Kinetics and mechanism of ethanol dehydration on γ-Al 2O3: The critical role of dimer inhibition

Article abstract of DOI:10.1021/cs400051k

Steady state, isotopic, and chemical transient studies of ethanol dehydration on γ-alumina show unimolecular and bimolecular dehydration reactions of ethanol are reversibly inhibited by the formation of ethanol-water dimers at 488 K. Measured rates of ethylene synthesis are independent of ethanol pressure (1.9-7.0 kPa) but decrease with increasing water pressure (0.4-2.2 kPa), reflecting the competitive adsorption of ethanol-water dimers with ethanol monomers; while diethyl ether formation rates have a positive, less than first order dependence on ethanol pressure (0.9-4.7 kPa) and also decrease with water pressure (0.6-2.2 kPa), signifying a competition for active sites between ethanol-water dimers and ethanol dimers. Pyridine inhibits the rate of ethylene and diethyl ether formation to different extents verifying the existence of acidic and nonequivalent active sites for the dehydration reactions. A primary kinetic isotope effect does not occur for diethyl ether synthesis from deuterated ethanol and only occurs for ethylene synthesis when the β-proton is deuterated; demonstrating olefin synthesis is kinetically limited by either the cleavage of a C_β-H bond or the desorption of water on the γ-alumina surface and ether synthesis is limited by the cleavage of either the C-O bond of the alcohol molecule or the Al-O bond of a surface bound ethoxide species. These observations are consistent with a mechanism inhibited by the formation of dimer species. The proposed model rigorously describes the observed kinetics at this temperature and highlights the fundamental differences between the Lewis acidic γ-alumina and Bronsted acidic zeolite catalysts.

Full text of DOI:10.1021/cs400051k

Research Article
pubs.acs.org/acscatalysis
Kinetics and Mechanism of Ethanol Dehydration on γ‑Al O : The
2
3
Critical Role of Dimer Inhibition
Joseph F. DeWilde,†,§ Hsu Chiang, Daniel A. Hickman, Christopher R. Ho, and Aditya Bhan*
†,§
‡
†
,†
†Department of Chemical Engineering and Materials Science, University of Minnesota-Twin Cities, 421 Washington Avenue SE,
Minneapolis, Minnesota 55455, United States
‡
Engineering and Process Science, The Dow Chemical Company, 1776 Building, Office A-17, Midland, Michigan 48674, United
States
*
S Supporting Information
ABSTRACT: Steady state, isotopic, and chemical transient
studies of ethanol dehydration on γ-alumina show unimolecular
and bimolecular dehydration reactions of ethanol are reversibly
inhibited by the formation of ethanol−water dimers at 488 K.
Measured rates of ethylene synthesis are independent of
ethanol pressure (1.9−7.0 kPa) but decrease with increasing
water pressure (0.4−2.2 kPa), reflecting the competitive
adsorption of ethanol−water dimers with ethanol monomers;
while diethyl ether formation rates have a positive, less than
first order dependence on ethanol pressure (0.9−4.7 kPa) and
also decrease with water pressure (0.6−2.2 kPa), signifying a
competition for active sites between ethanol−water dimers and
ethanol dimers. Pyridine inhibits the rate of ethylene and
diethyl ether formation to different extents verifying the
existence of acidic and nonequivalent active sites for the dehydration reactions. A primary kinetic isotope effect does not
occur for diethyl ether synthesis from deuterated ethanol and only occurs for ethylene synthesis when the β-proton is deuterated;
demonstrating olefin synthesis is kinetically limited by either the cleavage of a C -H bond or the desorption of water on the γ-
β
alumina surface and ether synthesis is limited by the cleavage of either the C−O bond of the alcohol molecule or the Al−O bond
of a surface bound ethoxide species. These observations are consistent with a mechanism inhibited by the formation of dimer
species. The proposed model rigorously describes the observed kinetics at this temperature and highlights the fundamental
differences between the Lewis acidic γ-alumina and Brønsted acidic zeolite catalysts.
KEYWORDS: ethanol dehydration, γ-alumina, parallel reactions, surface dimer species, ethylene, diethyl ether, Lewis acid sites
1
. INTRODUCTION
leading the authors to conclude that surface hydroxyl groups
are unable to protonate pyridine and, by extension, the less
basic alcohol molecules, thereby demonstrating that acid sites
required for alcohol dehydration over γ-Al O are likely to be
Gamma alumina (γ-Al O ) is an industrially relevant solid
catalyst because of its high surface area (50−300 m g ) and
thermal stability up to 873 K.
Al O to form either olefins or ethers through unimolecular or
bimolecular pathways, respectively; both pathways result in the
production of water as a byproduct. The dehydration of
alcohols over γ-Al O , thus, can serve as a probe to elucidate
the catalytic site requirements and molecular reaction
mechanisms on this important material.
2
3
2
−1
1
−5
2
3
Alcohols dehydrate over γ-
Lewis acidic rather than Brønsted acidic. Upon evacuation of
2
3
11
−1
the sample at 423 K, Parry observed the band at 1450 cm
6,7
split into two distinct bands verifying the existence of multiple
Lewis acid sites on γ-Al O . The conclusion that γ-Al O is not
2
3
2
3
2
3
15
Brønsted acidic is further supported by N Nuclear Magnetic
Resonance (NMR) measurements on pyridine exposed γ-Al O
2
3
14
performed by Ripmeester in which the pyridinium ion (peak
The surface of γ-Al O contains hydroxyl groups, Lewis acid
2
3
15
at 170 ppm) was also not observed. Additionally, Kwak et al.
sites in the form of surface aluminum (Al) atoms, as well as
surface oxygen (O) atoms capable of behaving as basic
sites.
titrate both Brønsted and Lewis acid sites on catalytic
surfaces.
observed a near 1:1 correlation with the decrease in the
27
8
−10
intensity of the Al NMR peak attributed to surface penta-
coordinated aluminum atoms (∼23 ppm) and the BaO loading
of BaO loaded γ-Al O samples, indicating that BaO acts as a
Pyridine is often used a basic probe molecule to
1
1−13
The infrared spectroscopic band associated with
2
3
the presence of pyridinium ions on the surface of pyridine-
−1
exposed γ-Al O samples (1540 cm ) was not observed by
Received: January 23, 2013
Revised: March 11, 2013
Published: March 26, 2013
2
3
1
1
Parry at ambient temperatures while a band associated with
−1
pyridine bonded to Lewis acid sites (1450 cm ) was observed,
©
2013 American Chemical Society
798
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
selective titrant of penta-coordinated alumina atoms. Further-
more, Kwak et al. observed the dehydration rates of methanol
pressure (in the range of 1.1 to 3.1 kPa) between 373 and 433
K. A mechanism consistent with the observed inhibitory kinetic
effects of water and the pressure dependence of both ether and
olefin synthesis rates on alcohol pressure has not been reported
in the literature.
16
into dimethyl ether on BaO/γ-Al O at 573 K decreased with
2
3
BaO loading, suggesting that Lewis acidic penta-coordinated
surface alumina atoms play an important role in the synthesis of
ethers from alcohol dehydration on γ-Al O .
2
3
In this work, steady state kinetic measurements demonstrate
ethylene and diethyl ether (DEE) synthesis from the
17
Pines and Haag found that the cis/trans ratio among 2-
butene isomers produced from 2-butanol dehydration (cis/
trans = 4.3) was nearly equivalent to that from 1-butene
double-bond isomerization (cis/trans = 4.4) over η-alumina at
dehydration of ethanol over γ-Al O3 at 488 K are both
2
inhibited by the formation of ethanol−water dimer species at
ethanol pressures between 1.9 and 7.0 kPa and water partial
pressures between 0.4 and 2.2 kPa. Water was determined to
irreversibly poison a fraction of the active sites for ethylene and
DEE formation. Pyridine was found to reversibly inhibit the
rate of both ethylene and DEE synthesis to different degrees,
indicating the active sites are acidic in nature and are not
identical for the two dehydration products. Kinetic mechanisms
consistent with these conclusions are presented and evaluated.
5
23 K; on this basis the authors concluded that the two
reactions occur through the same intermediate. The authors
proposed a proton olefin complex on the surface formed from
the decomposition of a surface bound oxonium ion as the
intermediate for olefin formation to explain the high cis/trans
1
8
̈
ratios observed. Knozinger and Scheglila alternatively
concluded olefin formation occurs through an E-2 type
elimination of the alcohol rather than through a proton-olefin
complex intermediate based upon the measured kinetic isotope
effects (KIE) for the dehydration of deuterated tert-butanol, sec-
butanol, and iso-butanol into butenes between 393 and 483 K.
2
. MATERIALS AND METHODS
.1. Catalyst Preparation. γ-Al O (Alfa Aesar, BET
2
2
3
19
2
−1
3
−1
surface area = 155 m /g , pore volume = 0.257 cm g ) was
treated for 3 h in a 1 M NH NO solution at 353 K prior to its
use in kinetic experiments as previously described by Roy et
al. Catalyst particle sizes between 180 and 425 μm (40−80
mesh) were obtained by pressing and sieving the γ-Al O
3
̈
Knozinger et al. proposed the alcohol undergoes this
elimination across a surface hydroxyl group and a basic center.
An olefin in absence of water, however, would be unable to
reform the proposed reaction intermediate; the active sites,
therefore, would not be appropriate for olefin double bond
4
3
24
2
isomerization. Thus the mechanism proposed by Kno
̈
zinger et
powder. Acid-washed quartz sand (0.5−0.7 g, 152−422 μm
particle size, Acros Organics) was mixed with the catalyst
samples (0.02−0.2 g) to form the reactor bed. The catalyst was
1
9
al. is unable to fully explain the same cis/trans ratios for
dehydration and double bond isomerization measured by Pines
1
7
3
−1
and Haag.
Steady state kinetic measurements performed by Kno
et al. on the rate of olefin formation from the dehydration of
then treated in dry air (1.67 cm s at NTP conditions, Ind.
Grade, Matheson Trigas) for 4 h at 723 K after heating the
̈
zinger
−1
catalyst from ambient conditions with a rate of 0.0167 K s .
2
0
3 −1
cyclohexanol (10−33 kPa at 433 and 453 K) and the rate of
The catalyst samples were then cooled in dry air (1.67 cm s )
to the reaction temperature (488 K). The regeneration of the
catalyst after kinetic experiments was also achieved using the
same treatment in air.
21
ether formation from the dehydration of methanol (7−35 kPa
methanol pressures at 433−468 K) over γ-Al O with varying
2
3
alcohol and water partial pressures showed that the rates of
olefin and ether synthesis were inhibited by water and increased
with alcohol pressure before becoming independent of alcohol
pressure. The empirical rate expression shown in eq 1 was
proposed by the authors to fit the resulting data for both
reactions.
2.2. Steady State Kinetic Measurements of Ethanol
Dehydration over γ-Al O . The rate of ethanol dehydration
2
3
was measured using a quartz tube packed bed reactor (10 mm
inner diameter, 1.6 cm3 bed volume) system. The bed
temperature was measured with a type K thermocouple located
on the external surface of the reactor and maintained at
reaction temperature (488 K) using a tube furnace (National
Electric Furnace FA120 type) and a Watlow temperature
controller (96 series).
PA
r = r0
PA + bP
W
(1)
Ethanol dehydration was carried out at 488 K and ambient
pressure under a carrier gas consisting of He (1.7−3.2 cm s
at NTP conditions, grade 4.7, Minneapolis Oxygen Company)
r is the rate of olefin or ether formation, r is the rate of olefin
or ether formation at the zero order in the alcohol pressure
0
3
−1
limit, P and P are the partial pressures of alcohol and water,
A
W
and an internal standard mixture for analysis (25.0% CH and
respectively, and b is an empirical constant. The square root
dependence on ethanol pressure of the proposed expression,
however, implies the dissociation of the alcohol molecule into
two equivalent surface species, which is not consistent with
their proposed mechanisms for dehydration. Conversely, Shi
4
3
−1
balance Ar, 0.017 cm s at NTP conditions, Minneapolis
oxygen) and under differential reaction conditions (<10%
conversion). The catalyst was exposed to 2.2 kPa of deionized
water diluted with the carrier gas (1.7 cm s ) for 1 h at 488 K
prior to reaction.
3
−1
2
2
and Davis observed, upon dehydration of 2-butanol and
methanol at a constant total alcohol pressure at 503 K, the
selectivity of di-2-butyl and dimethyl ether increased propor-
tionally with the square of the partial pressure of 2-butanol and
Liquid pyridine (99+%, Sigma Aldrich), C
Decon Laboratories, Inc.), C OD (99.5 at.% D, Sigma-
Aldrich), C OD (99.5 at.% D, Sigma-Aldrich), and deionized
H OH (99.5%,
2
5
H
5
2
D
5
2
23
methanol, respectively. Alternatively, De Morgues et al.
water were fed into the carrier gas stream at 405 K via syringe
pumps (KD scientific KDS −100 and Cole Parmer). Feed
partial pressures (0.0−0.3 kPa pyridine, 0.9−7.0 kPa C H OH,
proposed that olefin formation occurred over two sites, one of
which is inhibited by water surface species based upon their
observation that the rate of propene synthesis from 2-propanol
dehydration was inversely proportional to water pressure (in
the range of 0.0 to 1.2 kPa) and independent of 2-propanol
2
5
1.3 kPa C H OD, 1.0 kPa C D OD, and 0.4−2.3 kPa deionized
2
5
2
5
water) were controlled by adjusting the liquid flow rates into
the system. Condensation of the reactants and products was
7
99
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
avoided via the resistive heating of the transfer lines to
temperatures greater than 343 K.
Table 1. Ethylene and DEE Synthesis Rates for Ethanol
Dehydration at an Ethanol Pressure of 1.2 kPa and a Total
3
−1
The composition of the reactor effluent was analyzed using
an online mass spectrometer (MKS Cirrus 200 Quadrupole)
and a gas chromatograph (Agilent 6890 N GC) with both a
flame ionization detector fed through a methyl-siloxane
capillary column (HP-1, 50.0 m × 320 μm × 0.52 μm) and a
thermal conductivity detector fed through a packed column
Gas Flow Rate of 3.2 cm s over 0.02 g of γ-Al O at 488 K
2 3
for a Catalyst Sample Which Was Not Exposed to Water
Prior to Reaction, A Catalyst Sample Exposed to 2.2 kPa
Water for 1 h Prior to Reaction, and the Water Exposed
3
−1
Sample Regenerated in Air (1.67 cm s for 4 h at 723 K)
ethylene synthesis rate
DEE synthesis rate
(
SUPELCO HAYESEP R 80/100 mesh packed column, 12 ft).
−7
−1 −1
−6
−1 −1
catalyst sample
(/10 mol s
g
)
(/10 mol s
g )
Error bars in the reported figures correspond to the 95%
confidence intervals based upon successive gas chromatograph
injections taken without changing the experimental conditions.
A reactor containing only quartz sand diluent (0.793 g)
generates, on average, 1.1 × 10 mol s of DEE and no
ethylene upon feeding ethanol (2.6 kPa) and water (1.1 kPa) at
No H O exposure
2.22
0.90
2.42
5.29
3.49
5.49
2
H O exposed
2
regenerated
−9
−1
unchanged between the samples, leading the authors to
conclude water simply blocked adsorption sites on the γ-
Al O surface rather than altering the structure of the active
4
88 K. The production of DEE on the quartz sand diluent
(∼10% of the total rate of DEE production) was accounted for
2
3
and removed when calculating the reported rates of DEE
synthesis. Exposure to pyridine (0.1 kPa) was observed to
completely and irreversibly deactivate the production of DEE
on the diluent alone; the rate of DEE production was,
therefore, not adjusted in measurements in which pyridine was
cofed with ethanol (Section 3.2).
sites. Density Functional Theory (DFT) calculations (Perdew−
Wang 91 generalized gradient-corrected exchange-correlation
25,26
27
functional) by Digne et al.
and by Wischert et al. both
demonstrate water is capable of irreversibly disassociating on
the under-coordinated aluminum atoms on the catalyst surface
to form stable surface hydroxyl species, preferentially on highly
under-coordinated (three- or four-coordinated) aluminum
2.3. In-Situ Chemical Titration of Diethyl Ether Active
Sites Using Pyridine. Pyridine (0.02 and 0.05 kPa) was
introduced to a reactant stream containing ethanol (1.5 kPa)
and water (1.1 kPa) flowing over 0.2 g of catalyst (exposed to
25,26
27
atoms.
Furthermore, Wischert et al. observed that the
number of sites for methane and hydrogen adsorption on γ-
Al O decreased with hydroxyl group coverage for samples
2
3
2
.2 kPa of water at 488 K for 1 h) at 488 K, and the resulting
pretreated at temperatures less than 923 K, demonstrating that
these hydroxyl groups are capable of blocking potential catalytic
active sites.
transient production of DEE was monitored using an in-line
mass spectrometer (MKS Cirrus 200 Quadrupole mass
spectrometer system). The pyridine uptake that would
eliminate DEE synthesis was determined by extrapolating the
deactivation profile (see section 3.2).
The rates of ethylene and DEE synthesis are restored upon
3
−1
treatment of the water exposed catalyst in air (1.67 cm s ) for
h at 723 K (Table 1), demonstrating that thermal treatment
4
2.4. Parameter Estimation Techniques for Kinetic
can recover active sites for ethanol dehydration.
All subsequent experiments and measurements were
performed on catalyst samples that have undergone water
treatment to investigate the kinetics of ethanol dehydration in
the absence of the irreversible deactivation caused by the water
product.
Modeling. Kinetic parameters were optimized and uncertain-
ties (95% marginal highest posterior density intervals) were
determined using the Athena Visual Studio (v14.2, W. E.
Stewart and M. Caracotsios) statistical software package and
Bayesian statistical estimation techniques. Replicates for fit
analysis were provided by independent kinetic measurements at
the same ethanol and water pressure but for experimental
measurements taken at constant water or constant ethanol
pressure discussed in Section 3.4.
3
.2. Titration of Active Sites for Ethanol Dehydration.
The rates of ethylene and DEE synthesis were observed to be
inhibited by pyridine, but were found to return to the values
measured prior to the introduction of pyridine after the
pyridine feed was stopped (Figure 1a). These observations
demonstrate the following: (1) either Lewis acidic Al atoms or
hydroxyl groups comprise the active sites for both ethylene and
DEE formation and (2) pyridine can reversibly bind to these
active sites during ethanol dehydration, thus, inhibiting the
reaction, but is incapable of irreversibly adsorbing onto these
sites under reaction conditions.
The rate of formation of both ethylene and DEE normalized
to the formation rate measured in the absence of a pyridine co-
feed at different pyridine pressures is shown in Figure 1b. The
rate of ethylene formation was found to be inhibited to a
greater extent than the rate of DEE formation; indicating a
difference in the equilibrium surface coverage of pyridine on the
active sites of DEE and ethylene synthesis. The acidic active
sites for ethylene and DEE synthesis are, therefore, not
equivalent, although their uniqueness cannot be proven with
inhibition experiments alone.
3. RESULTS AND DISCUSSION
3
.1. Effects of Prior Exposure of γ-Al O to Water on
2
3
Ethanol Dehydration Rates. The rates of ethylene and
diethyl ether synthesis from the dehydration of ethanol at 488
K were found to be 59 and 34% lower, respectively, for catalyst
samples that have undergone treatment in water prior to
reaction (exposed to 2.2 kPa of water at 488 K for 1 h) than
catalyst samples which have not (Table 1), demonstrating water
is capable of deactivating some of the active sites and inhibiting
the rate of ethylene and DEE formation either by irreversibly
adsorbing onto the active sites or by causing a structural change
in the surface of the catalyst. This conclusion is supported by
the Temperature Programmed Desorption (TPD) experiments
24
performed by Roy et al. in which the amount of 2-propanol
adsorbed onto the surface of their γ-Al O samples was found
2
3
to be 73% lower for samples with prior exposure to water
(
exposed to 0.4 kPa water vapor at 373 K) than for samples
The amount of pyridine it would take to completely
deactivate DEE synthesis was estimated by linearly extrapolat-
ing the initial slope of the measured DEE synthesis rate,
24
without any prior water exposure. Additionally, Roy et al.
observed the temperature at which olefin was formed was
8
00
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
Figure 1. Ethylene (green triangles) and DEE (blue diamonds) (a) synthesis rates with respect to reaction time and (b) reduced synthesis rates
normalized to the measured synthesis rates with no pyridine co-feed as a function of pyridine pressure for the dehydration of 0.9 kPa of ethanol over
3
−1
0
.02 g of γ-Al O (total volumetric flow rate = 3.2 cm s ) at 488 K. The dashed lines serve as a guide for the eye.
2 3
considering pyridine adsorption to be irreversible and all DEE
active sites to exhibit similar activity in the linear regime
pyridine adsorption sites for complete deactivation may not
correlate directly with the number of active sites for DEE
formation for two reasons: (1) it is possible that pyridine could
adsorb onto sites other than the active sites for ethanol
dehydration; and (2) the quantity of sites for either ethylene or
DEE formation cannot be independently inhibited using an
unspecific titrant such as pyridine.
(Figure 2). The residual synthesis of DEE is most likely a result
Electron Paramagnetic Resonance measurements performed
2
8
by Zotov et al. on anthracene-exposed γ-Al O samples
2
3
−2
determined ∼0.002 molecules nm of anthracene underwent
radicalization on weak electron acceptor sites, which have been
demonstrated to possess a linear correlation with the rate of
ethylene formation from ethanol dehydration on doped γ-Al O
2
3
28
samples. The site density estimated by Zotov et al. (∼0.002
−2
sites nm ) provides a lower bound to the active site density, as
additional active sites for dehydration may exist that are not
active for anthracene radicalization, while the estimation based
−2
upon in situ pyridine titration (∼0.1 sites nm ) provides an
upper bound to the active site density.
Figure 2. DEE synthesis rates as a function of time after introduction
of 0.03 kPa of pyridine for the dehydration of ethanol (1.5 and 1.1 kPa
ethanol and water partial pressures, respectively) over 0.2 g of γ-Al O
3
.3. Kinetic Isotope Effects for Ethanol Dehydration.
2
3
The rates of ethylene and DEE formation from C H OH
relative to the rates of formation using C D OD and C H OD
r /r ) reactants at 488 K are presented in Table 3. No
H D
3
−1
2
5
(total volumetric flow rate = 1.7 cm s ) at 488 K. The dashed line
2
5
2
5
shows the linear extrapolation used to determine pyridine uptake for
complete deactivation.
(
Table 3. Measured Kinetic Isotope Effects for Ethylene and
Diethyl Ether Formation at 488 K for the Dehydration of
C H OD and C D OD over γ-Al O
of the dynamic reversibility of pyridine adsorption and
−5
inhibition demonstrated in Figure 1. On average, 2.7 × 10
mol g− (∼0.10 sites nm ) of pyridine is estimated to
completely deactivate DEE synthesis; this value is consistent
across different pyridine pressures (0.02 and 0.05 kPa) and
1
−2
2
5
2
5
2
3
reactant
a
b
product
ethylene KIE (r /r )
C H OD
C D OD
2 5
−1
2
5
space velocities (1.1 and 2.0 s ) (Table 2). The number of
0.89 ± 0.14
0.97 ± 0.12
2.42 ± 0.19
1.01 ± 0.14
H
D
diethyl ether KIE (r /r )
H
D
Table 2. Pyridine Uptake Necessary for Complete
a
Reactions performed under 1.3 kPa of C H OH or C H OD with no
Deactivation of DEE Synthesis Rate Determined from in-situ
Titrations Performed at Differing Pyridine Pressures and
2
5
2
5
b
water co-feed. Reactions performed under 1.0 kPa of C H OH or
2
5
a
C D OD with 1.2 kPa water co-feed.
2 5
Space Velocities over 0.2 g of γ-Al O at 488 K
2
3
space velocity
pyridine pressure
(/kPa)
pyridine uptake
−
1
‑5
−1
(/s
)
(/10 mol g
)
statistically significant effect on the rate of DEE formation is
observed when using either deuterated reactant, demonstrating
the rate-limiting step of DEE formation does not involve the
cleavage of a C−H bond or the cleavage of one of the O−H
bonds. Bimolecular dehydration of ethanol to form DEE,
1.1
1.1
2.0
2.0
0.05
0.02
0.05
0.02
2.7 ± 0.5
2.2 ± 0.9
3.2 ± 1.6
2.7 ± 0.3
therefore, involves either the cleavage of the C -O bond or an
α
a
The 95% confidence intervals were determined based upon
independent titrations.
Al−O bond in the rate-limiting step. Similarly, no kinetic
29
̈
isotope effects were observed by Knozinger et al. for dimethyl
8
01
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
ether formation from the dehydration of methanol, CH OD,
The rate of ethylene synthesis upon co-feeding ethanol and
water at 488 K was found to be zero order with respect to
ethanol pressure for the majority of the ethanol and water
pressures investigated (between 1.9 and 7.0 kPa ethanol partial
pressure and 0.4 and 2.2 kPa water partial pressure); becoming
of positive order (0.4) in ethanol only at low ethanol pressures
(≤2 kPa) and at the tested highest water pressure (2.2 kPa) as
shown in Figure 4a. A negative order dependence on water
pressure for the rate of ethylene formation is still observed,
however, in the regime where ethylene synthesis is invariant
with ethanol pressure. These observations are consistent with
the conclusion that the formation of ethylene must be inhibited
at least partially by the formation of ethanol−water dimers to
maintain the observed dependencies.
The inhibition of the rate of ethylene synthesis by water (0.4
to 1.9 kPa) was found to decrease with increasing ethanol
pressure (0.9 to 4.2 kPa) and was found to have, on average, an
order less than 1 (between −0.9 and −0.5) (Figure 5a). A
competitive occupation of the surface sites by single molecule
ethanol species, ethanol−water dimers, and water dimer surface
species provides a consistent kinetic scheme capable of
describing these observations.
3
and CD OD over γ-Al O .
3
2
3
No statistically significant difference was observed between
the rates of ethylene formation from the dehydration of
C H OH and C H OD, demonstrating O−H bond cleavage is
2
5
2
5
not kinetically limiting for ethylene formation. A primary
isotope effect, however, was observed for the dehydration of
C D OD, verifying the rate-limiting step of ethylene formation
2
5
involves either the breaking of a C−H bond (likely the more
acidic C -H bond) or the desorption of a water molecule from
β
the surface of the alumina in which a surface O−H bond is
broken with the hydrogen atom originating from the ethanol
18
̈
molecule. Knozinger and Scheglila similarly observed a
primary kinetic isotope effect for tert-butanol, sec-butanol, and
iso-butanol dehydration over γ-Al O only when the C atom
2
3
β
was deuterated.
.4. Kinetics and Mechanism of Ethanol Dehydration.
3
The formation rates of both ethylene and DEE formation were
observed to be inhibited by water at 488 K; the rates of both
DEE and ethylene synthesis, however, recovered to their initial
value upon restoring the initial water pressure as shown in
Figure 3, demonstrating water is capable of reversibly inhibiting
ethylene and DEE synthesis in addition to irreversibly
deactivating a fraction of the active sites of γ-Al O .
A positive order dependence (between 0.0 and 0.8) in
ethanol partial pressure was observed for the rate of DEE
formation over the tested range of ethanol and water partial
pressures (Figure 4b), with the exception of the highest tested
ethanol pressures (>4 kPa) and the lowest tested water partial
pressure (0.4 kPa), where the rate of DEE formation was found
to be independent of ethanol pressure. The observed reaction
order for DEE synthesis with respect to ethanol decreased with
increasing ethanol pressure and decreasing water pressure.
These observations can be explained by the competition
between ethanol−water dimers and adsorbed ethanol dimers.
This conclusion is further supported by the observation that the
reaction order with respect to water pressure decreases with
increasing ethanol pressure (0.9 to 4.2 kPa) from −1.3 to −0.7
2
3
(
Figure 5b).
The rate of ethylene and DEE synthesis was found to be in
Figure 3. Ethylene (green triangles) and DEE (blue diamonds)
synthesis rates from the dehydration of 4.2 kPa of ethanol over 0.02 g
the zero order regime with respect to ethanol pressure above 4
kPa of ethanol and at 0.4 kPa of water (Figure 4) despite
ethylene synthesis being a unimolecular reaction and DEE
synthesis being a bimolecular reaction; implying the active sites
for ethylene and DEE formation are inhibited by different
3
−1
of γ-Al O (total volumetric flow rate = 3.2 cm s ) at 488 K as a
2
3
function of reaction time and water pressure.
Figure 4. (a) Ethylene and (b) DEE synthesis rates for ethanol dehydration at 488 K on γ-Al O as a function of ethanol partial pressure with 0.4
2
3
3
−1
kPa (blue diamonds), 0.6 kPa (orange circles), 1.2 kPa (green triangles), and 2.2 kPa (red squares) water co-feeds (total gas flow rate = 3.2 cm s ).
The solid lines represent the model fits to equations (a) 3 and (b) 5.
8
02
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
Figure 5. (a−c) Ethylene and (d−f) DEE synthesis rates for ethanol dehydration at 488 K over γ-Al O as a function of water partial pressure with
2
3
3
−1
(
a,d) 4.2 kPa (red squares), (b,e) 3.0 kPa (green triangles), and (c,f) 0.9 kPa (orange circles) ethanol co-feeds (total gas flow rate = 3.2 cm s ). The
solid lines represent fits to the experimental data with the models presented in equations (a−c) 3 and (d−f) 5.
Scheme 1. Ethoxide Desorption Limited Mechanism for Ethylene Formation from Ethanol Dehydration over γ-Al O
2
3
surface species. This observation, in addition to the observed
differences in pyridine inhibition of the synthesis rates for the
two dehydration products, leads us to postulate that the site
requirements for ethylene and DEE synthesis are distinct.
8
03
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
Table 4. Estimated Values for the Kinetic Parameters of Ethylene Formation over γ-Al O at 488 K Using the Model Presented
2
3
in eq 3 and the data from Figures 4 and 5
kC2H4 (molC2H4 s− g−1
1
)
8
K′AW/(K K ) (kPa−1)
K K /(K K ) (kPa−1)
KW1KW2/(K K ) (kPa−1)
A1 d
parameter
A1
d
A1 A2
A1
d
estimated value
(6.42 ± 0.90) × 10−
0.81 ± 0.27
0.035 ± 0.028
0.57 ± 0.19
Figure 6. Parity plots for the kinetic models presented in (a) eq 3 for ethylene formation and (b) eq 5 for diethyl ether formation at 488 K. The solid
line represents a perfect prediction of the observed synthesis rates.
The observed site requirements and the measured kinetic
dependencies for ethylene formation can be explained with the
proposed ethoxide mediated mechanism shown in Scheme 1. In
this mechanism, an ethanol molecule initially adsorbs onto a
Lewis acid site of γ-Al O (Step 1, Scheme 1) and dissociates
on the surface to form an ethoxide species and a hydroxyl group
composed of the hydrogen atom from ethanol and a surface
oxygen atom on the same Lewis acid site (Step 2, Scheme 1).
PEtOH and PH2O denote the partial pressures of ethanol and
water, respectively. The kinetic parameter kC2H4 represents the
rate constant associated with the formation of ethylene by
desorption of ethoxide species. The parameters K , K , K ,
A1
d
A2
2
3
K , K denote the equilibrium constants for the formation of
W1
W2
adsorbed ethanol species, dissociation of the ethanol species to
form a surface ethoxide species, formation of ethanol dimer
complexes, the adsorption of water onto the active sites, and
subsequent adsorption of additional water molecules to form
water dimer complexes, respectively. KAWa and KAWb are the
equilibrium constants for the kinetically indistinguishable
routes for the formation of the ethanol−water dimers from
the adsorption of an ethanol molecule to an active site occupied
by a surface bound water molecule or the adsorption of a water
molecule to an active site occupied by an ethanol molecule,
respectively.
If surface bound ethoxide species, ethanol−water dimer
species, water dimer species, and surface bound ethanol dimer
species are considered to be the dominant surface species under
the conditions investigated in this kinetic study, eq 2 can be
simplified into eq 3.
The C -H and C -O bonds are subsequently cleaved as the
β
α
ethoxide species desorbs from the surface to form an ethylene
molecule and a hydroxyl group on the γ-Al O surface in the
2
3
rate-limiting step (Step 3, Scheme 1). Desorption of the
hydroxyl group and a surface proton forms a water molecule;
regenerating the catalyst surface and completing the catalytic
cycle (Step 4, Scheme 1). The active site can be inhibited by
the adsorption of a water molecule (Step 5, Scheme 1) along
with the subsequent adsorption of an additional water or
ethanol molecule to form either a water dimer (Step 6, Scheme
1
) or an ethanol−water dimer (Step 7a, Scheme 1),
respectively. The ethanol−water dimer could alternatively be
formed from the adsorption of a water molecule onto an active
site occupied by a surface bound ethanol molecule (Step 7b,
Scheme 1). These two routes for the formation of the ethanol−
water dimer, however, cannot be kinetically distinguished. The
active site could also be inhibited by an ethanol dimer species
formed from the adsorption of a second ethanol molecule after
the formation of the surface bound ethanol species (Step 8,
Scheme 1).
⎡
K′AW
r
= k
PEtOH/⎢P
+
P
P
C H
C H
EtOH
EtOH H O
2
4
2
4
2
⎣
K K
A1 d
⎤
K K
K K
W1 W2
K K
A1 d
A1 A2
2
2
+
P
+
P
⎥
EtOH
H O
2
K K
⎦
A1
d
(3)
The new parameter K′ represents the sum of the equilibrium
The kinetic expression for the rate of ethylene formation
AW
constants for the two indistinguishable mechanisms for the
(rC2H4) presented in eq 2 can be derived from the mechanism
formation of the ethanol−water dimer species (K′
=
AW
presented in Scheme 1 assuming quasi-equilibrium is achieved
for all steps prior to the proposed rate limiting desorption of
the ethylene molecule (Step 3, Scheme 1).
K K
W1 AWa
+ K K ). Equation 3 is consistent with the
A1 AWb
observation that the rate of ethylene synthesis is largely
independent of ethanol pressure (between 1.9 and 7.0 kPa
ethanol partial pressure and 0.4 and 2.2 kPa water partial
pressure), becoming slightly of positive order only at low
ethanol pressures (less than 1.9 kPa ethanol) at the highest
water pressure tested (2.2 kPa), while being inhibited by water
with an average order less than 1. The parameters in eq 3 were
fit to the experimental data, and their values are shown in Table
⎡
r
= k
K K P
/ 1 + K K P
+ K P
C H
C H A1 d EtOH ⎣
A1 d EtOH
W1 H O
2
4
2
4
2
2
+
(K K
+ K KAWb)P
P
+ K K P
W1 AWa
A1
EtOH H O
A1 A2 EtOH
2
2
⎤
+
K KW2P
W1
H O⎦
2
(2)
8
04
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
Scheme 2. Bimolecular Mechanism for Diethyl Ether Formation from Ethanol Dehydration over γ-Al O
2
3
4
. The parity plot of the fitted model, shown in Figure 6a,
proposed and cannot be distinguished with kinetic measure-
ments alone.
demonstrates the proposed model is capable of describing the
data accurately. The normal probability plot and lag plot for the
residual error in the model are presented and discussed in the
Supporting Information (Figures S1a and S2a, respectively).
A similar mechanism (Scheme 2) can be proposed for the
formation of DEE. Initially, an ethanol molecule adsorbs onto
the active site of the catalyst (Step 1, Scheme 2). The
adsorption of another ethanol molecule forms an ethanol dimer
complex on the surface (Step 2, Scheme 2). This dimer
complex then decomposes in a rate-limiting step involving the
cleavage of either a C−O or Al−O bond to form a gas phase
DEE molecule and a water molecule, regenerating the catalyst
surface in the process (Step 3, Scheme 2); although this process
is expected to be composed of many fundamental steps, these
steps are not kinetically observable. Alternatively, a water
molecule can competitively adsorb onto the active site (Step 4,
Scheme 2), and subsequent adsorption of an additional water
molecule or an ethanol molecule forms a water dimer (Step 5,
Scheme 2) or an ethanol−water dimer species (Step 6a,
Scheme 2), respectively, capable of inhibiting the rate of DEE
formation. As was the case with ethylene formation, the
ethanol−water dimer could alternatively be formed from the
adsorption of a water molecule onto a site occupied by a surface
bound ethanol molecule (Step 6b, Scheme 2).
−1
The value of (K K )/(K K ) (0.57 kPa ) relative to the
W1 W2
A1 d
−1
value of (K′ )/(K K ) (0.81 kPa ) signifies that water
dimers will become more abundant on the surface of γ-Al O
AW
A1 d
2
3
than the ethanol−water dimers when the ratio of water to
ethanol pressure exceeds 1.5, which explains why a positive
order dependence of ethanol on the rate of ethylene formation
was observed only when the water pressure was equivalent to
or exceeded the ethanol partial pressure (Figure 5a).
−1
The lower value of (K K )/(K K ) (0.035 kPa ) relative
A1 A2
A1 d
to both (K K )/(K K ) and (K′ )/(K K ) signifies that
W1 W2
A1
d
AW
A1 d
only a small fraction of the surface is occupied with ethanol
dimers except at the highest ethanol pressures and lowest water
pressure tested. For this reason, the number of tested
conditions in which this term is significant is much smaller
than the other terms in the denominator of the rate expression
which results in a relatively large confidence interval (80% of
the parameter value).
Equation 4 shows the kinetic expression for the rate of DEE
formation (rDEE) derived from the proposed mechanism.
The proposed desorption of the ethoxide species is
consistent with the E2-like rate-limiting step for olefin
formation proposed by Kno
18
̈
zinger and Scheglila based upon
2
⎡
rDEE = kDEEK K P
/ 1 + K P
+ K P
A1 A2 EtOH ⎣
A1 EtOH
W1 H O
2
the observed KIE. The formation of ethoxides is supported by
30
2
the observation of Greenler that the infrared (IR) spectra of
alumina exposed to ethanol before evacuation contained similar
bands to those found in the IR spectra of Al(OCH CH ) .
Furthermore, Gabrienko et al. observed chemical shifts
indicative of the formation of alkoxide species on different
+ (K K
+ K KAWb)P
P
+ K K P
W1 AWa
A1
EtOH H O
A1 A2 EtOH
2
2
⎤
+
K KW2P
W1
H O⎦
2
3 3
2
(4)
31
The rate constant for DEE formation is denoted by kDEE
.
1
3
Equation 4 can be simplified into eq 5 assuming that ethanol−
water and ethanol dimer species are more abundant than the
other surface species depicted in Scheme 2.
Lewis acid sites in C solid state NMR measurements of γ-
Al O exposed to C labeled propene, n-butene, and isobutene
1
3
2
3
(
63 and 79 ppm, 69 and 74 ppm, and 67 ppm, respectively).
Thus, the desorption of a surface ethoxide species (step 3) is
proposed to be rate-limiting rather than the subsequent
desorption of water (step 4) which could also produce the
observed KIE (Section 3.3). The existence of surface ethoxides
does not, however, guarantee their involvement in the synthesis
of ethylene; thus, a kinetically equivalent mechanism in which
the adsorbed ethanol monomer formed in step 2 directly
undergoes an E2-elimination to form ethylene could also be
2
kDEE
P
EtOH
rDEE = P2
K ′AW
+
P
P
2
EtOH
EtOH H O
K K
(5)
A1 A2
The rate expression presented in eq 5 can describe the
observed less than 1 order reaction order in ethanol pressure on
the rate of DEE formation and the observation that the order
decreased with increasing ethanol pressure and decreasing
8
05
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
water pressure until becoming independent of ethanol pressure.
The parameters in eq 5 were estimated with a Bayesian fit to
the experimental data collected at 488 K and are shown in
Table 5. The model accurately describes the data between 0.9
between these catalytic systems. Our findings demonstrate the
kinetic importance of molecular dimers for the dehydration of
ethanol over γ-Al O and provide important considerations for
2
3
future catalytic research with this common catalytic system.
Table 5. Estimated Values for the Kinetic Parameters of DEE
4. CONCLUSIONS
Formation over γ-Al O at 488 K Using the Model Presented
2
3
Kinetic measurements of ethanol dehydration over γ-Al O at
2
3
in eq 5 and the Data from Figures 4 and 5
4
88 K on both samples exposed to water before reaction and
1
parameter
kDEE (molDEE s− g−1
)
(K′AW)/(K K )
untreated samples established water is capable of irreversibly
deactivating some of the active sites involved in the formation
of both ethylene and DEE. Steady state kinetic measurements
on samples treated with water demonstrated water is capable of
reversibly inhibiting the rate of both ethylene and DEE
synthesis on the remaining active sites. Pyridine and ethanol co-
feed kinetic experiments indicated pyridine reversibly inhibits
the rate of ethylene and DEE formation to different extents,
demonstrating the active sites for ethylene and DEE formation
are acidic and nonequivalent. In-situ pyridine titration studies
A1 A2
(1.60 ± 0.18) × 10−6
estimated value
5.05 ± 1.24
and 3.0 kPa in ethanol partial pressure and between 1.2 and 2.2
kPa water partial pressures as well as describing the zero order
regime at ethanol pressures >4.5 kPa and at a water pressure of
0
.4 kPa (Figures 4b and 5b). The parity plot for the model and
the collected data is presented in Figure 6b. The normal
probability plot and lag plot associated with the residual error in
the model are presented and discussed in the Supporting
Information (Figures S1b and S2b, respectively).
−5
−1
estimated 2.7 × 10 mol g of pyridine are required to adsorb
onto the surface of water-treated γ-Al O3 to completely
2
The value of the combined parameter (K′ )/(K K )
AW
A1 A2
deactivate the synthesis of DEE; this value provides an upper
estimate of active site density for the formation of DEE (0.10
(
5.05) signifies that, under the experimental conditions,
ethanol−water dimers are the most abundant surface species
unless the ethanol pressure is 4 or more times greater than the
water pressure, explaining why the rate of DEE formation was
found to be independent of ethanol partial pressure only when
the ethanol partial pressure was an order of magnitude larger
than the water partial pressure.
−2
sites nm ). A primary kinetic isotope effect was observed for
the rate of ethylene formation when C D OD was used as the
reactant but not when C H OD was employed, indicating the
rate-limiting step for ethylene formation from ethanol
dehydration over γ-Al O involves either the cleavage of a
C−H bond (likely the acidic C -H bond) of ethanol or the
subsequent desorption of water from the surface of γ-Al O . No
2
5
2
5
2
3
β
Pyridine titration studies elucidated both the acidic and the
nonequivalent nature of the active sites for the formation of
ethylene and DEE; whether the active sites consist of a single
adsorption site or two adjacent adsorption sites, however, could
not be determined based upon kinetic experiments or pyridine
titration studies alone. The ethanol dimers, water dimers, and
ethanol−water dimers could represent either two molecules
adsorbed onto the same adsorption site, as shown to form on
2
3
significant kinetic isotope effect was observed for the rate of
DEE formation for either of the deuterated reactants,
elucidating the formation of DEE from ethanol dehydration
over γ-Al O is limited by either C -O or Al−O bond cleavage.
2
3
α
Steady state kinetic measurements of the formation of ethylene
and DEE at various ethanol and water partial pressures at 488 K
demonstrated the following: (1) the rate of ethylene formation
was independent of ethanol pressure for most of the ethanol
and water pressures tested while being inhibited by water, (2)
the rate of DEE formation was found to be a positive, less than
first order dependence in ethanol pressure except at very high
ethanol pressure and low water pressures where the rate of
DEE formation became independent of ethanol pressure, and
3
2−34
acidic zeolites,
or adsorbed onto two adjacent adsorption
sites both of which are necessary for the formation of ethylene
and DEE and comprise the active site for dehydration. These
two possibilities, however, are kinetically equivalent and, thus,
cannot be distinguished using the information presented in this
work. The existence of two adjacent adsorption sites has been
23
previously suggested by De Mourgues et al. to explain the
observed independence in isopropanol pressure and simulta-
neous inhibition by water for the rate of propene formation
(
3) the order of the inhibition by water decreased with
increasing ethanol pressure for both ethylene and DEE
formation. The proposed mechanisms and associated rate
equations, in which, at the reaction conditions, the formation of
ethylene and DEE is inhibited mostly by bimolecular surface
species, are consistent with the experimental observations.
from isopropanol dehydration over γ-Al O . De Mourges et
2
3
2
3
al. proposed that one site exclusively adsorbs ethanol, while
the other is exclusively inhibited by water. De Mourges’s model,
however, does not allow for the inhibition of olefin formation
by water dimers and, thus, is unable to fully explain the
observed −1.6 order inhibition of the rate of ethylene synthesis
by water at the lowest ethanol pressure (0.9 kPa) and highest
measured water pressures (1.5 and 1.9 kPa).
ASSOCIATED CONTENT
■
*
S
Supporting Information
Analysis of the residual error in the kinetic models for ethanol
dehydration. This material is available free of charge via the
Internet at http://pubs.acs.org.
The measured inhibitory effects of water and the kinetic
effects of ethanol for ethylene and DEE synthesis demonstrate
the kinetic importance of water, ethanol−water, and ethanol
molecular pairs, either in the form of adsorbed dimers or in the
form of molecules adsorbed on adjacent active sites, in the
dehydration of ethanol over γ-Al O . The observed inhibition
AUTHOR INFORMATION
■
Corresponding Author
2
3
*
E-mail: abhan@umn.edu. Phone: 612-626-3981. Fax: 612-
26-7246.
by ethanol−water dimers for both ethylene and DEE formation
6
on Lewis-acidic γ-Al O was not observed for the same
2
3
32
chemistry performed over Brønsted acidic zeolite catalysts,
highlighting the important fundamental mechanistic differences
Author Contributions
§
Equal contribution to this work was given by these authors.
8
06
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807

ACS Catalysis
Research Article
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank The Dow Chemical Company for their financial
support. H.C. was supported as part of the Catalysis Center for
Energy Innovation, an Energy Frontier Research Center funded
by the U.S. Department of Energy, Office of Science, Office of
Basic Energy Sciences under Award no. DE-SC0001004.
REFERENCES
■
(
1) Chorkendorff, I.; Niemantsverdriet, J. W. Concepts of Modern
Catalysis and Kinetics; Wiley-VCH: Weinheim, Germany, 2007.
(
2) Kogel, J. E.; Trivedi, N. C.; Barker, J. M. Industrial Minerals &
Rocks: Commodities,Markets, and Uses; Society for Mining Metallurgy:
Littleton, CO, 2006.
(
3) Lippens, B. C.; De Boer, J. H. Acta Crystallogr. 1964, 17, 1312−
321.
4) Slade, R. C. T.; Southern, J. C.; Thompson, I. M. J. Mater. Chem.
991, 1, 563−568.
5) Cesteros, Y.; Salagre, P.; Medina, F.; Sueiras, J. E. Chem. Mater.
999, 6, 123−129.
6) Winfield, M. E. In Catalysis; Emmett, P. H., Ed.; Reinhold: New
York, 1960; Vol. VII.
1
(
1
(
1
(
(
(
7) Kno
̈
zinger, H. Angew. Chem., Int. Ed. Engl. 1968, 7, 791−805.
8) Coster, D.; Blumenfeld, A. L.; Fripiat, J. J. J. Phys. Chem. 1994, 98,
6
(
201−6211.
9) Morterra, C.; Magnacca, G. Catal. Today 1996, 27, 497−532.
10) Liu, X.; Truitt, R. E. J. Am. Chem. Soc. 1997, 119, 9856−9860.
11) Parry, E. P. J. Catal. 1963, 2, 371−379.
(
(
(
12) Soled, S. L.; McVicker, G. B.; Murrell, L. L.; Sherman, L. G.;
Dispenziere, N. C., Jr.; Hsu, S. L.; Waldman, D. J. Catal. 1988, 295,
86−295.
13) Baertsch, C. D.; Komala, K. T.; Chua, Y. H.; Iglesia, E. J. Catal.
2
(
2
(
(
002, 205, 44−57.
14) Ripmeester, J. A. J. Am. Chem. Soc. 1983, 105, 2925−2927.
15) Kwak, J.; Hu, J.; Kim, D.; Szanyi, J.; Peden, C. J. Catal. 2007,
2
(
51, 189−194.
16) Kwak, J. H.; Mei, D.; Peden, C. H. F.; Rousseau, R.; Szanyi, J.
Catal. Lett. 2011, 141, 649−655.
(
(
(
(
(
(
(
17) Pines, H.; Haag, W. O. J. Am. Chem. Soc. 1961, 82, 1526−1532.
18) Knozinger, H.; Scheglila, A. J. Catal. 1970, 17, 252−263.
19) Knozinger, H.; Buhl, H.; Kochloefl, K. J. Catal. 1972, 24, 57−68.
20) Knozinger, H.; Buhl, H.; Ress, E. J. Catal. 1968, 12, 121−128.
21) Kallo, D.; Kno
̈
̈
̈
̈
̈
́
̈
zinger, H. Chem.-Ing. Techn. 1967, 39, 676−680.
22) Shi, B.; Davis, B. H. J. Catal. 1995, 157, 359−367.
23) De Mourgues, L.; Peyron, F.; Trambouze, Y.; Prettre, M. J.
Catal. 1967, 7, 117−125.
24) Roy, S.; Mpourmpakis, G.; Hong, D. Y.; Vlachos, D. G.; Bhan,
A.; Gorte, R. J. ACS Catal. 2012, 2, 1846−1853.
25) Digne, M.; Sautet, P.; Raybaud, P.; Euzen, P.; Toulhoat, H. J.
Catal. 2002, 211, 1−5.
26) Digne, M.; Sautet, P.; Raybaud, P.; Euzen, P.; Toulhoat, H. J.
Catal. 2004, 226, 54−68.
27) Wischert, R.; Laurent, P.; Coper
Am. Chem. Soc. 2012, 134, 14430−14449.
28) Zotov, R. a.; Molchanov, V. V.; Volodin, A. M.; Bedilo, A. F. J.
Catal. 2011, 278, 71−77.
29) Knozinger, H.; Scheglila, A.; Watson, A. M. J. Phys. Chem. 1968,
2, 2770−2774.
(
(
(
(
́
et, C.; Delbecq, F.; Sautet, P. J.
(
(
̈
7
(
(
30) Greenler, R. G. J. Chem. Phys. 1962, 37, 2094−2100.
31) Gabrienko, A. A.; Arzumanov, S. S.; Toktarev, A. V.; Stepanov,
A. G. J. Phys. Chem. C 2012, 116, 21430−21438.
(32) Chiang, H.; Bhan, A. J. Catal. 2010, 271, 251−261.
(33) Zecchina, A.; Bordiga, S.; Spoto, G.; Scarano, D.; Spano, G.;
Geobaldo, F. J. Chem. Soc., Faraday Trans. 1996, 92, 4863−4875.
34) Lee, C. C.; Gorte, R. J.; Farneth, W. E. J. Phys. Chem. B 1997,
01, 3811−3817.
(
1
8
07
dx.doi.org/10.1021/cs400051k | ACS Catal. 2013, 3, 798−807