Microporous and Mesoporous Materials 181 (2013) 38–46

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials

journal homepage: www.elsevier .com/locate /micromeso

Dehydration of fructose to 5-hydroxymethyl furfural over orderedAlSBA-15 catalysts

1387-1811/$ - see front matter � 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.micromeso.2013.07.015

⇑ Corresponding author. Tel.: +91 20 25902019; fax: +91 20 25902633.E-mail address: sv.chilukuri@ncl.res.in (S. Chilukuri).

Nishita Lucas, Ganesh Kokate, Atul Nagpure, Satyanarayana Chilukuri ⇑Catalysis Division, National Chemical Laboratory, Dr Homi Bhabha Road, Pune 411 008, India

a r t i c l e i n f o

Article history:Received 8 October 2012Received in revised form 22 April 2013Accepted 9 July 2013Available online 17 July 2013

Keywords:FructoseDehydrationHydroxymethyl furfuralAlSBA-15Acidity

a b s t r a c t

5-Hydroxymethyl furfural is an important platform chemical. It is successfully synthesized from renew-able carbohydrates using mesoporous AlSBA-15 catalysts under biphasic conditions. Fine tuning of thecatalyst acidity is important to drive the reaction to give good yields of furan compound. Al-SBA-15catalysts with different Si/Al ratios were prepared and characterized by XRD, 27Al MASNMR, SEM, TEM,N2 sorption, ICP-OES and TPD of ammonia. Results show that part of aluminium is substituted intotetrahedral positions. The catalyst with lower acid site density but medium to strong acid strength favoursselective formation of HMF. Under the optimized conditions, HMF selectivity was as high as 88% at 59mol% conversion of fructose. Leaching of part of aluminium occurs under hydrothermal conditions, ifsolvent mixtures containing water are used, while activity can be retained if DMSO is used as solvent.

� 2013 Elsevier Inc. All rights reserved.

1. Introduction

It is well accepted that present petroleum based chemicals needto be replaced with renewables as a result of the dwindling re-serves of the former. Moreover, rapid ecological changes takingplace as a result of fossil fuel use makes it imperative to replacethem with renewable fuels to attain a sustainable and renewablesociety. In this context, effective utilization of biomass for the pro-duction of fuels and chemicals has emerged as an important area ofresearch. Saccharides have attracted greater attention as a promis-ing carbon based alternative source as it is a renewable chemicalfeedstock [1,2]. Hence, their conversion to useful chemicals hasbeen intensely explored [3]. However, problems of low volatility,poor solubility in desired medium and polyfunctionality of carbo-hydrates is hindering their effective conversion to value addedchemicals. This necessitates development of innovative methodsand catalyst systems to facilitate and accelerate their utilization.Among the various transformations, the synthesis of 5-hydroxym-ethylfurfural (HMF) through dehydration of hexoses is extensivelystudied due to importance of HMF as a platform chemical [4]. Thepotential applications of HMF includes its selective oxidation to2,5-furandicarboxylic acid which can be used as a replacementfor terephthalic acid in the production of polyesters [5–8], itsreduction to 2,5-dihydroxymethylfuran and 2,5-bis(hydroxy-methyl)tetrahydrofuran [9,10], which can serve as alcohol compo-nents in the production of polyesters, its hydrogenolysis to

dimethyl furan, with excellent fuel properties to replace gasoline[11], its rehydration to levulinic acid an important chemical build-ing block with versatile applications [12,13]. Moreover, numerousphenolic resins and polymerizable furanic compounds with prom-ising properties have been prepared from HMF [14,15].

As a result of above described versatile applications, successfulefforts were made to produce HMF using a variety of reaction sys-tems e.g., ionic liquids [16–19], high boiling organic solvents likedimethylsulfoxide [20–23], water [24] and biphasic systems[25,26], using both homogeneous and heterogeneous catalysts. Io-nic liquid based systems have emerged as excellent mode for selec-tive formation of HMF but suffer from drawbacks such as masstransfer limitations and high cost. Reactions in water medium offera green protocol for HMF preparation, but HMF yields are lower asrehydration to LA is favorable. Whereas, in DMSO, formation of LAis suppressed leading to high yields of HMF. However, this ap-proach necessitates difficult and energy intensive product isolationprocedures [25]. In this context, biphasic systems are superior interms of high HMF yields and energy efficiency. Currentlyresearchers are focusing on the use of mesoporous materials as cat-alysts or catalyst supports for such reactions, as the mesoporosityis found to be beneficial, especially when reactant/product mole-cules are bulky [27]. In addition to mesoporosity, moderate acidity,pore structure along with their good thermal stability offers fur-ther possibilities for designing a more efficient and environmentalfriendly process for the production of HMF.

Visualizing the above mentioned potential of mesoporousmaterials, we have explored aluminium incorporated SBA-15 forthe dehydration of fructose to HMF. We have incorporated

N. Lucas et al. / Microporous and Mesoporous Materials 181 (2013) 38–46 39

aluminium into SBA-15 by post synthesis methods. The acidity wastuned by varying the Al content in the sample to study its influenceon the reaction. The synthesized materials were subjected to de-tailed characterization to help in understanding the reactivityand selectivity during dehydration of fructose. Various processparameters were also optimized for the dehydration reaction withan aim to improve the HMF selectivity and yield. Attempt wasmade to correlate the acidity with HMF yield.

2. Experimental

2.1. Materials

Fructose, glucose, sucrose, maltose, methyl isobutyl ketone(MIBK), 25% NH3 solution, zirconium oxychloride (ZrOCl2.8H2O),dodecatungstophosphoric acid (hereafter TPA) and toluene werepurchased from Loba Chemie Pvt. Ltd., aluminium chloride, tetra-ethyl orthosilicate (TEOS) and poly(ethylene glycol)-block-poly(propylene glycol)-block-poly(ethylene glycol) (averagemolecular weight, 5800), were procured from Sigma–Aldrich, USA.All the chemicals were of research grade and were used after dryingfollowing standard procedures. Zeolites NH4

+-mordenite (Si/Al = 10) and NH4

+-beta were (Si/Al = 19) obtained from ZeolystInternational, USA. Zeolite NH4

+-ZSM5 (Si/Al = 23) was obtainedfrom catalyst pilot plant, National Chemical Laboratory, India. Thesewere calcined at 450 �C for 4 h in air before using for the reaction.

2.2. Catalyst preparation

2.2.1. Synthesis of SBA-15Mesoporous siliceous SBA-15 was synthesized according to the

procedure reported by Stucky and co-workers [28]. About 4 g ofamphiphilic triblock copolymer, poly(ethylene glycol)-block-poly(propylene glycol)-block-poly(ethylene glycol) (average molec-ular weight, 5800) was dispersed in 30 mL of water, to which 120 g of2 M HCl solution was added to get a homogeneous solution. To thissolution, 8 g of TEOS was added slowly under stirring. This mixturewas continuously stirred at 40 �C for 24 h and transferred to a Tef-lon-lined autoclave for crystallization at 100 �C for 2 days. The solidobtained after crystallization was filtered, washed with distilledwater and dried at 100 �C for 12 h. The crystalline material was thencalcined in flowing air at 550 �C for 6 h in order to decompose and re-move the tri-block copolymer to yield mesoporous SBA-15.

2.2.2. Synthesis of AlSBA-15A post-synthesis route similar to the procedure reported by

Luan et al. [29] was employed for the preparation of AlSBA-15. Ina typical synthesis, anhydrous AlCl3 was dissolved in dry ethanol(25 mL). To this solution, SBA-15 (1 g) was added and refluxedfor 10 h under stirring. The material was then filtered, washedrepeatedly with dry ethanol and dried at room temperature fol-lowed by calcination at 550 �C for 5 h to give Al-SBA15 catalysts.By varying the AlCl3 content, a series of AlSBA-15 catalysts (withSi/Al ratio ranging 10–40) were prepared.

2.2.3. Synthesis of 15 wt% TPA on ZrO2

This catalyst was prepared by wet impregnation method fol-lowing the procedure described elsewhere [30]. The support zirco-nium oxyhydroxide was prepared by adding 5 wt% NH3 solution to0.5 M solution of ZrOCl2.8H2O, drop wise under continuous stirringtill a final pH of 9.5 was obtained. The precipitate was filtered,washed repeatedly (till free from chloride ions), dried at 110 �Cfor 14 h and crushed to get white zirconium oxyhydroxide powder.To methanolic solution of TPA, zirconium oxyhydroxide was addedand stirred for 12 h at room temperature. The solvent was

evaporated to dryness, dried at 80 �C overnight and calcined at750 �C to get 15 wt% TPA/ZrO2-750 (hereafter 15TZ750).

2.3. Characterization

The specific surface area of the catalysts was measured by N2

physisorption at liquid nitrogen temperature using a Quanta-chrome Nova-1200 surface area analyzer and standard multi pointBET analysis method. Samples were degassed at 300 �C in vacuumfor 2 h before N2 physisorption measurements. Pore volumes werecalculated using the t-plot method of De Boer. The BJH method andthe corrected Kelvin equation was used to determine the pore sizedistributions from the adsorption and desorption branches of theisotherms [31].

Low-angle X-ray diffraction patterns were collected in the 2hrange 0.5–5�, with a Philips X’pert Pro diffractometer using Ni fil-tered CuKa (k = 1.5406 Å, 40 kV, 30 mA), radiation, with an X’celer-ator detector, using the real time multiple strip (RTMS) detectiontechnique, at a scan rate of 0.5� min�1.

The acidity of the catalysts was investigated by temperature-programmed desorption of NH3 (NH3–TPD) using a MicromeriticsAutochem-2920 instrument. Prior to TPD run, the sample was acti-vated at 450 �C in He flow (40 mL min�1) for 1 h. Subsequently, thetemperature was brought down to 80 �C and NH3 was sorbed byexposing the samples to a stream of 10% NH3 in He (30 mL min�1)for 0.5 h. The temperature was then raised to 100 �C and flushedwith He for 1 h at 100 �C to remove the physisorbed NH3. Thedesorption of NH3 was carried out in He flow (40 mL min�1) byincreasing the temperature to 550 �C at 10 �C min�1, while moni-toring the concentration of NH3 desorbed using a thermal conduc-tivity detector.

Scanning electron microscopy of the samples was carried outusing Quanta 200, 3D instrument model of FEI. The samples wereprepared by dispersing them ultrasonically in isopropyl alcohol,transferring a portion onto silicon wafers which were dried andsubjected to gold coating in vacuum.

Transmission electron microscopy of the samples was carriedout using FEI Technai TF-30 instrument operating at 300 kV. Thesamples for TEM measurements were prepared by placing a drop-let of the sample suspension, prepared in isopropyl alcohol usingultrasonication on a carbon coated copper grid by leaving themfor drying at room temperature.

27Al Magic-angle spinning (MAS) NMR spectra were recordedon a Bruker DSX300 spectrometer at 7.05 T magnetic field. Thesample was rotated at 6000 Hz at the magic angle, while collectingthe spectral data. A delay of 2 S was maintained between two 45�pulses. External Al(H2O)6

3+ was used as a reference.

2.4. Catalytic activity

The dehydration reaction was carried out in 300 mL Parr (SS316) autoclave with a Teflon liner. In a typical reaction, 1 g of fruc-tose was dissolved in 10 mL of Millipore water, to which 50 mL ofMIBK and freshly activated catalyst were added. The closed vesselwas purged with nitrogen and the reaction was conducted at thedesired temperature, while stirring the mixture constantly at500 RPM. At the end of reaction, the aqueous phase was filteredusing nylon 0.22 lm filter and the filtrate was analyzed usingHPLC, equipped with RI detector and H+ Aminex column(305 mm � 7.8 mm) with 0.1% H3PO4 as the mobile phase at a flowrate of 0.6 mL min�1. The organic phase was analyzed by Varian3800 gas chromatograph, equipped with flame ionization detectorand a HP-5 capillary column (50 m � 0.32 mm � 1.05 lm filmthickness). The products were identified by GC–MS and comparedwith authentic samples.

Fig. 1. N2 adsorption and desorption isotherm of (a) AlSBA-15 (Si/Al = 10), (b)AlSBA-15 (Si/Al = 20), (c) AlSBA-15 (Si/Al = 30), (d) AlSBA-15 (Si/Al = 40) and (e)SBA-15.

40 N. Lucas et al. / Microporous and Mesoporous Materials 181 (2013) 38–46

3. Results and discussion

3.1. Characterization of the catalysts

3.1.1. Surface areaTextural characteristics of the catalyst materials such as specific

surface area, pore volume and pore size distribution were obtainedfrom low temperature (�196 �C) nitrogen adsorption isotherms,the results of which are given in Table 1. The adsorption–desorp-tion isotherms are shown in Fig. 1. All isotherms were of type IV,as defined by IUPAC and exhibit a H1-type broad hysteresis loop.The presence of large mesopores is confirmed from the sharp steeprise at a relative pressure (P/P0) of 0.63–0.9, for all isotherms due tothe capillary condensation of nitrogen within uniform mesoporeswhere the P/P0 position of the inflection point is correlated to thediameter of the mesopore. The synthesized SBA-15 has a high sur-face area of 808 m2 g�1, with an average mesopore size of 60.5 Å(calculated from Kelvin equation). The surface areas and the porevolumes of aluminium containing AlSBA-15 samples were lowerthan parent SBA-15. The pore volume also decreased with increas-ing aluminium content, probably as a result of non-framework Alspecies as evidenced from 27Al NMR results. The presence of theseextra framework species also leads to reduction in the surface area.However, no particular trend was observed in the variation of sur-face area values, which are in accordance with earlier findings [32].

3.1.2. X-ray diffractionThe XRD powder patterns of the AlSBA-15 (Si/Al = 10, 20, 30, 40)

samples, prepared through secondary synthesis are shown in Fig. 2.These consist of three well resolved peaks in the 2h range of 0.8–1.8�, which are similar to those recorded for SBA-15 in literature[28]. The presence of one strong reflection of (100) plane at2h = 0.8 and two weaker reflections of (110), (200) at higher 2hare associated with the p6 mm hexagonal symmetry in the materi-als. The similarity in XRD pattern of AlSBA15 and SBA-15 clearlyshows that with the post modification (alumination), no perturba-tion of the hexagonal mesoporous structure occurred. Additionally,the XRD profile of the samples does not contain any other phase oramorphous material.

3.1.3. Electron microscopyThe morphology of samples obtained using SEM is shown in

Fig. 3. The rope like morphology of the parent SBA-15 is retained

Table 1Textural and physio-chemical properties of catalysts in dehydration of fructose.

Catalyst Si/Al$ S.A. (m2/g) P.D. (Å)

SBA-15 – 808 60.5Al-SBA15 (Si/Al = 10) 19 549 53.7Al-SBA15 (Si/Al = 20) 42 560 54.1Al-SBA15 (Si/Al = 30) 58 541 53.9Al-SBA15 (Si/Al = 40) 65 611 60.1H-Beta 19 700 5.7 � 7.5 (linear)

6.5 � 5.6 (tortuous)[36]H-Mordenite 10 490 6.5 � 7.0

2.6 � 5.7[36]H- ZSM5 23 350 0.54 � 0.56

0.55 � 0.51[36]15TZ750 – 53 –

Reaction conditions: fructose = 1 g (in 10 mL water), temperature = 165 �C, catalyst = 0.1Legend: S. A. = surface area, P. D. = pore diameter, P. V. = pore volume, Conv. = conversio

$ Determined by ICP-OES.* Time = 35 min.+ Time = 45 min.# Time = 50 min.

even after alumination. All the AlSBA-15 samples with varyingSi/Al ratio retain the morphology indicating that morphology re-mains unchanged with variation in the aluminium content.

TEM images of the catalysts (Fig. 4) clearly demonstrate theretention of the periodic structure, even after modification withaluminium, thus preserving the long range hexagonal orderedmesopores, which confirms the earlier findings of X- ray diffractionand SEM.

3.1.4. TPD of NH3

The acidity of the AlSBA-15 samples was determined by TPD ofNH3. The TPD profiles of the samples with different Si/Al ratios aregiven in Fig. 5. The total acidity of materials is expressed inmmol g�1 of desorbed NH3 under each temperature maximumand these values are listed in Table 1. All AlSBA-15 samples showtwo peaks, one pertaining to weak acidity (100–250 �C) and theother to moderately strong acidity (250–400 �C). With the decreasein the Si/Al ratio (or increasing aluminium content) from 40 to 10,the total acidity, in terms of ammonia desorbed, increased from

P.V. (cc/g) Acidity (mmol/g) Fructose conv. (mol %) HMF sel. (%)

0.870 0.002 – –0.727 0.267 68 590.703 0.243 65 650.685 0.188 63 780.784 0.147 59 88– 0.796 94 48

62* 501.380 79 49

– 61+ 52– 0.570 70 40

58# 41– 0.290 84 28

63+ 30

5 g, water: MIBK = 1:5 (v/v), reaction time = 1 h, RPM = 500.n, Sel. = selectivity.

Fig. 2. X-ray diffractograms of (a) AlSBA-15 (Si/Al = 10), (b) AlSBA-15 (Si/Al = 20),(c) AlSBA-15 (Si/Al = 30), (d) AlSBA-15 (Si/Al = 40) and (e) SBA-15.

N. Lucas et al. / Microporous and Mesoporous Materials 181 (2013) 38–46 41

0.147 to 0.267 mmol g�1. This shows that the acidity increasedwith aluminium content. The parent mesoporous SBA-15 has veryweak acidity due to the presence of surface silanol groups. How-ever, this acidity is very low as compared to aluminium containingAlSBA-15 samples, suggesting the formation of acid sites on postmodification with aluminium. Thus the total acidity is dependenton the amount of aluminium present in the samples, but the acid-ity did not increase linearly with aluminum content, particularly inthe higher aluminium containing samples.

3.1.5. 27Al MAS NMRSolid state 27Al MASNMR is an excellent technique to probe the

local environment of aluminium species. The spectra of the AlSBA-15 samples are shown in Fig. 6. All samples show two peaks, onepeak at 53 ppm is indicative of aluminium in tetrahedral coordina-tion, in which it is covalently bound to four Si atoms via oxygenbridges and the other peak at nearly 0 ppm is attributed to octahe-drally coordinated aluminium species in extra framework posi-tions [33]. With the decreasing Si/Al ratio (increasing Al content),higher proportion of aluminium is substituted in the octahedralcoordination, suggesting lower incorporation of Al into the tetrahe-dral framework and the presence of extra framework/non-frame-work species at higher Al loadings. It is expected that onlytetrahedral Al generates Bronsted acid sites [34]. Hence, 27AlMASNMR results show that the acidity is not expected to increase

Fig. 3. Scanning electron microscopy of (a) AlSBA

linearly with Al content. The TPD of NH3 results confirm thisaspect.

3.2. Catalytic activity in dehydration

The dehydration of fructose to HMF was used as a model reac-tion for testing the catalytic activity and selectivity of various cata-lysts. This reaction can yield numerous products (Scheme 1), but wehad targeted for high HMF selectivity. The results of the catalyticactivity are given in Table 1. We have used three zeolites for thisreaction, namely H-beta (Si/Al = 19), H-ZSM5 (Si/Al = 23) and H-mordenite (Si/Al = 10), the later is reported to be selective forHMF in biphasic conditions [35]. Under the reaction conditions em-ployed, the highly acidic zeolites gave higher fructose conversion(94, 79 and 70 mol% for beta, mordenite and ZSM5 respectively)but resulted in lower selectivity for HMF (�40–50%). The higherconversion on H-beta can be attributed to its stronger acidity andthree dimensional pore structure. Mordenite has two dimensionalpore system, 12-membered ring large pores and 8-membered ringsmall pores, while H-beta has three dimensional pore system withall 12-membered ring large pores [36]. This along with smallercrystallite size of H-beta facilitates easier diffusion of reactants/products, leading to high conversion of fructose. Though ZSM-5has three dimensional pore structure along with strong acidity, itsmedium pores may be playing a role in lowering the conversion.Moreover, selectivity is affected because of strong acidity and con-finement of reactants/products in the medium pores of the zeolitefor longer duration. Even, increasing the stirring speed from 500to 800 RPM had only a marginal effect with the conversion increas-ing from 70 to 73 mol%). The results of the catalytic activity of zeo-lites are in good agreement with the earlier findings which showsthat targeted furan compounds undergo humin formation in theintrazeoloitic channels thus lowering its yield [37]. Hence, the poorefficiency of this zeolite can be attributed to its strong acidity andlocal constraints within the pore system. In contrast, the mesopor-ous materials have weaker and lower acidity than zeolites, but gavesuperior catalytic performance. In order to thoroughly investigatethe superiority in catalytic activity of mesoporous AlSBA-15, theselectivity of other catalyst were compared at similar conversionas that of AlSBA-15 (Si/Al = 40) and results are shown in Table 1.In case of zeolites, HMF selectivity was poor even at lower conver-sion levels compared to mesoporous catalysts. The inferior catalyticperformance of zeolites is attributed to strong acidity as it has beenobserved that heterocyclic molecules reflect specific interactionswith Bronsted acid sites undergoing strong adsorption and oligo-merization in the zeolitic channels [38]. Thus weaker acidity of AlS-BA-15 prevents further reaction of the targeted furan compound,resulting in its selective formation. On the other hand, when we

-15 (Si/Al = 10) and (b) AlSBA-15 (Si/Al = 40).

Fig. 4. Transmission electron microscopy of (a) AlSBA-15 (Si/Al = 10) and (b) AlSBA-15 (Si/Al = 40).

Fig. 5. Temperature programmed desorption of (a) AlSBA-15 (Si/Al = 10), (b) AlSBA-15 (Si/Al = 20), (c) AlSBA-15 (Si/Al = 30) and (d) AlSBA-15 (Si/Al = 40).

Fig. 6. 27Al MAS NMR of (a) AlSBA-15 (Si/Al = 10), (b) AlSBA-15 (Si/Al = 20), (c)AlSBA-15 (Si/Al = 30) and (d) AlSBA-15 (Si/Al = 40).

O

H

HO

H

HO

H

H

OHHOH

OH

H

CH2OH

H

HOH2C

H H

HO OH

O

OOHC CH2OH

HMF

LevulinicAcid

Humins

Polymerization

Isomerization

Fructose

Glucose

Dehydration

Oligomers

CondenastionRehydration

H3C COOH

O

HCOOHFormicAcid

Scheme 1. Dehydration of C-6 sugars to HMF.

42 N. Lucas et al. / Microporous and Mesoporous Materials 181 (2013) 38–46

have employed 15TZ750 (tungstated zirconia) as catalyst, fructoseconversion was 84 mol% while HMF selectivity was only 28%. Atlower conversion (62 mol%) the selectivity was almost the same.This shows that surface acidity has to be optimum for attaininggood HMF selectivity. The superiority of the aluminated SBA-15 cat-alysts with tunable acidity can be inferred from the results. For theAlSBA-15 with varying Si/Al ratio, the conversion of fructose wasfound to decrease from 68 to 59 mol% with increasing Si/Al ratio,but the HMF selectivity increased from 59% to 88%. The reductionin conversion can be explained with decrease in total acidity as a re-sult of less aluminium content with increasing Si/Al ratio of the cat-alyst samples. The HMF selectivity increased with decrease in thenon-framework aluminium at higher Si/Al ratio (as evident from27Al MAS NMR profile). It is known that non-framework aluminiumcations in porous materials acts as Lewis acid centers [39–41],which accelerate the formation of humins from the carbohydrateand furan compounds [37], thereby reducing the HMF selectivity.Selectivity to HMF was plotted against acidity for understandingthe role of acidity on the reaction. Fig. 7 shows plot of total acidityagainst fructose conversion and HMF selectivity. The plot clearly

Fig. 7. Effect of total acidity on fructose conversion and HMF selectivity.

N. Lucas et al. / Microporous and Mesoporous Materials 181 (2013) 38–46 43

shows that though the conversion increased with total acidity,selectivity to HMF follows exactly opposite trend. To understandthe variation of HMF selectivity with weak and strong acidity, thesewere plotted against HMF selectivity as a ratio to total acidity, as gi-ven in Fig. 8. It is interesting to see that HMF selectivity increasedagainst moderately strong acidity/total acidity ratio showing thatmoderately strong acidity contributes to selective formation ofHMF (Fig. 8b). On the other hand, against weak acidity/total acidityratio (Fig. 8a), it follows opposite trend. This clearly shows that highacidity and weaker acid sites do not aid HMF selectivity, though thehigher acidity leads to increased fructose conversion. As a result,among the samples studied, AlSBA-15 (Si/Al = 40) showed maxi-mum HMF selectivity. Hence, it was used for further optimizationof reaction conditions.

The liquid phase dehydration of fructose with AlSBA-15 (Si/Al = 40) was studied by varying the reaction temperatures in the155–185 �C range; the results are shown in Fig. 9. Increasing tem-perature from 155 to 185 �C, leads to rise in fructose conversionfrom 59 to 91 mol% respectively. The product distribution shows

Fig. 8. HMF selectivity versus (a) weak acidity/total acidity ra

that at higher temperature, selectivity to HMF decreased from88% to 54%, whereas selectivity for others most of it being huminsand oligomers has increased (Scheme 1). Additionally, very low LAselectivity was seen (<5%); as MIBK suppresses the rehydration ofHMF [42]. Furfural which can be obtained by decarboxylation ofHMF [35] was almost insignificant. At higher reaction tempera-tures the color of the reaction mixture darkened, which may beattributed to the presence of soluble oligomers. Thus it can be con-cluded that at higher reaction temperatures, in presence of MIBK,formation of humins and oligomerization is dominant among theother side reactions.

The effect of catalyst weight on fructose dehydration reactionis shown in Fig. 10. It is observed that when the catalyst amountincreased from 0.075 g to 3 g the conversion also increased from33 to 77 mol%. However, the HMF selectivity of 88% was constantup to a catalyst weight of 0.15 g, which then decreased with fur-ther increase in catalyst content. The increase in conversion isattributed to increase in the number of active sites; while avail-ability of more active sites drives the reaction towards more con-densation products. Furfural and LA are not formed in significantquantities clearly demonstrating that the acidity is not suitablefor their formation. But, availability of more active sites appearsto favor condensation products thus decreasing the selectivityof HMF. Hence, 0.15 g of catalyst was found to be the optimalfor this reaction.

The effect of solvent on the reaction is studied by using differentpolar and non-polar solvents; results are shown in Fig. 11. Thefructose conversion is almost same for MIBK and toluene (59 and58 mol% respectively), however in water the conversion of fructosewas lower (49 mol%), as water is a polar solvent and a Lewis base, ithas high affinity for the active acidic sites leading to blockage ofacid sites thereby reducing conversion. Among the various solventsMIBK gave maximum HMF selectivity (88%) since its polar naturemaximizes the dissolution of HMF. Water being most polar solventgives less selectivity of HMF (52%) and more of LA (18.5%) showingthat rehydration of HMF to LA is dominant in water. With toluene,more of insoluble humins and soluble oligomers were obtained asthe selectivity of others increased. The values of extraction coeffi-cient R = HMForg/HMFaq is 3.2, 0.236 and 0 for MIBK, toluene andwater respectively, revealing that MIBK is most efficient extractingsolvent [43].

tio and (b) moderately strong acidity/total acidity ratio.

Fig. 9. Effect of temperature on dehydration of fructose. Reaction conditions:Fructose = 1 g (in 10 mL water), catalyst = 0.15 g, water: MIBK = 1:5 (v/v), reactiontime = 1 h, RPM = 500.

Fig. 10. Effect of catalyst weight on dehydration of fructose. Reaction conditions:fructose = 1 g (in 10 mL water), temperature = 165 �C, water: MIBK = 1:5 (v/v),reaction time = 1 h, RPM = 500.

Fig. 11. Effect of solvent on dehydration of fructose. Reaction conditions: fruc-tose = 1 g (in 10 mL water), temperature = 165 �C, catalyst = 0.15 g, water: organicsolvent = 1:5 (v/v), reaction time = 1 h, RPM = 500.

Fig. 12. Effect of time on dehydration of fructose. Reaction conditions: Fruc-tose = 1 g (in 10 mL water), temperature = 165 �C, catalyst = 0.15 g, water:MIBK = 1:5 (v/v), RPM = 500.

44 N. Lucas et al. / Microporous and Mesoporous Materials 181 (2013) 38–46

The effect of time on reaction was also investigated and the re-sults are shown in Fig. 12. It was seen that with the increase in timefrom 0.5 to 2 h conversion increased from 44 to 86 mol%, whereasthe HMF selectivity decreased from 88% to 57%. This shows thatwith increase in duration of the reaction, the HMF formed gets de-graded or undergoes further condensation reactions leading to areduction in selectivity.

In order to evaluate the effectiveness of the catalytic system toconvert other saccharides to HMF, various carbohydrates wereused for dehydration reaction, the results are shown in Fig. 13. Itwas found that fructose gave maximum selectivity of HMF, i.e.88%, while sucrose; a disaccharide of fructose also gave moderateselectivity of 43%. Glucose and maltose gave lower conversionand HMF selectivity (28%) since the isomerization of glucose to

fructose is slow over acid catalysts, it reduces the dehydration.Thus ketose sugars can be selectively converted to HMF whereasaldose sugars scarcely yielded the furan compound.

To check the recyclability of the catalyst, the used catalyst wascalcined in air at 550 �C for 5 h in order to oxidize the carbona-ceous species deposited on the catalyst (C = 13.2%, H = 1.5%,N = nil, estimated by micro analysis). The calcined (AlSBA-15, Si/Al = 40) catalyst was used for the reaction that gave a fructoseconversion of 51 mol% and HMF selectivity of 59%. In order toinvestigate the reasons for this decrease in catalytic activity, thecatalyst was subjected to 27Al MAS NMR and XRD studies, the re-sults are shown in Figs. 14 and 15 respectively. From the NMRstudies it can be seen that the peak around 0 ppm assigned tooctahedral aluminium is absent, showing the leaching of non-framework species. As a result of aluminium leaching, the overallcatalytic activity detoriated. The XRD patterns of fresh and used

Fig. 13. Effect of saccharide on dehydration. Reaction conditions: saccharide = 1 g(in 10 mL water), temperature = 165 �C, catalyst = 0.15 g, water: MIBK = 1:5 (v/v),reaction time = 1 h, RPM = 500.

Fig. 14. 27Al MAS NMR of (a) fresh catalyst and (b) used catalyst.

Fig. 15. X-ray diffraction pattern of (a) fresh catalyst (b) used catalyst.

Table 2Recyclability studies of AlSBA-15(Si/Al = 40).

S.N. Recycle Fructose conv. (mol%) HMF Sel. (%)

1 – 96 772 1st 91 803 2nd 80 754 3rd 79 76

Reaction conditions: fructose = 0.5 g, DMSO = 20 mL, temperature = 120 �C, catalystweight = 0.075 g, time = 2 h, RPM = 500.

N. Lucas et al. / Microporous and Mesoporous Materials 181 (2013) 38–46 45

catalysts reveal that the peaks in the used catalyst shifted tohigher 2h values for all the three reflections indicating the struc-tural deformation, as a result of hydrothermal treatment. WhenDMSO was used as solvent under milder conditions, fructose con-version of 98 mol% and HMF selectivity of 76% was obtained. Theused catalyst was washed with acetone, dried and used for nextcycle. The results of recyclability of AlSBA-15 (Si/Al = 40) in DMSOare shown in Table 2. There was a slight drop in fructose conver-sion; with only little change in HMF selectivity’s. Thus, the cata-lyst could be recycled effectively up to 3 cycles without much lossin catalytic activity. These results show the efficacy of AlSBA-15catalysts for dehydration of saccharides, though the recyclabilityof catalysts in mixed solvents containing water is not good. Keep-ing this aspect in mind, utilization of this catalyst system fordevelopment of continuous process employing milder reactionconditions and low boiling solvents as reported recently [44] willbe explored in near future.

4. Conclusions

The selective formation of HMF, an important platform chemi-cal has been achieved using aluminium incorporated AlSBA-15 cat-alysts. The effect of Si/Al ratio on the catalytic activity andselectivity was systematically investigated. A good linear correla-tion between the moderately strong acidity/total acidity ratioand HMF selectively was obtained. These findings demonstratedthat the presence of moderately strong acidity is crucial for selec-tive dehydration of fructose. AlSBA-15 works truly as heteroge-neous catalyst if milder reaction conditions and non aqueoussolvent systems are employed.

Acknowledgements

Nishita Lucas Atul Nagpure acknowledges Council of Scientificand Industrial Research, New Delhi, for providing senior researchfellowship.

References

[1] P. Gallezot, Green Chem. 9 (2007) 295–302.[2] F.W. Lichtenthaler, S. Peters, C.R. Chim. 7 (2004) 65–90.[3] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502.[4] A. Rosatella, S. Simeonov, R. Frade, C. Afonso, Green Chem. 13 (2011) 754–793.[5] S.E. Davis, L. Houk, E.C. Tamargo, A.K. Datye, R.J. Davis, Catal. Today 160 (2011)

55–60.[6] N.K. Gupta, S. Nishimura, A. Takagaki, K. Ebitani, Green Chem. 13 (2011) 824–

827.[7] M.A. Lilga, R.T. Hallen, M. Gray, Top. Catal. 53 (2010) 1264–1269.[8] O. Cassanove, S. Iborra, A. Corma, ChemSusChem 2 (2009) 1138–1144.[9] T. Thannatthanachon, T.B. Rauchfuss, Angew. Chem. Int. Ed. 49 (2010) 6616–

6618.[10] W. Yang, A. Sen, ChemSusChem 3 (2010) 597–603.

46 N. Lucas et al. / Microporous and Mesoporous Materials 181 (2013) 38–46

[11] Y.R. Leshkov, C.J. Barett, Z.Y. Liu, J.A. Dumesic, Nat. Lett. 447 (2007) 982–986.[12] R. Leonard, Ind. Eng. Chem. 48 (1956) 1331–1341.[13] V. Ghorpade, M. Hanna, US Patent 5859263 (1996).[14] A. Gandini, M.N. Belcacem, Prog. Polym. Sci. 22 (1997) 1203–1379.[15] H. Koch, F. Krause, R. Steffan, H.U. Woelk, Starke 35 (1983) 304–313.[16] J.Y. Chan, Y. Zhang, ChemSusChem 2 (2009) 731–734.[17] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith, Green Chem. 11 (2009) 1327–1331.[18] T. Stalberg, M.G. Sorensen, A. Riisager, Green Chem. 12 (2010) 321–325.[19] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Science 316 (2007) 1597–1599.[20] K. Shimizu, R. Uozumi, A. Satsuma, Catal. Commun. 10 (2009) 1849–1853.[21] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith, Ind. Eng. Chem. Res. 47 (2008) 9234–

9239.[22] H. Yan, Y. Yang, D. Tong, X. Xiang, C. Hu, Catal. Commun. 10 (2009) 1558–1563.[23] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith, Catal. Commun. 10 (2009) 1771–

1775.[24] C. Carlini, P. Patrono, A.M. Galleti, G. Sbrana, Appl. Catal. A 275 (2004) 111–

118.[25] Y.R. Leshkov, J.A. Dumesic, Top. Catal. 52 (2009) 297–303.[26] Y. Román-Leshkov, J.N. Chheda, J.A. Dumesic, Science 312 (2006) 1933–1937.[27] O. Casanova, S. Iborra, A. Corma, J. Catal. 275 (2010) 236–242.[28] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fedrickson, B.F. Chemlka, G.D. Stucky,

Science 279 (1998) 548–552.[29] Z. Luan, M. Hartmann, D. Zhao, W. Zhou, L. Kevan, Chem. Mater. 11 (1999)

1621–1627.

[30] B.M. Devassy, S.B. Halligudi, S.G. Hegde, A.B. Halgeri, F. Lefebvre, Chem.Commun. 10 (2002) 1074–1075.

[31] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380.[32] G.V. Shanbhag, T. Joseph, S.B. Halligudi, J. Catal. 250 (2007) 274–282.[33] H. Kosslick, G. Lischke, G. Walther, W. Storek, A. Martin, R. Fricke, Microporous

Mater. 9 (1997) 13–33.[34] R.A. Beyerlein, C. Choi-Feng, J.B. Hall, B.J. Huggins, G.J. Ray, Top. Catal. 4 (1997)

27–42.[35] C. Moreau, R. Durand, S. Razigade, J. Duhamet, P. Faugeras, P. Rivalier, P. Ros, G.

Avignon, Appl. Catal. 145 (1996) 211–244.[36] W.M. Meier, D.H. Olson, Atlas of zeolite structure types, 3rd ed., on behalf of

structure commission of the international zeolites association, 1992.[37] R. Weingarten, J.A. Tomsett, W.C. Conner, G.W. Huber, J. Catal. 279 (2011) 174–

182.[38] A. Chica, K.G. Strohmaier, E. Iglesia, Appl. Catal. B 60 (2005) 223–232.[39] G.H. Kuehl, H.K. Timken, Microporous Mesoporous Mater. 35–36 (2000) 521–

532.[40] A. Lucas, P. Canizares, A. Duran, A. Carrero, Appl. Catal. A 154 (1997) 221–240.[41] N.E. Poh, H. Nur, M. Muhid, H. Hamdan, Catal. Today 114 (2006) 257–262.[42] T.S. Hansen, J. Meilby, A. Riisager, Green Chem. 13 (2011) 109–114.[43] A. Torres, P. Dautidis, M. Tsapatsis, Environ. Sci. 3 (2010) 1560–1572.[44] C. Aellig, I. Hermans, ChemSusChem 5 (2012) 1737–1742.