Insights into the ceria-catalyzed ketonization reaction for biofuels applications

Article abstract of DOI:10.1021/cs400003n

The ketonization of small organic acids is a valuable reaction for biorenewable applications. Ceria has long been used as a catalyst for this reaction; however, under both liquid and vapor phase conditions, it was found that given the right temperature regime of about 150-300 °C, cerium oxide, which was previously believed to be a stable catalyst for ketonization, can undergo bulk transformations. This result, along with other literature reports, suggest that the long held belief of two separate reaction pathways for either bulk or surface ketonization reactions are not required to explain the interaction of cerium oxide with organic acids. X-ray photon spectroscopy, scanning electron microscopy, and temperature programmed decomposition results supported the formation of metal acetates and explained the occurrence of cerium reduction as well as the formation of cerium oxide/acetate whiskers. After thermogravimetry/mass spectrometry and FT-IR experiments, a single reaction sequence is proposed that can be applied to either surface or bulk reactions with ceria.

Full text of DOI:10.1021/cs400003n

Research Article
pubs.acs.org/acscatalysis
Insights into the Ceria-Catalyzed Ketonization Reaction for Biofuels
Applications
,
†
Ryan W. Snell and Brent H. Shanks*
†Department of Chemical and Biological Engineering, Iowa State University, 1140L BRL, Ames, Iowa, 50011, United States
*
S Supporting Information
ABSTRACT: The ketonization of small organic acids is a
valuable reaction for biorenewable applications. Ceria has long
been used as a catalyst for this reaction; however, under both
liquid and vapor phase conditions, it was found that given the
right temperature regime of about 150−300 °C, cerium oxide,
which was previously believed to be a stable catalyst for
ketonization, can undergo bulk transformations. This result,
along with other literature reports, suggest that the long held
belief of two separate reaction pathways for either bulk or surface
ketonization reactions are not required to explain the interaction
of cerium oxide with organic acids. X-ray photon spectroscopy, scanning electron microscopy, and temperature programmed
decomposition results supported the formation of metal acetates and explained the occurrence of cerium reduction as well as the
formation of cerium oxide/acetate whiskers. After thermogravimetry/mass spectrometry and FT-IR experiments, a single
reaction sequence is proposed that can be applied to either surface or bulk reactions with ceria.
KEYWORDS: ketonization, cerium oxide, acetic acid, metal carboxylate, bio-oil
1
. INTRODUCTION
underlying chemistry involved in the reaction. Unfortunately,
despite recent advances, much about ketonization remains
Although a number of renewable energy alternatives, such as
solar and wind, are increasingly being utilized, there remains a
tremendous challenge in replacing a trillion dollar infrastructure
built around liquid fuels. Therefore, the creation of a renewable
liquid fuel able to successfully make use of the current
infrastructure would be extremely valuable. Because biomass
provides a renewable source of carbon, it is a potential
alternative to current petroleum use. Although there are a
number of different techniques being explored for converting
biomass to liquid fuels, fast pyrolysis has been increasingly
discussed because of the simplicity of the process, direct
creation of a liquid bio-oil, and the flexibility to use different
4
,5
debated, including the catalytic mechanism.
This is
particularly true regarding the reaction in the condensed phase.
Adding to the challenge of understanding the ketonization
reaction is the postulate that the reaction proceeds through
different mechanisms, depending on the metal oxide catalyst
4
,6
used. For metal oxide catalysts with a low metal−oxygen
bond strength or low lattice energy, the reaction purportedly
progresses through the formation of a bulk metal carboxylate,
which then acts as the working catalyst and whose thermal
decomposition results in ketone production. In contrast, for
catalysts with a high metal−oxygen bond strength or large
lattice energy, a surface-catalyzed reaction is reported to
1
feedstocks. However, there are a number of undesirable
characteristics of this oil, such as a low pH and high oxygen
content, both caused in part by its large quantity of small
4
promote the formation of the ketone. For reference, the
value of the lattice energy for ceria relative to other potential
ketonization catalysts is given in Table 1. However, Mekhemer
et al. reported that MgO could catalyze the ketonization
reaction through either bulk or surface reactions, depending on
the temperature. They reported that MgO underwent trans-
formation to a bulk acetate during room temperature exposure
to acetic acid while higher temperatures favored a reaction
2
organic acids. Since bio-oil contains a number of thermally
unstable compounds, it would be preferred to perform
condensed phase catalytic reactions so as to address some of
these issues.
One reaction that would be potentially beneficial is
ketonization. The ketonization reaction has been attracting
increasing attention for biorenewable fuel applications since
Kunkes et al. reported its use during the conversion of glucose
7
proceeding through surface adsorption. This apparent
deviation from the earlier suggested lattice-energy-directed,
two-pathway model was proposed to occur in this particular
instance because of the unique basicity of the oxide. Given
3
to liquid fuels. Regrettably, the high temperatures, typically
>
300 °C, utilized for the reaction make it impractical for
condensed phase applications, as would be preferred with bio-
oil. Increasing catalyst activity is thus vital for obtaining
reasonable liquid phase reaction conditions. To achieve this
task, it would be helpful to have a good understanding of the
Received: January 2, 2013
Revised: March 4, 2013
Published: March 8, 2013
©
2013 American Chemical Society
783
dx.doi.org/10.1021/cs400003n | ACS Catal. 2013, 3, 783−789

ACS Catalysis
Research Article
Table 1. Lattice Energies of Potential Ketonization
Catalysts
different reaction temperatures, it was found that at 150 °C,
only traces of acetone were observed after 24 h using an acid/
ceria ratio of 5, but increasing the reaction temperature to 300
8
material
lattice energy (kJ/mol)
°
C led to complete conversion of the acetic acid in 30 min.
A number of reports in which the ketonization reaction was
15916
15276
12970
12150
11188
9627
2
Al O3
2
Cr O3
performed at temperatures above 300 °C have reported that the
structure of the ceria catalyst was stable during the ketonization
reaction, as determined by postreaction characterization using
MnO2
TiO2
ZrO2
6,13,15
XRD.
However, in the current work, postreaction XRD
CeO2
characterization of the ceria catalyst used at 230 °C
demonstrated that the oxide underwent bulk restructuring
during the course of the reaction. Further reactions were run to
BeO
4514
ZnO
4142
MgO
PbO
3795
see if this crystal structure change would occur at lowe7r
3520
temperatures, as was found by Mekhemer et al. with MgO.
CaCO3
2804
Interestingly, the original fluorite phase for the ceria was
maintained when exposed to acetic acid at temperatures lower
than 200 °C, demonstrating that bulk absorbance of the acid
was not occurring at these temperatures. In addition, the
fluorite crystal structure was preserved when the ceria was used
at a reaction temperature of 300 °C. Therefore, as shown in
Figure 2 for the postreaction ceria, there appeared to be an
these results, the authors proposed that the reaction mechanism
was different for low-temperature ketonization, which purport-
edly progressed through the pyrolytic route, from that at high
temperatures, which catalyzed the reaction through the use of
7
both acid and base surface sites.
In a large screening study of metal oxides in the ketonization
reaction, ceria was found to be a ketonization catalyst superior
9
to MgO, and ceria has been the focus of a number of
1
0−14
studies.
Because we were interested in exploring
ketonization at lower temperatures in the condensed phase, a
high-activity catalyst was desirable, so the focus in the current
work was how the reaction of the model bio-oil compound,
9
acetic acid, proceeded in the presence of ceria. In particular, we
examined the question of bulk transformation versus surface
adsorption for ceria.
Figure 2. XRD figures of postreaction ceria catalysts used at different
temperatures in the ketonization of acetic acid (1.0 g of acetic acid, 24
h, 0.20 g of catalyst).
2
. RESULTS AND DISCUSSION
2.1. Formation of Cerium Acetate. The synthesized ceria
used in the reaction studies was first characterized in its fresh
2
state. The material, which had a BET surface area of 73 m /g,
intermediate temperature range during which ceria restructur-
ing occurred. It has been proposed that metal oxide catalysts
with metal−oxygen bonds weak enough to undergo bulk
was found from X-ray diffraction (XRD) to be in the cerianite
form (Figure S1, Supporting Information). Consistent with the
XRD results, X-ray photoelectron spectroscopy (XPS) analysis
showed that the cerium oxidation state was primarily Ce4+
16
carboxylate formation have low activity in ketonization.
However, these results appeared to contradict this postulate
because ceria, one of the most active ketonization catalysts, was
clearly found to undergo bulk crystal changes. To validate that
this change in the bulk structure was due to interaction with
acetic acid and not a solvent effect, a control was performed in
which no acetic acid was added. In this experiment, no bulk
rearrangement occurred, which unquestionably demonstrated
the requirement of acetic acid interaction with ceria to cause
the bulk transformation.
To further confirm that bulk metal acetate was being formed
during the course of the reaction, a temperature-programmed
decomposition (TPD) experiment was performed in which the
spent catalyst was heated at a 5 °C/min ramp rate and the
evolved products were measured through the use of mass
spectrometry. As shown in Figure 3, acetone was evolved in
two steps, the first beginning at ∼250 °C and reaching a
maximum at ∼300 °C, which was near the temperatures that
the ketonization reaction readily occurred. To provide a
comparison, TPD was also performed with purchased cerium
acetate hydrate. Figure 4 demonstrates that a similar evolution
of products was found for cerium acetate, and the acetone
evolution appeared to occur at approximately the same
temperatures. The one slight difference with the cerium acetate
hydrate was the appearance of three acetone evolution peaks, as
(
Supporting Information Figure S2). From scanning electron
spectroscopy (SEM) images (Supporting Information Figure
S3), the ceria particles were in the form of nonuniformly
shaped agglomerated crystals.
The primary focus of the current work was exploring lower
temperatures that could be utilized for reaction in the
condensed phase. As shown in Figure 1, the acetic acid
ketonization reaction proceeded in toluene, even at the
comparatively low temperature of 230 °C. From exploring
Figure 1. Ketonization of acetic acid indicating that the reaction can
proceed at low temperatures (230 °C, 50 mL of toluene, 1.0 g of acetic
acid, 0.055 g of ceria).
7
84
dx.doi.org/10.1021/cs400003n | ACS Catal. 2013, 3, 783−789

ACS Catalysis
Research Article
exposed to the same reaction conditions as was the ceria. After
this exposure, both the cerium acetate hydrate and cerium
carbonate hydrate were found to yield diffractograms very
similar to that of the postreaction ceria, lending more evidence
to the formation of a metal−carboxylate intermediate during
ceria catalyzed ketonization (Figure 6).
Figure 3. TPD of cerium oxide catalyst after use for 24 h in the
ketonization of acetic acid at 230 °C.
Figure 6. XRD figures of (a) fresh cerium acetate hydrate, (b) reaction
exposed cerium acetate, and (c) reaction-exposed cerium carbonate
hydrate (0.85 or 1.0 g of acetic acid for acetate or carbonate, 2 h
reaction, 230 °C, 0.28 g of cerium acetate, 0.20 g of cerium carbonate).
SEM imaging of the ceria catalyst used at 230 °C showed the
formation of large rods or whiskers (Figure 7). A control in
Figure 4. TPD of cerium acetate hydrate.
Figure 7. SEM images of ceria catalysts after use in (a) 24 h reaction
with an acid/catalyst weight ratio of 18.1, 230 °C, 50 mL of toluene
b) 24 h and 230 °C in toluene solvent, and (c) cerium acetate hydrate
after use in reaction with an acid/acetate weight ratio of 3.1, 2 h
reaction, 230 °C, 50 mL of toluene.
opposed to the two observed with the spent catalyst. A possible
cause for this difference may be that the spent ceria catalyst was
not fully transformed into bulk acetate, as seen in Figure 2, and
instead was in an intermediate cerium oxyacetate form. Because
the catalyst used at 300 °C did not show the bulk restructuring
during the course of reaction as determined by XRD, TPD was
performed on this used catalyst to see whether acetone would
be evolved at similar temperatures. As shown in Figure 5, this
result was, indeed, the case.
(
which ceria was exposed to the reaction conditions without the
presence of acetic acid demonstrated that this restructuring was
growth promoted not by the solvent but, rather, through
interaction with the acetic acid. Interestingly, the cerium acetate
hydrate when used under reaction conditions in the presence of
acetic acid also displayed the formation of rods, but in this case,
the rods were significantly aggregated. The SEM results
complemented and corroborated the previous XRD and TPD
findings in that some form of metal carboxylate was being
formed at the intermediate reaction temperatures.
Because reports on the ketonization of small organic acids
have almost exclusively been performed in the gas phase as a
result of the high temperatures being used, it was evaluated
whether the condensed phase results were unique to this phase
or were more generally applicable to the reaction independent
of the reaction phase. To probe this question, in situ XRD was
used in which acetic acid vapors were passed over the ceria
catalyst at different temperatures while continuously scanning
over a specified 2θ range. The results are shown in Figure 8. As
the sample was taken through the different temperature
regimes, bulk acetate formation was observed, demonstrating
that the phase change behavior of the ceria when exposed to
acetic acid in the gas phase was similar to that seen in the
condensed phase. Taken together, the results from examining
behavior in either reaction phase suggested that acetate
formation occurred independently of the reaction phase within
Figure 5. Ceria catalyst TPD profile after being used in the
ketonization of acetic acid at 300 °C for 30 min.
XRD of the cerium acetate hydrate sample did not result in a
diffractogram that was equivalent to that found with the ceria
catalyst used at 230 °C, which supported the observation that
during the course of the reaction, ceria was not transformed
into a pure carboxylate but, instead, some apparent
intermediate form. To further explore the possible connection,
both cerium acetate hydrate and cerium carbonate hydrate were
7
85
dx.doi.org/10.1021/cs400003n | ACS Catal. 2013, 3, 783−789

ACS Catalysis
Research Article
not the result of metal−oxygen bond strength differences.
Therefore, this material property alone does not dictate the
route for the ketonization reaction pathway. Furthermore,
recent reports have shown that oxides with even higher lattice
energies than ceria such as La O and Nd O undergo partial
2
3
2
3
bulk acetate formation during ketonization of acetic acid, but
13
ceria did not. The reason that ceria was found to be stable in
the study was likely due to the high reaction temperature being
used. It is worth considering whether two different reaction
pathways are required to reconcile previously reported
ketonization results. It is well-known that the pyrolysis of
17
select metal acetates leads to the formation of acetone;
however, clear evidence that the surface mechanism is catalytic
and different from the bulk thermal decomposition mechanism
is not as apparent.
It is possible the proposed surface catalytic pathway is just
the result of surface carboxylate pyrolysis. This hypothesis
would explain the ceria XRD observations in Figure 2. At high
temperatures, metal carboxylate formation in the bulk would
not occur because of the thermal instability of this species at
these conditions. Moreover, properties of the particular metal
oxide, such as its lattice energy, could prevent the formation of
carboxylates in the bulk until sufficient energy is available to
break M−O bonds in the metal oxide. If the reaction
temperature required to break M−O bonds and form
carboxylates in the bulk is higher than the value for which
the metal carboxylate decomposes, then disruption of the
crystal structure would not be observed. However, this situation
does not require different reaction mechanisms because the
metal carboxylates could still form on the surface or, depending
on transport limitations, possibly even within the first few
atomic layers. After formation, these carboxylates could
pyrolytically decompose to form the ketone. Hence, in either
the bulk or surface-promoted cases, a necessary intermediate
species would be a metal carboxylate. A schematic diagram
depicting this hypothesis for ceria is shown in Figure 10.
Figure 8. In situ XRD characterization of ceria behavior during the
course of acetic acid vapor phase ketonization at various temperatures.
Temperatures were increased after every 9 scans so that scans 1−9,
1
0−18, 19−27, 28−36, and 36−45 occurred at temperatures of 150,
00, 230, 270, and 300 °C, respectively.
2
the temperature window of ∼200−300 °C. Of note from Figure
8
was that at 300 °C (scans 36−45), the diffractogram had not
reverted back to the original ceria form, as might be expected
from the results shown in Figure 2.
To explore this apparent difference, a condensed phase
reaction was performed at 230 °C with 1.0 g of acetic acid and
0
.20 g of ceria for 24 h. After reaction, the catalyst was
separated and analyzed by XRD, and then the postreaction
ceria catalyst was placed in toluene and held for 1 h at 300 °C.
Following this treatment, the ceria was again separated and
analyzed by XRD, with the results compared with a catalyst
exposed to reaction conditions at 300 °C for 30 min with 1.0 g
of acetic acid, and 0.2 g of ceria. Figure 9 shows that results
Figure 10. Different temperature regimes for ceria-catalyzed
ketonization. At low temperatures, there is insufficient energy for
the acetic acid to break the Ce−O bonds; at intermediate
temperatures, acetates are able to form but are stable enough to
form carboxylates in the bulk; and at high temperatures, acetates form
but are rapidly decomposed to form acetone, leaving the bulk structure
maintained.
Figure 9. XRD profiles of ceria exposed to 24 h reactions at 230 °C,
rerunning after the catalyst was placed in toluene at 300 °C for 1 h,
and ceria after only 30 min of reaction at 300 °C.
found in the condensed phase were analogous to those
discovered in the vapor phase. For the experiments in both
phases, the ceria did not revert back to its original crystal
structure at 300 °C after the initial formation of metal
carboxylates in the bulk. However, as shown in Figure 2 and
Supporting Information Figure S4, when ceria was exposed to
acetic acid in either the liquid or vapor phase at 300 °C, the
material did not show the formation of a bulk metal
carboxylate. Since previously reported research examining the
stability of ceria in the ketonization reaction was performed at
temperatures at or greater than 300 °C, the lack of bulk
2.2. Decomposition of Cerium Carboxylates. It has
been commonly reported that an α-hydrogen is necessary for
ketonization to proceed, which lends support to the formation
4
,18
of a ketene-like intermediate.
To examine whether this
concept would fit into the hypothesis of thermal carboxylate
decomposition, the ketonization of pivalic acid (trimethylacetic
acid) was attempted in the condensed phase. Consistent with
the literature, none of the expected ketonization product
(2,2,4,4-tetramethyl-3-pentanone) was found at temperatures
of up to 315 °C in 24 h reaction runs (1.7 g of pivalic acid, 0.2 g
of catalyst). Moreover, when the postreaction ceria catalyst was
separated from the reaction mixture using vacuum filtration and
9
,13
transformation would be expected.
From the results in the current work, it is apparent that the
lack of carboxylate formation in the bulk at high temperatures is
7
86
dx.doi.org/10.1021/cs400003n | ACS Catal. 2013, 3, 783−789

ACS Catalysis
Research Article
analyzed by XRD at temperatures up to 300 °C, the original
cerianite structure was maintained. In contrast, when the
ketonization reaction was run at 315 °C, no catalyst powder
could be observed in the reaction mixture post reaction. After
drying in air to remove the solvent and reactant, a powder
precipitate was formed. Averaging over four EDS spectra, the
content of Ce, C, and O in the powder were found to be 39, 43,
and 17 wt %, respectively. This composition is approximately
that of cerium(III) pivalate (34% Ce, 43% C, 23% O). The
electron-donating methyl groups or sterics of pivalic acid could
explain the higher temperature necessary for bulk carboxylate
formation when ceria was exposed to pivalic acid. SEM images
of the recovered precipitate from the reaction did not show
rod-shaped morphology (see Supporting Information Figure
S5), as was evident for the ceria after reaction with acetic acid.
This result was reasonable because cerium pivalate is readily
soluble in toluene, but cerium acetate is not.
pivalate because it was not decomposed until such a high
temperature that the formed ketone decomposed immediately.
These results suggested that the necessity of an α-hydrogen
could actually be related more to a pyrolytic mechanism than a
catalytic mechanism.
2.3. Proposed Acetic Acid Ketonization Reaction
Sequence with Ceria. Kinetic modeling of the ketonization
reaction has shown that Langmuir−Hinshelwood kinetics with
two adsorbed species reacting result in a good fit with
2
1,22
experimental data.
Carboxylate species can bind with
23
metal cations in a number of different conformations. The
type of binding may influence the ease with which the
intermediate metal acetate is thermally decomposed. Metal
acetate infrared spectra and its connection to carboxylate
coordination has previously been studied. Differences in
carboxylate bond symmetry were found to occur depending on
the type of binding, that is, unidentate, bidentate, or ionic.
These differences create varying locations for the υasym(CO2)
24
XPS of the precipitate sample (Supporting Information
4+
Figure S6) showed the cerium, which was in the Ce state
and υsym(CO ) bands, thus changing the Δ
values and
2
asy‑sym
3
+
prior to reaction, had been fully reduced to Ce in the course
of cerium pivalate formation. The observation of metal ion
reduction might explain some reports of cerium reduction
during the course of ketonization by demonstrating that
reduction of the metal oxide occurred during carboxylate
formation and not necessarily during evolution or further
potentially allowing for determination of carboxylate coordina-
24,25
tion.
Several reports have shown that although there is
some value in this method, caution should still be used so as
25−27
not to reach unfounded conclusions.
Therefore, the
spectra evaluation presented here will be generalized.
The FTIR spectrum of cerium acetate after heating to 230
°C, is shown in Figure 12, along with the difference spectra of
1
9,20
reactions of the ketone.
This explanation would be
consistent with our XPS characterization of the postreaction
ceria. The material was dried in air at 100 °C, which would
likely oxidize any reduced oxide that had formed; however, the
bulk cerium carboxylate formation could stabilize the cerium in
the 3+ state and would not decompose in air at the low
temperature used for drying.
The observation that no pivalic acid ketonization over cerium
oxide occurred at temperatures up to 315 °C in the condensed
phase was consistent with the proposed need for an α-
hydrogen. It has been shown previously that branching and,
thus, a loss in the number of α-hydrogen atoms for a given acid
18
Figure 12. FTIR spectra of cerium acetate after (c) heat treatment at
compound decreases its activity in the ketonization reaction,
2
30 °C, (b) ceria difference spectra after exposure to acetic acid at 230
so the thermal stability of metal carboxylates should increase in
the order of acetate < propionate < isobutyrate < pivalate.
Figure 11 shows the decomposition behavior of the respective
°
°
C, and (a) ceria difference spectra after exposure to acetic acid at 300
C.
cerium oxide after exposure to small amounts of acetic acid
vapor at either 230 or 300 °C. Although only one set of peaks
appeared for cerium acetate, it is evident from the figure that
the asymmetric peak was split for the ceria compounds. From
−
1
peak maxima, the Δ
was found to be equal to 153 cm
asy−sym
1
for the acetate, Δ
= 109/129 cm− for the ceria exposed
asy−sym
= 104/121 cm− for the
1
to acetic acid at 230 °C, and Δ
asy−sym
ceria exposed at 300 °C. The acetate value agreed quite well
of 150 cm− for cerium
1
with the previously reported Δ
asy−sym
Figure 11. Normalized MS results obtained during the thermogravi-
metric analysis of cerium carboxylates with different degrees of
branching (a) cerium pivalate m/z = 58, (b) cerium isobutyrate m/z =
acetate hydrate, and the Δ
for ceria exposed to acetic acid
asy−sym
−1
at 300 °C was very close to the Δ
of 105 cm
asy−sym
determined by Hasan et al. after exposure of their ceria catalyst
to acetic acid at the same temperature. According to Deacon
71, (c) cerium propionate m/z = 86, and (d) cerium acetate m/z = 58.
11
2
1
−1
and Phillips,
Δ
> 200 cm generally corresponds to
asy−sym
< 150 cm− corresponds to
1
cerium carboxylates as measured by thermogravimetry/mass
spectrometry. The relative decomposition order followed
exactly what was expected in which substitution of methyl
groups for the α-hydrogens dramatically increased the stability
of the carboxylate. Cerium acetate, propionate, and isobutyrate
decomposed and formed their corresponding ketones. How-
ever, it was difficult to probe the ketone formation for cerium
unidentate binding, Δ
bridging or chelating coordination, and Δ
asy−sym
< 105 cm−
1
asy−sym
is typically more connected to chelation. Although cerium
acetate has been reported to have bridging carboxylate bonds, it
is unlikely that acetic acid on the surface of ceria could
coordinate in a bridging fashion because the bond length
25,28
between Ce cation pairs is too great.
Therefore, it is
7
87
dx.doi.org/10.1021/cs400003n | ACS Catal. 2013, 3, 783−789

ACS Catalysis
Research Article
Figure 13. Proposed ketonization reaction sequence.
possible that the acid is binding in a chelating fashion on the
surface of ceria. Ceria can expose crystallographic facets of
α-carbon on the molecule whose carboxylate group will
eventually form carbon dioxide; however, the reaction sequence
in Figure 13 would also alleviate concerns that Nagashima et
{
111}, {100}, and {110}, but typically, the {111} surface is
29
18
prevalent due to its enhanced stability. Because the ideal
termination of the {111} surface exposes only cerium with one
dangling bond, it seems unlikely that a chelating coordination
could occur without surface reconstruction. However, recon-
struction of the surface seems quite feasible because Figure 2
shows even bulk crystallite changes can occur given the right
conditions.
al. expressed about the model by Pestman et al. because the
mechanism has the C−C cleavage occurring in the same step as
ketone production, thereby prohibiting formation of side
product, such as methanol. It is worth noting that the proposed
reaction sequence could potentially be generalized to any type
of metal oxide ketonization catalyst as well as to being able to
occur on the surface or throughout the bulk of the respective
material. As such, differences seen between catalysts could be
explained through the formation of different types of
carboxylate coordination as well as resulting from the
electronegativity of the respective metal, creating discrepancies
as to the ease in which the C−C bond is broken. The strength
of the M−O bonds in the metal oxide is also important in the
proposed mechanism. Strong M−O bonds would make
formation of the needed metal carboxylate groups difficult,
whereas weak M-O bonds would result in the formation of
thermally stable metal carboxylates.
Carboxylate binding modes may explain some of the
phenomena observed with the ceria catalysts that were
characterized post reaction. Figure 7 demonstrated that as
ceria began to form a bulk carboxylate, the overall macro-
structure of the particles was modified toward whisker
morphology. The binding of cerium acetate in a bridging
arrangement could facilitate this type of elongation, but
chelating would not. The different form of metal carboxylate
coordination might also help explain the slight differences in
Figures 3−5 for cerium acetate and the ceria catalysts used at
different temperatures. Ceria used at 230 °C, as would be
expected from the XRD results, decomposed in a pattern more
closely resembling cerium acetate than did ceria used at 300 °C,
since the higher temperature would not allow for bulk acetate
formation.
3
. CONCLUSIONS
Characterization techniques such as XPS, SEM, FTIR, and
XRD of fresh and spent ceria catalysts were found to be
necessary to explain some of the phenomena, such as cerium
reduction and morphology changes, seen in this work as well as
in previous reports on the ketonization reaction. Degradation of
the ceria crystal structure was found to be dependent on the
temperature regime used for the ketonization reaction.
Therefore, the long postulated belief that ketonization with
ceria occurs through multiple catalytic routes, based upon
observance of a bulk carboxylate phase, does not need to be
invoked. A mechanism involving the coupling of two
carboxylate species, either on the surface or throughout the
bulk, would explain the results observed in the current work as
well as experimental results given in previous work by other
groups. The postulate that pyrolytic decomposition of
carboxylates is the primary mechanism for ketonization
provides a direction for developing improved catalytic
materials. Although the postulated ketonization reaction
sequence can be extended to explain results for a range of
metal oxides, further investigations into ketonization catalysts
are needed, particularly with respect to more complex systems,
such as supported metal oxides or mixed metal oxides.
Taken together, the results from the current work suggested
a carboxylate mechanism similar to that proposed by Dooley,
17
as shown in Figure 13. This reaction scheme first involves the
abstraction of an α-proton by a carboxyl group bound to a
metal. Surface basic sites in a cerium oxyacetate or oxide may
also perform this abstraction when available. This difference
between the acetate/oxyacetate/oxide system may explain the
enhanced stabilities of metal carboxylates observed as shown in
Figures 8 and 9. The abstraction step would also be consistent
with the reports claiming the necessity of an α-hydrogen.
Branched carboxylates would have greater stability (Figure 11)
because the added substituents are electron donors that would
lower the acidity. In this mechanism, the reaction would
proceed through the formation of a six-membered ring
intermediate before the simultaneous formation of CO and
2
acetone, agreeing with Figures 3−5.
Water and carbon dioxide have been commonly cited as
21
inhibitors for the ketonization reaction. This effect would be
consistent with the proposed mechanism because water would
hydrolyze the carboxylate bonds and carbon dioxide would
competitively bind metal sites either in the bulk or on the
surface. The proposed mechanism is somewhat similar to that
4
of Pestman et al. in that a proton is initially removed from the
7
88
dx.doi.org/10.1021/cs400003n | ACS Catal. 2013, 3, 783−789

ACS Catalysis
ASSOCIATED CONTENT
Research Article
(
25) Deacon, G. B.; Phillips, R. J. Coord. Chem. Rev. 1980, 33, 227−
■
2
50.
* Supporting Information
S
(
(
26) Edwards, D. A.; Hayward, R. N. Can. J. Chem. 1968, 46, 3443.
27) Deacon, G. B.; Huber, F.; Phillips, R. J. Inorg. Chim. Acta 1985,
Figures S1−S3 showing the XRD, XPS, and SEM image of the
fresh ceria. In situ XRD of ceria exposed to acetic acid vapors at
104, 41−45.
300 °C in Figure S4. Figures S5 and S6 showing SEM image
(28) Stubenrauch, J.; Brosha, E.; Vohs, J. M. Catal. Today 1996, 28,
431−441.
and XPS of ceria after exposure to pivalic acid. This material is
available free of charge via the Internet at http://pubs.acs.org.
(29) Baudin, M.; Woj
́
cik, M.; Hermansson, K. Surf. Sci. 2000, 468,
5
1−61.
AUTHOR INFORMATION
■
Corresponding Author
E-mail: bshanks@iastate.edu.
*
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was funded by the National Science Foundation
through PIRE (OISE 0730227) and ERC (EEC-0813570)
grants. Jim Anderegg (Ames Laboratory) was very kind in
helping with the XPS results. Dan Weis, an ISU undergraduate,
also helped perform some of the experiments.
REFERENCES
■
(
(
(
1) Bridgwater, A. V. Biomass Bioenergy 2012, 38, 68−94.
2) Czernik, S.; Bridgwater, A. V. Energy Fuels 2004, 18, 590−598.
3) Kunkes, E. L.; Simonetti, D. A.; West, R. M.; Serrano-Ruiz, J. C.;
Gartner, C. A.; Dumesic, J. A. Science 2008, 322, 417−421.
(
4) Pestman, R.; Koster, R. M.; van Duijne, A.; Pieterse, J. A. Z.;
Ponec, V. J. Catal. 1997, 168, 265−272.
(
(
5) Rajadurai, S. Catal. Rev.: Sci. Eng. 1994, 36, 385−403.
6) Yakerson, V. I.; Fedorovskaya, E. A.; Klyachko-Gurvich, A. L.;
Rubinshtei, A. M. Kinet. Katal. 1961, 2, 907−915.
7) Mekhemer, G. A. H.; Halawy, S. A.; Mohamed, M. A.; Zaki, M. I.
J. Catal. 2005, 230, 109−122.
8) Lide, D. R. CRC Handbook of Chemistry and Physics; CRC Press
LLC: Boca Raton, 2003−2004.
9) Glinski, M.; Kijenski, J.; Jakubowski, A. Appl. Catal., A 1995, 128,
09−217.
10) Gaertner, C. A.; Serano-Ruiz, J. C.; Braden, D. J.; Dumesic, J. A.
(
(
(
2
(
J. Catal. 2009, 266, 71−78.
(
(
11) Deng, L.; Fu, Y.; Guo, Q.-X. Energy Fuels 2009, 23, 564−568.
12) Kobune, M.; Sato, S.; Takahashi, R. J. Mol. Catal. A: Chem.
2
(
008, 279, 10−19.
13) Yamada, Y.; Segawa, M.; Sato, F.; Kojima, T.; Sato, S. J. Mol.
Catal. A: Chem. 2011, 346, 79−86.
(
(
14) Snell, R. W.; Shanks, B. H. Appl. Catal., A 2013, 451, 86−93.
15) Hasan, M. A.; Zaki, M. I.; Pasupulety, L. Appl. Catal., A 2003,
2
(
43, 81−92.
16) Dooley, K. M. Catalysis of Acid/Aldehyde/Alcohol Con-
densations to Ketones. In Catalysis; Spivey, J. J., Roberts, G. W., Eds.;
The Royal Society of Chemistry, Cambridge, 2004, pp 293−319.
(
(
17) Friedel, C. Justus Liebigs Ann. Chem. 1858, 108, 122−125.
18) Nagashima, O.; Sato, S.; Takahashi, R.; Sodesawa, T. J. Mol.
Catal. A: Chem. 2005, 227, 231−239.
19) Rubinshtein, A. M.; Slinkin, A. A.; Yakerson, V. I.;
Fedorovskaya, E. A. Russ. Chem. Bull. 1961, 10, 2090−2092.
20) Gangadharan, A.; Shen, M.; Sooknoi, T.; Resasco, D. E.;
Mallinson, R. G. Appl. Catal., A 2010, 385, 80−91.
21) Gaertner, C. A.; Serrano-Ruiz, J. C.; Braden, D. J.; Dumesic, J. A.
J. Catal. 2009, 266, 71−78.
22) Rajadurai, S.; Kuriacose, J. C. Mater. Chem. Phys. 1987, 16, 17−
9.
23) Mehrotra, R. C.; Bohra, R. Metal Carboxylates; Academic Press:
London, 1983.
24) Nakamoto, K.; Fujita, J.; Tanaka, S.; Kobayashi, M. J. Am. Chem.
Soc. 1957, 79, 4904−4908.
(
(
(
(
2
(
(
7
89
dx.doi.org/10.1021/cs400003n | ACS Catal. 2013, 3, 783−789