RuO2 clusters within LTA zeolite cages: Consequences of encapsulation on catalytic reactivity and selectivity

Article abstract of DOI:10.1002/anie.200700128

(Figure Presented) Trapped! The title system (ca. 1 nm diameter; left) catalyzes methanol oxidation with higher turnover rates than clusters on SiO₂ supports. Spatial constraints lead to the preferential oxidation of methanol over larger alcohols. Restricted access to active sites also protects encapsulated Ru clusters (right) against inhibition of ethene hydrogenation by organosulfur compounds.

Full text of DOI:10.1002/anie.200700128

Angewandte
Chemie
DOI: 10.1002/anie.200700128
Encapsulated Catalysts
RuO Clusters within LTA Zeolite Cages: Consequences of
2
Encapsulation on Catalytic Reactivity and Selectivity**
Bi-Zeng Zhan and Enrique Iglesia*
The consequences of cluster size and spatial constraints are
critical for the design of selective catalysts. Zeolites and other
microporous solids provide essential scaffolds for controlling
the size and accessibility of active structures and the nature of
[
1,2]
spatial constraints around such sites.
The size-dependent
redoxproperties of oi xd e nanoclusters lead to diverse
[
3–5]
catalytic properties for a given composition,
but the
synthesis and stability of such structures remain significant
challenges. Encapsulation in zeolite cages or channels is an
attractive route to control and maintain cluster size, but it has
[
6–9]
been confirmed in only a few cases.
RuO2 clusters of
uniform size (ca. 1 nm) were deposited predominantly in
FAU-X (faujasite-X zeolite, 1.3 nm) during hydrothermal
synthesis and showed high reactivity for aerobic alcohol
[
6]
oxidation. A similar strategy led to RuO clusters within
2
[
7]
MFI (ZSM-5) channels, which after reduction catalyzed 1-
hexene hydrogenation with higher rates than for 2,4,4-
trimethyl-1-pentene, a larger molecule that diffuses slowly
in MFI channels. Coating Pt–Fe/SiO with LTA (zeolite-A)
2
films led to selective oxidation of CO in the presence of larger
[
10]
butane molecules.
Herein, we report the encapsulation of RuO clusters
2
within sodium-LTA cages during hydrothermal synthesis,
producing a caged catalyst that will be referred to as
Figure 1. Near-edge X-ray absorption spectra for Na(RuO )A and
2
Na(RuO )A, and the catalytic consequences of encapsulation.
RuO
2
/SiO
2
(as-synthesized) and after treatment in O
2
(Praxair, UHP)
2
and H (Praxair, UHP). For clarity, traces are shifted vertically.
In situ X-ray absorption spectra (XAS) were used to probe
the ruthenium oxidation state in LTA-encapsulated clusters
after treatment with H or O . Na(RuO )A and RuO /SiO
2
2
2
2
2
2
[
6]
[7]
samples gave identical near-edge spectra (Figure 1), suggest-
ing that RuCl3 precursors were fully converted to RuO2
within FAU and MFI. Treatment of Na(RuO )A in O up
2 2
to 673 K for 1 h did not cause detectable changes in the near-
edge spectra. Na(RuO )A treated in H at 393 K gave near-
during hydrothermal synthesis, as in the case of RuO clusters
2
2
2
edge (Figure 1) and fine-structure (Figure S1 in the Support-
ing Information) spectra similar to those for reduced Ru/SiO
2
[
+]
[
*] Dr. B.-Z. Zhan, Prof. E. Iglesia
Department of Chemical Engineering
University of California at Berkeley
Berkeley, CA 94720 (USA)
[11]
(
treated in H at 573 K) and for Ru metal powders (not
2
shown). Thus, RuO nanoclusters within LTA cages are
2
4
+
0
reduced from Ru to Ru in H below 393 K, consistent with
2
H2 consumption rates measured during this treatment
Fax: (+1)510-642-4778
E-mail: iglesia@berkeley.edu
(Figure 2).
+
Na(RuO )A diffractograms showed only crystalline LTA
[
] Current address:
2
Chevron Research and Technology Company
Richmond, CA 94802 (USA)
(Figure S2 in the Supporting Information). The absence of
RuO lines (ca. 28.4 and 35.58) suggests that RuO domains
2
2
[
**] This work was supported by the Director, Office of Basic Energy
Sciences, Chemical Sciences Division of the US Department of
Energy under Contract DE-AC02-05CH11231. We are grateful for the
electron microscopy data provided by Dr. M. Avalos, L. Rendon, and
F. Ruiz from the Centro de Ciencias de la Materia Condensada,
UNAM, Mexico. We also acknowledge with thanks valuable
technical discussions with S. I. Zones from the Chevron Research
and Technology Company.
are small or amorphous. High-resolution transmission elec-
tron micrographs of Na(RuO )A (Figure S3 in the Supporting
2
Information) confirmed the predominant presence of RuO2
clusters with diameters of about 1 nm, similar to the size of
the cages in LTA (1.1 nm). This result is consistent with
[
6]
[7]
preferential encapsulation, as also found in FAU and MFI
[
12]
and as supported by N2 adsorption uptakes.
Energy-
dispersive X-ray spectra (EDS; Figure S4 in the Supporting
Information) showed uniform Ru distributions among and
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 3697 –3700
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3697

Communications
2
7
Magic angle spinning Al NMR spectra (Figure S7 in the
Supporting Information) for Na-LTA and Na(RuO )A were
2
identical, suggesting that RuO2 clusters did not perturb
framework aluminum centers. The amount of H consumed
2
(
per Ru atom) was 1.9 Æ 0.2 on Na(RuO )A and 2.0 Æ 0.2 on
2
RuO /SiO (Figure 2), consistent with the full reduction of
Ru to Ru , as is also evident from XAS data (Figure 1) and
from previous studies on SiO supports.
2
2
4
+
0
[
11]
2
Table 1 shows methanol oxidative dehydrogenation
ODH) turnover rates (TORODH; per surface Ru atom) and
(
selectivities on Na(RuO )A and RuO /SiO . ODH reaction
2
2
2
rates are limited by CÀH bond activation of chemisorbed
[
11]
methoxide species on RuO domains. This step requires
2
4
+
reduction of Ru centers within Mars–vanKrevelen redox
cycles. Initial HCHO products then react with methanol to
form methoxymethanol (CH OCH OH) intermediates that
3
2
dehydrogenate to methylformate (MF) or condense to form
dimethoxymethane (DMM). HCHO also reacts to form
HCOOH and CO . These sequential pathways require one
x
Figure 2. H uptake on Na(RuO )A and RuO /SiO during reduction
2
2
2
2
oxidative CH OH dehydrogenation (ODH) event for each
3
treatments (H /Ruratios shown above each peak).
2
HCHO, MF, DMM, and CO molecule formed. Methanol
x
ODH turnover rates on RuO /SiO agreed with those in
2
2
À1
[11]
within LTA crystals, but also detected a few crystallites
containing mostly Ru; their size and location suggest that they
are present outside LTA crystals and that they do not
contribute significantly to the surface area of Ru or RuO2
structures. Treatment with dilute (0.005m) oxalic acid
removed some of these larger clusters and decreased their
contribution to catalytic reactions (Figure S5 in the Support-
ing Information).
previous studies (32 h at 393 K), but ODH turnover rates
were significantly higher on Na(RuO )A (121 h ). These
À1
2
reactivity differences reflect the highly reducible nature of
RuO2 clusters in LTA, which are evident from their H2
reduction rates (Figure 2).
Competitive oxidation of methanol (C ) and 2-methyl-1-
1
propanol (C ) was used to probe the selectivity of the
4
encapsulation synthetic protocols and the potential for
reactant-shape selectivity using LTA-encapsulated clusters.
2-Methyl-1-propanol does not diffuse through eight-mem-
The dynamics of RuO reduction were measured from H2
2
consumption rates during thermal treatment. Reduction rates
[
13]
(
(
Figure 2) and Arrhenius plots of incipient reduction rates
Figure S6 in the Supporting Information) showed that RuO2
bered ring windows in LTA. The ratio of C and C alcohol
1 4
ODH rates was much higher on Na(RuO )A (2.1) than on
2
domains in Na(RuO )A were reduced more rapidly than in
RuO /SiO2 (0.33; Table 2), consistent with RuO2 surfaces
2
2
RuO /SiO (ca. 6 nm; 18% dispersion, Table 1). The temper-
residing predominantly within LTA cages and being inacces-
sible to 2-methyl-1-propanol. Mea-
2
2
[
a]
sured C4 alcohol ODH rates
Table 1: Methanol oxidation rates and selectivities on RuO catalysts.
2
À1
(1.7 h ; Table 2) reflect trace con-
Catalysts
Rudisper-
sion [%]
TORMeOH [mol (g
TORODH [mol (g
Production selectivity [%]
HCHO MF DMM CO2
À1 À1
À1 À1
centrations of large RuO clusters
2
atomRusurface
)
h
]
atomRusurface) h ]
on external surfaces of Na(RuO )A.
2
Methanol ODH rates on Na-
RuO /SiO2
18
48
61
32
21
71
4
4
2
(
RuO )A were much lower in com-
2
(
4.3 wt% Ru)
petitive reactions than when meth-
anol was used as the sole alcohol
Na(RuO )A
171
121
35
52
6
7
2
(
4.7 wt% Ru)
À1
reactant (3.5 vs. 117 h , Table 2); 2-
[
(
a] 4 kPa MeOH, 9 kPa O , 1 kPa N , balance He, 393 K. Catalysts treated in 20% O in He
100 cm min ) at 423 K for 2 h before reactions; TORODH is the oxidative dehydrogenation rate per
2
2
2
3
À1
methyl-1-propanol ODH rates,
however, were unaffected by the
presence of methanol co-reactants.
Thus, the ratio of C and C alcohol
surface Ru atom at 20% conversion. No dimethyl ether detected.
1
4
atures required to achieve reduction rates of 4 mol (gatom
ODH rates was much higher with individual reactants (65)
than in competitive reactions (2.1), because 2-methyl-1-
propanol at channel entrances or external surfaces interferes
with methanol diffusion into zeolite channels.
À1 À1
Ru)
h
were 368 K and 396 K for Na(RuO )A and RuO /
2 2
SiO , respectively. These trends are unusual, because smaller
2
oxide domains dispersed on mesoporous supports are typi-
[
5,14,15]
cally harder to reduce than larger domains.
The
Reactant-shape selectivity was also observed in compet-
remarkable reducibility of encapsulated RuO clusters may
itive oxidation of methanol (C ) and isopropyl alcohol (IPA)
2
1
reflect their weak interactions with Na-terminated walls in
LTA cages, in contrast with the strong Ru-O-Si linkages
formed upon grafting RuO onto hydroxylated SiO surfaces.
(Figure S5 in the Supporting Information). The ratio of ODH
rates for methanol and IPA was 1.5 on RuO /SiO and 3.0 on
2
2
Na(RuO )A. This value further increased to 3.7 after treating
2
2
2
3
698 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 3697 –3700

Angewandte
Chemie
Table 2: ODH rates and rate ratios for methanol (C ) and 2-methyl-1-
more, Ru clusters formed by reduction of RuO -LTA hydro-
genate ethene in the presence of organosulfur compounds,
1
2
[
a]
propanol (C ) oxidation on RuO catalysts.
4
2
À1 À1
Catalyst
ODH rates [mol (gatomRusurface
)
h
]
C /C ODH
which poison Ru clusters on SiO . These results represent the
2
1
4
rate ratio
first report of encapsulation within zeolites with eight-
membered ring channels and demonstrate the marked
catalytic consequences of encapsulation for reactivity, reac-
tant selectivity, and protection of active sites against sintering
and poisons.
C1
C4
RuO /SiO2
2.2
3.5
117
6.7
0.33
2.1
65
2
(
4.3 wt% Ru)
Na(RuO )A
1.7
1.8
2
(
4.7 wt% Ru)
[
b]
Na(RuO )A
2
(
4.7 wt% Ru)
Experimental Section
[a] 2 kPa MeOH, 2 kPa 2-methyl-1-propanol, 9 kPa O , 1 kPa N , balance
Encapsulation of RuO within LTA cages was carried out using
2
2
2
3
À1
[19]
He, 413 K. Catalysts treated in 20% O in He (100 cm min ) at 423 K
hydrothermal crystallization.
The hydrothermal synthesis was
2
for 2 h before reactions. [b] MeOH and 2-methyl-1-propanol reactions
carried out separately.
conducted at 373 K for 16 h while stirring (400 rpm; see also the
Supporting Information). The sample contained 4.7 wt% Ru and a Si/
Al atomic ratio of 1.3 after drying at 393 K for 8 h in ambient air.
Silica-supported RuO was prepared by incipient wetness impregna-
2
[
11]
Na(RuO )A with 0.005m aqueous oxalic acid at ambient
tion with [Ru(NO)(NO ) ] using methods reported previously.
2
3 3
temperature, a treatment that partially removes RuO2
clusters on external zeolite surfaces. X-ray photoelectron
spectra (XPS) showed that Ru/Si atomic ratios decreased
from 0.054 to 0.040 after oxalic acid treatment; this value is
much lower than expected from the bulk compositions
RuO /SiO (4.3 wt%) was treated at 673 K for 2 h in flowing dry air
2 2
before characterization and catalytic measurements.
In situ X-ray absorption spectra were recorded at the Stanford
Synchrotron Radiation Laboratory using beamline 6–2. Samples
(
10 mg, 80–120 mesh) were held within a quartz capillary (0.8 mm
[8]
inner diameter, 0.1 mm wall thickness). Spectra were recorded at
ambient temperature and on samples heated from ambient to the
target temperature at 0.167 Ks and held for 1 h. XPS spectra were
(
0.098), suggesting that exposed Ru surfaces lie predomi-
À1
nantly within LTA crystallites.
Encapsulated Ru clusters hydrogenate alkenes in the
presence of organosulfur compounds that typically inhibit
such reactions by competitive coadsorption or sulfida-
measured with a Kratos (Axis HS) instrument. The dispersion of
Ru clusters was determined from H uptakes at 313 K, assuming a 1:1
2
[
11]
H/Rusurface stoichiometry (Autosorb-1, Quantachrome). H reduc-
2
[
16–18]
tion rates of supported RuO domains were measured during thermal
2
tion.
Eight-membered ring windows in LTA prevent
3
À1
treatment with H2 (Praxair, 20% H /Ar; 1.33 cm s ) and heated
from 243 to 873 K at 0.167 Ks . The complete reduction of CuO
2
thiophene access to encapsulated Ru clusters and maintain
active Ru surfaces in the presence of poisons that deactivate
unprotected Ru clusters. Thiophene hydrogenation rates
À1
powder (Sigma-Aldrich, 99.995%) was used for calibration. Catalytic
rates and selectivities were measured in a packed-bed quartz micro-
reactor using samples (60–120 mesh, 0.03–0.3 g) diluted with quartz
powder to prevent temperature gradients. Reactant and product
concentrations were measured by on-line gas chromatography
decreased with time on both Ru/SiO and Na(Ru)A (reduced
2
from RuO ). The hydrogenation products (2,3-dihydrothio-
2
phene, 3,4-dihydrothiophene, tetrahydrothiophene) became
undetectable after 3 h (Figure S8 in the Supporting Informa-
tion), indicating that all Ru sites accessible to thiophene in
(Hewlett-Packard 6890GC).
Received: January 10, 2007
Published online: April 5, 2007
[
16,17]
both catalysts are poisoned by thiophene-derived species.
Ethene and thiophene hydrogenation rates decreased in
parallel on Ru/SiO . Ethene hydrogenation rates (per surface
2
Keywords: cluster compounds · encapsulation ·
heterogeneous catalysis · hydrothermal synthesis · zeolites
À1
À1
.
Ru atom) decreased from 5.3 h to 3.4 h within 3 h. The
residual ethene hydrogenation rates reflect the hydrogenation
reactivity of sulfided Ru species, as also detected on [bis(te-
0
[16]
+[17]
tramethylthiophene)Ru ], [(thiophene)Ru(Cp)]
(Cp =
[
1] W. Hölderich, M. Hesse, F. Näumann, Angew. Chem. 1988, 100,
232 – 251; Angew. Chem. Int. Ed. Engl. 1988, 27, 226 – 246; .
[
18]
C H ), and RuS . In contrast, Na(Ru)A gave stable ethene
hydrogenation rates (ca. 21 h ) after an initial decrease
from 25 to 21 h ), apparently caused by the poisoning of
5
5
x
À1
[2] M. E. Davis, Ind. Eng. Chem. Res. 1991, 30, 1675 – 1683.
[3] A. T. Bell, Science 2003, 299, 1688 – 1691.
À1
(
[
4] D. G. Barton, S. L. Soled, G. D. Meitzner, G. A. Fuentes, E.
Iglesia, J. Catal. 1999, 181, 57 – 72.
Ru sites at external surfaces or channel entrances. Thus, LTA
frameworks provide significant protection from thiophene
adsorption and poisoning in Ru nanoclusters and lead to
ethene hydrogenation rates much higher than on unprotected
Ru clusters.
[
[
5] K. Chen, A. T. Bell, E. Iglesia, J. Catal. 2002, 209, 35 – 42.
6] B.-Z. Zhan, M. A. White, T.-K. Sham, J. A. Pincock, R. J.
Doucet, K. V. R. Rao, K. N. Robertson, T. S. Cameron, J. Am.
Chem. Soc. 2003, 125, 2195 – 2199.
In conclusion, RuO2 nanoclusters (ca. 1 nm diameter)
were predominantly encapsulated within LTA cages during
[7] S. Altwasser, R. Glaser, A. S. Lo, P. H. Liu, K. J. Chao, J.
Weitkamp, Microporous Mesoporous Mater. 2006, 89, 109 – 122.
[
8] H. S. Lacheen, E. Iglesia, J. Am. Chem. Soc. 2006, 128, 15082 –
5083.
9] N. Herron, C. A. Tolman, J. Am. Chem. Soc. 1987, 109, 2837 –
839.
hydrothermal crystallization of LTA. These RuO clusters are
2
1
more easily reduced than comparable clusters on mesoporous
supports and give much higher turnover rates for reactions
involving reduction–oxidation cycles, such as methanol oxi-
dation. RuO2 clusters encapsulated on Na-LTA oxidize
methanol preferentially over 2-methyl-1-propanol. Further-
[
2
[
10] P. Collier, S. Golunski, C. Malde, J. Breen, R. Burch, J. Am.
Chem. Soc. 2003, 125, 12414 – 12415.
[11] H. C. Liu, E. Iglesia, J. Phys. Chem. B 2005, 109, 2155 – 2163.
Angew. Chem. Int. Ed. 2007, 46, 3697 –3700
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 3699

Communications
[
12] Uptake (Autosorb-6, Quantachrome) of N at its boiling point
[13] D. W. Breck in Adsorption by Dehydrated Zeolite Crystals,
Zeolite Molecular Sieves (Ed.: D. W. Breck), Robert E. Krieger
Publishing Company, Florida, 1984, Chapter 8.
[14] C. A. Müller, M. Maciejewski, R. A. Koeppel, A. Baiker, J.
Catal. 1997, 166, 36 – 43.
2
2
À1
gave a BET surface area of 11.2 m g for Na(RuO )A. This
2
value reflects only the external area of zeolite crystals (LTA
[
13]
micropores are inaccessible to N2 under these conditions ),
consistent with the size of LTA crystals (ca. 500 nm by scanning
electron microscopy), which would have a surface area of
[15] Y. Xu, W. A. Shelton, W. F. Schneider, J. Phys. Chem. B 2006,
2
À1
110, 16591 – 16599.
1
0 m g (for cubic crystals). The RuO2 clusters (ca. 2 nm
[
16] S. Luo, T. B. Rauchfuss, S. R. Wilson, Organometallics 1992, 11,
3497 – 3499.
diameter; 48% dispersion, Table 1) alone would give a surface
2
À1
area of about 50 m g if they were present at external surfaces
[17] N. N. Sauer, R. J. Angelici, Organometallics 1987, 6, 1146 – 1150.
[18] B. C. Wiegand, C. M. Friend, Chem. Rev. 1992, 92, 491 – 504.
[19] R. W. Thompson, K. C. Franklin in Verified Synthesis of Zeolitic
Materials, 2nd revised ed. (Ed.: H. Robson), Elsevier, Amster-
dam, 2001, p. 179.
and accessible to N . Thus, BET data are consistent with
2
preferential encapsulation of RuO clusters within LTA cages in
2
these samples.
3
700 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 3697 –3700