451022 Strategies for the Encapsulation and Stabilization of Monodisperse Au Clusters within Zeolites

Thursday, November 17, 2016: 8:30 AM
Imperial B (Hilton San Francisco Union Square)
Trenton Otto, Chemical and Biomolecular Engineering, University of California, Berkeley, Berkeley, CA, Stacey I. Zones, Chevron Energy and Technology Company, Richmond, CA and Enrique Iglesia, Department of Chemical and Biomolecular Engineering, University of California, Berkeley, Berkeley, CA

The encapsulation of nanoparticles within microporous solids can improve the inherent instability of highly dispersed nanoclusters by sequestering them within voids that prevent their coalescence with other clusters, while also exploiting spatial constraints to restrict the size to which they can grow [1,2]. Confinement within such voids also precludes access by certain reactants or poisons to active sites present within intracrystalline voids, retains undesired products until they convert to smaller species that can egress by diffusion, or stabilizes specific transition states; in all cases, these enabling strategies are dictated by the size of the voids or channels in a specific microporous framework [2]. The encapsulation of such clusters within small-pore and medium-pore zeolites typically cannot be achieved through post-synthetic exchange, because solvated metal precursors are too large to enter the apertures of these zeolites [3].

This study describes successful strategies and guiding principles for the synthesis of small (1-2 nm) and monodisperse Au clusters protected against coalescence and poisoning by their uniform dispersion throughout the void space of LTA (0.42 nm apertures) and MFI (0.55 nm apertures) zeolites. The synthesis protocols involve hydrothermal zeolite crystallization around cationic Au3+ precursors. These precursors are stabilized by mercaptosilane ligands, which bind with the metal cations via their terminal sulfur atoms and protect against premature reduction or precipitation during the nucleation and growth of the zeolite frameworks. The silane groups in these ligands simultaneously promote the formation of siloxane bridges with nucleating silicate oligomers, thus enforcing the uniform dispersion of precursors through each crystal and the ultimate size homogeneity of the encapsulated clusters after thermal treatments required for ligand removal. UV-visible spectra show that Au3+ forms Au0 clusters in O2 or H2 over a narrow temperature range that sets the dynamics of nucleation and growth and thus the mean size and uniformity of the clusters. H2 treatments at the mildest temperatures within this range lead to small clusters uniform in size (1.0-2.0 nm) that are evident in electron micrographs, and to clean and accessible metal surfaces as demonstrated by the infrared spectra of chemisorbed CO. Their unprecedented size and monodispersity are preserved after oxidative treatments (>773 K) that significantly sinter Au clusters on mesoporous supports. Oxidative dehydrogenation rates of small (ethanol) and large (isobutanol) alkanols and the poisoning of unprotected clusters by organosulfur titrants show that >90% of the Au surfaces reside within intracrystalline LTA and MFI voids. This study illustrates how confinement favors small, uniquely stable, and monodisperse clusters, even for Au, a metal prone to cluster growth at conditions often required for its catalytic use.

X-Ray diffractograms of the synthesized zeolites confirmed the formation of the intended frameworks (LTA, MFI) and the absence of large (>10 nm) metal clusters. Elemental analysis showed that the incorporation of the ligand-stabilized metal precursors into the solids recovered was essentially complete and gave metal contents consistent with those predicted from synthesis reagents. The intensity of the surface plasmon resonance band in the UV-visible spectra of these zeolites was monitored during heating in O2 or H2 to track the formation of Au0 clusters from the Au3+ precursors occluded within the intracrystalline voids. The intensity of this band serves as a diagnostic of Au0 clusters larger than 2 nm, the diameter at which they incipiently exhibit plasmon resonance [4]. The H2 treatments led to a sharp increase in the intensity of the plasmon resonance band between 540-623 K, reflecting the reduction of Au3+ by H2 and the subsequent nucleation of Au0 clusters throughout the zeolites. Transmission electron micrographs (TEM) showed that H2 reduction treatments at 573 K form small Au clusters (LTA: 1.3 nm, MFI: 2.0 nm) narrowly distributed in size and embedded uniformly throughout zeolite crystals. These zeolite-encapsulated clusters retained their sizes even after subsequent treatments up to 773-823 K in O2 or H2. Such stability reflects the constraints imposed on cluster growth by kinetic hurdles brought forth by the large number and small size of intervening zeolite apertures that separate nucleation points during the decomposition processes; we also surmise that a limiting size is imposed by thermodynamic barriers originating from the energy required to disrupt the zeolite framework to accommodate clusters larger than a few cages or channel intersections in LTA or MFI, respectively. Au clusters dispersed on mesoporous supports such as TiO2 or SiO2, in contrast, form larger clusters (>10 nm) that are heterogeneously distributed in size after similar or even milder thermal treatments [5,6]. The zeolite encapsulation technique described here thus allows the systematic study of Au clusters at reaction conditions that strongly favor the sintering of Au particles into larger and much less reactive agglomerates.

Infrared bands of CO chemisorbed on Au clusters in the zeolite samples were examined to assess the cleanliness and accessibility of their metal surfaces. This infrared technique was used as an alternative to chemisorptive titrations, which are challenging to implement with Au clusters because of high dissociation activation barriers for common titrants, such as H2 or O2 [7]. The integrated intensities of Au-CO infrared bands on Au-MFI and Au-LTA samples, normalized by the number of surface Au atoms (as estimated by TEM), were compared to those measured on Au/SiO2 samples prepared with methods known to fully remove contaminating reagents and form clean metal surfaces [8]. Normalized integrated Au-CO band intensities on Au-LTA and Au-MFI samples prepared using mercaptosilane ligands were essentially identical to those on Au/SiO2 at all CO pressures, consistent with Au surfaces that are fully accessible and free of synthetic debris.

The selective placement of the Au clusters within the LTA zeolite (0.42 nm apertures) was confirmed by measuring oxidative dehydrogenation (ODH) turnover rates of ethanol (0.40 nm kinetic diameter, [3]) and isobutanol (0.55 nm kinetic diameter, [3]). The windows in LTA allow ethanol diffusion into the void space but exclude isobutanol; thus, comparison of ODH turnover rates for these small and large alkanols on unrestricted surfaces (e.g. Au/SiO2) and on Au-LTA can be used to confirm that the exposed Au atoms reside predominantly within the LTA void space. Ethanol ODH rates on Au-LTA samples were also measured in the presence of thiophene, an organosulfur moiety that irreversibly titrates Au metal surfaces [3]. Thiophene (0.46 nm, [9]) is too large to enter the apertures of LTA; consequently, the extent to which thiophene decreases ethanol ODH rates provides a rigorous measure of the fraction of the Au surfaces accessible to thiophene and thus unprotected by intracrystalline voids. MFI channels (0.55 nm apertures) cannot exclude isobutanol or thiophene from intracrystalline voids. Consequently, encapsulation in MFI was instead probed by determining the fraction of Au cluster surfaces that are protected by titration with an organosulfur molecule (dibenzothiophene; DBT, 0.9 nm, [9]) that cannot enter MFI crystals.

The ratio of the ethanol to isobutanol ODH turnover rate on Au-LTA (133) was much larger than that on Au/SiO2 (1.5), showing that the LTA zeolite framework indeed precludes access of isobutanol to encapsulated Au surfaces. Ethanol ODH rates on Au clusters supported by mesoporous SiO2 were almost fully suppressed by exposure to thiophene or DBT, consistent with unrestricted access of the sulfur compounds to these unconstrained clusters. LTA-encapsulated Au, in contrast, retained >90% of its initial ethanol ODH rate following exposure to thiophene. Analogous experiments that exposed Au clusters within MFI to DBT also showed that >95% of the initial ODH activity was preserved in the presence of the sulfur titrant. This evident size selectivity confirms that the Au clusters reside predominantly within the LTA or MFI zeolite voids.

This work provides a general synthetic procedure to encapsulate Au clusters within the voids of small-pore and medium-pore zeolites. The shape selectivity shown is yet another demonstration of the reactant selection properties that have made zeolites such ubiquitously useful catalysts, while also demonstrating how confinement can lead to highly dispersed and uniquely stable Au clusters that are uniform in size, making such clusters useful at temperatures previously inaccessible in their synthesis and catalytic use.

The authors acknowledge the Chevron Energy Technology Company for financial support.


[1] Weisz, P.; Frilette, V.; Maatman, R.; Mower, E. J. Catal. 1962, 1, 307.

[2] Wu, J. C. S.; Goodwin, J. G.; Davis, M. J. Catal. 1990, 125, 488.

[3] Wu, Z.; Goel, S.; Choi, M.; Iglesia, E. J. Catal. 2014, 311, 458.

[4] Peng, S.; McMahon, J. M.; Schatz, G. C.; Gray, S. K.; Sun, Y. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 14530.

[5] Otto, T.; Zones, S. I.; Iglesia, E. J. Catal., in press.

[6] Veith, G. M.; Lupini, A. R.; Pennycook, S. J.; Mullins, D. R.; Schwartz, V.; Bridges, C. A.; Dudney, N. J. J. Catal. 2009, 262, 92.

[7] Carabineiro, A. A. C.; Nieuwenhuys, B. E.; Gold Bull. 2009, 42, 288.

[8] Zhu, H.; Ma, Z.; Clark, J. C.; Pan, Z.; Overbury, S. H.; Dai, S. Appl. Catal., A 2007, 326, 89.

[9] Van de Voorde, B.; Hezinova, M.; Lannoeye, J.; Vandekerkhove, A.; Marszalek, B.; Gil, B.; Beurroies, I.; Nachtiqall, P.; De Vos, D. Phys. Chem. Chem. Phys. 2015, 17, 10759.

Extended Abstract: File Not Uploaded