Highly Dispersed SiOx/Al2O3 Catalysts Illuminate the Reactivity of Isolated Silanol Sites

Article abstract of DOI:10.1002/anie.201505452

The reaction of γ-alumina with tetraethylorthosilicate (TEOS) vapor at low temperatures selectively yields monomeric SiO_x species on the alumina surface. These isolated (-AlO)₃Si(OH) sites are characterized by PXRD, XPS, DRIFTS of adsorbed NH₃, CO, and pyridine, and ²⁹Si and ²⁷Al DNP-enhanced solid-state NMR spectroscopy. The formation of isolated sites suggests that TEOS reacts preferentially at strong Lewis acid sites on the γ-Al₂O₃ surface, functionalizing the surface with "mild" Bronsted acid sites. For liquid-phase catalytic cyclohexanol dehydration, these SiO_x sites exhibit up to 3.5-fold higher specific activity than the parent alumina with identical selectivity.

Full text of DOI:10.1002/anie.201505452

.
Angewandte
Communications
International Edition: DOI: 10.1002/anie.201505452
German Edition: DOI: 10.1002/ange.201505452
Heterogeneous Catalysis
Highly Dispersed SiO /Al O Catalysts Illuminate the Reactivity of
x
2
3
Isolated Silanol Sites
Aidan R. Mouat, Cassandra George, Takeshi Kobayashi, Marek Pruski, Richard P. van Duyne,
Tobin J. Marks,* and Peter C. Stair*
Abstract: The reaction of g-alumina with tetraethylorthosili-
cate (TEOS) vapor at low temperatures selectively yields
monomeric SiO species on the alumina surface. These isolated
x
(
-AlO) Si(OH) sites are characterized by PXRD, XPS,
3
29
DRIFTS of adsorbed NH , CO, and pyridine, and Si and
3
2
7
Al DNP-enhanced solid-state NMR spectroscopy. The for-
mation of isolated sites suggests that TEOS reacts preferen-
tially at strong Lewis acid sites on the g-Al O3 surface,
2
functionalizing the surface with “mild” Brønsted acid sites.
For liquid-phase catalytic cyclohexanol dehydration, these
3
+
Figure 1. A) The pentasil unit of zeolite H-ZSM-5 with Al substitution
(
all lattice points, except Al, represent Si atoms, with SiÀO bonds
SiO sites exhibit up to 3.5-fold higher specific activity than the
x
omitted for clarity). B) A proposed Brønsted-acidic site in amorphous
parent alumina with identical selectivity.
SiO –Al O .
2
2
3
S
ilica–aluminas are well-known solid acid catalysts and are
[
8]
employed in both laboratory organic synthesis and in large-
coking. The attraction of tuning solid acid strength has
[
1]
[9]
scale industrial processes. These materials include amor-
phous silica–aluminas, typically prepared by coprecipitation
or sol–gel methods, and ion-exchanged, crystalline alumi-
nosilicate zeolites such as H-ZSM-5 or H-Y. Solid acids
catalyze a variety of transformations, including dehydration,
skeletal isomerization, and cracking, and are used in
processes such as fluid catalytic cracking (FCC) and meth-
anol-to-gasoline (MTG) synthesis. In crystalline zeolites,
been demonstrated in several organic transformations,
including recent work by Gazit and Katz
the efficacy of mildly acidic silanol groups in b-glucan
hydrolysis. In recent work, van Bokhoven and co-workers
demonstrated a controlled process for the synthesis of SiO2–
Al O catalysts via chemical liquid deposition (CLD) grafting
of Al and Si species on SiO and Al O supports.
report a new vapor phase approach to creating isolated,
uniform, and weakly acidic SiO sites on alumina, and explore
[10]
who reported
[2]
[
3]
[10]
[4]
2
3
[7b, 11]
Here we
2
2
3
[
5]
3+
“
strong” Brønsted-acidic sites arise from isomorphous Al
x
substitution in the SiO lattice, yielding “framework” protons
their catalytic properties with respect to a transformation of
biomass relevance. Cyclohexanol dehydration is diagnostic of
2
[6]
as in Figure 1a. Amorphous silica–aluminas also possess
Brønsted acid sites (Figure 1b). Solid acids typically present
a range of acid site strengths according to local coordination
environments. Significant spectroscopic and physicochemical
studies have been undertaken to characterize the strength and
[12]
solid acid catalytic activity
and permits assessment of
selectivity with respect to unwanted side reactions and coking,
[13]
a traditional challenge to solid acid catalysis.
We employ a strategy here that combines the mild
conditions of a low temperature vapor-phase deposition
process and the specific surface reactivity of g-Al O to
[7]
catalytic relevance of the different acid sites.
In spite of advances in the synthesis and post-synthetic
modification of solid acids, the non-uniformity of acid site
strengths on solid acid catalysts remains a challenge. This non-
uniformity can significantly compromise selectivity, leading to
unwanted side products and ultimate deactivation by
2
3
realize a new class of SiO /Al O catalysts featuring well-
x
2
3
defined, isolated Si active sites having the structure (-AlO) Si-
3
(OH) (Scheme 1). These species are characterized by powder
X-ray diffraction (PXRD), X-ray photoelectron spectroscopy
29
27
(
XPS), Si and Al solid-state NMR enhanced by dynamic
[
*] A. R. Mouat, C. George, Prof. R. P. van Duyne, Prof. T. J. Marks,
Prof. P. C. Stair
Chemistry Department and the Center for Catalysis and Surface
Science, Northwestern University
2
145 Sheridan Rd., Evanston, IL 60208 (USA)
E-mail: t-marks@northwestern.edu
pstair@northwestern.edu
Dr. T. Kobayashi, Prof. M. Pruski
U.S. DOE Ames Laboratory, and Department of Chemistry
Iowa State University
Ames, IA 50011 (USA)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201505452.
Scheme 1. Vapor-phase synthesis of (-AlO) Si(OH) sites on g-Al O .
3
2
3
Proposed active site in SiO /Al O catalysts, A.
x
2
3
1
3346
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 13346 –13351

Angewandte
Chemie
nuclear polarization (DNP), and diffuse
reflectance infrared Fourier transform
spectroscopy (DRIFTS) of adsorbed
NH , CO, and pyridine, followed by
3
catalytic studies. A linear relationship is
demonstrated between the Si site density
and surface Brønsted acidity. We also
show that these SiO /Al O catalysts are
x
2
3
competent for the catalytic dehydration
of cyclohexanol to cyclohexene, exhibit-
ing up to 3.5-fold higher specific activity
than the parent g-Al O with no loss in
2
3
selectivity and with negligible coking.
Silica–alumina catalysts with multi-
ple SiO surface layers were previously
2
prepared by chemical vapour deposition
[
14]
(
(
CVD).
Atomic layer deposition
ALD) growth of SiO2 on Si(100)
wafers was achieved by George using
[15]
SiCl + H O, while Ferguson obtained
4
2
similar results using Si(OEt) (TEOS) +
4
[
16]
H O. Due to the low reactivity of SiO2,
2
an NH3 co-feed was required in both
cases to achieve efficient growth. How-
[17]
ever, Ma reported that more nucleo-
philic Au/TiO substrates nucleate SiO2
2
growth using Si(OMe) (TMOS) + H O
without NH3.
4
2
2
Figure 2. A) Si surface coverage in Si/nm as determined by ICP. B) Si (2p) XPS spectrum of
2
9
In the present strategy, SiO /Al O
SiO catalysts. C) DNP-enhanced Si CPMAS NMR spectra before and after calcining. Spectra
x
x
2
3
catalysts are prepared at 508C by were obtained at 100 K using MAS rate n =10 kHz, recycle delay tRD =1.5 s, contact time
R
tCP =3 ms, number of scans NS=2048, acquisition time AT =50 min. The surface structures
Si(3Al,OH), Si(Al,Si,2OH), Si(2Al,2OH), Si(2Al,Si,OH), and Si(3Al,Si) are depicted on the right.
Note that Si(3Al,OH) is the principle species in the calcined sample.
TEOS ALD on phase-pure g-alumina
2
À1
(
Alfa-Aesar, 60 m g ) in the viscous
[18]
flow reactor described previously.
The ALD A-B type growth sequence
Figure S9; Supporting Information) is:
) gaseous TEOS flow over the g-Al O for 120 s; 2) chamber
(
1
ment of SiO as a surface species, neither intercalating into the
2
3
x
purge for 240 s; 3) water vapor flow over the alumina for 300 s
from a 258C reservoir; 4) chamber purge for 300 s (see
Supporting Information for more details). After the ALD
growth, the catalysts are calcined for 18 h at 5508C in flowing
g-Al O3 structure nor altering the catalyst crystallite size
(Figure S12).
2
Next, an alternative growth mode in which all TEOS
pulses were performed in sequence, followed by H O pulses,
2
O . Before deposition, no surface species other than C, Al,
was investigated to determine if Si saturation occurs without
A-B-cycling. While Si growth is again observed in this
procedure, coverages reach only ca. 60% of those observed
for comparable numbers of A-B growth cycles. We speculate
that deposition is sterically hindered by the presence of
surface EtOH groups, consistent with previous observations
2
and O are detected on the alumina by XPS. The coverages of
2
the deposited Si on the alumina (1.5–3.2 Si/nm ) were assayed
by inductively coupled plasma (ICP) analysis (Figure 2A)
with confirmation by XPS. Note here that the Si deposition
saturates, independent of TEOS exposure time. Despite the
A-B-type reaction sequence, substrate rehydroxylation does
not result in further silica growth. Thus, the highest loading
[19]
in ALD processes.
The SiO /Al O sites were next characterized by XPS. The
x
2
3
2
obtained is 3.2 Si/nm . Due to the low loading of Si, little
adventitious C(1s) carbon peak is referenced here to a binding
energy (BE) of 285.0 eV. A higher energy feature assignable
change in Al surface character is observed spectroscopically
[20]
(
see below), as only a small fraction of surface sites are
to CÀO species is not detected, even in uncalcined catalysts.
The absence of residual surface ÀOEt moieties is further
corroborated by solid-state NMR spectroscopy (see below).
The Al(2p) BE occurs at 74.4 Æ 0.2 eV for all catalysts after
calcination, while the Al(2p) BE of reference g-Al O is found
occupied by Si atoms. Brunauer–Emmett-Teller (BET) sur-
face area analysis indicates constant surface area after ALD
across all SiO loadings (Figure S9, representative isotherms
x
in Figure S10), and scanning electron microscopy (SEM)
evidences no obvious morphological changes, with the
particle size remaining < 3 mm by SEM (Figures S9 and
S11). X-ray powder diffraction patterns of the catalysts
remain unchanged after Si saturation, confirming the assign-
2
3
at 74.4 eV. Representative Al XPS spectra are shown in the
Supporting Information (Figures S15–S17). The Si(2p) BE of
the calcined SiO /Al O sites is found at 102.4 Æ 0.2 eV for all
x
2
3
Si loadings (Figure 2B) and differs from that of bulk SiO ,
2
Angew. Chem. Int. Ed. 2015, 54, 13346 –13351
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13347

.
Angewandte
Communications
[
21]
found here at 104.1 eV, in agreement with the literature.
of Al-OH to Al-O-Si bonds; however, due to the low Si
loading only a small fraction of Al sites have Al-O-Si bonds,
precluding further interpretation of peak shape.
IV
Lower Si BEs relative to bulk silica are observed in other
aluminosilicates, and have been assigned to Si-O-M struc-
[
22]
tures where M ¼
6
Si. Increased acidity in the silanol proton
Ambient DRIFTS spectra of the catalysts were recorded
at 1008C after drying samples at 5008C for 1 hour, using
a KBr background (Figure S18). The spectral features are
is attributable to this electronic effect.
The surface geometry of the ALD-derived SiO /Al O
3
sites was next probed by DNP-enhanced Si{ H} cross-
polarization magic angle spinning (CPMAS) NMR spectros-
x
2
29
1
[27]
consistent with those previously noted in the literature.
Importantly, there is no shift in the position of the hydroxy
[23]
copy (Figure 2C, i–ii).
The DNP technique boosts the
signatures, hence no attenuation of the OÀH bond strength,
À1
sensitivity of conventional CPMAS by ca. 2 orders of
magnitude through excitation of the exogenously adminis-
tered biradicals at their ESR resonance frequency and
concomitant transfer of magnetization to the nuclear spins
although there is lessened intensity of the bands at 3790 cm
À1
and 3730 cm similar to that previously reported for Si
[27b]
deposition on g-Al O .
This contrasts with the analogous
2
3
but liquid-phase CLD process, which reveals attenuation
2
9
[11a]
(
due to the low Si loadings of these catalysts, conventional Si
under comparable conditions.
Notable is the appearance
À1
CPMAS NMR gave no detectable signals even after 40000
scans). In the uncalcined materials, the spectrum is assignable
to sites Si(2Al,Si,OH), Si(Al,Si,2OH) and Si(3Al,Si) species
of a sharp feature at 3723 cm in the present SiO /Al O
x 2 3
catalyst, assignable to surface SiÀOH species. Previously this
À1 [27b,28]
peak was assigned at 3725–3745 cm ,
with the lower
[24]
which are all expected at d = À85 to À90 ppm. Contribu-
frequency mode associated with greater OÀH bond ionicity,
À1
tions from other sites, such as Si(2Al,2OH) and Si(3Al,OH),
signifying enhanced Brønsted acidity. The 3723 cm position
[25]
cannot be excluded a priori. Furthermore, surface ethoxy
is thus consistent with the present assignment of the Si-O-H
sites as Brønsted acid sites.
13
1
groups cannot be detected by DNP-enhanced C{ H}
CPMAS NMR (Figure S13), arguing that hydrolysis is
essentially complete under the growth conditions.
NH DRIFTS experiments were next conducted to probe
3
surface acid sites, with catalyst surfaces exposed to NH at
3
The NMR results indicate that calcination enhances the Si
site homogeneity and increases the density of Si-O-Al
linkages, evidenced by narrowing and downfield displacement
1008C, then purged with Ar. The resulting spectra (Fig-
ure 3A,B) show that NH interaction with the neat g-alumina
3
hydroxy groups gives rise to Al -OH···NH features at 3769,
n
3
2
9
À1
of the Si signal. Moreover the d = À81 ppm position suggests
3727, and 3671 cm (Figure 3A, i, ii) in agreement with the
On ALD-derived SiO /Al O , an additional
x 2 3
band at 3741 cm (Figure 3A, iii, iv) is assigned to Si-
[29]
that the major Si surface species after calcination has an
literature.
[25b]
À1
Si(3Al,OH) geometry (Figure 2C, ii).
Minor responses
from the other species noted above are in principle possible;
however, the spectroscopic and catalytic data (see below)
argue that their contributions are minimal. Signals char-
acteristic of Si(4Si) (ca. d = À110 ppm) and Si(3Si,OH)
OH···NH species by analogy to similar modes in amorphous
3
[8a,30]
silica–aluminas.
The acid-site DRIFTS spectral region
[24]
(
ca. d = À100 ppm) sites
are also negligible, demon-
strating, in agreement with the XPS data, that the Si
deposition is limited by surface reactivity. Note that the
2
9
present Si NMR findings are consistent with other
studies arguing that the final stable geometry of surface
[26]
silica is a tridentate Si(3Al,OH) structure. We speculate
that here the geometry after calcination is a result of Si
atom spatial isolation. Although the deposition process
appears to yield some Si-O-Si linkages, polymeric SiO2
does not form, so that during the 5508C calcination, the Si
surface atoms reconstruct in a final geometry dictated
solely by the underlying g-Al O struture.
2
1
3
27
The DNP-enhanced Al{ H} CPMAS NMR measure-
ments were carried out to selectively detect Al sites on the
surface while excluding signals from underlying bulk
2
Al O . The spectra of g-Al O and 3.2 Si/nm SiO /Al O
2
3
2
3
x
2
3
before and after calcination (Figure S14) show two Al
VI
peaks assignable to octahedral sites (Al , d = 5 ppm) and
IV
[24]
tetrahedral sites (Al , d = 65 ppm).
Deposition of Si
VI
IV
converts a fraction of Al sites to Al sites, with no peak
shift or new signal apparent. The spectral intensities do not
undergo any further change upon calcination, suggesting
that the surface reconstruction by calcination involves
primarily the Si species, a result corroborated by infrared
Figure 3. A) DRIFTS spectra of NH treated SiO /Al O for: i) uncalcined
3
x
2
3
2
g-Al O , ii) calcined g-Al O , iii) uncalcined SiO /Al O with 3.2 Si/nm , and
2
3
2
3
x
2
3
2
iv) calcined SiO /Al O with 3.2 Si/nm . B) DRIFTS acid region spectra of
x
2
3
spectroscopy (see below). Note that a small upfield shift NH -treated SiO /Al O . C) Ratio of Brønsted-to-Lewis acid DRIFTS peak
3
x
2
3
IV
(
ca. 5 ppm) in the Al peak is expected upon conversion areas for NH
treated SiO /Al O samples.
3 x 2 3
1
3348 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 13346 –13351

Angewandte
Chemie
(
1
Figure 3B, i–iv) shows three NH bands on neat g-Al O at
3 2 3
620, 1471, and 1268 cm . Those at 1620 and 1268 cm are
À1
À1
assignable to NH bound to weak and strong Al Lewis acid
3
À1
sites, respectively, while the feature at 1471 cm is assigned to
a Brønsted acid site-derived NH4 mode. In contrast, the
+
[31]
DRIFTS spectrum of NH on uncalcined SiO –Al O exhibits
3
x
2
3
À1
only an SiÀOH mode at 3741 cm (Figure 3A, iii), indicating
that the majority of the g-Al O hydroxy groups are consumed
2
3
À1
in the ALD process. While the band at 3769 cm is usually
associated with the most reactive Al-OH groups,
[27a]
here we
find that all the alumina hydroxy groups react with gaseous
TEOS. The SiO /Al O DRIFTS acid region spectrum also
x
2
3
indicates selective consumption of the strong Lewis acid sites
À1
related to the band at 1268 cm (Figure 3B, iii), leaving the
other acid sites unperturbed. It thus appears that the present
ALD process involves coordination to, and presumably
TEOS cleavage at, strong Lewis acid sites. When these sites
are consumed, film growth ceases since the SiÀOH units are
insufficiently reactive to protonolyze the TEOS SiÀOR
linkage.
NH adsorption on either uncalcined or calcined g-Al O
3
3
2
reveals no obvous DRIFTS spectral differences (Fig-
Figure 4. Cyclohexanol dehydration kinetics over SiO /Al O . A) Turn-
x
2
3
ure 3A,B, i, ii). For NH adsorportion on calcined Si-saturated
2
3
over frequencies for 0.0–3.2 Si/nm SiO /Al O catalysts. Units: mol
x
2
3
SiO /Al O , similar AlÀOH features at 3769, 3727, and
2
x
2
À1
3
substrate/(m cats). B) Rate difference per mol Si on SiO /Al O
x 2 3
3
671 cm are visible (Figure 3A, , iv), however the SiÀOH
catalysts. The rate of pure alumina is subtracted from the rate of each
catalyst to approximate the specific activity of the Si sites, expressed
as the slope m of the linear regression. Units: mol substrate/(mol
Sis).
À1
group signature at 3741 cm remains strong, arguing that the
Si species are not incorporated into the bulk g-Al O . A
2
3
À1
strong Lewis acid peak at 1268 cm is also evident (Fig-
ure 3B, iv) and resembles that of pure g-Al O . The Brønsted
2
3
acidity of the SiO /Al O series was next assayed from the
DRIFTS peak area ratio of Brønsted NH4 to Lewis
formation but without formation of carbonaceous deposits
x
2
3
+
[34]
typical of strong solid acids. For ALD-derived SiO /Al O
3
x
2
site···NH species (Figure 3C). The linear correlation of this
catalysts, no products except cyclohexene are observed by
GC-MS, and the carbon balance determined by an independ-
ent calibration of both cyclohexanol and cyclohexene is
invariably 95–102%. Since the cyclohexanol dehydration rate
over SiO /Al O exhibits a first-order dependence on [cyclo-
hexanol], the rate constant k is determined from the slope of
the cyclohexanol conversion versus time plot under differ-
ential conditions, i.e., when [cyclohexanol] is not significantly
altered (Figure S22).
3
ratio with the Si surface density argues for enhanced Brønsted
acidity of the new active sites. Additionally, the DRIFTS of
adsorbed pyridine (Figure S20) confirms that there is no
perturbation of the catalyst Lewis acid sites after Si deposition
and calcination. DRIFTS of NH3 spectra for individual
catalyst loadings are shown in Figure S21.
x
2
3
CO DRIFTS was used to further probe surface hydroxy
groups, since the OÀH stretchinig frequency observed upon
CO adsorption is sensitive to the Brønsted acidity, due to Al-
Note that the present SiO /Al O catalysts provide higher
x
2
3
[32]
OH···CO hydrogen-bonding.
At 1008C, the DRIFTS
initial turnover frequencies (normalized to catalyst surface
spectrum of CO on pure g-Al O (Figure S19) reveals bands
at 3794 and 3774 cm , assignable to terminal tetrahedral and
area, Figure 4A) versus the parent g-Al O . The SiO /Al O
3
TOF reaches a maximum of 6.6 ” 10 mol substrate/(m cats)
2
3
2
3
x
2
À1
À8
2
octahedral AlÀOH groups, in excellent agreement with the
at the highest Si loading, a 3.5-fold increase over that of g-
Al O , 1.9 ” 10 mol substrate/(m cats). SiO was found to be
2 3 2
catalytically inert under the same conditions, differentiating
[
33]
À8
2
literature. ALD deposition of Si on the catalyst surface
does not affect these peak positions, although a reduction in
À1
the 3794 cm band intensity is noticable. A new OH band
the Brønsted acidity of the isolated SiO site from polymeric
x
À1
2
appears at 3726 cm in the 3.2 Si/nm catalyst spectrum,
which is assigned here to isolated Si-OH groups on the basis
SiO . A commercially available 6.5 wt% SiO –Al O catalyst
2
2
2
3
was screened for comparison. While the TOF of the commer-
[
3b]
À7
2
of previous reports. Thus, the CO DRIFTS confirms the
presence of isolated SiÀOH sites, with no discernable
attenuation of existing AlÀOH groups.
cial catalyst (1.4 ” 10 mol substrate/(m cats)) was found to
À8
be 2 times that of our highest TOF (6.6 ” 10 mol substrate/
(m cats)), the carbon balance of the reaction was only 68%,
2
The catalytic properties of the calcined SiO /Al O
3
indicating significant carbon loss consistent with heavy
product formation (Figure S22). We hypothesize that this
difference in carbon balance is a result of the milder acidity of
x
2
materials were next evaluated for cyclohexanol dehydration
in a batch reactor (see Figures S6 and S7 for details); catalytic
data are compiled in Table S22, and summarized in Figure 4.
It is seen that the present SiO /Al O catalysts possess
the SiO species relative to strong Brønsted sites in conven-
x
tional SiO –Al O and demonstrates the value of tuning the
x
2
3
2
2
3
sufficiently strong Brønsted acid sites to catalyze this trans-
strength and homogeneity of solid acid catalysts.
Angew. Chem. Int. Ed. 2015, 54, 13346 –13351
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13349

.
Angewandte
Communications
Since the DNP-enhanced CPMAS NMR and infrared
spectroscopies indicate (see above) that the SiO /Al O
3
gratefully acknowledged. GC-TOF instrumentation at the
Integrated Molecular Structure and Research Center
(IMSERC) was supported by the NSF (CHE-0923236). This
work made use of the Epic Facility (NUANCE Center-
Northwestern University), which has received support from
the MRSEC program (NSF DMR-1121262) at the Materials
Research Center, The Nanoscale Science and Engineering
Center (EEC-0118025/003), both programs of the National
Science Foundation; the State of Illinois; and Northwestern
University. The Clean Catalysis Facility of the Northwestern
University Center for Catalysis and Surface Science is
supported by a grant from the DOE (DE-SC0001329).
Solid-state NMR studies at Ames Laboratory were supported
by the U.S. DOE, Office of Basic Energy Sciences, Division of
Chemical Sciences, Geosciences, and Biosciences through the
Ames Laboratory under Contract No. DE-AC02-07CH11358.
We thank A. Weingarten for helpful contributions.
x
2
alumina character is largely unchanged after calcination, it
is reasonable to assume the parent alumina catalytic activity
remains approximately constant at all Si loadings. This allows
for subtraction of the alumina rate contribution from the
apparent SiO /Al O rates (Figure 4B). The resulting linear
x
2
3
increase in rate per mol Si strongly argues that all surface Si
sites are uniformly active in the cyclohexanol dehydration,
and the slope of the linear regression yields a turnover
À2
frequency per Si site of 1.5 ” 10 mol substrate/(molSis).
As a further comparison, a surface-acid SiO /Al O
3
2
2
catalyst was prepared according to the wet chemical CLD
[11b]
technique of van Bokhoven.
This catalyst was selected
because it represents an umambiguous, direct comparison to
our own, being prepared from the same precursors, TEOS
and g-Al O , but under different synthetic conditions. In
2
3
addition, the physical and acidic properties of such catalysts
[7b, 11]
have been extensively characterized.
The catalyst was
Keywords: alcohol dehydration · atomic layer deposition ·
prepared with a single Si deposition reaction using an g-Al O3
heterogeneous catalysis · silica-alumina · solid acid
2
identical to that in the ALD syntheses (further description in
Figure S9). The liquid-phase technique yields a catalyst of
How to cite: Angew. Chem. Int. Ed. 2015, 54, 13346–13351
Angew. Chem. 2015, 127, 13544–13549
2
À1
substantially similar surface area (56 m g ), particle size
2
(
< 3 mm), and Si content (2.8 Si/nm ). However, the catalytic
results (Figures S22 and S23) deviate from those observed
with the ALD-derived catalysts. Analysis of catalytic activity
and selectivity is hampered by apparent leaching of Si from
the catalyst under reaction conditions, as Si-containing
species are observed by GC-MS and Si NMR (principally
cyclohexyl ethers of Si). Selectivity to cyclohexene is low
[
1] K. Tanabe, in Solid Acids and Bases (Ed.: K. Tanabe), Academic
Press, San Diego, 1970, pp. 1 – 3.
29
[2] a) C. L. Thomas, Ind. Eng. Chem. 1949, 41, 2564 – 2573; b) B. E.
Yoldas, J. Non-Cryst. Solids 1984, 63, 145 – 154; c) N. Yao, G.
Xiong, M. He, S. Sheng, W. Yang, X. Bao, Chem. Mater. 2002, 14,
(< 20%), and other products, including dicyclohexyl ether
1
22 – 129.
and cyclohexylcyclohexane, are observed. Despite the leach-
ing, catalytic activity comparable to SiO /Al O catalysts on
[3] a) C. J. Plank, E. J. Rosinski, W. P. Hawthorne, I&EC Prod. Res.
Dev. 1964, 3, 165 – 169; b) E. G. Derouane, J. C. VØdrine, R. R.
Pinto, P. M. Borges, L. Costa, M. A. N. D. A. Lemos, F. Lemos,
F. R. Ribeiro, Catal. Rev. 2013, 55, 454 – 515.
x
2
3
À8
2
a surface area basis (4.7 ” 10 mol substrate/(m cats)) was
observed, but based on the catalyst Si content, the Si sites do
not show the same activity as those in SiO /Al O (< 1.5 ”
[
4] a) F. F. Roca, L. De Mourgues, Y. Trambouze, J. Catal. 1969, 14,
07 – 113; b) M. Guisnet, N. S. Gnep, S. Morin, Microporous
x
2
3
1
2
1
0 mol substrate/(molSis)). The origin of this difference is at
Mesoporous Mater. 2000, 35 – 36, 47 – 59.
5] a) J. H. Lunsford, Catal. Today 2000, 63, 165 – 174; b) D. Chen,
K. Moljord, A. Holmen, Microporous Mesoporous Mater. 2012,
164, 239 – 250.
present unclear and promises a rich question for future study.
In addition some structural differences in previously prepared
[
[11a]
CLD catalysts are noted, as characterized by FTIR
DNP-enhanced CPMAS SSNMR.
and
[11c]
[6] H. Kawakami, S. Yoshida, T. Yonezawa, J. Chem. Soc. Faraday
Trans. 2 1984, 80, 205 – 217.
In summary, we report the synthesis of a novel class of
highly dispersed SiO /Al O catalysts by the vapor-phase
[
7] a) D. G. Poduval, J. A. R. van Veen, M. S. Rigutto, E. J. M.
Hensen, Chem. Commun. 2010, 46, 3466 – 3468; b) M. Caillot,
A. Chaumonnot, M. Digne, J. A. van Bokhoven, J. Catal. 2014,
316, 47 – 56.
x
2
3
reaction of TEOS with g-Al O . Significant surface Brønsted
2
3
acidity is observed that tracks the Si coverage in a linear
fashion, implying along with XPS, NMR, and DRIFTS
[8] a) M. Trombetta, G. Busca, S. Rossini, V. Piccoli, U. Cornaro, A.
Guercio, R. Catani, R. J. Willey, J. Catal. 1998, 179, 581 – 596;
b) W. Suprun, M. Lutecki, T. Haber, H. Papp, J. Mol. Catal. A
spectroscopic data, an (-AlO) Si(OH) active site assignment.
3
The silanol proton is catalytically competent, and the SiOx/
2
009, 309, 71 – 78.
9] a) G. P. Heitmann, G. Dahlhoff, W. F. Hçlderich, J. Catal. 1999,
86, 12 – 19; b) R. Martinez, M. C. Huff, M. A. Barteau, Appl.
Catal. A 2000, 200, 79 – 88.
Al O sites exhibit higher catalytic activity than g-Al O sites
2
3
2
3
[
in cyclohexanol dehydration. In contrast to stronger Brønsted
acid sites in a conventional silica–alumina catalyst, no carbon
loss is detected, demonstrating the attraction of more uni-
1
[10] a) O. M. Gazit, A. Katz, Langmuir 2012, 28, 431 – 437; b) O. M.
Gazit, A. Katz, J. Am. Chem. Soc. 2013, 135, 4398 – 4402.
form/less acidic SiO sites.
x
[
11] a) M. Caillot, A. Chaumonnot, M. Digne, J. A. V. Bokhoven,
ChemCatChem 2013, 5, 3644 – 3656; b) M. Caillot, A. Chau-
monnot, M. Digne, C. Poleunis, D. P. Debecker, J. A. van Bok-
hoven, Microporous Mesoporous Mater. 2014, 185, 179 – 189;
c) M. Valla, A. J. Rossini, M. Caillot, C. Chizallet, P. Raybaud,
M. Digne, A. Chaumonnot, A. Lesage, L. Emsley, J. A.
van Bokhoven, C. CopØret, J. Am. Chem. Soc. 2015.
Acknowledgements
Financial support from the NSF (CHE-1213235), IACT (DE-
AC02-06CH11357), and ICEP (DE-FG02-03ER15457) is
1
3350 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 13346 –13351

Angewandte
Chemie
[
12] J. Datka, B. Gil, O. Vog, J. Rakoczy in Stud. Surf. Sci. Catal.,
Vol. 125 (Eds.: I. Kiricsi, G. Pµl-BorbØly, J. B. Nagy, H. G.
Karge), Elsevier, Amsterdam 1999, pp. 409 – 415.
13] J. He, C. Zhao, J. A. Lercher, J. Catal. 2014, 309, 362 – 375.
14] a) M. Niwa, S. Kato, T. Hattori, Y. Murakami, J. Chem. Soc.
Faraday Trans. 1 1984, 80, 3135 – 3145; b) S. Sato, M. Toita, Y.-Q.
Yu, T. Sodesawa, F. Nozaki, Chem. Lett. 1987, 16, 1535 – 1536.
15] J. W. Klaus, S. M. George, Surf. Sci. 2000, 447, 81 – 90.
16] J. D. Ferguson, E. R. Smith, A. W. Weimer, S. M. George, J.
Electrochem. Soc. 2004, 151, G528 – G535.
[24] G. Engelhardt, D. Michel, High-Resolution Solid-State NMR of
Silicates and Zeolites, Wiley, Chichester, 1987.
[25] a) M. McMillan, J. S. Brinen, J. D. Carruthers, G. L. Haller,
Colloids and Surfaces 1989, 38, 133 – 148; b) N. Katada, M. Niwa,
Res. Chem. Intermed. 1998, 24, 481 – 494.
[26] M. Casarin, D. Falcomer, A. Glisenti, M. M. Natile, F. Poli, A.
Vittadini, Chem. Phys. Lett. 2005, 405, 459 – 464.
[27] a) G. Busca, Catal. Today 2014, 226, 2 – 13; b) E. Finocchio, G.
Busca, S. Rossini, U. Cornaro, V. Piccoli, R. Miglio, Catal. Today
1997, 33, 335 – 352.
[
[
[
[
[
[
17] Z. Ma, S. Brown, J. Y. Howe, S. H. Overbury, S. Dai, J. Phys.
Chem. C 2008, 112, 9448 – 9457.
18] J. W. Elam, M. D. Groner, S. M. George, Rev. Sci. Instrum. 2002,
[28] T. C. Sheng, S. Lang, B. A. Morrow, I. D. Gay, J. Catal. 1994, 148,
341 – 347.
[29] a) J. B. Peri, R. B. Hannan, J. Phys. Chem. 1960, 64, 1526 – 1530;
b) M. Niwa, N. Katada, Y. Murakami, J. Phys. Chem. 1990, 94,
6441 – 6445.
[30] M. R. Basila, T. R. Kantner, J. Phys. Chem. 1967, 71, 467 – 472.
[31] D. P. Sobczyk, J. J. G. Hesen, J. van Grondelle, D. Schuring,
A. M. de Jong, R. A. van Santen, Catal. Lett. 2004, 94, 37 – 43.
[32] K. I. Hadjiivanov, G. N. Vayssilov in Advances in Catalysis,
Vol. 47, Academic Press, San Diego, 2002, pp. 307 – 511.
[33] a) L. Oliviero, A. Vimont, J.-C. Lavalley, F. Romero Sarria, M.
Gaillard, F. Mauge, Phys. Chem. Chem. Phys. 2005, 7, 1861 –
1869; b) K. Hadjiivanov in Advances in Catalysis, Vol. 57 (Ed.:
C. J. Friederike), Academic Press, San Diego 2014, pp. 99 – 318.
[34] F. Garcia-Ochoa, A. Santos, Ind. Eng. Chem. Res. 1993, 32,
7
3, 2981 – 2987.
[
[
19] R. L. Puurunen, J. Appl. Phys. 2005, 97, 121301.
20] A. F. Lee, D. E. Gawthrope, N. J. Hart, K. Wilson, Surf. Sci. 2004,
5
48, 200 – 208.
[
[
[
21] a) E. Paparazzo, Surf. Interface Anal. 1988, 12, 115 – 118; b) L.
Galµn, I. Montero, F. Rueda, J. M. Albella, Surf. Interface Anal.
1
992, 19, 473 – 477.
22] a) T. L. Barr, Zeolites 1990, 10, 760 – 765; b) I. BçszçrmØnyi, T.
Nakayama, B. McIntyre, G. Somorjai, Catal. Lett. 1991, 10, 343 –
3
55.
23] a) T. Maly, G. T. Debelouchina, V. S. Bajaj, K.-N. Hu, C.-G. Joo,
M. L. Mak-Jurkauskas, J. R. Sirigiri, P. C. A. van der Wel, J.
Herzfeld, R. J. Temkin, R. G. Griffin, J. Chem. Phys. 2008, 128,
2
626 – 2632.
0
52211 – 052219; b) A. Lesage, M. Lelli, D. Gajan, M. A.
Caporini, V. Vitzthum, P. MiØville, J. Alauzun, A. Roussey, C.
Thieuleux, A. Mehdi, G. Bodenhausen, C. Coperet, L. Emsley, J.
Am. Chem. Soc. 2010, 132, 15459 – 15461.
Received: June 13, 2015
Revised: August 31, 2015
Published online: September 23, 2015
Angew. Chem. Int. Ed. 2015, 54, 13346 –13351
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13351