Ordered mesoporous and macroporous inorganic films and membranes

Ordered mesoporous and macroporous inorganic films and membranes

(Parte 1 de 8)

Journal of Membrane Science 235 (2004) 53–72

Ordered mesoporous and macroporous inorganic films and membranes

Department of Chemical and Materials Engineering, University of Cincinnati, Cincinnati, OH 45221-0012, USA Received 21 April 2003; received in revised form 29 December 2003; accepted 12 January 2004


This article reviews synthesis, structures and emerging applications of ordered inorganic meso- and macroporous films and membranes.

Thin films or membranes of these groups of porous materials can be prepared by various techniques commonly used for film formation from liquid solution, including solvent evaporation, in situ growth from solution and other alternative methods. Although these techniques allow control over the thickness and pore structures in thin films or membranes, they currently offer limited control over the pore alignment relative to the substrate surface. The ordered mesoporous and macroporous thin films and membranes are highly promising for a range of potential applications in separations, chemical sensing, heterogenous catalysis, microelectronics and photonics as insulating layers of low dielectric constant and photonic bandgap materials. Furthermore, functionalization of the internal pore surfaces in these films and deposition of functional nanoparticles within the pores offer numerous new possibilities for molecular engineering of catalytic, photonic, and magnetic materials for advanced nanotechnological applications based on quantum confinement effect. © 2004 Elsevier B.V. All rights reserved.

Keywords: Mesoporous; Macroporous; Pore structures

1. Introduction

According to IUPAC [1], mesoporous and macroporous films or membranes refer to materials with pore diameters in the 2–50 and >50nm ranges, respectively. The term “film” is used to describe a material in a specific planar geometry, while the membrane is defined as a thin physical barrier allowing selective transport of mass species. Thin films can be formed on any support (substrate), whereas membranes must be prepared on a porous support (or substrate), which allows transport of mass species. Mesoporous and macroporous films and membranes have been studied extensively in the past decades and many of them have found commercial applications [2,3]. Examples of these applications include a large number of inorganic membrane microfiltration and ultrafiltration processes being used in industry for beverage and drinking water purification and wastewater treatment. The majority of these mesoporous and macroporous

E-mail addresses: vguliant@alpha.che.uc.edu (V.V. Guliants), jlin@alpha.che.uc.edu (Y.S. Lin). 1 Co-corresponding author.

films and membranes were made by compacting nano- or micron-sized dense crystallites (such as -o r -alumina crystals). Their pore structure is defined by the crystallite size and the manner in which they are packed. The pores of these films or membranes are not arranged in an ordered fashion as opposed to those in ordered mesoporous and macroporous films or membranes.

The first ordered mesoporous materials, belonging to aluminosilicate molecular sieves called M41S, were reported by researchers at Mobil over a decade ago [4]. The mesoporous aluminosilicate and other metal oxide materials with well-defined pore sizes up to 33nm break past the pore-size constraint (<1.5nm) of microporous zeolites. The extremely high surface areas (>1000m2/g) and precise tuning of pore sizes are among the many desirable properties that have made such materials the focus of great attention. The M41S materials also ushered in a new approach in materials synthesis where, instead of the use of single molecules as templating agents (as in the case of zeolites), self-assembled molecular aggregates or supramolecular assemblies are employed as the structure-directing agents. Since the initial discovery of the M41S materials, many other ordered pore structures corresponding to different surfactant assemblies have been synthesized. Various aspects of these ordered

0376-7388/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2004.01.019 mesoporous materials have been reviewed by Brinker (recent advances in porous inorganic materials) [5], Vartuli et al. (synthesis of the M41S family) [6], Stucky et al. (biomimetic synthesis of mesoporous materials) [7], Raman et al. (porous silicates templated by surfactants and organo-silicate precursors) [8], Antonelli and Ying (mesoporous transition metal oxide (TMS) family) [9], Behrens (mesoporous transition metal oxides) [10], Zhao et al. (aluminosilicate MCM-41) [1], and Sayari (catalytic applications of MCM-41 and HMS) [12].

The most common ordered mesoporous structure is the two-dimensional hexagonal phase with the P6mm symmetry, consisting of close-packed hexagonal arrays of cylindrical surfactant micelles. Lamellar silicate-surfactant phases are also obtained, which are not stable to the removal of the surfactant which leads to a collapse of the silica layers. Various cubic phases have also been reported. The bicontinuous cubic gyroid phase with the Ia3d symmetry is found in alkaline-catalyzed syntheses. This phase with its network of interconnected pores is much more attractive than the two-dimensional hexagonal phase for applications requiring diffusion of species into and out of the pore network. The Pm3n phase composed of spherical micelles in a cubic close-packed arrangement is found in both acid- and alkaline-catalyzed preparations.

The initial work of the Mobil researchers involved only alkaline-catalyzed preparations and materials reported in these papers are referred to as MCM-41 (P6mm two-dimensional hexagonal phase), MCM-48 (Ia3d cubic phase) and MCM-50 (lamellar phase). Two-dimensional hexagonal mesoporous materials prepared by Inagaki and co-workers [13] from surfactant-intercalated kanemite are referred to as FSM-16. Subsequent preparation of materials in acidic medium led to ordered mesoporous materials with similar symmetries, but distinct wall properties: SBA-1

(Pm3n, cubic phase), SBA-2 (P63/mmc, three-dimensional hexagonally packed spherical micelle phase) and SBA-3

(P6mm, two-dimensional hexagonally packed cylinders) [14]. Other surfactant-templated phases include disordered materials with a worm-like network of channels: KIT-1 [15], MSU-1 [16], a well-defined intergrown cubic hexagonal phase STAC-1 [17] and ordered non-ionic alkyl-EOx-templated materials, e.g. cubic SBA-1 (Pm-3m), three-dimensional hexagonal SBA-12 (P63/mmc) [18]. Using the same methodology, triblock-copolymer-templated materials with much larger pores have also been synthesized in acidic systems, e.g. two-dimensional hexagonal SBA-15 (P6mm) [19] and SBA-16 cubic cage structure (Im-3m) [18].

The technological potential of mesoporous materials for chemical separations and heterogeneous catalysis was recognized at the time of their discovery [20]. An important feature of mesoporous materials is the recently discovered ability to form thin films. While the early work focused on the synthesis of bulk materials, methods for preparation of mesoporous materials in a thin film configuration at a thickness range of tens of nanometers to micrometers were reported in the last few years. The motivation for synthesis of mesoporous thin films originates from the appreciation of their technological potential as membranes, sensors [20], heterogeneous catalysts [20], and insulating layers of low dielectric constant for microelectronics [21]. These applications require the ordered material in the form of thin film. Mesoporous thin films or membranes are characterized by bulk properties, such as symmetry, pore diameter, surface area and stability, and film-related parameters, such as pore orientation, film thickness, continuity and surface roughness.

Although a number of reviews have appeared on bulk ordered mesoporous materials, only one review on ordered mesoporous films has emerged to date with the emphasis on self-standing, unsupported films prepared at the liquid and air interface [2] despite increasing number of published studies on thin film ordered mesoporous materials and their technological importance. Ordered macroporous materials templated by ordered arrays of colloidal spheres and other shapes have also recently received interest from scientific community [23]. The present review is focused on solid substrate-supported thin films of the ordered mesoporous and macroporous materials that are promising for applications in chemical sensing, separations and catalysis. We will first discuss the major preparation routes and mechanisms of formation of mesoporous thin films and then review potential applications of these materials in membrane separations, heterogeneous catalysis and nanotechnology. The subjects of mesoporous and macroporous films will be discussed in separate sections because of the fundamentally different formation mechanisms and resulting pore structures.

2. Mechanisms of self-assembly 2.1. Ordered mesoporous materials

2.1.1. Liquid crystal templating mechanism

The original M41S family of mesoporous molecular sieves is synthesized by reacting a silica source (e.g. tetraethylorthosilicate (TEOS), Ludox, fumed silica, sodium silicate), an alkyltrimethylammonium halide surfactant (e.g. cetyltrimethylammonium bromide (CTAB)), a base (e.g. sodium hydroxide or tetramethylammonium hydroxide (TMAOH)), and water at ≥100◦C for 24–144h. Aluminosilicate M41S is synthesized by addition of an aluminum source to the synthesis mixture. The as-synthesized product contains occluded organic surfactant, which is removed by calcination at ∼500◦C in air to yield an ordered mesoporous material.

A “liquid crystal templating” (LCT) mechanism of formation was proposed by the Mobil researchers (Fig. 1), based on the similarity between liquid crystalline surfactant assemblies (i.e. lyotropic phases) and M41S [4]. The common traits were the mesostructure dependence on the hydrocarbon chain length of the surfactant tail group [24,26], the effect of surfactant concentration, and the influence of organic

Fig. 1. Formation of mesoporous structures by liquid crystal templating mechanism (redrawn from [2]): (path a) via pre-existing surfactant liquid crystal phase; (path b) via the formation of silica-coated surfactant species that (path c) either form micelles which agglomerate to form ordered and disordered arrays or (path d) a lamellar phase which undergoes a phase transition into the final hexagonal phase. Mesoporous phase exhibiting a hexagonal array of ordered pores is obtained after calcination to remove the surfactant.

swelling agents. Two mechanistic pathways were proposed by the Mobil researchers to explain formation of MCM-41 containing two-dimensional hexagonal arrays of cylindrical mesopores: (1) via condensation and cross-linking of the inorganic species at the interface with a pre-existing hexagonal lyotropic liquid crystal (LC) phase; and (2) via ordering of the surfactant molecules into the two-dimensional hexagonal mesophase mediated by the inorganic species and followed by their condensation and cross-linking.

In the both pathways, the inorganic species, which are negatively charged at the high synthesis pH preferentially interact with the positively charged alkyl ammonium head groups of the surfactants and condense into a solid, continuous framework. The resulting organic-inorganic mesostructure could be alternatively viewed as a hexagonal array of surfactant micellar rods embedded in a silica matrix. The surfactant removal produces the open mesoporous MCM-41 framework. Typical TEM images of mesoporous two-dimensional hexagonal mesostructures are shown in Fig. 2 [25]. These mesophases with pore sizes >2.5nm generally display type IV nitrogen adsorption–desorption isotherms at 77K characteristic of mesoporous materials with ordered unimodal pore size distribution shown in Fig. 3. M41S type materials with smaller mesopores (1.7–2.5nm) exhibit type I isotherms characteristic of microporous materials. It is now known that pathway 1 did not take place in the Mobil syntheses because the surfactant concentrations used were far below the critical micelle concentration (CMC) required for the hexagonal LC formation [26]. This mechanism was recently shown to be valid under different synthesis conditions (see Section 2.2).

The second mechanistic pathway of LCT was proposed as a cooperative self-assembly of the alkyl ammonium surfactant and the silicate precursor species below the CMC. It is well established that no preformed LC phase is necessary for the MCM-41 formation, but to date the mechanistic details of MCM-41 synthesis have not yet been fully agreed upon. Several mechanistic models have been advanced which share the basic idea that the silicate species promoted LC phase formation below the CMC.

2.1.2. Inorganic species-driven self-assembly

Davis and co-workers [27] found that the hexagonal LC phase did not develop during the MCM-41 synthesis, based

Fig. 2. High resolution TEM image of two-dimensional hexagonal mesoporous SBA-15 silica in the direction along (A) and perpendicular (B) to the c-axis [25].

on in situ 14N NMR spectroscopy. They proposed that, under the synthesis conditions reported by Mobil, the formation of MCM-41 began with the deposition of two to three monolayers of silicate precursor onto isolated surfactant micellar

BJH Adsorption dV/dlog(D) Pore Volume

Pore Volume (cm3/g)

Fig. 3. Pore size distribution of a typical two-dimensional hexagonal SBA-15 mesoporous silica [25].

rods. The silicate-encapsulated rods were randomly ordered, eventually packing into a hexagonal mesostructure. Heating and aging completed the condensation of the silicates into the as-synthesized MCM-41 mesostructure.

Instead of the formation of silicate-covered micellar rods,

Steel et al. [28] postulated that surfactant molecules assembled directly into the hexagonal LC phase upon addition of the silicate species, based on 14N NMR spectroscopy. The silicates were organized into layers with rows of the cylindrical rods intercalated between the layers. Aging the mixture caused the layers to pucker and collapse around the rods, which then transformed into the surfactant-containing two-dimensional hexagonal MCM-41 structure.

A “charge density matching” mechanistic model proposed by Monnier et al. [29] and Stucky et al. [30] suggested that MCM-41 could be derived from a lamellar phase. The initial phase of the synthesis mixture was layered as detected by X-ray diffraction (XRD) formed due to the electrostatic attraction between the anionic silicates and the cationic surfactant head groups. As the silicate species began to condense, the charge density was reduced. Accompanying this process, the curvature was introduced into the layers to maintain the charge density balance with the surfactant head groups, which transformed the lamellar mesostructure into the hexagonal mesostructure.

Under synthesis conditions that prevented condensation of the silicate species, such as low temperatures and high pH (∼14), a true cooperative self-assembly of the silicates and surfactants was found possible. Firouzi et al. [31] conclusively showed, through 2H and 29Si NMR spectroscopy and neutron scattering that a micellar solution of CTAB transformed to a hexagonal phase in the presence of silicate anions. This was consistent with the effect of electrolytes on micellar phase transitions [32]. The silicate anions ion-exchanged with the surfactant halide counterions, to form a “silicatropic liquid crystal” (SLC) phase that involved silicate-encrusted cylindrical micelles. The SLC phase exhibited behavior very similar to typical lyotropic systems, except that the surfactant concentrations were much lower and the silicate counterions were reactive [3]. Heating the SLC phase caused the silicates to condense irreversibly into MCM-41. Firouzi et al. [31,3] also demonstrated that in addition to the charge balance requirement (i.e. electrostatic interaction) there was preferential bonding of the alkyl ammonium head groups to multi-charged D4R (double four-ring, [Si8O20]8−) silicate anions under the high pH conditions. The interaction was so strong that an alkyltrimethy- lammonium surfactant solution could force a silicate solution that did not contain D4R oligomers to re-equilibrate and form D4R species. It was suggested that this behavior came from the closely matched projected areas of a D4R anion and an ammonium head group (0.098nm2 versus 0.094nm2) and the correct distribution of charges on the projected surfaces.

Fyfe and Fu [34] were able to prepare mesostructured silicates with D4R silicates. Combination of the D4R precursors with cetyltrimethylammonium chloride (CTAC) surfactants produced mesostructured precipitates. Control of the condensation of the silicates within the mesostructure by acidic vapor treatment led to the observation of cubic, lamellar and hexagonal phases as intermediate transformation phases.

The previous theories have regarded the formation of

MCM-41 as a series of events that occur homogeneously throughout an aqueous solution. Recent work has shown that MCM-41 might be formed heterogeneously. Regev [35] found evidence for MCM-41 intermediate structures in the form of clusters of rod-like micelles “wrapped” by a coating of silicate through low-temperature transmission electron microscopy (TEM) and small-angle X-ray scattering. The clusters of elongated micelles were found before precipitation occurred. According to Regev [35], as the reaction progressed, the silicate species diffused to and deposited onto the individual surfaces of the micelles within the cluster. The clusters of elongated micelles eventually became clusters of silicate-covered micelles. Thus, the clusters of micelles served as nucleation sites for MCM-41 formation.

2.2. Ordered macroporous materials

Colloidal crystals consisting of three-dimensional ordered arrays of monodispersed spheres, represent novel templates for the preparation of highly ordered macroporous inorganic solids, exhibiting precisely controlled pore sizes and highly ordered three-dimensional porous structures. This macroscale templating approach typically consists of three steps as shown in Fig. 4. First, the interstitial voids of the monodisperse sphere arrays are filled with precursors of various classes of materials, such as ceramics, semiconductors, metals, monomers, etc. In the second step, the precursors

Fig. 4. Experimental procedure to generate three-dimensional macroporous inorganic materials by templating against crystalline arrays of colloidal spheres (modified from [124]).

Fig. 5. SEM image of colloidal array of ∼400nm polystyrene spheres [130].

condense and form a solid framework around the spheres. Finally, the spheres are removed by either calcination or solvent extraction.

The success of forming macroporous ordered structures is mainly determined by van der Waals interactions, wetting of the template surface, filling of the voids between the spheres and volume shrinkage of the precursors during solidification process. The colloidal crystal templating method may be used in combination with sol–gel, salt solution, nanocrystalline and other precursors to produce the inorganic three-dimensional macrostructures [23]. The colloidal crystal templates used to prepare three-dimensional macroporous materials include monodisperse polystyrene (PS), poly(methyl methacrylate) (PMMA) and silica spheres.

Fig. 6. SEM image of macroporous inorganic material. The walls are composed from vanadium–phosphorus oxide crystals (modified from [130]).

A typical SEM image of a colloidal array of polystyrene spheres used as a template in the synthesis of macroporous inorganic materials is shown in Fig. 5. Prior to precursor infiltration, these monodisperse spheres are ordered into close-packed arrays by sedimentation, centrifugation, vertical deposition or electrophoresis [36]. The final inorganic macroporous structure after the removal of spheres contains ordered interconnected pore structure shown in Fig. 6.

(Parte 1 de 8)