Mechanism and kinetics of 1-dodecanol etherification over tungstated zirconia

Article abstract of DOI:10.1016/j.jcat.2017.08.001

Growing interest in finding renewable alternatives to conventional fossil fuels and petroleum-derived specialty chemicals has motivated the investigation of biomass-derived alcohols to make ethers as diesel additives or lubricants. To optimize the direct etherification of long chain alcohols in the liquid phase, it is necessary to develop an understanding of the kinetics and mechanism of etherification and dehydration reactions. In this study, tungstated zirconia was identified as a selective solid-acid catalyst for the liquid-phase etherification of 1-dodecanol. Investigations of the mechanism and kinetics of this reaction suggest that cooperation between Br?nsted- and Lewis-acid sites on tungstated zirconia enhances the selectivity to ether by increasing the surface concentration of adsorbed alcohol, thereby promoting bi-molecular ether formation relative to unimolecular alcohol dehydration. The suggested rate limiting step for etherification is the formation of a C–O bond between two adsorbed alcohol molecules, and the suggested rate-limiting step for dehydration is the cleavage of the C–H bond of the β-carbon atom in an adsorbed alcohol. Measurements of the kinetic isotope effects for etherification and dehydration support the proposed mechanism. A microkinetic model based on the proposed mechanism for dodecanol etherification and dehydration over tungstated zirconia accurately describes the observed effects of alcohol concentration and product inhibition.

Full text of DOI:10.1016/j.jcat.2017.08.001

Journal of Catalysis 354 (2017) 13–23
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Mechanism and kinetics of 1-dodecanol etherification over tungstated
zirconia
Julie Rorrer a,b, Ying He a,c, F. Dean Toste a,c, Alexis T. Bell a,b,
⇑
a Energy Biosciences Institute, University of California Berkeley, CA 94720, United States
b Department of Chemical and Biomolecular Engineering, University of California Berkeley, CA 94720, United States
c Department of Chemistry, University of California Berkeley, CA 94720, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Growing interest in finding renewable alternatives to conventional fossil fuels and petroleum-derived
specialty chemicals has motivated the investigation of biomass-derived alcohols to make ethers as diesel
additives or lubricants. To optimize the direct etherification of long chain alcohols in the liquid phase, it is
necessary to develop an understanding of the kinetics and mechanism of etherification and dehydration
reactions. In this study, tungstated zirconia was identified as a selective solid-acid catalyst for the liquid-
phase etherification of 1-dodecanol. Investigations of the mechanism and kinetics of this reaction suggest
that cooperation between Brønsted- and Lewis-acid sites on tungstated zirconia enhances the selectivity
to ether by increasing the surface concentration of adsorbed alcohol, thereby promoting bi-molecular
ether formation relative to unimolecular alcohol dehydration. The suggested rate limiting step for ether-
ification is the formation of a CAO bond between two adsorbed alcohol molecules, and the suggested
rate-limiting step for dehydration is the cleavage of the CAH bond of the b-carbon atom in an adsorbed
alcohol. Measurements of the kinetic isotope effects for etherification and dehydration support the pro-
posed mechanism. A microkinetic model based on the proposed mechanism for dodecanol etherification
and dehydration over tungstated zirconia accurately describes the observed effects of alcohol concentra-
tion and product inhibition.
Received 3 June 2017
Revised 2 August 2017
Accepted 3 August 2017
Keywords:
Dodecanol
Etherification
Tungstated zirconia
Di-dodecylether
Dodecene
Ó 2017 Elsevier Inc. All rights reserved.
1. Introduction
Ethers are attractive products because they can be used as die-
sel additives and automotive lubricant base oils. Diesel-range lin-
Continued use of fossil energy resources to produce fuels con-
tributes to an increase in atmospheric CO2 and in turn to changes
in the global climate [1]. This concern has motivated the explo-
ration of biomass as a possible source of renewable carbon for
the production of fuels and lubricants [2]. A central question is
how to convert biomass into synthons that could be used to pro-
duce fuels and lubricants. One of the appealing approaches is to
ferment the sugars derived by hydrolysis of cellulose and hemicel-
lulose into alcohols such as ethanol and butanol. The condensation
of furfural with acetone under hydrogen can also produce 1-
octanol [3]. Other longer-chain alcohols such as 1-dodecanol can
be produced by hydrolysis of triglycerides and fatty acids. These
linear alcohols can also be converted to higher carbon-number
alcohols with branched carbon chains via the Guerbet reaction,
and the corresponding Guerbet alcohols can be converted to ethers
[4].
ear ethers produced from biomass are of interest because they
have high cetane numbers as well as high energy density [5–8].
Shorter chain branched ethers have high octane numbers and can
be added to gasoline [9]. Ethers are also of interest as lubricants.
For example, alkylated di-phenyl ether and glyceryl ethers have
excellent properties [10,11], and branching in the alkyl portions
of ethers lowers the pour points and raises the viscosity of the
ether [12,13]. Therefore, ether-based lubricants sourced from bio-
mass provide a stable, tunable, and renewable alternative to
poly-alpha-olefins produced from petroleum [14,15].
Ethers can be prepared either by direct, acid-catalyzed etherifi-
cation of alcohols, or by reductive etherification of an alcohol and
an aldehyde or ketone. Direct etherification of alcohols is advanta-
geous because it does not require the use of hydrogen and precious
metal catalysts for reduction [16]. The competing reaction in the
presence of an acid catalyst is alcohol dehydration to produce
the corresponding alkene, a process that is favored thermodynam-
ically at temperatures above approximately 350 K. Since higher
temperatures are desirable for increasing reaction rates, this raises
⇑
Corresponding author at: Energy Biosciences Institute, University of California
Berkeley, CA 94720, United States.
E-mail address: alexbell@berkeley.edu (A.T. Bell).
http://dx.doi.org/10.1016/j.jcat.2017.08.001
0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

14
J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
the question of how one can favor etherification over alkene
formation.
catalysts were heated at 10 K/min and treated in air at 1073 K
for 3 h, then cooled to room temperature. To prepare pure zirconia,
amorphous zirconium oxyhydroxide was calcined at 1073 K under
the same incipient wetness impregnation conditions but without
the addition of ammonium metatungstate. Each catalyst was then
crushed to <250 mm mesh using a mortar and pestle.
Selective etherification has been reported for the liquid-phase
reactions of 1-octanol, 1-hexanol, and 1-pentanol over acid cata-
lysts such as Amberlyst 70 [17,18], Nafion NR-50 [18,19], and H-
BEA zeolite [20]. While polymeric resins are selective to ether, they
are not as thermally stable as metal oxides [18,21]. Zeolites, on the
other hand, are thermally stable, but catalyze unwanted side prod-
ucts and deactivate due to coking [22].
2.3. Catalyst characterization
In this study, we screened a series of solid-acid catalysts for the
direct etherification of 1-dodecanol and identified tungstated zir-
conia as a highly active and selective catalyst. Tungstated zirconia
has been employed previously for acid-catalyzed reactions includ-
ing gas-phase isomerization of n-butane [23], gas-phase dehydra-
tion of alcohols to alkenes [24], and liquid-phase reactions, such
as esterification, transesterification, and alkylation [24–26].
Investigations of the gas-phase kinetics for the dehydration of
short-chain, linear alcohols over tungstated zirconia indicate that
dehydration occurs via a unimolecular mechanism but that these
alcohols are not converted to ethers [27,28]. By contrast, we found
that the etherification of dodecanol in the liquid phase is highly
selective over tungstated zirconia. Motivated by this finding, we
undertook an investigation of the mechanism and kinetics of the
etherification and dehydration of 1-dodecanol over tungstated zir-
conia with the objective of developing an explanation for the high
selectivity of this catalyst for etherification in the liquid phase.
Powder X-ray diffraction (XRD) patterns for WOx/ZrO2 (0–22 wt
% W) were taken with a Bruker D8 GADDS diffractometer equipped
with a CuꢀKa source (40 kV, 40 mA). Raman spectra were obtained
with a LabRAM HR Horiba Scientific Raman spectrometer equipped
with a 633 nmꢀ1 laser. BET surface area measurements were per-
formed with a Micrometrics TriStar BET and pretreated with a
Micrometrics FlowPrep 060. Brønsted- and Lewis-acid sites were
identified and the ratio of these sites was determined from IR spec-
tra of adsorbed pyridine. Spectra were acquired using a Thermo
Scientific Nicolet 6700 Fourier Transform Infrared Spectrometer
(FT-IR) equipped with a Diffuse Reflectance Infrared Fourier Trans-
form Spectroscopy (DRIFTS) cell. A mixture of catalyst (50 mg)
diluted with KBr (250 mg) was added to the DRIFTS cell and pre-
treated at 573 K for 2 h under helium. Background scans of the cat-
alyst were taken at 393 K, 423 K, 473 K, 523 K, and 573 K. Pyridine
was introduced into the He flow at 393 K, and spectral data was
taken after stabilization of adsorbed pyridine at 393 K. The temper-
ature was then raised to measure the amount of pyridine that
remained adsorbed at 423 K, 473 K, 523 K, and 573 K. Spectral
intensities were calculated using the Kubelka-Munk function. The
concentration of Brønsted-acid sites was determined by titration
with NH4OH. The protons on the catalysts were first exchanged
for Na+ cations by placing the catalyst in a 1 M NaCl solution over-
night, and then the solution was titrated with NH4OH until the pH
was neutral, using phenolphthalein as an indicator. The moles of
base added was used as a metric for the number of H+ ions in solu-
tion [32]. ICP Elemental analysis was performed by Galbraith Lab-
oratories, Inc. in order to determine tungsten weight loadings.
2. Materials and methods
2.1. Materials
All chemicals obtained commercially were used without further
purification. The following compounds were obtained from Sigma-
Aldrich: 1-hexanol (>98%), 1-dodecanol (>98%), decane (>95%),
dodecane (>99%), hexane (>99%), 1-hexene (>99%), and pyridine
(99.8%). N-tetradecane was obtained from Spectrum Chemical
(>99%), and was used as an internal standard for analytical pur-
poses. Di-dodecyl ether (>95%) and 1-dodecene (>95%) were
obtained from TCI. Di-n-hexyl ether (>98%), and 2,6-di-tert-butyl-
pyridine (97%) were obtained from Alpha Aesar. Hexan-1,1-d2-1-
ol (>99%) was obtained from CDN Isotopes Inc. Hexan-2,2-d2-1-ol
was synthesized and purified to >98% according to Ref. [29]
Amberlyst 15 (hydrogen form, dry), and Amberlyst 36 were
obtained from Sigma-Aldrich. Amberlyst 70 was obtained from
Dow Chemical, and was dried at 368 K and stored in a desiccator
before use. Zeolite BEA was obtained from Alpha Aesar, and was
calcined at 873 K for 6 h before use. Gamma-alumina was obtained
from Strem Chemicals. Mesostructured silica (MCM-41 hexagonal
type), Nafion NR-50, and mesostructured aluminosilicate (MCM-
41, hexagonal) were obtained from Sigma-Aldrich. Para-toluene
sulfonic acid was obtained from Spectrum Chemical.
2.4. Isotopic labeling and NMR
Isotopic labeling of the alpha and beta hydrogen atoms in 1-
hexanol was used to support the proposed mechanisms of etheri-
fication and dehydration. Hexan-2,2-d2-1-ol was prepared accord-
ing to literature and the structure was confirmed using 1H NMR
[29]. Kinetic isotope effects (kH/kD) were determined by measuring
the initial rates of alkene and ether formation for 1-hexanol,
hexan-1,1-d2-1-ol, and hexan-2,2-d2-1-ol. NMR spectra of the reac-
tion products were recorded with a Bruker AVQ-400 spectrometer.
2.5. Batch reactions
All reactions were carried out in sealed 12 mL Q-Tube batch
reaction vessels with magnetic stirring at 600 RPM using an IKA
C-MAG HS 10 digital hot plate with temperature control accurate
to within 1 K. For determination of the reaction kinetics, a sepa-
rate batch reaction was carried out for each time point to ensure
consistency of volume and concentration of each sample. All reac-
tions over tungstated zirconia were carried out either in the
absence of solvent or in decane with 100 mg of catalyst and
250 mL of dodecanol (unless otherwise noted). N-tetradecane
(100 mL) was added post-reaction as a standard for analysis. The
reactants and catalyst were added to the Q-Tube in the pre-
heated hot plate with an aluminum heating block, and after the
specified reaction time, the vials were removed from the hot plate
and placed in an ice bath to stop the reaction. Products were
diluted with 5 mL of acetone, and centrifuged at 4000 RPM for
2.2. Synthesis of zirconia and tungstated zirconia
Porous amorphous zirconia, monoclinic zirconia, and
tungstated zirconia were synthesized using previously reported
methods.[23,30,31]
Amorphous
zirconium
oxyhydroxide
(ZrOx(OH)4-2x was formed by adding ammonium hydroxide
)
(Spectrum, 28–30%) dropwise to a stirred solution of 0.5 M zirconyl
chloride octahydrate (Sigma Aldrich, 98%) at 298 K. The precipitate
was filtered and rinsed with 10% ammonium hydroxide and dried
at 383 K for 24 h. Tungstated zirconia (4.1, 7.7, 10.2, 12.6, 15.4,
22.3 wt% W) was prepared via incipient wetness impregnation of
amorphous zirconium oxyhydroxide with aqueous ammonium
metatungstate hydrate (Spectrum). After impregnation, all

J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
15
8 min to remove the catalyst. A 200 mL aliquot of this solution was
then diluted with another 1.3 mL of acetone and added to a GC vial
for analysis with gas chromatography and mass spectrometry.
advantageous since the catalyst can be readily regenerated through
re-calcination in air, eliminating the need for solvents such as
toluene required for resin regeneration [33]. In addition, resins
such as Nafion NR-50 are known to swell in the reaction medium,
thereby introducing inconsistent mass transfer limitations due to
variation in the number of accessible acid sites with time [33].
Because of its high activity, thermal stability, and selectivity,
WOx/ZrO2 (12.6 wt% W) was selected for further investigation.
2.6. Product analysis
Product analysis was carried out using a Varian CP-3800 Gas
Chromatograph/Mass Spectrometer (GC/MS). The products were
detected using a flame ionization detector (FID) to ensure high a
signal to noise ratio, and were identified using the Varian 320 triple
quadrupole mass spectrometer (MS). N-tetradecane was used as an
internal standard to ensure accurate product quantification. FID
response factors were determined using commercially purified
compounds. For all reactions, the main products observed were
ether, alkenes, and water. At longer reaction times, trace amounts
of branched ether and oligomerized olefins were detected, but
were considered to be negligible. Mass balances for all reactions
were achieved to within 5%, with the assumption that one mole
of water is formed for every mole of ether or alkene formed.
3.2. Catalyst characterization
Fig. 1a and b show X-ray Diffraction (XRD) and Raman spectra
for different weight loadings of WOx/ZrO2. For low weight loadings
of tungstated zirconia, the zirconia exists primarily in the mono-
clinic phase. From the XRD in Fig. 1a, the 0 wt% W sample (i) has
peaks at 2h angles of 24°, 28°, 32°, and 56°, which are characteristic
of monoclinic zirconia [26,30,34]. These peaks can also be observed
for the 4 wt% W sample (ii). As the weight loading increases, fea-
tures for tetragonal zirconia become predominant, as evidenced
by the XRD peaks at 2h angles of approximately 30°, 35°, 50°,
and 59° for the 4–22 wt% catalysts [30,34,35]. The presence of
tetragonal zirconia has been attributed to tungsten oxide inhibi-
tion of zirconia sintering and transformation to the monoclinic
phase during calcination [24,36]. When the weight loading of tung-
sten oxide becomes high enough to completely cover the surface of
the zirconia (v-vii), bulk tungsten oxide is formed, as evidenced by
the presence of XRD peaks at 2h angles between 23° and 25° [37].
The Raman spectra for the catalysts are shown in Fig. 1b. Spec-
2.7. Analysis of reaction kinetics
Initial rates of etherification and dehydration of 1-dodecanol
were determined by measuring the initial rates of formation of
ether and alkene. Individual batch reactions for each time point
were carried out at each temperature over the temperature range
of 388–403 K. Mass transfer limitations for linear alcohols were
found to be negligible at 600 RPM for catalyst particle sizes less
trum (i) exhibits bands at 178–220, 307–379, and at 635 cmꢀ1
which confirm that the zirconia is monoclinic [38]. As the tungsten
oxide weight loading increases, a small band appears at 1020 cmꢀ1
,
than 250 lm in diameter (Supporting Information, Fig. S1).
.
3. Results and discussion
This band has been attributed to terminal W@O bonds, character-
istic of oligomeric tungsten oxide species [26]. At weight loadings
above 10.2 wt% W (spectrum (iv)), bands at 274, 715, and 807 cmꢀ1
appear that are characteristic of the WAO bond stretches for bulk
WO3 [39]. For the 12.6 wt% tungstated zirconia, it is estimated that
the Lewis-acid sites on the surface are predominantly due to W
because the exposed Zr has been covered.
Fig. 2a shows the relationship between W weight loading and
Brønsted-acid site density. The surface concentration of
Brønsted-acid sites rises to a maximum at about 12.6 wt% and then
declines. This pattern is very similar to previously reported studies,
in which the appearance of a maximum in the surface concentra-
3.1. Catalyst selection
A number of solid-acid catalysts were screened for the liquid-
phase etherification of 1-dodecanol because of their excellent sta-
bility and recoverability. Table 1 summarizes the results of these
experiments, which were carried out at 393 K and low alcohol con-
versions (<15%), for which the inhibiting effects of products on cat-
alytic activity and selectivity are negligible [28]. Amberlyst 70,
Nafion NR-50, and WOx/ZrO2 (12.6 wt% W) exhibited the highest
ether selectivities (97%, 98%, and 94%, respectively). Amberlyst 70
is a Brønsted-acidic styrene–divinylbenzene resin with a high den-
sity of sulfonic acid groups, and Nafion NR-50 has Brønsted-acidic
sulfonic acid groups enhanced by an electron-withdrawing
polyfluorocarbon backbone [18,21]. While Amberlyst 70 and
Nafion NR-50 are highly selective to ether, they are unstable above
463 K [18,21]. The high thermal stability of tungstated zirconia is
tion of Brønsted-acid sites at
a tungsten weight loading of
12.6 wt% was attributed to the formation of a polytungstate mono-
layer [28,31,40]. Fig. 2b demonstrates that the surface area also
reaches a maximum at a tungsten weight loading of 12.6%. This
effect is attributed to the fact that the addition of tungsten oxide
inhibits the sintering of zirconia to the denser monoclinic phase,
Table 1
Screening of acid catalysts for the etherification of 1-dodecanol to di-n-dodecyl ether.
Entry
Acid catalyst
Dodecanol conversion (%)
Selectivity to ether (%)
Selectivity to alkenes/other (%)
1
2
3
4
5
6
7
8
Zeolite Beta
Para-Toluene Sulfonic Acid
Amberlyst 15
Amberlyst 36
Amberlyst 70
5
3
4
6
11
10
14
0
0
0
61
40
46
65
97
98
94
0
0
0
0
0
39
60
54
35
3
2
6
0
0
Nafion NR-50
WOx/ZrO2 (12.6 wt% W)
ZrO2, Amorphous
ZrO2, Monoclinic
Gamma-Alumina
Mesostructured Silica
Mesostructured Aluminosilicate
9
10
11
12
0
0
0
0
0
Reaction Conditions: 393 K, 4 h, 600 RPM, 500 mg 1-dodecanol, 150 mg n-tetradecane as internal standard. Entry 1: 3.3 * 10ꢀ5 mol eq. H+ acid sites, entries 2–5: 1*10ꢀ4 mol
eq. H+ acid sites, entry 6: 0.043 g. cat, entry 7: 9.1*10ꢀ6 mol eq. H+ sites, entries 8–12: 100 mg catalyst.

16
J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
Fig. 1. (a) X-ray diffraction spectra of catalysts with tungsten weight loadings of (i) 0% (ii) 4.1%, (iii) 7.7%, (iv) 10.2%, (v) 12.6%, (vi) 15.4%, and (vii) 22.3%. (b) Raman spectra of
catalysts with tungsten weight loadings of (i) 0%, (ii) 4.1%, (iii) 7.7%, (iv) 10.2%, (v) 12.6%, (vi) 15.4%, and (vii) 22.3%. Abbreviations: monoclinic zirconia (MZ), bulk WO3 (BW),
tetragonal zirconia (TZ).
Fig. 2. (a) Brønsted-acid site density [mol H+ mꢀ2] versus W weight loading [%], (b) Surface area [m2 gꢀ1] versus W weight loading [%].
Table 2
resulting in a higher surface area [41,42]. However, once the mono-
Characterization of 4–22 wt% W/ WOX/ZrO2.
layer of polymeric tungsten oxide is exceeded, the formation of
bulk WO3 increases the mass of the catalyst but does not provide
any additional active surface area, and is thus responsible for the
decrease in surface area per mass of catalyst for the 15% and 22%
W catalysts. The BET surface area, the concentration of Brønsted-
acid sites, and the ratio of Brønsted- to Lewis-acid sites as func-
tions of W weight loading are given in Table 2. The DRIFTS spectra
of adsorbed pyridine used to determine the ratio of Brønsted- to
Lewis-acid sites on the 4–22 wt% catalysts are provided in Fig. S4
of the Supporting Information.
Weight loading
of W (wt% W,
WOx/ZrO2)
BET
Brønsted Acid Sites
(mol eq. H+ mꢀ2
Ratio of Brønsted/
Lewis Acid Sites
(393 K)
Surface
Area
)
(m2 gꢀ1
)
4.1
7.7
10.2
12.6
15.4
22.3
33
46
50
51
50
41
2.09 * 10ꢀ7 1.2 * 10ꢀ7 0.33 0.01
1.26 * 10ꢀ7 6.5 * 10ꢀ8 0.57 0.24
1.74 * 10ꢀ6 4.0 * 10ꢀ8 0.92 0.05
1.78 * 10ꢀ6 1.4 * 10ꢀ7 0.90 0.02
1.38 * 10ꢀ6 1.6 * 10ꢀ7 0.30 0.11
1.17 * 10ꢀ6 1.7 * 10ꢀ7 0.28 0.04
The surface concentrations of active Brønsted- and Lewis-acid
sites can be determined by poisoning them with a strongly bound
adsorbate. DRIFTS spectra of adsorbed pyridine were obtained and
used to determine the ratio of Brønsted- to Lewis-acid sites for the
4–22 wt% catalysts. As observed in Fig. 3a for the 12.6 wt% catalyst,
bands at 1609 and 1444 cmꢀ1 indicate the presence of Lewis-acid
sites, and bands at 1639 and 1540 cmꢀ1 indicate the presence of
Brønsted acid sites upon addition of pyridine [30,43]. Extinction
coefficients published by Emeis were used to determine the ratio
of Brønsted- to Lewis-acid sites from the intensities of the bands
at 1444 cmꢀ1 and 1540 cmꢀ1 [44]. As shown in Table 2, the ratio
of Brønsted- to Lewis-acid sites increases with increasing weight
loading of W until 12.6 wt%, beyond which the ratio decreases.
Below 12.6 wt% W the Lewis-acid sites come from both tungsten
and zirconia, but above 12.6 wt% W the Lewis-acid sites are pre-
dominately from tungsten. The decrease in the ratio of Brønsted-

J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
17
Fig. 3. (a) DRIFTS spectra for the adsorption of pyridine onto WOx/ZrO2 (12.6 wt% W) at (i) 393 K, (ii) 423 K, (iii) 473 K, (iv) 523 K, and (v) 573 K. (b) DRIFTS spectra for the
adsorption of 2,6-di-tert-butyl-pyridine onto WOx/ZrO2 (12.6 wt% W) at (i) 393 K, (ii) 423 K, (iii) 473 K, (iv) 523 K, and (v) 573 K. Intensities normalized by the Kubelka-Munk
function.
to Lewis-acid sites above 12.6 wt% is attributed to the formation of
bulk tungsten oxide, as suggested by the structural
characterization.
other biomass-derived alcohols [48,49]. As discussed below, we
propose that the mechanism for liquid-phase dodecanol etherifica-
tion over tungstated zirconia that is most consistent with the
kinetics reported here involves both Brønsted- and Lewis-acid sites
and that the mechanism of dodecanol dehydration involves only
Brønsted-acid sites.
Pyridine poisons both Brønsted- and Lewis-acid sites, whereas
2,6-di-tert-butyl pyridine,
a hindered base, will only poison
Brønsted-acid sites [45]. Fig. 3b shows the diffuse reflectance infra-
red spectra for 2,6-di-tert-butyl pyridine adsorbed on the 12.6 wt%
W catalyst. The spectrum of adsorbed pyridine in Fig. 3a exhibits
bands at 1540 cmꢀ1 and 1444 cmꢀ1, attributable to pyridine inter-
acting with Brønsted- and Lewis-acid sites, respectively. By con-
trast, the spectrum for adsorbed 2,6-di-tert-butyl pyridine,
shown in Fig. 3b, only exhibits a band at 1540 cmꢀ1 for pyridine
adsorption onto Brønsted-acid sites, but the band at 1444 cmꢀ1
for pyridine adsorption onto Lewis-acid sites is absent.
The proposed mechanisms for etherification and dehydration of
dodecanol over tungstated zirconia are shown in Scheme 1. We
envision that the first step for both etherification and dehydration
is the reversible adsorption of the alcohol onto a Brønsted-acid site.
While the alcohol in Scheme 1 is shown as molecularly adsorbed, it
is possible that it may adsorb dissociatively to form alkoxide spe-
cies [46,50]; however, the differentiation between associative
and dissociative adsorption was not considered in the present
study. For etherification, the second alcohol molecule is assumed
to adsorb onto an adjacent Lewis-acid site, and the rate-limiting
step (RLS) is taken to be the reaction of the two adsorbed species
to form adsorbed water and ether, which then desorb in the final
step. The rate-limiting step for dehydration is the reaction of an
alcohol molecule adsorbed at a Brønsted-acid site with a vacant
basic site adjacent to the Brønsted-acid site, resulting in the
abstraction of a hydrogen atom from the b-carbon of the adsorbed
alcohol. The final step in the dehydration pathway is the desorp-
tion of water. As will be discussed below, ether and water can inhi-
bit the rates of dodecanol etherification and dehydration.
3.3. Mechanisms and kinetics of dodecanol etherification and
dehydration
A number of mechanisms have been proposed to explain the
etherification and dehydration of alcohols over heterogeneous acid
catalysts; these include dual-site Langmuir-Hinshelwood type
mechanisms, Eley-Rideal type mechanisms with only one adsorbed
alcohol, and mechanisms with more complex dimers and trimers
of adsorbed reactants and products. Cunill et al. have proposed a
modified Eley-Rideal model for the Brønsted-acid catalyzed
liquid-phase etherification of 1-hexanol over Amberlyst 70 in
which water remains adsorbed on the catalyst surface and, there-
fore, competes with the alcohol for acid sites [17]. Bhan et al. have
proposed that the gas-phase etherification of ethanol over the
To identify the rate-limiting steps for the liquid-phase etherifi-
cation and dehydration of dodecanol, experiments were carried out
using deuterium-labeled alcohols. The catalyst for these experi-
ments was WOx/ZrO2 (12.6 wt% W). NMR was used to identify
which H(D) atoms are involved in each reaction. Dehydration
and etherification reactions of deuterium-labeled 1-hexanol
(hexan-1,1-d2-1-ol and hexan-2,2-d2-1-ol) were used to probe
Lewis-acidic c-Al2O3 is inhibited by alcohol-water dimers on the
catalyst surface [46]. Iglesia et al. have proposed that the gas-
phase dehydration of ethanol over polyoxometalates involves
adsorbed dimer and trimer species [47]. In the case of tungstated
zirconia, different authors have claimed that the active centers
for alcohol dehydration are Lewis- or Brønsted-acid sites. Larsen
et al. have proposed that the dehydration of 1-propanol, 2-
propanol, and tert-butanol over tungstated zirconia proceeds
solely over Lewis-acid sites [27], whereas Iglesia et al. have pro-
posed that 2-butanol dehydration occurs only over Brønsted-acid
sites [28]. Theoretical and experimental evidence for synergistic
effects between Lewis- and Brønsted-acid sites have also been sug-
gested for metal oxide catalysis in the etherification of glycerol and
the movement of hydrogen atoms on the
a and b carbons of the
alcohol. 1H NMR of the ether and alkene products of the hexan-
2,2-d2-1-ol reaction showed no evidence of hydrogen on the b-
carbon of the alkene, and no hydrogen present on either of the b-
carbons of the ether. These results indicate that the hydrogens
on the b-carbon are not involved in etherification, but are involved
in the irreversible dehydration step. Also, the 1H NMR of the prod-
ucts from the reaction of hexan-1,1-d2-1-ol showed no change in
the position of deuterium atoms on the
a-carbon. In both deuter-
ated alcohols, the lack of D/H scrambling indicates that etherifica-

18
J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
Scheme 1. Proposed mechanism for liquid phase primary linear alcohol etherification and unimolecular dehydration over tungstated zirconia.
Table 3
Measured kinetic isotope effects for hexene and di-n-hexyl ether synthesis at 418 K
for the dehydration of hexan-1,1-d2-1-ol and hexan-2,2-d2-1-ol.
hydrogen bond cleavage was the rate limiting step for unimolecu-
lar dehydration of propanol over -Al2O3 [51].
c
Product
Reactant
The kinetics of dodecanol etherification and dehydration based
on Scheme 1 involve the equilibrium adsorption of alcohol on
Brønsted- and Lewis-acid sites and the rates of the reactions lead-
ing to alcohol etherification and dehydration. These relationships
are given by Eqs. (1)-(4):
hexan-1,1-d2-1-ol
hexan-2,2-d2-1-ol
Di-n-Hexyl Ether KIE (kH/kD)
Hexene KIE (kH/kD)
1.00 0.09
1.01 0.09
1.03 0.08
1.64 0.09
KBA
½ROHꢁ þ ½BAꢁ $ ½ROH ꢀ BAꢁ
ð1Þ
tion and dehydration are irreversible under the conditions of the
experiment. The NMR spectra are provided in Figs. S5a–S5f in the
supporting information.
KLA
½ROHꢁ þ ½LAꢁ $ ½ROH ꢀ LAꢁ
ð2Þ
ð3Þ
Measurement of kinetic isotope effects was used to confirm that
the rate-limiting step for 1-hexanol dehydration is cleavage of the
b-carbon CAH bond. Table 3 shows that there is a kinetic isotope
effect for unimolecular dehydration of hexan-2,2-d2-1-ol to form
hexene (kH/kD = 1.64 0.09), but there is no kinetic isotope
effect for bimolecular etherification to form di-hexyl ether
(kH/kD = 1.03 0.08), confirming that the rate limiting step for
alkene formation is b-carbon hydrogen bond cleavage. No kinetic
isotope effect was observed for etherification or unimolecular
dehydration reactions over hexan-1,1-d2-ol, confirming that the
re ¼ ke½ROH ꢀ LAꢁ½ROH ꢀ BAꢁ
rd ¼ kd½ROH ꢀ BAꢁ½LBꢁ
ð4Þ
Substitution of Eqs. (1) and (2) into Eqs. (3) and (4) gives:
2
re ¼ keKLAKBA½ROHꢁ ½LAꢁ½BAꢁ
ð5Þ
ð6Þ
rd ¼ kdKBA½ROHꢁ½BAꢁ½LBꢁ
a
carbon-hydrogen bond is not involved in the rate limiting step.
The overall rates of etherification and dehydration based on this
set of elementary steps and site balances on the Brønsted-acid sites
[BA], Lewis-acid sites [LA], and Lewis-base sites [LB] are given by
the following expressions (see Supporting Information for the
derivation of these expressions):
Therefore, the rate limiting step for etherification must involve
either formation of the CAO bond or desorption of the ether from
the acid site. This finding is in agreement with the results of Bhan
et al., who used kinetic isotope effects to confirm that the b-carbon
2
keKLAKBA½ROHꢁ
re ¼
ð7Þ
ð1 þ KBA½ROHꢁ þ KRORꢀBA½RORꢁ þ KH2OꢀBA½H2OꢁÞð1 þ KLA½ROHꢁ þ KRORꢀLA½RORꢁ þ KH2OꢀLA½H2OꢁÞ
kdKBA½ROHꢁ
rd ¼
ð8Þ
ð1 þ KBA½ROHꢁ þ KRORꢀBA½RORꢁ þ KH2OꢀBA½H2OꢁÞð1 þ KLB½ROHꢁ þ KRORꢀLB½RORꢁ þ KH2OꢀLB½H2OꢁÞ

J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
19
In Eqs. (7) and (8), ke is the rate coefficient for the rate-limiting
step of etherification, KLA is the equilibrium constant for adsorption
of an alcohol onto a Lewis-acid site, KBA is the equilibrium constant
for adsorption of an alcohol onto a Brønsted-acid site, KROR-BA is the
equilibrium constant for ether adsorption onto a Brønsted-acid site,
KH2O-BA is the equilibrium constant for water adsorption onto a
Brønsted-acid site, KROR-LA is the equilibrium constant for ether
adsorption onto a Lewis-acid site, KH2O-LA is the equilibrium con-
stant for water adsorption onto a Lewis-acid site, kd is the rate con-
stant for the rate-limiting step of dehydration, KLB is the equilibrium
constant for adsorption of an alcohol onto a basic site, KROR-LB is the
equilibrium constant for adsorption of an ether onto a basic site, and
KH2O-LB is the equilibrium constant for adsorption of water onto a
basic site. For low conversions where [ROR] and [H2O] are negligi-
ble, and with the assumption that the Brønsted-acid sites are satu-
rated with reactant (KBA[ROR] ꢂ 1), Eqs. (7) and (8) simplify to:
rates of di-dodecyl ether and dodecene formation at varying initial
concentrations of dodecanol in decane over a temperature range of
388–403 K, and the results are shown in Fig. 4. The solid curves in
these figures represent a fit to the data by Eqs. (9) and (10). The
rate coefficients and equilibrium constants required to achieve
these fits are given in Table 4. It is clear from Fig. 4a and b, and
from the parity plots shown in Fig. 4c and d, that there is good
agreement between the predicted and observed rates for ether
and alkene formation for temperatures between 388 and 403 K.
The rate coefficients and equilibrium constants ke, kd, KLA, and
KLB and the respective activation energies and enthalpies of
adsorption were determined from Arrhenius and Van’t Hoff plots
(Supporting Information, Fig. S2), and the values of these parame-
ters are given in Table 4. The apparent activation energy for ether-
ification (128 kJ molꢀ1), is significantly lower than the activation
energy for unimolecular dehydration (152 kJ molꢀ1). The value of
the apparent activation energy for etherification lies in the range
of values reported for liquid-phase etherification of 1-hexanol
and 1-octanol over other solid acid catalysts such as H-BEA,
Amberlyst 70, and Nafion NR-50, which range from 118 to 150
kJ molꢀ1 [17,19,20]. The unimolecular dehydration activation bar-
rier also falls within the range of values reported for the dehydra-
keKLA½ROHꢁ
re ¼
ð9Þ
1 þ KLA½ROHꢁ
kd
rd ¼
ð10Þ
1 þ KLB½ROHꢁ
tion of linear alcohols over solid acids such as TiO2, ZrO2, and c-
3.4. Rates of dodecanol etherification and dehydration
Al2O3, which range from 141 to 171 kJ molꢀ1.[52,53] The relatively
high activation barrier for unimolecular dehydration compared to
etherification observed in this study over tungstated zirconia
The effect of alcohol concentration on the initial rates of ether-
ification and dehydration was determined by measuring the initial
Fig. 4. (a) Effect of initial dodecanol concentration in decane [M] on initial rates of di-dodecyl ether formation [M hꢀ1 mꢀ2], (b) effect of initial dodecanol concentration in
decane [M] on initial rates of dodecene formation [M hꢀ1 mꢀ2] for reactions from 388 to 403 K with 100 mg WOx/ZrO2 (12.6 wt% W) and a reaction volume of 2.4 * 10ꢀ4 L, (c)
parity plot for ether formation, and (d) parity plot for alkene formation.

20
J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
Table 4
Kinetic parameters.
Temperature [K]
ke [M hꢀ1 mꢀ2
]
kd [M hꢀ1 mꢀ2
]
KLA [Mꢀ1
]
KLB [Mꢀ1
]
388
393
398
403
0.010
0.018
0.026
0.046
0.004
0.007
0.013
0.024
11.5
7.50
7.06
5.53
0.459
0.391
0.341
0.400
Ea [kJ molꢀ1
]
128
35
8
152
42
2
ln[A] [M hꢀ1 mꢀ2
]
3
1
D
D
Hads [kJ molꢀ1
S [J molꢀ1 Kꢀ1
]
]
ꢀ59 12
ꢀ15 14
ꢀ44 35
ꢀ132 31
explains the higher selectivity to ether in the liquid phase than the
gas phase.
distinguish between ether added initially and ether formed, di-
decyl ether was used for ether inhibition experiments instead of
di-dodecyl ether. The results indicate that water inhibits both ether
and alkene formation, while the di-decyl ether only inhibits ether
formation and not dehydration. These data are given in the Sup-
porting Information, Fig. S3. With the assumption that saturation
of Brønsted-acid sites is unaffected by the presence of water and
ether, but ether and water inhibit adsorption on Lewis-acid and
base sites, the rate expressions for etherification and dehydration
now become:
3.5. Inhibition of dodecanol etherification and dehydration by water
and ether
Water has been shown to inhibit Brønsted-acid sites during the
liquid-phase etherification of 1-hexanol over Amberlyst 70 [17],
and it is reasonable to expect that ether may have a similar inhibi-
tory effect. To assess the influence of water and ether on the rates
of etherification and dehydration of dodecanol, the initial rates of
ether and alkene formation were measured at 393 K with varying
initial concentrations of water and ether added to the reactor. To
keKLA½ROHꢁ
re ¼
ð11Þ
ð1 þ KLA½ROHꢁ þ KRORꢀLA½RORꢁ þ KH2OꢀLA½H2OꢁÞ
Fig. 5. Concentration [M] versus time [h] for experimental data and model from 393 K – 403 K for (a) 1-dodecanol, (b) di-dodecyl ether, (c) dodecene, and (d) water over
WOx/ZrO2 (12.6 wt% W). Linear portions of the time-course curves for low conversions are shown in the insets.

J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
21
kd
Brønsted acid site density reaches a maximum. It is clear from
Fig. 6b that the initial rates of 1-dodecanol etherification and dehy-
dration have a nearly linear relationship with Brønsted-acid site
density and that the selectivity to ether does not change signifi-
cantly with Brønsted-acid site density. Characterization of the cat-
alysts (see Table 2) indicates that the ratio of Brønsted- to Lewis-
acid sites changes with weight loading of tungsten, suggesting that
the selectivity to ether is independent of the ratio of Brønsted- to
Lewis-acid sites. However, as the Brønsted-acid site density
increases, so does the fraction of Lewis-acid sites attributed to W,
thus the rate of reaction may also increase with increasing W
Lewis-acid site density.
As demonstrated from the DRIFTS characterization in
Fig. 3a and b, pyridine poisons both Brønsted- and Lewis-acid sites,
whereas 2,6-di-tert-butyl pyridine poisons only Brønsted acid
sites. To test the effects of acid-site poisoning on the rates of ether
and alkene formation, the initial rates of etherification and dehy-
dration were measured for reactions with different initial concen-
trations of pyridine or 2,6-di-tert-butyl pyridine added to the
reaction mixture. These results are shown in Fig. 7. The rate of
ether formation is observed to decrease as the concentration of
the base increases, with the decrease in the rate being somewhat
smaller for pyridine than for 2,6-di-tert-butyl pyridine poisoning.
These results clearly demonstrate that both poisons are effective
in blocking the adsorption of alcohol onto Brønsted-acid sites, a
critical initial step for etherification.
Because the Brønsted- to Lewis-acid site ratio for WOx/ZrO2
(12.6 wt% W) is approximately 0.9, if Lewis-acid sites were not at
all active in the etherification reaction, we would predict that
approximately 1.8 mol of pyridine would be necessary to achieve
the same decrease in rate as 1 mol of 2,6-di-tert-butyl pyridine.
However, for this reaction, the decrease in the etherification rate
is nearly the same for pyridine and hindered pyridine. This obser-
vation is consistent with the proposed mechanism, indicating that
both Brønsted- and Lewis-acid sites play a role in the etherification
reaction. Upon poisoning both Brønsted- and Lewis-acid sites with
pyridine, the rate of ether formation reaches zero; however, after
poisoning only the Brønsted-acid sites on the surface with hin-
dered pyridine, some ether formation can still occur. This suggests
that the Lewis-acid sites may be active for etherification on their
own. The fact that Lewis-acid sites are still active supports the
hypothesis that the surface is dominated by W Lewis-acid sites
as opposed to Zr Lewis-acid sites because zirconia itself is not
active for etherification (Table 1, entries 8–9). We also note that
rd ¼
ð12Þ
ð1 þ KLB½ROHꢁ þ KRORꢀLB½RORꢁ þ KH2OꢀLB½H2OꢁÞ
The values of the rate coefficients and equilibrium constants
appearing in Eqs. (11) and (12), ke, kd, KLA, and KLB, were fixed at
the values given in Table 4 and the equilibrium constants KROR-LA
H2O-LA, KROR-LB, and KH2O-LB were then fitted to the data using
the following constraints: First, the value of KROR-LB was taken to
be approximately equal to 0 based on ether-inhibition experiments
indicating that ether only inhibits the rate of etherification. Second,
when considering acid site balances over both Lewis-acid sites and
Lewis-base sites, it was assumed that water adsorbs onto an acid-
base pair; therefore, KH2O-LA must be equal to KH2O-LB. Finally, inhi-
bition studies suggested that water has a greater inhibiting effect
,
K
on both reactions, implying that KROR-LA < KH2O-LA and KH2O-LB
.
Using these constraints, data for the complete conversion of pure
dodecanol at 393 K, 398 K, and 403 K was used to estimate values
for KROR-LA, KH2O-LA, and KH2O-LB. The experimental curves for con-
centration of reactants and products versus time and the fits of
the model to these data are shown in Fig. 5. The equilibrium con-
stants and adsorption energies obtained in this manner are shown
in Table 5.
3.6. Role of Brønsted- and Lewis-acid sites
The role of the Brønsted-acid sites was probed by examining the
relationship between reaction rates and Brønsted-acid site density,
whereas poisoning experiments were used to probe the roles of
both Brønsted- and Lewis-acid sites.
Initial rates of etherification and dehydration were measured
over tungstated zirconia with varying Brønsted-acid site densities.
The initial rates of etherification and dehydration across the 4–
22 wt% catalysts are shown in Fig. 6a. The rates of etherification
and dehydration peak between 10.2 and 12.6 wt% W, where the
Table 5
Kinetic parameters for water and ether inhibition.
Temperature [K]
KROR-LA [Mꢀ1
]
KH2O-LA [Mꢀ1
]
KH2O-LB [Mꢀ1
]
393
398
403
3
2.95
2.9
ꢀ4
4
3.8
3.6
4
3.8
3.6
D
D
H [kJ molꢀ1
S [J molꢀ1 Kꢀ1
]
ꢀ14
ꢀ14
]
ꢀ2.2 0.14
ꢀ44 35
ꢀ24 0.78
Fig. 6. (a) Initial rates of ether and alkene formation [M hꢀ1 mꢀ2] versus W weight loading, (b) Brønsted-acid site density [mol H+ mꢀ2] versus rates of ether and alkene
formation [M hꢀ1 mꢀ2]. Reaction conditions: 393 K, 100 mg WOx/ZrO2 (12.6 wt% W), 600 RPM.

22
J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
tion and dehydration. This model was then modified to include the
inhibiting effects of ether and water on both reactions. It was found
that the rates of etherification and dehydration correlate linearly
with Brønsted-acid site density, supporting the hypothesis that
both reactions require Brønsted-acid sites. Due to the absence of
activity for etherification over pure zirconia, and the relationship
between tungsten oxide weight loading and reaction rate, it was
concluded that the active Lewis-acid sites are associated with the
dispersed tungsten oxide as opposed to zirconia. Poisoning of the
catalyst with pyridine and hindered pyridine indicates that
Lewis-acid sites are active for etherification but not dehydration,
which is consistent with the proposed mechanism.
Acknowledgements
This work was funded by the Energy Biosciences Institute. J.
Rorrer would also like to acknowledge funding from the National
Science Foundation Graduate Research Fellowship under Grant
No. DGE 1106400. The authors thank Chris Ho, Lin Louie, and Joa-
quin Resasco for assistance with XRD, FTIR, and Raman spec-
troscopy, respectively. The authors also thank Adam Grippo,
Konstantinos Goulas, and Miriam Eulers for useful discussion,
and Alan Liu for assisting in reaction preparation.
Fig. 7. Initial rates of ether formation [M hꢀ1 mꢀ2] with varying initial concentra-
tions [M] of 2,6-di-tert-butyl pyridine (DTBP) (black square) and pyridine (blue
diamond), and initial rates of alkene formation with varying initial concentrations
of DTBP (orange circle) and pyridine (red triangle).
no alkene is formed once all of the Brønsted-acid sites are poi-
soned, consistent with the proposed mechanism. Based on the pro-
posed mechanism and the resulting expressions for the reaction
rates, we anticipate that the overall selectivity to ether would
decrease if the tungsten Lewis-acid sites could be poisoned with-
out altering the Brønsted-acid site density.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2017.08.001.
The dotted vertical line in Fig. 7 represents the concentration of
base equal to the measured concentration of Brønsted-acid sites
determined by base titration. There are several possible reasons
why the concentration of Brønsted-acid sites measured by ion
exchange and titration is not equal to the concentration of hin-
dered pyridine at which the rate becomes zero. First, given that
there is a distribution in acid site strength, there may be some acid
sites that are not active for etherification or dehydration. Second,
the presence of a base on a Brønsted- or Lewis-acid site may alter
the acidity of adjacent Lewis- or Brønsted-acid sites, respectively.
Finally, there may be steric factors responsible for the enhanced
decrease in rate. Bond lengths for WAO are on the order of 2 ang-
stroms, and OAH bonds are on the order of 0.8 angstroms [54],
while the width of pyridine is on the order of 2.8 angstroms [55],
making it possible that a single pyridine molecule could inhibit
multiple acid sites.
References:
[1] L. Chapman, J. Transp. Geogr. 15 (2007) 354–367.
[2] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044–4098.
[3] M. Balakrishnan, E.R. Sacia, S. Sreekumar, G. Gunbas, A.A. Gokhale, C.D. Scown,
F.D. Toste, A.T. Bell, Proc. Natl. Acad. Sci. 112 (2015) 7645–7649.
[4] C. Carlini, A. Macinai, A.M. Raspolli Galletti, G. Sbrana, J. Mol. Catal. A Chem.
212 (2004) 65–70.
[5] R.J.J. Nel, A. De Klerk, Ind. Eng. Chem. Res. 48 (2009) 5230–5238.
[6] M.J. Truitt, Condensation of Alcohols for Biofuel Production, US 8845766 B2,
2014.
[7] A.L. Maximov, A.I. Nekhaev, D.N. Ramazanov, Pet. Chem. 55 (2015) 1–21.
[8] G. Olah, Cleaner Burning and Cetane Enhancing Diesel Fuel Supplements, US
Patent 5520710 A, 1996.
[9] M.J. Pilling, Low-Temperature Combustion and Autoignition, Elsevier, 1997.
[10] K.W. Wolfgang Beilfuss, Ralf Gradtke, Wolfgang Siegert, Karl-Heinz Diehl,
Glyceryl Ethers as Preservatives for Cooling Lubricants, US Patent 7268102 B2,
2007.
[11] D.A.L. Leslie, R. Rudnick, Ross A. Kremer, Alkylated Diphenyl Ether Lubricants,
Patent US 5552071 A, 1996.
[12] L.R. Rudnick, Synthetics, Mineral Oils, and Bio-Based Lubricants: Chemistry
and Technology, Second Edition., CRC Press, Taylor And Francis Group, Boca
Raton, 2013.
4. Conclusions
[13] D. Jadhav, A.M. Grippo, S. Shylesh, A.A. Gokhale, J. Redshaw, A.T. Bell,
ChemSusChem 10 (2017) 2527–2533.
[14] T. Zolper, Z. Li, C. Chen, M. Jungk, T. Marks, Y.-W. Chung, Q. Wang, Tribol. Lett.
48 (2012) 355–365.
[15] R. Benda, J. Bullen, A. Plomer, J. Synth. Lubr. 13 (1996) 41–57.
[16] M. Balakrishnan, E.R. Sacia, A.T. Bell, Green Chem. 14 (2012) 1626–1634.
[17] R. Bringué, E. Ramírez, M. Iborra, J. Tejero, F. Cunill, Chem. Eng. J. 246 (2014)
71–78.
[18] R. Bringué, M. Iborra, J. Tejero, J. Izquierdo, F. Cunill, C. Fite, V. Cruz, J. Catal. 244
(2006) 33–42.
[19] E. Medina, R. Bringué, J. Tejero, M. Iborra, C. Fité, Appl. Catal. A Gen. 374 (2010)
41–47.
[20] I. Hoek, T.A. Nijhuis, A.I. Stankiewicz, J.A. Moulijn, Appl. Catal. A Gen. 266
(2004) 109–116.
[21] K. Arata, Adv. Catal. 37 (1990) 165–211.
[22] G. Perot, M. Guisnet, J. Mol. Catal. 61 (1990) 173–196.
[23] M. Hino, K. Arata, J. Chem. Soc. Chem. Commun. (1988) 1259–1260.
[24] W. Zhou, N. Soultanidis, H. Xu, M.S. Wong, M. Neurock, C.J. Kiely, I.E. Wachs,
ACS Catal. 7 (2017) 2181–2198.
[25] M.K. Patil, A.N. Prasad, B.M. Reddy, Curr. Org. Chem. 15 (2011) 3961–3985.
[26] D.E. López, K. Suwannakarn, D.A. Bruce, J.G. Goodwin, J. Catal. 247 (2007) 43–
50.
Tungstated zirconia is shown to be active and highly selective
for the etherification of 1-dodecanol. It is proposed that the mech-
anism for liquid-phase etherification of 1-dodecanol over this cat-
alyst involves a cooperation between Brønsted- and Lewis-acid
sites that facilitates the bi-molecular etherification reaction. Iso-
topic labeling studies suggest that the rate-limiting step for ether-
ification involves the formation of a carbon oxygen bond between
two adsorbed alcohol molecules, and that the rate-limiting step for
dehydration is the cleavage of the bond between the b-carbon and
a hydrogen atom in an adsorbed alcohol. The proposed mechanism
suggests that both etherification and dehydration require alcohol
adsorption onto
a Brønsted-acid site and that etherification
requires an additional alcohol adsorbed onto an adjacent W
Lewis-acid site, while dehydration requires an open Lewis base
site. A kinetic model based on the proposed mechanism (Scheme 1)
was developed and fitted to initial reaction rate data for etherifica-
[27] G. Larsen, E. Lotero, L.M. Petkovic, D.S. Shobe, J. Catal. 169 (1997) 67–75.

J. Rorrer et al. / Journal of Catalysis 354 (2017) 13–23
23
[28] J. Macht, C.D. Baertsch, M. May-Lozano, S.L. Soled, Y. Wang, E. Iglesia, J. Catal.
227 (2004) 479–491.
[29] S. Peng, N.M. Okeley, A.-L. Tsai, G. Wu, R.J. Kulmacz, W.A. Van der Donk, J. Am.
Chem. Soc. 124 (2002) 10785–10796.
[30] Y. Zhao, W. Li, M. Zhang, K. Tao, Catal. Commun. 3 (2002) 239–245.
[31] C.D. Baertsch, S.L. Soled, E. Iglesia, J. Phys. Chem. B 105 (2001) 1320–1330.
[32] A. Wang, N.P. Balsara, A.T. Bell, Green Chem. 18 (2016) 4073–4085.
[33] J.A. Melero, J. Iglesias, G. Morales, Green Chem. 11 (2009) 1285–1308.
[34] S.R. Ginjupalli, S. Mugawar, P. Rajan, N., P. Kumar Balla, V. R. C. Komandur, J.
Chem. Technol. Biotechnol. 89 (2014) 1890–1897.
[35] G. Larsen, E. Lotero, S. Raghavan, R.D. Parra, C.A. Querini, Appl. Catal. A Gen.
139 (1996) 201–211.
[36] D.G. Barton, S.L. Soled, G.D. Meitzner, G.A. Fuentes, E. Iglesia, J. Catal. 181
(1999) 57–72.
[37] R. Balzer, V. Drago, W.H. Schreiner, L.F.D. Probst, J. Braz. Chem. Soc. 25 (2014)
2026–2031.
[38] M. Li, Z. Feng, G. Xiong, P. Ying, Q. Xin, C. Li, J. Phys. Chem. B 105 (2001) 8107–
8111.
[41] K. Shimizu, T.N. Venkatraman, W. Song, Appl. Catal. A Gen. 225 (2002) 33–41.
[42] E.I. Ross-Medgaarden, W.V. Knowles, T. Kim, M.S. Wong, W. Zhou, C.J. Kiely, I.E.
Wachs, J. Catal. 256 (2008) 108–125.
[43] R. Kourieh, S. Bennici, A. Auroux, React. Kinet. Mech. Catal. 105 (2012) 101–
111.
[44] C.A. Emeis, J. Catal. 141 (1993) 347–354.
[45] C. Baertsch, K.T. Komala, Y. Chua, E. Iglesia, J. Catal. 205 (2002) 44–57.
[46] J.F. DeWilde, H. Chiang, D.A. Hickman, C.R. Ho, A. Bhan, ACS Catal. 3 (2013)
798–807.
[47] W. Knaeble, E. Iglesia, J. Phys. Chem. C 120 (2016) 3371–3389.
[48] G.S. Foo, D. Wei, D.S. Sholl, C. Sievers, ACS Catal. 4 (2014) 3180–3192.
[49] M. Calatayud, A.M. Ruppert, B.M. Weckhuysen, Chemistry 15 (2009) 10864–
10870.
[50] A.J. Jones, E. Iglesia, Angew. Chemie Int. Ed. 53 (2014) 12177–12181.
[51] M. Kang, J.F. DeWilde, A. Bhan, ACS Catal. 5 (2015) 602–612.
[52] P. Kostestkyy, J. Yu, R.J. Gorte, G. Mpourmpakis, Catal. Sci. Technol. 4 (2014)
3861.
[53] S. Roy, G. Mpourmpakis, D. Hong, D.G. Vlachos, A. Bhan, R.J. Gorte, ACS Catal. 2
(2012) 1846–1853.
[39] S. Kuba, P. Concepción Heydorn, R.K. Grasselli, B.C. Gates, M. Che, H. Knözinger,
Phys. Chem. Chem. Phys. 3 (2001) 146–154.
[40] A. Galano, G. Rodriguez-Gattorno, E. Torres-García, Phys. Chem. Chem. Phys.
10 (2008) 4181–4188.
[54] J.T. Szymanski, A.C. Roberts, Can. Mineral. 22 (1984) 681–688.
[55] D. Mootz, H.G. Wussow, J. Chem. Phys. 75 (1981) 1517–1522.