
determine the reactivity of several potential intermediate
compounds and determine the reaction pathways. As
summarized in the Experimental Section, we constructed a
new apparatus for the pulse experiments in which various
substrates could be introduced into the carrier gas flows
containing water or hydrogen because the presence of water
and hydrogen in the reactant flows was previously reported to
substantially improve the propene yield and catalyst durabil-
ity.12 In addition, surface adsorbates on the Sc/In2O3 catalyst
were investigated using FT-IR measurements to confirm the
reaction pathways.
EXPERIMENTAL SECTION
IR Measurements of Surface Adsorbates. The FT-IR
spectra were recorded at room temperature using a JASCO
FTIR-6300 spectrophotometer with a MCT detector (accu-
mulation 32 scans, resolution 4 cm−1). The powdered sample
was pressed into a self-supporting disk with a diameter of 10
mm and placed in a quartz cell with KRS-5 windows. The cell
was capable of in situ treatment and gas introduction. The
■
Materials. Indium oxide, In2O3 (4.2 m2g−1, BET surface
area after calcination at 1073 K for 5 h), was commercially
obtained from the Kanto Chemical Co. Japan. Scandium-loaded
In2O3 samples were prepared by a conventional impregnation
method using acetate salt. The detailed preparation method is
described in a previous paper.12 The scandium loading was 3
3464
dx.doi.org/10.1021/cs5006822 | ACS Catal. 2014, 4, 3463−3469

ACS Catalysis
Research Article
adsorption measurement of ethanol or acetic acid was
performed by introducing the gas at room temperature to the
Sc3/In2O3 sample pre-evacuated at 673 K.
determined at 773 K as a function of the gas hourly space
velocity (GHSV) and are summarized in Figure 3a. Unreacted
RESULTS AND DISCUSSION
■
Confirmation of Overall Reaction Yielding Propene.
The conversion of ethanol to propene on Sc3/In2O3 has been
proposed to proceed via an acetone intermediate. The overall
reactions for the formation of propene can be described as
shown in eq 7. The acetone formation is described by eq
8.12,15,16
2CH3CH2OH → CH2 = CHCH3 + CO2 + 3H2
2CH3CH2OH + H2O → CH3C(O)CH3 + CO2 + 4H2
3CH3CH2OH + H2O → CH2 = C(CH3)2 + 2CO2 + 6H2
CO2 + H2 → CO + H2O
(7)
(8)
(9)
(10)
Figure 3. Product distributions upon introduction of ethanol (a) and
acetaldehyde (b) pulses at 773 K in a N2 flow. The distributions were
measured as a function of GHSV. The symbols are the same as those
shown in Figure 1.
A
C3/3 + AC4/2 = ACO
x
(11)
The molar ratio of the total amounts of propene and acetone,
AC3, with those of carbon dioxide, ACO , were 3:1 based on eqs
2
ethanol was detected at 500 000 h−1 but it disappeared at
approximately 80 000 h−1. The first observed product was
acetaldehyde whose amount was maximized at 300 000 h−1 and
decreased at lower SVs. Acetone became the main product at
50 000 h−1, which was accompanied by the coproduction of
CO2. At 5000 h−1, the amounts of acetone and CO2 decreased
and those of propene, isobutene, and CO increased. The
reactivity of acetaldehyde was studied separately and is
summarized in Figure 3b. The change in the product
distributions with SV was similar to that of ethanol even
though the amount of propene produced was smaller than that
in the reaction of ethanol (Figure 3a) due to the smaller
amount of hydrogen, as mentioned below. These results
indicated that acetaldehyde was the true intermediate in the
conversion of ethanol to propene. All of the results clearly
indicated the reaction sequence of ethanol → acetaldehyde →
acetone (and CO2) → propene and isobutene, which supports
the proposed schemes shown in Figure 1.
7 and 8. In the calculation, the ACO values should be replaced
2
with the total amounts of carbon dioxide and carbon monoxide,
A
CO , because the progress of eq 10 was confirmed on the
x
catalyst. In addition, because the isobutene formation was
accompanied by CO2 formation (eq 9), the AC3 values should
be corrected for the quantities of isobutene, AC4. eq 11 needs to
be confirmed.
The product distributions in the flow system were employed
to confirm the validity of the above equations. Typical results
are shown in Figure 2 where the catalysis was investigated
Next, the reaction sequences for the formation of C2x−1
ketones from Cx-aldehydes on the Sc/In2O3 were investigated.
Three types of pathways have been proposed for the oxide-
based catalysts including the (i) aldol reaction,15 (ii)
Tishchenko reaction,15 and (iii) ketonization reaction of
carboxylic acids,17 as shown in Figure 1. The aldol reaction
of acetaldehyde yields 3-hydroxybutylaldehyde, which can be
converted to CH3C(O)CH2CHO or CH3CHCHCHO by
dehydrogenation (Pathway 14) or dehydration (16). Although
the latter cannot produce acetone, the reactivity was studied. As
shown in Figure 4a, crotonaldehyde was easily converted to
propene with the release of CO (and CO2), but no acetone was
produced. The results indicated that crotonaldehyde was not
involved in the reaction pathways over the Sc/In2O3. β-
Ketoaldehyde has the potential to yield acetone through the
release of CO (Pathway 15) or by reaction with water (yielding
acetone, CO2, and H2). Because unstable β-hydroxyaldehyde
and β-ketoaldehyde cannot be used as substrates in the current
experiments, the products produced in Figure 2 were carefully
analyzed using an off-line capillary GC system. However, no
production of these compounds was observed, which suggested
that the aldol reaction route was unlikely. The formation of
ethyl acetate, acetic acid, and 2-propanol was confirmed in the
current experiments. Therefore, the possibilities of the ethyl
Figure 2. Catalytic activity of Sc3/In2O3 (the upper row) and a
comparison of AC3/3 + AC4/2 with ACO (the lower) as a function of
x
the reaction time and temperature. Reaction conditions: catalyst
weight, 2.0 g; total flow rate, 12.8 mL min−1 (GHSV 687 h−1); 1 atm;
PEtOH, 30 vol %; PH O, 8.5 vol %; PH , 30 vol %; N2 balance.
2
2
under various reaction temperatures and times over Sc3/
In2O3.12 The upper column of the figure shows the amounts of
ethanol reacted and the distributions of hydrocarbons
produced, whereas the lower one summarizes the amounts of
CO2 + CO and the values of AC3/3 + AC4/2. Equation 11 was
confirmed by the results in the lower column irrespective of the
experimental conditions, which supports the occurrence of the
total reactions shown in eqs 7−9.
Reaction Pathways of Ethanol and Acetaldehyde. The
product distributions in the pulse experiments of ethanol were
3465
dx.doi.org/10.1021/cs5006822 | ACS Catal. 2014, 4, 3463−3469

ACS Catalysis
Research Article
the resulting acetone was converted to propene and isobutene
in the presence and absence of hydrogen, respectively, with
increasing contact time. The reaction of acetone will be
discussed in more detail below, and the acetic acid formation
pathway will be discussed here.
Acetic acid is known to be generated from acetaldehyde
under oxidative conditions, but this reaction was not observed
here due to the absence of oxygen.18 However, the direct
reaction of an adsorbate (CH3CH2O) with a surface OH group
has been proposed to generate the CH3COO species on the
oxide surface,4,18 which might lead to the formation of acetic
acid through one of several possible pathways such as Pathway
12. This possibility was examined using the pulse technique,
and the results are summarized in Figure 6 as a function of
Figure 4. Product distributions upon introduction of crotonaldehyde
(a) and ethyl acetate (b, c) pulses at 773 K in a N2 (a, b) or H2 (c)
flow. The distributions were measured as a function of GHSV. The
symbols are the same as those shown in Figure 1.
acetate route15,16 and the direct path forming acetic acid17 were
investigated.
Figure 4b,c show the results from the pulse experiments of
ethyl acetate as a function of the space velocity. In the absence
of hydrogen (Figure 4b), the major product was acetone at high
SVs and ethene at low SVs, and the yields of propene were
lower than those in the flow system. In the presence of
hydrogen (Figure 4c), the product distributions (acetone, CO2,
and ethene) were essentially similar to those without hydrogen
(Figure 4b), though the propene and isobutene selectivities
were substantially changed with/without hydrogen and almost
replaced each other. The selective formation of ethene
indicated the progress of Pathway 8, which was very different
from the product distributions in the flow system shown in
Figure 2. Therefore, ethyl acetate was not an intermediate in
propene formation on the Sc/In2O3 catalysts. The small
amount of ethyl acetate that was observed in the flow system as
described above is a byproduct from acetic acid and ethanol due
to the Fisher esterification (eq 1).
Acetic Acid as an Intermediate. The introduction of
acetic acid in the absence or presence of hydrogen resulted in
the products shown in Figure 5a,b. The amount of unreacted
Figure 6. Product distributions upon introduction of acetaldehyde
pulses at GHSV of 36,311 h−1 in a N2 flow (a, b) or a mixed flow of N2
and H2O (PH O = 10 vol %) (c, d). The distributions were measured as
2
a function of reaction temperature. Panels (b) and (d) provide
enlarged views of panels (a) and (c). The symbols are the same as
those shown in Figure 1.
reaction temperature because the very high reactivity of acetic
acid at 773 K might hide the appearance of acetic acid as an
intermediate (product). A small amount of acetic acid was
produced at 623 and 673 K in the absence of water, and its
amount increased in the presence of water. In addition, the
conversion levels of acetaldehyde also increased at these
temperatures due to the addition of water to the reaction
system. These observations indicated the true progress of the
reaction of acetaldehyde with water to form acetic acid
(Pathway 12), in which the water addition increased the
reaction rate of acetaldehyde. The high reactivity of acetic acid
would result in its minor appearance as the intermediary
product in the experiments (Figures 2 and 3).
Figure 5. Product distributions upon introduction of acetic acid pulses
at 773 K in a N2 (a) or H2 (b) flow. The distributions were measured
as a function of GHSV. The symbols are the same as those shown in
Figure 1.
acetic acid was very small even at high SVs, which indicated a
very high reactivity of acetic acid over the catalyst. Acetic acid
was selectively converted to acetone and CO2 with a short
contact time irrespective of the presence and absence of
hydrogen. The ratio of acetone to CO2 was approximately 3:1
(carbon basis) except for the ratio obtained at long contact
times. The results confirmed the very easy and selective
ketonization of acetic acid (Pathways 9 and 10). In addition,
The surface intermediates were directly observed on the Sc3/
In2O3 catalyst using FT-IR measurements. Typical results are
shown in Figure 7 as a function of sample temperature. The
3466
dx.doi.org/10.1021/cs5006822 | ACS Catal. 2014, 4, 3463−3469

ACS Catalysis
Research Article
very rapid conversion of 2-propanol to propene or the low
desorption rate of the 2-propanol intermediate. In contrast, in
the absence of hydrogen (Figure 8a), isobutene was the major
product along with a small amount of propene byproduct. This
result indicated the progress of the selective conversion of
acetone to isobutene, which has been previously reported for
the ZnxZryOz catalyst.16 Similar results were observed in Figure
5a,b where acetic acid was used as the substrate. The
conversion of acetone to isobutene can be formally expressed
by eq 12 even though the reaction mechanism is unknown and
the acetic acid byproduct cannot be confirmed due to its very
rapid conversion. The total reaction from ethanol to isobutene
is expressed in eq 9.
Figure 7. FT-IR spectra of the surface species adsorbed on the Sc3/
In2O3 catalyst. Ethanol was adsorbed at room temperature on the self-
standing wafer of the catalyst after the pretreatment mentioned in the
text, and then, the sample temperature was gradually increased under a
dynamic vacuum. The vibrations of the C−H, O−C−O, and C−O
bonds are depicted in the panels (a), (b), and (c).
2CH3C(O)CH3 → CH2 = C(CH3)2 + CH3C(O)OH
(12)
The reactivity of 2-propanol is summarized in Figures 9a,b.
Even at the highest space velocity studied, the remaining 2-
ethanol adsorption induced the appearance of surface ethoxide
species19,20 at 2967 cm−1 (assignable to νas(CH3)), 2930
(νas(CH2)), 2899 (νs(CH3), 2871 (νs(CH2)), 1447 (δas(CH3)),
and 1385 (δs(CH3)), 1104 (ν(CO)mono), 1061 (ν(CO)bi), and
884 (ν(CCO)). This species was converted to acetate
species21−24 upon heating the sample to 473 K or higher,
which was characterized by the absorptions at 2935 cm−1
(νs(CH3)), 1568 (νas(OCO)), and 1443 (νs(OCO)). The
results indicated the surface reaction involved the conversion of
ethanol to acetic acid, which is consistent with the results from
the pulse experiments. All of the results in Figures 5−7
indicated that Pathway 12 was the primary route for the
conversion of ethanol to acetic acid (then to acetone) on the
Sc/In2O3 catalyst.
Conversion of Acetone to Propene. Finally, the
conversion of acetone to propene was investigated. As reported
previously, on the Y/CeO2 catalyst, acetone was reduced to 2-
propanol by unreacted ethanol (i.e., the MPV reduction) and
not by gaseous hydrogen. The reaction of acetone on the
current Sc/In2O3 catalyst was studied, and the results are
summarized in Figure 8a,b. In the presence of hydrogen (Figure
Figure 9. Product distributions upon introduction of 2-propanol
pulses at 773 K in a N2 (a) or H2 (b) flow. The distributions were
measured as a function of GHSV. The symbols are the same as those
shown in Figure 1.
propanol was barely observed irrespective of the presence/
absence of hydrogen, which indicated its very high reactivity.
The characteristic product in both experiments was acetone.
The major product was acetone at a high space velocity, and the
amount decreased as the SV decreased. Therefore, the reverse
reaction of eq 5 (eq 13), which has a large rate constant,
occurred on Sc/In2O3. The produced acetone was hydro-
genated to 2-propanol in the presence of hydrogen and then
dehydrated to propene (Figure 9b) or converted to isobutene
in the absence of hydrogen (Figure 9a). The results indicated
that the reactions between acetone + hydrogen and 2-propanol
were reversible and the equilibrium was biased to the acetone
side.
CH3CH(OH)CH3 → CH3C(O)CH3 + H2
(13)
On the basis of the above results and discussion, the
following reaction pathways yielding propene from ethanol was
proposed for the Sc-modified In2O3 catalyst. Ethanol was
dehydrogenated to acetaldehyde, and the resulting acetaldehyde
was converted to acetic acid via oxidation with water or surface
hydroxyl groups. The acetic acid yielded acetone and carbon
dioxide via the ketonization reaction. Propene was produced by
the hydrogenation of acetone with hydrogen followed by
dehydration. It would be worthy to discuss which reaction is
the rate-determining step in the consecutive reactions on Sc/
In2O3. As shown in Figure 3a, the observed intermediates in the
pulse experiments of ethanol were acetaldehyde and acetone,
Figure 8. Product distributions upon introduction of acetone pulses at
773 K in a N2 (a) or H2 (b) flow. The distributions were measured as a
function of GHSV. The symbols are the same as those shown in Figure
1.
8b), acetone was selectively converted to propene along with a
small amount of isobutene byproduct. The maximum yield of
propene reached 94% at low SV, which indicated that the
reaction of acetone with hydrogen was facile, in contrast to the
reaction on Y/CeO2, and yielded propene. The production of
2-propanol was not significant, which was most likely due to
3467
dx.doi.org/10.1021/cs5006822 | ACS Catal. 2014, 4, 3463−3469

ACS Catalysis
Research Article
indicating that the respective subsequent reactions of
acetaldehyde and acetone would be slow. Thus, the reaction
rates of these compounds were compared. The amounts of
unreacted ethanol at GHSV 3 × 105 h−1 was 18% (Figure 3a)
and that of unreacted acetaldehyde at the same conditions was
59% (Figure 3b). In contrast, that of unreacted acetone at
GHSV 1 × 105 h−1 was 69% (Figure 9b), although the GHSV
value applied was one-third of those for ethanol and
acetaldehyde, indicating the low reactivity of acetone. It follows
that the slowest step of the current reaction on Sc/In2O3 would
be the hydrogenation of acetone to form 2-propanol.
Correlation of Property of Surface Acetate Species
with the Ketonization Reaction. Surface acetate species,
which are the key intermediate for the production of acetone,
have been observed on various oxide catalysts. The product
yields resulting from the ketonization reaction are dependent
on the oxides. Therefore, the correlation of properties of the
surface acetate species with the ketonization activity was
investigated, and the results revealed the following interesting
relationship. At first, the angle of the O−C−O bond in the
acetate species produced by the adsorption of acetic acid on the
current Sc3/In2O3 catalyst was compared to that observed
during the ethanol adsorption (Figure 7). The bond angles,
which were calculated by eq 1425 using the peak areas of the
1568 and 1444 cm−1 IR bands (abbreviated Aasym and Asym),
were 87.0 and 88.8 deg, respectively. The good agreement
between these values indicated that the acetate species
produced in the reaction of ethanol adsorbed on the same
active site with the same configuration as that of the acetate
species originating from the acetic acid adsorption. The results
indicated that the values measured upon adsorption of acetic
acid can be employed as an index of the surface adsorbates in
the ethanol reaction.
Figure 10. Dependence of catalytic activity on the bond angle of
acetate species for the production of acetone and propene.
considerably smaller than the angle of free acetic acid, which is
126 degrees. This angle might be related to a M2O(CH3COO)4
(M: an active metal ion) intermediate,26 which may be present
on the oxide surfaces for the ketonization of acetic acid. The
detailed study of the volcano-shaped dependence is currently
under way.
CONCLUSIONS
■
The reaction mechanism for the conversion of ethanol to
propene on Sc/In2O3 was studied using a pulse reaction
technique. Acetaldehyde was produced via dehydrogenation of
ethanol and directly converted to acetic acid with surface
hydroxyl groups, the latter of which was newly confirmed in the
current study. Acetic acid was transformed to acetone and
carbon dioxide via ketonization. The acetone was reduced to
propene with hydrogen via 2-propanol or converted to
isobutene without hydrogen. The hydrogenation of acetone
to 2-propanol was the slowest step among the consecutive
reactions on the Sc/In2O3 catalyst. The great difference among
the reaction pathways over Ni-M41, Y/CeO2, and Sc/In2O3
may be due to ability of the catalysts to activate the water and
hydrogen molecules. It should be noted that the direct
conversion of acetaldehyde to acetic acid on Sc/In2O3 did
not produce ethene as byproduct and resulted in better
selectivity to propene than that over Y/CeO2, because ethyl
acetate produced from acetaldehyde on the latter catalyst was
an intermediate to yield ethene and acetic acid. In addition, the
bond angle of the acetate species adsorbed on the catalyst
surface was suggested to be an important factor to determine
the reactivity for the ketonization.
0.5
2θ = 2 arctan (Aasym/Asym
)
(14)
The catalytic activity of various oxides for the ketonization of
ethanol was examined at 623−823 K in separate experiments
without the cofeed of water and hydrogen for comparison
under the same conditions. The results are summarized in
Supporting Information, Figure S1 as a function of the reaction
temperature and the reaction time. It should be noted that the
tested oxides possessed neutral or weak basic−weak acidic
surface sites because oxides with strong basic sites or strong
acidic sites would not be suitable for the complicated catalysis
described above. The respective active temperature ranges were
dependent on the oxides, and their catalytic activity increased,
decreased, or was stable as the reaction time increased.
Therefore, the mean catalytic activity after 2.5 h at the reaction
temperature, which affords the maximum activity, was
employed as the index for the catalysis. The O−C−O bond
angles adsorbed on Sc/In2O3 and ZnO were measured here,
and the values on SnO2, TiO2, CeO2, and MgO were obtained
from the literature,21−24 because the oxide wafers were very
difficult to prepare for the current IR experiments. The catalytic
activity is plotted as a function of the O−C−O bond angles of
adsorbed acetate species in Figure 10 based on Supporting
Information, Table S1. The vertical axis is the total amount of
acetone and propene produced based on the assumption that
the reaction mechanism was the same as the one on the oxide
catalysts examined here. Surprisingly, a typical volcano-shaped
dependence was observed. The best O−C−O angle for the
ketonization reaction was approximately 90 deg, which was
ASSOCIATED CONTENT
* Supporting Information
■
S
The ethanol conversions over six oxide catalysts and the results
of IR measurements are summarized in Table S1 and Figure S1.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: iwamotom@tamacc.chuo-u.ac.jp. Tel.: +81-3-3817-
1612.
■
Notes
The authors declare no competing financial interest.
3468
dx.doi.org/10.1021/cs5006822 | ACS Catal. 2014, 4, 3463−3469

ACS Catalysis
Research Article
(15) (a) Nakajima, T.; Yamaguchi, T.; Tanabe, K. J. Chem. Soc. Chem.
Commun. 1987, 394−395. (b) Nakajima, T.; Tanabe, K.; Yamaguchi,
T.; Matsuzaki, I.; Mishima, S. Appl. Catal. 1989, 52, 237−238.
(c) Sreerama-Murthy, R.; Patnaik, P.; Sidheswaran, P.; Jayamani, M. J.
Catal. 1988, 109, 298−302. (d) Kamimura, Y.; Sato, S.; Takahashi, R.;
Sodesawa, T.; Akashi, T. Appl. Catal., A 2003, 252, 399−410.
(e) Kamimura, Y.; Sato, S.; Takahashi, R.; Sodesawa, T.; Fukui, M.
Chem. Lett. 2000, 232−233. (f) Tsuchida, T.; Kubo, J.; Yoshioka, T.;
Sakuma, S.; Takeguchi, T.; Ueda, W. J. Catal. 2008, 259, 183−189.
(16) Sun, J.; Zhu, K.; Gao, F.; Wang, C.; Liu, J.; Peden, C. H. F.;
Wang, Y. J. Am. Soc. Chem. 2011, 133, 11096−11099.
(17) Renz, M. Eur. J. Org. Chem. 2005, 979−988.
ACKNOWLEDGMENTS
■
We wish to acknowledge Grants-in-Aid from the Advanced
Low Carbon Technology Program (JST), the Japan Society for
the Promotion of Science (JSPS), and the New Energy and
Industrial Technology Development Organization (NEDO) of
Japan. The authors wish to thank Dr. Tetsuo Suzuki and Mrs.
Osamu Takahashi, Hiroshi Ohashi, and Takahiro Kakinuma of
the NEDO research group for helpful discussions.
REFERENCES
■
(18) Bi, J.; Liu, M.; Song, C.; Wang, X.; Guo, X. Appl. Catal., B 2011,
107, 68−76.
(1) Morschbaker, A. Polym. Rev. 2009, 49, 79−84.
(2) For example, Olah, G. A.; Molnar, A. Hydrocarbon Chemistry, 2nd
ed.; John Wiley & Sons, Inc.: Hoboken, NJ, 2003; pp 30−84.
(3) (a) Phillips, C. B.; Datta, R. Ind. Eng. Chem. Res. 1997, 36, 4466−
4475. (b) Brandao, P.; Philippou, A.; Rocha, J.; Anderson, M. W.
Catal. Lett. 2002, 80, 99−102. (c) Aguayo, T.; Gayubo, A. G.; Atutxa,
A.; Olazar, M.; Bilbao, J. Ind. Eng. Chem. Res. 2002, 41, 4216−4224; J.
Chem. Technol. Biotechnol. 2002, 77, 211−216. (d) Gayubo, A. G.;
Alonso, A.; Valle, B.; Aguayo, A. T.; Bilbao, J. Ind. Eng. Chem. Res.
2010, 49, 10836−10844. (e) Gayubo, A. G.; Alonso, A.; Valle, B.;
Aguayo, A. T.; Olazar, M.; Bilbao, J. Chem. Eng. J. 2011, 167, 262−277.
(4) (a) Takahara, I.; Saito, M.; Inaba, M.; Murata, K. Catal. Lett.
2005, 105, 249−252; 2007, 113, 82−85. (b) Inaba, M.; Murata, K.;
Saito, M.; Takahara, I. Green Chem. 2007, 9, 638−646. (c) Song, Z.;
Takahashi, A.; Mimura, N.; Fujitani, T. Catal. Lett. 2009, 131, 364−
369. (d) Song, Z.; Takahashi, A.; Nakamura, I.; Fujitani, T. Appl.
Catal., A 2010, 384, 201−205.
(19) Sheng, P.-Y.; Bowmaker, G. A.; Idriss, H. Appl. Catal., A 2004,
261, 171−181.
(20) Idriss, H.; Diagne, C.; Hindermann, J. P.; Kiennemann, A.;
Barteau, M. A. J. Catal. 1995, 155, 219−237.
(21) Pei, Z.-F.; Ponec, V. Appl. Surf. Sci. 1996, 103, 171−182.
(22) Hasan, M. A.; Zaki, M. I.; Pasupulety, L. J. Phys. Chem. B 2002,
106, 12747−12756.
(23) Yee, A.; Morrison, S. J.; Idriss, H. J. Catal. 1999, 186, 279−295.
́
(24) Kukulska-Zając, E.; Gora-Marek, K.; Datka, J. Microporous
Mesoporous Mater. 2006, 96, 216−221.
(25) Delgass, W. N.; Haller, G. L.; Kellerman, R.; Lunsford, J. H.
Spectroscopy in Heterogeneous Catalysis; Academic Press: New York,
1979; pp 19−57.
(26) Yamada, Y.; Segawa, M.; Sato, F.; Kojima, T.; Sato, S. J. Mol.
Catal. A 2011, 346, 79−86.
(5) (a) Arias, D.; Colmenares, A.; Cubeiro, M. L.; Goldwasser, J.;
Lopez, C. M.; Machado, F. J.; Sazo, V. Catal. Lett. 1997, 45, 51−58.
(b) Oikawa, H.; Shibata, Y.; Baba, T. Appl. Catal., A 2006, 312, 181−
185.
(6) (a) Golay, S.; Doepper, R.; Renken, A. Chem. Eng. Sci. 1999, 5,
4469−4474. (b) Bakoyannakis, D. N.; Zamboulis, D.; Stalidis, G. A.;
Deliyanni, E. A. J. Chem. Technol. Biotechnol. 2001, 76, 1159−1164.
(c) Doheim, M. M.; El-Shobaky, H. G. Coll. Surf. A 2002, 204, 169−
174. (d) Zaki, T. J. Colloid Interface Sci. 2005, 284, 606−613.
(e) Varisli, D.; Dogu, T.; Dogu, G. Chem. Eng. Sci. 2007, 62, 5349−
5352; Ind. Eng. Chem. Res. 2008, 47, 4071−4076; Ind. Eng. Chem. Res.
2009, 48, 9394−9401. (f) Carrasco-Marin, F.; Mueden, A.; Moreno-
Castilla, C. J. Phys. Chem. B 1998, 102, 9239−9244. (g) Kamiguchi, S.;
Chihara, T. Catal. Lett. 2003, 85, 97−100. (h) Kamiguchi, S.;
Nagashima, S.; Komori, K.; Kodomari, M.; Chihara, T. J. Cluster Sci.
2007, 18, 414−430.
(7) (a) Kamimura, Y.; Sato, S.; Takahashi, R.; Sodesawa, T.; Akashi,
T. Appl. Catal., A 2003, 252, 2399−2407. (b) Nagashima, O.; Sato, S.;
Takahashi, R.; Sodesawa, T. J. Mol. Catal. A 2005, 227, 231−239.
(c) Tsuchida, T.; Kubo, J.; Yoshioka, T.; Sakuma, S.; Takeguchi, T.;
Ueda, W. J. Catal. 2008, 259, 183−189.
(8) (a) Kasai, K.; Haishi, T.; Iwamoto, M. Shokubai 2007, 49, 126−
128. (b) Iwamoto, M.; Kasai, K. Japanese Patent No. P2008−
255104A.
(9) Iwamoto, M.; Kasai, K.; Haishi, T. ChemSusChem 2011, 4, 1055−
1058.
(10) Sugiyama, S.; Kato, Y.; Wada, T.; Ogawa, S.; Nakagawa, K.;
Sotowa, K. Top. Catal. 2010, 53, 550−554.
(11) (a) Hayashi, F.; Iwamoto, M. ACS Catal. 2013, 3, 14−17.
(b) Hayashi, F.; Tanaka, M.; Lin, M.; Iwamoto, M. J. Catal. 2014, 316,
112−120.
(12) (a) Mizuno, S.; Kurosawa, M.; Tanaka, M.; Iwamoto, M. Chem.
Lett. 2012, 41, 892−894. (b) Iwamoto, M.; Mizuno, S.; Tanaka, M.
Chem.Eur. J. 2013, 19, 7214−7220.
(13) (a) Iwamoto, M.; Kosugi, Y. J. Phys. Chem. C 2007, 111, 13−15.
(b) Ikeda, K.; Kawamura, Y.; Yamamoto, T.; Iwamoto, M. Catal.
Commun. 2008, 9, 106−110. (c) Haishi, T.; Kasai, K.; Iwamoto, M.
Chem. Lett. 2011, 40, 624−626.
(14) Klomp, D.; Mashmeyer, T.; Hanefeld, U.; Peters, J. A. Chem.
Eur. J. 2004, 10, 2088−2093.
3469
dx.doi.org/10.1021/cs5006822 | ACS Catal. 2014, 4, 3463−3469