Mark E. Davis Research Group

California Institute of Technology

Zeolite Synthesis

Cooperative Catalysis

Zeolites are ideal catalytic materials because they have nanostructure on the same scale as the molecules that react within their pores. Their nanostructure is a function of the conditions under which they are created, and because of the molecular scale of the catalytic surfaces, their macroscopic properties are thus a function of those same conditions.

Zeolites are microporous crystalline oxides with a high surface to volume ratio. Strictly speaking, they are aluminosilicates of form Mx*(Si1-x+Alx)O2*yH2O, but the Davis group is also interested in many zeolite-like materials that are also microporous and have important catalytic properties, such as aluminophosphates, zincosilicates, and titanosilicates.

Though the catalytic properties of zeolites are of central interest to the Davis group, they are primarily used (by mass) as adsorbents and ion-exchange materials. They were first used for organic synthesis in the 1960s for catalytic cracking of petroleum. ZSM-5, first synthesized in 1967, was applied to numerous refining and petrochemical processes with great success.

Zeolites are actually natural products. Not all naturally occuring zeolites have been recreated in the laboratory, and conversely, many zeolites have been synthesized that do not occur in nature.

Making zeolites is a delicate art. Typically, the experimenter will combine non-molecular building blocks like alumina and silica with some alkali metal hydroxide and water, then incubate the mixture in an enclosed environment at 50-200 degree Celcius for a certain period of time. Organic chemicals called Structure Directing Agents (SDAs) are often added to steer the self-assembly of the zeolite in a desired direction.

Zeolite synthesis is not similar to synthetic organic chemistry. The zeolite is assembled through a series of spontaneous molecular recognition events that cannot be understood in the same dot-and-arrow schema that synthetic organic schemes are constructed from. In fact, this process is not very well understood at all.

The final structure is a product of synthetic conditions and the post-synthetic treatment. One of the most interesting factors the experimenter controls is the hydrophobicity and structure of the SDA. Though this relationship is not facile, several lines of research in the Davis group have turned up explanations and successful predictions of final nanostructure as a function of SDA choice. See scanning electron micrographs of zeolites synthesized by the Davis group (below).

You can read more about zeolites and other ordered porous materials in recent reviews:

Corma A; Davis, ME; (2004) ChemPhysChem 5 304-313.

Davis, ME; (2002) Nature 417 813-821

Before the synthesis of VPI-5, the largest zeolites had pore sizes below 10 angstroms. The pores of the zeolites contained 12 oxygen atoms or fewer. In 1987, the Davis group synthesized the aluminophosphate VPI-5, an 18-ring microporous material with a 13 angstrom pore diameter. This was the first of a series of extra large pore zeolite-like materials.

In 1997 at the California Institute of Technology, the Davis group synthesized CIT-5, an extra-large pore 14-ring zeolite with a pore diameter of just over 10 angstroms--the second extra-large pore zeolite discovered. This represented an important development in the field because CIT-5 has the high thermal stability of a high-silica zeolite, but large pore size like VPI-5 and its successors. And unlike UTD-1, the first extra-large pore zeolite, CIT-5 does not require an organometallic SDA and the lengthy post-synthetic workup that follows. CIT-5 is structure directed by N(16)-methylsparteinium with lithium cations:

The following nanostructure results:

Why are extra-large pore zeolites important?

The range of reactions over which a zeolite can apply shape selective pressure is a function of the zeolites pore size and structure. Zeolites with pore sizes between ten and twenty angstroms would impact fine chemicals, pharmecuticals, and life sciences.

Davis, ME; Saldarriaga, C; Montes, C; Garces, J; Crowder, C; (1988) Nature 331, 698.

Freyhardt, CC; Tsapatsis, M; Lobo, RF; Balkus, KJ; Davis, ME; (1996) Nature 381, 295.

Wagner, P; Yoshikawa, M; Lovallo, M; Tsuji, K; Taspatsis, M; Davis, ME; (1997) Chem. Commun. 2179.

Nery JG ; Hwang SJ; Davis, ME; (2002) Micro. Meso. Mat. 52 19-28.

Ogino, I ; Davis, ME; (2004) Micro. Meso. Mat. 67 67-78.

Current methods for zeolite synthesis use organic molecules as their SDAs. Once the synthesis is complete a high temperature combustion step is required to remove these SDAs from the zeolite framework. This process both destroys the SDA, a high cost material, and the energy release and the water produced in the combustion can damage the inorganic structure.

In 2003, the Davis group reported a methodology for the synthesis of zeolite ZSM-5 using a SDA with an attached cyclic ketal group that can be fragmented inside the zeolite. The fragmentation allows the parts of the SDA to be removed without a high combustion step in the treatment. Also, once removed the fragments can be recombined to form the original SDA, hence the “recyclable SDA”. The concept of the methodology is shown here.

In the diagram above the methodology follows 5 basic steps:

• Assemble an SDA from at least two components using covalent bonds or non-covalent interactions that will survive the zeolite formation conditions (e.g. high pH).
• Form the zeolite, thus trapping the SDA inside the framework.
• Cleave the SDA components in ways as to not destroy the framework. (e.g. a low pH wash)
• Remove all the fragments.
• Reassemble the SDA.

As noted the “recyclable SDA” methodology provides a way to produce zeolite with out destroying the high cost SDA. Also removing the combustion step is beneficial for forming various zeolite products such as molecular sieve membranes that can be damaged by the mechanical stresses of thermal expansion or low dielectric components that need air to fill the microporous space to achieve the desired properties.

The Davis group is currently using the methodology to prepare larger SDAs and testing it with the synthesis of other zeolites.

Lee H ; Zones SI; Davis, ME; (2003) Nature, 425 385-387.

Hybrid inorganic–organic materials comprising organic functional groups tethered from silica surfaces are versatile, heterogeneous catalysts. Recent advances have led to the preparation of silica materials containing multiple, different functional groups that can show cooperative catalysis; that is, these functional groups can act together to provide catalytic activity and selectivity superior to what can be obtained from either monofunctional materials or homogeneous catalysts.

A number of lessons from the basic principles of enzyme catalysis can be applied to preparing more efficient synthetic catalysts. One important way that enzymes accelerate chemicals reactions is through cooperative interactions between precisely positioned reactive groups in the active site. With functional groups (metal centers, nucleophiles, acids, bases, hydrogen bond donors, hydrogen bond accepters) positioned at fixed distances to one another in the active site, these groups are capable of interacting through electrostatic, hydrogen bonding, and covalent interactions to influence their reactivity. Through these types of interactions, enzymes are capable of significantly accelerating reactions (rate enhancement of many orders of magnitude) as well imparting dramatic effects on selectivity. By adapting these approaches of spatially positioning functional groups and modifying the surface of the catalyst, it may be possible to further improve upon current heterogeneous catalysts.

A catalytic reaction where bifunctional cooperativity has been long known is the synthesis of bisphenol A from acetone and phenol (below). This reaction can be catalyzed by strong acids alone, but the activity and selectivity can both be increased dramatically by adding a thiol (either homogeneous or heterogeneous) as a co-catalyst.

Davis group incorporated arylsulfonic acid and alkylthiol groups into mesoporous silica (SBA-15) by a direct synthesis method. The resulting randomly-distributed acid–thiol catalysts exhibited far greater activity and selectivity than materials containing only the acid. Furthermore, a physical mixture of separate acid-containing and thiol-containing silicas gave only modest results, suggesting that the acid and thiol groups must be in proximity to one another for enhanced catalytic activity. The origin of this cooperativity is sequential activation of the acetone first by protonation, then by thiol attack, forming a highly electrophilic sulfonium intermediate (below).

Selected Recent Publication:

R. K. Zeidan, V. Dufaud and M. E. Davis, J. Catal., 2006, 239, 299–306.

Materials containing functionalities that are incompatible with one another in solution have been reported by the Davis group. In these materials, incompatible functional groups were immobilized together on a silica surface and interesting chemical reactivity was observed. Therefore, materials functionalized with both acid and base groups that would be incompatible if not tethered to a surface have been reported. These materials were prepared through direct synthesis and the two incompatible functional groups were randomly distributed throughout the silica material. In this report, strongly acidic aryl sulfonic acid groups were simultaneously incorporated into SBA-15 along with primary amine groups, generating a bifunctional acid–base material that was a good catalyst in the aldol condensation reaction between acetone and p-nitrobenzaldehyde.

In this case, the bifunctional material significantly outperformed materials functionalized individually with either group, as well as a physical mixture of the independently functionalized materials, further illustrating the cooperative effect of these groups that must be near enough to one another to be capable of interacting (below). This material is particularly unique since when these two groups are used together homogeneously, no reaction is observed. Furthermore, adding homogeneous acid to a catalyst containing only the heterogeneous base rendered the catalyst inactive. Also, a dramatic solvent effect was observed wherein the reactivity of the catalysts was found to vary greatly with the solvent of the reaction as protic solvents presumably promote proton transfer and neutralization of the incompatible acid and base groups.

Selected Recent Publication:

R. K. Zeidan, S. J. Hwang and M. E. Davis, Angew. Chem., Int. Ed., 2006, 45, 6332–6335.

In a follow-up to this report, we replaced the sulfonic acid groups with weaker acidic groups (phosphonic, carboxylic), again distributed on a surface with primary amines (below). The activities of these catalysts were found to increase as the strength of the acid component decreased. We attributed this trend to changes in the proton-transfer interactions between the acid and amine groups, where a weaker acid is more easily reprotonated than a stronger acid. The cooperative effect of these materials was profound: carboxylic acid alone was not strong enough to give measurable yield in the aldol reaction, and primary amine alone gave only 33% conversion. However, when both of these functionalities were present in close proximity, the reaction went to completion.

Selected Recent Publication:

R. K. Zeidan and M. E. Davis, J. Catal., 2007, 247, 379–382.

Cooperative surface catalysis relies on the two functional groups being close enough to each other on the surface to interact with each other or with the reacting molecules. Thus, one would anticipate that the catalytic activity of such materials would depend on the distances between catalytic surface sites.

1) Imprinting

The traditional way to position multiple functional groups in a solid involves imprinting, in which a template molecule is used to guide the organization of the relevant functionalities. Multiple-point covalent imprinting has been employed to form pairs or triplets of identical functional groups within a silica matrix. The catalysts reported by Davis group illustrate this approach. Different molecular templates were prepared incorporating carbamate-protected amine groups and triethoxysilyl groups. Using sol–gel polymerization, these templates were then imprinted into bulk, amorphous silica and in a second step the molecular template was removed. The templates were removed by reaction with trimethylsilyl iodide, and the resulting materials contained one, two or three amine functionalities spatially positioned in the pocket vacated by the imprint (see below)

Selected Publication:

A. Katz and M. E. Davis, Nature, 2000, 403, 286.

(2) Site pairing

SBA-15 materials functionalized with sulfonic acid sites grouped into pairs have been investigated. These materials were obtained by incorporating a disulfide bridging group onto the surface, followed by reduction of the disulfide to pairs of thiols and subsequent oxidation with peroxide to generate the acid pairs. The activity of the acid catalysts for bisphenol A synthesis was about two-fold higher when the acids were grouped into pairs; however, this increased activity is more likely attributable to the presence of residual thiol groups or other partially-oxidized species, and not to the effect of pairing.

Davis group developed a family of catalysts where alkylsulfonic acid sites were paired with thiol groups on a silica surface and used in the bisphenol A synthesis. Sultone rings were tethered to the surface of SBA-15 and were then opened by reaction with various nucleophiles, such as hydrosulfide ion or the monoanion of a dithiol. The ring-opening reaction served both to generate the sulfonic acid group and also to tether the second functionality to the same site. In this method, the acid/thiol ratio is fixed at 1 but the acid–thiol distance can be tuned by changing the nucleophile (see below). It has been reported how both the activity and selectivity of the acid–thiol catalysts are highly dependent on the distance between the two groups, and that the best catalyst is obtained when the acid and thiol groups are in very close proximity.

Selected Recent Publications:

V. Dufaud and M. E. Davis, J. Am. Chem. Soc., 2003, 125, 9403–9413.

E. L. Margelefsky, V. Dufaud, R. K. Zeidan and M. E. Davis, J. Am. Chem. Soc., 2007, 129, 13691–13697.

An ordered mesoporous silica catalyst containing arylsulfonic acid and alkylthiol groups organized in pairs was synthesized by grafting a designed bis-silane precursor onto SBA-15, followed by linker cleavage, followed by the synthesis of the randomly distributed arylsulfonic acid/tiol (bottom).

In the bisphenol A reaction, the random and paired arylsulfonic acid/thiol catalysts exhibit similar catalytic performance, whereas in the slower bisphenol Z reaction, the paired catalyst outperforms the randomly distributed catalyst. Compared to catalysts containing weaker alkylsulfonic acid sites, the arylsulfonic acid catalysts are more active while maintaining similar selectivities. Thus, it appears that the common tradeoff between activity and selectivity can be circumvented by appropriate catalyst design.

Selected Recent Publications:

E. L. Margelefsky, A. Bendjeriou, R. K. Zeidan, V. Dufaud and M. E. Davis, J. Am. Chem. Soc., 2008, 130, 13442.