The pathway to total isomer selectivity: N-hexane conversion (reforming) on platinum nanoparticles supported on aluminum modified mesoporous silica (MCF-17)

Article abstract of DOI:10.1021/ja509638w

When pure mesoporous silica (MCF-17) was modified with aluminum (Al modified MCF-17), Lewis acid sites were created, but this material was inactive for the catalytic conversion (reforming) of n-hexane to isomers. When colloidally synthesized platinum nanoparticles were loaded onto traditional MCF-17, the catalyst showed very low activity toward isomer production. However, when Pt nanoparticles were loaded onto Al modified MCF-17, isomerization became the dominant catalytic pathway, with extremely high activity and selectivity (>90%), even at high temperatures (240-360 °C). This highly efficient catalytic chemistry was credited to the tandem effect between the acidic Al modified MCF-17 and the Pt metal.

Full text of DOI:10.1021/ja509638w

Article
pubs.acs.org/JACS
The Pathway to Total Isomer Selectivity: n‑Hexane Conversion
(Reforming) on Platinum Nanoparticles Supported on Aluminum
Modified Mesoporous Silica (MCF-17)
Nathan Musselwhite, Kyungsu Na, Selim Alayoglu,* and Gabor A. Somorjai*
Department of Chemistry, University of California, Berkeley, Berkeley, California 94720, United States
Chemical Sciences Division, Lawrence Berkeley National Laboratory, One Cyclotron Road, Berkeley, California 94720, United States
ABSTRACT: When pure mesoporous silica (MCF-17) was modified with
aluminum (Al modified MCF-17), Lewis acid sites were created, but this material
was inactive for the catalytic conversion (reforming) of n-hexane to isomers. When
colloidally synthesized platinum nanoparticles were loaded onto traditional MCF-17,
the catalyst showed very low activity toward isomer production. However, when Pt
nanoparticles were loaded onto Al modified MCF-17, isomerization became the
dominant catalytic pathway, with extremely high activity and selectivity (>90%), even
at high temperatures (240−360 °C). This highly efficient catalytic chemistry was
credited to the tandem effect between the acidic Al modified MCF-17 and the Pt
metal.
Scheme 1. Possible Reaction Products for n-Hexane
Reforming
INTRODUCTION
■
The catalytic conversion (reforming) of linear hydrocarbons to
their corresponding branched isomers is a vital reaction in the
production of high octane gasoline from the naphtha feedstock
of crude oil.1 The target, like all catalytic reactions, is to
optimize both the overall activity and product selectivity,
simultaneously.1−3 In naphtha reforming, branched isomers are
the desired product, while cracking the hydrocarbon creates the
unwanted byproducts. This chemistry is normally accomplished
through the use of a platinum catalyst, alloyed with small
amounts of other transition metals, and loaded onto an acidic
support, such as a zeolite.4−7 It is believed that the platinum
catalyzes the dehydrogenation and hydrogenation of the
hydrocarbons (promoted by the alloying of additional metals),
while the isomerization occurs on the acidic support via
carbocation transition species.8,9
The reforming of n-hexane provides a good model reaction
in order to scientifically study the complex process of naphtha
refining. Hexane can react via five major pathways: (1) cracking
to shorter chain hydrocarbons; (2) isomerization to 2-
methylpentane, 3-methylpentane, or multibranched isomers;
(3) cyclization to methylcyclopentane or cyclohexane; (4)
aromatization to benzene; (5) dehydrogenation to hexene or
other unsaturated compounds. This possible reaction products
for n-hexane reforming are displayed in Scheme 1.
modified MCF-17 or Pt nanoparticles loaded onto unmodified
MCF-17 were tested separately, it was found that both catalysts
were relatively inactive. This behavior was attributed to metal−
oxide interface formed between the Al modified MCF-17 and
the Pt nanoparticles.
METHODS
■
MCF-17 type mesoporous silica was synthesized by previously
reported literature methods.10 Briefly, 1,3,5-trimethylbenzene
(TMB), which was utilized as a pore swelling agent, was added to
an aqueous solution of triblockcopolymer Pluoronic P123 and HCl.
After stirring of this solution for 2 h at 40 °C, Tetraethylorthosilicate
(TEOS) was then added and the solution was stirred for an additional
In the present paper, we report on the evaluation of several
catalysts in the vapor phase hydrogenation of n-hexane. When
Al modified mesoporous silica MCF-17 loaded with Pt
nanoparticles (Al modified MCF-17/Pt) was evaluated, it was
found to be highly active and selective (>90%) toward isomer
production, even at high temperatures. When either the Al
Received: September 18, 2014
Published: October 31, 2014
© 2014 American Chemical Society
16661
dx.doi.org/10.1021/ja509638w | J. Am. Chem. Soc. 2014, 136, 16661−16665

Journal of the American Chemical Society
Article
20 h. NH4F was then added, and the solution was allowed to
hydrothermally react at 100 °C for 24 h. The product was then
calcined for 6 h at 550 °C. The pore size of the material is 30−50 nm
and the surface area is ∼1000 m2/g.
RESULTS AND DISCUSSION
The overall activity of all tested catalysts is shown in Figure 1.
The activity is displayed as a turnover frequency, which is based
■
The mesopore surface of pure silica MCF-17 was aluminated
through utilization of the grafting method.11 The as-synthesized MCF-
17 was calcined at 550 °C for 4 h in air to remove organic molecules.
Then, the calcined sample was grafted with aluminum by slurring with
anhydrous AlCl3 in absolute ethanol, in order to give a Si/Al ratio of
10:1. The slurring solution was stirred overnight at room temperature,
and then the ethanol was removed by rotary-evaporation. The
precipitated Al-modified MCF-17 was dried at 130 °C for 1 h, and
subsequently calcined at 550 °C for 4 h in air. For the introduction of
acidic sites, the calcined Al-modified MCF-17 was slurred in 1 M of
NH4NO3 aqueous solution for 4 h at room temperature.12 The sample
was then filtered, washed with distilled water, and dried at 130 °C for 1
h. This process was repeated three times and the final sample was
calcined at 550 °C for 4 h in air to give H+-Al-modified MCF-17.
Polyvinylpyrrolidone (PVP) capped Pt nanoparticles with an
average particle size of 2.5 nm were synthesized and supported on
mesoporous silica (MCF-17), Al modified MCF-17, and mesoporous
zeolite (MFI) according to previously reported methods.13 Briefly,
H2PtCl6 precursor salt was dissolved in ethylene glycol in the presence
of PVP, then was allowed to react at 470 K, until particles were formed
and stable. The as-synthesized nanoparticles were then washed and
redispersed in ethanol. To load the nanoparticles on the various
support materials, a colloidal solution of PVP-capped Pt nanoparticles
in ethanol was mixed with the desired support material to give a
nominal metal loading of 0.5 wt %. The mixtures were then sonicated
for 3 h, and the supported catalysts were then collected by
centrifugation and washed with 20% ethanol in acetone.
After synthesis, the catalytic behavior of the materials was
investigated in the reforming of n-hexane. Catalytic measurements
were made utilizing a tubular fixed catalyst bed reactor at ambient
pressure, which has been described in previous publications.14−16
Briefly, a 1/4 in. diameter stainless steel reactor was loaded with 0.02−
0.5 g catalyst (which was pelletized and sieved to yield 60−100 μm
size granulates), then capped on each end with a purified thermal silica
filter. The remaining space in the reactor tube was filled with purified
fused aluminum granulate and capped with glass wool. To keep the
catalysis in a kinetic region and allow for selectivity comparisons, the
total hexane conversion was held between 1% and 5%.
Figure 1. Overall turnover frequency (moles hexane converted mole
Pt site per second) plotted against temperature for evaluated catalysts.
Pt loaded on MCF-17 is shown in red for reference and is scaled by a
factor of 10 to allow it to be plotted on the graph. Al modified MCF-
17 is catalytically inactive, so is not plotted. Al modified MCF-17
loaded with Pt is shown in blue. The TOF for the Al modified catalyst
is about 43.5× greater at the highest temperature studied (360 °C).
on the total surface area of platinum, which is calculated based
on geometrical methods and TEM analysis of average
nanoparticle size (2.5 nm). The reference catalyst of Pt
nanoparticles loaded on mesoporous silica MCF-17 (red)
displayed very low activity, and had to be scaled by a factor of
10 in order to be plotted on the same graph. The Al modified
MCF-17 is not shown in the plot because it was found to be
catalytically inactive at all temperatures studied. The Al
modified MCF-17/Pt, which is shown in blue, displayed high
activity compared to the metal free support and the unmodified
MCF-17/Pt catalyst.
After loading, the catalysts were first pretreated at 633 K under a gas
mixture of N2 (Praxair, 5.0 UHP, 10 sccm) and H2 (Praxair, 5.0 UHP,
10 sccm) for 2 h, with a heating rate of 2 K min−1. After pretreatment,
the reactor system was cooled to 513 K, under the same gas flow. The
gas flow was then changed to 16 sccm H2, and n-hexane (Fluka, ≥
99.0%) was introduced using a Teledyne ISCO 500D liquid flow
pump at a rate of 1.2 mL h−1 into the heated reactor head which was
maintained at 423 K. In the reactor head, hexane was evaporated and
mixed with H2, resulting in a two-component gas flow with a
hexane:H2 ratio of 1:5 entering the reactor at near ambient pressure. A
Baratron type (890B, MKS Instruments) manometer was used to
monitor the reactor inlet pressure. The reaction products were
sampled in the vapor phase at the reactor outlet and analyzed via an in-
line gas chromatograph (GC), with all flow lines heated to 433 K.
Quantitative analysis of product composition was accomplished with a
Hewlett-Packard (5890 Series II) GC which was equipped with an
Aldrich HP-1 capillary column and a flame ionization detector (FID).
A PC based GC Chemstation software (Hewlett-Packard) was utilized
for automatic GC sampling, data collection and postrun processing.
Postreaction characterization of the Al modified MCF-17/Pt catalyst
using scanning transmission electron microscopy (STEM)/energy
dispersive spectroscopy (EDS) was conducted using a Jeol 2100F
TEM. Point-to-point spatial resolution of the electron probe was 1.5
nm.
The overall isomer activity of the hexane reforming reaction
for each catalysts is shown in Figure 2; this data is recorded at
the highest temperature studied (360 °C). This data was
obtained by taking the product of the overall isomer selectivity
and the overall turnover frequency, in order to obtain the
isomer turnover frequency (moles isomer produced per mole
Pt site per second). The MCF-17/Pt catalyst (red) shows very
low isomer activity (9.66 × 10−4 s−1), due to the lack of acidity.
The Al Modified MCF-17 (orange “x”) catalyst shows no
activity when Pt is not present. However, when Pt is loaded
into the Al modified MCF-17 catalyst (blue), the isomer
production is highly increased to 1.18 × 10−1 s−1, which is
about 120 times the isomer activity of the MCF-17/Pt catalyst
and the Al modified MCF-17 support separately.
The specific product selectivity of the reaction is shown in
Figure 3; this data is recorded at the highest temperature
reactions (360 °C). The data shown is at approximately 5%
total hexane conversion. Al modified MCF-17 without Pt
loading showed no activity at this temperature. Isomerization is
the dominant pathway on the Al modified MCF-17/Pt catalyst,
with high selectivity (>90%) toward 2-methylpentane (2MP)
and 3-methylpentane (3MP). The MCF-17/Pt (red) shows
∼20% selectivity toward C1−C5 cracking products, and ∼7%
16662
dx.doi.org/10.1021/ja509638w | J. Am. Chem. Soc. 2014, 136, 16661−16665

Journal of the American Chemical Society
Article
Al modification and subsequent Pt loading of the MCF-17
support.
The STEM, TEM, and HR-TEM images of the postreaction
Al modified MCF-17/Pt catalyst are shown in Figure 4, in parts
a, b, and c, respectively. It can be seen from the STEM and
TEM images that the particles are well dispersed throughout
the catalyst, with little agglomeration of the metal. The
representative STEM images and corresponding color-coded
EDS spectral maps at Si K (green), Al K (blue), and Pt M (red)
lines are shown in Figure 4d. It can be seen that the Al is well
dispersed among the catalyst, and segregated Al domains do
not exist on either the silica support or around the Pt
nanoparticles. It is also important to note that no evidence of
chlorine was observed in the EDS analysis, as residual Cl− from
the AlCl3 used for support modification could have an effect on
the Brønsted acidity of the catalyst.
The high temperature isomer selectivity of the Al modified
MCF-17 catalyst is a remarkable finding, and most likely a
direct effect of the interface between the oxide support and the
Pt nanoparticles. Zeolites contain catalytically active micropores
with both Lewis and Brønsted acid sites.17 These acidic
micropores are capable of catalyzing the n-hexane reaction
without the use of any metal catalyst. Zeolite catalysis favors
isomerization at low temperature, but at high temperature,
cracking is the dominant pathway, even when Pt is loaded onto
the zeolite.18 The high temperature catalytic behavior of the Al
modified MCF-17/Pt catalyst is strikingly different from the
zeolite catalysts. From Figures 2 and 3, it can be seen that the
Al modified MCF-17/Pt catalyst is highly selective for
isomerization, at a temperature at which traditional zeolite
catalysts perform predominantly cracking reactions. This
favorable selectivity is attributed to the Al modified MCF-17
support providing acidity, without catalyzing the reaction by
itself. It is well-known that aluminum modification of silica
provides strong Lewis acid sites which are not present in
normal silica.19−21 These Lewis acid sites are known to stabilize
carbocation transition state species, which in turn with the
dehydrogenation/hydrogenation chemistry of the Pt metal, act
to promote the production of isomers through a bifunctional
mechanism on a traditional reforming catalyst.22−26
Figure 2. The product of the overall catalytic turnover and the isomer
selectivity at 360 °C reaction in order to show the enhancement of
adding acidity to catalysis. The MCF-17/Pt catalyst (red) shows very
low activity (9.66 × 10−4 s−1) without the presence of acidity. The Al
Modified MCF-17 catalyst shows no activity when Pt is not present.
However, when Pt is loaded into the Al modified MCF-17 catalyst
(blue) the isomer production is highly increased to 1.18 × 10−1 s−1,
which is about 120 times the activity of the catalyst and support
separately.
To determine the type of mechanism present in this system,
a catalytic experiment using a physical mixture of Pt supported
on unmodified MCF-17 and Pt free Al modified MCF-17 was
run. If a bifunctional promotion of isomerization between the
Pt nanoparticles and the Al modified MCF-17 support exists,
the catalytic results should be identical to the results obtained
from the Al modified MCF-17/Pt catalyst. It was found that the
selectivity for this reaction is similar to that for the Al modified
MCF-17 with Pt loaded (selectivity data shown in Figure 3).
However, the activity of the mixed catalysts was an order of
magnitude lower than that of the Al modified MCF-17 catalyst
(1.08 × 10−2 s−1). This could be explained by the fact that a
bifunctional mechanism may be occurring; however, it appears
that a majority of isomer products formed from the Al modified
MCF-17/Pt catalyst require the direct contact of the Al
modified support and the Pt metal. This result suggests that the
primary active sites in this reaction, the Pt metal surface and the
acidic Al sites, must be in a very close proximity to each other
(nanometers not microns) in order to produce the high
isomerization activity.
Figure 3. Product selectivity at high temperature reactions (360 °C).
Al modified MCF-17 without Pt loading showed no activity at this
temperature. Isomerization is the dominant pathway on the Al
modified MCF-17/Pt catalyst, with high selectivity (>90%) to 2-
methylpentane (2MP) and 3-methylpentane (3MP). The MCF-17/Pt
(red) shows relatively higher (∼20%) selectivity toward C1−C5
cracking products, and ∼7% selectivity to olefins (the major product
being hexene); the remaining products are isomerization products. A
physical mixture of the MCF-17/Pt and the Al modified MCF-17
(green) showed nearly identical selectivity to the Al modified MCF-
17/Pt catalyst, indicating a bifunctional mechanism may be present.
However, the activity for the physically mixed catalyst was an order of
magnitude lower, indicating that the close proximity of the Pt and
acidic catalytic sites is necessary for the high isomer production in the
Al Modified MCF-17/Pt catalyst. All data shown are at similar
conversion.
selectivity to olefins (the major product being hexene); the
remaining products are isomerization products. It can be seen
that when the same support is previously modified with Al, the
selectivity is altered to very little of these products, but greater
than 90% selectivity toward the 2MP and 3MP isomers. It can
be clearly ascertained that the selectivity is vastly superior after
Infrared spectroscopy was utilized to investigate the amount
of acid sites in the Al modified MCF-17 support, and the
resulting spectrum is shown in Figure 5. When we utilize the
16663
dx.doi.org/10.1021/ja509638w | J. Am. Chem. Soc. 2014, 136, 16661−16665

Journal of the American Chemical Society
Article
Figure 4. (a) STEM, (b) TEM, and (c) HR-TEM images of the post-reaction 2.5 nm Pt NPs supported on Al-modified MCF-17. (d) Representative
STEM image and corresponding color-coded EDS spectral maps at Si K (green), Al K (blue), and Pt M (red) lines. The scale bars in the images
correspond to (a) 50, (b) 20, (c) 1, and (d) 50 nm.
quantify the amount of each type of acid site in the support to
0.4 and 1.3 mmol/g, respectively. To put this into relation with
the amount of metal sites, the number of surface Pt atoms was
calculated based on spherical TEM projected surface areas and
metal loading (as determined by elemental analysis). The
concentration of surface Pt atoms per gram catalyst was found
to be 0.011 mmol/g, giving an acid site to metal site ratio of
about 40:1 for Brønsted sites and about 120:1 for Lewis sites. It
appears that when Pt nanoparticles are added to the
catalytically inactive modified silica support, the Pt metal
accomplishes the necessary dehydrogenation/hydrogenation
reactions to the hydrocarbon, while the Lewis acidity of the
support promotes isomerization through an enhanced for-
Figure 5. FT-IR spectrum of Al-modified MCF-17. Due to the large
density of silanol on the mesopore wall, an intense peak is observed at
3720 cm−1, which indicates a stretching band of the Si−OH bond.
After pyridine adsorption, both Brønsted and Lewis acid sites are
distinguishable at 1550 and 1450 cm−1, respectively (see inset
spectrum for magnification). When we utilize the adsorption
coefficients of Brønsted and Lewis acid sites (1.13 and 1.28 cm/
μmol, respectively),23 the amount of Brønsted and Lewis acid sites can
be quantified to 0.4 and 1.3 mmol/g, respectively.
mation of carbocation intermediates.
It was found that neither the Pt nanoparticle size nor the
concentration of Al added to the mesoporous silica altered
either the catalytic activity or selectivity. No changes in Pt
nanoparticle stability were found between the modified and
unmodified MCF-17 catalysts (based on catalytic deactivation
rates and post reaction TEM images). The Al modified MCF-
17/Pt catalyst also shows only a 20% deactivation after a
second cycle. It is well-known that Pt dissociates molecular
hydrogen, which can spill over onto the support and prevent
coke formation, which is the primary reason for catalytic
adsorption coefficients for both Lewis and Brønsted acid sites
27
(1.13 and 1.28 cm/umol, respectively),
it is possible to
16664
dx.doi.org/10.1021/ja509638w | J. Am. Chem. Soc. 2014, 136, 16661−16665

Journal of the American Chemical Society
Article
deactivation in zeolites.28,29 The catalyst studied in this work is
less susceptible to this type of catalytic deactivation.
(13) Alayoglu, S.; Aliaga, C.; Sprung, C.; Somorjai, G. A. Catal. Lett.
2011, 141, 914.
(14) Pushkarev, V. V.; An, K.; Alayoglu, S.; Beaumont, S. K.;
Somorjai, G. A. J. Catal. 2012, 292, 64.
(15) Pushkarev, V. V.; Musselwhite, N.; An, K.; Alayoglu, S.;
Somorjai, G. A. Nano Lett. 2012, 12, 5196.
(16) Musselwhite, N.; Alayoglu, S.; Melaet, G.; Pushkarev, V. V.;
Lindeman, A. E.; An, K.; G.A. Somorjai, G. A. Catal. Lett. 2013, 143,
907.
(17) Bhatia, S. Zeolite Catalysts: Principles and Applications; CRC
Press: Boca Raton, FL, 1990.
(18) Corma, A.; Planelles, J.; Sanchez-Marin, J.; Tomas, F. J. Catal.
1985, 93, 30.
(19) Fraile, J. M.; Garcia, J. I.; Mayoral, J. A.; Pires, E.; Salvatella, L.;
Ten, M. J. Phys. Chem. B 1999, 103, 1664.
(20) Fraile, J. M.; Garcia, J. I.; Mayoral, J. A.; Pires, E. J. Mol. Catal. A:
Chem. 1997, 119, 95.
(21) Xu, M.; Arnold, A.; Buchholz, A.; Wang, W.; Hunger, M. J. Phys.
Chem. B 2002, 106, 12140.
(22) Fung, J.; Wang, I. J. Catal. 1996, 164, 166.
(23) Triwahyono, S.; Jalil, A. A.; Ruslan, N. N.; Kamarudin, N. H. N.
J. Catal. 2013, 303, 50.
(24) Ramirez, S.; Viniegra, M.; Dominguez, J. M.; Schacht, P.; De
Menorval, L. C. Catal. Lett. 2000, 66, 25.
(25) Pajonk, G. M. Appl. Catal., A 2000, 202, 157.
(26) Roessner, F.; Roland, U. J. Mol. Catal. A: Chem. 1996, 112, 401.
(27) Ayrault, P.; Datka, J.; Laforge, S.; Martin, D.; Guisnet, M. J. Phys.
Chem. B 2004, 108, 13755.
CONCLUSIONS
■
In summary, three model catalysts were studied for the
isomerization of n-hexane. The catalysts were colloidally
synthesized Pt nanoparticles (2.5 nm) loaded onto mesoporous
silica (MCF-17), MCF-17 which had been modified with Al,
and the third catalyst was Al modified MCF-17 mesoporous
silica loaded with Pt nanoparticles (2.5 nm). It was found that
the unmodified MCF-17/Pt catalyst was selective toward the
undesired cracking products, and showed low overall activity
for n-hexane reforming, indicating that the Pt metal alone was
not a major contributor to the chemistry. The Al modified
MCF-17 catalyst without Pt was found to be catalytically
inactive at all studied temperatures, indicating the acidity alone
could not accomplish the isomerization chemistry. However,
when Pt nanoparticles were loaded onto this Al modified MCF-
17 support, the catalyst was found to be remarkably active and
selective for the isomerization of n-hexane. This remarkable
selectivity was attributed to the close proximity between the
acidic sites in the catalytically inactive oxide support and the Pt
metal, which act in tandem to significantly promote the
isomerization of the hydrocarbon.
AUTHOR INFORMATION
Corresponding Authors
salayoglu@lbl.gov
(28) Srinivas, S. T.; Rao, P. K. J. Catal. 1994, 148, 470.
(29) Guisnet, M.; Magnoux, P. Catal. Today 1997, 36, 477.
■
somorjai@berkeley.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work is funded by The Chevron Energy Technology
Company. We acknowledge support from the Director, Office
of Science, Office of Basic Energy Sciences, Division of
Chemical Sciences, Geological and Biosciences of the U.S.
DOE under contract DE-AC02-05CH11231. K.N. thanks the
Basic Science Research Program through the National Research
Foundation of Korea (NRF) funded by the Ministry of
Education (2012R1A6A3A03039602). Work at the Molecular
Foundry was supported by the Director, Office of Science,
Office of Basic Energy Sciences, Division of Material Sciences
and Engineering, of the U.S. Department of Energy under
Contract No. DE-AC02-05CH11231.
REFERENCES
■
(1) Somorjai, G.A.; Li, Y. Introduction to Surface Chemistry and
Catalysis; Wiley: Hoboken, NJ, 2010.
(2) Somorjai, G. A.; Park, J. Y. Angew. Chem., Int. Ed. 2008, 47, 9212.
(3) Clark, J. H. Green Chem. 1999, 1, 1.
(4) Sinfelt, J. H. Catal. Lett. 1991, 9, 159.
(5) Han, Y.; Lee, S. S.; Ying, J. Y. Chem. Mater. 2007, 19, 22922.
(6) Yang, M.; Somorjai, G. A. J. Am. Chem. Soc. 2004, 126, 7698.
(7) Davis, S. M.; Zaera, F.; Somorjai, G. A. J. Am. Chem. Soc. 1982,
104, 7453.
(8) Iglesia, E.; Soled, D. G.; Kramer, G. M. J. Catal. 1993, 144, 238.
(9) Antos, G.J.; Aitani, A.M. Catalytic Naphtha Reforming; Marcel
Dekker: New York, 2004.
(10) Han, Y.; Lee, S. S.; Ying, J. Y. Chem. Mater. 2007, 19, 22922.
(11) Jun, S.; Ryoo, R. J. Catal. 2000, 195, 237.
(12) Na, K.; Jo, C.; Kim, J.; Cho, K.; Jung, J.; Seo, Y.; Messinger, R. J.;
Chmelka, B. F.; Ryoo, R. Science 2011, 333, 328.
16665
dx.doi.org/10.1021/ja509638w | J. Am. Chem. Soc. 2014, 136, 16661−16665