AlPO - American Chemical Society

Synthesis of Mesoporous Aluminophosphate (AlPO) and Investigation of Zirconium. Incorporation into Mesoporous AlPOs. Jian-Ming Lu1, Koodali T. Ranjit,...
0 downloads 0 Views 216KB Size
9284

J. Phys. Chem. B 2005, 109, 9284-9293

Synthesis of Mesoporous Aluminophosphate (AlPO) and Investigation of Zirconium Incorporation into Mesoporous AlPOs Jian-Ming Lu1 , Koodali T. Ranjit,*,† Pesak Rungrojchaipan, and Larry Kevan‡ UniVersity of Houston, Houston, Texas 77204-5003 ReceiVed: NoVember 10, 2004; In Final Form: March 4, 2005

Mesoporous aluminophosphate materials with variable amounts of zirconium have been synthesized at room temperature using a nonionic surfactant tri-block copolymer (PEO20PPO70PEO20) as the structure-directing agent. Powder X-ray diffraction of as-synthesized and calcined AlPO and ZrAlPO mesoporous materials shows a single broad peak near 2θ ) 2.5°, indicative of the average pore-pore correlation distance. Electron probe microanalysis shows that the ratio of P/Al in the powders is ∼0.5, far lower than 1.0 for an ideal aluminophosphate framework. XRD, TEM, and N2 adsorption data indicate that the calcined samples consist of wormlike tubular materials having surface areas >350 m2/g and pores in the mesopore range. Electron spin resonance (ESR) studies of the γ-irradiated and evacuated ZrAlPO samples show signals due to Zr3+ that increase with Zr content in addition to signals due to framework defects (i.e., V centers) and H atoms. The line shape and g values observed for Zr3+ are best explained as arising from a trivalent zirconium ion situated at the framework tetrahedral sites.

Introduction The discovery of the family of M41S ordered mesoporous materials by Mobil scientists has stimulated research in the synthesis and design of mesoporous materials.1,2 Mesoporous materials in general have high specific surface areas (usually ∼1000 m2/g), large pore volumes (>0.6 cm3/g), and large pore sizes (2-10 nm). These mesoporous materials are synthesized by a surfactant templating process, and their large and uniformsized pores ensure a facile diffusion of the reactants and products. The pore size of the mesoporous materials extends beyond the traditional realm of zeolites; hence, these materials find use as catalysts in the transformation of large molecules. Thus, mesoporous materials are attractive for applications in the field of catalysis and adsorption. Similar to zeolites, aluminophosphate molecular sieves (AlPO4-n) are crystalline materials having Al and P ions in tetrahedral framework positions.3-5 The AlPO materials are microporous in nature; hence, large and bulky organic molecules cannot be incorporated into their pores, and this limits their usefulness in catalytic reactions of large molecules. To extend the pore size of the AlPO materials, several research groups have attempted to prepare aluminophosphate-based mesoporous molecular sieves. Oliver et al.6 and Sayari et al.7-9 employed a high-temperature (100-180 °C) hydrothermal process to prepare mesoporous AlPO in a water-free tetraethylene glycol solvent and in water, respectively. In both of these approaches, a neutral surfactant, such as decylamine or dodecylamine, was employed as the structure-directing agent, and Catapal B alumina and phosphoric acid were used as the aluminum and phosphorus sources. Preparation of lamellar mesostructured aluminophosphates using * Corresponding author. Tel. (785) 532-6829. Fax: (785) 532-6666. E-mail: [email protected]. † Present Address: Department of Chemistry, Kansas State University, Manhattan, KS 66506. ‡ Deceased June 4, 2002.

surfactants has been reported by several research groups.10-13 However, these methods led to the formation of lamellar phases with very poor stability. Some modified syntheses based on this approach have also been reported in the literature. Kimura et al.14 have reported a mesostructured aluminophosphate having a hexagonal phase mixed with a lamellar phase by using hexadecyltrimethylammonium chloride as the surfactant. By using long alkyl chain length surfactants (C22) and a solubilizing agent, namely, 1,3,5-trimethylbenzene, mesostructured aluminophosphate materials were synthesized. The calcined material had a large surface area (>700 m2/g); however, the Al/P ratios of the products were ca. 1.5. This value is different from 1, which would be the ideal value expected for AlPO-n. In a further study, the formation of a lamellar mesostructured aluminophosphate material called APW-1 was reported.15 Only under very restricted and stringent experimental conditions could hexagonal phases be obtained. Feng et al.16 also reported a hexagonal mesostructured material prepared in the presence of F- ions; however, the structure collapsed on calcination. Holland et al.17 reported the preparation of mesoporous aluminophosphate based materials via layered intermediates; however, the materials were found to be thermally unstable. A series of metal (Ti, Fe, Co, V, and Ni) substituted disordered mesoporous aluminophosphate molecular sieves has been synthesized using hexadecyltrimethylammonium chloride as the surfactant and tetramethylammonium hydroxide as the charge balancing organic base.18-22 We have also reported the synthesis of hexagonal mesoporous aluminophosphate and silicoaluminophosphate molecular sieves with 4.0 nm pore sizes; however, these materials have only moderate thermal stability and decompose on heating to temperatures greater than 600 °C.23-26 The use of triblock copolymers as the structure-directing agent has further extended the pore size and more importantly the stability of the mesoporous AlPO materials.27 Ordered large pore (up to 12 nm) and stable mesoporous aluminophosphates (AlPO) were synthesized by using a block copolymer (EO106PO70EO106, Pluronic F127) as a structure-directing agent.28 Hence, we were

10.1021/jp0448393 CCC: $30.25 © 2005 American Chemical Society Published on Web 04/26/2005

Synthesis of Mesoporous Aluminophosphate interested in preparing mesoporous aluminophosphate materials using triblock copolymers as the structure-directing agent. New aluminophosphate materials with large pores and that are thermally stable are in demand for potential use as catalysts for the transformation of large molecules. Zirconia-based materials have wide applications in the field of catalysis in dehydration, hydrogenation, and hydroxylation reactions.29 The low surface area of zirconia (typically below 50 m2/g) restricts their industrial use.30 Isomorphous substitution of framework Si ions by Zr ions in zeolites of the MFI and MEL structure types was found to result in a Bronsted and Lewis acidity.31 The Zr containing MFI types of zeolites were found to have moderate acidity and selective oxidizing capability as compared to their pure siliceous counterpart. Zr-substituted AlPO-5 was found to be active in the isomerization of m-xylene.32 Zr-containing ZSM-5 microporous materials were found to be active for the hydroxylation of benzene to phenol.29 Isomorphous substitution of Si4+ by Zr4+ ions was inferred (based on an increase in unit cell volume and due to the presence of a band at 963 cm-1 from IR studies probably due to SiO-Zr linkages) but not conclusively proven. Other types of microporous materials such as ZSM-11, ZSM-12, and APO-11 have been explored for the incorporation of Zr4+ ions, but no definite proof for the incorporation of Zr4+ into framework Si4+positions was presented.33-35 Thermally stable, microporous, and mesoporous zirconium-doped silica were prepared via supramolecular templating under low pH conditions. The zirconium cations were found dispersed throughout the material and inferred from solid state NMR studies to be preferentially located at the pore wall surface.36 Zirconium containing mesoporous silica (Zr-MS/Zr-HMS) was synthesized using a hexadecylamine surfactant at room temperature; however, the study did not contain any information about the geometry, coordination, and location of the Zr ions.37 The nature and location of Zr in the mesoporous MCM-41 materials were identified by electron spin resonance (ESR) studies after the reduction of the Zr4+ ions to Zr3+.38 Thus, there are very few detailed reports on the location, coordination, and geometry of Zr4+ in mesoporous materials. There exist no reports in the literature regarding the incorporation of Zr ions into mesoporous aluminophosphate material. In the present study, we have attempted the synthesis of mesoporous aluminophosphate using triblock copolymers as the structure-directing agent. Since the incorporation of transition metal ions such as Zr4+ can create a Bronsted acidity for catalytic applications, we also investigated the incorporation of Zr ions in these mesoporous AlPO materials. The characterization of the pore structure and the identification of the local coordination and geometry of the incorporated Zr4+ ions require the use of several techniques. Thus, the present study includes a comprehensive characterization of the mesoporous AlPO materials. The Zr-containing AlPO materials were characterized by powder X-ray diffraction, nitrogen adsorption measurements, electron microprobe analysis (EMPA), ESR, and electron spinecho modulation (ESEM) spectroscopy. Experimental Procedures Synthesis. Mesoporous AlPO and Zr-AlPO were prepared by using a poly(ethylene oxide)20-poly(propylene oxide)70-poly(ethylene oxide)20 surfactant (average Mn ca. 5800; ca. 30 wt. % ethylene glycol, Aldrich), aluminum-tri-sec-butoxide (Aldrich, 99.99%), zirconium(IV) propoxide solution (Aldrich, 70 wt. % in 1-propanol), and phosphoric acid (85 wt %, EM Industries) as the aluminum, zirconium, and phosphorus sources, respec-

J. Phys. Chem. B, Vol. 109, No. 19, 2005 9285 tively. In a typical preparation of the synthesis gel, 6.0 g of poly(ethylene oxide)20-poly(propylene oxide)70-poly(ethylene oxide)20 surfactant was dissolved in 30 mL of 2-butanol under stirring. Then, 5.0 g of aluminum butoxide was slowly added under vigorous stirring. This solution was stirred for 30 min. To this solution, 0.7 mL of water and 0.5 mL of phosphoric acid in 35 mL of 2-butanol was added slowly. Zirconiumsubstituted mesoporous aluminophosphate (ZrAlPO) was synthesized by using a similar procedure, except that the required amount of zirconium propoxide solution for the desired Zr/Al ratio was added prior to the addition of phosphoric acid. The gels obtained using the previous synthetic protocols were stirred for 48 h at room temperature to promote crystallization. The solid product (white in color) obtained was filtered, washed repeatedly with ethanol and deionized water, and then dried in air overnight at 100 °C. The surfactant was then removed by calcination of the air-dried AlPO/ZrAlPO mesoporous materials in static air at 550 °C for 12 h. Characterization. X-ray diffraction patterns were recorded on a Philips 1840 X-ray diffractometer using CuKR radiation of a wavelength of 1.541 Å in the range of 1.5° < 2θ < 10°. The samples were spread as thin layers on metal holders. Chemical analysis was performed by electron microprobe analysis on a JEOL JXA-8600 electron beam superprobe operated at a beam voltage of 15 kV and current of 30 mA. Oxygen was calibrated with diopsite, CaMgSi2O6; Al with anorthite, CaAl2Si2O8; Zr with baddeleyite, ZrO2; and P with monazite, LaCeNdPO4. Prior to measurement, the samples were prepared as pressed pellets to make a dense material with a reasonably smooth surface. The composition of the AlPO and ZrAlPO mesoporous materials was determined by calibration with known standards and by averaging over several defocused areas (four randomly chosen regions) to give the bulk composition. The electron beam was defocused to a 10 µm diameter to minimize the damage caused to the sample by heating. No significant differences were observed between the selected regions, which indicate a uniform elemental composition over the entire sample. The precision was 700 m2/g) and narrow pore size distributions have been prepared by several research groups.44,45 The broadening of the XRD peaks has been observed by Ryoo and his workers in the case of KIT-1 materials.46 A similar type of broadening has been observed by Pinnavaia and his group for aluminosilicate MSU-X (X ) 1-4) materials, where X refers to the surfactant molecules that can be alkylPEO, alkyl-aryl-PEO, or PEO-PPO block copolymers.47-49 Such broadening has been attributed either to finite size effects (due to very fine particle morphology) or more commonly to a more disordered framework structure. The pore structure of the MSU-X materials is believed to be a three-dimensional disordered network consisting of short wormlike tubular channels. The channel widths are, however, uniform in these materials. Since the TEM images of our AlPO and ZrAlPO materials (discussed later) also show a large disorder, we attribute the broadening due to the presence of a disordered framework structure rather than due to finite size effects. In recent years, wormhole-like mesoporous materials have been the focus of attention of several research groups.50,51 These materials are more active heterogeneous catalysts in comparison to their ordered hexagonal analogues. Their enhanced reactivity has been partly explained based on the framework connectivity in three dimensions, thus allowing the guest molecules to more easily access active sites. The uniform channel in three dimensions is thought to give an advantage for preventing channel blockage over the linear channel in MCM-41. The synthesis of three-dimensional mesoporous aluminophosphate has been attempted by Sayari et al.7-9 They reported a lack of success despite trying to modify the synthesis by changing the composition of the gel and other experimental

Synthesis of Mesoporous Aluminophosphate conditions. On the basis of the present study and our previous results, we believe that their lack of success could be due to the use of high synthesis temperature, low gel pH, Catapal alumina, or a combination of these factors. The synthesis of AlPO and ZrAlPO was successful only at room temperature with the following composition (0.02-0.12) of ZrO2: 1.0 AlO2/0.25 P2O5/0.1 S/35 sec-BuOH/2 H2O, where S represents the triblock copolymer nonionic surfactant. Synthesis at higher temperatures (60, 100, and 120 °C) resulted only in amorphous materials, and no peaks in the low angle (2θ < 10°) were observed. Another factor that influences the synthesis is the aluminum source. The synthesis of mesoporous AlPO and ZrAlPO is successful with aluminum isopropoxide, aluminum butoxide, and aluminum hydroxide as the Al precursors but not with Catapal alumina. The P/Al ratio in the gel is 0.5, when gels with higher ratios of P/Al were used, and an amorphous material was obtained and no peaks in the low angle (2θ < 10°) were observed. In our previous studies, UHM-1 with a P/Al ratio of 0.60 was found to contain AlO4, AlO6, PO4, and P(OH)x(OAl)4-x, structural units where x is 1-3.25 This suggests that the AlPO and ZrAlPO hexagonal mesoporous framework arrangement cannot be formed from ideal AlO4 and PO4 structural units. Thus, we believe that such mixed coordination environments are possible in AlPO and ZrAlPO in the present study. This indicates that the condensation of either aluminum or phosphorus sources is incomplete and that the mesoporous AlPO and ZrAlPO have a disordered framework structure as compared to the MCM-41 analogue. The mechanism involving the synthesis of mesoporous materials employing nonionic surfactants (N0) is explained based on an electrically neutral N0I0 assembly mechanism between the inorganic precursor (I0) and the ethylene oxide groups of the nonionic surfactant that occurs through hydrogen bonding. This is in contrast to electrostatic interactions that occur between a cationic surfactant S+ and a negatively charged inorganic precursor I- responsible for the formation of ordered mesoporous materials in the case of the SBA-X materials (X ) 1-3). Since hydrogen bond interactions are weaker then electrostatic interactions, the final morphology, textural property, and quality of the material synthesized using nonionic surfactants are highly dependent on the initial local interactions created by the lipophile/hydrophile balance, the presence and diffusion of the various ions (Al3+, P5+, and Zr4+ in this instance), as well as the relative rates of hydrolysis and polymerization reactions of the inorganic precursors. Thus, a reliable and reproducible synthesis of mesoporous materials requires careful control of the experimental conditions. Further, Brownian motion of the precursor ions tends to destroy the order in the framework (that is temperature dependent too). This may explain why we were able to synthesize disordered AlPO and ZrAlPO only at room temperatures and unable to synthesize the mesoporous aluminophosphates at higher temperatures. The electrical neutrality is strictly valid when P/Al ) 1 in the gel and in the final product. However, in the present study, since the ratio of P/Al is 0.5 in the gel and in the final solid product, the mechanism based on N0I0 is not applicable. Calcination. On calcination of the as-synthesized AlPO and ZrAlPO materials, the resolution of their XRD patterns improves as shown in Figure 1. The mesoporous materials after calcination at 550 °C thus retain the integrity of the tubular structure upon removal of the organic structure-directing agent. The resulting calcined material was found to have a high surface area and very narrow pore size distribution (see discussion later).

J. Phys. Chem. B, Vol. 109, No. 19, 2005 9287 TABLE 1: Elemental Compositions of the Gel Mixture and Solid Products for AlPO and ZrAlPO gel solid a

ratio

AlPO

ZrAlPO

ZrAlPO

ZrAlPO

ZrAlPO

P/Al Zr/Al P/Al Zr/Al

0.5 0 0.4 0

0.5 0.02 0.48 N.D.a

0.5 0.04 0.48 N.D.

0.5 0.08 0.48 0.04

0.5 0.12 0.42 0.07

N.D.: not determined.

Figure 2. Transmission electron microscopy image of as-synthesized AlPO.

Electron Probe Microanalysis. EPMA analysis shows that both AlPO and ZrAlPO have unique compositions. The overall P/Al ratio in the gel and the resulting solid powder is comparable, and the EMPA analysis of the sample indicates that the distribution of the Al, P, and Zr atoms in the entire sample is homogeneous and evenly distributed. The overall P/Al ratio is ∼0.5. This suggests that there are some Al species that are independent in the mesoporous material without a neighboring phosphorus atom. This may be due to the incomplete condensation of the Al precursor. This indicates that excess hydroxyl aluminum species are involved in the assembly on the micelle interface to form mesoporous AlPO material. Thus, the mesoporous AlPO and ZrAlPO materials possess a nonideal structure AlPO framework significantly different from microporous aluminophosphate molecular sieves. Since the P/Al ratio is far lower than the ideal ratio of 1, the presence of extraframework Al2O3 species cannot be discounted. Table 1 shows the elemental composition of the gel and the resulting calcined dry solid ZrAlPO mesoporous materials. Transmission Electron Microscopy. The TEM image of calcined ZrAlPO is shown in Figure 2. The TEM picture does not clearly depict the pores, suggesting that those portions of the samples are not mesoporous. However, the nitrogen adsorption isotherms (discussed subsequently) clearly suggest the mesoporous nature of the pores. Also, the TEM picture indicates that the mesoporous possess a disordered structure in contrast to ordered hexagonal mesoporous MCM-41 materials. The disorder in the mesopore materials is similar to the ones observed for UHM-1, UHM-3, KIT-1, and MSU-1 mesoporous materials.25,46,47 N2 Adsorption Isotherms. Specific Surface Area and Pore Volumes. The BET surface area of the calcined AlPO material is very high (i.e., 438 m2/g) as indicated in Table 2. The high surface area of the calcined material support the XRD and TEM results that these materials possess a tubular structure rather than a lamellar structure. Figure 3 shows the nitrogen adsorption-

9288 J. Phys. Chem. B, Vol. 109, No. 19, 2005

Lu¨ et al.

Figure 3. Adsorption-desorption isotherms of nitrogen at 77 K of (a) AlPO, inset shows pore size distribution curve and (b) ZrAlPO (Zr/ Al ) 0.02), inset shows the pore size distribution curve for ZrAlPO (Zr/Al ) 0.02).

TABLE 2: Surface Area, Pore Size, and Pore Volume of Mesoporous AlPO and ZrAlPO mesoporous material

BET surface area (m2/g)

BJH pore size (Å)

pore volume (cm3/g)

AlPO ZrAlPO (Zr/Al ) 0.02) ZrAlPO (Zr/Al ) 0.04) ZrAlPO (Zr/Al ) 0.08) ZrAlPO (Zr/Al ) 0.12)

438 479 488 500 368

35 34 33 35 32

0.55 0.56 0.55 0.60 0.50

desorption isotherm of the calcined AlPO and ZrAlPO materials. Both the materials show isotherms typical of irreversible type IV adsorption isotherms with a H 1 hysteresis loop according to the IUPAC classification.52 More recently, it has become established that H1 hysteresis loops are characteristic of materials that have a high degree of pore uniformity and that possess cylindrical pore geometry. Hence, the presence of a H1 hysteresis loop on the adsorption isotherm for AlPO and ZrAlPO mesoporous material indicates the high degree of pore size uniformity and facile pore connectivity. N2 adsorption at low relative pressure (P/P0 < 0.25) is due to the monolayer adsorption of nitrogen on the walls of the mesopores. This is not indicative of the presence of micropores. As the relative pressure increases (P/P0 > 0.4), the isotherms exhibit sharp inflections in the P/P0 range of 0.45-0.80, characteristic of capillary condensation within the uniform mesopores. The P/P0 positions of the inflection points clearly lie in the diameter of the mesopore range, and the sharpness of the steps indicates the uniformity of the tubular channels. The broad hysteresis loop in the isotherms also is indicative of the disorder in the shape of these mesoporous materials. This is consistent with the XRD

and TEM results. On incorporation of Zr in the mesoporous ZrAlPO materials, the isotherms show an inflection similar to the AlPO materials. This is in contrast to our previous results. Zhao et al.25 have observed that in silicon incorporated SAPO materials, the isotherm exhibited a sharper inflection, suggesting that the mesoporous SAPO materials were more ordered as compared to the AlPO materials. There was no uniform trend in the BET surface area and pore volume with increase in Zr content. Pore Size Distribution. There are many methods for calculating the pore size distributions. These are usually calculated by using the Kelvin equation53 or the HorvathKawazoe method54 and its modifications.55 The first group includes the method of Barrett-Joyner-Halenda (BJH).56 The BJH method has been used extensively for the calculation of the pore size distribution of mesoporous materials. However, the original BJH work does not clearly explain the choice of the branch of the isotherm. Tanev et al.57 proposed that the desorption branch of the isotherm carries more information about the degree of pore blocking than does the adsorption branch. More recently, it is has been demonstrated that the BJH computation procedure (cylindrical pore geometry assumed in the algorithm) can be applied successfully for calculating the pore size distributions in mesoporous materials.58 We have based our calculations on these arguments, and the insert in Figure 3 shows the pore size distribution in AlPO and ZrAlPO materials. The BJH plot for both AlPO and ZrAlPO gives a remarkably narrow pore size distribution with a pore size of about 35 Å. There was no significant change in the pore size distribution on increasing the Zr content in the AlPO materials. Also, worth mentioning is that despite the presence of disordered tubular channels in these materials, the channel width is very uniform as in the ordered MCM-41 materials. This is based on the fact that the full width at half-maximum of the pore size distribution curve is ∼35 Å, which is similar to ∼30 Å in the case of MCM41. Thus, the AlPO and ZrAlPO mesoporous materials have truly uniform channel widths, while the channels are completely disordered. These results once again confirm the presence of tubular mesoporous channels in these disordered materials. The pore size calculations by nitrogen physisorption is smaller than the repeat distance a0 ) 42 Å determined by XRD (a0 ) 2d(100)/x3, where d(100) ) 35 Å for most of the AlPO and ZrAlPO materials in the present study). This is because the latter includes the thickness of the pore wall. Thus, the pore wall of these materials is around 7 Å. The thicker wall observed for materials prepared with nonionic PPO-PEO-PPO surfactants is believed to be due to weaker PEO-alumina or PEO-phosphate interactions and the weaker microphase separating tendency of the surfactant PEO-PPO relative to the CTAB surfactant in the MCM-41 system.59 Thermogravimetric Analysis (TGA). Thermogravimetric analysis of AlPO shows a weight loss of ∼ 46% on heating to 700 °C. The TGA results of AlPO show three distinct weight losses, one endothermic and two exothermic peaks as shown in Figure 4a. The first weight loss (3% weight loss) near 70 °C is attributed to the loss of physisorbed water. The second weight loss (20% weight loss) near 165 °C and the third weight loss (23% weight loss) in the region of 210-270 °C are attributed to the desorption and the decomposition of the surfactant species. Thus, by 300 °C, essentially all of the nonionic organic surfactant species is removed from the tubular channels of the mesoporous AlPO material. Thus, the decomposition of the nonionic surfactant species occurs at far lower temperatures as compared to 500 °C normally required for the removal of the

Synthesis of Mesoporous Aluminophosphate

J. Phys. Chem. B, Vol. 109, No. 19, 2005 9289

Figure 5. ESR spectra of mesoporous AlPO and ZrAlPO recorded after thermal evacuation and γ-irradiation at 77 K for (a) AlPO, (b) ZrAlPO (Zr/Al ) 0.02), (c) 0.04 (Zr/Al ) 0.04), (d) 0.08 (Zr/Al ) 0.08), and (e) 0.12 (Zr/Al ) 0.12).

Figure 4. Thermal gravimetric analysis curves for (a) AlPO and (b) ZrAlPO (Zr/Al ) 0.02).

cationic surfactants such as CTAB used for the synthesis of MCM-41 under basic conditions. This is consistent with the weaker interactions expected between the nonionic surfactant and the inorganic alumina or phosphorus wall as compared to that of cationic CTAB and silica under basic MCM-41 synthesis. The TGA results of ZrAlPO also showed three weight losses similar to AlPO. The total weight loss was ∼48% on heating to 700 °C. Figure 4b shows the TG curve for ZrAlPO (Zr/Al ) 0.02). Diffuse Reflectance (DR) Studies. Diffuse reflectance spectroscopy is one of the techniques to distinguish framework species from extraframework species.60 DR studies have hence been used to identify and characterize zirconium species in various matrixes.61,62 Zr containing mesoporous silica Zr-MS is found to have an absorption in the region of 205-215 nm.37 In Zr-ZSM-5 and Zr-SBA-15, the absorption bands are observed at 212 and 210 nm, respectively.61,62 These absorption bands are usually attributed to ligand-to-metal charge transfer (LMCT) from an O2- to an isolated Zr4+ ion in a tetrahedral configuration. In bulk ZrO2, where there are Zr-O-Zr linkages, the LMCT band shifts to a lower energy (i.e., to ∼230 nm). For the DR studies in the present investigation, the DR spectra were recorded with reference to the pure AlPO sample. The DR studies of ZrAlPO containing low amounts of Zr (Zr/Al ) 0.02 and 0.04) in the present study did not show well-resolved absorption bands and were difficult to observe because of very low signal-to-noise ratios. However, for ZrAlPO samples with

Zr/Al ratios of 0.08 and 0.12, small but very distinct bands could be seen in the 210-220 nm region. The DR results suggest that the zirconium ion is probably substituted in a tetrahedral site (i.e., in framework position) in accordance with previous results. ESR Studies on ZrAlPO. Calcined samples of ZrAlPO were ESR silent both at room temperature and 77 K. However, on evacuation (ca. 10-4 Torr at 550 °C for 5 h), a strong anisotropic signal with a six line hyperfine splitting centered at gav ) 2.007 with a peak-to-peak line width of ∼6 G appeared for both AlPO and ZrAlPO as shown in Figure 5. The origin of these signals could be attributed to (a) the presence of nonionic surfactant used as the structure directing agent in the synthesis of AlPO and ZrAlPO, (b) hydrocarbons adsorbed from the vacuum system (vacuum grease or oil), or (c) V centers. The presence of unburnt template molecules in the calcined AlPO and ZrAlPO is ruled out since the TGA of the calcined samples indicate only a ∼5% weight loss due to physisorbed water only. Since necessary care was taken to prevent any contamination or adsorption of hydrocarbons from the vacuum system onto the samples, the origin of these signals is attributed to the presence of framework defects (i.e., the presence of electrons trapped in oxygen vacancies (V centers)). The strong signal with gav ) 2.007 in ZrAlPO is undoubtedly due to V centers created by the irradiation as a similar signal was observed also in AlPO upon the same treatment. The shift of the signal from the free electron value of g ) 2.0023 is positive for a hole and negative for an unpaired electron. The shift observed in the present study, ∆g ∼0.005 is positive and consistent with hole trapping. These are radiation-induced hole centers trapped in the lone pair p orbital of the associated oxygen atoms. Further, the V centers did not show power saturation with the microwave power.63 The lack of microwave power saturation for the V center indicates that the defects are associated with the framework.

9290 J. Phys. Chem. B, Vol. 109, No. 19, 2005

Figure 6. Reaction mechanisms for the formation of V centers in AlPO and ZrAlPO.

Microwave power saturation is expected if the V centers are due to ionized impurities, interstitial lattice defects, or trapped holes. The stability of the V centers in zeolites also depends on the presence of electron scavengers such as H+ or Na+. The V center intensity is larger in ZrAlPO as compared to AlPO materials, (compare Figure 5a,b-e). The formation of such defect centers has been observed by us previously in SiMCM41 and TiMCM-41 materials.64 A reaction mechanism for the formation of V centers in aluminosilicate zeolites has been proposed by Abou-Kais et al.65 The formation of V centers in mesoporous ZrAlPO materials can be explained by the reaction mechanisms proposed in Figure 6. In AlPO, the framework is not a good electron acceptor; hence, the equilibrium is shifted to the left resulting in a lower stability of the V centers and a reduced intensity of the ESR signal due to the V centers as shown in Figure 6 (mechanism I). The observation of six-line hyperfine splitting is due to Al-O-P framework units (27Al nuclear spin is 5/2). In ZrAlPO, two additional types of mechanisms are possible for the formation of V centers. The H+ ions are good electron scavengers and can easily react with the ejected electron to form H atoms as seen in the ESR spectra (Figure 5). However, they are unstable and can recombine with the V centers or can be trapped in some cavities. Thus, the large increase in the V centers in ZrAlPO in comparison with AlPO can be explained by the presence of H+ ions produced by the substitution of tetravalent Zr for a phosphorus site in the AlPO framework (mechanism II). In ZrAlPO, another mechanism by which V centers can be stabilized is by the reduction of Zr4+ to Zr3+ using the ejected electron (mechanism III).This mechanism is consistent with an increase in the intensity of the signal due to V centers together with formation of a Zr3+ signal in the ESR spectrum of ZrAlPO in comparison with the spectrum of AlPO. This provides indirect evidence for the substitution of zirconium in the AlPO framework. In ZrAlPO, a reduction of Zr4+ to Zr3+ can stabilize the V centers; hence, the ESR signal due to V centers is higher as compared to AlPO. In zirconia and zirconium hydroide-oxide materials, thermal evacuation led to the formation of ESR signals due to Zr3+

Lu¨ et al. having g⊥ ) 1.978 and g| ) 1.953 in addition to the defect centers.66 However, in the present study, we did not observe any signals due to Zr3+ after evacuation. This is probably because of the difference in the redox behavior of Zr in bulk ZrO2 and ZrAlPO. Similarly, Chaudhari et al.38 also did not observe any ESR signal due to Zr3+ in Zr-MCM-41 after evacuation and attributed this behavior due to the role of silica in stabilizing Zr4+. Similarly, we believe that Al3+ and/or P5+ may be responsible for stabilizing Zr4+. Attempts to reduce Zr4+ in ZrAlPO by dry hydrogen or carbon monoxide were not successful in the present study. This is in contrast to the literature report in which Zr-MCM-41 samples after reduction with dry hydrogen exhibited ESR signals corresponding to two species, I and II.38 The species I with ESR signals at g1 ) 1.943, g2 ) 1.970, and g3 ) 1.876 was attributed to Zr3+ in rhombohedral geometry located within the pore walls. Species II was assigned to the Zr4+(O2•-) by the authors due to the interaction of Zr3+ with atmospheric oxygen and located at the surface of the pores. In the present work, γ-irradiation of the evacuated ZrAlPO samples resulted in the formation of three distinct ESR signals, and no signals due to Zr4+(O2•-) were observed. The first strong signal having a six-line hyperfine splitting with gav ) 2.007 is attributed to V centers as discussed earlier. The second ESR signal with weak axially symmetric signals at g| ) 1.966 and g⊥ ) 1.934 is assigned to Zr3+ in tetrahedral geometry. As will be discussed later, we assign this signal to Zr3+ produced in distorted tetrahedral geometry by the reduction of Zr4+ by γ-irradiation. The considerable departure of the g components from the free spin value ge ) 2.0023 reflects an appreciable spin-orbit interaction for this species. It is interesting to note that the g values of Zr3+ ions in ZrAlPO are different from those observed for pure zirconia samples. Figure 5 shows the ESR spectra of AlPO and ZrAlPO (Zr/Al ) 0.02, 0.04, 0.08, and 0.12) samples at 77 K after evacuation and γ-irradiation. All the spectra show signals at gav ) 2.007 due to defect centers; however, the magnitude of the signal at gav ) 2.007 is higher for all ZrAlPO samples as compared to AlPO as discussed earlier. Also, it is clear from Figure 5a that the AlPO sample after γ-irradiation does not show any peaks at g| ) 1.966 and g⊥ ) 1.934 corresponding to Zr3+. The intensity of the signal at g| ) 1.966 and g⊥ ) 1.934 is found to increase with increase in Zr content suggesting that the assignment of these signals due to Zr3+ is reasonable. The g⊥ component usually has a weak absorption profile and hence is poorly resolved. The g⊥ component is therefore not well-resolved for ZrAlPO samples having low amounts of Zr (i.e., ratios of Zr/Al < 0.04). In addition to the previous two signals, the third signal in the spectra contains a doublet separated by 500 G due to trapped H atoms. Figure 7 shows the ESR spectra of ZrAlPO (Zr/Al ) 0.04) after γ-irradiation and subsequent annealing to room temperature for ∼5 s. The signal due to H atoms vanishes immediately upon exposure to room temperature for a short duration ( δ. However, g⊥ is larger than g|. In the present study, however, we have observed that g| is larger than g⊥. Also, the g values calculated for an octahedral geometry differ considerably from the experimentally observed values of g| ) 1.966 and g⊥ ) 1.934. In a tetrahedral crystal field, a d1 ion has the 2E state in the ground level. A further tetrahedral distortion removes the orbital degeneracy of the ground state as shown in Figure 8b. If the distortion is positive (i.e., tetragonal compression), the |0> or dz2 will lie low in energy. Conversely, if the distortion is negative (i.e., tetragonal elongation), the 1/x2 (|2> + |-2>) or dx2-y2 state lies lower. Considering the distortion axis as the z direction, the expressions for the g components in a tetragonally distorted tetrahedral field,67 the g values obtained to first order for an axially symmetry, are g| ≈ 2.0023 and g⊥ ≈ 2.0023 - 6λ/δ for tetragonal compression and g| ≈ 2.0023 - 8λ/δ and g⊥ ≈ 2.0023 - 2λ/δ for tetragonal elongation Assuming ∆ ) 4/9(∆0) ) 4889 cm-1 for tetrahedral coordination, the corresponding g values calculated are g| ) 2.0023 and g⊥ ) 1.702 for compression and g| ) 1.602 and g⊥ ) 1.902 for elongation. In the former case (δ > 0), g⊥ < g| ≈ 2.0023, and in the latter (δ > 0), g| < g⊥ < 2.0023, and both the g components are smaller than 2.0023. From the experimentally observed g values, we can conclude that the Zr3+ ions are most likely in a tetrahedral coordination (tetragonal compression) rather than octahedral coordination. If Zr occupies a framework site, then one expects tetrahedral symmetry. Considerable deviations of ESR g values from the calculated ones are observed for Zr3+ ions in many systems including the present one. Indeed, very rarely do the experimental and calculated g values match for Zr3+ ions and d1 ions such as Ti3+.69,70 This suggests that additional effects must be taken into account to fit the experimental data. Thus, it is clear that coupling to other electronic levels is also important. However, these are beyond the scope of the present investigation. Table 3 shows the comparison of the ESR spectroscopic parameters of various Zr substituted molecular sieves. It is interesting to compare the spin Hamiltonian parameters in ZrAlPO mesoporous materials and TiMCM-41 materials. Both Zr4+ and Ti4+ are isoelectronic, but the unpaired electron of Ti4+occupies a 3d metal orbital, whereas that of Zr4+ occupies a 4d orbital. Ti in the MCM-41 structure could only be reduced by γ-irradiation similar to Zr in ZrAlPO in our present study. We have earlier observed that γ-irradiation at 77 K of TiMCM41 having a high Si/Ti ratio produced an orthorhombic signal

9292 J. Phys. Chem. B, Vol. 109, No. 19, 2005 due to V centers and an axial with g| ) 1.971 and g⊥ ) 1.901 attributed to Ti3+ in framework tetrahedral sites.64 The smaller g| value for Zr3+ as compared to Ti3+ is due to the differences in the spin-orbit coupling constants of Zr (500 cm-1) and Ti (155 cm-1). Similarly, γ-irradiation of TS-1 produced Ti3+ with similar characteristics. The very similar ESR signals for Ti3+ in TiMCM-41, TS-1, and Zr3+ in ZrAlPO indicate that the coordination environments of Ti and Zr ions are the same. The Zr4+ ions in ZrAlPO do not reduce on thermal evacuation as compared to some Zr4+ ions that are reduced in Zr-MCM-41. In Zr-MCM-41, these were ascribed to Zr ions located at surface sites or the surface of the pores.38 The lack of observation of Zr3+ ions after thermal evacuation in ZrAlPO clearly suggests the absence of Zr4+ ions at surface sites in ZrAlPO mesoporous materials. Further, DR spectra suggest that the Zr ions are located at framework sites. Thus, both ESR and DR studies indicate that Zr ions are located at framework sites. ESEM Studies. To obtain more information about the location of Zr3+ species in ZrAlPO, three-pulse ESEM spectra of γ-irradiated ZrAlPO material at 5 K was done. The spectrum was recorded at an external magnetic field of 3410 G corresponding approximately to g ) 1.96, where the echo intensity reaches a maximum and τ was set to 0.25 µs to eliminate modulation from 27Al nuclei and maximize modulation from 31P nuclei. No attempt was made to fit the strong aluminum modulation because of the additional parameters needed for the large quadrupole interaction. Typically, 27Al modulation has not been too useful in determining local environmental structure in powder samples. Echo detected field swept spectra at 5 K gave absorption spectra rather than derivative spectra and were very poorly resolved; hence, no useful information could be discerned from the ESEM studies. Hence, it was not possible to make valid comparisons with the ESR data obtained at 77 K. Threepulse deuterium ESEM spectra of ZrAlPO with adsorbed D2O and CD3OH showed a very weak deuterium modulation that was poorly resolved; hence, ESEM studies failed to shed further light on the coordination and geometry of Zr ions in the mesoporous ZrAlPO materials. Conclusions Mesoporous AlPO and ZrAlPO were successfully synthesized using tri-block copolymers as the structure-directing agent. The framework contains a nonideal P/Al ratio less than 1.0, making these materials different from conventional microporous AlPO materials such as AlPO-5 and AlPO-11. The results indicate that the mesoporous AlPO materials consist of wormlike tubular channels that contain a disordered framework structure but with channel widths that are highly uniform. Thus, the mesoporous materials prepared in the present study exhibit morphologies similar to KIT-1 and MSU-X (where X ) 1-4). On evacuation and γ-irradiation, ESR studies show signals due to Zr3+ in addition to framework defects. The line shape and width of the ESR signals due to Zr3+ indicate that the Zr ions isomorphously substitute for either Al or P. Acknowledgment. The authors (J.M.L., K.T.R., and P.R.) are thankful to the late Prof. Larry Kevan for giving us an opportunity to work in his laboratory and for his constant support and motivation for pursuing research in the areas of ESR and mesoporous materials. This research was supported by the Robert A. Welch Foundation and the University of Houston Energy Laboratory. References and Notes (1) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710.

Lu¨ et al. (2) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Schmitt, K. D.; Chu, C. T. W.; Olsen, D. H.; Shepard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834. (3) Prakash, A. M.; Hartmann, M.; Zhu, Z.; Kevan, L. J. Phys. Chem. B 2000, 104, 1610. (4) Ranjit, K. T.; Kevan, L. J. Phys. Chem. B 2003, 107, 2610. (5) Hartmann, M.; Kevan, L. Chem. ReV. 1999, 99, 635. (6) Oliver, S.; Kuperman, A.; Coombs, N.; Louth, A.; Ozin, G. A. Nature 1995, 378, 47. (7) Sayari, A.; Moudrakovski, I. L.; Reddy, J. S. Chem. Mater. 1996, 8, 2080. (8) Sayari, A. Studies in Surface Science and Catalysis, Vol. 105. In Progress in Zeolite Microporous Materials; Chon, H., Ihm, S.-K., Uh, Y. S., Eds.; Elsevier: Amsterdam, 1997; pp 37-44. (9) Sayari, A.; Karra, V. R.; Reddy, J. S.; Moudrakovski, I. L. J. Chem. Soc., Chem. Commun. 1996, 411. (10) Cheng, S.; Tzeng, J.-N.; Hsu, B.-Y. Chem. Mater. 1997, 9, 1788. (11) Gao, Q.; Chen, J.; Xu, E.; Yue, Y. Chem. Mater. 1997, 9, 457. (12) Pophal, C.; Schnell, R.; Fuess, H. Studies in Surface Science and Catalysis, Vol. 105. In Progress in Zeolite Microporous Materials; Chon, H., Ihm, S.-K., Uh, Y. S., Eds.; Elsevier: Amsterdam, 1997; pp 101-108. (13) Gao, Q.; Chen, J.; Li, S.; Xu, R. Studies in Surface Science and Catalysis, Vol. 105. In Progress in Zeolite Microporous Materials; Chon, H., Ihm, S.-K., Uh, Y. S., Eds.; Elsevier: Amsterdam, 1997; pp 389-396. (14) Kimura, T.; Sugahara, Y.; Kuroda, K. Chem. Commun. 1998, 559. (15) Kimura, T.; Sugahara, Y.; Kuroda, K. Chem. Mater. 1999, 11, 508. (16) Feng, P.; Xia, Y.; Feng, J.; Bu, X.; Stucky, G. D. J. Chem. Soc., Chem. Commun. 1997, 949. (17) Holland, B. T.; Isbester, P. K.; Blanford, C. F.; Munson, E. J.; Stein, A. J. Am. Chem. Soc. 1997, 119, 6796. (18) Selvam, P.; Mohapatra, S. K. Microporous Mesoporous Mater. 2004, 73, 137. (19) Mohapatra, S. K.; Selvam, P. Chem. Lett. 2004, 33, 198. (20) Selvam, P.; Sonavane, S. U.; Mohapatra, S. K.; Jayaram, R. V. AdV. Synth. Catal. 2004, 346, 542. (21) Mohapatra, S. K.; Selvam, P. Catal. Lett. 2004, 93, 47. (22) Selvam, P.; Mohapatra, S. K.; Sonavane, S. U.; Jayaram, R. V. Tetrahedron Lett. 2004, 45, 2003. (23) Zhao, D.; Luan, Z.; Kevan, L. J. Chem. Soc., Chem. Commun. 1997, 1009. (24) Zhao, D.; Luan, Z.; Kevan, L. J. Phys. Chem. B 1997, 101, 6943. (25) Luan, Z.; Zhao, D.; He, H.; Klinowski, J.; Kevan, L. J. Phys. Chem. B 1998, 102, 1250. (26) Bae, J. Y.; Ranjit, K. T.; Luan, Z.; Krishna, R. M.; Kevan, L. J. Phys. Chem. B 2000, 104, 9661. (27) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. J. Am. Chem. Soc. 1998, 120, 6024. (28) Wang, L.; Tian, B.; Fan, J.; Liu, X.; Yang, H.; Yu, C.; Tu, B.; Zhao, D. Microporous Mesoporous Mater. 2004, 67, 123. (29) Tanabe, K.; Yamaguchi, T. Catal. Today 1994, 20, 185. (30) Yamaguchi, T. Catal. Today 1994, 20, 199. (31) Dongare, M. K.; Singh, P.; Moghe, P.; Ratnasamy, P. Zeolites 1991, 11, 690. (32) Dongare, M. K.; Sabde, D. P.; Shaikh, R. A.; Kamble, K. R.; Hegde, S. G. Catal. Today 1999, 49, 267. (33) Ji, S.; Li, H.-l.; Liao, S.-j.; Wang, L.-f. Fenzi Cuihua 2001, 15, 273. (34) Yu, L.; Pang, W. Shiyou Xuebao, Shiyou Jiagong 1994, 10, 56. (35) Wang, Z.-M.; Yan, Z. F. Preprints-American Chemical Society, DiVision of Petroleum Chemistry 2001, 46, 292. (36) Wong, M. S.; Huang, H. C.; Ying, J. Y. Chem. Mater. 2002, 14, 1961. (37) Tuel, A.; Gontier, S.; Teissier, R. J. Chem. Soc., Chem. Commun. 1996, 651. (38) Chaudhari, K.; Bal, R.; Das, T. Kr.; Chandwadkar, A.; Srinivas, D.; Sivasanker, S. J. Phys. Chem. B 2000, 104, 11066. (39) Fauth, J. M.; Schweiger, A.; Brauschweiler, L.; Forrer, J.; Ernst, R. R. J. Magn. Reson. 1986, 66, 74. (40) Kevan, L. In Time Domain Electron Spin Resonance; Kevan, L., Schwartz, R. N., Eds.; Wiley-Interscience: New York, 1979; Ch. 8. (41) Kevan, L.; Bowman, M. K.; Narayana, P. A.; Boeckman, R. K.; Yudanov, V. F.; Tsvetkov, Yu. D. J. Chem. Phys. 1975, 63, 409. (42) Mamak, M.; Coombs, N.; Ozin, G. A. J. Am. Chem. Soc. 2000, 122, 8932. (43) Xing, W.; Li, F.; Yan, Z.-Y.; Lu, G. Q. J. Power Sources 2004, 134, 324. (44) Luan, Z.; He, H.; Zhou, W.; Cheng, C.-F.; Klinowski, J. J. Chem. Soc., Faraday Trans. 1995, 91, 2955. (45) Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865. (46) Ryoo, R.; Kim, J. M.; Ko, C. H.; Shin, C. H. J. Phys. Chem. 1996, 100, 17718. (47) Bagshaw, S. A.; Prouzet, E.; Pinnavaia, T. J. Science 1995, 269, 1242.

Synthesis of Mesoporous Aluminophosphate (48) Bagshaw, S. A.; Pinnavaia, T. J. Angew. Chem., Int. Ed. Engl. 1996, 35, 1102. (49) Prouzet, E.; Pinnavaia, T. J. Angew. Chem., Int. Ed. Engl. 1997, 36, 516. (50) Wei, Y.; Dong, H.; Feng, Q. Chem. Phys. Chem. 2002, 3, 803. (51) Polarz, S.; Smarsly, B.; Bronstein, L.; Antonietti, M. Angew. Chem., Int. Ed. 2001, 40, 4417. (52) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscow, L.; Pierotti, R. A.; Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603. (53) Broekhoff, J. C. P.; de Boer, J. H. J. Catal. 1967, 9, 8. (54) Horvath, G.; Kawazoe, K. J. Chem. Eng. Jpn. 1983, 16, 470. (55) Saito, A.; Foley, H. C. AIChE J. 1991, 37, 429. (56) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73, 373. (57) Tanev, P. T.; Vlaev, L. T. J. Colloid Interface Sci. 1993, 160, 110. (58) Kruk, M.; Jaroniec, M. Chem. Mater. 2001, 13, 3169. (59) Tian, B.; Liu, X.; Tu, B.; Yu, C.; Fan, J.; Wang, L.; Xie, S.; Stucky, G. D.; Zhao, D. Nature Mater. 2003, 2, 159. (60) Morey, M. S.; Stucky, G. D.; Schwarz, S.; Fro¨ba, M. J. Phys. Chem. B 1999, 103, 2037.

J. Phys. Chem. B, Vol. 109, No. 19, 2005 9293 (61) Wang, G.; Wang, X. Stud. Surf. Sci. Catal. 1993, 83, 67. (62) Newalkar, B. L.; Olanrewaju, J.; Komarneni, S. J. Phys. Chem. B 2001, 105, 8356. (63) Stamires, D. N.; Turkevich, J. J. Am. Chem. Soc. 1964, 86, 757. (64) Prakash, A. M.; Sung-Suh, H. M.; Kevan, L. J. Phys. Chem. B 1998, 102, 857. (65) Abou-Kais, A.; Vedrine, J. C.; Massardier, J. J. Chem. Soc., Faraday Trans. 1975, 71, 1697. (66) Liu, H.; Feng, L.; Zhang, X.; Xue, Q. J. Phys. Chem. 1995, 99, 332. (67) Weil, J. A.; Bolton, J. R.; Wertz, J. E. Electron Paramagnetic Resonance: Elementary Theory and Practical Applications; John Wiley: New York, 1994; Ch. 8, pp 213-238. (68) Kim, Y. M.; Bray, P. J. J. Chem. Phys. 1970, 53, 716. (69) Ramaswamy, V.; Tripathi, B.; Srinivas, D.; Ramaswamy, A. V.; Cattaneo, R.; Prins, R. J. Catal. 2001, 200, 250. (70) Bencini, A.; Gatteschi, D. In Transition Metal Chemistry; Melson, G. A., Figgis, B. N., Eds.; Marcel Dekker: New York, 1982; Vol. 8, pp 1-178.