Aqueous-phase dehydration of xylose to furfural in the presence of MCM-22 and ITQ-2 solid acid catalysts

Applied Catalysis A: General 417–418 (2012) 243–252

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

Aqueous-phase dehydration of xylose to furfural in the presence of MCM-22 and ITQ-2 solid acid catalysts Margarida M. Antunes a , Sérgio Lima a , Auguste Fernandes b , Martyn Pillinger a , Maria F. Ribeiro b , Anabela A. Valente a,∗ a b

Department of Chemistry, CICECO, University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal Institute for Biotechnology and Bioengineering, Centre for Biological and Chemical Engineering, Instituto Superior Técnico, Av. Rovisco Pais, 1049-001 Lisboa, Portugal

a r t i c l e

i n f o

Article history: Received 26 July 2011 Received in revised form 28 December 2011 Accepted 29 December 2011 Available online 10 January 2012 Keywords: Xylose Furfural Acid dehydration MCM-22 zeolite ITQ-2

a b s t r a c t The H-MCM-22 zeolite possessing an MWW (medium-pore) framework and its delaminated counterpart, ITQ-2, with enhanced external surface area, are effective and recyclable solid acid catalysts in the batch-wise, aqueous-phase dehydration of xylose, at 170 ◦ C. Up to 71% furfural (Fur) yield is reached at 98% conversion using a biphasic aqueous–organic solvent system (for the simultaneous extraction of Fur); using solely water as the solvent gives up to 54% Fur yield at 97% conversion. Sulfuric acid, used in an approximately equivalent amount to the total acid site concentration of the solid acids, gave 55% Fur yield at 93% conversion. Decreasing the Si/Al ratio of H-MCM-22 zeolite improves the acid properties and consequently the catalytic activity, without affecting significantly the Fur selectivity. While the delamination process considerably enhanced the external surface area of ITQ-2 in comparison to H-MCM-22, it caused modifications in the acid properties, leaving the two prepared materials with the same Si/Al atomic ratio of 24, on a comparable footing in terms of catalytic performance in the studied catalytic reaction. Nevertheless, these solid acid catalysts are fairly stable (similar Fur yields are reached in recycling runs; no structural modifications and no leaching phenomena were detected). © 2012 Elsevier B.V. All rights reserved.

1. Introduction Furfural (Fur; C5 H4 O2 ) is a platform chemical derived from lignocellulosic biomass and has been commercialised since the beginning of the last century (from cereal waste stockpiles) [1]. The Fur market extends to different sectors of the chemical industry (e.g., oil refining, resins, agrochemicals and pharmaceuticals) and is attracting increasing attention in the biofuels sector [1–4]. The synthesis of Fur involves the primary hydrolysis of pentosans (polysaccharides, which may be found in agricultural/forestry wastes/surpluses) to give the respective pentoses (monosaccharides), followed by the dehydration of the latter to Fur (Scheme 1); theoretically, water is the sole co-product. The hydrolysis/dehydration reactions are promoted by acid catalysts and sulfuric acid is most commonly used in the industrial production of Fur. While being quite effective, H2 SO4 poses risks to human health, the environment (difficult catalyst recovery/recycling, possible formation of sulfur-containing by-products, neutralisation of effluents) and the process equipment (e.g., corrosion hazards).

∗ Corresponding author. Tel.: +351 234 370603; fax: +351 234 401470. E-mail addresses: [email protected], [email protected] (A.A. Valente). 0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2011.12.046

There has been growing interest in the search for heterogeneous catalytic routes for the production of Fur. Different types of porous solid acids have been investigated as catalysts in the reaction of xylose in the aqueous phase, such as: (organic) acid resins [5] and modified carbonaceous spheres [6], (hybrid) silicas functionalised with sulfonic acid groups [7–9], (inorganic) bulk/supported heteropolyacids (HPAs) [10,11], micro/mesoporous transition metal oxides (Zr, Zn, Ti [12], ZrW [5], Ti/Nb [13], NbSi [14]), sulfated metal oxides (ZrAl [15,16]), silicoaluminophosphates [17] and aluminosilicates [5,18–21]. Important requirements to be put on solid acid catalysts for this reaction system include water-tolerance (minimal levelling-off of the acid strength in water), hydrothermal stability (crystalline structure integrity and stability towards metal leaching) and good thermal stability if the catalyst regeneration requires thermal decomposition of accumulated carbonaceous deposits (typically in the range 350–550 ◦ C), or chemical stability if carbonaceous deposits are to be removed by harsh chemical treatments (e.g., lower temperature, liquid-phase oxidising conditions). From the literature data, certain families of porous inorganic solids can be pinpointed as fairly robust materials for this catalytic application, and in particular aluminosilicas (bulk catalysts) are relatively cheap and versatile materials with respect to the type of crystalline and porous structures (micro/meso/macropores; 1,

244

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252

2 or 3-D pore systems), the acid properties and surface polarity (varying Si/Al ratio), and they may be prepared with particle/crystallite sizes down to the nano-scale. In pioneering work by Moreau et al. [18], zeolites revealed to be effective solid acids for the conversion of saccharides into Fur; in the case of H-Mordenite and H-Y faujasite with Si/Al atomic ratio in the range 2–15, 90–96% selectivity was reached at 27–37% conversion, 170 ◦ C, albeit selectivity dropped considerably as conversion increased [18]. Aiouache and co-workers found a similar trend in Fur selectivity with conversion of xylose for zeolite ZSM-5 as catalyst, and explained these results on the basis of enhanced Fur loss reactions inside the pore system, which may eventually cause pore blockage and deactivation of the catalyst [19]. Various synthetic approaches have improved the catalytic performances of zeolites, such as decreasing the zeolite crystallite sizes to the nano-scale to increase the surface/pore volume ratio [21], and delaminating lamellar precursors of zeolites into aggregates of sheets with zeolitic nature and enhanced specific surface area [22]. The zeolite MCM-22, which possesses an MWW type framework and was discovered by Mobil in 1990 [23], may be prepared by calcination of a lamellar precursor herein denoted as Pre-MCM22. The MWW structure consists of two independent pore systems and specific surface area is commonly in the range 300–500 m2 g−1 : one pore system is formed by large cages (∼7.1 A˚ Ø, 18.2 A˚ height) which are accessed through 10-membered ring (10-MR) apertures, and the other one is defined by a circular 10-MR sinusoidal channel system (medium-pore). The maximum diameter of the pore openings of MCM-22 is 5.5 A˚ or 6.2 A˚ based on atomic or Norman radii, respectively [24]. Since xylose molecules possess an approximate molecular diameter of 6.8 A˚ [25], access to the internal catalyst surface of MCM-22 through the 10-MR apertures will likely be impeded. In order to enhance the external surface area available for the catalytic reactions, Corma and co-workers [26–29] developed a delamination procedure to transform Pre-MCM-22 into a delaminated material known as ITQ-2, possessing very high external specific surface area (∼600–800 m2 g−1 ). ITQ-2 consists of 2.5 nm thick sheets possessing a hexagonal array of cup-shaped open pores facing out of each side of the sheets (∼7.1 A˚ Ø, 7.0 A˚ height, with bottoms connected by a double 6-MR unit), and a circular 10-MR sinusoidal channel system running between the cups, inside the sheet [26–30]. In the present work, the dehydration of xylose to Fur is investigated in the presence of MCM-22 and ITQ-2 catalysts, using a biphasic water–toluene solvent system or solely water as solvent, at 170 ◦ C.

2. Experimental 2.1. Preparation of the catalysts The layered zeolite precursors, denoted Pre-MCM-22(X), where X is the Si/Al molar ratio of 30 or 50 used in the synthesis gel, were prepared as described in the literature [27,28]. The Si/Al ratios were chosen to achieve a compromise between enhanced acid site density and successful delamination: decreasing the Al content of the zeolite precursor leads to a lower charge density and therefore the interactions between the zeolitic layers are weaker, facilitating the delamination process [31–34]. For Pre-MCM-22(30) sodium aluminate (Riedel-De Haën, 53% Al2 O3 , 47% Na2 O, 0.587 g, 6.10 mmol) and sodium hydroxide (Prolab, 98%, 1.03 g, 25.7 mmol) were dissolved in water (milli-Q, 156 g, 8.67 mol). Hexamethyleneimine (Aldrich, 99%, 8.36 g, 84.3 mmol) was then added and the mixture stirred for 45 min, followed by the addition, under agitation, of silica (Aerosil 200, Degussa, 11.0 g, 183.1 mmol). The mixture was stirred for a further 2 h to give a gel which was transferred to a 250 mL PTFE lined stainless-steel autoclave, rotated at 50 rpm, and heated at 135 ◦ C for 11 days; after this period of time the autoclave was quenched in cold water. The solid phase was separated by centrifugation and washed thoroughly with deionised water until pH < 9.0, and subsequently dried at 25 ◦ C overnight. Finally, part of the solid was calcined under static air at 540 ◦ C for 6 h, giving Na-MCM-22(24) (in parenthesis, the bulk Si/Al atomic ratio measured by ICP-AES). A similar procedure was followed to prepare Pre-MCM-22(50), which upon calcination gave Na-MCM-22(38). The delaminated aluminosilicate, ITQ-2, was prepared from Pre-MCM-22(30) as follows: Pre-MCM-22(30) (10 g) was dispersed in milli-Q H2 O (20 g), followed by the addition of aqueous cetyltrimethylammonium hydroxide/bromide (100 g, 20 wt%; 40% exchanged Br/OH, prepared by ion-exchange of cetyltrimethylammonium bromide (Fluka, ≥96%) using Amberlite IRA-400(OH)) and aqueous tetrapropylammonium hydroxide/bromide (30 g, 40 wt%; 50% exchanged Br/OH, prepared from a mixture of tetrapropylammonium bromide (Fluka, ≥99%) and tetrapropylammonium hydroxide (40 wt%, Alfa Aesar)). The resultant mixture possessed pH ≈ 13.5 and was heated at 80 ◦ C with vigorous stirring for 16 h in order to facilitate the swelling of the layers of the precursor material; the final pH was 11.9. The suspension was sonicated using an ultrasound bath (50 W, 50 Hz) during 1 h; after decantation, the supernatant colloid was separated. The pH of the colloid was lowered to ca. 2.0 by adding 6 M HCl in order to facilitate

Scheme 1. Simplified representation of the conversion of carbohydrate biomass into furfural.

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252

the flocculation of the delaminated solid. The solid was separated by centrifugation and washed with distilled water. After drying at 60 ◦ C for 12 h, the solid was calcined at 540 ◦ C for 3 h under a flow of N2 (300 mL min−1 ), and then during 6 h under air (300 mL min−1 ) [29,35]. The catalysts were manually ground using an agate pestle and mortar and subsequently sieved to give a powder with particle sizes of less than 106 ␮m width. An ion-exchange/calcination procedure was applied to NaMCM-22(24) and Na-MCM-22(38) to give H-MCM-22(24) and H-MCM-22(38), respectively. The ion-exchange procedure consisted of suspending 1 g of the solid material in 15 mL of 1 M NH4 NO3 and stirring the resultant suspension for 24 h at 80 ◦ C; the ion-exchanged materials were washed thoroughly with deionised water, dried at 50 ◦ C overnight, and finally calcined in air at 540 ◦ C (heating rate of 1 ◦ C min−1 ) for 6 h.

2.2. Catalyst characterisation ICP-AES measurements for Si and Al (error of 5–10%) were carried out at the Central Laboratory for Analysis, University of Aveiro (by L. Carvalho). Powder XRD data were collected at room temperature on a Philips X’Pert MPD diffractometer, equipped with an X’Celerator detector, a graphite monochromator (Cu K␣ ˚ and a flat-plate sample holder, in a X-radiation,  = 1.54060 A) Bragg–Brentano para-focusing optics configuration (40 kV, 50 mA). Samples were step-scanned in 0.04◦ 2 steps with a counting time of 6 s per step. Scanning electron microscopy (SEM) was carried out on a Hitachi SU-70 UHR Schottky instrument. Transmission electron microscopy (TEM) was carried out on Hitachi H9000 and JEOL 2200FS instruments. Samples were prepared by spotting holey amorphous carbon film-coated 400-mesh copper grids (Agar Scientific) with a suspension of the solid sample in ethanol. The following textural parameters were estimated from N2 adsorption isotherms measured at −196 ◦ C using a Micromeritics instrument Corp Germini model 2380 (sample pre-treatment: 250 ◦ C, vacuum): BET specific surface area (SBET , calculated for relative pressures in the range 0.01–0.10), microporous external specific surface area (Sext ; t-plot method), and BJH pore size distribution curve calculated from the adsorption branch of the isotherm. Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) were carried out under air, with a heating rate of 10 ◦ C min−1 , using Shimadzu TGA-50 and DSC-50 systems, respectively. The 27 Al magic-angle spinning (MAS) NMR spectra were measured at 104.26 MHz with a Bruker Avance 400 (9.4 T) spectrometer, using a contact time of 0.6 ␮s, a recycle delay of 0.8 s, and a spinning rate of 15 kHz; chemical shifts are quoted in ppm from Al(H2 O)6 3+ . Infrared spectra were recorded on a Unican Mattson Mod 7000 FTIR spectrophotometer using KBr pellets. The liquid state 1 H NMR and 13 C NMR spectra were recorded on a Bruker DRX 300 MHz spectrometer at 20 ◦ C. It has been reported in the literature for dehydration reactions in aqueous solutions with solid acid catalysts that gas-phase characterisation of acid sites can be used to predict catalytic activity in the aqueous phase [5]. The acid properties of the prepared catalysts were examined by FT-IR studies of adsorbed basic probe molecules, namely pyridine and collidine (2,4,6-trimethylpyridine). Pyridine possesses a critical dimension of ca. 6.5 A˚ [36] which is comparable with the molecular diameters of xylose (6.8 A˚ along the longest axis [25]). For pyridine as basic probe, the acid properties were measured using a Nexus-Thermo Nicolet apparatus (64 scans and resolution of 4 cm−1 ) equipped with a specially designed cell, using self-supported discs (5–10 mg cm−2 ). After in situ outgassing at 450 ◦ C for 3 h (10−6 mbar), pyridine (99.99%) was contacted with the sample at 150 ◦ C for 10 min and then evacuated at 150 ◦ C (30 min) under vacuum (10−6 mbar). The IR bands at ca. 1540 and

245

1455 cm−1 are related to pyridine adsorbed on B and L acid sites, respectively [37]. For collidine as basic probe, the FT-IR spectra were recorded with a resolution of 2 cm−1 on a Nicolet Nexus spectrophotometer equipped with an MCT detector. The powdered solids were pressed into self-supported disks (10 mg cm−2 ) and placed in a quartz IR cell with KBr windows. After in situ outgassing at 450 ◦ C for 4 h, collidine was contacted with the sample at 200 ◦ C for 30 min and then evacuated at 200 ◦ C (30 min) under vacuum. The IR band at ca. 1635 cm−1 is related to collidine adsorbed on B acid sites. 2.3. Catalytic tests Batch catalytic experiments were performed under nitrogen atmosphere in tubular glass micro-reactors with pear-shaped bottoms and equipped with an appropriate PTFE-coated magnetic stirring bar and a valve for gas purging. In a typical procedure, dxylose (30 mg), powdered catalyst (20 mg) and either water (1 mL, denoted W) or a solvent mixture (denoted W-T) comprising H2 O (0.3 mL) and toluene (0.7 mL) were added to the reactor. The reaction mixtures were heated with a thermostatically controlled oil bath under magnetic stirring at 700 rpm. Zero time was taken to be the instant the micro-reactor was immersed in the oil bath. It may be assumed that external mass transfer limitations are avoided at this stirring rate, since tests showed that the initial reaction rates (based on conversion at 1 h reaction) were similar for stirring rates at or above 700 rpm: 1.6, 2.4 and 2.3 mmol gcat −1 h−1 at 600, 700 and 800 rpm, for H-MCM-22(24); 2.1, 2.0 and 2.1 mmol gcat −1 h−1 at 600, 700 and 800 rpm, for ITQ-2. The products present in the aqueous phase were analysed using a Knauer K-1001 HPLC pump and a PL Hi-Plex H 300 mm × 7.7 mm (i.d.) ion exchange column (Polymer Laboratories Ltd., UK), coupled to a Knauer 2300 differential refractive index detector (for xylose) and a Knauer 2600 UV detector (280 nm, for Fur). The mobile phase was 0.001 M H2 SO4 . Analysis conditions: flow rate 0.6 mL min−1 , column temperature 65 ◦ C. The Fur present in the organic phase was quantified using a Gilson 306 HPLC pump and a Spherisorb ODS S10 C18 column, coupled to a Gilson 118 UV/Vis detector (280 nm). The mobile phase consisted of 37% (v/v) methanol and 63% (v/v) H2 O (flow rate 0.5 mL min−1 ). Authentic samples of d-xylose and Fur were used as standards and calibration curves were used for quantification. The Fur yield (%) was calculated using the formula: [(moles of Fur formed)/(initial moles of xylose) × 100]. For each reaction time, at least two replicates of an individual experiment were made; the reported results are the average values. The maximum average absolute deviation calculated for xylose conversion and Fur yield was less than 4%. 3. Results and discussion 3.1. Characterisation of the catalysts The ion-exchange/calcination treatments applied to Na-MCM22(24) and Na-MCM-22(38) to give H-MCM-22(24) and H-MCM22(38), respectively, did not affect the Si/Al atomic ratio. The delamination procedure applied to Pre-MCM-22(30) gave the material ITQ-2 with a slightly lower Si/Al atomic ratio, possibly due to slight dissolution of silica during the delamination process [33,34]. The powder XRD patterns of the (Na,H)-MCM-22 samples are in good agreement with the literature data for a MWW framework topology (Fig. 1) [26,38]. The XRD pattern of ITQ-2 reveals a significant reduction in the long-range order as a result of the delamination procedure, in agreement with literature data [26,28,33,34,39]. The SEM analyses show that the (Na,H)-MCM-22 samples consist of thin plate-like crystallites of a few hundreds

246

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252

Fig. 3. TEM images of Na-MCM-22(24) (a) and an ITQ-2 layer (b), viewed along the 10-ring channels, perpendicular to the c-direction.

Fig. 1. Powder XRD patterns of as-prepared and (for H-MCM-22(24) and ITQ-2) used/calcined (after catalysis) catalysts.

of nanometers to ca. 1 ␮m wide, and ITQ-2 consists of particles of irregular shape (exemplified in Fig. 2 for Na-MCM-22(24) and ITQ-2; Figure S1 for H-MCM-22(24) and H-MCM-22(38)). Fig. 3 shows a representative HRTEM image of a Na-MCM-22(24)

Fig. 2. SEM images of (a) Na-MCM-22(24) and (b) ITQ-2.

crystallite viewed along the 10-MR channels (bright spots), perpendicular to the c-direction. The images generally showed the stacking of MWW sheets with a thickness ranging from 150 to 300 nm, corresponding to the thickness of the platelet crystals. As expected, the 10-MR channels are separated by 1.2 nm, and the thickness of the layers is ca. 2.5 nm. HRTEM images of ITQ-2 showed much more fragmented crystals consisting of fewer sheets than NaMCM-22(24), and even single layers of 2.5 nm thickness as shown in Fig. 3(b). The (Na,H)-MCM-22 materials exhibited type I adsorption isotherms, typical of microporous materials, with an enhanced increase in the amount of adsorbed N2 at high relative pressures (p/p0 ), which is most likely due to multilayer adsorption on the external surface of the crystallites (exemplified for Na-MCM-22(24) in Fig. 4; Figure S2 for H-MCM-22(24), Na-MCM-22(38) and HMCM-22(38)). The external surface areas (Sext ) of these samples are significant (48–74 m2 g−1 , Table 1), which is consistent with the rather small crystallite sizes (Fig. 2). The ion-exchange/calcination treatments did not affect significantly the textural properties of the Na-MCM-22 samples. The adsorption-desorption isotherm for ITQ2 exhibited a hysteresis loop at p/p0 > 0.5, indicating the presence of mesoporosity; the mesopore size distribution curve exhibited a

Fig. 4. Nitrogen adsorption isotherms measured for Na-MCM-22(24) and ITQ-2, at −196 ◦ C.

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252

247

Table 1 Textural and acid properties, and catalytic performances of aluminosilicates tested as catalysts in the reaction of xylose, using a biphasic W-T solvent mixture, at 170 ◦ C.a SBET (Sext ) (m2 g−1 )

Catalyst (Si/Al)b Na-MCM-22(38) H-MCM-22(38) Na-MCM-22(24) H-MCM-22(24) ITQ-2(24) H-Nu-6(2)(32) del-Nu-6(1)(29) BEA(12) BEA/TUD-1 Al-TUD-1(21) H-ZSM-5(46) a b c d e f g h

524 (62) 497 (48) 361 (52) 333 (74) 623 (611) 20 (–) 151 (6) 643 (176) 712 (–) 757 (–) 425

[L] + [B] (␮mol g−1 )c

[L]/[B]d

t (h)e

Conv. (%)f

YFur (%)g

Ref.

– 168 – 204 198 ndh nd 351 209 197 –

– 0.6 – 0.4 1.0 nd nd 1.3 1.2 2.3 –

– 24/32 24 16 16 6 6 4/6 6/8 6 6/16

– 82/97 98 92 99 59 87 98/100 94/98 90 88/98

– 60/68 47 70 66 28 46 54/49 69/74 60 60/61

– – – – – [22] [22] [21] [21] [20] –

Reaction conditions: 0.67 M xylose, xylose:catalyst mass ratio of 3:2, biphasic W-T solvent mixture, 170 ◦ C. The Si/Al atomic ratio of the catalyst is given in parenthesis. Sum of the total amount of Brönsted acid sites [B] plus Lewis acid sites [L]. Ratio of the amounts of Lewis to Brönsted acid sites. Reaction time. Xylose conversion at the specified reaction times. Furfural yield at the specified reaction times. Not determined.

maximum at 3.7 nm (Fig. 4), in agreement with the literature data [28,40]. The ITQ-2 sample possesses much higher Sext (611 m2 g−1 ) than the (Na,H)-MCM-22 materials, which is consistent with a successful delamination procedure (Table 1). The 27 Al MAS NMR spectra of the Na-MCM-22 samples exhibit an asymmetric broad peak with maximum at ca. 56 ppm and a shoulder at ca. 50 ppm (Fig. 5), assigned to tetrahedrally coordinated framework aluminium (denoted FAl), which may be nonequivalent chemical environments with different T–O–T angles [30,39–42]. In the case of the H-MCM-22 samples an additional (narrow, weaker) peak appears at ca. 0 ppm, assigned to extraframework aluminium species (denoted EFAl) formed during the ion-exchange/calcination treatments [39,43,44]. The relative amounts of the FAl and EFAl species were determined from the

Fig. 5.

27

Al MAS NMR spectra of the as-prepared catalysts.

areas of the respective peaks (AFAl = area of the peak in the range of 35–70 ppm; AEFAl = area of the peak centered at 0 ppm): the ratio AFAl /AEFAl equals 8.4 for H-MCM-22(24) and 9.4 for H-MCM-22(38); these results are in agreement with the literature data in that higher Si/Al ratios can lead to higher relative amounts of tetrahedral aluminium [34]. The 27 Al NMR spectrum of ITQ-2 reveals the presence of FAl (ca. 55 ppm) and EFAl species (0 ppm), in agreement with the literature data [34,39]. The AFAl /AEFAl ratio for ITQ-2 equals 8.5 which is similar to that observed for its counterpart H-MCM-22(24). The FT-IR spectrum of ITQ-2 shows a band at ca. 956 cm−1 assigned to terminal Si-OH groups, which is poorly resolved in the case of H-MCM-22(24), further suggesting that delamination occurred to a significant extent in the preparation of ITQ-2 (Fig. 6) [26,28,39]. The acid properties of the prepared catalysts were examined by FT-IR studies of adsorbed pyridine (Table 1). The H-MCM-22 materials possess mainly Brönsted acid sites ([B]), and comparable ratio of Lewis to Brönsted acid sites ([L]/[B] = 0.4–0.6). The H-MCM-22(24) sample possesses a higher total amount of Lewis and Brönsted acid sites ([L] + [B]) in comparison to H-MCM-22(38), which is consistent with the lower Si/Al ratio for the former. The ITQ-2 sample possesses a similar total amount of acid sites

Fig. 6. FT-IR spectra, in the framework region of H-MCM-22(24) and ITQ-2.

248

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252

to H-MCM-22(24), while the [L]/[B] ratio is higher for ITQ-2 (a similar trend has been reported in the literature [27]). The ratios of the amounts of acid sites measured using pyridine adsorbed at 150 and 250 ◦ C ([L]150 /[L]250 and [B]150 /[B]250 ) were both equal to 1.1 for H-MCM-22(24), whereas for ITQ-2 [L]150 /[L]250 = 1.4 and [B]150 /[B]250 = 2.6. Hence, the surface acidity (particularly the B acidity) for ITQ-2 is weaker. For comparison, the acid properties of ITQ-2 and H-MCM-22(24) were measured using collidine as the basic probe molecule. Due to ˚ the collidine molecule is not expected to its steric bulk (ca. 7.4 A), enter the pores of the MWW framework structure [45,46], i.e., it should only interact with B acid sites (denoted [B]col ) located on the external surface or at the pore entrances. On the other hand, collidine is a stronger base (pKa ∼ = 7.4) than pyridine (pKa ∼ = 5.3) [47]. The [B]col for H-MCM-22(24) and ITQ-2 was 45 and 153 ␮mol g−1 , respectively (the reproducibility was checked in duplicate experiments). The lower [B]col for H-MCM-22(24) than for ITQ-2 is most likely due to steric constraints (inaccessibility of the acid sites on the internal surface to the bulky collidine molecules). For ITQ-2, [B]col is greater than [B] determined using pyridine, suggesting that the range of interaction energies with (stronger base) collidine is wider than that with pyridine. Based on the above results, it seems that while the delamination procedure led to enhanced external surface area of ITQ-2, the surface acidity became relatively weak. 3.2. Catalytic studies 3.2.1. Catalytic performance of H-MCM-22 in water-organic solvent system The catalysts were tested in the aqueous-phase reaction of xylose to give furfural (Fur), using a biphasic water–toluene (0.3:0.7, v/v) solvent system (denoted W-T), at 170 ◦ C. Xylose dissolves completely in water and is insoluble in toluene, whereas Fur is distributed in the two liquid phases with a partition ratio ((moles of Fur in T)/(moles of Fur in W)) in the range of 8–10, at room temperature. Hence, the use of toluene as co-solvent allows the simultaneous extraction of Fur as it is formed, into the (upper) organic phase, which may avoid consecutive Fur-loss reactions. The reaction of xylose in the presence of H-MCM-22(24) gave 70% Fur yield at 92% conversion, reached at 16 h (Fig. 7). The kinetic curves for H-MCM-22(24) and its parent basic form, Na-MCM22(24), are roughly coincident (Fig. 7(a)); the initial reaction rate is 2.4 mmol h−1 gcat −1 (based on conversion at 1 h reaction). However, higher Fur yields are reached for H-MCM-22(24) than for Na-MCM22(24) (Fig. 7(b)). The reaction mechanism of the dehydration of xylose to Fur involves a series of elementary steps leading to the liberation of three water molecules per Fur molecule formed, and these reactions take place over acid sites which may be of Brönsted or Lewis type [12,20,21,48]. The presence of basic sites in the catalysts can promote undesirable reactions such as the aldolisation decomposition of the saccharide into oligomeric acid products [49,50]. The influence of the Si/Al ratio of the catalyst on the reaction of xylose was investigated by comparing the catalytic performances of H-MCM-22(24) and H-MCM-22(38). For H-MCM-22(38), the initial reaction rate was 1.0 mmol h−1 gcat −1 and a conversion of 71% was reached at 16 h reaction (Fig. 7(a)). The higher catalytic activity observed for H-MCM-22(24) correlates with the higher total amount of acid sites ([L] + [B]) determined by IR studies of adsorbed pyridine (Table 1). The Fur yields reached at ca. 98% xylose conversion were 68% for H-MCM-22(38) and 71% for H-MCM-22(24) (Fig. 7(b)). Hence, the decrease in the Si/Al ratio (in the range of 24-38) of the H-MCM-22 zeolite leads to an increase in the total amount of acid sites which, in turn, improves the catalytic activity in the reaction of xylose, without affecting significantly the Fur selectivity.

Fig. 7. Kinetic profiles of the reaction of xylose (a), and Fur yield versus xylose conversion curves (b), for Na-MCM-22(24) (), H-MCM-22(24) (×), H-MCM-22(38) (+) and ITQ-2 (), used as catalysts. Reaction conditions: 0.67 M xylose, xylose:catalyst mass ratio of 3:2, biphasic W-T solvent mixture, 170 ◦ C.

The catalytic activity of (medium-pore) H-MCM-22(24) is intermediate between that previously reported for the small-pore zeolite H-Nu-6(2) (Si/Al = 32) [22] and large-pore zeolite BEA (Si/Al = 12) [21], used as catalysts in the same reaction under similar conditions (Table 1). Nevertheless, the maximum Fur yields reached are highest for H-MCM-22(24). For comparison, the reaction of xylose was carried out, under similar reaction conditions, in the presence of the protonic form of a commercial sample of ZSM-5 (Alfa Aeser; ammonium form; Si/Al = 46; 425 m2 g−1 ), which is a medium-pore zeolite with the MFI framework type. The as-received zeolite was heated under static air at a ramp rate of 1 ◦ C min−1 to 550 ◦ C, and maintained at this temperature for 10 h, giving H-ZSM-5(46). The xylose conversions and Fur yields at 6 h/16 h were 88%/98% and 60%/61%, respectively. Compared with the prepared H-MCM-22 samples, the H-ZSM-5(46) catalyst is more active (e.g., for the more active H-MCM-22(24), 73%/92% conversion at 6 h/16 h), but led to lower Fur yields at high conversions (at ca. 98% conversion: 71%, 69% and 61% Fur yield for H-MCM-22(24), H-MCM-22(38) and H-ZSM-5(46), respectively). O’Neill et al. investigated the reaction of xylose in the presence of H-ZSM-5 with Si/Al = 28 [19]. In that study, a maximum of ca. 33% Fur yield was reached within 60 min, at 180 ◦ C, and afterwards the yield tended to

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252

249

drop with time (reaction conditions: 10 wt% xylose, catalyst:xylose ratio of 0.3 (w/w), under pressurised He atmosphere); the Fur yields reported for the reaction temperature of 160 ◦ C were lower. The average pore size of the investigated ZSM-5 catalyst was ca. 1.2 nm which is far greater than the molecular diameters of both xylose and Fur, and this was considered as a possible cause of the formation of considerable amounts of by-products (oligomers). 3.2.2. ITQ-2 versus H-MCM-22(24) in water–organic solvent system The reaction of xylose was further investigated in the presence of ITQ-2 (the delaminated counterpart of H-MCM-22(24)), using the biphasic W-T solvent system at 170 ◦ C, which gave a Fur yield of 66% at 99% conversion (Fig. 7(b)). In terms of the maximum Fur yield reached, these results are superior to those reported previously for a delaminated solid acid (denoted del-Nu-6(1)) prepared by exfoliation of the lamellar precursor Nu-6(1) (46% Fur yield [22]), comparable to a mesoporous Al-TUD-1 material (60% Fur yield [20]), and lower than that for a BEA/TUD-1 composite (74% Fur yield [21]), used as catalysts in the same reaction under similar conditions (ITQ-2 possesses comparatively lower catalytic activity), Table 1. It is difficult to establish clear structure–activity relationships between different types of aluminosilicate catalysts; the catalytic performances may be due to an interplay of different factors such as morphology, crystalline structure, porosity, surface polarity and acid properties. A common feature, however, is the apparent fairly high stability of the aluminosilicate catalysts investigated, under the applied reaction conditions (discussed below for H-MCM-22 and ITQ-2). These comparisons merely summarise literature data obtained under similar reaction conditions, and do not serve to “rate” the different families of catalysts (the physical–chemical properties of different types of catalysts may be optimised). A comparison of the catalytic results for ITQ-2 and H-MCM22(24) shows no major differences in terms of reaction rate (Fig. 7(a)): ITQ-2 gave an initial reaction rate of 2.0 mmol h−1 gcat −1 and a conversion at 8 h of 78%, while H-MCM-22(24) gave 2.4 mmol h−1 gcat −1 and a conversion of 83%. These results correlate fairly well with the similar total concentration of acid sites ([L] + [B]) measured for the two catalysts using pyridine as probe molecule (Table 1). The catalytic activity does not correlate with the [B]col which is higher for ITQ-2 than for H-MCM-22(24); possibly the stronger base collidine interacts with some acid sites which are too weak for catalysing the reaction of xylose. The Fur yield versus xylose conversion profiles are also comparable for the two catalysts, and the Fur yield reached at ca. 99% xylose conversion is 71% for H-MCM-22(24) and 66% for ITQ-2 (Fig. 7(b)). Hence, ITQ-2 stands on a similar footing to H-MCM-22(24) (the Si/Al ratio is the same for the two materials) in terms of catalytic performance in the reaction of xylose. The catalytic performances of ITQ-2 and H-MCM-22(24) are similar despite their considerably different Sext . A clear assessment of the location (external/internal surfaces) of the effective acid sites is not trivial. As mentioned in Section 1, the intrazeolite void space of the MWW pore structure is accessible through the 10-MR apertures [26], and the maximum pore diameter is 5.5 A˚ or 6.2 A˚ based on atomic or Norman radii, respectively [24]. The xylose molecules (the hemiacetal isomers are predominant in solution) possess an approximate molecular diameter of 6.8 A˚ [25], and the critical, max˚ 5.99 A˚ and 5.5 A, ˚ imum and kinetic diameters of Fur are 4.56 A, respectively [24]. Based on these data, and taking into consideration that in the liquid phase the solute molecules are solvated by the solvent, the access of xylose (and possibly Fur) molecules to the internal surface of the MWW-type framework may be severely hindered. It has been reported for benzene as substrate and zeolite MCM-22 as catalyst that the reaction takes place essentially on

Fig. 8. DSC curves for the as-prepared H-MCM-22(24) (grey line) and ITQ-2 (black line) catalysts, and the respective solids recovered from the reaction of xylose after ca. 98% conversion was reached (used catalysts). Reaction conditions: 0.67 M xylose, xylose:catalyst mass ratio of 3:2, biphasic W-T solvent mixture, 170 ◦ C.

the external surface [45,51]; molecular dynamics simulations indicated that benzene (kinetic diameter of 5.85 A˚ [52]) presents low diffusivity in either of the two pore systems of the MWW structure [53]. Assuming that the reaction of xylose takes place essentially on the external surface of the catalysts, the similar catalytic performances of ITQ-2 and H-MCM-22(24) despite their considerably different Sext may be due to weaker overall acidity in the former case (some acid sites may be inactive or possess relatively low intrinsic catalytic activity). After at least 98% xylose conversion was reached, the H-MCM22(24) and ITQ-2 catalysts were separated from the reaction mixtures by centrifugation, washed with methanol and dried at 50 ◦ C overnight, giving pale brown powders. For each solid, the DSC analysis showed an endothermic band below 200 ◦ C assigned to desorption of physisorbed water and strong exothermic bands above 200 ◦ C (which were not observed for the fresh catalysts) assigned to the combustion of organic matter (Fig. 8). The amount of organic matter in the used catalysts (measured by the weight loss in the temperature range of 220–700 ◦ C) was similar for HMCM-22(24) and ITQ-2 (ca. 12 wt%). However, the DSC profiles are different, suggesting that the chemical nature of the carbonaceous matter is different for the two catalysts. The spectrum of by-products formed may be influenced by the [L]/[B] acid site ratio [5] and/or textural properties (by-products may become strongly adsorbed/entrapped inside the MWW microporous structure). The stability of the H-MCM-22(24) and ITQ-2 catalysts was investigated by recycling the solid acids, under biphasic solvent conditions, at 170 ◦ C. The washing of the used catalysts with different solvents (methanol, acetone, toluene, water) failed to efficiently remove the organic matter from the catalysts. Therefore, prior to reuse, the solids were calcined at either 450 ◦ C (ITQ-2) or 550 ◦ C (HMCM-22(24)) for 5 h (heating rate of 1 ◦ C min−1 ) to leave a residual amount of organic matter of less than 1 wt%; the higher temperature required for the complete combustion of the organic matter in the case of H-MCM-22(24) is consistent with the thermal analysis data. The Fur yields in four consecutive 6 h-batch runs were similar for H-MCM-22(24) and ITQ-2 (Fig. 9). When the used HMCM-22(24) catalyst was treated at 450 ◦ C instead of 550 ◦ C, the Fur yield decreased in recycling runs (Fig. 9), most likely due to the incomplete removal of organic matter; the calcined solid was very light brown in color and contained ca. 4 wt% organic matter.

250

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252

Fig. 9. Furfural yields in four consecutive 6 h-batch runs of the reaction of xylose in the presence of H-MCM-22(24) (regenerated by applying thermal treatment at either 450 or 550 ◦ C) or ITQ-2. Reaction conditions: 0.67 M xylose, xylose:catalyst mass ratio of 3:2, biphasic W-T solvent mixture, 170 ◦ C.

The powder XRD patterns (Fig. 1) and the Si/Al ratios (measured by ICP-AES) of the used/calcined H-MCM-22(24) and ITQ-2 catalysts were similar to those of the respective fresh catalysts (Si/Al = 25 and 24, respectively). The reaction of xylose in the presence of ITQ-2 at 155 ◦ C was slower (an initial reaction rate of 0.8 mmol h−1 gcat −1 ) and led to lower Fur yield (55% yield at 98% conversion, reached at 24 h reaction) than that observed at 170 ◦ C (initial reaction rate of 2.0 mmol h−1 gcat −1 ; 66% Fur yield at 99% conversion, reached at 16 h reaction). For the reaction temperature of 170 ◦ C, and at constant ITQ-2 catalyst bulk density and catalyst/xylose mass ratio, changing the W:T solvents ratio affects the Fur yields: for a W:T ratio of 0.5:0.5, 51% Fur yield was reached at 98% conversion and 16 h reaction, which is lower than that observed for the W:T ratio of 0.3:0.7. 3.2.3. Catalytic performances of H-MCM-22(24) and ITQ-2 using solely water as solvent The reaction of xylose was further investigated in the presence of ITQ-2 and H-MCM-22(24), using solely water as solvent (denoted W reaction system), at 170 ◦ C; the catalyst bulk density and catalyst/xylose ratio is the same as that for the W:T biphasic system (the initial concentration of xylose in water is lower in the case of the W reaction system). The kinetic profiles for ITQ-2 and HMCM-22(24) are very similar until 8 h reaction (Fig. 10(a)), which correlates with the similar total concentration of acid sites ([L] + [B]) for the two catalysts (Table 1). The Fur yield versus xylose conversion curves are comparable, giving Fur yield of 52–54% at 97% conversion (Fig. 10(b)). The comparable performances of these two catalysts using the W reaction system parallels that observed for the biphasic W-T system. For the two solvent systems, at reaction times longer than 8 h, ITQ-2 gives slightly higher conversions than H-MCM-22(24), possibly due to slower catalyst deactivation (coking) in the former case. The maximum Fur yields reached at high xylose conversions are somewhat lower when the simultaneous extraction of Fur is not performed, due to the enhanced formation of by-products. Accordingly, TGA analyses of the used catalysts revealed higher amounts of organic deposits using the W reaction system (18–21 wt% for the solids recovered after 24 h reaction) in comparison to the W-T one (12 wt%). The catalytic contribution to the decomposition of Fur was minor: less than 5%, obtained in

Fig. 10. Kinetic profiles of the reaction of xylose (a), and furfural yield versus xylose conversion curves (b), for H-MCM-22(24) (×), ITQ-2 (), and H2 SO4 (), used as catalysts. Reaction conditions: xylose:solid acid catalyst mass ratio of 3:2 (or 4 mM H2 SO4 ), solely water as solvent, 170 ◦ C.

separate catalytic tests using Fur as the substrate instead of xylose. Hence, Fur loss reactions may be essentially due to its reaction with xylose or intermediates of the reaction of xylose. To get insight into the type of by-products formed, the reaction of xylose was carried out in the presence of ITQ-2 using D2 O as solvent at 170 ◦ C for 24 h, and the reactions mixture was analysed by 1 H and 13 C NMR spectroscopy. The presence of xylose in the reaction solution was hardly detected in the spectra (not shown), which is consistent with the catalytic results (97% conversion at 24 h). The main peaks were relative to Fur: 1 H NMR, 9.4, 7.8, 7.5 and 6.6 ppm; 13 C NMR, 183.6, 155, 153.1, 128.7 and 116.2 ppm. Formic acid was detected (169 ppm and 8.1 ppm in the 13 C and 1 H NMR spectra, respectively), which may be formed via the decomposition of Fur [19,54–56]. The 1 H NMR spectrum exhibited weak to very weak signals in the range of 3–4.5 ppm which may be due to carbohydrate H–C–O or H2 C–O groups; weak peaks below 2 ppm may be assigned to methyl/methylene carbon atoms. Fragmentation reactions of xylose can take place to form oxygenated aliphatic compounds [56,57]. Scheme 2 shows possible pathways for the formation of by-products in the xylose-to-Fur reaction systems [1,19,54–56,58].

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252

251

Acknowledgements This work was partly funded by the FCT, POCTI and FEDER (project POCTI/QUI/56112/2004). The authors thank Dr U. D. Morales (Instituto de Tecnología Química, CSIC, Universidad Politécnica de Valencia) for the valuable help with the recipe of ITQ2; Marta A. C. Ferro for help with the TEM studies, and acknowledge the Portuguese network of electron microscopy, the RNME, FCT Project REDE/1509/RME/2005; Prof. C.P. Neto (University of Aveiro) for helpful discussions; Dr. F. Domingues (University of Aveiro) for access to HPLC equipment. S.L. (SRFH/BPD/23765/2005) and M.M.A. (SFRH/BD/61648/2009) are grateful to the FCT for grants. Scheme 2. Simplified representation of possible reaction pathways for the formation of by-products in the reaction of xylose to furfural, with reference to literature data [1,19,54–56,58].

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.apcata.2011.12.046.

For comparison, the reaction of xylose was carried out in the presence of 4 mM H2 SO4 as catalyst instead of the solid acids; the initial amount of liquid acid was comparable to the total amount of acid sites in the loaded solid acid catalysts. Although the kinetic profile until 8 h reaction was similar for H2 SO4 and the solid acid catalysts, H2 SO4 gave much lower conversions at longer reaction times (Fig. 10(a)). The pronounced retardation of the reaction of xylose in the presence of H2 SO4 may be due to the partial decomposition of the catalyst through its possible participation in the formation of sulfur-containing by-products [59,60] and/or to the decrease in the concentration of active Brönsted acid species due to the protonation of Fur [60]. The Fur yield versus conversion profile for H2 SO4 (55% Fur yield at 93% conversion) is comparable to those for the solid acid catalysts (Fig. 10(b)).

4. Conclusions The aqueous phase dehydration of xylose to furfural (Fur) was investigated under batch mode in the presence of H-MCM-22 zeolite and its delaminated counterpart (ITQ-2) possessing enhanced external surface area, using a biphasic water–toluene solvent mixture or solely water as the solvent (W-T and W reaction system, respectively), at 170 ◦ C. Fur yields of up to 71% and 54% are reached at more than 96% conversion for the W-T and W systems, respectively. Sulfuric acid as catalyst (4 mM; W reaction system) gives comparable Fur yields (55% at 93% conversion). Decreasing the Si/Al ratio (in the range of 38–24) of H-MCM22 increases the total amount of Lewis plus Brönsted acid sites ([L] + [B]), which leads to an improvement in the catalytic activity, without affecting significantly the Fur selectivity. For the two W and W-T reaction systems, the ITQ-2 catalyst exhibits comparable catalytic performance to its H-MCM-22 counterpart (with the same Si/Al of 24), which correlates with the similar total amounts of [L] + [B] acid sites of these materials. No structural modifications or leaching phenomena were detected for the used catalysts (thermally regenerated to remove organic deposits), and Fur yields in consecutive batch runs are similar. A difference between HMCM-22(24) and ITQ-2 is the less energy intensive conditions required for the thermal regeneration of ITQ-2 (some differences in the chemical nature of the carbonaceous matter may exist, based on DSC analyses), and an apparently slower catalyst deactivation by coking. Although the enhanced Sext /SBET ratio was effectively accomplished for ITQ-2 through the delamination procedure, a concomitant weakening of the surface acidity seems to have occurred, which should be avoided in order to optimise the catalytic performance and take the highest value/profit possible out of the application of a more refined catalyst preparation procedure.

References [1] K.J. Zeitsch, The Chemistry and Technology of Furfural and its Many ByProducts, vol. 13, first ed., Elsevier, The Netherlands, 2000 (Sugar Series). [2] J.-M. Lee, Y.-C. Kim, I.T. Hwang, N.-J. Park, Y.K. Hwang, J.-S. Chang, J.-S. Hwang, Biofuels Bioprod. Bioref. 2 (2008) 438–454. [3] “Avantium steps ahead with its Biofuels program. Engine test demonstrates potential of Furanics”, Avantium Research and Technology, Press Release 22 October 2007. [4] J. Ebert, Biomass Power and Thermal, September 2008, pp. 30–34. [5] R. Weingarten, G.A. Tompsett, W. Curtis Conner Jr., G.W. Huber, J. Catal. 279 (2011) 174–182. [6] Y. Jiang, X. Li, Q. Cao, X. Mu, J. Nanopart. Res. 13 (2011) 463–469. [7] A.S. Dias, M. Pillinger, A.A. Valente, J. Catal. 229 (2005) 414–423. [8] G.D. Yadav, R.V. Sharma, IN Patent 201002442-I3 (2010). [9] X.J. Shi, Y.L. Wu, H.F. Yi, G. Rui, P.P. Li, M.D. Yang, G.H. Wang, Energies 4 (2011) 669–684. [10] A.S. Dias, S. Lima, M. Pillinger, A.A. Valente, Carbohydr. Res. 341 (2006) 2946–2953. [11] A.S. Dias, M. Pillinger, A.A. Valente, Microporous Mesoporous Mater. 94 (2006) 214–225. ˜ N. Durán, Bioresour. [12] H.D. Mansilla, J. Baeza, S. Urzúa, G. Maturana, J. Villasenor, Technol. 66 (1998) 189–193. [13] A.S. Dias, S. Lima, D. Carriazo, V. Rives, M. Pillinger, A.A. Valente, J. Catal. 244 (2006) 230–237. [14] A.S. Dias, S. Lima, P. Brandão, M. Pillinger, J. Rocha, A.A. Valente, Catal. Lett. 108 (2006) 179–186. [15] A.S. Dias, S. Lima, M. Pillinger, A.A. Valente, Catal. Lett. 114 (2007) 151–160. [16] X. Shi, Y. Wu, P. Li, H. Yi, M. Yang, G. Wang, Carbohydr. Res. 346 (2011) 480–487. [17] S. Lima, A. Fernandes, M.M. Antunes, M. Pillinger, F. Ribeiro, A.A. Valente, Catal. Lett. 135 (2010) 41–47. [18] C. Moreau, R. Durand, D. Peyron, J. Duhamet, P. Rivalier, Ind. Crop Prod. 7 (1998) 95–99. [19] R. O’Neill, M.N. Ahmad, L. Vanoye, F. Aiouache, Ind. Eng. Chem. Res. 48 (2009) 4300–4306. [20] S. Lima, M.M. Antunes, A. Fernandes, M. Pillinger, M.F. Ribeiro, A.A. Valente, Molecules 15 (2010) 3863–3877. [21] S. Lima, M.M. Antunes, A. Fernandes, M. Pillinger, M.F. Ribeiro, A.A Valente, Appl. Catal. A: Gen. 388 (2010) 141–148. [22] S. Lima, M. Pillinger, A.A. Valente, Catal. Commun. 9 (2008) 2144–2148. [23] M.K. Rubin, P. Chu, US Patent 4,954,325 (1990). [24] J. Jae, G.A. Tompsett, A.J. Foster, K.D. Hammond, S.M. Auerbach, R.F. Lobo, G.W. Huber, J. Catal. 279 (2011) 257–268. [25] E. Sjöman, M. Mänttäri, M. Nyström, H. koivikko, H. Heikkil, J. Membr. Sci. 292 (1–2) (2007) 106–115. [26] A. Corma, V. Fornés, S.B. Pergher, J.G. Bouglas, T.L.M. Maesen, Nature 396 (1998) 353–356. [27] A. Corma, V. Fornés, J. Martínez-Triguero, S.B. Pergher, J. Catal. 186 (1999) 57–63. [28] A. Corma, U. Díaz, V. Fornés, J.M. Guil, J. Martínez-Triguero, E.J. Creyghton, J. Catal. 191 (2000) 218–224. [29] A. Corma, V. Fornés, S.B. Pergher, Pat WO 9717290 (1997). [30] S. Maheshwari, E. Jordan, S. Kumar, F.S. Bates, R.L. Penn, D.F. Shantz, M. Tspatsis, J. Am. Chem. Soc. 130 (2008) 1507–1516. [31] C. Delitala, M.D. Alba, A.I. Becerro, D. Delpiano, D. Meloni, E. Musu, I. Ferino, Microporous Mesoporous Mater. 118 (2009) 1–10. [32] V. Ayala, A. Corma, M. Iglesias, J.A. Rincón, F. Sánchez, J. Catal. 224 (2004) 170–177. [33] R. Schenkel, J.-O. Barth, J. Kornatowski, J.A. Lercher, Stud. Surf. Sci. Catal. 142 (2002) 69–76. [34] P. Frontera, F. Testa, R. Aiello, S. Candamano, J.B. Nagy, Microporous Mesoporous Mater. 106 (2007) 107–114.

252 [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48]

M.M. Antunes et al. / Applied Catalysis A: General 417–418 (2012) 243–252 U. Díaz, V. Fornés, A. Corma, Microporous Mesoporous Mater. 90 (2006) 73–80. C.E. Webster, R.S. Drago, M.C. Zerner, J. Am. Chem. Soc. 120 (1998) 5509–5516. C.A. Emeis, J. Catal. 141 (1993) 347–354. V.V. Narkhede, H. Gies, Chem. Mater. 21 (2009) 4339–4346. H.J. Jung, S.S. Park, C.-H. Shin, Y.-K. Park, S.B. Hong, J. Catal. 245 (2007) 65–74. P. Concepción, C. López, A. Martínez, V.F. Puntes, J. Catal. 228 (2004) 321–332. M. Hunger, S. Ernst, J. Weitkamp, Zeolites 15 (1995) 188–192. G.J. Kennedy, S.L. Lawton, A.S. Fung, M.K. Rubin, S. Steuernagel, Catal. Today 49 (1999) 385–399, and references cited therein. K. Okumura, M. Hashimoto, T. Mimura, M. Niwa, J. Catal. 206 (2002) 23–28. D. Mao, J. Xia, Q. Chen, G. Lu, Catal. Commun. 10 (2009) 620–624. H. Du, D.H. Olson, J. Phys. Chem. B 106 (2002) 395–400. N.S. Nesterenko, F. Thibault-Starzyk, V. Montouilliout, V.V. Yushchenko, C. Fernandez, J.-P. Gilson, F. Fajula, I.I. Ivanova, Kinet. Catal. 47 (2006) 40–48. I.I. Grandberg, G.K. Faizova, A.N. Kost, Khim. Geterotsiklich. Soedinenii 2 (1966) 561–566. J.B. Binder, J.J. Blank, A.V. Cefali, R.T. Raines, ChemSusChem 3 (2010) 1268–1272.

[49] C. Moreau, Agro-Food-Industry Hi-Tech 13 (2002) 17–26. [50] J.M. De Bruijn, Monosaccharides in Alkaline Medium: Isomerization, Degradation, Oligomerization, Ph.D. Thesis, TH Delft, 1986. [51] A. Corma, V. Martínez-Soria, E. Schnoeveld, J. Catal. 192 (2000) 163–173. [52] H. Talu, C.J. Guo, D.T. Hayhurst, J. Phys. Chem. 93 (1989) 7294–7298. [53] G. Sastre, C. Richard, A. Catlow, A. Corma, J. Phys. Chem. B 103 (1999) 5187–5196. [54] D.L. Williams, A.P. Dunlop, Ind. Eng. Chem. 40 (1948) 239–241. [55] A.P. Dunlop, Ind. Eng. Chem. 40 (1948) 204–209. [56] M.J. Antal Jr., T. Leesomboon, W.S. Mok, G.N. Richards, Carbohydr. Res. 217 (1991) 71–85. [57] G. Bonn, M. Rinderer, O. Bobleter, J. Carbohydr. Chem. 4 (1) (1985) 67–77. [58] R. Weingarten, J. Cho, W.C. Conner Jr., G.W. Huber, Green Chem. 12 (2010) 1423–1429, and references cited therein. [59] E. Gilbert, E.P. Jones, Ind. Eng. Chem. 53 (1961) 501–507. [60] J. Frost, V., Bui, J.W. Frost, Patents US2011046412-A1, WO2011022588-A1 (2010).