Highly efficient conversion of glucose into methyl levulinate catalyzed by tin-exchanged montmorillonite

Renewable Energy 120 (2018) 231e240

Contents lists available at ScienceDirect

Renewable Energy journal homepage: www.elsevier.com/locate/renene

Highly efficient conversion of glucose into methyl levulinate catalyzed by tin-exchanged montmorillonite Jie Liu, Xue-Qian Wang, Bei-Bei Yang, Chun-Ling Liu**, Chun-Li Xu, Wen-Sheng Dong* Key Laboratory of Applied Surface and Colloid Chemistry, MOE, School of Chemistry and Chemical Engineering, Shaanxi Normal University (SNNU), Xi'an 710062, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 7 November 2016 Received in revised form 25 December 2017 Accepted 28 December 2017 Available online 28 December 2017

Tin-exchanged montmorillonite catalysts were prepared by an ion exchange method and examined as solid acid catalysts. The synthesized catalysts were characterized by X-ray fluorescence spectroscopy, inductively coupled plasma optical emission spectroscopy, N2 adsorptionedesorption analysis, powder X-ray diffraction, Fourier transform infrared spectroscopy, X-ray photoelectron spectroscopy, NH3 temperature-programmed desorption analysis, and pyridine adsorption Fourier transform infrared spectroscopy. Subsequently, the catalysts were examined for use in the conversion of glucose into methyl levulinate in methanol. A high yield of methyl levulinate of 59.7% was obtained upon conversion of 0.3 g glucose in 24 g methanol over 0.15 g catalyst at 220
C under 2 MPa N2 for 6 h. The recyclability of the catalyst was also examined, and the conversions of glucose and methanol remained mostly unchanged under repeated usage of the catalyst in five catalytic runs; in contrast, the yield of methyl levulinate decreased slightly. The excellent catalytic performance of the tin-exchanged montmorillonite catalyst was attributed to a combination of the presence of a large amount of acidic sites and balanced amounts of Lewis and Brønsted acid sites on the catalyst. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Glucose Methyl levulinate Montmorillonite Tin

1. Introduction Biomass, a renewable carbon resource, is a promising alternative to diminishing fossil resources for the provision of chemicals and fuels [1,2]. However, currently, the development of economical, efficient, and environmentally friendly technologies to convert biomass into chemicals and fuels represents a huge challenge. Recently, the transformation of carbohydrate biomass into valueadded chemicals, such as 5-hydroxymethylfurfural [3,4], levulinic acid [5], and levulinate esters [6], has drawn enormous attention. Levulinate esters, such as methyl levulinate, ethyl levulinate, and butyl levulinate, are used as additives for the transportation fuels, as substrates for various condensation and addition reactions, and as feedstock in the flavoring and fragrance industries [7]. Levulinate esters can be produced from various materials, such as sugars (xylose, fructose, glucose) [8e24], cellulose [25e35], and even raw biomass materials [36]. Much effort has been devoted to

* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (W.-S. Dong).

(C.-L.

https://doi.org/10.1016/j.renene.2017.12.104 0960-1481/© 2017 Elsevier Ltd. All rights reserved.

Liu),

[email protected]

improve the production of levulinate esters from carbohydrate biomass. To date, various homogeneous and heterogeneous catalysts have been developed to convert biomass materials into levulinate esters. Mineral acids, metal salts, and sulfonic acidfunctionalized ionic liquids (SO3H-ILS) have been used as efficient homogeneous catalysts for biomass alcoholysis to methyl levulinate. For instance, Peng et al. [13] reported the use of sulfuric acid at low concentrations (0.01 mol L1) as a catalyst for the conversion of glucose into methyl levulinate at 200
C in methanol; a methyl levulinate yield of 50% was obtained. Saravanamurugan et al. [14] reported the use of SO3H-ILS in the catalytic transformation of sugars to ethyl levulinate. Specifically, the ILS displayed a high catalytic activity toward the conversion of fructose into ethyl levulinate, achieving a yield of 77%. However, the catalyst was not effective in the conversion of glucose because of the high stability of the glucose rings. Conversely, glucose was converted into ethyl-dglucopyranoside (63% yield) primarily, with a lower yield of ethyl levulinate. Zhou et al. [15] demonstrated the efficacy of Al2(SO4)3 as a catalyst in the alcoholysis of glucose and a methyl levulinate yield of 64% was obtained at 160
C. In another study, a mixed acids system containing both Brønsted and Lewis acids was examined and displayed superior performance in the alcoholysis of cellulose

232

J. Liu et al. / Renewable Energy 120 (2018) 231e240

and glucose. Tominaga et al. [27] used a triflate salt as a Lewis acid and an organic acid as a Brønsted acid to synthesize methyl levulinate from cellulose and glucose. Specifically, methyl levulinate (75% yield) was obtained from cellulose using In(OTf)3 and 2naphthalenesulfonic acid as the catalyst system at 180
C for 5 h. In contrast, a lower methyl levulinate yield of 58% was obtained from glucose using In(OTf)3 as the sole catalyst at 160
C. Although the above-exemplified homogeneous catalysts are effective for the conversion of carbohydrate biomass into levulinate esters, the catalysts are difficult to separate and recycle, thereby making the process potentially costly. Various solid catalysts have been used to produce levulinate esters from carbohydrates. Peng et al. [17] investigated ZSM-5, NaY, 2 H-mordenite, Zr3(PO4)4, SO2 4 /ZrO2, SO4 /TiO2, and TiO2 as solid acid catalysts in the conversion of glucose into methyl levulinate in near-critical methanol. Methyl levulinate was produced with a yield of 33% when the reaction was conducted at 200
C for 2 h using SO2 4 /TiO2 as the catalyst. Saravanamurugan et al. [19] employed H-USY as a catalyst to obtain methyl levulinate from glucose, and methyl levulinate esters were obtained with a yield of ~49% at 160
C for 20 h. In a subsequent study, Zhou et al. [22] investigated the catalytic performance of H-USY treated with nitric acid for use in the conversion of glucose in methanol; methyl levulinate was obtained with a yield of 54% at 180
C for 20 h. Xu et al. [23] reported the use of sulfated montmorillonite as an efficient catalyst in the conversion of glucose and achieved a methyl levulinate yield of 48% at 200
C for 4 h. Recently, Kuo et al. [24] reported that acidic TiO2 nanoparticles exhibited superior activity and selectivity towards methyl levulinate from glucose in methanol, methyl levulinate was obtained with a yield of 61% at 175
C for 9 h. However, the catalyst deactivated gradually during recycling. Despite the progress made to date, the development of an efficient and environmental benign catalyst that can afford high yields of levulinate esters and good stability is still imperative for costeffective conversion of biomass feedstocks. Of particular interest, montmorillonite (Mt) has a large surface area, significant cation-exchange ability, expansible interlayer space, a good adsorption capacity, and flexible and tunable acidity. Mt is composed of aluminosilicate layers, where one octahedral alumina sheet is sandwiched between two tetrahedral silica sheets. In the aluminosilicate layers, partial isomorphous substitution of Si(IV) ions by trivalent metal cations and Al(III) ions by divalent metal cations causes a charge deficit. To balance this charge deficit, a number of exchangeable hydrated alkali and alkaline earth metal cations occupy the interlayer space of montmorillonite [37].Various types of metal cations can be readily introduced in the expansible interlayers. Metal cations-exchanged Mt comprises both Lewis and Brønsted acid sites that enable Mt to operate as an effective and environmentally benign heterogeneous catalyst in acid-catalyzed organic transformations with excellent product selectivity [38e42]. On the other hand, it has been reported that the incorporation of tin to the frame of zeolites results in highly efficient catalysts for converting monosaccharides or different biomass sources into high-valued chemicals [43,44]. In our previous study, various metal ion-exchanged Mt catalysts were examined in the conversion of glucose to methyl levulinate. We found that Al-exchanged Mt with well-balanced amounts of Lewis and Brønsted acid sites showed excellent activity and stability in this reaction [45]. In the present study, we demonstrate the excellent catalytic activity of tin-exchanged Mt toward the conversion of glucose in methanolda methyl levulinate yield of ~60% was obtained. Moreover, the tin-exchanged Mt catalyst could be reused at least five times without significant losses in the product yield. Detailed process parameters including reaction time,

temperature, substrate amount, and the reusability of the catalyst were investigated in terms of catalytic performance. Moreover, different carbohydrate sources were studied. 2. Experimental 2.1. Chemicals SnCl4$5H2O (99%), D-glucose (99%), sucrose (99%), D-fructose (99%), starch (99%), cellobiose (99%), methyl formate (99.5%), and methanol (99.5%) were purchased from Sinopharm Chemical Reagent Co. Ltd. (China). Microcrystalline cellulose (degree of polymerization 250, crystalline index 70%, and average particle size 20 mm) was obtained from Sigma-Aldrich. Methylglucoside (99%) was purchased from Aladdin Reagent Co. Ltd.. Methyl levulinate (99%) and methyl lactate (99%) were obtained from Tokyo Chemical Industry Co. Ltd.. Na/Mt was purchased from Zhejiang San-Ding Technological Reagent Co. Ltd. (China). Inulin (99%) and K10 Mt were purchased from Alfa Aesar. 2.2. Preparation of the catalysts Ion exchange was used to prepare tin-exchanged Mt catalysts. In a typical synthesis, 1 g Na/Mt was ion-exchanged with 150 mL aqueous solution of SnCl4$5H2O (12.9 mmol L1) at 60
C for 12 h with stirring. The resulting slurry was filtered and repeatedly washed with deionized water until the pH of the filtrate was neutral. Subsequently, the sample was dried at 120
C. The resultant product is denoted as Sn/Mt. Then, Sn/Mt was calcined at 300, 400, or 500
C in static air for 3 h to obtain a series of Sn/Mt catalysts which are correspondingly denoted as Sn/Mt-300, Sn/Mt-400, and Sn/Mt-500. A similar method was used to synthesize tin-exchanged K10 Mt (Sn/K10). 2.3. Characterization of the catalysts The relative content of each element in the catalysts was determined via X-ray fluorescence (XRF) spectroscopy on a Shimadzu XRF-1800 X-ray fluorescence spectrometer (40 kV, 95 mA). The tin content was more accurately determined using inductively coupled plasma optical emission spectroscopy (ICP-OES) on a Teledyne Leeman Labs Prodigy7 apparatus. The specific surface areas and pore size distribution of the samples were obtained by N2 adsorptionedesorption measurements at 196.2
C using a Micromeritics ASAP 2020M system. Prior to the measurements, the samples were outgassed at 150
C for 4 h. The surface area (SBET) values of the samples were calculated using the BrunauereEmmetteTeller (BET) equation. The pore size distributions were derived from the desorption branch of the N2 isotherms using the BarretteJoynereHalenda (BJH) method. The total pore volume (Vp) values were estimated by considering the nitrogen uptake at a relative pressure (P/P0) of ~0.99. The average pore diameter (Dp) of the samples was calculated from the surface area and the total pore volume (Dp ¼ 4Vp/SBET). The crystal structures of the catalysts were characterized by powder X-ray diffraction (XRD) on a Rigaku D/MAX2550VB X-ray diffractometer (35 kV, 40 mA) with a Cu Ka source. The diffraction patterns were recorded in 2q ranging from 10
to 80
at a scan rate of 8
min1 and from 2
to 12
at a scan rate of 1
min1. The UVevisible diffuse reflectance spectra were recorded on a PerkinElmer Lambd-950 spectrophotometer using BaSO4 as a reference. The IR spectra were recorded on a Thermo Fisher Nicolet IS10 spectrometer using the KBr disk technique. The concentration of the sample in KBr was 1.0 wt%, and 0.1 g KBr was used in the preparation of the reference and sample disks. The chemical states

J. Liu et al. / Renewable Energy 120 (2018) 231e240

233

of the elements in the samples were examined via X-ray photoelectron spectroscopy (XPS), which was performed on an Axis Ultra, Kratos (UK) system using an Al Ka source (15 kV, 1486.6 eV). The vacuum in the spectrometer was <109 Torr. The binding energy was calibrated relative to the C 1s peak (284.8 eV). The acidity of the catalysts was assessed by conducting temperature-programmed desorption of ammonia (NH3-TPD). The analysis was performed using a Micromeritics AutoChem II 2920 apparatus equipped with a Cirrus™2 quadrupole mass spectrometry detector. Typically, 50 mg sample was loaded into the sample tube and pretreated in He at 150
C for 1 h to remove adsorbed species on the surface, then cooled to 100
C in a flow of He. Subsequently, a flow of 5% NH3 in He was introduced and allowed to adsorb onto the sample at 100
C for 2 h. Then, the sample was purged with dry He at the same temperature to remove the physically adsorbed NH3. Finally, the sample was heated at a linear rate of 10
C min1 to 600
C under dry He. The desorbed ammonia (m/ z ¼ 16) was monitored using the quadrupole mass spectrometer. Evaluation of Brønsted acid and Lewis acid sites on the surface of the catalysts was conducted via pyridine adsorption. Fourier transform infrared (FT-IR) spectra of pyridine adsorbed onto the catalyst samples were recorded on a Thermo Fisher Nicolet IS10 spectrometer, equipped with a deuterium triglycine sulfate detector. Each catalyst sample (30 mg) was compressed into a selfsupporting wafer with a diameter of 13 mm, which was subsequently placed in a quartz IR cell equipped with a CaF2 window and a vacuum system. The cell was pretreated in situ under vacuum (6  103 Pa) at 150
C for 60 min and then cooled to room temperature, followed by exposure to pyridine vapor for 30 min. The IR spectra were recorded at room temperature after subsequent evacuation at 150
C for 1 h. Quantitative evaluation of Brønsted (B) and Lewis (L) acid sites was performed using the integrated area of the adsorption bands at ~1540 and 1450 (or 1445) cm1 and the molar extinction coefficient (εB ¼ 1.67 cm mmol1 and 1 εL ¼ 2.22 cm mmol ) method using the following equations [46]:

an ice-water bath and depressurized. The solid catalyst and the liquid products were separated by filtration. Methyl levulinate, methyl lactate, methanol, and 5methoxymethylfurfural in the reaction products were analyzed on a gas chromatography (Agilent 6820, FID detector) equipped with a HP-FFAP capillary column (30 m  0.32 mm  0.25 mm). nButyl alcohol was used as an internal standard. Glucose and methylglucoside were determined on a high-performance liquid chromatography (Shimadzu LC-20AT, RID-10A detector) with and a Shim-pack SPR-Ca(G) column (50 cm  7.8 mm), using ultrapure water as the mobile phase at a flow rate of 0.3 mL min1. The amount of glucose and products were determined using calibration curves. The conversion of glucose and methanol were calculated using the following equations:

IA pR2 CX ¼ X ; WεX

Methyl lactate yield ð%Þ ¼

Methanol conversion ð%Þ ¼

moles of converted methanol moles of initial methanol  100: (3)

The yields of products were calculated using the following equation:

Yi ð%Þ ¼

moles of product i  100; moles of initial glucose

(4)

where Yi refers to the yield of methyl levulinate, methylglucoside, or 5-methoxymethylfurfural. The yields of methyl lactate were calculated using the following equations:

(1)

where X represents B or L; C is the concentration of the acid sites (mmol g1); IA is the integrated absorbance; R is the radius of the catalyst disk (0.65 cm); and W is the weight of the catalyst (30 mg).

moles of converted glucose  100; moles of initial glucose (2)

Glucose conversion ð%Þ ¼

moles of methyl lactate 1   100; moles of initial glucose 2 (5)

3. Results and discussion

2.4. Catalytic reaction procedure and products analysis

3.1. Structural characteristics of the catalysts

All reactions were performed in a stainless steel autoclave (35 mL). In a typical experiment, 0.3 g glucose, 0.15 g catalyst, and 24 g methanol were charged into the reactor. The autoclave was purged thrice with pure N2 and then pressurized to 2.0 MPa with N2 at room temperature. The temperature of the reactor was then raised to 220
C and held for 6 h under a stirring rate of 500 rpm. After the reaction, the reactor was cooled to room temperature in

The elemental compositions of the catalysts obtained by ICPOES and XRF are presented in Table 1. Na/Mt and K10 Mt primarily consisted of octahedral Al3þ and tetrahedral Si4þ with Naþ, Kþ, and Ca2þ as exchangeable interlayer cations, and small amounts of Fe, Mg, and Ti impurity cations [45]. Following exchange with Sn4þ ions, the Naþ and Ca2þ cations in Na/Mt were no longer detected, indicating that Sn4þ ions occupied the internal vacancies originally

Table 1 The compositions of catalysts as determined by XRF. Sample

Sn content (wt%)a

Content (wt %) SnO2

SiO2

Al2O3

MgO

Na2O

CaO

Fe2O3

K2O

TiO2

K10 Sn/K10 Sn/Mt Sn/Mt-300 Sn/Mt-400 Sn/Mt-500

e 15.1 14.2 19.5 19.6 19.3

e 23.7 27.7 26.4 25.6 25.4

79.4 61.3 51.1 49.8 50.2 50.2

15.2 11.2 19.2 19.3 19.8 19.8

1.4 0.9 2.6 2.6 2.8 2.8

e e e e e e

0.4 e e e e e

1.5 1.1 1.2 1.1 1.1 1.1

1.63 1.30 0.49 0.51 0.44 0.45

0.34 0.30 0.12 0.14 0.09 0.10

a

Measured by ICP-OES.

234

J. Liu et al. / Renewable Energy 120 (2018) 231e240

occupied by exchangeable Naþ and Ca2þ ions. For Sn/K10, Sn4þ ions were introduced into the interlayer of Mt by replacing Hþ and Ca2þ. Sn/K10 and Sn/Mt had similar Sn contents (~15 wt%). After calcination of Sn/Mt at different temperatures (300e500
C), aqua ions of tin in Sn/Mt may gradually transform into tin oxide, and the Sn content increased to ~19.5 wt% owing to the removal of interlayer H2O molecules. The low-angle XRD patterns of the catalysts are presented in Fig. 1A. The (001) reflection of Sn/Mt was located at 2q of 5.9
, which shifted to a lower 2q position in comparison with that of Na/ Mt at 2q of 7.2
[45], indicating an increase in the basal spacing owing to the displacement of Naþ and Ca2þ in the interlayer space

Fig. 1. X-ray diffraction patterns (A: low-angel; B: wide-angel) for various catalysts: (a) Sn/Mt, (b) Sn/Mt-300, (c) Sn/Mt-400, (d) Sn/Mt-500, (e) Sn/K10, (f) K10.

of Na/Mt with larger hydrated metal Sn4þ cations. The (001) reflection was not observed in the XRD patterns of Sn/Mt-300, Sn/ Mt-400, and Sn/Mt-500. This result indicated that the Mt layered structure collapsed upon calcination. In contrast, the (001) reflection of K10 Mt was observed at 2q ¼ 5.8
, corresponding to a d001 spacing of 15.1 Å. Following exchange with Sn4þ ions (Sn/K10), the intensity of the (001) reflection decreased, suggesting a reduced crystallinity, which is probably caused by partial delamination of the Mt layered structure owing to the acidic nature of the SnCl4$5H2O aqueous solution. The wide-angle XRD patterns of the catalysts are shown in Fig.1B. Sn/Mt featured some diffraction peaks at 2q of 19.8
, 34.9
, 61.9
, and 73.4
, that were assigned to the Mt sheets. The peak at 26.6
was assigned to quartz. Sn/Mt showed similar XRD pattern with Na/Mt as observed in our previous study [45]. However, with the introduction of Sn4þ ions to Na/Mt, the intensity of the diffraction decreased, indicating a reduced crystallinity. Moreover, the intensity of the diffraction peaks decreased further upon calcination of Sn/Mt at increasing temperatures. K10 Mt displayed a more complex phase structure that contained impurities i.e., quartz, tridymite-O, and actinolite in addition to the mineral clay Mt. Sn/K10 featured characteristic diffractions of the Mt sheets with considerably reduced intensity, and similarly indicated a reduced crystallinity. In all the tin-exchanged samples prepared, no diffraction peaks corresponding to Sn compounds were detected, implying that the Sn species were either amorphous or highly dispersed. The UVevis spectra of samples (see Fig. S1) show that both Sn/ Mt and Sn/K10 display neither a broad electronic absorption with a maximum around 290 nm due to octahedrally coordinated Sn4þ in SnO2 nor an electronic absorption band at 220 nm due to tetrahedrally coordinated Sn4þ species in the framework positions [47,48], which led us to believe that hydrated forms of isolated SnIV located in the interlayer space of Na/Mt or K10 are responsible for the absorption in the range of 200e270 nm. However, for the calcined Sn/ Mt samples, with increasing calcined temperature the electronic absorption band was broaden, and became similar to that of SnO2, suggesting that the gradual transformation of hydrated Sn4þ cations to SnO2 during calcination. The N2 adsorptionedesorption results in Fig. S2 show that all the prepared catalysts displayed Type IV isotherms, with H4shaped hysteresis loops in the P/P0 range of 0.45e1.0, which are characteristic of narrow slit-like pores. The average pore diameters of the samples were ~5e7 nm. The detailed textural properties of the catalysts as determined by the nitrogen physisorption analysis are summarized in Table 2. K10 Mt featured a relatively high specific surface area of 103 m2 g1, a total pore volume of 0.178 cm3 g1, and an average pore diameter of 6.9 nm. After the introduction of Sn4þ ions into K10 Mt, the surface area and pore volume increased to 160 m2 g1 and 0.224 cm3 g1, respectively, whereas the average pore diameter decreased to 5.6 nm. Na/Mt featured a specific surface area of 57 m2 g1, a total pore volume of 0.102 cm3 g1, and an average pore diameter of 7.1 nm [45]. The introduction of Sn4þ ions in Na/Mt considerably enhanced the specific surface area and pore volume, but reduced the average pore diameter. The increase in the specific area and pore volume could be due to delamination or destruction of the laminar structure of the Mt sheets in an acidic environment [49]. In contrast, the decrease in the average pore diameter of Sn/K10 and Sn/Mt may be due to filling of the pores by tin cations during exchange. Relative to those of Sn/Mt, the specific surface area and pore volume of Sn/Mt-300, Sn/Mt-400, and Sn/Mt500 were lower. The observed reduction in the specific surface area and pore volume could be due to partial collapse of the Mt layered structure upon calcination. The FT-IR spectra of the samples are presented in Fig. 2. The overall spectral features of the tin-exchanged Mt catalysts were

J. Liu et al. / Renewable Energy 120 (2018) 231e240

235

Table 2 The textual properties and acidities of catalysts. Catalyst

SBET (m2 g1)

Vp (cm3 g1)

Dp (nm)

Total acidity (mmol NH3/gcat)a

Brønsted Acidity (mmol/gcat)b

Lewis Acidity (mmol/gcat)b

B/L acidity ratio

K10 Sn/K10 Sn/Mt Sn/Mt-300 Sn/Mt-400 Sn/Mt-500

103 160 178 151 126 133

0.178 0.224 0.182 0.166 0.154 0.173

6.9 5.6 4.1 4.4 4.9 5.2

0.63 0.78 0.63 0.44 0.36 0.24

57.8 23.4 176.7 56.2 7.3 0

127.3 194.5 159.6 101.9 131.4 142.0

0.45 0.12 1.11 0.55 0.06 0

a b

The amount of total acid sites determined by NH3-TPD. The amount of acid sites determined by pyridine adsorption FT-IR.

comparable with those of the parent Mts. Specifically, the band at 3626 cm1 was assigned to OH (AlAleOH) stretching vibrations in the Mt lattice [50]. The bands at 3441 and 1635 cm1 were attributed to the stretching vibrations and deformation of interlayer water molecules in Mt, respectively. The most intense bands at 1094 and 1044 cm1 were assigned to the SieO out-of-plane and in-plane stretching vibrations, respectively [50]. The band at 917 cm1 was attributed to OH bending vibrations in octahedral AlAleOH [50]. The bands at 528 and 467 cm1 were assigned to SieOeAl and SieOeSi bending vibrations, respectively [51]. The bands at 792 and 704 cm1 were ascribed to tridymite and quartz, respectively [50]. No bands at ~1440 cm1 were observed in the tinexchanged Mt samples. The band at 1440 cm1 assigned to contaminant CO2 species on Na/Mt were removed during ex3 change with tin ions [45]. The intensity of the bands at 3626, 3441, and 1635 cm1 decreased gradually for the Sn/Mt samples calcined at increasing temperatures, suggesting a reduction in the amount of hydroxyl groups coordinated to the octahedral cations and water molecules in the Mt interlayers, possibly resulting in a decrease in Brønsted acidity. XPS analysis (see Table S1) showed that the tin-exchanged Mt catalysts contained Sn, Al, Si, O, and C only, thus indicating that the impurities (Fe, Mg, Cr, and Ti) deposited within the clay interlayers. The Sn 4d5/2 binding energy for all samples was 487.1 eV (see Fig. S3), confirming that Sn in the tin-exchanged catalysts was present as Sn(IV).

Fig. 2. FTIR spectra of various catalysts: (a) Sn/Mt, (b) Sn/Mt-300, (c) Sn/Mt-400, (d) Sn/Mt-500, (e) K10, (f) Sn/K10.

3.2. Acidic characteristics of the catalysts The NH3-TPD curves of the parent and tin-exchanged Mt samples are shown in Fig. 3. The amount of total acid sites is shown in Table 2. As observed in our previous study [45], Na/Mt displayed only a small desorption peak at ~125
C, suggesting that this sample contained only weak acidic sites on the surface, and the total acid amount was 0.02 mmol NH3/gcat. However, the total acidity significantly increased to 0.63 mmol NH3/gcat following incorporation of the tin ions into the Na/Mt interlayers. In contrast, the NH3-TPD curve of Sn/Mt displayed a strong peak centered at ~140
C and a relatively weak and broad peak centered at 280
C, indicating that Sn/Mt contained substantial amounts of weak acid sites and some medium strong acidic sites. The amount of total acidic sites on the tin-exchanged Mt samples calcined at different temperatures decreased gradually with increasing calcination temperatures from 300 to 500
C, indicating that calcination destroyed both weak and strong acidic sites on the catalysts. K10 Mt displayed three NH3 desorption peaks centered at 130, 300, and 510
C, indicative of the existence of three different types of acidic sites: weak, medium, and strong. Conversely, Sn/K10 exhibited only two NH3 desorption peaks, a peak centered at ~130
C and a broad NH3 desorption band centered at 330
C. The results revealed that the introduction of tin into the K10 Mt interlayers modified the acidity of K10 by increasing the amount of medium strong acidic sites and depleting the content of strong acidic sites.

Fig. 3. NH3-TPD-MS curves of various catalysts: (a) Sn/Mt, (b) Sn/Mt-300, (c) Sn/Mt400, (d) Sn/Mt-500, (e) K10, (f) Sn/K10.

236

J. Liu et al. / Renewable Energy 120 (2018) 231e240

attributed to the loss of water molecules attached to the hydration shell of the tin cation [54]. K10 Mt displayed bands corresponding to pyridine adsorbed on both Brønsted and Lewis acid sites. Brønsted acidity in K10 arises from Hþ ions occupying exchange sites on the surface or interlayers. Lewis acidity arises from A13þ, Fe3þ/Fe4þ, and Mg2þ present in the octahedral sheets of K10 [49]. After exchange with Sn ions, the intensity of the band at 1540 cm1 decreased, indicating a reduction in the amount of Brønsted acid sites because of the exchange of Hþ with Sn4þ. In contrast, the band at 1445 cm1 shifted to higher wavenumbers (~1449 cm1), suggesting an increase in Lewis acid strength. Furthermore, the amount of Lewis acid sites increased from 127.3 to 194.5 mmol g1 cat. 3.3. Catalytic activity

Fig. 4. FTIR spectra of pyridine adsorbed on various catalysts: (a) Sn/Mt, (b) Sn/Mt300, (c) Sn/Mt-400, (d) Sn/Mt-500, (e) K10, (f) Sn/K10.

The FT-IR spectra of pyridine adsorption on the prepared catalysts are shown in Fig. 4. The amount of Brønsted and Lewis acid sites on the samples was calculated based on the bands at ~1540 and 1450 or 1445 cm1, respectively, and the results are presented in Table 2. The bands were assigned according to the literature [52e54].The bands at 1540 and 1640 cm1 were attributed to pyridinium ion (PyHþ) produced upon reaction between pyridine and Brønsted acid sites. (The latter sites show a characteristic band at 1540 cm1) The bands at ~1450 and 1610 cm1 were assigned to coordinatively bound pyridine on Lewis acid sites, and the 1450 cm1 band is typical of such sites. The band at 1490 cm1 is typically associated with vibrations owing to PyHþ and coordinatively bound pyridine. No bands corresponding to pyridine adsorbed onto Brønsted and Lewis acid sites were observed on Na/ Mt, indicating the absence of both Brønsted and Lewis acidity on this catalyst [45]. After the introduction of Sn ions into Na/Mt, the spectral features associated with Brønsted and Lewis acid sites were more prominent, as evidenced by the appearance of absorbance at 1540 and 1445 cm1 in the spectrum of Sn/Mt. Upon calcination of Sn/Mt, the band at 1540 cm1 weakened gradually and eventually disappeared, indicating a reduction in the amount of Brønsted sites on the calcined Sn/Mt catalysts. In contrast, the band at 1452 cm1 shifted to lower wavenumbers (~1445 cm1), suggesting a decrease in Lewis acid strength. It is known that clay catalysts contain both Brønsted and Lewis acid sites, with the Brønsted sites being mainly associated with the interlamellar region [53]. For the tin-exchanged Mt catalysts, the acidity in clays arises from Hþ ions produced upon dissociation of water molecules in the hydration spheres of metallic Sn4þ cation under the polarizing effect of the Sn4þ cation [49,53]: [Sn(H2O)x]4þ / [Sn(OH)(H2O)x1]3þ þ Hþ,

(6)

where x is the number of water molecules that directly coordinate to the metal Sn cation. When coordination of an organic compound to the metal Sn4þ cation is accompanied by expulsion of water molecules that are originally coordinated to the Sn4þ cation, the Sn4þ cation becomes an electron pair acceptor and acts as a Lewis acid site [51].The acidity of tin-exchanged Mt is influenced by the amount of water retained between its clay layers. The reduced Brønsted acidity of the Sn/Mt catalysts upon thermal treatment above 300
C can be

The conversion of glucose in methanol over the prepared Mt catalysts was assessed, and the results are shown in Table 3. For comparison, the reported values of H-USY [19], H-USY treated with nitric acid [22] and sulfated Mt [23] were also added in Table 3. As observed, the small molecular liquid products obtained from the conversion of glucose included methyl levulinate, methylglucoside, methyl lactate, 5-methoxymethylfurfural and methyl formate. Methylfructoside was not detected in the present system. Besides, some dark-brown insoluble substances i.e. humins were also observed, which were formed by side reactions of the acidcatalyzed decompositions of reactant and/or certain products under the experimental conditions. During reaction, methanol was mainly converted into dimethyl ether through the intermolecular dehydration of methanol, which is a useful precursor to other organic compounds and an aerosol propellant. The experimental results and literature data suggest two possible reaction pathways for the conversion of glucose to methyl levulinate [9,13,17,19,22,23] (see Scheme 1). Pathway A involves the isomerization of glucose to fructose, dehydration of fructose to 5hydroxymethylfurfural, and subsequent etherification to form 5methoxymethylfurfural, followed by rehydration and methanol addition to form methyl levulinate. In pathway B, glucose directly reacts with methanol to form methylglucoside, which undergoes dehydration to form 5-methoxymethylfurfural. The latter undergoes rehydration and methanol addition to form methyl levulinate. However, the conversion of methylglucoside to 5methoxymethylfurfural is slow and need higher temperatures, which was confirmed by the following studies. We found that methylglucoside was only detected below 200
C; above 200
C methylglucoside was converted completely. Both Brønsted acid and Lewis acid sites are involved in the methanolysis of glucose. The Lewis acid sites are mainly responsible for the isomerization of glucose to fructose, whereas the Brønsted acid sites catalyze the other routes either by itself or in the cooperation with Lewis acid sites [9,13e15]. Hence, the correct balance between Lewis and Brønsted acid sites is critical to the success of this complex tandem transformation [20,24]. On the other hand, fructose can also be converted into trioses i.e dihydroxyacetone and glyceraldehyde through the retro-aldol condensation. These trioses are readily converted to methyl lactate through sequential dehydration and methanol addition, followed by a 1,2-hydride shift [55]. The data in Table 3 show that glucose was converted completely at 220
C in all the cases. At 150
C, similar glucose conversions (80%85%) were obtained on all the catalysts. These results suggest that glucose conversion mainly depends on the reaction temperature, whereas the acidity of the catalyst did not have significantly influence on glucose conversion. However, the yield of methyl levulinate strongly depends upon the acidity of the catalysts.

J. Liu et al. / Renewable Energy 120 (2018) 231e240

237

Table 3 The conversion of glucose to methyl levulinate over various catalystsa. Temp (
C)

Catalysts

No catalyst SnCl4$5H2O

b

K10 Sn/K10 Sn/Mt Sn/Mtc Sn/Mt-300 Sn/Mt-400 Sn/Mt-500 H-USYd H-USY-0.2e f 20-SO2 4 /Mt a b c d e f

220 150 220 150 220 150 220 150 220 150 220 220 150 220 150 220 150 160 180 200

Conversion (%)

Yield (%)

Glucose

Methanol

Methyl levulinate

Methyl glucoside

Methyl lactate

5-methoxymethyl-furfural

100 78 100 82 100 84 100 83 100 85 100 100 84 100 83 100 81 >99 100 >99

5.5 4.4 12.0 6.7 21.1 10.0 22.6 11.0 26.1 14.4 27.6 23.9 12.6 19.5 11.5 15.2 10.1 e e e

e e 11.5 3.4 36.9 21.5 42.7 22.7 59.7 32.4 56.1 51.0 28.2 35.5 14.2 23.8 5.9 49 54 48

3.8 45.5 8.2 19.0 e 29.8 e 30.9 e 34.2 e e 24.6 3.5 25.1 12.6 24.0 e e e

10.2 1.1 30.3 18.3 7.6 0.7 5.0 0.9 4.2 1.0 5.4 5.6 1.6 9.7 1.8 9.1 1.6 e e e

e e 0.36 0.1 1.96 0.43 1.00 0.74 0.24 0.51 0.12 0.32 0.53 0.80 1.16 1.09 1.50 e e e

Reaction conditions: methanol 24 g, glucose 0.3 g, catalyst 0.15 g, 2 MPa N2, 6 h. 0.0628 g SnCl4$5H2O containing the equal moles of Sn in 0.15 g Sn/Mt. The addition of 0.5 g H2O in the reaction system. The data was obtained from Ref. [19] under the conditions of 0.25 g glucose, 10 mL methanol, 0.15 g catalyst, 20 bar Ar, 20 h. The data was obtained from Ref. [24] under the conditions of 1.12 g glucose, 36 g methanol, 0.3 g catalyst, 0.1 MPa N2, 20 h. The data was obtained from Ref. [25] under the conditions of 1 mmol glucose, 20 mL methanol, 0.15 g catalyst, 200
C, 4 h.

In the absence of catalysts, at 220
C, only a small amount of methyl lactate (10.2%) and methyl glucoside (3.8%) were obtained. The main product was attributed to insoluble matter generated upon polymerization of glucose at high temperatures [56]. In the presence of SnCl4$5H2O, a typical Lewis acid catalyst, the main liquid products were methyl lactate (30.3%), methyl levulinate (11.5%), methylglucoside (8.2%), together with trace amounts of 5methoxymethylfurfural. Using K10 Mt as the catalyst, at 150
C, the conversion of glucose was 84%, and the main detected products were methyl levulinate (21.5%) and methylglucoside (29.8%). Similar results were obtained for the reaction conducted over Sn/ K10 at 150
C. However, when the reaction temperature was raised to 220
C, methyl levulinate was obtained as the main product with a yield of 42.7% on Sn/K10 catalyst, which was higher than that

obtained using K10 (36.9%) as catalyst. Furthermore, methylglucoside was not detected, suggesting that the higher temperature promoted the conversion of methylglucoside. The higher yield of methyl levulinate obtained in the presence of Sn/K10 when compared with that obtained in the presence of K10 was mainly attributed to the higher amounts of Lewis acid sites on Sn/K10, which promoted the isomerization of glucose to fructose. Additionally, the reduction of strong acid sites on Sn/K10 as shown by NH3-TPD measurments could suppress the formation of humintype polymers. Previously, Kuo et al. also confirmed that the relatively stronger acidities on H-type zeolites were not suitable for fructose conversion into methyl levulinate [24]. In the presence of Na/Mt as the catalyst, at 220
C, only small amounts of methyl levulinate (2.6%) and methyl lactate (12.7%)

Scheme 1. Proposed reaction pathways for glucose conversion in methanol.

238

J. Liu et al. / Renewable Energy 120 (2018) 231e240

Fig. 5. Effect of reaction temperature on glucose conversion (reaction conditions: methanol 24 g, glucose 0.3 g, catalyst 0.15 g, 2 MPa N2, 6 h).

Fig. 7. Effect of glucose amount on the yields of products (reaction conditions: methanol 24 g, catalyst 0.15 g, 220
C, 6 h, 2 MPa N2).

were detected [45]. The low selectivity to methyl levulinate on Na/ Mt could be ascribed to the absence of both Brønsted and Lewis acid sites on the catalyst. However, after introducing Sn4þ ions into Na/Mt, the yield of methyl levulinate improved significantly from 2.6 to 59.7% at 220
C because of the increase in the amounts of both Brønsted and Lewis acid sites. However, for the calcined Sn/Mt catalysts, with the increasing calcination temperatures, the yield of methyl levulinate decreased gradually from 51.0 to 23.8% at a reaction temperature of 220
C. At 150
C, the yields of both methyl levulinate and methylglucoside decreased with increasing calcination temperatures. These results were attributed mainly to the decrease in the amount of total acid sites, especially Brønsted acid sites, upon calcination. Among all the catalysts studied, Sn/Mt afforded the highest yield of methyl levulinate of 59.7% under the following conditions: 0.3 g glucose, 24 g methanol, 0.15 g catalyst, 220
C, 2 MPa N2, 6 h. The highest methyl levulinate yield of 59.7% was obtained. The corre1 sponding methyl levulinate productivity was 1.1 mmol g1 cat h , which is higher than those obtained from the conversion of glucose 1 in methanol catalyzed by H-USY (0.227 mmol g1 cat h ) [19], H-USY 1 treated with nitric acid (0.55 mmol g1 h ) [22] and sulfated Mt cat 1 (0.8 mmol g1 cat h ) [23]. Such a performance was attributed to large amount of total acid sites as well as an appropriate balance

between the Brønsted and Lewis acid sites on the catalyst as confirmed by the FT-IR spectra of pyridine adsorption. Because role of water is crucial in the sequential transformation of glucose into levulinate derivatives, the effect of small amounts of water on the reaction was investigated. The results in Table 3 showed that the yield of methyl levulinate decreased slightly; whereas the conversions of both glucose and methanol remained almost constant, suggesting that small amounts of water do not have significant effect on the reaction. The reaction temperature dependence of glucose conversion catalyzed by Sn/Mt was investigated, and the results are shown in Fig. 5. As observed in Fig. 5, the conversions of both glucose and methanol increased with increasing reaction temperatures. In contrast, the yield of methyl levulinate increased initially and then decreased. This result was probably due to the decomposition of methyl levulinate to some extent at higher temperatures [9]. The optimal yield of methyl levulinate was obtained at 220
C. The yield of methylglucoside decreased gradually with increasing reaction temperatures. At reaction temperatures above 200
C, methylglucoside was not detected among the products. These results suggest that high temperatures could promote the transformation of 5-methoxymethylfurfural into methyl levulinate through

Fig. 6. Effect of reaction time on glucose conversion (reaction conditions: methanol 24 g, glucose 0.3 g, catalyst 0.15 g, 220
C, 2 MPa N2).

Fig. 8. Recycling of Sn/Mt for the conversion of glucose (reaction conditions: methanol 24 g, glucose 0.3 g, catalyst 0.15 g, 150
C, 6h, 2 MPa N2).

J. Liu et al. / Renewable Energy 120 (2018) 231e240

239

Table 4 The conversion of various substrates using the Sn/Mt catalyst. Substrate

Fructose Sucrose Starch Inulin Cellubiose Cellulose

Temp (
C)

220 220 220 220 220 220

Conversion (%)

Yield (%)

Substrate

Methanol

Methyl levulinate

Methyl glucoside

Methyl lactate

5-methoxymethyl -furfural

100 100 100 100 100 100

28.3 28.7 22.9 22.8 28.0 32.5

65.6 62.0 45.5 55.1 55.7 19.4

e e e e e 0.98

3.1 4.2 7.8 2.5 3.7 4.1

e 0.65 3.31 0.06 0.46 1.69

Reaction conditions: methanol 24 g, catalyst 0.15 g, substrate 0.3 g, 2 MPa N2, 6 h.

rehydration and methanol addition. Subsequently, the effect of reaction time on the glucose conversion performance was examined. As shown in Fig. 6, at 220
C, with increasing reaction times, the yield of methyl levulinate gradually increased from 50.6 to 59.7% with increasing reaction times from 2 to 6 h, after which a plateau was observed. Methylglucoside was not detected under the reaction time conditions studied. It can be deduced that the reaction time has little influence on the glucose and methanol conversion and methyl lactate yield. In contrast, the conversions of glucose and methanol and the yields of methylglucoside and methyl lactate mainly depend on the reaction temperature. The effect of the glucose amount on the conversion performance was then examined. As shown in Fig. 7, with increasing amounts of glucose, the yield of methyl levulinate gradually decreased although glucose was fully converted. When the amount of glucose was increased to 0.6 g, the yield of methyl levulinate decreased to 48%, and the total yields of detectable liquid products also decreased. However, a higher content of dark, insoluble matter was observed in the autoclave. We believe that the insoluble matter was likely formed by partial polymerization of glucose because lesser amounts of acidic sites were available in that case. The re-usability of Sn/Mt catalyst was examined by conducting successive reaction cycles. After each cycle, the catalyst was recovered from the reaction mixture by centrifugation and subjected to treatment with H2O2 solution (40 mL, 28.5%) at 35
C overnight under stirring to remove deposited carbon species. Subsequently, the catalyst was filtered, washed with distilled water, and dried at 120
C overnight, and then used in the next run. The results are summarized in Fig. 8. In the first run, the conversion of glucose was 84.6%, the yields of methyl levulinate and methyl glucoside were 32.4 and 34.2%, respectively. During recycling, the conversion of glucose increased in the 2nd and 3rd runs, and then decreased slightly, subsequently remaining stable (~83%) during the 4th and 5th runs. The yield of methylglucoside changed in the same order and remained at ~28% after the 4th run. In contrast, the yield of methyl levulinate decreased in the 2nd run, then increased slightly in the 3rd run, subsequently remaining stable at ~28%. The results suggest the occurrence of some structural changes during recycling or treatment of the Sn/Mt with H2O2. ICP analysis confirmed that none of Sn, Al, and Si, other species leached into the reaction solution at both 150 and 220
C. The XRD pattern of the used catalyst (see Fig. S4) did not change considerably when compared with that of fresh Sn/Mt. Both the fresh and used catalysts featured similar FT-IR spectra (see Fig. S5). XPS analysis (see Table S1) showed that the oxidation state of Sn did not change after reaction. However, the Si/Al atomic ratio decreased and the Sn/Si atomic ratio increased, confirming that surface reconstruction on the used Sn/Mt occurred. Additionally, the catalytic performance of Sn/Mt in the conversion of a variety of carbohydrates was examined including monosaccharide, disaccharide, and polysaccharide, and the results are

shown in Table 4. All the substrates were fully converted at 220
C. The yield of methyl levulinate decreased in the order of fructose > sucrose > inulin z cellobiose > starch > cellulose. The result implies that fructose and its polymer are more prone to conversion into methyl levulinate than glucose and its polymer because a demanding isomerization step is avoided when starting with fructose and its polymer. 4. Conclusions We demonstrated that tin-exchanged Mt could effectively catalyze the conversion of monosaccharide, disaccharide, and polysaccharide in methanol into methyl levulinate. A high methyl levulinate yield of ~60% was obtained when glucose was used as the substrate. The prepared catalyst can be re-used for at least five times with slight losses observed in the methyl levulinate yield. The excellent catalytic performance of the tin-exchanged Mt was attributed to the large amount of acid sites and a good balance of Brønsted and Lewis acid sites on the catalyst. Acknowledgments The authors gratefully acknowledge financial support from the National Natural Science Foundation of China (Grant no. 21576161) and the Fundamental Research Funds for the Central Universities (GK201706011). Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.renene.2017.12.104. References
rez, I.T. Horva
th, R.A. Sheldon, M. Poliakoff, Valorization of [1] C.O. Tuck, E. Pe biomass: deriving more value from waste, Science 337 (2012) 695e699. [2] P. Gallezot, Conversion of biomass to selected chemical products, Chem. Soc. Rev. 41 (2012) 1538e1558. [3] S. Zhao, M. Cheng, J. Li, J. Tian, X. Wang, One pot production of 5hydroymethylfurfural with high yield from cellulose by a Brønsted-Lewissurfactant-combined heteropolyacid catalyst, Chem. Commun. 47 (2011) 2176e2178. [4] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Metal chlorides in ionic liquid solvents convert sugars to 5-hydroxymethylfurfural, Science 316 (2007) 1597e1600. [5] D.W. Rackemann, W.O.S. Doherty, The conversion of lignocellulosics to levulinic acid, Biofuels, Bioprod Biorefin 5 (2011) 198e214. [6] J. Zhang, S.B. Wu, B. Li, H.D. Zhang, Advances in the catalytic production of valuable levulinic acid derivatives, ChemCatChem 4 (2012) 1230e1237. [7] A. Demolis, N. Essayem, F. Rataboul, Synthesis and applications of alkyl levulinates, ACS Sustainable Chem Eng 2 (2014) 1338e1352. [8] X. Hu, Y. Song, L. Wu, M. Gholizadeh, C.-Z. Li, One-pot synthesis of levulinic acid/ester from C5 carbohydrates in a methanol medium, ACS Sustainable Chem Eng 1 (2013) 1593e1599. [9] Y. Liu, C.-L. Liu, H.-Z. Wu, W.-S. Dong, An efficient catalyst for the conversion of fructose into methyl levulinate, Catal. Lett. 143 (2013) 1346e1353. [10] R. Liu, J. Chen, X. Huang, L. Chen, L. Ma, X. Li, Conversion of fructose into 5hydroxymethylfurfural and alkyl levulinates catalyzed by sulfonic acid-

240

J. Liu et al. / Renewable Energy 120 (2018) 231e240

functionalized carbon materials, Green Chem. 15 (2013) 2895e2903. [11] L. Peng, H. Li, L. Lin, K. Chen, Effect of metal salts existence during the acidcatalyzed conversion of glucose in methanol medium, Catal. Commun. 59 (2015) 10e13. [12] J. Chen, G. Zhao, L. Chen, Efficient production of 5-hydroxymethylfurfural and alkyl levulinate from biomass carbohydrate using ionic liquid-based polyoxometalate salts, RSC Adv 4 (2014) 4194e4202. [13] L. Peng, L. Lin, H. Li, Extremely low sulfuric acid catalyst system for synthesis of methyl levulinate from glucose, Ind. Crop. Prod. 40 (2012) 136e144. [14] S. Saravanamurugan, O.N.V. Buu, A. Riisager, Conversion of mono- and disaccharides to ethyl levulinate and ethyl pyranoside with sulfonic acid functionalized ionic liquids, ChemSusChem 4 (2011) 723e726. [15] L. Zhou, H. Zou, J. Nan, L. Wu, X. Yang, Y. Su, T. Lu, J. Xu, Conversion of carbohydrate biomass to methyl levulinate with Al2(SO4)3 as a simple, cheap and efficient catalyst, Catal. Commun. 50 (2014) 13e16. [16] C. Chang, G. Xu, W. Zhu, J. Bai, S. Fang, One-pot production of a liquid biofuel candidate-ethyl levulinate from glucose and furfural residues using a combination of extremely low sulfuric acid and zeolite USY. Fuel 140 (2015) 365e370. [17] L. Peng, L. Lin, H. Li, Q. Yang, Conversion of carbohydrates biomass into levulinate esters using heterogeneous catalysts, Appl. Energy 88 (2011) 4590e4596. [18] L. Peng, L. Lin, J. Zhang, J. Shi, S. Liu, Solid acid catalyzed glucose conversion to ethyl levulinate, Appl. Catal. A: Gen. 397 (2011) 259e265. [19] S. Saravanamurugan, A. Riisager, Zeolite catalyzed transformation of carbohydrates to alkyl levulinates, ChemCatChem 5 (2013) 1754e1757.
ndez, J. Iglesias, J.A. Melero, M. Paniagua, [20] G. Morales, A. Osatiashtiani, B. Herna D.R. Brown, M. Granollers, A.F. Leeb, K. Wilson, Conformal sulfated zirconia monolayer catalysts for the one-pot synthesis of ethyl levulinate from glucose, Chem. Commun. 50 (2014) 11742e11745. [21] E.C. Njagi, H.C. Genuino, C.-H. Kuo, S. Dharmarathna, A. Gudz, S.L. Suib, Highyield selective conversion of carbohydrates to methyl levulinate using mesoporous sulfated titania-based catalysts, Micropor Mesopor Mat 202 (2015) 68e72. [22] L. Zhou, H. Zhao, L. Cui, Y. Bai, J. Bian, T. Lu, Y. Su, X. Yang, Promotion effect of mesopore on the conversion of carbohydrates to methyl levulinate over HUSY zeolite, Catal. Commun. 71 (2015) 74e78. [23] X. Xu, X. Zhang, W. Zou, H. Yue, G. Tian, S. Feng, Conversion of carbohydrates to methyl levulinate catalyzed by sulfated montmorillonite, Catal. Commun. 62 (2015) 67e70. [24] C.-H. Kuo, A.S. Poyraz, L. Jin, Y. Meng, L. Pahalagedara, S.-Y. Chen, D.A. Kriz, C. Guild, A. Gudz, S.L. Suib, Heterogeneous acidic TiO2 nanoparticles for efficient conversion of biomass derived carbohydrates, Green Chem. 16 (2014) 785e791. [25] K. Tominaga, A. Mori, Y. Fukushima, S. Shimada, K. Sato, Mixed-acid systems for the catalytic synthesis of methyl levulinate from cellulose, Green Chem. 13 (2011) 810e812. [26] A.S. Amarasekara, B. Wiredu, Acidic ionic liquid catalyzed one-pot conversion of cellulose to ethyl levulinate and levulinic acid in ethanol-water solvent system, Bioenerg Res 7 (2014) 1237e1243. [27] Y. Hishikawa, M. Yamaguchi, S. Kubo, T. Yamada, Direct preparation of butyl levulinate by a single solvolysis process of cellulose, J. Wood Sci. 59 (2013) 179e182. [28] B. Wiredu, J.N. Dominguez, A.S. Amarasekara, The co-catalyst effect of zeolites on acidic ionic liquid catalyzed one-pot conversion of cellulose to ethyl levulinate and levulinic acid in aqueous ethanol, Curr. Catal. 4 (2015) 143e151. [29] G. Xu, C. Chang, S. Fang, X. Ma, Cellulose reactivity in ethanol at elevated temperature and the kinetics of one-pot preparation of ethyl levulinate from cellulose, Renew. Energy 78 (2015) 583e589. [30] S. Kang, J. Yu, Effect of methanol on formation of levulinates from cellulosic biomass, Ind. Eng. Chem. Res. 54 (2015) 11552e11559. [31] D. Ding, J. Xi, J. Wang, X. Liu, G. Lu, Y. Wang, Production of methyl levulinate from cellulose: selectivity and mechanism study, Green Chem. 7 (2015) 4037e4044. [32] F. Rataboul, N. Essayem, Cellulose reactivity in supercritical methanol in the presence of solid acid catalysts: direct synthesis of methyl levulinate, Ind. Eng. Chem. Res. 50 (2011) 799e805. [33] X. Wu, J. Fu, X. Lu, One-pot preparation of methyl levulinate from catalytic alcoholysis of cellulose in near-critical methanol, Carbohydr. Res. 358 (2012)

37e39. [34] H. Li, L. Peng, L. Lin, K. Chen, H. Zhang, Synthesis, isolation and characterization of methyl levulinate from cellulose catalyzed by extremely low concentration acid, J Energy Chem. 22 (2013) 895e901. [35] J. Feng, J. Jiang, J. Xu, Z. Yang, K. Wang, Q. Guan, S. Chen, Comment on processes for the direct conversion of cellulose or cellulosic biomass into levulinate esters, Appl Energ 154 (2015) 520e527. [36] M. Mascal, E.B. Nikitin, Preparation of methyl levulinate from fractionation of direct liquefied bamboo biomass, ChemSusChem 3 (2010) 1349e1351. [37] B.S. Kumar, A. Dhakshinamoorthy, K. Pitchumani, K10 montmorillonite clays as environmentally benign catalysts for organic reactions, Catal Sci Technol 4 (2014) 2378e2396. [38] J. Wang, J. Ren, X. Liu, J. Xi, Q. Xia, Y. Zu, G. Lu, Y. Wang, Direct conversion of carbohydrates to 5-hydroxymethylfurfural using Sn-Mont catalyst, Green Chem. 14 (2012) 2506e2512. [39] J. Wang, Y. Masui, M. Onaka, Conversion of triose sugars with alcohols to alkyl lactates catalyzed by Brønsted acid tin ion-exchanged montmorillonite, Appl. Catal. B: Environ. 107 (2011) 135e139. [40] H. Li, J. Ren, L. Zhong, R. Sun, L. Liang, Production of furfural from xylose, water-insoluble hemicelluloses and water-soluble fraction of corncob via a tin-loaded montmorillonite solid acid catalyst, Bioresour. Technol. 176 (2015) 242e248. [41] F.-F. Wang, J. Liu, H. Li, C.-L. Liu, R.-Z. Yang, W.-S. Dong, Conversion of cellulose to lactic acid catalyzed by erbium-exchanged montmorillonite K10, Green Chem. 17 (2015) 2455e2463. [42] K. Higuchi, M. Onaka, Y. Izumi, Solid acid and base-catalyzed cyanosilylation of carbonyl compounds with cyanotrimethylsilane, Bull. Chem. Soc. Jpn. 66 (1993) 2016e2032.
n-Leshkov, M.E. Davis, Activation of carbonyl-containing molecules [43] Y. Roma with solid Lewis acids in aqueous media, ACS Catal. 1 (2011) 1566e1580. [44] M. Moliner, State of the art of Lewis acid-containing zeolites: lessons from fine chemistry to new biomass transformation processes, Dalton Trans. 43 (2014) 4197e4208. [45] J. Liu, B.-B. Yang, X.-Q. Wang, C.-L. Liu, R.-Z. Yang, W.-S. Dong, Glucose conversion to methyl levulinate catalyzed by metal ion-exchanged montmorillonites, Appl. Clay Sci. 141 (2017) 118e124. [46] K. Saravanana, B. Tyagia, R.S. Shuklaa, H.C. Bajaj, Esterification of palmitic acid with methanol over template-assisted mesoporous sulfated zirconia solid acid catalyst, Appl. Catal. B: Environ. 172e173 (2015) 108e115. [47] P. Li, G. Liu, H. Wu, Y. Liu, J. Jiang, P. Wu, Postsynthesis and selective oxidation properties of nanosized Sn-beta zeolite, J. Phys. Chem. 115 (2011) 3663e3670. €ls, M. Dusselier, F. de Clippel, P. Vanelderen, [48] J. Dijkmans, D. Gabrie K. Houthoofd, A. Malfliet, Y. Pontikes, B.F. Sels, Productive sugar isomerization with highly active Sn in dealuminated b zeolites, Green Chem. 13 (2013) 2777e2785. [49] P. Kumar, R.V. Jasra, T.S.G. Bhat, Evolution of porosity and surface acidity in montmorillonite clay on acid activation, Ind. Eng. Chem. Res. 34 (1995) 1440e1448. [50] B. Tyagi, C.D. Chudasama, R.V. Jasra, Determination of structural modification in acid activated montmorillonite clay by FT-IR spectroscopy, Spectrochim. Acta 64 (2006) 273e278. [51] J. Madejov
a, FTIR techniques in clay mineral studies, Vib. Spectrosc. 31 (2003) 1e10. [52] K. Shimizu, T. Higuchi, E. Takasugi, T. Hatamachi, T. Kodama, A. Satsum, Characterization of Lewis acidity of cation-exchanged montmorillonite K-10 clay as effective heterogeneous catalyst for acetylation of alcohol, J. Mol. Catal. Chem. 284 (2008) 89e96. [53] L. Jankovi c, P. Komadel, Metal cation-exchanged montmorillonite catalyzed protection of aromatic aldehyde with Ac2O, J. Catal. 218 (2003) 227e233. [54] C. Ravindra Reddy, Y.S. Bhat, G. Nagendrappa, B.S.J. Prakash, Brønsted and Lewis acidity of modified montmorillonite clay catalysts determined by FT-IR spectroscopy, Catal. Today 141 (2009) 157e160. [55] M.S. Holm, S. Saravanamurugan, E. Taarning, Conversion of sugars to lactic acid derivatives using heterogeneous zeotype catalysts, Science 328 (2010) 602e605. [56] X. Hu, C. Lievens, A. Larcher, C.-Z. Li, Reaction pathways of glucose during esterification: effects of reaction parameters on the formation of humin type polymers, Bioresour. Technol. 102 (2011) 10104e10113.