Following solid-acid-catalyzed reactions by MAS NMR spectroscopy in liquid phase - Zeolite-catalyzed conversion of cyclohexanol in water

Article abstract of DOI:10.1002/anie.201306673

A microautoclave magic angle spinning NMR rotor is developed enabling insitu monitoring of solid-liquid-gas reactions at high temperatures and pressures. It is used in a kinetic and mechanistic study of the reactions of cyclohexanol on zeolite HBEA in 130 °C water. The ¹³Cspectra show that dehydration of 1-¹³C-cyclohexanol occurs with significant migration of the hydroxy group in cyclohexanol and the double bond in cyclohexene with respect to the ¹³C label. A simplified kinetic model shows the E1-type elimination fully accounts for the initial rates of 1- ¹³C-cyclohexanol disappearance and the appearance of the differently labeled products, thus suggesting that the cyclohexyl cation undergoes a 1,2-hydride shift competitive with rehydration and deprotonation. Concurrent with the dehydration, trace amounts of dicyclohexyl ether are observed, and in approaching equilibrium, a secondary product, cyclohexyl-1-cyclohexene is formed. Compared to phosphoric acid, HBEA is shown to be a more active catalyst exhibiting a dehydration rate that is 100-fold faster per proton. Hot analysis: The catalytic conversion of cyclohexanol on zeolite HBEA in hot liquid water leads to dehydration as well as alkylation products. A novel microautoclave suitable for application in magic angle spinning (MAS) NMR at high temperatures has been successfully applied to obtain new insight into the mechanistic pathway leading to an understanding of the reactions under selected experimental conditions.

Full text of DOI:10.1002/anie.201306673

Angewandte
Chemie
DOI: 10.1002/anie.201306673
Heterogeneous Catalysis
Following Solid-Acid-Catalyzed Reactions by MAS NMR
Spectroscopy in Liquid Phase—Zeolite-Catalyzed Conversion of
Cyclohexanol in Water**
Aleksei Vjunov, Mary Y. Hu, Ju Feng, Donald M. Camaioni, Donghai Mei, Jian Z. Hu,
Chen Zhao, and Johannes A. Lercher*
Abstract: A microautoclave magic angle spinning NMR rotor
is developed enabling in situ monitoring of solid–liquid–gas
reactions at high temperatures and pressures. It is used in
a kinetic and mechanistic study of the reactions of cyclo-
hexanol on zeolite HBEA in 1308C water. The 13C spectra
show that dehydration of 1-13C-cyclohexanol occurs with
significant migration of the hydroxy group in cyclohexanol
and the double bond in cyclohexene with respect to the 13C
label. A simplified kinetic model shows the E1-type elimination
fully accounts for the initial rates of 1-13C-cyclohexanol
disappearance and the appearance of the differently labeled
products, thus suggesting that the cyclohexyl cation undergoes
a 1,2-hydride shift competitive with rehydration and deproto-
nation. Concurrent with the dehydration, trace amounts of
dicyclohexyl ether are observed, and in approaching equilib-
ized to the concentration of protons is about two orders of
magnitude higher in a zeolite than in aqueous solution of
a mineral acid.[4] It is also remarkable that hydroalkylation of
phenol on zeolite HBEA can occur in water, while other
macroporous solid acids do not catalyze this reaction.[1,5] It
can be speculated that zeolites achieve this by providing an
environment (sometimes called confinement and nest
effects),[6] which uniquely stabilizes the sorbed and transition
states.[7] To better understand the catalysis we have under-
taken substantial efforts to monitor the state of the reacting
molecules by spectroscopic methods. The information
obtained potentially allows one not only to understand the
mechanistic pathways, but provides also the basis to design
new and improve existing zeolites with respect to activity and
selectivity.
rium,
a
secondary product, cyclohexyl-1-cyclohexene is
The dehydration of secondary alcohols in aqueous phase
catalyzed by mineral acids generally proceeds by the E1
mechanism.[8] Also, alcohol dehydration in the gas phase
using polyoxometalate[9] and zeolite[10] catalysts has been
suggested to follow strictly the E1 elimination pathway. The
catalyst acidity determined by the zeolite lattice composi-
tion[7] and the pore size[11] jointly control activity and
selectivity for catalytic gas phase reactions. Herein we
elucidate the mechanism of the reaction of aqueous cyclo-
hexanol on zeolite HBEA by magic angle spinning (MAS)
nuclear magnetic resonance (NMR) spectroscopy, which
enables simultaneous monitoring of the state of reactants,
intermediates, and products, thereby providing mechanistic
insight through following the variations in isotopomer
concentrations during catalytic transformations.
formed. Compared to phosphoric acid, HBEA is shown to
be a more active catalyst exhibiting a dehydration rate that is
100-fold faster per proton.
T
he efficient catalytic hydrodeoxygenation (HDO) of phe-
nols over metals, such as Pt,[1] Pd,[2] or Ni,[3] requires an acid
function to catalyze dehydration of the intermediately formed
cycloalkanols. Surprisingly, the rate of dehydration normal-
[*] A. Vjunov, M. Y. Hu, Dr. J. Feng, Dr. D. M. Camaioni, Dr. D. Mei,
Dr. J. Z. Hu, Prof. Dr. J. A. Lercher
Institute for Integrated Catalysis
Pacific Northwest National Laboratory
P.O. Box 999, Richland, WA 99352 (USA)
E-mail: johannes.lercher@pnnl.gov
While in situ MAS-NMR spectroscopy has been devel-
oped to study gas–solid reactions in controlled environ-
ments,[12] following reactions in liquid phase during the
formation of solids was the limit for studying reactions
during the synthesis of zeolites.[13] It has not been possible, to
date, to monitor two- and three-phase reactions involving
solids, gases, and liquids at elevated temperatures and
pressures.[14] To address this we have constructed a micro-
autoclave that functions as MAS-NMR rotor and allows batch
reactions with pressures of up to 20 bar independent of the
gas atmosphere and temperatures of up to 2008C, thus
allowing mechanistic studies of most reactions in organic
synthesis. The detailed design of the microautoclave that can
be repeatedly sealed is illustrated in Figure 1. The rotor can
withstand spinning the sample at elevated temperature and
pressure with the design yielding negligible 13C background
signal (see the Supporting Information for detail). The reactor
Dr. C. Zhao, Prof. Dr. J. A. Lercher
Department of Chemistry and Catalysis Research Institute
TU Mꢀnchen
Lichtenbergstrasse 4, 85748 Garching (Germany)
[**] We thank Zizwe A. Chase and Junming Sun (both from Washington
State University) for pretreating the catalyst. C.F. Peden is
acknowledged for his support in developing the NMR capability.
This work was supported by the U. S. Department of Energy (DOE),
Office of Basic Energy Sciences, Division of Chemical Sciences,
Geosciences & Biosciences. All experiments were performed at the
Environmental Molecular Sciences Laboratory, a national scientific
user facility sponsored by the DOE’s Office of Biological and
Environmental Research located at Pacific Northwest National
Laboratory (PNNL). PNNL is a multiprogram national laboratory
operated for DOE by Battelle Memorial Institute under Contract no.
DE-AC06-76L0-1830.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201306673.
Angew. Chem. Int. Ed. 2014, 53, 479 –482
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
479

.
Angewandte
Communications
is also suitable for operation with exo-
A nearly identical value is observed for cyclohexene, thus
suggesting that the dispersion forces account for the largest
fraction of the interaction energy.[16] Note that water fills the
zeolite pores, while cyclohexanol occupies only up to 25% of
the pore volume during the reaction. Nevertheless this
corresponds to a significant (20-fold) enhancement of the
concentration of cyclohexanol compared to the aqueous
phase. This exemplifies the potential to determine not only
changes in concentrations, but also how it is possible to
quantify the distribution between the adsorbed and the
mobile phase at elevated pressures and temperatures.
geneous high-pressure gas using a load-
ing chamber that allows sealing and
opening of the rotor in a controlled
atmosphere.[15] The lid of the microau-
toclave and the valve adaptor to pres-
surize it, are removable. Thus, operation
at pressures exceeding 20 bar is possible
with marginal gas leaking for up to 72 h.
Let us now turn to a typical experi-
ment using
a
solid catalyst, i.e.,
HBEA150 zeolite (about 25 mg for
a typical example) and 125 mL of 0.33m
1-13C-cyclohexanol in water. The cata-
lyst and the reactants were sealed in the
rotor, while NMR data were acquired at
an estimated reaction temperature of
1308C, as function of time. After 23 h
this led to 36% cyclohexene, 3% dicy-
clohexyl ether, and 1% cyclohexyl-1-
cyclohexene. The forward rate constant
for cyclohexene formation was kf = 3.4 ꢀ
At the start of the reaction, formation of cyclohexene and
dicyclohexyl ether is observed together with scrambling of the
13C label in the alicyclic ring. As the reaction progresses
a secondary product, cyclohexyl-1-cyclohexene, is detected.
At long reaction time quantification of products appearing in
low concentration is hindered by the diversity of labeled
products and long spin lattice relation times (T1) for the
carbon atoms without attached hydrogen atoms. However,
ꢀ 96% of the 13C could be accounted for as cyclohexanol and
cyclohexene during the initial reaction stage (2 h); the slight
signal loss is attributed to low initial concentrations of
dicyclohexyl ether and cyclohexyl-1-cyclohexene that would
require a higher signal-to-noise ratio to measure accurately.
To determine which mechanism, E1 or E2, operates in the
dehydration of cyclohexanol, the 13C-label scrambling was
analyzed and the initial rates of disappearance of 1-13C-
cyclohexanol and the appearance of the differently labeled
products were determined. A simplified kinetic model (see
the Supporting Information) shows the E1-type elimination
via carbenium ions fully accounting for all initial rates,
suggesting that the cyclohexyl cation undergoes a 1,2-hydride
shift competitive with rehydration and deprotonation.[17] The
model shows that the rate of the 1,2-hydride shift approx-
imately equals that of deprotonation and is approximately
half the rate of the rehydration of cyclohexene (Figure S7 in
the Supporting Information). As dehydration approaches
equilibrium, the probability for an electrophilic attack of
a cyclohexyl carbenium ion on cyclohexene to form cyclo-
hexyl-1-cyclohexene increases. The cyclohexyl cation is con-
cluded to be the central intermediate, which explains the
reaction products, as well as the rapid migration of the
hydroxy group and double bond away from the 13C-labeled
carbon. An E2-type elimination kinetic model cannot account
for these rates. The mechanism determined by virtue of in situ
NMR analysis and proposed herein is shown in Scheme 1.
Figure 1. The
9.5 mm outer diame-
ter high-temperature
and high-pressure
MAS rotor is shown.
a) The fully assem-
bled rotor; and
b) the various com-
ponents, i.e., the zir-
conia rotor sleeve
(1), the ceramic
insert made of mate- 10À5 sÀ1 (see the Supporting Information
rials such as Macor
(2), the sample cell
space (3) and thread
for detail); the turnover frequency
(TOF) was determined as the initial
forward rate normalized by the concen-
(4), O ring (5), and
tration of Brønsted acid sites, TOF =
Torlon screw (6). The
outside surface of
4.2 ꢀ 10À4 molcyclohexene molBASÀ1 sÀ1.
Figure 2 shows the series of NMR
spectra following the conversion of
cyclohexanol and an example for
a single spectrum with the assignment
of individual 13C signals. Initially 1-13C-
cyclohexanol is observed as a narrow
peak at about 70 ppm, corresponding to
1-13C-cyclohexanol in the aqueous
the insert is fixed to
the inner surface of
the zirconia rotor
sleeve by high-tem-
perature glue.
phase, and a broad peak at about 70.8 ppm, which is
attributed to 1-13C-cyclohexanol interacting with the zeolite.
Using these two peaks we estimate that about 50% of the
cyclohexanol present is initially adsorbed in the zeolite pores.
Figure 2. Stacked plot containing 80 MAS-NMR spectra acquired for
a mixture of 22 mg HBEA150 and 120 mL of 0.33m 1-13C-cyclohexanol
at 1308C in liquid water and as a function of time.
Scheme 1. Proposed reaction pathway of cyclohexanol reacting in
liquid water on HBEA.
480
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 479 –482

Angewandte
Chemie
In the absence of water, alcohol dehydration on zeolites
has been suggested to proceed through the E1-mechanistic
pathway,[18] with the formation of surface-bound alkoxide
species and water constituting the rate-determining step.[19] In
contrast, a corresponding species was not observed in MAS-
NMR (zeolite+-OÀR) in presence of larger concentrations of
liquid water. In passing it should also be mentioned that such
alkoxides have been suggested to be unstable even in the
presence of gaseous water.[20]
shows that decomposition of the protonated cyclohexanol
À
into cyclohexyl cation and water through C O bond scission
is an endothermic step and rate limiting with an activation
À1
À
barrier of 97 kJmol . After C O bond scission a b-methyl-
ene proton of the cyclohexyl cation is transferred back to
HBEA, thereby reforming the zeolite hydroxy group and
leading to cyclohexene, which is favored over the carbenium
ion by 15 kJmolÀ1 in the zeolite.
Thus, we conclude that sorbed cyclohexanol is bound
primarily through dispersion forces and specifically adsorbed
at Brønsted acid sites in protonated form. Under the present
experimental conditions only approximately 5% of cyclo-
hexanol in the pores is adsorbed in the latter form. Water
elimination follows an E1 mechanism forming a cyclohexyl
carbenium ion, which either undergoes rapid 1,2-hydride shift
or is rehydrated, thus, scrambling the label in cyclohexanol.
Transfer of the proton back to HBEA closes the catalytic
cycle and leads to the primary product, cyclohexene. Alter-
natively, the cyclohexyl oxonium ion reacts with another
cyclohexanol forming dicyclohexyl ether, in analogy to
dimethyl ether formation.[22] In approaching the equilibrium,
the cyclohexyl carbenium ion also undergoes nucleophilic
addition of cyclohexene leading to C–C coupling and
formation of cyclohexyl-1-cyclohexene.
À
À
The E2 mechanism, requiring the C O and b C H bonds
to be cleaved concertedly, is not relevant in the present case.
If this pathway were dominant, then the initial rate of
disappearance of the starting 1-13C-cyclohexanol should equal
the initial rate of appearance of cyclohexene. Since this step is
reversible, the 13C scrambling would be explained by hydra-
tion of cyclohexene leading to equal probability for the label
to be in either the 1- or 2-labeled position of cyclohexanol.
However, the initial rate for disappearance of 1-13C-cyclo-
hexanol is 2.5 times the initial rate of formation of 1-13C-
cyclohexene. Furthermore, the initial rates for formation of 2-
13C-cyclohexanol and 3-13C-cyclohexene are 0.2 and 0.1 times
the initial rate of disappearance of 1-13C-cyclohexanol,
respectively. Thus, on basis of these individual rates an E2-
type elimination mechanism can be excluded.
Given that hydronium ions or Brønsted acid sites are the
catalytically active species, it is important to compare the
specific catalytic activity of Brønsted acid sites (BAS) in
zeolites with the activity of hydronium ions in acidic aqueous
solutions. The comparison of the reactivity of zeolite HBEA
with that of phosphoric acid at 1608C shows under identical
The results show that the novel microautoclave NMR
rotor developed allows kinetic and mechanistic studies of
solid–liquid–gas and liquid–liquid–gas reactions at high
temperatures and pressures to an unprecedented extent.
The breadth and depths of mechanistic studies will only be
limited by the sensitivity of the NMR spectrometer to
differentiate the isotopomers of the reacting species.
reaction
conditions
a
TOF
value
of
1.0 ꢀ
10À2 molcyclohexene molBASÀ1 sÀ1 for the zeolite and 1.4 ꢀ
10À4 molcyclohexene molBASÀ1 sÀ1 for H3PO4, i.e., the zeolite has
over a 100 times higher turnover frequency (see the Support-
ing Information for details). Because the dehydration displays
a first-order dependence in cyclohexanol for the zeolite as
well as the mineral acid, it is concluded that the concentration
of acid sites that interact with cyclohexanol must be low. This
raises the question, why the rate is substantially higher in the
case of the zeolite. While the presence of sufficient water
leads to hydronium ions in both cases, the environment of the
zeolite pore is shown to stabilize the transition state such that
the activation entropy for the dehydration is significantly
higher in the case of the zeolite.[21] A further interpretation is
beyond the scope of the present contribution. Note, that
preliminary results of theoretical calculations indeed point to
a stabilization of the hydroxonium ion transition state.
Because C–C coupling products are observed in the
presence of zeolite, but not when catalyzed by aqueous
H3PO4, we conclude that also in this case the zeolite pores
exert a specific influence stabilizing the transition state of the
bimolecular reaction.[19]
Experimental Section
NMR experiments were performed on a 500 MHz wide bore NMR
spectrometer equipped with a homemade 9.5 mm MAS probe.[12c]
A
single pulse sequence with a 45-degree pulse angle and high-power
proton decoupling was used. As example experiment: in situ 13C
MAS-NMR spectra were acquired as function of time while spinning
the sample, a mixture of HBEA150 (22 mg) and 1-13C-cyclohexanol
(120 mL of 0.33m) in water at estimated 1308C, at 2.4 kHz. A recycle
delay of 5 s and an accumulation number of 256 scans was applied
resulting in acquisition times for a single spectrum of approximately
0.29 h. The temperature in the microautoclave was calibrated prior to
the experiment using Pb(NO3)2.[23] To minimize conversion of 1-13C-
cyclohexanol prior to a kinetic run, the temperature of the sample was
raised to 1308C in two steps, first to 808C, at which time spectrometer
settings were checked, and then to the reaction temperature. The
heat-up time for this second step was approximately 5 min. In this way
less than 2% of 1-13C-cyclohexanol was converted prior to acquiring
the first spectrum in Figure 2 (see the Supporting Information for
details). To assure quantitative analysis of obtained 13C NMR spectra
T1 values for cyclohexanol and cyclohexene were measured at RTand
reaction temperature. Results are available in the Supporting
Information. HBEA150 (Si/Al = 75) was obtained from Sꢁd
Chemie AG (Clariant) in hydrogen form (see the Supporting
Information). As an example of the batch autoclave experiment:
HBEA150 (17.4 g) and cyclohexanol (80 mL of 0.33m) in water were
placed in the 300 mL Hastelloy Parr reactor, heated to 1608C, and
stirred vigorously as temperature setting is reached. After a set time,
the reactor was cooled with ice, leading to a rapid cool down and
reaction quench. Contents of the reactor were extracted with
These reaction pathways are in line with computational
modeling in the absence of water indicating that cyclohexanol
adsorbs on Brønsted acid sites with a binding energy of
116 kJmolÀ1, which is significantly higher than the binding
energy of water (45 kJmolÀ1). Cyclohexanol is spontaneously
protonated upon adsorption, with proton transfer being
thermodynamically favored by 29 kJmolÀ1. Calculation
Angew. Chem. Int. Ed. 2014, 53, 479 –482
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
481

.
Angewandte
Communications
dichloromethane and dried with sodium sulfate. The extract was
analyzed on an Agilent 7890 A GC equipped with a HP-5MS 25 m
0.25 mm inner diameter column, coupled with an Agilent 5975C MS.
Periodic density functional calculations (DFT) were performed
using the CP2K code.[24] Calculations employ a mixed Gaussian and
plane wave basis sets. Core electrons were represented with norm-
conserving Goedecker–Teter–Hutter pseudopotentials[25] and the
valence electron wave function was expanded in a double-zeta basis
set with polarization functions[26] along with an auxiliary plane wave
basis set with an energy cutoff of 360 eV. The generalized gradient
approximation exchange-correlation functional of Perdew, Burke,
and Enzerhof (PBE)[27] was used for all calculations. Each config-
uration was optimized with Broyden–Fletcher–Goldfarb–Shanno
(BGFS) algorithm with SCF convergence criteria of 10À8 au. To
compensate long-range van der Waals (VdW) interaction between
adsorbate and zeolite, DFT-D3 scheme[28] with an empirical damped
potential term was added into the energies obtained from exchange-
correlation functional. Transition states for protonation, configura-
tion transformation, and dehydration steps in HBEA pore were
located using the IT-NEB method[29] with seven images between
initial and final state.
[12] a) J. F. Haw, B. R. Richardson, I. S. Oshiro, N. D. Lazo, J. A.
Speed, J. Am. Chem. Soc. 1989, 111, 2052 – 2058; b) W. Wang, J.
Jiao, Y. Jiang, S. S. Ray, M. Hunger, ChemPhysChem 2005, 6,
1467 – 1469; c) J. Z. Hu, J. A. Sears, H. S. Mehta, J. J. Ford, J. H.
Kwak, K. Zhu, Y. Wang, J. Liu, D. W. Hoyt, C. H. F. Peden, Phys.
Chem. Chem. Phys. 2012, 14, 2137 – 2143; d) A. Aerts, C. E. A.
Kirschhock, J. A. Martens, Chem. Soc. Rev. 2010, 39, 4626 – 4642;
e) H. Ernst, D. Freude, T. Mildner, I. Wolf, Solid State Nucl.
Magn. Reson. 1996, 6, 147 – 156.
[13] a) F. Taulelle, M. Haouas, C. Gerardin, C. Estourned, T. Loiseau,
G. Ferey, Colloids Surf. A 1999, 158, 299 – 311; b) O. B. Vistad,
D. E. Akporiaye, F. Taulelle, K. P. Lillerud, Chem. Mater. 2003,
15, 1639 – 1649.
[14] F. Zaera, Chem. Rev. 2012, 112, 2920 – 2986.
[15] a) D. W. Hoyt, R. V. F. Turcu, J. A. Sears, K. M. Rosso, S. D.
Burton, A. R. Felmy, J. Z. Hu, J. Magn. Reson. 2011, 212, 378 –
385; b) R. V. F. Turcu, D. W. Hoyt, K. M. Rosso, J. A. Sears, J. S.
Loring, A. R. Felmy, J. Z. Hu, J. Magn. Reson. 2013, 226, 64 – 69.
[16] a) F. Eder, M. Stockenhuber, J. A. Lercher, J. Phys. Chem. B
1997, 101, 5414; b) B. de Moor, M. F. Reyniers, J. A. Lercher, G.
Marin, J. Phys. Chem. C 2011, 115, 1204.
[17] T. S. Sorensen, S. M. Whitworth, J. Am. Chem. Soc. 1990, 112,
6647 – 6651.
Received: July 31, 2013
[18] a) P. A. Jacobs, M. Tielen, J. B. Uytterhoeven, J. Catal. 1977, 50,
98 – 108; b) S. J. Gentry, R. Rudham, J. Chem. Soc. Faraday
Trans. 1 1974, 70, 1685 – 1692.
Revised: October 8, 2013
Published online: November 26, 2013
[19] H. Chiang, A. Bhan, J. Catal. 2010, 271, 251 – 261.
[20] B. Hunger, S. Matysik, M. Heuchel, W. D. Einicke, J. Therm.
Anal. Calorim. 2001, 64, 1183 – 1190.
[21] T. Maihom, P. Khongpracha, J. Sirijaraensre, J. Limtrakul,
ChemPhysChem 2013, 14, 101 – 107.
Keywords: alcohol dehydration · heterogeneous catalysis ·
NMR spectroscopy · reaction mechanisms · zeolites
.
[1] D.-Y. Hong, S. J. Miller, P. K. Agrawal, C. W. Jones, Chem.
Commun. 2010, 46, 1038 – 1040.
[2] C. Zhao, J. He, A. A. Lemonidou, X. Li, J. A. Lercher, J. Catal.
2011, 280, 8 – 16.
[3] C. Zhao, S. Kasakov, J. He, J. A. Lercher, J. Catal. 2012, 296, 12 –
23.
[4] C. Zhao, J. A. Lercher, ChemCatChem 2012, 4, 64.
[5] C. Zhao, D. M. Camaioni, J. A. Lercher, J. Catal. 2012, 288, 92 –
103; C. Zhao, W. Song, J. A. Lercher, ACS Catal. 2012, 2, 2714 –
2723.
[6] a) E. G. Derouane, J. Catal. 1986, 100, 541; b) E. Kikuchi, H.
Nahano, K. Shinomura, Y. Morita, Sekiyu Gakkaishi 1985, 28,
210.
[7] M. Brꢂndle, J. Sauer, J. Am. Chem. Soc. 1998, 120, 1556 – 1570.
[8] F. Carey, R. Sundberg, Advanced Organic Chemistry, Springer,
New York, 2007, pp. 473 – 577.
[9] M. Janik, J. Macht, E. Inglesia, M. Neurock, J. Phys. Chem. C
2009, 113, 1872 – 1885.
[10] A. Corma, H. Garcia, Catal. Today 1997, 38, 257 – 308.
[11] D. E. Bryant, W. L. Kranich, J. Catal. 1967, 8, 8 – 13.
[22] a) S. R. Blaszkowski, R. A. van Santen, J. Am. Chem. Soc. 1996,
118, 5152 – 5153; b) S. Namuangruk, J. Meeprasert, P. Khem-
thong, K. Faungnawakij, J. Phys. Chem. C 2011, 115, 11649 –
11656; c) X. Liang, A. Montoya, B. S. Haynes, J. Phys. Chem. B
2011, 115, 8199 – 8206.
[23] T. Takahashi, H. Kawashima, H. Sugisawa, T. Baba, Solid State
Nucl. Magn. Reson. 1999, 15, 119 – 123.
[24] CP2K Open Source Molecular Dynamics Program, http://
www.cp2k.org/.
[25] a) S. Goedecker, M. Teter, J. Hutter, Phys. Rev. B 1996, 54, 1703;
b) C. Hartwigsen, S. Goedecker, J. Hutter, Phys. Rev. B 1998, 58,
3641; c) M. Krack, Theor. Chem. Acc. 2005, 114, 145.
[26] J. VandeVondele, J. Hutter, J. Chem. Phys. 2007, 127, 114105.
[27] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77,
3865.
[28] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010,
132, 154104.
[29] a) G. Henkelman, B. P. Uberuaga, H. Jonsson, J. Chem. Phys.
2000, 113, 9901; b) G. Mills, H. Jonsson, G. K. Schenter, Surf. Sci.
1995, 324, 305.
482
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 479 –482