Stabilization of Self-Assembled Alumina Mesophases - American ...

Article pubs.acs.org/cm

Stabilization of Self-Assembled Alumina Mesophases Lidia López Pérez,† Sébastien Perdriau,‡ Gert ten Brink,§ Bart J. Kooi,§ Hero Jan Heeres,† and Ignacio Melián-Cabrera*,† †

Chemical Reaction Engineering (ITM), ‡Stratingh Institute of Chemistry, and §Zernike Institute for Advanced Materials and Materials Innovation Institute M2i, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands S Supporting Information *

ABSTRACT: An efficient route to stabilize alumina mesophases derived from evaporation-induced self-assembly is reported after investigating various aspects in-depth: influence of the solvent (EtOH, s-BuOH, and t-BuOH) on the textural and structural properties of the mesophases based on aluminum tri-sec-butoxide (ATSB), synthesis reproducibility, role of nonvolatile acids, and the crystallization and thermal stability of the crystalline counterparts. Mesophase specific surface area and pore uniformity depend notably on the solvent; sBuOH yields the highest surface area and pore uniformity. The optimal mesophase synthesis is reproducible with standard deviations in the textural parameters below 5%. The most pore-uniform mesophases from the three solvents were thermally activated at 1023 K to crystallize them into γ-alumina. The s-BuOH mesophase is remarkably thermally stable, retaining the mesoscopic wormhole order with 300 m2/g (0.45 cm3/g) and an increased acidic site density. These features are not obtained with EtOH or t-BuOH, where agglomerated γ-Al2O3 crystallites are formed with lower surface areas and broader pore size distributions. This was rationalized by the increase of the hydrolysis rate using EtOH and t-BuOH. t-BuOH dehydrates under the synthesis conditions or reacts with HCl, situations that increase the water concentration and rate of hydrolysis. It was found that EtOH exchanges rapidly, producing a highly reactive Al-ethoxide, thus enhancing the hydrolysis rate as well. Particle heterogeneity with random packing of fibrous and wormhole morphologies, attributed to the high hydrolysis rate, was observed for mesophases derived from both solvents. Such a low particle coordination favors coarsening with enlargement of the pore size distribution upon thermal treatment, explaining the lower thermal stability. Controlled hydrolysis and formation of low-polymerized Al species in s-BuOH are possibly responsible for the adequate assembly onto the surfactant. This was verified by the formation of a regular distribution of relatively size-uniform nanoparticles in the mesophase; high particle coordination prevents coarsening, favors densification, and maintains a relatively uniform pore size distribution upon thermal treatment. The acid removal in the evaporation is another key factor to promote network condensation in this route. KEYWORDS: evaporation-induced self-assembly, γ-alumina, mesoporosity, alkoxide chemistry, sol−gel, thermal stability, coarsening, solvent effects controlled hydrolysis and condensation conditions. The first mesoporous aluminas were found by Pinnavaia10,11 and Davis12 and co-workers. Soon after, a remarkable development in the field took place; nowadays mesoporous aluminas can be synthesized with amorphous,10−23 partially crystallized,24−27 and most interestingly for catalysis, atomically well-crystallized γ-walls.28−42 In order to make the crystallization possible without structural collapse, various approaches have been proposed. The surfactant-induced fiber formation28−32,40 renders γ-Al2O3 with notable surface areas between 250 and 450 m2/g, which are obtained upon a relatively mild thermal activation at 600−823 K. The pore size distribution is typically broad. The introduction of pore narrowness via 2D hexagonal

1. INTRODUCTION Aluminas are one of the largest volume produced inorganic materials. In catalysis, γ-alumina is the most employed crystalline phase,1−4 serving as support for catalysts in largescale processes such as hydroprocessing, catalytic reforming, steam reforming, and for automotive exhaust catalytic converters. Bohemite is thermally dehydrated into the γ-phase at temperatures between 723 and 1000 K.1,5 The porosity originates from the interparticle space that produces relatively low surface areas; the pores lay in the high mesopore region and the pore size distribution is broad. Increasing the surface area of such materials will result in improved volumetric reactor efficiency.6 This has been the challenge for business and academia, not only for commercial processes but also for emerging applications.7 The surface area has been enhanced by means of supramolecular surfactants8,9 and Al alkoxides in kinetically © 2013 American Chemical Society

Received: October 1, 2012 Revised: February 17, 2013 Published: March 8, 2013 848

dx.doi.org/10.1021/cm303174r | Chem. Mater. 2013, 25, 848−855

Chemistry of Materials

Article

Preparation. Mesoporous alumina mesophases were synthesized using a procedure based on Niesz et al.19 in acidic medium (HCl, citric, oxalic or p-toluenesulfonic acids) with a nonionic surfactant (Pluronic P123, EO20PO70EO20) and aluminum tri-sec-butoxide (ATSB) as aluminum source. Three alcohols were investigated systematically (ethanol, sec-butanol, and tert-butanol) while the synthesis temperature was varied from 313, 333, and 353 K. The amount of water was fixed to H2O/Al = 6 (mol). The resulting gel molar composition is 1/0.017/30/6/1.8 Al3+:Pluronic P123:R− OH:H2O:HCl. In a typical experiment, approximately 170 μmol (0.986 g) of Pluronic P123 was dissolved in 150 mmol of alcohol (e.g., 14 mL for s-BuOH) and stirred for 15 min at the corresponding synthesis temperature. In a second solution, 18 mmol of HCl (1.49 mL of HCl 37 wt.%) and 150 mmol of alcohol were mixed at room temperature; the necessary water was supplied by the HCl source. Aluminum tri-sec-butoxide (10 mmol, 2.54 g) was gradually added to the latter solution under vigorous stirring. After 15 min the two solutions were mixed and additionally stirred at the corresponding synthesis temperature. The stirring speed in all pre-evaporation steps was set at 100 rpm. The homogeneous sol was left 3 days at the synthesis temperature under N2 flow (30 mL STP/min) resulting in a white gel slurry. The condensation was induced by evaporating the solvent at the corresponding synthesis temperature by increasing the nitrogen flow (200 mL STP/min) and the stirring speed at 250 rpm until dryness. Thermal Activation. The resulting dry materials were calcined at 823, 923, 973, 1023, and 1173 K in a conventional oven employing a mild heating rate of 1 K/min and an intermediate step at 673 K for 4 h to decompose the organic template, based on the TGA patterns in Figure S-1 (Supporting Information). This was followed by a second step at 1 K/min up to the final temperature, which was maintained for 4 h, after which the sample was cooled down. Table 1 summarizes the synthesis variables and materials nomenclature.

order and crystalline walls was implemented by nanocasting SBA-15 silicas as hard template;35 yielding surface areas of ca. 400 m2/g at 823 K. More recently, a bimodal mesoporous carbon has been used36 as well as polyurethane; this gives rise to a combined meso- and macroporous structure.42 Softtemplating has also shown promising results by yielding poreuniform aluminas mesophases, which by definition involves a more direct pathway. Niesz et al.19 reported the first softtemplated 2D hexagonal mesostructure with p6mm space group, combining the evaporation-induced self-assembly (EISA)43,44 and (S0H+)(X−I+) matching45,46 of an Al alkoxide derived gel. Chemically, aluminum tri-tert-butoxide is acidhydrolyzed in the presence of block copolymers as soft templates with a limited amount of water in ethanol as solvent. The material showed organized pores and a relatively wide pore size distribution; despite the fact that no XRD structural data was reported, the walls must still be amorphous. Yuan et al.,38 introduced an upgrading route to make the walls crystalline, which has been confirmed by others.39,41 The addition of carboxylic acids (citric, tartaric, and DL-malic acids) retards the hydrolysis by blocking reactive sites by chelation. This created ordered structures with narrow pores. However, crystalline walls were only observed at 1073 K; unfortunately, this produced a reduction of the surface area from 434 to 226 m2/g. Hence, it is a challenge from the synthetic point of view to investigate other routes to promote further thermal stability. Niesz’s route19 makes use of the EtOH as solvent; the role of the solvent has not been investigated afterward. Although ethanol is a common solvent in EISA for silica-based precursors, it is perhaps not the most suitable solvent when hydrolyzing complex Al-alkoxidestri-tert-butoxide, triisopropoxide, or tri-sec-butoxidebecause of its reactivity. Exchange reactions between EtOH with the alkoxo groups can take place:

Table 1. Synthesis Variables and Texture of the Resulting Mesophases

M(OR)n + m EtOH → [M(OR)n − m (OEt)m ] + mROH (1)

The more accessible ethoxo ligands facilitate the attack of nucleophiles, thereby increasing the hydrolysis rate; precipitates can be formed instead of gels. When the exchange reaction is not complete, ill-defined alkoxides are formed, a situation that promotes the anisotropic material growth. We think that the solvent can play a major role because in the upgraded route38,39,41 carboxylic chelating agents are employed with EtOH; possibly this avoids the in situ formation of Al-ethoxide as well. In this work, the solvent effect for the synthesis and activation of EISA mesoporous aluminas has been investigated. As a result, an effective upgrading strategy has been found where the solvent itself can provide stability to the alkoxide in the hydrolysis. When the crystallization takes place, the material retains the mesoscopic order, it is highly porous and the pores are uniform; thus crystallization of the walls occurs without structural collapse. A rationalization on the synthesis conditions, pore structure, and particle coarsening is proposed.

material

solvent, T (K)

SBET (m2/g)

VT (cm3/g)

WBJH (Å)

OM-1 OM-2 OM-3 OM-4 OM-5 OM-6 OM-7 OM-8 OM-9 OM-10 COM

EtOH, 313 EtOH, 333 EtOH, 353 s-BuOH, 313 s-BuOH, 333 s-BuOH, 353 t-BuOH, 313 t-BuOH, 333 t-BuOH, 353 EtOH, 333

321 329 336 336 419 432 356 331 156 388 141

0.652 0.534 0.546 0.480 0.685 0.397 0.548 0.508 0.829 0.649 0.254

80 51 55 55 48 34 54 67 289 48 63

The term organized mesophase (OM) is employed for materials calcined at 823 K that still has an amorphous structure. Occasionally, the calcination was carried out at 673 K to compare textural data with reported values; when applicable, this temperature will be indicated. Characterization. Small-angle X-ray scattering (SAXS) measurements were carried out at room temperature using a Bruker NanoStar instrument. A ceramic fine-focus X-ray tube, powered with a Kristallflex K760 generator at 35 kV and 40 mA, has been used in point focus mode. The primary X-ray flux is collimated using crosscoupled Göbel mirrors and a pinhole of 0.1 mm in diameter providing a Cu Kα radiation beam with a full width at half-maximum of about 0.2 mm at the sample position. The sample−detector distance was 1.04 m. The scattering intensity was registered by a Hi-Star position-sensitive area detector (Siemens AXS) in the q-vector range of 0.1−2.0 nm. Thermogravimetric analysis (TGA) was carried out on a MettlerToledo TGA/SDTA851e using synthetic air (100 mL STP/min) at 10 K/min from 303 to 1173 K.

2. EXPERIMENTAL SECTION Chemicals. Pluronic P123 (Mav = 5800, EO20PO70EO20), aluminum tri-sec-butoxide (97%), oxalic acid anhydrous (≥99.0%), citric acid monohydrate (≥99.0%), p-toluenesulfonic acid monohydrate (≥98.5%), from Sigma-Aldrich; sec-butyl alcohol (99%), from Acros Organics; and hydrochloric acid (37 wt %), ethanol (99%), tertbutyl alcohol (99%), and an ultrapure commercial γ-alumina (denoted as COM), from Merck, were used in this study. 849



Article

Nitrogen physisorption analyses (77 K) were carried out in a Micromeritics ASAP 2420. The samples were degassed in vacuum at 623 K for 10 h. The surface area was calculated using the standard BET method (SBET).47 The single point pore volume (VT) was estimated from the amount of gas adsorbed at a relative pressure of 0.98 in the desorption branch. The pore size distributions (PSD) were obtained from the BJH method48 using the adsorption branch of the isotherms (WBJH), while the t-plot method was employed to quantify the micropore volume (Vμ).49 TEM images were taken on a JEOL 2010F FEG transmission electron microscope at magnifications from 800 000 up to 15 000 000. The sample was suspended in 2-propanol and then deposited on a holey carbon grid until dryness. Powder X-ray diffraction (XRD) measurements were done on a Bruker D8 powder X-ray diffractometer using Cu Kα radiation, λ = 1.540 56 Å. The spectra were recorded with a step size of 0.02° for 3 s accumulation time, in the 2θ angle range of 10°−80°. 27 Al magic angle spinning nuclear magnetic resonance (MAS NMR) measurements were conducted on Bruker Avance-400 spectrometer using a 4 mm zirconium holder, applying a spinning frequency of 11 kHz at 298 K. The 27Al MAS NMR spectra were obtained at 104.201 MHz, with acquisition delay of 1 s and acquisition time of 0.08 s. Typically 4000 scans were collected. The spectra were referenced with respect to 1.0 M aqueous solution of Al(NO3)3 assigned on 0 ppm. Liquid 1H and 13C NMR spectra were recorded in a Varian AMX400 instrument at 399.93 MHz for 1H and 100.59 MHz for 13C. Chemical shifts are reported in δ units (ppm) and referenced to tetramethylsilane (TMS). The splitting patterns are designated as follows: s (singlet), d (doublet), t (triplet), q (quartet), and m (multiplet). A qualitative amount of aluminum tri-sec-butoxide was dissolved in EtOH-d6 (Acros, anhydrous, 99% atom D) in a vacuum glass line using modified NMR tubes. This allows one to work under moisture-free conditions. After 2 h, an excess of sec-butanol was added to the previous mixture to verify the presence of sec-butanol. The water content was determined by Karl Fischer titration using a Metrohm 702 SM instrument and the Hydranal products (SigmaAldrich Biotechnology).

narrow PSDs compared to the commercial material. However, s-BuOH achieves the narrowest PSDs; OM-5 has a relatively steep capillary condensation associated with the pore uniformity. The degree of pore uniformity was quantified by the full width at half-maximum (fwhm) of the BJH adsorption branch (Figure 1); it shows that the pore uniformity is clearly

Figure 1. Pore uniformity of the OMs measured as full width at halfmaximum (Å) from the BJH adsorption PSD (Figure S-2b,d,f, Supporting Information).

influenced by the alcohol and temperature applied in the EISA process. With difference, s-BuOH at 333 K shows the optimal mesophase in terms of regularity of the pores, followed by 313 K. The presence of mesoscopic ordering was characterized by small-angle X-ray diffraction (Figure 2). The s-BuOH-based material synthesized at 313 K was fairly amorphous (not shown). At 333 K a well-defined diffraction peak at ca. 1.5° was identified, which notably broadened at 353 K. The featureless 110 and 200 diffractions point out the absence of 2D hexagonal

3. RESULTS AND DISCUSSION Impact of the Alcohol in the EISA Synthesis of Alumina Mesophases. A systematic study varying the type of alcohol and temperature was carried out; aluminum tri-secbutoxide and the acid type were kept constant with Al/HCl = 1.8 and Al/H2O = 6. The nitrogen adsorption isotherms of the mesophases are given in Figure S-2 (Supporting Information), including a reference γ-alumina. The isotherms were generally of type IV with hysteresis H1, representing solids with cylindrical pore geometry with relatively high pore size uniformity and facile pore connectivity.50 H2 hysteresis loops were occasionally observed, associated with pores with narrow mouths. Ethanol-based mesophases evidenced isotherms with IV−H1 shape at all temperatures, while s-BuOH- and t-BuOHderived mesophases presented hysteresis H2 as well, at 353 and 333 K, respectively. The textural parameters are given in Table 1; in all cases the texture is truly mesoporous (Vμ < 0.005 cm3/ g). Comparison reveals more evident differences among the alcohols. At 313 K the BET surface areas are very similar for the three alcohols; at 333−353 K, materials from s-BuOH yield notably higher surface areas than those from EtOH and tBuOH. The optimal SBET was found for the mesophase synthesized under s-BuOH at 333K (OM-5), resulting in 419 m2/g with 0.685 cm3/g. Equally interesting is the impact on the pore size distribution (PSD) (Figure S-2, Supporting Information). t-BuOH-derived mesophases exhibit broad PSDs with large differences in size with the synthesis temperature. Ethanol produced relatively

Figure 2. SAXS patterns (left) of s-BuOH-based OMs after calcination at 823 K: (a) OM-6 (353 K) and (b) OM-5 (333 K). SEM picture for the OM-5 (right, top) as well as TEM (bottom and right middle). 850



Article

p6mm symmetry. This was confirmed by TEM analysis (Figure 2, bottom and right-middle) that revealed a regular wormhole pore structure in nanoagglomerates (SEM, Figure 2 right-top). TEM pictures covering broader regions for OM-5 confirm the presence of such wormhole pore structure throughout the particle (Figures S-3 and S-4, Supporting Information). The narrowest mesophases obtained with EtOH (OM-2) and t-BuOH (OM-7) were analyzed by SAXS and TEM. Figure S-5 (Supporting Information) shows an overview picture of OM-2; two types of particle morphologies were observed with wormhole and fiber-type domains and mixtures thereof. The SAXS pattern (inset in Figure S-5, Supporting Information) shows a weak 100-related reflection due to the less regular particle-to-particle distance regarding the sample heterogeneity. OM-7 (Figure S-6, Supporting Information) is composed by a random distribution of particles of different morphologies and sizes, with a low-intensity SAXS pattern as well (inset in Figure S-6, Supporting Information). Regarding OM-5, combining the textural analysis, SAXS, and TEM results, it is concluded that this optimal mesophase possesses a disordered network (not being 2D hexagonally packed) with highly uniform mesopores. Thus, having a highly ordered structure is not a prerequisite to obtain high surface areas (and thermal stability, as discussed later); OM-5 (498 m2/ g, calcined at 673 K) is comparable to the highly structured hexagonal mesophases reported by Yuan et al.38 (434 m2/g, 673 K), Morris et al.39 (228 m2/g, 673 K), Wu et al.41 (275 m2/ g, 673 K), and Li et al. (300 m2/g, 823 K).42 Consequently, the regularity of the pores (or high particle coordination) and degree of condensation (both discussed later) appear to play the major roles in obtaining high surface areas and stability. Synthesis Reproducibility. The reproducibility of the synthesis for the optimal mesophase was evaluated by comparing four independent batches that were calcined at 823 K. The results of N2 gas adsorption are given in Figure 3

upgraded EISA synthesis: controlling the hydrolysis rate and maintaining the medium acidic during the evaporation to keep the hydrolysis rate unaltered during the self-assembly process. We verified whether the addition of carboxylic acids is beneficial in the synthesis of the optimal mesophase, OM-5. Addition of citric and oxalic acids was investigated in two manners at acid/Al = 0.25 and HCl always present: addition directly from the beginning of the synthesis or before the evaporation. Figures S-7 and S-8 (Supporting Information) show the N2 adsorption isotherms of these mesophases. In none of these modifications was observed a beneficial influence of the carboxylic acid in terms of surface area or pore volume (Table S-2, entries 2−13, Supporting Information). The isotherms have IV−H1 or −H2 hysteresis; the steepness of the capillary condensation nevertheless decreases substantially, which is reflected in a reduced total pore volume. In a second systematic study, the chelation and volatility effects were decoupled by using a nonchelating and nonvolatile acid, ptoluenesulfonic (p-TSAM). The concentration was screened slightly wider at p-TSAM/Al = 0.10, 0.25, and 1.00. The textural properties of the mesophases are given in Figure S-9 and Table S-2 (entries 14−21) (Supporting Information). It can be clearly deduced that p-TSAM also provides a negative impact on the surface area and total pore volume. It was generally observed that when these materials are thermally treated from 1023 K upward the collapse was even more pronounced (Figures S-7−9 and Table S-2, Supporting Information) with a substantial reduction of the surface area, pore volume, and PSD broadening. These results suggest two mechanistic insights about our upgrading route: the chelation effect does not play a role in the presence of a stable alcohol (s-BuOH) or the nonvolatility of the acid, i.e., low pH, provokes an undesired effect during the evaporation as the material loses thermal stability during the calcination, or both. It appears that the gel does not condense adequately under the acidic evaporation. It is known that extreme pHs decrease the rate of condensation reactions.51 Therefore, it seems that before the evaporation the hydrolysis is completed and the low pH avoids the proper condensation of the network. When using HCl, its evaporation increases the pH, promotes the condensation, and provides stability to the material. From this, the self-assembly seems to be essentially a condensation process; cross-linked aluminum species can arrange around the triblock copolymer to form the ordered mesophases. It can be suggested that condensation can occur when using nonvolatile carboxylic acids;38 this can be induced by the esterification of the carboxylic acids with EtOH, leading to a pH increase. For our approach, it is evident that the addition of nonvolatile acids produces thermally unstable structures involving a different upgrading mechanism to that reported by Yuan et al.38 Crystallization and Thermal Stability. The crystallization temperature and stability were evaluated in order to appreciate the practical benefits of this upgrading route. The optimal mesophase OM-5 was calcined at increasingly higher temperatures from 673 up to 1173 K, and the textural properties were compared. The shape of the N2 adsorption isotherms remains unchanged upon heating with a reduction of the steepness of the capillary condensation. This trend changes notably only at 1073 K, which represents a breaking point after which the structure collapses at 1173 K. The thermal stability becomes more obvious when the BET surface areas are represented as a

Figure 3. Reproducibility of the OM-5 synthesis: (a) nitrogen sorption isotherms and (b) BJH PSDs of the adsorption branch for four independent synthesis batches after calcination at 823 K.

and Table S-1 (Supporting Information). The isotherms and BJH PSDs indicate that the reproducibility is remarkably high. This was quantified by means of the standard deviations in the BET surface area (σBET), position of the maximum of the BJH PSD (σBJH), and the total pore volume (σVT). These three parameters account for the reproducibility at the three porosity scales: low mesopores and multilayer adsorption (σBET), medium pores (σBJH), and the external texture (σVT). The calculated standard deviations are notably low, σBET = 1.9%, σBJH = 2.8%, and σVT = 4.1%), which proves the high reproducibility of the synthesis. Role of Nonvolatile Acids. According to Yuan et al.,38 nonvolatile carboxylic acids have two functions in their 851



Article

most relevant diffraction peaks for the (440) and (400) crystallographic planes as well as an unresolved contribution of the (220), (311), and (222) planes. SAXS analysis (Figure 5a) reveals that mesoscopic ordering is well-maintained at 1023 K, a result that agrees with the narrow pore size distribution with a small shift of the pore maximum from 42 to 46 Å. The γ-phase becomes more crystalline at higher temperature, and no other transitional aluminas were observed. The conversion from amorphous into γ-alumina was confirmed by solid-state 27Al MAS NMR (Figure 5c). The starting mesophase showed three broad resonances located at ca. 0, 35, and 70 ppm associated with octa- (AlO6), penta- (AlO5), and tetrahedrally (AlO4) coordinated Al3+.52−54 The broadness is typical for Al nuclei having strong quadrupolar side effects. The high concentration of AlO5 at 823 K is attributed to the large number of crystallographic cationic defects associated with the high surface area.55,38 The crystallization at higher temperature is accompanied by a drop in AlO5 and a concomitant increase of AlO4 (ca. 40%) and AlO6 (ca. 60%), which is consistent with the dehydration of the surface. The acidic properties were evaluated by temperature-programmed desorption (TPD) of NH3. The NH3-TPD profiles, shown in Figure S-10 (Supporting Information), present two distinct broad desorption peaks centered at 523 and 873 K, attributed to weakly and strongly adsorbed NH3. The shape of the profiles coincides with the commercial alumina; the higher density of acidic sites in the 1023 K sample comes from a higher percentage of active surface area. The narrowest mesophases obtained with EtOH (OM-2) and t-BuOH (OM-7) were subjected to thermal activation at 1023 K as well and compared to that obtained with s-BuOH (OM-5). XRD, SAXS, and PSDs results are given in Figure 6, sorption isotherms are in Figure S-11 (Supporting Information), and the derived textural parameters are compiled in Table S-4 (Supporting Information). γ-Alumina is obtained in all cases (Figure 6a); the crystallinity varies due to the sintering or coarsening degree of the nanoparticles in the order OM-7 >

function of the calcination temperature (Figure.4b); the BET values slightly decrease until the collapse at 1173 K. The PSDs

Figure 4. OM-5 (s-BuOH/333 K) calcined at high temperatures: (a) nitrogen isotherms and PSD and (b) histogram of the BET surface areas and pore volumes. Raw data are given in Table S-3 (Supporting Information).

(Figure 4a) remain narrow with the calcination temperature up to 1023 K with no substantial pore enlargement, having welldefined cylindrical pores of 46 Å with 300 m2/g. The crystallization was evaluated by XRD; its evolution as a function of the calcination temperature is given in Figure 5b. The material is amorphous up to 973 K; the crystallization of the walls into γ-alumina takes place at 1023 K, showing the

Figure 6. Optimal mesophases activated at 1023 K: (a) XRD patterns, (b) SAXS patterns, and (c) BJH pore size distributions (dashed lines correspond to the mesophase counterparts calcined at 823 K) for (1) OM-5, (2) OM-2, and (3) OM-7.

Figure 5. (a) Small- and (b) wide-angle XRD patterns and (c) solidstate 27Al MAS NMR spectra of the OM-5 calcined at 673 K (0), 823 K (1), 923 K (2), 973 K (3), 1023 K (4), 1073 K (5), and 1173 K (6). 852



Article

Figure 7. SAXS and TEM images of the mesophases (823 K, left) and after thermal activation at 1023 K (right): (a) OM-7, (b) OM-2, and (c) OM5.

formation (bp 324 K) or the dehydration of t-BuOH to produce di-tert-butyl ether57 or both. The first pathway can justify only part of the water formation with ca. 1.4 wt % H2O. However, under the synthesis conditions, the dehydration equilibrium is easily shifted, as the water produced is continuously consumed in the hydrolysis. The amount of water produced by dehydration can then be high, and an acceleration of the hydrolysis rate is expected. Highly polymerized Al species or precipitates may not assemble well around the block copolymers to form an ordered mesophase. This is deduced from Figure 7a (left) that shows relatively large and disordered particles with different morphologies, as deduced from the SAXS and TEM pictures. Such assembly has a low particle coordination, i.e., low packing density, that favors coarsening rather than densification (sintering)58 upon thermal treatment at 1023 K, with a consequent broadening of the pore size distribution (as seen in Figures 6c and 3). Particle coarsening is observed in Figure 7a (right), where particles are bonded to form long-chain open pore structures. Hence, a fast hydrolysis in t-BuOH gives rise to poorly assembled Al species with low particle coordination, which provokes particles coarsening, resulting in an open pore structure at 1023 K. Ethanol could also be dehydrated by mineral acids, but higher temperatures (typically ≥453 K) are normally needed.59 It can alternatively participate in the exchange reaction with the alkoxide to produce substituted or pure Al ethoxide, which is more reactive than the starting sec-butoxide. Indeed, alcohol exchange reactions are employed to prepare different alkoxides under reflux.60,61 It is not evident, nevertheless, that these exchange reactions can occur at our synthesis temperature, i.e. 313−353 K. To shed light on this, control experiments were carried out in which aluminum tri-sec-butoxide was mixed with deuterated ethanol at room temperature in inert atmosphere. The 1H and 13C H-decoupled NMR spectra of the liquid product after 2 h reaction are given in Figures S-13a,c (Supporting Information). The spectra shows pure s-BuOD62 coming from the exchange reaction between ethanol and aluminum tri-sec-butoxide (Table S-6, Supporting Information, compiles the chemical shifts). To confirm this, the liquid product was peaked with external s-BuOH (cf. spectra a,b and c,d, in Figure S-13, Supporting Information). The comparison evidences the complete matching of the compound released

OM-2 > OM-5. Mesoscopic ordering is lost when using EtOH and t-BuOH (Figure 6b), in agreement with the much broader pore size distribution (Figure 6c) and lower surface areas. Consequently, well-defined mesopores with mesoscopic crystalline walls and high surface area are obtained with sBuOH only. Figure 6c also compares the PSDs of the corresponding mesophase counterparts (823 versus 1023 K). For t-BuOH the PSD dramatically broadens and shifts from 54 to 193 Å. Ethanol-derived mesophases widen substantially into a bimodal distribution centered at 62 and 95 Å. On the other hand, s-BuOH virtually retains the size (42−46 Å) with a reduction of the PSD maxima associated with the densification of the network. This comparison evidenced that the solvent notably influences the thermal stability and is associated with their inertness during the sol−gel reactions, as discussed in the next section. Crystallized OM-5 is compared with equivalent EISA-based γ-Al2O3 materials38,39 as well as with commercial products56 (Figure S-12, Supporting Information). It appears that our upgrading route achieves relatively higher surface areas, and this effect is not due to a smaller pore size (cf. pore sizes in Figure S-12, Supporting Information) but is possibly attributed to higher thermal stability. It is worth discussing that the surfactant-induced aluminas28−32,40 normally give rise to higher surface areas. This can be explained by two effects: better mesoscopic preservation, as attributed to the lower crystallization temperatures required, and their wider pore size distribution. A substantial fraction of small mesopores contributes most to the surface area. The EISA approaches produce, on the other hand, different type of mesoporous γAl2O3 with higher thermal stability and narrower pores. Solvent Impact on the Self-Assembly and Thermal Stability. Alcohols are normally employed in sol−gel to solubilize the alkoxide, but their inertness has been scarcely investigated. Concerning t-BuOH, we conducted an experiment in which it was heated to the corresponding synthesis temperature in the presence of HCl, employing the same ratio as in the synthesis conditions. We observed the release of water, with a concentration starting from 4.50 wt % increased up to 4.97−6.83 wt % (Table S-5, Supporting Information). The increase is moderate and it points out to an equilibrium boundary. Such release of water comes from either t-BuCl 853



Article

4. CONCLUSIONS The applied solvent influences the stability of EISA alumina mesophases based on aluminum tri-sec-butoxide. It can negatively dehydrate under the sol−gel synthesis conditions (t-BuOH), react with HCl (t-BuOH), or exchange, leading to a more reactive alkoxide (EtOH); all these situations will increase the rate of the hydrolysis. Particle heterogeneity with random packing of fibrous and wormhole morphologies attributed to the high hydrolysis rate was observed, in mesophases derived from both solvents. Such a low particle coordination favors coarsening with enlargement of the pore size distribution upon thermal treatment, explaining the lower thermal stability. The solvent can, however, provide stability to the alkoxide (sBuOH). The controlled hydrolysis and formation of lowpolymerized Al species in s-BuOH are likely responsible for the adequate assembly onto the surfactant; experimentally, a regular distribution of relatively size-uniform nanoparticles was observed in the mesophase. This high particle coordination prevents coarsening and favors densification upon thermal treatment, hence maintaining a relatively uniform pore size distribution. Besides the alcohol, the acid removal in the evaporation is another key factor in this route to promote the network condensation. As a result, a stable mesoscopic crystallized γ-Al2O3 with wormhole ordering having 300 m2/ g, 0.450 cm3/g, and an increased acidic site density was obtained.

with s-BuOH. Ethanol will exchange even faster with aluminum tri-sec-butoxide at synthesis temperature (313−353 K) in the presence of a mineral acid. Consequently, Al-triethoxide is formed at the beginning of the synthesis. Taking into account that the hydrolysis rate decreases with the steric bulk of the alkoxo ligands,51,60,61,63 the formation of an Al-triethoxide will very possibly increase the hydrolysis rate. Figure 7b (left) shows particle heterogeneity with random packing of fibrous and wormhole morphologies, which can be ascribed to the fast formation of highly polymerized Al species that do not assemble adequately onto the block copolymer. The packing coordination is low and this provokes coarsening58 upon thermal treatment at 1023 K (Figure 7b, right); the SAXS scattering pattern almost disappears and the pore size distribution broadens (Figures 6c and 2). These coarsened particles are on average smaller than OM-7 (t-BuOH, Figure 7a, right), which is consistent with the relatively narrower pore size distribution. When comparing the various alcohols, it must be pointed out that the molar ratio was kept constant, which implies using different volumes of EtOH (18 mL/mol Al) and the butanols (28 mL/mol Al). Thus the partial concentrations of the active species (Al3+, P123, H2O, and HCl) are higher when using EtOH. This can affect the hydrolysis rate as well; we conducted an additional synthesis whereby the EtOH volume was kept identical to those of the butanols (i.e., 28 mL/mol Al). The mesophase (OM-10) was analyzed by gas adsorption. The PSD does not show differences with regard to the concentration effect (cf. OM-2 and OM-10 in Figure S-14, Supporting Information). It appears that the exchange reaction and modification of the alkoxide reactivity play major roles in the type of polymerized Al species. The solvent can intrinsically affect the reactivity of the alkoxides by the coordination extension effect, namely, by solvate formation or changing the oligomerization degree by alkoxo bridging. Examples of these effects have been discussed for alkoxides mixtures64 and Si-,63 Ti-,63 Zr-, V-, Ta-, and Cealkoxides.65 Concerning Al-alkoxides, although their chemical modification has been reported,66,67 the solvent effects on the sol−gel products have not been documented to the best of our knowledge. From the coordination standpoint, s-BuOH stabilizes the trimers−tetramers of ATSB by breaking the bridging of the OR groups, and the degree of oligomerization is limited.68 Therefore, we can assume that such stabilization of the alkoxide controls the hydrolysis rate under sol−gel hydrothermal conditions. TEM pictures in Figure 7c (left) show the pores formed by a relatively regular assembly of particles where the packing density is high, which gives rise to a relatively intense 100related SAXS reflection. Sintering theory58 indicates that under such circumstances densification is favored instead of coarsening. Indeed, after calcination at 1023 K the SAXS reflection is maintained, the pore structure remains wormhole, and no coarsened particles were found (TEM, Figure 7c, right). This is in agreement with the near lack of pore size broadening upon thermal treatment (42−46 Å in Figure 6c, 1); the lower intensity is ascribed to a densification of the material. Therefore, it appears that the controlled hydrolysis and formation of likely low-polymerized Al species in s-BuOH are responsible of the adequate assembly onto the surfactant. This produces a regular organization of small particles that are resistant to coarsening; the effect that this has on the pore size distribution is very important.

■

ASSOCIATED CONTENT

S Supporting Information *

Thermogravimetric analysis of the precursors; textural properties of the mesophases (isotherms and pore size distributions); TEM pictures for OM-5, OM-2, and OM-7; synthesis reproducibility (textural parameters); acid influence (citric, oxalic, and p-toluenesulfonic in combination with HCl) on the mesophases texture (isotherms, pore size distributions, and textural parameters); thermal stability (textural parameters); NH3-TPD patterns; activation of the optimal mesophases (isotherms, pore size distributions and textural parameters); textural comparison with reference materials; H2O content derived from t-BuOH dehydration; 1H and 13C NMR spectra and resonances for EtOH-d6 exchange study; and effect of the alcohol volume. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by Shell Global Solutions International B. V. under project No. 51008532. Dr. Marcello Rigutto and Dr. Marije Nijkamp (Shell) are thanked for fruitful discussions. L.L.P. acknowledges personal grants from Universidad Autónoma Metropolitana (México DF) and the National program PROMEP (103.5/07/1961) Mexico. Dr. Patricia Kooyman (TUD) is acknowledged for the supplementary TEM investigation during the revision. 854



■

Article

(35) Liu, Q.; Wang, A. Q.; Wang, X. D.; Zhang, T. Chem. Mater. 2006, 18, 5153. (36) Wu, Z. X.; Li, Q.; Feng, D.; Webley, P. A.; Zhao, D. Y. J. Am. Chem. Soc. 2010, 132, 12042. (37) Park, H.; Yang, S. H.; Jun, Y. S.; Hong, W. H.; Kang, J. K. Chem. Mater. 2007, 19, 535. (38) Yuan, Q.; Yin, A. X.; Luo, C.; Sun, L. D.; Zhang, Y. W.; Duan, W. T.; Liu, H. C.; Yan, C. H. J. Am. Chem. Soc. 2008, 130, 3465. (39) Morris, S. M.; Fulvio, P. F.; Jaroniec, M. J. Am. Chem. Soc. 2008, 130, 15210. (40) Zhang, Z.; Pinnavaia, T. J. Langmuir 2010, 26, 10063. (41) Wu, Q. L.; Zhang, F.; Yang, J. P.; Li, Q.; Tu, B.; Zhao, D. Y. Microporous Mesoporous Mater. 2011, 143, 406. (42) Li, L. L.; Duan, W. T.; Yuan, Q.; Li, Z. X.; Duan, H. H.; Yan, C. H. Chem. Commun. 2009, 6174. (43) Brinker, C. J.; Lu, Y. F.; Sellinger, A.; Fan, H. Y. Adv. Mater. 1999, 11, 579. (44) Grosso, D.; Cagnol, F.; Soler-Illia, G. J. D. A.; Crepaldi, E. L.; Amenitsch, H.; Brunet-Bruneau, A.; Bourgeois, A.; Sanchez, C. Adv. Funct. Mater. 2004, 14, 309. (45) Zhao, D. Y.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G.; Chmelka, B.; Stucky, G. Science 1998, 279, 548. (46) Zhao, D. Y.; Huo, Q.; Feng, J.; Chmelka, B.; Stucky, G. J. Am. Chem. Soc. 1998, 120, 6024. (47) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309. (48) Barret, E. P.; Joyner, L. G.; Hallenda, P. H. J. Am. Chem. Soc. 1951, 73, 373. (49) Lippens, B. C.; de Boer, J. H. J. Catal. 1965, 4, 319. (50) Kruk, M.; Jaroniec, M. Chem. Mater. 2001, 13, 3169. (51) Brinker, C. J.; Scherer, G. W. Sol-Gel Science. The Physics and Chemistry of Sol-Gel Processing; Academic Press: San Diego, 1990. (52) Chen, F. R.; Davis, J. G.; Fripiat, J. J. J. Catal. 1992, 133, 263. (53) Kentgens, A. P. M. Geoderma 1997, 80, 271. (54) Duvel, A.; Romanova, E.; Sharifi, M.; Freude, D.; Wark, M.; Heitjans, P.; Wilkening, M. J. Phys. Chem. C 2011, 115, 22770. (55) Chupas, P. J.; Grey, C. P. J. Catal. 2004, 224, 69. (56) Düuvel, A.; Romanova, E.; Sharifi, M.; Freude, D.; Wark, M.; Heitjans, P.; Wilkening, M. J. Phys. Chem. C 2011, 115, 22770. (57) Knifton, J. F.; Sanderson, J. R.; Stockton, M. E. Catal. Lett. 2001, 73, 55. (58) German, R. M. Sintering Theory and Practice; John Wiley & Sons, Inc.: New York, 1996. (59) Takahara, I.; Saito, M.; Inaba, M.; Murata, K. Catal. Lett. 2005, 105, 249. (60) Bradley, D. C.; Mehrotra, R. C.; Gaur, D. P. Metal Alkoxides; Academic Press: New York, 1978. (61) Bradley, D. C.; Mehrotra, R. C.; Rothwell, I. P.; Singh, A. Alkoxo and Aryloxo Derivatives of Metals; Academic Press: London, 2001. (62) Jones, I. C.; Sharman, G. J.; Pidgeon, J. Magn. Reson. Chem. 2005, 43, 497. (63) Sanchez, C.; Livage, J.; Henry, M.; Babonneau, F. J. Non-Cryst. Solids 1988, 100, 65. (64) Okuyama, M.; Fukui, T.; Sakurai, C. J. Non-Cryst. Solids 1992, 143, 112. (65) Nabavi, M.; Doeuff, S.; Sanchez, C.; Livage, J. J. Non-Cryst. Solids 1990, 121, 31. (66) Wengrovius, J. H.; Carbauskas, M. F.; Williams, E. A.; Going, R. C.; Donahue, P. E.; Smith, J. F. J. Am. Chem. Soc. 1986, 108, 982. (67) Lichtenberger, R.; Schubert, U. J. Mater. Chem. 2010, 20, 9287. (68) Peterlik, H.; Rennhofer, H.; Torma, V.; Bauer, U.; Puchberger, M.; Huesing, N.; Bernstorff, S.; Schubert, U. J. Non-Cryst. Solids 2007, 353, 1635.

REFERENCES

(1) Misra, C. Industrial Alumina Chemicals; ACS Monograph 184; American Chemical Society: Washington, DC, 1986. (2) Cejka, J. Appl. Catal., A 2003, 254, 327. (3) Márquez-Alvarez, C.; Zilkova, N.; Perez-Pariente, J.; Cejka, J. Catal. Rev. Sci. Eng. 2008, 50, 222. (4) Topsoe, H.; Clausen, B. S.; Massoth, F. E. Hydrotreating Catalysis, Science and Technology; Springer: Berlin, 1996. (5) Yoldas, B. E. Amer. Ceram. Soc. Bull. 1975, 54, 286. (6) Knudsen, K. G.; Cooper, B. H.; Topsøe, H. Appl. Catal. A. Gen. 1999, 189, 205. (7) Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106, 4044. (8) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710. (9) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C.T-W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenkert, J. L. J. Am. Chem. Soc. 1992, 114, 10834. (10) Bagshaw, S. A.; Prouzet, E.; Pinnavaia, T. J. Science 1995, 269, 1242. (11) Bagshaw, S. A.; Pinnavaia, T. Angew. Chem., Int. Ed. 1996, 35, 1102. (12) Vaudry, F.; Khodabandeh, S.; Davis, M. E. Chem. Mater. 1996, 8, 1451. (13) Zhang, W.; Pinnavaia, T. J. Chem. Commun. 1998, 1185. (14) Cabrera, S.; Haskouri, J. E.; Alamo, J.; Beltrán, A.; Betrán, D.; Mendioroz, S.; Marcos, M. D.; Amorós, P. Adv. Mater. 1999, 11, 379. (15) Cejka, J.; Zilkova, N.; Rathousky, J.; Zukal, A. Phys. Chem. Chem. Phys. 2001, 3, 5076. (16) González-Peña, V.; Márquez-Á lvarez, C.; Díaz, I.; Grande, M.; Blasco, T.; Pérez-Pariente, J. Microporous Mesoporous Mater. 2005, 80, 173. (17) Deng, W. H.; Toepke, M. W.; Shanks, B. H. Adv. Funct. Mater. 2003, 13, 61. (18) Tian, B. Z.; Yang, H. F.; Liu, X. Y.; Xie, S. H.; Yu, C. Z.; Fan, J.; Tu, B.; Zhao, D. Y. Chem. Commun. 2002, 1824. (19) Niesz, K.; Yang, P.; Somorjai, G. A. Chem. Commun. 2005, 1986. (20) Xu, B. J.; Xiao, T. C.; Yan, Z. F.; Sun, X.; Sloan, J.; GonzálezCortés, S. L.; Alshahrani, F.; Green, M. L. H. Microporous Mesoporous Mater. 2006, 91, 293. (21) Dacquin, J. P.; Dhainaut, J.; Duprez, D.; Royer, S.; Lee, A. F.; Wilson, K. J. Am. Chem. Soc. 2009, 131, 12896. (22) Cai, W. Q.; Yu, J. G.; Anand, C.; Vinu, A.; Jaroniec, M. Chem. Mater. 2011, 23, 1147. (23) Cai, W. Q.; Yu, J. G.; Jaroniec, M. J. Mater. Chem. 2011, 21, 9066. (24) Yada, M.; Kitamura, H.; Machida, M.; Kijima, T. Langmuir 1997, 13, 5252. (25) Yada, M.; Hiyoshi, H.; Ohe, K.; Machida, M.; Kijima, T. Inorg. Chem. 1997, 36, 5565. (26) Yada, M.; Ohya, M.; Machida, M.; Kijima, T. Chem. Commun. 1998, 1941. (27) González-Peña, V.; Díaz, I.; Márquez-Á lvarez, C.; Sastre, E.; Pérez-Pariente, J. Microporous Mesoporous Mater. 2001, 44, 203. (28) Zhu, H. Y.; Riches, J. D.; Barry, J. C. Chem. Mater. 2002, 14, 2086. (29) Zhang, Z.; Hicks, R. W.; Pauly, T. R.; Pinnavaia, T. J. J. Am. Chem. Soc. 2002, 124, 1592. (30) Zhang, Z.; Pinnavaia, T. J. J. Am. Chem. Soc. 2002, 124, 12294. (31) Hicks, R. W.; Pinnavaia, T. J. Chem. Mater. 2003, 15, 78. (32) Lee, H. C.; Kim, H. J.; Chung, S. H.; Lee, K. H.; Lee, H. C.; Lee, J. S. J. Am. Chem. Soc. 2003, 125, 2882. (33) Kuemmel, M.; Grosso, D.; Boissiere, C.; Smarsly, B.; Brezesinski, T.; Albouy, P. A.; Amenitsch, H.; Sanchez, C. Angew. Chem., Int. Ed. 2005, 44, 4589. (34) Boissiere, C.; Nicole, L.; Gervais, C.; Babonneau, F.; Antonietti, M.; Amenitsch, H.; Sanchez, C.; Grosso, D. Chem. Mater. 2006, 18, 5238. 855
