Efficient dehydration of fructose into 5-hydroxymethylfurfural in aqueous medium over silica-included heteropolyacids

Accepted Manuscript Efficient dehydration of fructose into 5-hydroxymethylfurfural in aqueous medium over silica-included heteropolyacids Guangqiang Lv, Liangliang Deng, Boqiong Lu, Jinlong Li, Xianglin Hou, Yongxing Yang PII:

S0959-6526(16)31891-1

DOI:

10.1016/j.jclepro.2016.11.053

Reference:

JCLP 8442

To appear in:

Journal of Cleaner Production

Received Date: 23 August 2016 Revised Date:

6 November 2016

Accepted Date: 8 November 2016

Please cite this article as: Lv G, Deng L, Lu B, Li J, Hou X, Yang Y, Efficient dehydration of fructose into 5-hydroxymethylfurfural in aqueous medium over silica-included heteropolyacids, Journal of Cleaner Production (2016), doi: 10.1016/j.jclepro.2016.11.053. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

RI PT

Word Count: 6251

Efficient dehydration of fructose into 5-hydroxymethylfurfural in aqueous

SC

medium over silica-included heteropolyacids

Guangqiang Lva,b Liangliang Dengc, Boqiong Lua,b, Jinlong Lia,b, Xianglin Houa, Yongxing

a

M AN U

Yanga *

Shanxi Engineering Research Center of Biorefinery, Institute of Coal Chemistry, Chinese

Academy of Sciences, Taiyuan, 030001 People’s Republic of China

c

University of Chinese Academy of Sciences, Beijing, 100039 People’s Republic of China

TE D

b

Shanxi Provincial Research Institute of Communications, Taiyuan, Shanxi, 030006 People’s

Republic of China.

AC C

EP

Tel: (+86) 351-4049501, E-mail: [email protected]

ABSTRACT: The dehydration of fructose in aqueous and biphasic system was studied using heteropolyacids, immobilized in silica, as catalyst. The catalysts were characterized by ICP-AES, nitrogen sorption, NH3-TPD, TEM, FT-IR, XRD, and TGA. The effect of different parameters, such as reaction temperature, reaction time, catalyst loading and volume ratio of water with methyl isobutyl ketone (MIBK) was studied. Silica included tungstosilisic acid

1

ACCEPTED MANUSCRIPT

(HSiW) was found to function as an efficient catalyst for the transformation of fructose into 5hydroxymethylfurfural in aqueous phase.

RI PT

Addition of MIBK into the reaction system can increase both of fructose conversion and HMF selectivity. Besides, MIBK showed a positive effect in increasing the stability of the

catalyst by suppressing the formation and deposition of humins over catalyst surface. After the

SC

third catalysis cycles, a stabilization of the catalytic activity was observed.

KEYWORDS: silica-included HPAs, HMF, fructose dehydration, aqueous phase,

M AN U

heterogeneous catalysis 1. Introduction

With environmental concerns rising and fossil fuel resources diminishing, the production of transportation fuel and fine chemical from biomass has attracted great political and technical

TE D

interests in recent years. 5-hydroxymethyfurfural (HMF) and furfural derived from cellulose, glucose, fructose, inulin and some other biomass has been envisaged as important platform chemicals in biorefinery processes(Yan et al., 2014). HMF, and furfural can be applied as

EP

platform chemicals or precursor to be further upgraded into value-added chemicals(Yan et al., 2014), polymers(Chheda et al., 2007a), and fuel(Balakrishnan et al., 2012; Chheda et al., 2007a;

AC C

Sutton et al., 2013).

Fructose, for its high activity and efficiency towards to HMF production, has been studied widely as a start material for the production of HMF. Kinds of heterogeneous and homogeneous catalytic system such as ionic liquids (Kotadia and Soni, 2013; Tong and Li, 2010), mineral acids (Chheda et al., 2007b), ion exchange resins (Li et al., 2013b), niobium phosphate (Zhang et al., 2012), functionalized heteropolyacids (Qu et al., 2012), graphene oxide (Wang et al., 2014),

2

ACCEPTED MANUSCRIPT

carbon-based solid acid (Wang et al., 2011) and zeolites (Ordomsky et al., 2012a, b) with water (Carlini et al., 2004; Carniti et al., 2006), high boiling point aprotic solvents (Tong and Li, 2010) or mediated organic solvent (Chheda et al., 2007b; Wang et al., 2014) as reaction medium have

RI PT

been applied for the fructose dehydration reaction. High yield and selectivity of HMF can be obtained in solvent like dimethyl sulfoxide (DMSO) (Kotadia and Soni, 2013; Tong and Li, 2010) and ionic liquids (Li et al., 2013b), but it is a high energy-consuming process for HMF

SC

isolation because of high boiling point of these solvents. Water is the most economically and environment friendly solvent. However, it was shown that water as a reaction medium promotes

M AN U

many side effects, such as oligomerization of fructose or HMF, transformation of unstable intermediate of HMF into polymers, humins, and rapid rehydration of HMF into levilininc acid, formic acid. Recently, modified aqueous solution is used as the reaction medium for HMF production in order to stabilize the produced HMF in aqueous by extract it into organic phase

TE D

(Gurbuz et al., 2012). This method allows the HMF separated easily from reaction medium and the reactive aqueous phase, containing spent homogeneous or heterogeneous catalysts, can be reused repeatedly. Vitaly.V. Ordomsky et al.(Ordomsky et al., 2012a) have studied fructose

EP

dehydration reaction in aqueous and biphasic system. A yield of HMF 33% was obtained in aqueous phase over zeolite MOR, the addition of methyl isobutyl ketone (MIBK) to the system

AC C

increased HMF selectivity significantly. In production of HMF with very low reactivity substrate like cellulose, a high HMF yield of 53% was obtained in direct degradation of cellulose in waterTHF biphasic system (Shi et al., 2013). Aqueous-organic biphasic system shows potential in high yield HMF production in lab and industry. Heterogeneous catalysts offer several advantages over homogeneous catalysts such as easy separation of product, reusability of catalyst and no corrosion of equipment, which make them

3

ACCEPTED MANUSCRIPT

more suitable for an industrial application. According to literatures (Moreau et al., 1996; Ordomsky et al., 2012b), the catalyst acid type, strength and structural properties, such as the surface area, pore size, microporous and mesoporous volume ratio have great influence on the

RI PT

catalytic performance in fructose conversion. Brønsted acid led to high HMF selectivitys, while Lewis acid often resulted in the intensive production of humins (Ordomsky et al., 2012b).

Heteropolyacids (HPAs) with Keggin type are typical strong Bronsted acids, which catalyze

SC

a wide variety of reactions (Caetano et al., 2008; Kozhevnikov, 1998a). The major disadvantages

M AN U

of HPAs as catalysts are their low thermal stability, low surface area (1–10 m2/g) (Kozhevnikov, 1998b). Even though heteropolyacids have been widely used, however, it dissolves in most organic solvent and it is difficult to recycle, thus it is more prefer to confine it into the mesoporous structure as previously reported(Yan et al., 2013). A variety of supports like zeolite (Tran et al., 2005), active carbon (Obalı and Doğu, 2008), silica (Caetano et al., 2008) and

TE D

polymers (Castanheiro et al., 2005) have been used as support to immobilize HPAs, in order to increase the specific area and the accessible acid sites number of HPAs in catalysis. Compared with supported HPAs, heteropolyacids included in silica matrix have attracted much interest as a

EP

less leaking solid acid catalyst, especially suitable for reaction in polar media (Árpád Molnár ∗,

AC C

1999; Ferreira et al., 2010; Izumi et al., 1999). Here in, silica-included heteropolyaicds were applied in fructose dehydration reaction in aqueous and biphasic system. The influence of acid strength, reaction temperature, reaction time and ratio of water with MIBK in reaction medium on the transform of fructose into HMF was investigated and discussed.

2. Materials and methods

2.1. Materials

4

ACCEPTED MANUSCRIPT

Fructose (Aladdin Chemistry Co., Ltd. Shanghai, China), 5-hydroxymethylfrufural (98%, DEMO Medical Tech Co., Ltd. China), Acetonitrile (Kermel Chem. Reagent Co., Ltd. China). H3PW12O40, H4SiW12O40 and tetraethyl orthosilicate (TEOS) were purchased from Amresco, J&K Scientific Company. All the regents were

RI PT

used as received without further purification.

2.2. Catalyst preparation

Silica-included heteropolyacids (HPAs, denoted as HPAs/SiO2 below) with 11 wt% nominal

SC

HPAs loading (silica to HPAs weight ratio 8) were prepared through the hydrolysis the tetraethyl orthosilicate (TEOS). The procedure described by Izumi et al. was applied with slight

M AN U

modifications (Árpád Molnár ∗, 1999; Ferreira et al., 2010; Izumi et al., 1999). Typically, HPAs (phosphotungstic acid, HPW; tungstosilicic acid, HSiW) were dissolved in 20 mL of water, 0.23 mol tetraethyl orthosilicate and 0.62 mol ethanol were added into the mixtures under magnetic stirring at 40 oC for 1 h in an oil bath. After that, the mixture was heated to 85 oC and maintained

TE D

at this temperature for another 4 h until the TEOS in mixture was hydrolyzed completely. The HPAs was immobilized into the in situ produced silica matrix. However, it takes more time to complete the hydrolysis of HSiW (85 oC, 10 h), because of the weaker acidity of HSiW than

EP

HPW. The resulting hydrogel was dehydrated at 45 oC in a vacuum drying oven (25 Torr) for 1 h. Subsequently, the temperature in vacuum oven was set at 150 oC and maintained for another 5 h.

AC C

After that, the dehydrated material was cooled to room temperature slowly, ground into 60-100 mesh, and extracted with hot water at 100 oC for 10 h in Soxhlet extractor. Finally, the obtained material was calcined again at 150 oC for another 3 h in vacuum drying oven (25 Torr) prior to use in catalytic reaction. 2.3 Catalytic reaction In typical procedures, a certain amount of investigated catalyst with 1 mmol fructose, 5 mL

5

ACCEPTED MANUSCRIPT

solvent, (pure water or different volume ratio of water and methyl isobutyl ketone (MIBK), were charged in a 10 mL stainless steel vessel with a Teflon lining and sealed by a screw cap. The reaction mixture was vigorous stirred during reaction with a magnetic stirrer inside the vessel.

RI PT

The schematic diagram of experimental set-up is provided in supporting information Fig. S1. A thermocouple dipped into the thermostatic oil with an electro-thermal pot was applied as the heating sources. Experiments were conducted at 140, 150, 160, 170 and 180 oC to investigate the

SC

effect of reaction temperature in this work. After the temperature of the oil bath rose to the

selected reaction temperature, the reaction vessels were then immersed into the oil bath and

M AN U

stirred magnetically at 300 rpm in the vessel. Zero time was taken as soon as the reaction vessels were put into the oil bath. A blank control experiment indicates that it takes ca. 15-20 min for the inner mixture in the vessel to reach the investigated reaction temperature (oil temperature in the pot). When the scheduled time reached, the reactor was quickly removed from the oil bath and

the vessels dip in the oil.

TE D

placed into ice water to quench the reaction. The reaction time means the period during which

For catalyst recycle tests, the spent catalyst was separated from reaction mixture by

EP

centrifugation, washed by water (4×10 mL) and MIBK (4×10 mL) to remove the adsorbed

AC C

fructose, reaction products like HMF, polymers and humins. After that, the catalyst was dried at 80 oC in oven and used for the next cycle catalytic reaction. 2.4 Analytical method

After the reaction, the mixture was cooled and diluted with 95% ethyl alcohol and filtered with a 0.45 µm syringe filter prior to analysis. As to the water-MBIK two-phase reaction mixture, the residual fructose and produced HMF amount in the two phases was analyzed separately, and the amount of fructose and HMF in two phases were analyzed and calculated together to obtain the

6

ACCEPTED MANUSCRIPT

final fructose conversion and HMF yield. All of the data were based on repeated runs. For the analysis of fructose conversion, a 4.6 mm ×

RI PT

250 mm Shodex sugar column (SC1011) and a refractive index detector (RID-10A) was used, and distilled water was used as the mobile phase at a flow rate of 1.0 mL min-1. The column temperature was maintained at 75 oC. For the analysis of HMF yield in reaction, a reversedphased C18 column (200 × 4.6 mm) at 25 oC with a UV detector at wavelength of 280 nm was

. The fructose conversion and HMF yield were calculated on the basis of external standard

M AN U

1

SC

used. The mobile phase was acetonitrile and distilled water (30:70 v/v) at a flow rate 0.6 mL min-

curves constructed with authentic standards. The produced HMF was determined by NMR method (NMR results are provided in supporting information, Fig. S2-Fig. S3).

TE D

3. Results and discussion

3.1 Physical-chemistry properties of the prepared HPAs/SiO2 Table 1 Results of characterization of silica included HPAs Sc

Calculated a

Found b

(m2/g) (cm3/g) (cm3/g) (Å)

(µmol/g)

(oC)

11.1

8.1

449.2

0.156

0.03

16.4

957.9

569.1

7.5

423.6

0.147

0.03

16.4

1050.8

539.6

AC C

HPW/SiO2

HPAs

HPAs

EP

Catalysts

HSiW/SiO2 11.1

Vmicc

Vmesc

Daveragec

TPD(NH3) Tmaxd

a

Calculated from the amounts of reagents charged in the preparation (wt%).

b

Determined by ICP analysis after extraction with hot water (wt%).

c

Surfaces area, pore volume and average pore size of catalysts was measured by BET method.

d

Maximum temperature of desorption of chemisorbed ammonia was determined by NH3-TPD method.

7

ACCEPTED MANUSCRIPT

The HPAs/SiO2 properties are shown in Table 1. As reported in literature (Árpád Molnár ∗, 1999), leakage of HPAs from HPAs/SiO2 by hot water extraction was considerable when the silica-to-acid weight ratio was less than 5. When the ratio was increased to 8, the leakage was

RI PT

substantially reduced. Therefore, we decide to prepare silica-included HPAs with a silica-to-acid weight ratio of 8 (HPA loading 11.1 wt%). Results in Table S1 show catalysts still experience a substantial loss of HPAs with hot water extraction. There are 8.1 wt% and 7.5 wt%

SC

phosphotungstic acid (HPW) and tungstosilicic acid (HSiW) retained in the catalysts by

determining the weight fraction of W by ICP. Larger silica-to-acid ratio, such as over 10, can

M AN U

reduced the leakage of the included HPAs in silica gel more efficiency, however, the HPAs weight percentage in the prepared material after thermal treatment in hot water also lower than that when the silica-to-acid ratio was 8. For instance, when the silica-to-HPW ratio was 10 in the catalyst preparation, ICP analysis indicated that 7.6 wt.% HPW was retained in the final obtained

TE D

HPW/SiO2 after hot water extraction. Considering these result, all the used HPAs/SiO2 below in this work was prepared with the initial calculated silica-to-acid at 8. Brunner−Emmet−Teller (BET) measurements BET analysis shows that the introduction of HPA

EP

into silica by the sol–gel method (HPW/SiO2 and HSiW/SiO2) creates a microporous material. The ratio of micropores volume with mesopores volume is about 5.2. Average pore diameter of

AC C

the two catalysts reaches 16.4 Å. The BET surfaces of samples are around 420-450 m2/g. The figure is larger than the results got by Arpad Molnar (250 m2/g) (Árpád Molnár ∗, 1999), but still lower than the figure (581 m2/g) found earlier by Izumi et al.(Izumi et al., 1999). The acidity of two catalysts was studied by temperature programmed desorption of ammonia (NH3-TPD, Fig. S4). Both catalysts exhibit the ammonia desorption at two temperatures: a lowtemperature peak at about 200 oC and a high-temperature peak at 530-570 oC, corresponding to

8

ACCEPTED MANUSCRIPT

the weak and strong acid sites of the samples, respectively. HPW/SiO2 has a stronger acidity (a desorption peak at temperature 569 oC) than HSiW/SiO2 (a desorption peak at temperature 539 o

C) (Table 1, Fig S4), though they have a similar weak acidity (desorption peaks at temperature

RI PT

around 200 oC). However, the results indicate HSiW/SiO2 has a medium acidity at temperature around 250-350 oC, which HPW/SiO2 does not. Modification of the SiO2 support with HSiW seems to induce new medium acidic sites around 250-350 oC as compared with SiO2 (SiO2 only

SC

appears a negligible ammonia desorption peak at temperature of approximately 320 oC, not

shown in Fig S4). The total amount of desorbed ammonia over HSiW/SiO2 is just slight higher

M AN U

than HPW/SiO2 (Table 1), maybe for the reason that HSiW has one more proton than HPW per Keggin uint though the loading of HPA in HSiW/SiO2 is slight lower than HPW/SiO2. Fieldemission transmission electron microscopy (TEM) technique was applied in the catalysts characterization. The included HPAs were found to be dispersed into silica matrix as fine and

TE D

uniform particles, as displayed in TEM images (supporting information Fig. S5). Other detailed characterization result, such as Fourier Transform Infrared Spectra (FT-IR), Powder X-ray diffraction (XRD), Thermogravimetric analyses (TGA), are provided in supporting information

EP

(Fig. S6 – Fig. S8).

AC C

3.2 Influence of reaction parameters on fructose dehydration 3.2.1 Effect of reaction temperature on fructose dehydration reaction Ana S. Dias has reported that mesoporous silica-supported HPW exhibit higher activity for xylose dehydration than non-support HPW (Dias et al., 2005, 2006). Much better result obtained with HPW and HSiW than with phosphomolybdic acid (HPMo) in fructose dehydration reaction (Dias et al., 2005, 2006). Herein, fructose dehydration under different reaction temperatures over silica-included HPW and HSiW was investigated and the results were shown in Table 2.

9

ACCEPTED MANUSCRIPT

Table 2. Fructose dehydration to HMF carried out in aqueous phase over silica included HPAs under different temperature a T (oC)

CFructoseb (%)

YHMFc (%)

SHMFd (%) 78.3

HSiW/SiO2

140

6.0

4.7

2

HSiW/SiO2

150

18.1

13.3

3

HSiW/SiO2

160

33.3

25.4

4

HSiW/SiO2

170

77.7

50.2

5

HSiW/SiO2

180

96.3

6

HPW/SiO2

140

29.2

7

HPW/SiO2

150

8

HPW/SiO2

160

9

HPW/SiO2

170

10

HPW/SiO2

180

SC

1

73.5 76.3

64.6

35.4

9.6

32.8

M AN U

34.1.

47.6

15.6

32.7

69.8

20.8

29.8

95.6

27.2

28.4

100

15.3

15.3

Reaction conditions: fructose, 1 mmol; deionized water 5 mL; reaction time, 160 min; HSiW/SiO2 or

HPW/SiO2, 18 mg. Fructose conversion.

c

HMF yield.

d

HMF selectivity.

AC C

EP

b

TE D

a

Catal.

RI PT

Entry

Scheme 1. Fructose dehydration reaction scheme in aqueous and biphasic system.

10

ACCEPTED MANUSCRIPT

Temperature plays an important role in the dehydration. In the presence of HPAs/SiO2, fructose conversion increased with the reaction temperature increasing over both HPW/SiO2 and HSiW/SiO2, however, HMF selectivity tends to be better at lower temperatures (Table 2, entry 1-

RI PT

5 and 6-10). Higher temperature promotes fructose dehydration, meanwhile, also fructose

polymerization (Scheme 1, reaction 1, 2) and HMF polymerization (Scheme 1, reaction 3)

(Ordomsky et al., 2012a). As shown in Scheme 1, the side-reactions of fructose, intermediates,

SC

and produced HMF into humins will lead a decreased HMF yield, on other side, the formation and deposition of humins will cover the surface of the catalyst and resulted in deactivation, as

M AN U

discussed in literature (Yan et al., 2015). Higher fructose conversion can be achieved over HPW/SiO2 than HSiW/SiO2 under all investigated reaction temperatures, because of the stronger acid strength of HPW/SiO2 than HSiW/SiO2. But only at lower reaction temperatures (Table 2, entry 1, 2, 7, 8), higher HMF yield over HPW/SiO2 can be obtained than HSiW/SiO2. Above 160 o

TE D

C, HPW/SiO2 resulted in a lower HMF selectivity and yield than HSiW/SiO2 (entry 3, 8; entry 4,

9; entry 5, 10). The highest HMF yield of 50.5% was obtained over HSiW/SiO2 at reaction of 170 oC, with the fructose conversion of 77.7% achieved in aqueous solution (entry 4).

EP

3.2.2 Effect of acid strength and reaction time on fructose dehydration The effect of reaction time on fructose dehydration was investigated at a fixed reaction

AC C

temperature 170 oC, as determined in previous. Fructose conversion and HMF yield over HPAs/SiO2 in different reaction time are shown in Fig 1. For comparison, the blank experiment without any catalyst was also provided in Fig. 1.

11

ACCEPTED MANUSCRIPT

90

No catalyst HSiW/SiO2

80

HPW/SiO2

a

70 60

RI PT

50 40 30 20 10 0 40

60

80 100 120 140 160 180 200 220 240

M AN U

No catalyst HSiW/SiO2

b

HPW/SiO2

TE D

HMF Yield (%)

Time (min)

55 50 45 40 35 30 25 20 15 10 5 0

SC

Fructose Conversion (%)

100

40

60

EP

20

80 100 120 140 160 180 200 220 Time (min)

AC C

Fig. 1. Fructose conversion (a) and HMF yield (b) versus reaction time over HSiW/SiO2, HPW/SiO2 and blank test in aqueous phase. Reaction condition: fructose, 1 mmol; deionized water, 5 mL; reaction temperature, 170 oC; catalyst loading, 18 mg. (The spline approximation method was applied in the curve plot.)

12

ACCEPTED MANUSCRIPT

48

HSiW/SiO2 with aqueous phase

45

HPW/SiO2 with MIBK/water=3 HPW/SiO2 with aqueous phase

42

RI PT

39 36 33 30 27 20

40

60

SC

HMF Conversion (%)

HSiW/SiO2 with MIBK/water=3

80 100 120 140 160 180 200 220 240

M AN U

Time (min)

Fig. 2. HMF conversion over HSiW/SiO2 and HPW/SiO2 in aqueous phase and with addition of MIBK with MIBK/water=3/1 vol. Reaction condition: HMF, 1 mmol; solvent, 5 ml; HSiW/SO2 or HPW/SiO2, 18 mg; reaction temperature, 170 oC. (The spline approximation method was applied in the curve plot.)

TE D

In Fig. 1, without any catalyst in aqueous phase at 170 oC, fructose conversion also reached almost 60% in 160 min. Over HPW/SiO2, fructose conversion always higher than HSiW/SiO2 during all the investigated reaction time for its stronger acid strength. This means acid strength

EP

has a facilitating effect on the transformation of fructose. Almost full conversion of fructose over HPW/SiO2 and 80% over HSiW/SiO2 can be reached in 3 h. Antal et al.(Antal et al., 1990)

AC C

reported that the fructose dehydration goes through cyclic intermediates, which might also be very reactive and take part in the transformation into humins as well. It is deduced that the parallel transformation of fructose into the oligomeric by-products in the presence of strong acid is responsible for the lower HMF yield in HPW/SiO2 catalytic system. Besides, we speculated that HMF is more variable in strong acid condition. To verify this hypothesis, HMF stability in the selected reaction condition was investigated. In the presences of HPW/SiO2 or HSiW/SiO2, 1 mmol HMF dissolved into 5 mL solvent was heated to 170 oC, the residual HMF concentration

13

ACCEPTED MANUSCRIPT

in different time was analyzed by HPLC method and the calculated HMF conversion was shown in Fig. 2. Results show that HMF is more stable under HSiW/SiO2 catalytic system than HPW/SiO2, in accordance with our speculation. For this reason, HMF yield show a sharp

RI PT

decrease in HPW/SiO2 catalytic system (Fig. 1, b). Compared with HPW/SiO2, over HSiW/SiO2, the precursors of HMF transforms into HMF are in priority than into oligomeric by-products and HMF is more stable. Thus, a higher HMF yield up to 50.2% was achieved in 160 min in

SC

HSiW/SiO2 catalytic system. Further experiments show that the addition of methyl isobutyl ketone (MIBK) into the reaction system can increase the HMF stability, this results provide us a

M AN U

method to achieve a high HMF production from fructose dehydration. Detailed discussion was provided below.

3.2.3 Effect of catalyst loading on fructose dehydration reaction

Data in Fig. 1 and Fig. 2 implied that addition of catalyst can promote fructose conversion

TE D

and HMF yield obviously, especially the prepared HSiW/SiO2. In order to study the effect of catalyst loading on the fructose conversion and HMF yield, experiments with different catalyst loading were conducted. The reaction temperature and time were kept constant. Fructose conversion HMF yieid

90 80

a

70

AC C

Fructose Conversion/ HMF Yield (%)

EP

100

60 50 40 30 20 10 0

0

6

12

18

24

30

36

HSiW/SiO2 (mg)

14

110 100 90 80 70 60 50 40 30 20 10 0

Frutose conversion HMF yield

0

6

SC

RI PT

b

24 12 18 HPW/SiO2 (mg)

30

36

M AN U

Fructose Conversion/HMF Yield (%)

ACCEPTED MANUSCRIPT

Fig. 3. Effect of catalyst loading (HSiW/SiO2 (a), HPW/SiO2 (b)) on fructose conversion and HMF yield in aqueous phase. Reaction condition: fructose, 1 mmol; deionized water, 5 mL; reaction temperature, 170 oC; reaction time, 160 min.

TE D

Fig. 3 shows the effect of catalyst loading on fructose conversion and HMF yield. Fructose conversion increases when the catalyst loading increases from 0 mg to 36 mg under investigated reaction condition. With same loading of HPAs/SiO2, fructose conversion always showed a high

EP

value over HPW/SiO2 than HSiW/SiO2. This clearly states that increasing the acid strength or amount of the catalyst can increase the activity sites in reaction, which boost fructose

AC C

dehydration reaction. Only 21.4% HMF yield was achieved in the absence of catalyst at 170 oC, 160 min. Nevertheless, the HMF yield reached to 42.3% when 6 mg HSiW/SiO2 was added. When the loading of catalyst increased to 18 mg, the HMF yield reached the maximum of 50.5%. However, with the continued increasing of catalyst, the HMF yield started to decrease, due to the HMF degradation reaction. Over HPW/SiO2, the HMF yield got the highest value of 33.6% while the catalyst loading was 6 mg. More catalyst resulted in the continued decreasing of HMF yield, due to the polymerization of fructose, intermediate and HMF degradation under strong

15

ACCEPTED MANUSCRIPT

acid condition. Above results demonstrates that for different acid strength HPAs/SiO2, the optimum catalyst usages are 18 mg and 6 mg for HSiW/SiO2 and HPW/SiO2 respectively to obtain the highest HMF yield.

RI PT

3.2.4. Effect of MIBK/water volume ratio on fructose dehydration over HPAs/SiO2

Fig. 2 shows the suppression of HMF degradation reaction over HPAs/SiO2 in the presence of MIBK. The HMF ratio between water and MIBK at reaction temperature ca. 160-170 oC is

SC

about 0.8 (Ordomsky et al., 2012a), it means at VMIBK/VH2O= 3, the amount of HMF presented in

M AN U

the organic phase is 2.7 times higher than the amount in water. This should lead to the decrease of HMF degradation rate in aqueous phase. To get a higher HMF yield, MIBK is introduced into the reaction medium. 100

a

90

60 50 40 30 20

TE D

70

EP

Fructose conversion (%)

80

10

Aqueous phase MIBK/water=1 MIBK/water=3 MIBK/water=5

0

AC C

20

40

60

80 100 120 140 160 180 200 220 240 Reaction time (min)

16

ACCEPTED MANUSCRIPT

80

b

70

50 40

RI PT

HMF Yield (%)

60

30

Aqueous phase MIBK/water=1 MIBK/water=3 MIBK/water=5

20 10 20

40

60

SC

0

80 100 120 140 160 180 200 220 240 Reaction time (min)

M AN U

Fig. 4. Influence of MIBK/water volume ratio on fructose conversion (a) and HMF yield (b) over HSiW/SiO2. Reaction condition: fructose, 1 mmol; HSiW/SiO2, 18 mg; solvent, 5 mL; reaction temperature, 170 oC.

The effects of MIBK addition on fructose conversion and HMF yield over HSiW/SiO2 is

TE D

shown in Fig. 4. Over HSiW/SiO2, in pure water, fructose conversion and HMF yield were relative low (77.7% and 50.5%, respectively) under optimized conditions. With the increasing of the amount of MIBK into biphasic system, both of the fructose conversion and HMF yield

EP

increased, reaching to 88.7% and 71.3% respectively when VMIBK/VH2O= 3. Increasing the volume ratio to 5:1 did not increase the fructose conversion obviously any more, but the HMF

AC C

yield still increased to 77.1%. HPLC analysis found that no fructose could be found in MIBK phase, indicating that fructose is concentrated in aqueous. The HSiW/SiO2 surface is hydrophilic, this property make the catalyst surrounded by concentrated fructose aqueous solution. This phenomenon makes the HSiW/SiO2 catalytic ability preserved and the concentrated fructose aqueous accelerates the fructose dehydration rate and gives a high fructose conversion under investigated reaction conditions. The extraction of produced HMF from aqueous solution into MIBK phase decreases the HMF degradation reaction, thus giving an increased HMF yield.

17

ACCEPTED MANUSCRIPT

Similar results were found in fructose dehydration over HPW/SiO2, fructose conversion reached 84.5% in pure water in 120 min. Addition of MIBK did not increase fructose conversion significantly any more (Fig. S9 a). However, the highest HMF yield increased from 30.1% (pure

RI PT

water) to 45.1% in VMIBK/VH2O= 3 biphasic system (Fig. S9 b). Further increasing MIBK/water ratio (5:1) increases HMF yield to 48.8%. These results demonstrate that addition of MIBK into reaction system is effective in obtaining a higher HMF yield.

SC

3.2.5 Effect of initial fructose concentration on fructose dehydration reaction.

M AN U

Previous work has indicated fructose dehydration reaction over HSiW/SiO2 can achieve higher HMF yield than HPW/SiO2 in aqueous and biphasic system. Thus, fructose dehydration reaction with different initial fructose concentration over HSiW/SiO2 was conducted in aqueous phase and biphasic system (VMIBK/VH2O= 3). Table 3 displays the experiments results. As shown in Table 3, in pure water, fructose conversion experiences an initial increase, and a succession

TE D

decrease with the increasing of the fructose concentration. The highest fructose conversion 82.1% was achieved with 2 mmol fructose dissolved in 5 mL water. In biphasic reaction system, highest fructose conversion was obtained in the lowest investigated fructose usage when MIBK was

EP

introduced into the reaction system. As discussed above, introduction of MIBK into the reaction system can increase the fructose concentration in aqueous around the catalyst, similar to increase

AC C

the fructose usage in pure water. As a consequence, in biphasic system, fructose dehydrate rate and final conversion was increased, compared with that in pure water in the same fructose usages. Table 3. Fructose dehydration reaction with different initial fructose concentration over HSiW/SiO2 in aqueous phase and biphasic system (VMIBK/VH2O= 3) a Usage of fructose

Pure water

MIBK/water=3

18

ACCEPTED MANUSCRIPT

YHMFc

SHMFd

CFructoseb

YHMFc

SHMFd

(%)

(%)

(%)

(%)

(%)

0.25

70.4

37.2

52.8

90.4

82.1

90.8

0.5

71.0

44.3

62.4

88.6

77.5

87.5

0.75

73.7

49.5

67.2

88.8

74.6

84.0

1.0

77.6

50.2

64.7

88.7

71.3

80.3

1.25

78.5

46.9

59.7

84.6

67.7

80.0

1.5

79.0

46.6

59.0

84.9

66.1

77.9

1.75

81.3

46.1

56.7

86.2

64.2

74.5

2.0

82.1

44.5

54.2

85.7

61.9

72.2

2.25

78.7

41.9

53.2

84.6

61.3

72.5

2.5

75.9

40.1

52.8

83.6

60.5

72.4

2.75

73.3

35.9

49.0

80.2

56.2

70.1

SC

M AN U

Reaction condition: solvent, 5 mL (1.25 mL water and 3.75 mL MIBK); reaction temperature, 170 o

TE D

a

RI PT

CFructoseb (%)

(mmol)

C; reaction time, 160 min; HSiW/SiO2, 18 mg

Fructose conversion

c

HMF yield

d

HMF selectivity

EP

b

AC C

In pure water, HMF selectivity experienced an initial increase with increasing the fructose concentration, the highest selectivity was achieved under 0.75 mmol fructose in 5 mL water. Low selectivity under low fructose concentration at this reaction condition can be attributed to the excess acid sites relative to the amount of fructose, which leads to oligomerization of HMF over acid sites on HSiW/SiO2. Extracting of HMF into organic phase by addition of MIBK had increased HMF selectivity to 90.8% (VMIBK/VH2O= 3) from 52.8% while 0.25 mmol fructose dissolved in 5 mL water.

19

ACCEPTED MANUSCRIPT

Overall, the HSiW/SiO2 heterogeneous catalyst is efficient for the transformation of fructose in aqueous phase. In transformation of ~9 wt% fructose solutions (2.75 mmol fructose in 5 ml water), the fructose conversion and HMF selectivity retained 73.3% and 49.0%. At

RI PT

VMIBK/VH2O= 3, the fructose conversion and HMF selectivity increased to 80.2% and 70.1% respectively.

8

7.49 6.02

6

5.52

5.34

5.27

5.21

4

5

M AN U

HSiW content (wt.%)

7

SC

3.3. Stability and reusability of HSiW/SiO2

5 4 3 2 1 0

1

2

TE D

0

3 Cycles

Fig. 5. Content of HSiW of each cycled HSiW/SiO2 determined by ICP. (0 stands for fresh

AC C

EP

HSiW/SiO2. 1~5 stand for the times processed in water under 170 oC and 160 min.)

20

ACCEPTED MANUSCRIPT

a

Fructose Conversion HMF Yield

70 60

RI PT

50 40 30 20 10 0 1

2

3

70 60 50 40

20 10 0

5

M AN U

80

30

4

Fructose Conversion HMF Yield

b

90

TE D

Fructose Conversion/HMF Yield (%)

Cycles

SC

Fructose Conversion/HMF Yield (%)

80

EP

1

2

3 Cycles

4

5

Fig. 6. Fructose conversion and HMF yield in aqueous (a) and MIBK/water=3 biphasic system

AC C

(b) in 5 subsequent cycles. Reaction condition: fructose, 1 mmol; solvent, 5 mL (1.25 mL water o

and 3.75 mL MIBK); reaction temperature, 170 C; reaction time, 160 min.

Leaching of HSiW from HSiW/SiO2 was investigated by hydrothermal treated in water at 170 oC for 160 min, without fructose was added. Fig. 5 shows that HSiW loading in HSiW/SiO2 catalyst decrease by up to 20% after the first run, followed by only minor reduction in subsequent runs, the overall leaking of HSiW after 5 runs is about 30%. The reactivity of these

21

ACCEPTED MANUSCRIPT

leached species in solution was checked by carrying out experiments in which the fresh HSiW/SiO2 catalyst was stirred in water, without fructose, for 160 min at 170 oC. Afterwards, the solids were removed by centrifugation, and fructose was added to the solutions, which was

RI PT

heated again to 170 oC and maintained for 160 min. The fructose conversion was just slightly higher (~60%) than that in pure water (Fig .1 a) without any catalyst, indicating the leached species are not much active in homogeneous phase might because of the low concentration of

SC

HSiW (HSiW content in solution before the addition of fructose was 3.5 × 10-7 g g-1,

determined by ICP method). In catalyst recycle tests, a HMF yield of 45~48% was achieved

in 5 mL water, under 170 oC, 160 min.

M AN U

when the hydrothermal treated HSiW/SiO2 was used in catalytic reaction with 1 mmol fructose

Deposition of insoluble humins on the catalyst surface is an important factor for deactivation of heterogeneous catalysts. Addition of MIBK into reaction medium, besides

TE D

increasing the fructose conversion and HMF selectivity, should have an effect in increasing the stability of the catalyst by suppressing humins formation and deposition on catalyst surface. In Fig. 6, HSiW/SiO2 recycles tests in aqueous phase shows that fructose conversion and HMF

EP

yield decrease clearly after first runs, this might be explained by the clearly leaking of HSiW

AC C

from HSiW/SiO2 and catalyst deactivation by humins deposition on the catalyst surface. Introduction of MIBK into the reaction system suppressed the decrease of fructose conversion and HMF yield after the first runs (Fig. 6 b). After the catalyst stabilized under hydrothermal treatment condition during reaction for 3-5 cycles, HMF yield can be remained above 40% and 60% in aqueous phase and biphasic system (VMIBK/VH2O= 3), respectively. Comparison of the HMF yields over HSiW/SiO2 with other heterogeneous system shows that HSiW/SiO2 is an effective catalyst in fructose dehydration reaction. A HMF yield up to 50.5%

22

ACCEPTED MANUSCRIPT

was achieved over HSiW/SiO2 in aqueous phase. Similar result, a HMF yield in the range 2044%, was obtained over other heterogeneous catalysts like zeolite MOR(Ordomsky et al., 2012a), vanadyl phosphate(Carlini et al., 2004) and niobium phosphates (Carniti et al., 2006). Addition

RI PT

of MIBK into the reaction system make the HMF yield increase from 50.5% in aqueous phase to 70.3% (VMIBK/VH2O= 3) and 77.1% (VMIBK/VH2O= 5) in biphasic system. This value is still lower than those reported in ionic liquid or high boiling organic solvents participated reaction system

SC

(73% ~ 99%)(Li et al., 2013a; Li et al., 2013b). However, it provides an opportunity to work with an easily prepared and efficient heterogeneous catalyst, and the utilization of water-MIBK

M AN U

biphasic system avoids the energy-demanding isolation procedures in HMF separation.

4. Conclusions

Silica included HSiW prepared by sel-gel method has been shown to be a highly reactive,

TE D

selective and reusable acid catalyst for fructose dehydration in aqueous and biphasic system. Characterization results show that the prepared silica-included HSiW is a kind of microporous material with Bronsted acid site, possessing large surface area, suitable pore size, appropriate

EP

acid strength and high acid density. All this factors favor its high catalytic performance in

AC C

fructose dehydration reaction.

The hydrophilic properties of the HPAs/SiO2 make the catalyst surface was surrounded by fructose aqueous solution during reaction, the addition of MIBK into the reaction system shows a positive effect in increasing fructose dehydrate rate, final fructose conversion. Introduction of MIBK into reaction system also can extract HMF into the MIBK pahse, and thus decreases HMF degradation reaction, improves HMF selectivity in reaction, suppress the formation and deposition of humins on catalyst surface. Experiments demonstrate HSiW/SiO2 and water-MIBK

23

ACCEPTED MANUSCRIPT

reaction system is effective to get a high fructose conversion and HMF yield under optimized conditions.

RI PT

Acknowledgments This work was financially supported by the National Natural Science Foundation of China (NNSFC) (Grant 21403273), the Shanxi Scholarship Council of China (2015-122), the

SC

Department of Human Resource and Social Security of Shanxi Province (Y6SW9613B1) and the Department of Science and Technology of Shanxi Province (201605D211006).

M AN U

Supporting information: Characterization method and results of the catalyst, NH3-TPD, TEM, FT-IR, TGA, XRD and additional reaction studies, NMR of produced HMF are provided. References

Antal, M.J., Jr., Mok, W.S.L., 1990. Mechanism of formation of 5-(hydroxymethyl)-2-furaldehyde from D-fructose and sucrose. Carbohydr. Res. 199, 91-109.

TE D

Árpád Molnár ∗, C.K., Béla Török, 1999. Heteropoly acids immobilized into a silica matrix: characterization and catalytic applications. Appl. Catal., A 189, 217-224.

EP

Balakrishnan, M., Sacia, E.R., Bell, A.T., 2012. Etherification and reductive etherification of 5(hydroxymethyl)furfural: 5-(alkoxymethyl)furfurals and 2,5-bis(alkoxymethyl)furans as potential bio-diesel candidates. Green Chem. 14, 1626-1634.

AC C

Caetano, C.S., Fonseca, I.M., Ramos, A.M., Vital, J., Castanheiro, J.E., 2008. Esterification of free fatty acids with methanol using heteropolyacids immobilized on silica. Catal. Commun. 9, 1996-1999. Carlini, C., Patrono, P., Galletti, A.M.R., Sbrana, G., 2004. Heterogeneous catalysts based on vanadyl phosphate for fructose dehydration to 5-hydroxymethyl-2-furaldehyde. Appl. Catal. A 275, 111-118. Carniti, P., Gervasini, A., Biella, S., Auroux, A., 2006. Niobic acid and niobium phosphate as highly acidic viable catalysts in aqueous medium: Fructose dehydration reaction. Catal. Today 118, 373-378. Castanheiro, J.E., Fonseca, I.M., Ramos, A.M., Oliveira, R., Vital, J., 2005. Hydration of αpinene over molybdophosphoric acid immobilized in hydrophobically modified PVA membranes. Catal. Today 104, 296-304.

24

ACCEPTED MANUSCRIPT

Chheda, J.N., Huber, G.W., Dumesic, J.A., 2007a. Liquid-phase catalytic processing of biomassderived oxygenated hydrocarbons to fuels and chemicals. Angew. Chem. Int. Ed. 46, 7164-7183.

RI PT

Chheda, J.N., Román-Leshkov, Y., Dumesic, J.A., 2007b. Production of 5hydroxymethylfurfural and furfural by dehydration of biomass-derived mono- and polysaccharides. Green Chem. 9, 342-350. Dias, A.S., Pillinger, M., Valente, A.A., 2005. Liquid phase dehydration of d-xylose in the presence of Keggin-type heteropolyacids. Appl. Catal., A 285, 126-131.

SC

Dias, A.S., Pillinger, M., Valente, A.A., 2006. Mesoporous silica-supported 12tungstophosphoric acid catalysts for the liquid phase dehydration of d-xylose. Microporous Mesoporous Mater. 94, 214-225.

M AN U

Ferreira, P., Fonseca, I.M., Ramos, A.M., Vital, J., Castanheiro, J.E., 2010. Valorisation of glycerol by condensation with acetone over silica-included heteropolyacids. Appl. Catal., B 98, 94-99. Gurbuz, E.I., Wettstein, S.G., Dumesic, J.A., 2012. Conversion of hemicellulose to furfural and levulinic acid using biphasic reactors with alkylphenol solvents. ChemSusChem 5, 383-387. Izumi, Y., Hisano, K., Hida, T., 1999. Acid catalysis of silica included heteropolyacids in polar reaction media. Appl. Catal., A 181, 277-282.

TE D

Kotadia, D.A., Soni, S.S., 2013. Symmetrical and unsymmetrical Brønsted acidic ionic liquids for the effective conversion of fructose to 5-hydroxymethyl furfural. Catal. Sci. Technol. 3, 469474. and

Multicomponent

Kozhevnikov, I.V., 1998b. Catalysis by Heteropoly Acids Polyoxometalates in Liquid-Phase Reactions. Chem. Rev. 98, 171-198.

and

Multicomponent

EP

Kozhevnikov, I.V., 1998a. Catalysis by Heteropoly Acids Polyoxometalates in Liquid-Phase Reactions. Chem. Rev. 98, 171-198.

AC C

Li, H., Zhang, Q., Liu, X., Chang, F., Zhang, Y., Xue, W., Yang, S., 2013a. Immobilizing Cr3+ with SO3H-functionalized solid polymeric ionic liquids as efficient and reusable catalysts for selective transformation of carbohydrates into 5-hydroxymethylfurfural. Bioresour. Technol. 144, 21-27. Li, Y., Liu, H., Song, C., Gu, X., Li, H., Zhu, W., Yin, S., Han, C., 2013b. The dehydration of fructose to 5-hydroxymethylfurfural efficiently catalyzed by acidic ion-exchange resin in ionic liquid. Bioresour. Technol. 133, 347-353. Moreau, C., Durand, R., Razigade, S., Duhamet, J., Faugeras, P., Rivalier, P., Ros, P., Avignon, G., 1996. Dehydration of fructose to 5-hydroxymethylfurfural over H-mordenites. Appl. Catal. A 145, 211-224. Obalı, Z., Doğu, T., 2008. Activated carbon–tungstophosphoric acid catalysts for the synthesis of tert-amyl ethyl ether (TAEE). Chem. Eng. J. 138, 548-555.

25

ACCEPTED MANUSCRIPT

Ordomsky, V.V., van der Schaaf, J., Schouten, J.C., Nijhuis, T.A., 2012a. The effect of solvent addition on fructose dehydration to 5-hydroxymethylfurfural in biphasic system over zeolites. J. Catal. 287, 68-75.

RI PT

Ordomsky, V.V., van der Schaaf, J., Schouten, J.C., Nijhuis, T.A., 2012b. Fructose dehydration to 5-hydroxymethylfurfural over solid acid catalysts in a biphasic system. ChemSusChem 5, 1812-1819. Qu, Y., Huang, C., Zhang, J., Chen, B., 2012. Efficient dehydration of fructose to 5hydroxymethylfurfural catalyzed by a recyclable sulfonated organic heteropolyacid salt. Bioresour. Technol. 106, 170-172.

SC

Shi, N., Liu, Q., Zhang, Q., Wang, T., Ma, L., 2013. High yield production of 5hydroxymethylfurfural from cellulose by high concentration of sulfates in biphasic system. Green Chem. 15, 1967-1974.

M AN U

Sutton, A.D., Waldie, F.D., Wu, R., Schlaf, M., Silks, L.A., 3rd, Gordon, J.C., 2013. The hydrodeoxygenation of bioderived furans into alkanes. Nat. Chem. 5, 428-432. Tong, X., Li, Y., 2010. Efficient and selective dehydration of fructose to 5hydroxymethylfurfural catalyzed by Bronsted-acidic ionic liquids. ChemSusChem 3, 350-355. Tran, M.H., Ohkita, H., Mizushima, T., Kakuta, N., 2005. Hydrothermal synthesis of molybdenum oxide catalyst: Heteropoly acids encaged in US-Y. Appl. Catal. A 287, 129-134.

TE D

Wang, H., Kong, Q., Wang, Y., Deng, T., Chen, C., Hou, X., Zhu, Y., 2014. Graphene Oxide Catalyzed Dehydration of Fructose into 5-Hydroxymethylfurfural with Isopropanol as Cosolvent. ChemCatChem 6, 728-732. Wang, J., Xu, W., Ren, J., Liu, X., Lu, G., Wang, Y., 2011. Efficient catalytic conversion of fructose into hydroxymethylfurfural by a novel carbon-based solid acid. Green Chem. 13, 26782681.

EP

Yan, K., Wu, G., Lafleur, T., Jarvis, C., 2014. Production, properties and catalytic hydrogenation of furfural to fuel additives and value-added chemicals. Renewable and Sustainable Energy Reviews 38, 663-676.

AC C

Yan, K., Wu, G., Wen, J., Chen, A., 2013. One-step synthesis of mesoporous H4SiW12O40SiO2 catalysts for the production of methyl and ethyl levulinate biodiesel. Catalysis Communications 34, 58-63. Yan, K., Yang, Y., Chai, J., Lu, Y., 2015. Catalytic reactions of gamma-valerolactone: A platform to fuels and value-added chemicals. Applied Catalysis B: Environmental 179, 292-304. Zhang, Y., Wang, J., Ren, J., Liu, X., Li, X., Xia, Y., Lu, G., Wang, Y., 2012. Mesoporous niobium phosphate: an excellent solid acid for the dehydration of fructose to 5hydroxymethylfurfural in water. Catalysis Science & Technology 2, 2485-2491.

26