M2O3 catalysts (M=Al, Ga, In)

M2O3 catalysts (M=Al, Ga, In)

Applied Catalysis A: General 249 (2003) 1–9 Reduction of ␣,␤-unsaturated aldehydes with basic MgO/M2 O3 catalysts (M=Al, Ga, In) Mar´ıa A. Aramend´ıa...

147KB Sizes 0 Downloads 9 Views

Applied Catalysis A: General 249 (2003) 1–9

Reduction of ␣,␤-unsaturated aldehydes with basic MgO/M2 O3 catalysts (M=Al, Ga, In) Mar´ıa A. Aramend´ıa, Victoriano Borau, César Jiménez∗ , José M. Marinas, José R. Ruiz∗ , Francisco Urbano Departamento de Qu´ımica Orgánica, Universidad de Córdoba, Campus de Rabanales, Edificio C-3, Ctra Nacional IV-A, km 396, 14014 Córdoba, Spain Received 9 January 2003; received in revised form 20 February 2003; accepted 21 February 2003

Abstract An overall nine magnesium–aluminium, magnesium–gallium and magnesium–indium hydrotalcite-like compounds (HTlc) were obtained by coprecipitation, the sol–gel technique and precipitation with urea. A pure hydrotalcite phase was obtained in all instances; by exception, the indium products were a mixture of the corresponding HTlc and In(OH)3 . Products were characterized by X-ray diffraction (XRD) and their specific surface areas determined. Calcination in a nitrogen atmosphere at 500 ◦ C yielded a periclase MgO phase—plus In2 O3 from the products containing In(OH)3 . Calcined solids were characterized by XRD, and their specific surface areas and surface chemical properties determined (the latter by CO2 chemisorption). These solids were used as catalysts for the Meerwein–Ponndorf–Verley (MPV) reaction between trans-crotonaldehyde and isopropyl alcohol, and the one obtained by calcination of the Mg/Al HTlc prepared by coprecipitation—which was the most basic—was found to be the most active in the process. This catalyst was also tested on the MPV reduction of various ␣,␤-unsaturated aldehydes with isopropyl alcohol. © 2003 Elsevier Science B.V. All rights reserved. Keywords: Basic catalysts; Hydrotalcite; Hydrogen transfer; Meerwein–Ponndorf–Verley

1. Introduction The reduction of unsaturated organic compounds by use of an organic molecule instead of hydrogen gas or a metal hydride as hydrogen donor is known as “hydrogen transfer” and is well-documented [1]. Our group has conducted extensive research into this process and used various supported metal catalysts in the reduction of nitro and carbonyl compounds, and alkenes, with hydrogen donors of variable nature [2–5]. One other way of reducing carbonyl compounds by hydro∗ Corresponding authors. Tel.: +34-9572-18638; fax: +34-9572-12066. E-mail address: [email protected] (J.R. Ruiz).

gen transfer involves using a metal alkoxide as catalyst and an alcohol as hydrogen donor. This process, known as the Meerwein–Ponndorf–Verley (MPV) reaction, is highly selective; in fact, because it leaves C=C and similar double bonds untouched, it is especially suitable for reducing ␣,␤-unsaturated carbonyl compounds. This method, however, is subject to some restrictions arising, among others, from the need to use an alkoxide excess of at least 100–200% and to subsequently neutralize residual alkoxide in the medium with a strong acid. Finally, because the process takes place in a homogeneous phase, isolating the products is very often a labour-intensive, time-consuming task. The above-described problems have lately been circumvented by using catalysts of variable nature

0926-860X/$ – see front matter © 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0926-860X(03)00163-7

2

M.A. Aramend´ıa et al. / Applied Catalysis A: General 249 (2003) 1–9

(both acid and basic) in a heterogeneous phase. Such catalysts, which include zeolites [6,7], silica [8] and alumina [9], allow the process to be dramatically simplified; in fact, these reactions take place in the gas phase but require the use of high temperatures and flow-through reactors. Some basic catalysts, however, allow the MPV process to occur at refluxing temperatures in the liquid phase. Such is the case with the magnesium–aluminium oxide mixture obtained by calcining hydrotalcite-like compounds (HTlc) [10]. Recently, our group succeeded in reducing citral with isopropyl alcohol in the liquid phase, using various catalysts obtained by calcining HTlc [11] or magnesium and calcium oxides in conjunction with different secondary alcohols [12]. In this work, our group extended previous research into the MPV process in the liquid phase by using various basic catalysts that were obtained from Mg/Al, Mg/Ga and Mg/In HTlc via three different synthetic procedures (viz. coprecipitation, the sol–gel technique and precipitation with urea). Once calcined, the solids provided catalysts spanning a wide range of surface chemical properties. 2. Experimental 2.1. Catalyst synthesis The Mg/Al, Mg/Ga and Mg/In HTlc [Mg(II)/M(III) = 3] used as catalyst precursors were obtained by using three different methods, namely: coprecipitation (method 1), the sol–gel technique (method 2) and precipitation by hydrolysis of urea (method 3). The solids obtained using the coprecipitation method were reported by our group elsewhere [13–15]. By contrast, the sol–gel technique and precipitation with urea have scarcely been used in this context; while some Al-containing HTlc have been synthesized using these two methods [16,17], there appears to be no reference to similar Ga- or In-containing solids. This is thus the first reported instance of Ga and In HTlc obtained using the sol–gel technique or precipitation by hydrolysis of urea. The procedure used to prepare HTlc with the sol–gel technique was as follows: 0.15 mol of magnesium ethoxide was dissolved in ethanol containing a small amount of HCl (35% in water). Following

refluxing under continuous stirring, the solution was supplied with 200 ml of acetone containing 0.05 mol of aluminium, gallium or indium acetylacetonate. The pH of the mixture was adjusted to 10 with ammonia (33% NH3 in water) and the solution refluxed under continuous stirring until a gel was formed. The gel was then filtered off, washed several times with distilled water and dried at 110 ◦ C in a stove. The procedure followed to obtained HTlc by precipitation of urea involved adding solid urea to a solution containing the respective metal nitrates in a Mg(II)/M(III) ratio of 3. The transparent solutions thus obtained were heated at 100 ◦ C to precipitate the corresponding HTlc. Once synthesized, the HTlc were ion-exchanged with carbonate to remove intercalated ions between layers. The procedure used the nine HTlc involved and suspending them in a solution containing 0.345 g of Na2 CO3 in 50 ml of bidistilled, de-ionized water per gram of HTlc at 100 ◦ C for 2 h, after which each solid was filtered off in vacuo and washed with 200 ml of bidistilled, de-ionized water. The HTlc thus obtained were subjected to a second ion-exchange operation under the same conditions as the first and, finally, calcined in a nitrogen atmosphere at 500 ◦ C for 8 h, using a gradient of 1 ◦ C min−1 . Table 1 shows the nomenclature used to designate the HTlc and their calcination products. 2.2. Catalyst characterization X-ray diffraction (XRD) patterns were recorded on a Siemens D-5000 diffractometer using Cu K␣ radiation. Scans were performed over the 2θ range from 5◦ to 70◦ . Table 1 Nomenclature used to designate the HTlc and their calcination products HTlc

Calcined HTlc

Synthetic method

HT-Mg/Al-1 HT-Mg/Al-2 HT-Mg/Al-3 HT-Mg/Ga-1 HT-Mg/Ga-2 HT-Mg/Ga-3 HT-Mg/In-1 HT-Mg/In-2 HT-Mg/In-3

HT-Mg/Al-1-500 HT-Mg/Al-2-500 HT-Mg/Al-3-500 HT-Mg/Ga-1-500 HT-Mg/Ga-2-500 HT-Mg/Ga-3-500 HT-Mg/In-1-500 HT-Mg/In-2-500 HT-Mg/In-3-500

Coprecipitation Sol–gel Urea Coprecipitation Sol–gel Urea Coprecipitaci´on Sol–gel Urea

M.A. Aramend´ıa et al. / Applied Catalysis A: General 249 (2003) 1–9

The textural properties of the solids were established from nitrogen adsorption–desorption isotherms at liquid nitrogen temperature, which were recorded on a Micromeritics ASAP-2010 instrument. Specific surface areas were determined using the BET method [18]. The amount of CO2 chemisorbed on each solid was measured on a Micromeritics 2900 TPD/TPR analyser. Prior to analysis, samples were heated at 500 ◦ C in an argon stream for 1 h. Measurements were made at room temperature by alternately passing argon, and the same gas containing 5% CO2 , over the sample; the amount of chemisorbed CO2 was calculated as the difference between the first adsorption peak (physisorbed plus chemisorbed CO2 ) and the arithmetic mean of the desorption peaks. Basicity was assessed on the assumption that one molecule of CO2 was adsorbed at one basic site. The number of basic sites obtained, nb , was thus a measure of basicity. Basic site density, defined as the ratio of the number of basic sites to the specific surface area, can also be used for correlation purposes as it provides a measure of the distance between adjacent basic sites. This method for determining basic sites has been extensively used by our group in examining catalytic solids of variable nature [13,19–21]. 2.3. Catalytic transfer hydrogenation Catalytic hydrogen transfer runs were conducted at 82 ◦ C in a two-mouthed flask containing 0.003 mol

3

of aldehyde, 0.06 mol of isopropyl alcohol and 1 g of freshly calcined catalyst. One of the flask mouths was fitted with a reflux condenser and the other was used for sampling at regular intervals. The system was shaken throughout the process. Products were identified from their retention times and by GC–MS analysis on an HP 5890 GC instrument furnished with a Supelcowax 30 m × 0.32 mm column and an HP 5971 MSD instrument.

3. Results and discussion 3.1. Characterization of HTlc Table 2 shows the nominal chemical composition, BET specific surface area and crystal phases detected by XRD for the different solids. The XRD patterns for the Mg/Al and Mg/Ga samples (Figs. 1 and 2, respectively) are consistent with simple crystal phases with x = M(III)/[M(III]) + Mg] ratios close to 0.25 (Table 2). The XRD patterns are also consistent with the proposed hydrotalcite structure, which consists of layered double hydroxides with brucite-like layers of [Mg1−x Mx (OH)2 ]x+ (CO3 )x/2 2− · mH2 O composition. The stoichiometric hydrotalcite structure, Mg6 Al2 (OH)16 CO3 ·4H2 O (JCPDS 22-700) is reached with x = 0.25. The unit cell constant, a, for each hydrotalcite structure with hexagonal 3R symmetry was calculated from the XRD patterns

Table 2 Chemical analysis, specific surface areas, and XRD characterization of the HTlc HTlc

HT-Mg/Al-1 HT-Mg/Al-2 HT-Mg/Al-3 HT-Mg/Ga-1 HT-Mg/Ga-2 HT-Mg/Ga-3 HT-Mg/In-1 HT-Mg/In-2 HT-Mg/In-3 a

Nominal molar composition x = M(III)/[M(III) + Mg]

SBET a (m2 g−1 )

0.25 0.23 0.26 0.22 0.21 0.24 0.21 0.23 0.22

50.0 151.0 20.2 68.6 173.9 9.1 85.4 97.4 93.7

Specific surface area. HT: hydrotalcite. c Lattice parameter. d Full-width at half-maximum. b

XRD analysis Phaseb

a (Å)c

FWHM (2θ)d

HT HT HT HT HT HT HT + In(OH)3 HT + non identified In(OH)3

3.060 3.070 3.058 3.086 3.090 3.081 3.081 3.091 –

0.43 0.53 0.35 0.62 0.97 0.19 1.01 1.17 –

4

M.A. Aramend´ıa et al. / Applied Catalysis A: General 249 (2003) 1–9

Fig. 1. XRD patterns for the Al-containing HTlc obtained by (a) coprecipitation, (b) the sol–gel technique and (c) decomposition of urea.

Fig. 2. XRD patterns for the Ga-containing HTlc obtained by (a) coprecipitation, (b) the sol–gel technique and (c) decomposition of urea.

[a = 2xd(1 1 0) ] and is shown in Table 2. As can also be seen from the XRD patterns, the HTlc obtained by decomposition of urea were the most crystalline, followed by those prepared by coprecipitation and those obtained using the sol–gel technique. Table 2 includes the full-width at half-maximum (FWHM) values for the (0 0 3) diffractions as quantitative measures of crystallinity along the c axis. From the previous results it follows that, whichever the synthetic method used, the structure of the Aland Ga-containing HTlc is consistent with that of hydrotalcite: it contains positively charged brucite-like [M6 (II)M2 (III)(OH)16 ]2+ layers, with carbonate ions and loosely bound water molecules occupying the interlayer region [22]. As can be seen, none of the In-based HTlc contained a pure hydrotalcite phase; in fact, the solid obtained by decomposition of urea exhibited a highly crystalline In(OH)3 phase (Dzhalindite, JCPDS 16-161). Because it was not a pure phase, its a values could not be compared with those for the previous solids, as this parameter depends on the metal ratio in the HTlc. Among the solids containing a pure hydrotalcite phase, a values were greatest in those obtained us-

Fig. 3. XRD patterns for the In-containing HTlc obtained by (a) coprecipitation, (b) the sol–gel technique and (c) decomposition of urea.

M.A. Aramend´ıa et al. / Applied Catalysis A: General 249 (2003) 1–9

5

ing the sol–gel technique, followed by coprecipitated solids and those prepared by decomposition of urea. These results are consistent with the fact that increasing x decreases this lattice parameter [23]. Likewise, Ga-containing solids exhibited greater a values than Al-containing solids obtained using the same procedure. The a values given for the two In-containing HTlc (Fig. 3) are not comparable to each other or the previous ones as the actual metal ratio in them could not be accurately determined owing to the formation of additional In and Mg phases. The specific surface areas for the solids containing a pure HTlc structure were greatest for those obtained using the sol–gel technique, followed by those prepared by coprecipitation and urea decomposition. As before, the values given for the In compounds are not significant owing to the presence of additional phases. 3.2. Characterization of calcined HTlc The thermal decomposition of the HTlc yielded mixtures of oxides most of which possessed large specific surface areas (Table 3). As a result, the removal of CO2 during calcination increased the porosity of the solids (particularly those obtained using the sol–gel technique). All solids except HT-Mg/In-3-500—the precursor of which, as noted earlier, possessed no HTlc structure—exhibited a periclase MgO crystal phase (JCPDS 4-829) (Figs. 4–6). In any case, the In-containing solids always exhibited the presence of In2 O3 (JCPDS 22-336), whereas those containing Al or Ga were found to contain no AlOx or GaOx crystal phase.

Fig. 4. XRD patterns for the calcined Al-containing HTlc obtained by (a) coprecipitation, (b) the sol–gel technique and (c) decomposition of urea.

CO2 chemisorption provides a qualitative and quantitative measure of the strength of basic sites in a solid. The CO2 thermal programmed desorption profiles for the Mg/Al and Mg/Ga HTlc (shown in Figs. 7 and 8, respectively) suggest the presence of various types of binding sites on the surface of the calcined HTlc. In fact, the curves consist of three overlapped CO2 desorption peaks. These results are consistent with those of Hatori and co-wokers [24] for MgO

Table 3 Specific surface area, XRD characterization and basic properties of the calcined HTlc Catalyst

HT-Mg/Al-1-500 HT-Mg/Al-2-500 HT-Mg/Al-3-500 HT-Mg/Ga-1-500 HT-Mg/Ga-2-500 HT-Mg/Ga-3-500 HT-Mg/In-1-500 HT-Mg/In-2-500 HT-Mg/In-3-500

XRD analysis Phase

Lattice parameter (Å)

MgO MgO MgO MgO MgO MgO MgO + In2 O3 MgO + In2 O3 In2 O3 + unidentified

4.216 4.196 4.151 4.212 4.218 4.218 4.232 4.220 –

SBET (m2 g−1 )

nb (␮mol CO2 g−1 )

235.2 224.6 159.2 138.4 141.9 130.0 80.9 65.4 113.0

330 296 201 182 170 116 193 85 46

6

M.A. Aramend´ıa et al. / Applied Catalysis A: General 249 (2003) 1–9

Fig. 5. XRD patterns for the calcined Ga-containing HTlc obtained by (a) coprecipitation, (b) the sol–gel technique and (c) decomposition of urea.

Fig. 6. XRD patterns for the calcined In-containing HTlc obtained by (a) coprecipitation, (b) the sol–gel technique and (c) decomposition of urea.

Fig. 7. CO2 TPD profiles for the calcined Al-containing HTlc obtained by (a) coprecipitation, (b) the sol–gel technique and (c) decomposition of urea.

Fig. 8. CO2 TPD profiles for the calcined Ga-containing HTlc obtained by (a) coprecipitation, (b) the sol–gel technique and (c) decomposition of urea.

M.A. Aramend´ıa et al. / Applied Catalysis A: General 249 (2003) 1–9

7

Scheme 1. Potential interactions of CO2 with HTlc surface sites.

and those of other authors for Al-containing [25] and Ga-containing HTlc [26]. The IR spectra for CO2 adsorbed on calcined HTlc [27] reveal that CO2 can be adsorbed in three different ways depending on the strength of the basic site concerned (Scheme 1). In our solids, the desorption peak at the lowest temperature corresponds to bicarbonate adsorbed on weakly basic OH groups; the desorption peak at the intermediate temperature to bidentate carbonate adsorbed on M–O pairs; and the desorption peak at the highest temperature to unidentate carbonate bound to strongly basic O2− sites. 3.3. Catalytic activity results The MPV reduction of trans-crotonaldehyde with isopropyl alcohol takes place as depicted in Scheme 2. During the reaction, isopropyl alcohol is oxidized

to acetone and crotonaldehyde to 2-butenol. Table 4 shows the catalytic activity of the solids as calculated using the initial rate method, as well as the reaction yield and selectivity—which, as can be seen, was greater than 90% for all solids except HT-Mg/In-2-500 and HT-Mg/In-3-500. As can also be seen, catalytic activity increased with increasing number of basic sites (i.e. the most basic catalysts were also the most active in the MPV process). Table 4 also shows the turnover frequency (TOF) for the solids, obtained as the ratio of their catalytic activity (ra ) to the number of basic sites as determined by CO2 chemisorption (nb ). As can be seen, TOF was similar for the MgO-containing solids but different for HT-Mg/In-3-500, which contained no MgO. Our catalytic activity results are consistent with a previously reported mechanism (see Scheme 3) where the hydrogen transfer occurs in a concerted manner

Scheme 2. MPV reaction of trans-crotonaldehyde with isopropyl alcohol. Table 4 Catalytic activity, yield and selectivity achieved in the MPV reaction of trans-crotonaldehyde Catalyst

ra (mmol h−1 gcat −1 )a

Yield (%)b

Selectivity (%)

TOF (mmol h−1 m−2 )

HT-Mg/Al-1-500 HT-Mg/Al-2-500 HT-Mg/Al-3-500 HT-Mg/Ga-1-500 HT-Mg/Ga-2-500 HT-Mg/Ga-3-500 HT-Mg/In-1-500 HT-Mg/In-2-500 HT-Mg/In-3-500

1.19 1.03 0.89 0.76 0.61 0.51 0.84 0.31 0.09

100 96 97 97 92 93 94 79 38

99 96 97 95 96 96 96 92 89

0.0051 0.0046 0.0053 0.0055 0.0043 0.0040 0.0103 0.0047 0.0008

a b

Catalytic activity at 10% conversion. Yield at 10 h.

8

M.A. Aramend´ıa et al. / Applied Catalysis A: General 249 (2003) 1–9

Table 5 Catalytic activity of solid HT-Mg/Al-1-500 in the MPV reactions of various ␣,␤-unsaturated aldehydesa Yield (%) (h)b

S (%)c

ra d

1

100 (9)

99

1.19

2

100 (9)

98

1.10

Entry

Reductant

Product

3

98 (10)

97

1.07

4

41 (10)

90

0.21

5

76 (10)

95

0.55

71 (10)

93

0.50

6 Reaction conditions: 3 mmol of aldehyde; 60 mmol of 2-propanol; T = Yield at time in bracket. c Selectivity. d Catalytic activity in mmol h−1 g −1 . cat a

82 ◦ C;

catalyst: 1 g.

b

and involves the formation of a six-member cyclic intermediate. In this scheme, isopropyl alcohol is adsorbed on an acid–base couple where a hydride ion is transferred from the alcohol to the aldehyde. The hydrogen transfer between the two adsorbed substrates has been found to be the rate-determining step of the process. On the other hand, the dehydrogenation of alcohols has been related to the presence of basic or acid–base sites in the catalyst. Accordingly, our results are consistent with the assumption that the most active site should be that most efficiently promoting the dehydrogenation of the alkoxide residue adsorbed on its basic sites. Once solid HT-Mg/Al-1-500 was confirmed to be the most active catalyst, it was used in the MPV reduction of various ␣,␤-unsaturated aldehydes with isopropyl alcohol. Table 5 shows the yield, selectivity and catalytic activity obtained. As can be seen, the solid exhibited similar selectivity and yield with linear

Scheme 3. Mechanism for the MPV reaction.

␣,␤-unsaturated aldehydes (viz. trans-crotonaldehyde, trans-2-pentenal and trans-2-hexenal). With branched and cyclic aldehydes, however, the reaction was much slower; such was the case with perilaldehyde (entry 4 in Table 5)—by contrast, trans-cinnamaldehyde and citral exhibited more similar values. Based on the proposed scheme, these results can be ascribed to steric hindrance (the bulkier the substituent on the ␣-carbon in the aldehyde is, the lower is the catalytic activity).

4. Conclusions An overall nine solids with a layered double hydroxide structure and containing magnesium plus aluminium, gallium or indium were obtained using three different synthetic methods (viz. coprecipitation, the sol–gel technique and precipitation by hydrolysis of urea). The coprecipitated solids were found to possess a brucite-like structure—in addition to an In(OH)3 phase in the case of indium. The sol–gel technique, which had previously never been employed to prepare Ga- or In-containing HTlc, also provided pure HTlc phases—by exception, In also exhibited an additional In(OH)3 phase. All phases were less crystalline than those in the solids obtained by coprecipitation. Finally, the Al and Ga HTlc obtained by precipitation and decomposition of urea were the most crystalline; also, the In solid was found to contain In(OH)3 alone.

M.A. Aramend´ıa et al. / Applied Catalysis A: General 249 (2003) 1–9

Calcination of the Al- and Ga-containing solids at 500 ◦ C yielded a single crystal phase (MgO); by contrast, the In-containing solid gave a mixture of crystalline MgO and In2 O3 . These calcined solids were characterized in terms of their surface chemical properties; the Al solids were found to be the most basic, followed by the Ga solids—the In solids were not comparable as they included additional phases. Also, the coprecipitated solids were found to be the most basic, followed by those obtained using the sol–gel technique and those prepared by decomposition of urea. The calcined solids were used as catalysts in the MPV reaction between trans-crotonaldehyde and isopropyl alcohol, the most basic solid (i.e. the Mg/Al precursor obtained by coprecipitation) also being the most active and selective in the process. The mechanism for the process is assumed to involve a hydrogen transfer between the two substrates, adsorbed at an acid–base site in the catalyst. The yield and selectivity were greater than 90% with all solids except those containing In2 O3 . Also, the reduction rate for various ␣,␤-unsaturated aldehydes was found to depend on the nature of the hydrocarbon chain. Acknowledgements The authors gratefully acknowledge funding by Spain’s Ministerio de Ciencia y Tecnolog´ıa and the Plan Nacional de Investigación, Desarrollo e Innovación Tecnológica (Project BQU-2001-2605), Fondos Feder and to the Consejer´ıa de Educación y Ciencia de la Junta de Andaluc´ıa. References [1] G. Brieger, T. Nestrick, Chem. Rev. 74 (1974) 567. [2] M.A. Aramend´ıa, V. Borau, C. Jiménez, J.M. Marinas, M.E. Sempere, J. Catal. 108 (1987) 487. [3] M.A. Aramend´ıa, V. Borau, C. Jiménez, J.M. Marinas, A. Porras, F.J. Urbano, Appl. Catal. A 172 (1998) 31.

9

[4] M.A. Aramend´ıa, V. Borau, C. Jiménez, J.M. Marinas, C. Santano, M.E. Sempere, J. Mol. Catal. 72 (1992) 221. [5] M.A. Aramend´ıa, V. Borau, J.F. Gómez, A. Herrera, C. Jiménez, J.M. Marinas, J. Catal. 140 (1993) 335. [6] E.J. Creyghton, R.S. Downing, J. Mol. Catal. A: Chem. 134 (1998) 47. [7] J.C. van der Waal, P.J. Kunkeler, K. Tan, H. van Bekkum, J. Catal. 173 (1998) 74. [8] F. Quignard, O. Graziani, A. Choplin, Appl. Catal. A 182 (1999) 29. [9] K. Ganesan, C.N. Pillai, J. Catal. 118 (1989) 371. [10] P.S. Kumbhar, J. Sanchez-Valente, J. Lopez, F. Figueras, J. Chem. Soc., Chem. Commum. (1998) 535. [11] M.A. Aramend´ıa, V. Borau, C. Jiménez, J.M. Marinas, J.R. Ruiz, F.J. Urbano, Appl. Catal. A 206 (2001) 95. [12] M.A. Aramend´ıa, V. Borau, C. Jiménez, J.M. Marinas, J.R. Ruiz, F.J. Urbano, J. Mol. Catal. A: Chem, in press. [13] M.A. Aramend´ıa, Y. Avilés, J.A. Benitez, V. Borau, C. Jiménez, J.M. Marinas, J.R. Ruiz, F.J. Urbano, Microporous Mesoporous Mater. 29 (1999) 319. [14] M.A. Aramend´ıa, Y. Avilés, V. Borau, J.M. Luque, J.M. Marinas, J.R. Ruiz, F.J. Urbano, J. Mater. Chem. 9 (1999) 1603. [15] M.A. Aramend´ıa, V. Borau, C. Jiménez, J.M. Luque, J.M. Marinas, J.R. Ruiz, F.J. Urbano, Mater. Lett. 43 (2000) 118. [16] T. López, P. Bosch, E. Ramos, R. Gómez, O. Navarro, D. Acosta, F. Figueras, Langmuir 12 (1996) 189. [17] U. Costantino, F. Marmottini, M. Nocchetti, R. Vivani, Eur. J. Inorg. Chem. (1998) 1439. [18] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1951) 73. [19] M.A. Aramend´ıa, J.A. Ben´ıtez, V. Borau, C. Jiménez, J.M. Marinas, J.R. Ruiz, F.J. Urbano, J. Solid State Chem. 144 (1999) 25. [20] M.A. Aramend´ıa, J.A. Ben´ıtez, V. Borau, C. Jiménez, J.M. Marinas, J.R. Ruiz, F.J. Urbano, Langmuir 15 (1999) 1192. [21] M.A. Aramend´ıa, V. Borau, C. Jiménez, F. Lafont, J.M. Marinas, A. Porras, F.J. Urbano, Rapid Commun. Mass Spectrom. 9 (1995) 193. [22] H.F.W. Taylor, Min. Mag. 39 (1973) 377. [23] S. Miyata, Clays Clay Miner. 28 (1980) 50. [24] H. Tsuji, F. Yagi, H. Hatori, H. Kita, J. Catal. 148 (1994) 759. [25] J.I. Di Cosimo, V.K. Diez, M. Xu, E. Iglesia, C.R. Apesteguia, J. Catal. 178 (1998) 499. [26] E. Lopez-Salinas, M. Garcia-Sanchez, M.E. Llanos-Serrano, J. Navarrete-Bolañoz, J. Phys. Chem. B 101 (1997) 5112. [27] M. Del Arco, C. Martin, I. Martin, V. Rives, R. Trujillano, Spectrochim. Acta 49 (1993) 1572.