Adsorption and reaction of acetaldehyde on shape-controlled CeO2 nanocrystals: Elucidation of structure-function relationships

Article abstract of DOI:10.1021/cs500611g

CeO₂ cubes with {100} facets, octahedra with {111} facets, and wires with highly defective structures were utilized to probe the structure-dependent reactivity of acetaldehyde. Using temperature-programmed desorption (TPD), temperature-programm

Full text of DOI:10.1021/cs500611g

Research Article
pubs.acs.org/acscatalysis
Adsorption and Reaction of Acetaldehyde on Shape-Controlled CeO2
Nanocrystals: Elucidation of Structure−Function Relationships
Amanda K. P. Mann, Zili Wu, Florencia C. Calaza,† and Steven H. Overbury*
Chemical Sciences Division and Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee
37831, United States
S
* Supporting Information
ABSTRACT: CeO2 cubes with {100} facets, octahedra with {111}
facets, and wires with highly defective structures were utilized to
probe the structure-dependent reactivity of acetaldehyde. Using
temperature-programmed desorption (TPD), temperature-pro-
grammed surface reactions (TPSR), and in situ infrared spectros-
copy, it was determined that acetaldehyde desorbs unreacted or
undergoes reduction, coupling, or C−C bond scission reactions,
depending on the surface structure of CeO2. Room-temperature
FTIR indicates that acetaldehyde binds primarily as η1-acetaldehyde
on the octahedra, in a variety of conformations on the cubes,
including coupling products and acetate and enolate species, and
primarily as coupling products on the wires. The percent
consumption of acetaldehyde ranks in the following order: wires
> cubes > octahedra. All the nanoshapes produce the coupling product crotonaldehyde; however, the selectivity to produce
ethanol ranks in the following order: wires ≈ cubes ≫ octahedra. The selectivity and other differences can be attributed to the
variation in the basicity of the surfaces, defects densities, coordination numbers of surface atoms, and the reducibility of the
nanoshapes.
KEYWORDS: CeO2 nanoshapes, structure dependence, acetaldehyde reaction, DRIFTS, temperature-programmed reaction,
Aldol condensation, Cannizzaro disproportionation
optimize these selective conversions, fundamental studies of
model systems are required. Acetaldehyde is of particular
interest as a model oxygenate compound, as it can undergo
acid- or base-catalyzed coupling reactions to produce higher-
molecular-weight species, forming value-added products. Still,
there remains a need to evaluate the influence of various shape-
dependent properties on catalytic performance to provide
information that can guide the design of more active and
selective catalysts.
To this end, experiments have been performed probing the
adsorption and reaction of acetaldehyde on CeO2(111) and
CeO2(100) single-crystal thin films.16−18 It was found that
these two faces vary with respect to the initial adsorption of
acetaldehyde and the resulting desorption products. Acetalde-
hyde primarily binds in a μ-C,O-conformation on CeO2(100)
facets, and in a η1-conformation on CeO2(111).16−18 Reactivity
also varies considerably between the different surface
terminations. CeO2(100) evolves unreacted acetaldehyde,
CO2, CO, acetylene, H2O, and trace amounts of crotonalde-
hyde, while CeO2(111) facets simply desorb unreacted
acetaldehyde. In addition, reduction of the ceria surface (to
INTRODUCTION
■
Cerium oxide (CeO2) has been shown to be an attractive
material for many applications, because of its oxygen storage
capability, rich redox chemistry, versatile acid and base catalytic
chemistry, and relative abundance. For example, it is an
important component in three-way catalytic converters,1
water−gas shift reactions,2 oxygen sensors,3,4 and fuel cells.5
In addition, it has been found to promote several organic
reactions,6 including ethanol oxidation,7 decomposition of
formic acid,8,9 and CO oxidation.10
Previously, experiments have shown that CeO2 is an active
catalyst to promote the coupling of acetaldehyde, an oxygenate
species, to form several products including crotonaldehyde,
crotyl alcohol, butane, butadiene, and acetone.11,12 Oxygenates
have garnered much attention as fuel additives, as components
obtained from the processing of biomass, and as an
intermediate for value-added products.13 However, the
conversion of oxygenates upon catalysis is not well-understood.
It is of critical importance to optimize these catalytic
transformations, and this goal requires detailed, fundamental
studies of structure−activity relationships using well-defined
catalysts. For example, oxygenate species derived from the
processing of biomass often require the removal of oxygen via
breakage of C−O bonds without scission of C−C bonds to
generate valuable products.14,15 In order to understand and
Received: May 5, 2014
Revised: June 10, 2014
© XXXX American Chemical Society
2437
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
CeO2 Cubes and Wires. CeO2 cubes and wires were prepared
based on a method published previously with some modifications.27
Briefly, 0.434 g of Ce(NO3)3·6H2O (0.001 mol) was dissolved in 2.5
mL of H2O. Then, 4.8 g of NaOH (0.120 mol) was dissolved in 17.5
mL of H2O and added to the Ce(NO3)3 solution. Upon the addition
of NaOH solution, an opaque precipitate formed, indicating hydrolysis
of Ce(NO3)3 and the formation of colloidal Ce(OH)3. The solutions
were heated to 90 or 180 °C for 24 h for the wires and cubes,
respectively, without stirring, in a 50-mL Teflon-lined stainless steel
4871 Parr reactor. The pressure reached ∼0.5 bar for the wires and ∼7
bar for the cubes. The resulting products were allowed to cool to room
temperature, dispersed in 10 mL of ethanol, and isolated via
centrifugation. The products were then dispersed in 10 mL of warm
DI water and isolated via centrifugation. This washing procedure was
repeated two more times. The products were dried under vacuum
overnight at 100 °C. The removal of Na impurities is important, since
previous studies have shown that ion impurities can affect the reactivity
of CeO2.33 Na impurities were removed by dispersing the products in
2 mL of 0.1 M NH4OH, which was suspended in an ultrasonic bath for
2 min. The product was isolated via centrifugation, followed by three
warm water washes. The products were then dispersed in 2 mL of 0.1
M HNO3, which was suspended in an ultrasonic bath for 2 min. The
product was isolated via centrifugation, followed by three warm water
washes. The products were dried under vacuum overnight, and then
calcined at 400 °C for 4 h in air. The cubes and wires have surface
areas of 17 and 71.5 m2/g, respectively. SEM and XRD of the cubes
and wires are shown in Figure S1 in the Supporting Information.
Characterization. Cerium oxide samples were characterized by
many techniques. Scanning electron microscopy (SEM) was
performed using a Zeiss Merlin system operating at 5 kV.
Transmission electron microscopy (TEM) was performed using a
Zeiss Libra 120 system operating at 120 kV. Powder X-ray diffraction
(XRD) was performed on a Panalytical X’Pert Pro system, using Cu
Kα radiation. Three-point Brunauer−Emmett−Teller (BET) analysis
of N2 adsorption isotherms was employed to obtain surface area
measurements using a Micromeretics Gemini system.
Temperature-Programmed Desorption (TPD). Acetaldehyde
TPD was performed on a temperature-controlled plug-flow reactor
(Altamira, Model AMI-200) coupled to a downstream gas sampling
quadrupole mass spectrometer (QMS, Pfeiffer-Balzer Omnistar).
Enough catalyst was loaded in a U-shaped quartz tube (4 mm I.D.)
supported by quartz wool to equal a surface area of 1.43 m2 of the
respective samples (i.e., 0.020 g wires, 0.0841 g cubes, 0.143 g
octahedra) and quartz sand was added to make the approximate
volumes equivalent. All samples were pretreated by heating to 400 °C
(10 °C/min) and holding for 1 h in a gas stream of 5% O2/He (30
sccm). The samples were then cooled to room temperature, followed
by purging in a gas stream of He (30 sccm). A gas stream of 0.5%
acetaldehyde/He (Air Liquide, 5000 ppm acetaldehyde) (30 sccm)
was flowed over the sample for 10 min, followed by a gas stream of He
(30 sccm) for 10 min. The temperature was then ramped to 400 °C (3
°C/min) under a gas stream of He (30 sccm). Because the cracking
fragments for acetaldehyde and the products have considerable overlap
in the mass spectrum, the analysis required that QMS mass intensities
be corrected. This correction was performed as described previously,7
by creating a matrix containing the mass fragment intensities for each
product (acetaldehyde, H2, methane, H2O, CO, ethanol, O2, CO2, 1,3-
butadiene, 2-butene, crotyl alcohol, acetone, furan, crotonaldehyde,
benzene, and toluene) contributing to the masses of 2, 15, 18, 28, 29,
31, 32, 44, 54, 56, 57, 58, 68, 70, 78, and 90 and solving the resulting
equation to convert mass intensity to product yield (partial pressure).
The mass fragment intensities were corrected for QMS sensitivity
parameters, including ionization efficiency, quadrupole transmissions,
and multiplier gain, as described previously.24 To determine TPD
yields, each product partial pressure was measured over the entire
temperature (time) range of the TPD experiment for each desorption
product. TPD carbon selectivity was obtained from eq 1, where Yi is
the temperature-integrated partial pressure, ni is the carbon number in
each product, and the sum is over all analyzed desorption products,
including acetaldehyde.
CeO2−x) alters the reactivity and results in the evolution of
different products. Nevertheless, the reactivity of single-
crystalline films is typically studied by temperature-pro-
grammed desorption (TPD) under ultrahigh vacuum (UHV)
conditions and the question remains whether these results can
be extrapolated to practical reactor conditions.
Syntheses have been reported to form CeO2 nanoparticles19
and shape-controlled nanocrystals,20−22 as well as various other
architectures.22−25 Here, nanosized CeO2 in the form of cubes,
wires, and octahedra generated via a hydrothermal method are
used to probe the catalytic adsorption and reaction of
acetaldehyde on well-defined CeO2 surfaces. CeO2 cubes
expose {100} facets and octahedra expose {111} facets. CeO2
{111} facets have been calculated to be the most
thermodynamically stable surface termination.26 CeO2 wires
or rodlike structures are commonly thought to expose a
combination of {110} + {100} facets.21,27,28 However, recent
results suggest that, depending on the annealing treatment,
CeO2 wires and rods could actually expose a variety of facets,
including {111} facets, and possess large numbers of nano-
scopic structural defects.29 Here, the structure of the CeO2
wires will be referred to as a defect structure. CeO2 is known to
possess strong basic sites and weaker acidic sites.6 The
geometry of the surface affects the positions and coordination
numbers of available acid and base sites, thereby modifying the
reactivity. The calculated oxygen-vacancy formation energies
are expected to follow the order of (110) < (100) < (111), and
is anticipated to influence the adsorption motifs for molecules
on their surfaces and redox activity.7,10,21,30−32
Therefore, the adsorption and reaction of acetaldehyde on
CeO2 cubes, wires, and octahedra are expected to exhibit shape-
dependent properties, resulting in diverse reactivity and
selectivity. TPD, diffuse reflectance infrared spectroscopy
(DRIFTS), and isothermal and temperature-programmed
surface reaction (TPSR) methods have been utilized to
determine reaction intermediates, products, and selectivities,
as well as to propose reaction pathways. Importantly, it was
found that crotonaldehyde is the dominant coupling product
for all nanoshapes; however, the selectivity to this product, as
well as the overall conversion of acetaldehyde, differs between
the various nanoshapes. These results contribute to a detailed
fundamental understanding of structure−reactivity relationships
and provide insight into the design and production of advanced
catalysts.
EXPERIMENTAL SECTION
■
Materials. All chemicals were handled in air and used as received.
Ce(NO3)3·6H2O (99%) was purchased from Sigma−Aldrich. NaOH
(≥97%) and HNO3 (68%) was purchased from Fisher Scientific.
NH4OH (28−30%) was purchased from Acros Organics. Deionized
water (18.2 MΩ) was used in all syntheses. Acetaldehyde (0.5% or
5000 ppm) in helium was purchased from Air Liquide America L.P.
CeO2 Octahedra. A quantity of 0.868 g of Ce(NO3)3·6H2O (0.002
mol) was dissolved in 80 mL of H2O. The solution was heated at 190
°C without stirring for 6 h in a 100 mL quartz-lined stainless steel
homemade autoclave reactor, and the pressure reached 14−14.5 bar.
The resulting products were allowed to cool to room temperature,
dispersed in 10 mL of ethanol, and isolated via centrifugation. The
products were then dispersed in 10 mL of warm DI water and isolated
via centrifugation. This washing procedure was repeated two more
times. The octahedra were dried under vacuum overnight and then
calcined at 400 °C for 4 h in air. It should be noted that the product
particles were a combination of discrete and agglomerated octahedra.
The octahedra have a surface area of 10 m2/g. SEM and XRD of the
octahedra are shown in Figure S1 in the Supporting Information.
2438
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
Figure 1. Temperature-programmed desorption mass spectrum profiles showing the characteristic mass fragments of different products resulting
from the room-temperature adsorption of acetaldehyde followed by heating in He for CeO2: (A) cubes, (B) wires, and (C) octahedra. The profiles
are offset vertically and multiplied as indicated for clarity.
Balzer Omnistar). Cerium oxide samples were pretreated by heating to
100 × (Yni)
i
Si (%) =
400 °C (10 °C/min) and holding for 1 h in a gas stream of 5% O2/He
(25 sccm). The samples were then cooled to room temperature, and
then purged in a gas stream of He (25 sccm). A pulse of 0.5%
acetaldehyde/He (sample loop 0.5 mL) was introduced to the sample
and continuous spectra (16 scans each) were collected over 5 min.
Then, a gas stream of 0.5% acetaldehyde/He (25 sccm) was flowed
over the sample for 2 min, followed by a gas stream of He (25 sccm)
for 2 min. The temperature was then ramped to 400 °C (10 °C/min)
and the resulting temperature-programmed DRIFTS spectra (TP-
DRIFTS) were recorded throughout.
∑
products Yjnj
(1)
Temperature-Programmed Surface Reaction (TPSR). The
same reactor, catalyst amounts, and pretreatment conditions as those
for TPD were used for TPSR. A gas stream of 0.5% acetaldehyde/He
(30 sccm) was flowed over the sample for 10 min at room
temperature, and then the temperature was ramped to 400 °C at 3
°C/min. The product gas stream was sampled continuously using the
QMS and the data was corrected for overlaps and QMS sensitivity
using the same methods described above to calculate the temperature
(time)-dependent product yield. Selectivity at each temperature was
obtained from these yields and computed from eq 1, where the sum is
over all components except acetaldehyde.
Conversion during Isothermal Reaction. The same reactor,
catalyst amounts, and pretreatment conditions as those for TPD were
used for isothermal reactions, except that after pretreatment in 5% O2/
He, the samples were cooled in He to a specified reaction temperature.
Then a gas stream of 0.5% acetaldehyde/He (30 sccm) was flowed
over the sample and 1 mL aliquots of the product gas stream was
collected periodically using an automatic sampling valve and analyzed
using a Buck Scientific gas chromatograph (GC) with a 2-m packed
RESULTS
■
Temperature-Programmed Desorption (TPD). The
product profiles resulting during TPD of adsorbed acetaldehyde
on CeO2 nanoshapes are shown in Figure 1. Here, major,
minor, and trace products are classified as products that
comprise >5%, 1%−5%, and <1% of the carbon-containing
product stream normalized to their carbon number over the
entire temperature range, as monitored by mass spectrometry
during TPD, respectively, and are highlighted in Table S1 in the
Supporting Information. Details can be found in the
Experimental Section. For all CeO2 nanoshapes, acetaldehyde
is a major desorption product and is observed between 30 °C
and 140 °C. Desorption of CO2, CO, H2O, and H2 from all
nanoshapes indicates that decomposition is a prevalent
pathway. Water is favored over H2 for the wires, but they are
comparable for the cubes and octahedra. The broad, non-
correlated desorption ranges are indicative of complex stepwise
processes, leading to these desorption products that vary
between the shapes.
1
15% Carbowax 20 M column (2.0 mm O.D., /8-in. I.D.), a thermal
conductivity detector (TCD), and a flame ionization detector (FID).
The product peak areas obtained from the GC were integrated to
determine the yield of each product based on calibration standards.
Acetaldehyde consumption was obtained from the measured
acetaldehyde eluted compared to the inlet acetaldehyde measured
while bypassing the reactor. Product selectivity was determined from
eq 1 where the sum is over all observed eluted species including
acetaldehyde. The carbon capture was deduced from the amount of
acetaldehyde consumed and the carbon-corrected yield of products
evolved at each temperature.
Surface Adsorbates by Fourier Transform IR (FTIR) Spec-
troscopy. FTIR spectroscopy was conducted in a diffuse reflectance
cell (Pike Technologies, Model HC-900, cell volume of ∼6 cm3) in a
Nicolet Nexus 670 FTIR spectrometer using a MCT/A detector with a
spectral resolution of 4 cm−1. A background spectrum was collected
using 256 scans at room temperature after the sample was pretreated.
Diffuse reflectance FTIR spectroscopy (DRIFTS) spectra were
collected by subtracting out the background from the resulting
spectra. Gases leaving the reaction cell were analyzed downstream
using a gas sampling quadrupole mass spectrometer (QMS, Pfeiffer−
Of greater interest are the other carbon carrying products.
Crotonaldehyde, formed by Aldol condensation, is a minor
product that is also desorbed in this temperature region. A trace
amount of furan is observed on the octahedra between 50 °C
and 100 °C. For the wires, a second acetaldehyde peak is
observed between 110 °C and 180 °C. Ethanol, formed via a
Cannizzaro disproportionation reaction shown in Scheme 1,34
is a major product on the wires and a minor product on the
2439
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
Scheme 1. Proposed Mechanism for the Formation of
Acetate and Ethoxide from the Cannizzaro Reaction
Involving a Vacancy and Surface Oxygen for Example on a
a
CeO2{111} Facet (Not to Scale)
a
Green spheres represent Ce atoms, gray spheres represent O atoms,
and white circles denote oxygen vacancies.
cubes below 275 °C, but is only observed as a trace product for
the octahedra. 1,3-Butadiene, which is formed by reductive
coupling, is observed on all nanoshapes, as a trace product on
the octahedra and cubes between 290 °C and 360 °C, and as a
minor product at lower temperatures on the wires between 125
°C and 300 °C. Crotyl alcohol, a coupling product derived
from the reduction of crotonaldehyde, is only evolved in trace
amounts on the wires. In addition, another reductive coupling
product, 2-butene, is only observed in trace amounts, primarily
on the wires above 325 °C. Finally, products formed from the
scission of C−C bonds (CO, CO2, methane, and acetone) are
only observed at higher temperatures. Acetone, which has been
proposed to form from two acetate groups undergoing
ketonization,11 is evolved from the wires and cubes as a
minor product above 275 °C, but as a trace product on the
octahedra. Methane is evolved as a major product above 350 °C
on the cubes and above 300 °C on the wires, but as a minor
product on the octahedra. CO and CO2 are evolved from all
nanoshapes above ∼270 °C as major products.
Summarizing, the various nanoshapes had different selectiv-
ities for the coupling products. All nanoshapes produced the
coupling product crotonaldehyde; however, the wires and cubes
additionally produce 1,3-butadiene and/or acetone as minor
products. Wires and cubes also desorbed acetaldehyde, and its
reduced form, ethanol. In comparison, the octahedra primarily
desorb acetaldehyde as a major product, and only trace
amounts of its reduced form, ethanol. The amount of ethanol
desorbed was greatest on the wires.
Surface Adsorbates via TP-DRIFTS. DRIFTS spectra
obtained after flowing acetaldehyde at room temperature over
the different catalyst nanoshapes, followed by helium flow to
remove gas phase and weakly adsorbed acetaldehyde, are
shown in Figure 2. Table 1 shows the corresponding peak
assignments. All samples exhibited bands at room temperature
in the 3000−2800 cm−1 region and are assigned to symmetric
and asymmetric ν(CH3) and ν(CH2) modes. Bands in the
1720−1704 cm−1 can be assigned to ν(CO) mode of η1-
acetaldehyde.16,35 Notably, the band in this region on the wires
is a very weak shoulder, compared to the strong band observed
on the cubes and the octahedra. Many of the bands observed at
room temperature correlate to species other than acetaldehyde.
All the nanoshapes show bands that can be assigned to red-
shifted ν(CO) and ν(CC) modes that likely correlate to
coupling products that form at room temperature. For example,
the coupling product crotonaldehyde is observed to desorb at
low temperatures in TPD experiments. Coupling products also
contribute to the ν(CH3) and ν(CH2) modes. These ν(CC)
and ν(CO) bands are the dominant surface species on the
wires. Though the small broad band at 1629 cm−1 on the cubes
Figure 2. Room-temperature DRIFTS spectra for CeO2 cubes, wires,
and octahedra after 2 min of acetaldehyde flow, followed by 2 min of
helium flow.
is assigned to coupling products, together with the band at
1311 cm−1, it could also correspond to an enolate species,36
which is not observed on the wires or octahedra at room
temperature. Enolate is expected to be the primary intermediate
to form coupling products via Aldol addition.16
Broad bands at 1579 and 1423 cm−1 on the cubes are
assigned to νas(O−C−O) and νs(O−C−O) of acetate.7,11
However, the room-temperature band on the cubes at 1579
cm−1 is sufficiently broad that it could also correspond to
νas(O−C−O) of a carbonate moiety. It often is difficult to
distinguish between acetate and carbonate groups by FTIR. If
the peak at 1579 cm−1 also correlates to a carbonate species, the
band at 1241 cm−1 can then be assigned to νs(O−C−O). The
split of ∼340 cm−1 indicates that the carbonate is bound in a
bidentate conformation.37 In contrast to the cubes, no acetate
bands are observed on the wires or octahedra at room
temperature. The bands at 1114 and 1060 cm−1 on the cubes,
1094 and 1069 cm−1 on the wires, and the broad band at 1137
cm−1 on the octahedra are assigned to the binding of
monodentate and bidentate ν(C−O) ethoxide species, as
represented in Scheme S1 in the Supporting Information. Idriss
and co-workers identified ethoxy species upon adsorption of
acetaldehyde on CeO2 at room temperature.11 Previous
experiments of ethanol adsorption showed bands at 1050
cm−1 for bidentate and 1120 and 1096 cm−1 for monodentate
species.7
TP-DRIFTS profiles measured as a function of increasing
temperature in helium are shown in Figure 3 for the various
CeO2 nanoshapes. The room-temperature band at 1704 cm−1
on the cubes and wires and 1720 cm−1 on the octahedra
corresponding to η1-acetaldehyde disappears as acetaldehyde
desorbs and the peak is largely gone by 100 °C on the cubes
and wires, and 150 °C on the octahedra. The bands correlating
to acetate species on the cubes become more prominent as the
temperature is increased, indicating that the amount of acetate
on the surface is increasing. Very small acetate bands begin to
grow on the octahedra above 100 °C at 1509 and 1432 cm−1.
Notably, acetate is not observed on the wires, even as the
temperature increases to 350 °C. Acetate groups can form upon
oxidation of acetaldehyde by the CeO2 surface or by a
2440
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
Table 1. Vibrational Assignments for Observed IR Peaks from the Adsorption of Acetaldehyde on CeO2 Nanoshapes
Vibrational Assignments (cm−1
)
a
vibrational mode
acetaldehyde on CeO2
cubes
wires
octahedra
νas(CH3)
νas(CH2)
νs(CH3)
νs(CH2)
3002, 2969, 2965
2964, 2925
2918, 2864
2836
2979
2923
2964
2978
2927
2934
2871
2867
2877
2839 (shoulder)
2745
2837 (shoulder)
N/O
ν(CH) η1-acetaldehyde
ν(CO) η1-acetaldehyde
ν(CO) and (CC) coupling products
νas(OCO) acetate and bidentate carbonate
ν(CC) enolate
2760−2746, 2759
1723
N/O
2734
1704
1704 (weak shoulder)
1720
1656, 1642, 1630
1584, 1611, 1565
1600
1666, 1629
1579 (broad)
1629
1662, 1629
N/O
1660 (shoulder), 1643
1592
N/O
N/O
νs(OCO) acetate
1443, 1429, 1406, 1422
1380, 1365
1423
N/O
1407
δs(CH3)
1371
N/O
N/O
δ(CH2) enolate
1317
1311
N/O
N/O
νs(OCO) bidentate carbonate
ν(C−O) monodentate
ν(C−O) bidentate
1269, 1260
1241
N/O
N/O
1134, 1120, 1096
1050, 1028
1114
1094
1137, 1120
N/O
1060
1069
a
Data taken from refs 7, 16, 38, and 56.
Cannizzaro disproportionation reaction. The carbonate bands
at 1579 and 1241 cm−1 on the cubes grow as temperature
increases, suggesting that the amount of carbonate on the
surface increases. Peaks at 1592 and 1210 cm−1 grow in on the
octahedra above 200 °C and likely correspond to the formation
of bidentate carbonate species. However, no carbonate species
are observed on the wires. Enolate bands that are observed on
the cubes at 1629 and 1311 cm−1 are gone by 200 °C, but do
not obviously correspond to evolution of any product. Instead,
the enolate moiety likely reacts to form some other more-stable
surface species (e.g., acetate).
Finally, bands at ∼2975 and 2870 cm−1, corresponding to
ν(CH3) modes, disappear at high temperatures on all the
nanoshapes as acetaldehyde and coupling products (e.g.,
crotonaldehyde, acetone) are desorbed. However, bands
corresponding to ν(CH2) modes at ∼2925 and 2839 cm−1
are maintained up to 350 °C on the cubes and wires. These are
likely coupling products and coke and indicate that deactivation
of the catalyst is a concern. Higher reaction temperatures may
be able to promote the reaction and subsequent desorption of
these surface species, especially as the oxygen mobility increases
at high temperature; however, heating to temperatures above
450 °C has been shown to induce morphological changes on
the cubes and wires. In general, little remains on the surface of
the octahedra at 350 °C, compared to the other nanoshapes,
indicating that most of the acetaldehyde and its products have
desorbed.
Temperature-Programmed Surface Reaction (TPSR).
TPSR profiles for the various nanoshapes under a flow of 0.5%
acetaldehyde in helium at 30 sccm are shown in Figure 4. In
general, there are two main regions observed in the TPSR.
First, there is the low-temperature region, T < T10, where the
consumption of acetaldehyde is <10%. T10 is ∼375 °C for the
cubes and ∼315 °C for the wires and octahedra, and this
temperature is indicated by a dashed red line in Figure 4.
Primarily, acetaldehyde, disproportionation, and coupling
products are observed at T < T10. Second, at T > T10, the
acetaldehyde consumption increases abruptly, so a larger
amount of products, including C1 products, are observed.
For all nanoshapes, unconverted acetaldehyde is the major
carbon-containing component detected in the product stream
when T < T10. However, there are many differences in the
product distribution and their variation with temperature,
which is dependent on the catalyst morphology. Figure 5
compares the product selectivities of the carbon-containing
products (see the Experimental Section) at 200 °C extracted
from the TPSR data, and similar results for other temperatures
are shown in Figure S2 in the Supporting Information. From
Figure 4, it can be seen that significant amounts of ethanol are
produced on the cubes and wires, and their profiles differ.
However, ethanol is formed beginning at room temperature on
both shapes; the cubes have a maximum formation rate at 180
°C, whereas the wires have multiple maxima, including at 60
The bands for bidentate ethoxide ν(C−O) seen at room
temperature on all the nanoshapes show varying trends. On the
octahedra, the band is small initially and is gone by 150 °C;
however, only trace amounts of ethanol evolve in TPD. In
contrast, for the cubes, the ethoxide bands decrease up to 200
°C, but never fully disappear and correlate to the evolution of
ethanol in the TPD below 225 °C as a minor product. On the
wires, the bands for bidentate ν(C−O) show a small decrease
below 250 °C. In the TPD, ethanol is observed as a major
product on the wires below 250 °C. On the cubes, wires, and
octahedra, stretches associated with ν(CC) and/or ν(CO)
increase sharply up to 200, 250, and 100 °C, respectively, and
could indicate the formation of coupling products. On the
cubes, these bands begin to decrease while 1,3-butadiene is
evolved from 280 °C to 360 °C in TPD. The bands associated
with coupling products on the wires begin to decrease up to
350 °C. In TPD, 1,3-butadiene is evolved between 125 °C and
250 °C, and above 300 °C, 2-butene is evolved on the wires.
1,3-Butadiene is desorbed in TPD on the octahedra from 300
°C to 375 °C. The bands associated with coupling products on
the octahedra begin to decrease by 150 °C. Peaks associated
with coupling products are largely gone on the octahedra and
cubes by 350 °C, but persist on the surface of the wires. On the
cubes and wires, a stretch in the aromatic region appears above
250 °C at 3037 cm−1 for the cubes and above 150 °C at 3020
cm−1 for the wires. While no aromatic compounds are observed
at high temperatures in TPD, this could be because too little is
evolved to be detected. As will be discussed later, in TPSR,
benzene and toluene are observed; this band could be the
surface species relating to these types of products.
2441
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
and 235 °C. The selectivity to form ethanol for both is ∼25% at
200 °C. In contrast, only a small amount of ethanol is evolved
from the octahedra, with a selectivity of <5% at 200 °C.
Crotonaldehyde is evolved from all three nanoshapes, and is
the dominant coupling product detected. Crotonaldehyde
formation starts ∼30 °C higher for the cubes and octahedra
than for the wires. Production is almost constant from 200 °C
up to T > T10 on the cubes, which have ∼55% selectivity to
produce crotonaldehyde at 200 °C (Figure 5). On the wires,
crotonaldehyde production exhibits two maxima near 150 and
275 °C with a selectivity of ∼50% for crotonaldehyde at 200 °C
(Figure 5). On the octahedra, the crotonaldehyde selectivity is
comparable to the cubes and wires near 200 °C, but
dramatically increases at T > T10, compared to the other
shapes (Figure 4). The formation of crotonaldehyde through an
aldol condensation reaction should produce one equivalent of
H2O, and the initial maximum in the water profile correlates
well with the maxima of the crotonaldehyde evolution on each
nanoshape. At T < T10, ethanol and crotonaldehyde are formed
with similar selectivities on the cubes and wires, with 2−4 times
higher selectivity for crotonaldehyde, compared to ethanol.
However, the octahedra have ∼10 times higher selectivity for
crotonaldehyde than ethanol at 200 °C, and ∼25 times higher
selectivity for crotonaldehyde than ethanol at T > T10.
Interestingly, the nanoshapes also produce reduced and/or
deoxygenated coupling products at T < T10 with selectivities of
<5%, namely, crotyl alcohol, 1,3-butadiene, and 2-butene.
Water formation could also occur via dehydration of
acetaldehyde to produce acetylene or, if surface H is available,
ethylene. However, neither ethylene nor acetylene is observed
in TPD or TPSR. These products may not form; or their
intermediates may remain on the surface and contribute to the
deactivation of the catalysts.
When T > T10, there is a greater consumption of
acetaldehyde on all shapes. The amount of acetaldehyde in
the product stream at 425 °C is decreased by ∼25%, 70%, and
10% from its initial concentration for the cubes, wires, and
octahedra, respectively. The cubes and wires have a reduced
selectivity for crotonaldehyde, less ethanol is desorbed, and the
amount of coupling products such as crotyl alcohol and 2-
butene formed in this region increases. In addition, there is
increased evolution of products that are associated with the
scission of C−C bonds, including methane, CO, CO2, and
acetone. The onset of methane evolution begins at T < T10 for
all the nanoshapes and the selectivity to produce methane at T
> T10 on the cubes and wires increases, compared to T < T10.
However, no such increase in selectivity is observed on the
octahedra. The amount of CO2 and CO produced on cubes,
wires, and octahedra is still increasing at 425 °C. H2O, which
can form from dehydration, Aldol condensation, or surface
reduction, also maximizes at 425 °C for all three nanoshapes.
No evolution of O2 is observed on any of the nanoshapes.
Despite significant acetate formation indicated in TP-DRIFTS
(especially on the cubes), neither acetic acid nor formic acid,
possible oxidation products, were observed. Also, no ethylene, a
possible direct dehydration product, was observed for any of
the nanoshapes.
Surprisingly, several cyclic products (≤5% selectivity between
100 and 400 °C) were evolved from the nanoshapes, including
toluene, benzene, and furan (see Figure S2 in the Supporting
Information). Furan is formed at T < T10 on all nanoshapes.
Ring closure of crotonaldehyde to form furan would create one
equivalent of H2. H2 formation is observed throughout the
Figure 3. TP-DRIFTS spectra recorded in He following the room-
temperature adsorption of acetaldehyde on CeO2: (A) cubes, (B)
wires, and (C) octahedra.
2442
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
Figure 4. Temperature-programmed surface reaction (TPSR) profiles under a constant acetaldehyde stream for CeO2 ((A, B) cubes, (C, D) wires,
and (E, F) octahedra). Products are separated by C1−C2 products and C3−C4 products and the MS intensity is scaled as indicated for clarity. The
conversion increases significantly above T10 (red dashed lines, see text). The product distributions in Figure 5 are taken from T = 200 °C, indicated
by the black dashed lines.
by ∼30 °C in the cubes and octahedra compared to the wires.
In contrast, benzene and toluene are only observed at T > T10,
and their evolution is closely correlated to the increase in
formation of 2-butene, 1,3-butadiene, and acetone as the energy
barrier to promote multiple reaction pathways is overcome.
The generation of products as a result of multiple coupling
reactions may result in deactivation of the catalyst. Coupling
products with high molecular weights have low volatility and
may bond strongly to the oxide, poisoning the surface, and will
therefore not be detected by MS. Indeed, at the end of TPSR,
the wire and cubes had significant amounts of oily brown-black
residue on the quartz U-tube and for all catalysts shapes, the
catalyst changed from a whitish-yellow color to a brown color.
Heating the catalysts in O2/He at 400 °C for 1 h caused the
catalyst to revert to its original color. Repeating the TPSR
experiments with the O2/He-treated catalysts resulted in good
reproducibility of the TPSR results. In addition, examination of
the catalysts via SEM after TPSR and O2/He treatment showed
that the catalysts retained their original morphology.
Figure 5. Product selectivities at 200 °C obtained from analysis of the
TPSR profiles (see the Experimental Section) and compared for the
three nanoshapes. Unconverted acetaldehyde constitutes the majority
component of the product stream, and the total acetaldehyde
conversion at 200 °C is indicated.
Isothermal Reaction and Deactivation. In order to
better quantify the reaction selectivity and amount of
entire TPSR temperature range for all the nanoshapes. Similar
to crotonaldehyde production, the evolution of furan is delayed
2443
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
acetaldehyde converted at different temperatures and times-on-
stream, reactions were run under isothermal conditions, and the
reactor eluent was analyzed using GC with FID and TCD. The
sample masses were adjusted to have the same total surface
area. To probe the product selectivity under isothermal
conditions, the product stream was analyzed after 11 min of
continuous flow of acetaldehyde over a fresh catalyst at the
indicated temperatures. The primary product observed in the
eluent was crotonaldehyde for all nanoshapes at 100, 200, and
300 °C (see Table S2 in the Supporting Information), but at
400 °C, the methane, acetone, and carbon oxides become
significant (Table 2).
buildup of carbonaceous residue on the surface results in
catalyst deactivation and complicates the analysis of reaction
kinetics.
The product stream was also sampled at times after 11 min,
but it was found that acetaldehyde conversion dropped rapidly,
indicating the effects of catalyst deactivation. Shown in Figure 7
is the decrease in the consumption (conversion and capture) of
acetaldehyde versus time-on-stream at 400 and 300 °C. The
sample masses were adjusted to have the same total surface
area. At 400 °C, the wires have the highest conversion, but also
a high rate of deactivation with a 74% decrease in activity over
60 min. Cubes have a similarly high rate of deactivation at 400
°C, with a 75% decrease in activity that primarily occurs in the
first 20 min of the reaction. Octahedra have the lowest rate of
deactivation, with a 33% decrease over 60 min. At 300 °C, the
wires and the cubes again have high initial rates of deactivation,
with the acetaldehyde consumption rate decreasing 62% and
86%, respectively, over 60 min. Again, the octahedra have the
lowest deactivation, with the acetaldehyde consumption
decreasing 40% over 60 min. It is expected that co-feeding
O2, CO2, or H2O may decrease the amount of deactivation, and
this will be the subject of future studies.
Table 2. Isothermal Reactivity of Nanoshapes for
Consumption of Acetaldehyde at 400 °C Collected after 11
a
Min, Corrected for the Carbon Number
Selectivity (%) at T = 400 °C
cubes
wires
octahedra
acetaldehyde consumed
crotonaldehyde
methane
60
4
94
8
36
17
<1
<1
<1
<1
2
2
3
acetone
5
15
11
1
DISCUSSION
CO and CO2
ethanol
3
■
TP-DRIFTS data indicate that acetaldehyde adsorbs in the η1-
conformation at room temperature on the cubes and octahedra.
This interaction is relatively weak, and much of acetaldehyde
adsorbed in this way will simply desorb rather than react as the
temperature is increased. Previously, the adsorption and
reaction of acetaldehyde was studied on oxidized and reduced
CeO2−x(100) and CeO2−x(111) thin films using TPD,
reflection absorption IR, soft x-ray photoelectron spectroscopy
and near edge x-ray absorption fine structure, with acetaldehyde
initially being introduced to the catalyst below room temper-
ature.16−18 For these films studied in a UHV environment,
acetaldehyde adsorbed in the η1-configuration on the fully
oxidized (111) films. Upon reduction of the surface, vacancies
on CeO2−x(111) thin films adsorb acetaldehyde to form enolate
species. Without these vacancies, acetaldehyde primarily
desorbs without undergoing reaction and coupling products
are not observed.16 However, on both oxidized and reduced
CeO2−x(100) thin films, acetaldehyde adsorbs in a μ-C,O-
acetaldehyde conformation (dioxyethylene) that dehydrogen-
ates above 100 °C to form acetate and enolate species.
Formation of μ-C,O-acetaldehyde is attributed to lower
coordination of both O and Ce on the fully oxidized (100)
surface.18 The DFT-calculated IR spectrum for μ-C,O-
acetaldehyde predicts a peak at 1002 cm−1 on CeO2(111).16
Such a peak is not clearly observed on the nanoshapes, but
could be slightly shifted and/or overlapping the modes for
another species (ethoxide) on the surface and therefore cannot
be eliminated as a possible bonding motif for acetaldehyde.
The enolate is expected to be a key intermediate for coupling
reactions (Scheme 2).38 Indeed, at room temperature, enolate
peaks are observed on the cubes and possibly on the octahedra
above 200 °C, and coupling products are produced both in
TPD and TPSR for all nanoshapes. Enolate is expected to form
on CeO2 octahedra with {111} facets if vacancies are present,
based on previous reports.16 Surface vacancies could form from
the low-temperature water desorption observed on wires and
octahedra, or may be present at the edges and corners.
1
unassigned
5
5
carbon capture
40
50
15
a
Product selectivities are given as a percentage of the inlet
acetaldehyde. Reaction is based on the equivalent total surface area
of 1.43 m2 for the three ceria samples.
Figure 6 shows the consumption of acetaldehyde versus
temperature for the three shapes. A portion of the inlet
Figure 6. Acetaldehyde consumed after 11 min of constant
acetaldehyde flow versus the temperature for CeO2 wires, cubes, and
octahedra (solid lines). The portion of consumption that leads to
carbonaceous buildup (C Capture) is shown for CeO2 wires, cubes,
and octahedra (dotted lines). The sample masses were adjusted to
have the same total surface area, equal to 1.43 m2.
acetaldehyde is converted to products that desorb, but also a
significant amount is captured on the catalysts. From
measurements of the decrease of acetaldehyde and analysis of
all products, the fraction of acetaldehyde that remains on the
sample is deduced (see the Experimental Section) and is shown
as dashed lines in Figure 6. The amount retained on the surface
is especially significant at 400 °C for all nanoshapes. The
Compared to the cubes, the wires show less enolate and
more coupling products, as indicated by the low-frequency
2444
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
Figure 7. Percent consumption of acetaldehyde by CeO2 cubes, wires, and octahedra over 1 h at (A) 400 °C and (B) 300 °C. The sample masses
were adjusted to have the same total surface area, equal to 1.43 m2.
υ(CO) modes (1629 cm−1). One explanation for the relative
Scheme 2. Proposed Mechanism for the Formation of
Crotonaldehyde from the Adsorption and Reaction of
Acetaldehyde Shown for Example on a CeO2 {111} Facet
lack of enolate is that it may undergo reaction at or below room
temperature and elude detection by FTIR. Adsorption below
room temperature may be required to isolate the enolate.
Indeed, the temperature as well as coverage may have a
significant effect on the reaction of acetaldehyde on CeO2. For
the nanoshapes under reactor environment, coupling can occur
readily upon adsorption of acetaldehyde at room temperature,
even on the {111} faceted octahedra. Therefore, the increased
number of defects, surface coverage, and the adsorption
temperature on CeO2 nanoshapes may contribute to the
observed differences in adsorption and reactivity, compared to
the single-crystalline films observed under UHV conditions.
Upon its formation, enolate can react with a second molecule
of acetaldehyde and undergo an Aldol condensation reaction to
form crotonaldehyde and 1 equiv of H2O (Scheme 2).11,18,35
Notably, this is the dominant coupling product observed in the
product stream in TPSR. However, other coupling products are
evolved from the cubes and wires at T < T10, including 1,3-
butadiene and 2-butene. Formation mechanisms are suggested
in Scheme 3. Based upon the proposed mechanisms, one
reason for the selectivity to crotonaldehyde may be that the
a
(Not to Scale)
a
Green spheres represent Ce atoms, gray spheres represent O atoms,
and white circles denote oxygen vacancies.
Scheme 3. Proposed Mechanism for the Formation of Several Coupling Products Evolved Shown for Example on CeO2{111}
a
Facet (Not to Scale)
a
Green spheres represent Ce atoms, gray spheres represent O atoms, and white circles denote oxygen vacancies.
2445
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
formation of various coupling products must first proceed
through a crotonaldehyde intermediate, which can desorb
rather than undergo further reaction (Scheme 3). In support of
this hypothesis, it is noted that the crotonaldehyde is produced
at a lower onset temperature than the other coupling products,
including, 1,3-butadiene, 2-butene, and crotyl alcohol. Crotyl
alcohol is formed upon the reduction of crotonaldehyde, and it
is only produced above 375 °C on the cubes and above 320 °C
on the wires with <4% selectivity. The small amount of crotyl
alcohol could be related to the lack of available H necessary to
reduce the crotonaldehyde. Another possibility is that crotyl
alcohol dehydration to 1,3-butadiene is rapid (see Step 2 in
Scheme 3).
When T > T10, C−C bond scission becomes more favorable,
resulting in the increased production of various C1 products,
such as CO, CO2, and methane, but also the coupling products
acetone and toluene. Bond energies for a free acetaldehyde
molecule are 96 kcal/mol for the acetyl C−H bond, 177.3 kcal/
mol for the CO bond, 84.4 kcal/mol for the C−C bond, and
99 kcal/mol for the methyl C−H bonds.50 While the C−C
bond is the weakest, the interaction of the acetyl O with the
CeO2 catalyst modifies the electron density and, therefore, the
bond energies. The room-temperature formation of crotonal-
dehyde suggests that it is easier to break the methyl C−H bond,
and the CO bond, than the C−C bond. In comparison, CO2
formation, which requires C−C bond scission, only occurs at
temperatures above 125 °C for all shapes.
Rapid dehydration of crotyl alcohol is a key factor in the
Lebedev and Ostromislensky processes.39−41 In the Lebedev
process, ethanol is used as a starting material to form 1,3-
butadiene. The reaction proceeds through a crotyl alcohol
intermediate that undergoes dehydration to form 1,3-
butadiene.42−44 The Ostromislensky process is similar but
uses acetaldehyde and ethanol to form 1,3-butadiene via a
crotyl alcohol intermediate.40,44 If there is excess ethanol in the
system, it could favor the further reduction of 1,3-butadiene to
form 2-butene and 1 equiv of acetaldehyde, as shown in
Scheme 3. Indeed, butene was found to be a byproduct in the
reaction of ethanol and acetaldehyde to produce 1,3-
butadiene.41 At T < T10, it appears that dehydration of crotyl
alcohol is favored since 1,3-butadiene and 2-butene are evolved
on the wires and cubes, respectively, but not crotyl alcohol.
However, at T > T10, crotyl alcohol desorption becomes more
favorable, compared to dehydration to 1,3-butadiene on the
cubes and wires. It should be noted that crotyl alcohol and 1,3-
butadiene are only observed when ethanol is present. On the
octahedra, the only coupling product observed in TPSR is
crotonaldehyde. According to Scheme 3, the formation of
crotyl alcohol (and, hence, 1,3-butadiene) is limited by ethanol,
which is needed for the reduction of crotonaldehyde.
Ethanol is observed in amounts that vary for cubes, wires,
and octahedra. Barteau et al. reported that ethanol could be
produced on TiO2(001) surfaces upon adsorption of
acetaldehyde by reacting with surface H, which is a direct
reduction.45 However, Mullins et al. reported that no ethanol
was observed on CeO2(100) single-crystalline thin films upon
adsorption of acetaldehyde, even upon dissociative adsorption
of water on the surface.18 A Cannizzaro disproportionation of
two acetaldehyde molecules would result in the formation of
ethoxy and acetate groups (Scheme 1). In this base-catalyzed
reaction, surface O can perform a nucleophilic attack on the
carbonyl of an aldehyde molecule, followed by a hydride
transfer to a second molecule of adsorbed aldehyde to form the
alkoxy. This pathway has been proposed for alcohol formation
from formaldehyde adsorption on CeO2−x(111),46 MgO, and
TiO2(001) films,47−49 and acetaldehyde adsorption on rutile
TiO2.34 Alkoxy can combine with hydrogen from a surface-
bound hydroxyl to desorb as the alcohol. We propose that this
disproportionation is the primary route to ethanol on all
shapes. Within this mechanism, the dependence on surface
structure in the production of ethanol would be related to the
base strength of the surface oxygen that performs the
nucleophilic attack.48 Higher ethanol production (and surface
acetate) on cubes compared to octahedra implies a more basic
character in (100) than (111). Therefore, the basicity of the
surface is a significant factor in the structure-dependent
selectivity.
It has been suggested that the production of acetone occurs
via ketonization of two acetate groups.11 Therefore, it is unique
among the coupling products in that it would not be expected
to proceed through an enolate intermediate. The acetate groups
can form via oxidation of acetaldehyde or a Cannizzaro
disproportionation reaction. Subsequently, the acetates can
undergo ketonization to form acetone and 1 equiv of CO2,
formally leaving a surface O atom. This O atom may assume a
lattice site or, if surface H is available, the O atom could form
OH or desorb as H2O. Indeed, water evolution increases
concurrently with acetone formation. The ability to promote
the oxidation of acetaldehyde to produce acetate species is
related to the basicity of the surface. A higher basicity has been
reported to correlate to a more ionic surface, which would be
able to donate electron density from surface O to the α-carbon
of acetaldehyde, therefore oxidizing the acetaldehyde and
leading to acetate groups.12 The cubes displayed bands
associated with acetate groups (and the octahedra, to a lesser
extent) in the TP-DRIFTS. However, the wires produce a large
amount of acetone in the TPD, compared to the small amounts
observed on the cubes and none on the octahedra. In the
integrated TPD, TPSR, and isothermal reactions at higher
temperatures, acetone production is ranked in the following
order: wires > cubes ≫ octahedra. Because acetate species are
not observed on the wires by DRIFTS, yet acetone is clearly
produced, this result could suggest that acetate is consumed
rapidly upon formation. The primary reason for the lack of
acetone product from octahedra is attributed to the relatively
low surface coverage of acetate caused by the lower base
strength predicted for the {111} surfaces.
In addition to undergoing ketonization to form acetone,
acetate groups can also undergo C−C bond scission to form
the C1 products methane, CO2, and CO, depending on the
selectivity between hydrogenation or dehydrogenation path-
ways and the amount of accessible surface O and H. Notably, at
400 °C in TPSR, the nanoshapes display different selectivities
to form these products. The CO2:CO selectivity ratio in TPSR
is ∼2, 5, and 10, for the octahedra, cubes, and wires,
respectively (see Figure S2 in the Supporting Information).
This result suggests that it is easier to remove oxygen from the
surface of the wires and cubes than the octahedra, and is likely
related to the lowest oxygen vacancy formation energies for the
{111} surface. At T > T10, the C−C bond scission, ketonization,
and coupling occur concurrently, as indicated by the
coevolution of C1 products, acetone, and crotonaldehyde,
respectively, which is clearly seen in TPSR (Figure 4) and
isothermal experiments (see Table S2 in the Supporting
Information). However, the distribution of these products is
very shape-sensitive. Roughly, octahedra produce more
2446
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
crotonaldehyde, compared to C1 and acetone; cubes produce
more C1 and crotonaldehyde, compared to acetone; wires
produce more CO2 and acetone, compared to crotonaldehyde
(see Figure S2 and Table S2 in the Supporting Information).
Cyclic products such as furan, benzene, and toluene are also
observed. Furan has been observed previously from reaction of
acetaldehyde on β-UO3.51 Furan selectivity is low for all
nanoshapes (<5%) and its evolution is expected to occur from
the dehydrogenation of crotonaldehyde (Scheme 3). At low
temperature T in TPSR, its evolution coincides approximately
with the onset of crotonaldehyde on all nanoshapes, but at T >
T10, it increases at the expense of crotonaldehyde on the wires.
Benzene and toluene are primarily observed when T > T10 (see
Figure S3 in the Supporting Information). Benzene formation
could occur via the addition of an acetaldehyde molecule to
crotonaldehyde, forming 2 equiv of water and 1 equiv of
benzene (Scheme 3). Benzene has been observed as a product
on TiO2, Al2O3, UO2 (111), and CeO2 previously, and been
proposed to form via this mechanism.35,52 Toluene has not
previously been reported as a product on CeO2 from the
catalysis of acetaldehyde, however, it could be formed from the
addition of a methyl group to benzene, in a mechanism similar
to that of acetone.11 The polymerization of acetaldehyde to
form high-molecular-weight species that do not easily desorb
could contribute to the deactivation of the catalyst. Indeed,
catalyst deactivation is a significant problem in the Lebedev and
Ostromislensky processes, with the amount of polymerization
increasing as the amount of acetaldehyde used in the reaction
was increased.41 Notably, a large amount of carbonaceous
material is calculated to be retained on the surfaces of the
various products (see Table 2 and Figure 6) and likely
contributes to the deactivation that is observed for the various
nanoshapes (Figure 7).
Differences between the nanoshapes in selectivity, acetalde-
hyde consumption, and deactivation were also observed
utilizing isothermal reactions. For example, at 400 °C (Table
2 and Figure 6), wires give the highest consumption rates of
acetaldehyde (94%). Yet, the wires produce a low ratio (1:4) of
crotonaldehyde, relative to the sum of the other products,
including acetone, ethanol, CO, and CO2. In contrast, the
octahedra produce a higher ratio of crotonaldehyde to the sum
of the other products (4:1). This higher selectivity to produce
crotonaldehyde over other products is countered by having
lower acetaldehyde consumption (36%). Finally, the cubes
produce a ratio of crotonaldehyde to the sum of the other
products of 1:4, similar to wires, but the amount of
acetaldehyde consumed is 60%. Significantly, the ethanol
selectivity was lower in isothermal experiments compared to
TPSR experiments. These differences in the product selectivity
could be related to factors such as different reaction times and
temperatures of initial exposure for the respective experiments
that result in variations in the amounts and types of surface
intermediates and carbonaceous buildup. Thus, the reaction
conditions and time-on-stream are significant factors in
determining the selectivity.
surfaces have a theoretical coordination number of 6, while Ce
atoms on {111} surfaces have a theoretical coordination
number of 7. O atoms on {111} and {110} surfaces have a
theoretical coordination number of 3, while O atoms on {100}
surfaces have a theoretical coordination number of 2. A lower
coordination number would be expected to correlate to a
higher reactivity and overall mirrors the higher reactivity
observed in the cubes with {100} facets, compared to octahedra
with {111} facets.53 In addition, it was previously reported that
CeO2 rods and cubes had greater surface O-vacancy densities,
compared to CeO2 octahedra, and that CeO2 rods had more
clustered defects than CeO2 cubes or octahedra.54 Defects are
high-energy sites and can be responsible for higher or unique
reactivity, although they are minority sites on the cubes and
octahedra. In particular, O-vacancy sites have been shown to be
the active sites for dissociative adsorption of methanol and
17,55
acetaldehyde on CeO2−x
.
Therefore, the conversion of
acetaldehyde on CeO2 is likely enhanced by clusters of
vacancies and defects and promotes the formation of enolate
species upon acetaldehyde adsorption.
CONCLUSIONS
■
Herein, shape-controlled CeO2 nanoshapes with well-defined
surface crystallographic orientations were used to evaluate the
structure-dependent properties for the adsorption and reaction
of acetaldehyde. Rich catalytic chemistry is observed, including
coupling reactions (to crotonaldehyde, crotyl alcohol, 1,3-
butadiene, 2-butene), ketonization (to acetone), Cannizzaro
disproportionation (to ethanol and acetate), ring closure (to
furan and benzene), and C−C bond scission (to methane and
CO2 or CO). Differences are observed in the reaction
selectivity, surface species, and desorption products, among
the different nanoshapes. These differences are attributed to
structure-related variation in defect density, surface oxygen
coordination, vacancy formation energy, and acid/base proper-
ties. Although rapid deposition of carbonaceous material on the
surface occurs, thereby altering the surface chemistry, the
following trends in structure−reactivity are observed:
• Reactive coupling of acetaldehyde to produce crotonalde-
hyde occurs by Aldol condensation on all nanoshapes, but with
higher selectivity on the octahedra at 400 °C in TPSR and
isothermal reactions.
• Enolate is observed by DRIFTS only on the cubes, but it is
the believed to be a short-lived intermediate in the Aldol
condensation to crotonaldehyde observed on all shapes.
• Ethanol is formed by a Cannizzaro disproportionation in
TPD and TPSR leaving acetate; wires ≈ cubes ≫ octahedra.
Higher base strength of defects and {100} surfaces leads to the
activation of acetaldehyde to initiate this reaction.
• Similarly, decreased acetone production from octahedra
compared to cubes and wires in TPD, TPSR, and isothermal
reaction is attributed to the relative lack of surface acetate
caused by the lower base strength predicted for the {111}
surfaces, compared to {100} and defects.
• C−C bond scission of acetate leads to correlated evolution
of CO, CO2, and methane above 400 °C in TPSR on all shapes,
with a selectivity at 400 °C that ranks in the following order:
wires ≈ cubes ≫ octahedra. This ordering is attributed to lower
coverage of acetate on the octahedra.
• Desorption of crotonaldehyde from the less-reactive {111}
facets of octahedra, in comparison with the increased activation
of adsorbates on defects and active {100} surfaces, may explain
The differences in how acetaldehyde adsorbs and reacts are
likely to be related to the basicity of the different surfaces, since
more basic surface oxygen groups would favor ethanol (and
acetate) formation. However, other parameters likely contribute
to the observed reactivity, including the coordination numbers
of surface Ce and O species, as well as the varying defect nature
of the nanoshapes. The theoretical coordination numbers vary
depending on the surface facets. Ce atoms on {100} and {110}
2447
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448

ACS Catalysis
Research Article
(19) Feng, X.; Sayle, D. C.; Wang, Z. L.; Paras, M. S.; Santora, B.;
Sutorik, A. C.; Sayle, T. X. T.; Yang, Y.; Ding, Y.; Wang, X.; Her, Y.-S.
Science 2006, 312, 1504−1508.
the lower rates of coking on the octahedra, compared to wires
and cubes.
̈
(20) Kepenekci̧ , O.; Emirdag-Eanes, M.; Demir, M. M. J. Nanosci.
ASSOCIATED CONTENT
■
Nanotechnol. 2011, 11, 3565−3577.
(21) Tana; Zhang, M.; Li, J.; Li, H.; Li, Y.; Shen, W. Catal. Today
2009, 148, 179−183.
(22) Yan, L.; Yu, R.; Chen, J.; Xing, X. Cryst. Growth Des. 2008, 8,
S
* Supporting Information
Scheme for monodentate and bidentate ethoxide; SEM, TEM,
and XRD of synthesized CeO2 nanoshapes; percent selectivities
for TPSR, TPSR-MS profiles for cyclic products; percent
composition of C-containing product stream for TPD by MS;
isothermal selectivity for the nanoshapes by GC. This material
is available free of charge via the Internet at http://pubs.acs.org.
1474−1477.
(23) Guo, Z.; Du, F.; Li, G.; Cui, Z. Inorg. Chem. 2006, 45, 4167−
4169.
(24) Sun, C.; Li, H.; Chen, L. Energy Environ. Sci. 2012, 5, 8475−
8505.
(25) Sun, C.; Li, H.; Chen, L. J. Phys. Chem. Solids 2007, 68, 1785−
1790.
AUTHOR INFORMATION
■
(26) Baudin, M.; Woj
51−61.
(27) Mai, H.-X.; Sun, L.-D.; Zhang, Y.-W.; Si, R.; Feng, W.; Zhang,
H.-P.; Liu, H.-C.; Yan, C.-H. J. Phys. Chem. B 2005, 109, 24380−
24385.
(28) Zhou, K.; Wang, X.; Sun, X.; Peng, Q.; Li, Y. J. Catal. 2005, 229,
206−212.
(29) Florea, I.; Feral-Martin, C.; Majimel, J.; Ihiawakrim, D.;
Hirlimann, C.; Ersen, O. Cryst. Growth Des. 2013, 13, 1110−1121.
(30) Paier, J.; Penschke, C.; Sauer, J. Chem. Rev. 2013, 113, 3949−
3985.
́
cik, M.; Hermansson, K. Surf. Sci. 2000, 468,
Corresponding Author
*E-mail: overburysh@ornl.gov.
Present Address
†Fritz Haber Institute Max Planck, Faradayweg 4−6, 14195
Berlin, Germany.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
(31) Nolan, M.; Parker, S. C.; Watson, G. W. Surf. Sci. 2005, 595,
223−232.
■
This research is sponsored by the Division of Chemical
Sciences, Geosciences, and Biosciences, Office of Basic Energy
Sciences, U.S. Department of Energy. Part of the work
including synthesis, XRD, TEM, and SEM was conducted at
the Center for Nanophase Materials Sciences, which is
sponsored at Oak Ridge National Laboratory by the Scientific
User Facilities Division, Office of Basic Energy Science, U.S.
Department of Energy.
(32) Wu, Z.; Li, M.; Mullins, D. R.; Overbury, S. H. ACS Catal. 2012,
2, 2224−2234.
(33) Gines, M. J. L.; Iglesia, E. J. Catal. 1998, 176, 155−172.
(34) Rekoske, J. E.; Barteau, M. A. Langmuir 1999, 15, 2061−2070.
́
(35) Rasko, J.; Kiss, J. Appl. Catal., A 2005, 287, 252−260.
(36) Takahashi, A.; Haneda, M.; Fujitani, T.; Hamada, H. J. Mol.
Catal. A: Chem. 2007, 261, 6−11.
(37) Busca, G.; Lorenzelli, V. Mater. Chem. 1982, 7, 89−126.
(38) Singh, M.; Zhou, N.; Paul, D. K.; Klabunde, K. J. J. Catal. 2008,
260, 371−379.
REFERENCES
■
(39) Makshina, E. V.; Janssens, W.; Sels, B. F.; Jacobs, P. A. Catal.
Today 2012, 198, 338−344.
(1) Kasp
285−298.
̌
ar, J.; Fornasiero, P.; Graziani, M. Catal. Today 1999, 50,
(40) Quattlebaum, W. M.; Toussaint, W. J.; Dunn, J. T. J. Am. Chem.
Soc. 1947, 69, 593−599.
(2) Shido, T.; Iwasawa, Y. J. Catal. 1992, 136, 493−503.
(3) Jasinski, P.; Suzuki, T.; Anderson, H. U. Sens. Actuators, B 2003,
95, 73−77.
(41) Toussaint, W. J.; Dunn, J. T.; Jachson, D. R. Ind. Eng. Chem.
1947, 39, 120−125.
(4) Beie, H. J.; Gnorich, A. Sens. Actuators, B 1991, 4, 393−399.
̈
(42) Lebedev, S. V. J. Gen. Chem. 1933, 3, 698−708.
(43) Lebedev, S. V. British Patent 331402, 1929.
(44) Angelici, C.; Weckhuysen, B. M.; Bruijnincx, P. C. A.
ChemSusChem 2013, 6, 1595−1614.
(5) Park, S.; Vohs, J. M.; Gorte, R. J. Nature 2000, 404, 265−267.
(6) Vivier, L.; Duprez, D. ChemSusChem 2010, 3, 654−678.
(7) Li, M.; Wu, Z.; Overbury, S. H. J. Catal. 2013, 306, 164−176.
(8) Senanayake, S. D.; Mullins, D. R. J. Phys. Chem. C 2008, 112,
9744−9752.
(45) Idriss, H.; Barteau, M. A. Catal. Lett. 1996, 40, 147−153.
(46) Zhou, J.; Mullins, D. R. Surf. Sci. 2006, 600, 1540−1546.
(47) Barteau, M. A. Chem. Rev. 1996, 96, 1413−1430.
(48) Peng, X. D.; Barteau, M. A. Langmuir 1989, 5, 1051−1056.
(49) Idriss, H.; Kim, K. S.; Barteau, M. A. Surf. Sci. 1992, 262, 113−
127.
(9) Lykhach, Y.; Happel, M.; Johan
́ ́
ek, V.; Skala, T.; Kollhoff, F.;
Tsud, N.; Dvora , F.; Prince, K. C.; Matolín, V.; Libuda, J. J. Phys.
̌ ́
k
Chem. C 2013, 117, 12483−12494.
(10) Wu, Z.; Li, M.; Overbury, S. H. J. Catal. 2012, 285, 61−73.
(11) Idriss, H.; Diagne, C.; Hindermann, J. P.; Kiennemann, A.;
Barteau, M. A. J. Catal. 1995, 155, 219−237.
(12) Kydd, R.; Teoh, W. Y.; Scott, J.; Ferri, D.; Amal, R.
ChemCatChem. 2009, 1, 286−294.
(50) Sanderson, R. Chemical Bonds and Bonds Energy; Academic
Press: London, 2012; pp 200−201.
(51) Madhavaram, H.; Idriss, H. Catal. Today 2000, 63, 309−315.
(52) Chong, S. V.; Idriss, H. J. Vac. Sci. Technol., A 2001, 19, 1933−
1937.
(13) Mavrikakis, M.; Barteau, M. A. J. Mol. Catal. A: Chem. 1998,
131, 135−147.
(14) Ren, H.; Yu, W.; Salciccioli, M.; Chen, Y.; Huang, Y.; Xiong, K.;
Vlachos, D. G.; Chen, J. G. ChemSusChem 2013, 6, 798−801.
(15) Alonso, D. M.; Bond, J. Q.; Dumesic, J. A. Green Chem. 2010,
12, 1493−1513.
(16) Calaza, F. C.; Xu, Y.; Mullins, D. R.; Overbury, S. H. J. Am.
Chem. Soc. 2012, 134, 18034−18045.
(17) Chen, T. L.; Mullins, D. R. J. Phys. Chem. C 2011, 115, 3385−
3392.
(18) Mullins, D. R.; Albrecht, P. M. J. Phys. Chem. C 2013, 117,
14692−14700.
(53) Alghamdi, K.; Hargreaves, J. S. J.; Jackson, S. D. Base Catalysis
with Metal Oxides. In Metal Oxide Catalysis; Jackson, S. D.,
Hargreaves, J. S. J., Eds.; Wiley−VCH: Weinheim, Germany, 2008;
pp 819−843.
(54) Wu, Z.; Li, M.; Howe, J.; Meyer, H. M.; Overbury, S. H.
Langmuir 2010, 26, 16595−16606.
(55) Namai, Y.; Fukui, K.-i.; Iwasawa, Y. Nanotechnology 2004, 15,
S49.
(56) Li, C.; Sakata, Y.; Arai, T.; Domen, K.; Maruya, K.-i.; Onishi, T.
J. Chem. Soc., Faraday Trans. 1 1989, 85, 929−943.
2448
dx.doi.org/10.1021/cs500611g | ACS Catal. 2014, 4, 2437−2448