Zeolite-supported rhodium complexes and clusters: Switching catalytic selectivity by controlling structures of essentially molecular species

Article abstract of DOI:10.1021/ja111749s

Precise synthesis and characterization of site-isolated rhodium complexes and extremely small rhodium clusters supported on zeolite HY allow control of the catalyst selectivity in the conversion of ethene to n-butene or ethane, respectively, as a result of tuning the structure of the active sites at a molecular level.

Full text of DOI:10.1021/ja111749s

COMMUNICATION
pubs.acs.org/JACS
Zeolite-Supported Rhodium Complexes and Clusters: Switching
Catalytic Selectivity by Controlling Structures of Essentially Molecular
Species
Pedro Serna and B. C. Gates*
Department of Chemical Engineering and Materials Science, University of California, One Shields Avenue, Davis, California 95616,
United States
S Supporting Information
b
diethene complexes anchored to the zeolite at acidic aluminum
ABSTRACT: Precise synthesis and characterization of site-
isolated rhodium complexes and extremely small rhodium
clusters supported on zeolite HY allow control of the
catalyst selectivity in the conversion of ethene to n-butene
or ethane, respectively, as a result of tuning the structure of
the active sites at a molecular level.
sites. The structure was determined by IR and extended X-ray
absorption fine structure (EXAFS) spectra, which showed that
the rhodium complex incorporating two ethene ligands π-
bonded to Rh (with a Rh-C coordination number of nearly
4) is linked to the zeolite by two Rh-O bonds [Table S1 and
Figure S1 in the Supporting Information (SI)]. The IR results
indicate that 20-25% of the protons initially present on the
zeolite are exchanged for the rhodium complexes in the synthesis
(Figure S2). No evidence of the formation of rhodium hydride
species was found (Figure S3).
Supported metals are a widely investigated and industrially
applied class of catalyst, and their properties are sensitive to
1,2
subtle changes in structure. The structural nonuniformity of
most solid catalysts limits the opportunities to control their
properties, but when the supported species incorporate only very
few metal atoms each, the surface chemistry becomes essentially
When the catalyst was treated with flowing H for 1.5 h at 1 bar
2
1
6
and room temperature, rhodium clusters formed, as expected.
The Rh-Rh coordination number determined by EXAFS spec-
troscopy was 1.9 (Table S1), showing an average nuclearity of
less than ∼3 and thus the presence of extremely small clusters
along with unconverted mononuclear rhodium species. The
cluster formation led to only a small decrease in the Rh-O
and Rh-Al coordination numbers, consistent with the smallness
of the clusters (Table S2). Thus, the data indicate a minimal
alteration of the metal-support bonding as a result of cluster
formation.
3
-6
molecular.
The supports then take on the role of ligands,
providing opportunities to fine-tune the catalytic properties that
7
-9
may rival those in organometallic solution catalysis.
Further-
more, when the metals are highly dispersed on a support, the
opportunities for involving the support in the catalysis are
1
0,11
optimized.
Here we report how to tune the catalytic properties of small,
nearly uniform clusters and site-isolated complexes of rhodium
on a highly uniform (crystalline) support, zeolite HY, for control
of the selectivity for alkene dimerization versus hydrogenation.
We show that complexes initially present as Rh(C H ) bonded
The catalytic behavior of each of the variously treated samples
was evaluated for the conversion of ethene in the presence of H2
in a once-through plug-flow reactor at atmospheric pressure and
303 K. The gas-phase products were monitored as a function of
time on stream with an online gas chromatograph. To provide
quantitative determinations of the catalytic activity, experiments
were done with the reactor operating with conversions in the
differential range (<10%, Figure S2), allowing the direct deter-
mination of reaction rates expressed as turnover frequencies
(TOFs) (Table 1). For each form of the catalyst, total numbers of
turnovers exceeding 1500 could always be attained within 4-10
h on stream, clearly demonstrating that the transformations are
2
4 2
to the zeolite catalyze both the dimerization and hydrogenation
of ethene, with a high selectivity for the former, and that
treatment of the catalyst to make rhodium clusters of only a
few atoms each, on average, boosts the activity for hydrogenation
and essentially eliminates that for dimerization. Moreover, these
changes are reversible.
Spectra of the working catalyst demonstrate the structural
changes that determine the selectivity, which occur at the very
early stages of cluster formation and breakup. The data also
demonstrate unprecedented performance of the site-isolated
rhodium complexes, as rhodium complexes in the absence of
labile ligands such as halides have been regarded as inactive for
17
catalytic.
In ethene-rich mixtures (C H /H molar ratio = 4), the
2
4
2
catalysts after each pretreatment were found to be active and
highly selective for ethene dimerization (Table 1); ethene
12-14
18
C-C bond formation.
We infer from our results that both
hydrogenation was only a minor side reaction. This result is
rhodium in the form of mononuclear species and the acidic
zeolite support are essential for catalysis of the dimerization
reaction.
surprising, as the activities of rhodium catalysts for formation of
C-C bonds are usually observed only when activating ligands
12-14
such as halides are bonded to the rhodium
and the
The structurally well-defined supported rhodium complex
selectivity for butenes characterizing other supported catalysts,
15
precursor was synthesized as before by reaction of Rh(C H ) -
2
4 2
(
(
acac) (acac = acetylacetonate) with dealuminated HY zeolite
Si/Al atom ratio = 30), leading to nearly uniform rhodium
Received: December 30, 2010
Published: March 10, 2011
r 2011 American Chemical Society
4714
dx.doi.org/10.1021/ja111749s
|J. Am. Chem. Soc. 2011, 133, 4714–4717

Journal of the American Chemical Society
COMMUNICATION
Table 1. Performance in a Once-Through Plug-Flow Reactor of the Catalyst Formed by Adsorption Of Rh(C H ) (acac) on
2
selectivity (mol %)
butenes
4 2
a
Dealuminated Zeolite HY
e,f
feed composition
/H molar ratio)
catalyst pretreatment
(C
2
H
4
2
TOF (s-
)
1 d
ethane
butane
b
none
4.00
0.25
4.00
0.25
0.075
0.260
0.071
0.932
18.9
22.0
20.4
89.3
78.2
72.5
76.5
7.6
1.9
4.5
3.1
3.1
b
none
c
2
H
H
c
2
a
b
Reaction conditions: once-through flow reactor operated at 303 K and 1 bar (C
2
H
4
and H
2
were present with helium at a partial pressure of 0.5 bar).
c
d
Catalyst used as prepared. Flowing H
2
at 303 K and 1 bar for 1.5 h prior to use. Turnover frequencies calculated from differential conversions (by
e
extrapolation to time = 0) assuming that all of the Rh atoms were accessible to the reactants. Selectivities were independent of conversion within the
range investigated (0.5-12%). As a secondary product, butane was invariably formed in low yields at the low observed conversions.
f
complexes was boosted by an increase in the H partial pressure
2
(
Figure S6), with only a moderate accompanying decrease in the
selectivity for butenes (from ∼78% with a C H /H molar ratio
2
4
2
22
of 4 to ∼72% with a C H /H molar ratio of 0.25).
2
4
2
Complementary experiments were carried out to characterize
the working catalyst in an EXAFS cell serving as a flow reactor
(
an integral reactor, not a differential one). To identify the
catalytically active rhodium species, we recorded spectra at the
rhodium K edge characterizing dynamic changes in the catalyst
structure after step changes in the feed C H /H ratio. The X-ray
2
4
2
2
3
beam was focused right at the inlet of the flow reactor/cell to
provide data representing the catalyst in an atmosphere essen-
tially matching that of the feed (and that throughout the
differential plug-flow reactor). EXAFS spectra were recorded at
3
min intervals during the experiment, and the composition of
the product from this integral flow reactor was determined
periodically by mass spectrometry. The composition of the feed
stream was cycled in the following sequence: (1) C H -rich; (2)
Figure 1. (top) Changes in selectivity of the supported catalyst initially
containing Rh(C supported on HY zeolite and (bottom) catalyst
2
4
2 4 2
H )
H -rich; (3) pure H ; and (4) H -rich. Specifics are given in the
2
2
2
structure represented by the Fourier transform function determined
from time-resolved EXAFS data as a function of the composition of the
feed (shown as molar ratios), which was cycled in the following
sequence: (1) 4:1:5 C H /H /He; (2) 1:4:5 C H /H /He; (3) pure
caption of Figure 1, which includes a summary of the time-
resolved EXAFS and mass spectrometric data recorded over
several cycles.
2
4
2
2
4
2
3
24
Fourier-transformed k -weighted EXAFS data are shown in
the bottom panel of Figure 1 as a function of time on stream for
the consecutive feed cycles (for details, see the figure caption).
The data show the evolution of the magnitude of the EXAFS
function (colored areas in the diagram) for Rh-backscatterer
contributions at short distances (∼1.6 Å, not phase-shift-
corrected) assigned to Rh-O and Rh-C together with the
H ; (4) 1:4:5 C H /H /He. In the bottom panel, the horizontal axis
2
2
4
2
represents time on stream, the vertical axis the Rh-backscatterer
distance (not corrected for phase shifts), and the colors the magnitudes
of various contributions (related to the abundance of backscatterer
atoms at a particular Rh-backscatterer distance; a change in color from
red to yellow to green to blue shows a continuing decrease in intensity of
the contribution). The peak at ∼2.5 Å includes the Rh-Rh contribution
together with Rh-Al and Rh-C contributions (comparatively much
long
longer-distance contribution (∼2.5 Å, not phase-shift-corrected)
less intense), and the peak at ∼1.6 Å includes the Rh-low-Z-scatterer
contributions, Rh-O and Rh-C. Details of the EXAFS data fitting are
provided in Table S1. The experiment was carried out in an EXAFS cell/
flow reactor at 303 K and 1 bar.
15
ascribed to Rh-Rh, Rh-Al, and Rh-C . For each step in the
long
cycle, changes in the magnitude of the Fourier transform for each
contribution were observed (Figure 1), indicating changes in the
structure/ligation of the Rh atoms in the catalytic sites, as
explained below. The catalyst selectivity data (represented by
fragments at m/z30 and 41 corresponding to ethane and butenes,
respectively, in the effluent), which are plotted above the EXAFS
results in Figure 1, are clearly correlated with changes in the
catalyst structure (these selectivity data determined with the
integral reactor only approximately match those obtained with
1
9
20
such as reduced cobalt oxides, nickel oxides, and ruthenium
2
1
clusters, in the presence of H rarely exceeds 1%.
2
In contrast, when the feed stream was H -rich (C H /H
2
2
4
2
molar ratio = 0.25), markedly different product distributions
were observed for the two pretreatments of the catalyst
25
(
Table 1). The sample initially incorporating rhodium clusters
the differential reactor). More detailed structural information
provided by the EXAFS results is given in Figure S8 and Table
S1; details of the data fitting are given in the SI.
(along with rhodium complexes) was highly active and selective
for hydrogenation, in contrast to the catalyst containing the
rhodium complexes without clusters, which was selective for
dimerization. Moreover, the rate of dimerization on the rhodium
The results in Figure 1 show that when the ethene-rich
mixture initially flowed through the bed of catalyst containing
4
715
dx.doi.org/10.1021/ja111749s |J. Am. Chem. Soc. 2011, 133, 4714–4717

Journal of the American Chemical Society
COMMUNICATION
the mononuclear rhodium complexes (cycle step 1), n-butenes
were the major catalytic reaction products. The data measured in
the differential reactor determined the rates of formation of n-
-
1
butenes and ethane to be 0.059 and 0.014 s , respectively
-
1
(
corresponding to a total TOF of 0.075 s ; Table 1, entry 1).
Under these conditions, the mononuclear rhodium complexes
were stable, with no Rh-Rh contributions (no clusters) detected
by EXAFS spectroscopy, although a decrease in the Rh-light-
scatterer contribution was observed with increasing time on
stream (Figure 1 and Figure S8), indicating a change in the
ligation of the rhodium (for details, see Table S1). Specifically,
the EXAFS data (Figure S8) indicate that during the selective
dimerization reaction, each Rh atom was, on average, bonded to
only 2.5 C atoms, consistent with the inference that a fraction of
the initially present π-bonded ethene ligands were converted to
reaction intermediates such as alkyls. Correspondingly, a weak
contribution (identified as a second Rh-C shell) was observed at
a distance longer than a bonding distance (Table S1), as expected
for butyl ligands formed on the rhodium sites during the
dimerization reaction. The formation of oligomeric intermedi-
ates on the surface of the catalyst was also evidenced by the IR
spectra, which showed conversion of the ethene ligands initially
Figure 2. Simplified schematic representation of the state of the
rhodium (green) in the catalyst attained at each step in the cycle, as
the catalytic performance of the solid is tuned.
A subsequent increase of the C H /H molar ratio to 4
2
4
2
(making H the limiting reactant) led to the rapid breakup of
2
2
6
the rhodium clusters by oxidative fragmentation by ethene, as
indicated by the disappearance of the Rh-Rh contribution
(Figure 1), accompanied by a sharp drop in the rate of hydro-
genation and a concomitant increase in the rate of n-butene
formation. The rate of ethane formation declined from 0.83 to
-
1
0.057 s , and the rate of dimer formation increased from 0.071
-
1
π-bonded to the rhodium sites (bands at 3084, 3060, and
to 0.19 s , corresponding to a marked increase in the selectivity
for dimerization over hydrogenation, with the ratio of the butene
and ethane formation rates increasing 40-fold.
1
3
016 cm- ) and the appearance of new bands at 2875, 2935,
-
1
and 2960 cm as a result of the dimerization reaction (see
Figure S8 for details).
The process described above was carried out through repeated
cycles (Figure 1); the changes were reversible.
When the H /ethene partial pressure ratio was subsequently
2
increased to 4 (cycle step 2; Table 1, entry 2), the catalyst was still
selective for n-butene formation, but the yields of the hydro-
genation products ethane and butane (a secondary product, not
shown in the plot) increased. Within the error of the EXAFS data,
there was no evidence of rhodium species other than mono-
nuclear complexes during this phase of the experiment, which
indicates that the presence of ethene in the gas phase, even at low
concentrations, stabilizes these mononuclear species, being a
continuous source of ligands to maintain the rhodium nuclearity.
However, a further decrease observed in the ratio of the short to
long Rh-C contributions (Table S2) suggests a higher conver-
sion of the adsorbed ethene ligands into reaction intermediates
The EXAFS data demonstrate that in H -rich reaction mix-
2
tures, both the rhodium complexes and the rhodium clusters are
stable (no changes were observed in the rhodium nuclearity
during the experiment); the complexes selectively produce n-
butenes and the clusters selectively produce ethane (see cycle
steps 2 and 4 in Figure 1, corresponding to entries 2 and 4 in
Table 1). Thus, our data allow a quantification of the changes in
the catalyst behavior at the very early stages of cluster formation,
when only few Rh-Rh bonds had been formed and only a portion
of the rhodium had been converted into clusters (Figure 2). When
hydrogenation was the predominant reaction, it was so fast in
comparison with dimerization that the latter was swamped. The
data show that the clusters are highly active catalysts for C-H bond
formation, much more active than the complexes.
such as butyl (or ethyl), corresponding to the higher H partial
2
pressure.
In the next phase of the experiment (cycle step 3), the feed was
The mononuclear Rh(I) species catalyze both hydrogenation
and dimerization. The latter has been observed for halide-
switched to pure H , with the catalyst still incorporating the
2
12,27
mononuclear rhodium species. During the next 1.5 h, a Rh-Rh
contribution gradually grew in (Figure 1) and reached a final
coordination number of 1.9 (Figure S9), indicating the formation
of clusters. Concomitantly, the Rh-C, Rh-Al, and Rh-O
contributions decreased slightly (Figure S8 and Table S1). After
the H treatment, the feed was switched to an H -rich mixture
containing rhodium complexes,
but our new observation of
such activity for the rhodium complex on the halide-free zeolite
indicates an important role of the acidic support in the formation
of C-C bonds. We emphasize that the precursor of our
supported catalyst, Rh(C H ) (acac), itself does not catalyze
2
4 2
12
ethene dimerization. Indeed, chemisorption of this precursor
on highly dehydroxylated MgO led to supported Rh(C H )
2 4 2
2
2
(
H /ethene molar ratio = 4). The rhodium clusters were then
2
stable and highly active catalytically, and the selectivity for
hydrogenation was high (Figure 1, corresponding to Table 1,
entry 4). Under these conditions, the stability of the rhodium
species that were isostructural to those in the zeolite-supported
rhodium complexes (Table S3) but shown by our data to be
inactive for the formation of butenes, being 100% selective for
hydrogenation in the presence of H2.
clusters in the working state corresponds to the presence of H in
2
large excess, and any ethene present is rapidly converted on the
clusters.
The results of blank experiments confirmed that the zeolite
alone did not provide the catalytic activity for dimer formation at
A comparison of the rates of dimerization and hydrogenation
the temperature of our experiments (with or without H co-fed
2
during catalysis in the H -rich mixture before and after cluster
formation indicated a dramatic increase in the rate of hydro-
with the alkene). Formation of C-C bonds on the zeolite started
2
only when the temperature was increased to 433 K, consistent
28,29
genation upon formation of clusters: the ethane formation rate
with previous reports,
orders of magnitude less than with the zeolite-supported rho-
and even then the dimer yield was 2
-
1
-1
increased from 0.057 s with only complexes to 0.83 s when
clusters were present.
30
dium complexes operating at room temperature.
4
716
dx.doi.org/10.1021/ja111749s |J. Am. Chem. Soc. 2011, 133, 4714–4717

Journal of the American Chemical Society
COMMUNICATION
In summary, our acidic zeolite-supported rhodium complex
catalyst is active for dimerization in the absence of halides and
selective for this reaction in a reducing atmosphere of H2
(17) Deactivation of the catalyst was observed, as shown in Figure
S5, and we infer that it was caused by the formation of oligomers that
adsorb strongly on the catalytically active species.
(18) As demonstrated later in the manuscript, the similar catalytic
(although the expected hydrogenation of the CdC bond occurs
performance of reduced and unreduced catalysts in flowing ethene-rich
mixtures (Table 1, entries 3 and 1, respectively) arises because the
rhodium clusters initially present in the former break up upon contact
with the feed, so both classes of samples are characterized by the
presence of mononuclear rhodium complexes under working
conditions.
(19) Kokes, R. J. J. Catal. 1969, 14, 83.
(20) Kokes, R. J.; Bartek, J. P. J. Catal. 1968, 12, 72.
(21) Rodriguez, E.; Leconte, M.; Basset, J.-M.; Tanaka, K. J. Catal.
as an accompanying reaction; Table 1). On the basis of our
results, we infer an important role of the zeolite as a macroligand
of the rhodium complexes, altering the electron density on the
rhodium or opening the possibility for cooperation between the
3
1
rhodium complexes and the acidic Al-OH sites of the zeolite.
Further, the results demonstrate how regulation of the structure
of the active sites of a supported catalyst at the molecular level
(switching between rhodium complexes and rhodium clusters)
1
989, 119, 230.
22) The role of H in dimerization reactions has often been ascribed
allows fine-tuning of the catalyst selectivity simply by variation of
the feed composition. To our knowledge, this is the first example
of precise control of the selectivity of a solid catalyst by tuning of
the structure of essentially molecular supported species.
(
2
to the generation (or regeneration) of metal hydride species that
facilitate the activation of the alkene or the desorption of dimeric
intermediates. For a comprehensive review, see: Pillai, S. M.; Ravindra-
nathan, M.; Sivaram, S. Chem. Rev. 1986, 86, 353.
(
23) Odzak, J. F.; Argo, A. M.; Lai, F. S.; Gates, B. C.; Pandya, K.;
Feraria, L. Rev. Sci. Instrum. 2001, 72, 3943.
24) Koningsberger, D. C.; Mojet, B. L.; van Dorssen, G. E.; Ra-
maker, D. E. Top. Catal. 2000, 10, 143.
25) Some differences in the distribution of reaction products shown
in Table 1 and Figure 1 were expected, as the EXAFS cell used as a
reactor to collect the data of Figure 1 could not be operated in the
differential conversion range because of the limitation that relatively
large masses of catalyst had to be used to optimize the quality of the
EXAFS data.
’
ASSOCIATED CONTENT
(
S
Supporting Information. Experimental procedures,
b
synthesis of catalysts, reactivity tests, and details of the EXAFS
fitting. This material is available free of charge via the Internet at
http://pubs.acs.org.
(
’
AUTHOR INFORMATION
Corresponding Author
bcgates@ucdavis.edu
(26) Oxidative fragmentation of small metal clusters by ethene in
solution has been demonstrated. See: (a) Adams, C. J.; Bruce, M. I.;
Liddell, M. J.; Tiekink, E. R. T.; Skelton, B. W.; White, A. H. J.
Organomet. Chem. 1993, 445, 187. (b) Kampe, C. E.; Boag, N. M.;
Kaesz, H. D. J. Mol. Catal. 1983, 21, 297. More recently, it has also been
demostrated on solid surfaces. See: (c) Uzun, A.; Gates, B. C. Angew.
Chem., Int. Ed. 2008, 47, 9245. The mechanism of this transformation
remains uncharacterized. On the basis of the changes that we deter-
mined by EXAFS spectroscopy upon breakup of the rhodium clusters,
’
ACKNOWLEDGMENT
We thank DOE (Basic Energy Sciences, Contract FG02-
7ER15600) for support and acknowledge beam time and
8
support of the DOE Division of Materials Sciences for its role
in the operation and development of beamline MR-CAT at the
Advanced Photon Source at Argonne National Laboratory.
the following approximate stoichiometry is tentatively proposed: Rh
3
-
(
C
2
H
5
)
3
þ 3C
2
H
4
þ 1.5H
2
f 3Rh(C . We emphasize the
2 5 2
H )
approximate nature of this statement, as EXAFS spectroscopy provides
only average information about the structure of the rhodium sites.
(
27) Takahashi, N.; Okura, I.; Keii, T. J. Am. Chem. Soc. 1975,
’
REFERENCES
9
7, 7489.
(
1) Lee, I.; Delbecq, F.; Morales, R.; Albiter, M. A.; Zaera, F. Nat.
(28) Derouane, E. G.; Gilson, J.-P.; Nagy, J. B. J. Mol. Catal. 1981,
10, 331.
(29) Cant, N. W. F.; Hall, W. K. J. Catal. 1972, 25, 161.
(30) Poisoning of the anchored rhodium complexes with small
amounts of CO, leading to the formation of anchored rhodium gem-
2
dicarbonyl [Rh(CO) ] species as shown by IR spectroscopy (Figure
Mater. 2009, 8, 132.
(
(
(
2) Bell, A. T. Science 2003, 299, 1688.
3) Fujdala, K. L.; Tilley, T. D. J. Catal. 2003, 216, 265.
4) DeVos, D. E.; Dams, M.; Sels, B. F.; Jacobs, P. A. Chem. Rev.
2
002, 102, 3615.
(
(
5) Thomas, J. M. Top. Catal. 2001, 15, 85.
6) Xu, Z.; Xiao, F.-S.; Purnell, S. K.; Alexeev, O.; Kawi, S.; Deutsch,
S1), led to complete deactivation of the zeolite-supported catalyst,
consistent with the inference that the rhodium species participate in
the C-C bond formation reaction (the interaction of CO with the acidic
sites of the zeolite is very weak at 303 K, as evidenced by the result that
the intensities of the IR bands characterizing the acidic -OH groups at
S. E.; Gates, B. C. Nature 2002, 372, 346.
7) Corker, J.; Lefebvre, F.; Evans, J.; L ꢀe cuyer, C.; Dufaud, V.;
Quignard, F.; Choplin, A.; Basset, J.-M. Science 1996, 271, 966.
8) Fujdala, K. L.; Drake, I. J.; Bell, A. T.; Tilley, T. D. J. Am. Chem.
Soc. 2004, 126, 10864.
9) Alacon, M. J.; Corma, A.; Iglesias, M.; Sanchez, F. J. Organomet.
Chem. 2002, 655, 134.
(
-
1
(
3630 and 3535 cm did not decrease upon contact of the sample with
CO).
(
(31) Sachtler, W. M. H.; Stakheev, A. Yu. Catal. Today 1992, 12, 283.
(
(
(
(
10) Sachtler, W. M. H.; Zhang, Z. Adv. Catal. 1993, 39, 129.
11) Stakheev, A. Yu.; Kustov, L. M. Appl. Catal., A 1999, 188, 3.
12) Cramer, R. J. Am. Chem. Soc. 1965, 87, 4717.
13) Alderson, T.; Jenner, E. L.; Lindsey, R. V. J. Am. Chem. Soc.
1
965, 87, 5638.
(
14) Fagnou, K.; Lautens, M. Chem. Rev. 2003, 103, 169.
(15) Liang, A. J.; Bhirud, V. A.; Ehresmann, J. O.; Kletnieks, P. W.;
Haw, J. F.; Gates, B. C. J. Phys. Chem. B 2005, 109, 24236.
16) Liang, A. J.; Gates, B. C. J. Phys. Chem. C 2008, 112, 18039.
(
4
717
dx.doi.org/10.1021/ja111749s |J. Am. Chem. Soc. 2011, 133, 4714–4717