Biorefineries are regarded as an alternative to use fossil resources to provide useful chemicals. Developing biorefineries has attracted growing interest because fossil resources are being depleted and are the cause of environmental concerns [1-3]. In the same way that petroleum refineries produce chemicals that are the building blocks for more complex molecules such as polymers,biorefineries will manufacture simple molecules from readily available feedstocks to allow the synthesis of biopolymers and other large molecules [4]. One approach that has been broadly studied is the production of furans from carbohydrates. 5-Hydroxymethylfurfural (HMF),a dehydration product of hexose,is one of the most important platform chemicals [5, 6] because of its potential to substitute petroleum-derived building blocks for the production of polyamides and fine products [7-9]. Developing an efficient catalytic system for the dehydration of carbohydrates to HMF has become a hot topic.
Heterogeneous catalysts such as zirconium phosphate,niobic acid,aluminosilicate,alumina and zeolites [10-13] have been investigated for the selective and catalytic conversion of carbohydrates into HMF,with promising results. Qi et al. [14] explored the dehydration of fructose to HMF with sulfated zirconia under mild conditions and achieved a yield of 72.8% in acetone-dimethylsulfoxide mixtures. Furthermore,Wang et al. [15] reported HMF preparation from fructose in 75.0% yield using a solid acid catalyst (sulfated SnO2-ZrO2) in DMSO at 120 °C for 2.5 h. Unlike conventional soluble acid catalysts or metal chlorides,insoluble acids can adjust the surface acidity of the catalytic particles to improve the selectivity for HMF [16]. Furthermore,clay-based heterogeneous catalysts have many practical and potential applications as green catalysts. Another potential catalyst is attapulgite (ATP),a hydrated octahedral layered magnesium aluminum phyllosilicate ((Mg,Al)4(Si)8(O,OH,H2O)26·nH2O). ATP is a low-cost mineral with a fibrous morphology of which China has large reserves [17, 18]. To the best of our knowledge,there are few reports of the use of natural clay particles in the conversion of carbohydrates to HMF.
γ-Valerolactone (GVL) is a valuable product that can be produced at a yield greater than 95% through hydrogenation of levulinic acid (LA) derived from HMF. GVL may be blended with gasoline in a capacity similar to ethanol or used to make organic solvents,chemicals,transportation fuels,valeric esters,and butene isomers [19-22]. Dumesic’s group [23] designed a biphasic reaction system using GVL as a solvent and a combination of Amb-70 and Sn-β as a solid acid catalyst to convert glucose to HMF in up to 59% yield.
InCl3 has emerged as an easily-handled and eco-friendly Lewis acid catalyst. In(Ⅲ) catalysts have been previously shown to enhance the conversion of cellulose to HMF [25, 26]. We propose a new solid acid In(Ⅲ) catalyst,SO42-/In2O3-ATP,using GVL as the solvent to prepare HMF from hexoses (Scheme 1). Herein,we report the preparation of catalysts by an impregnation-calcination method using ATP powder with indium as a metal precursor and investigate the selectivity and efficiency of the conversion of hexoses (glucose and fructose) to HMF. The effects of different ATP-based catalysts,reaction temperature,reaction time and acid concentration were investigated in detail.
Glucose (99%),D-fructose (99%),H2SO4 (95%-98%),and (NH4)SO4 (99%) were purchased from Sinopharm Chemical Reagent Co.,Ltd (Beijing,China) and used without further purification. Indium (99.99%) was purchased from Xiya Reagent (Shandong,China). GVL (98%) was obtained from Adamas Reagent Co.,Ltd (Shanghai,China). HMF (98%) and levulinic acid (98%) were supplied by Acros Organics.
SO42-/In2O3-ATP was prepared by an impregnation- calcination method. ATP powder (50 g) was placed in a 500-mL beaker containing concentrated H2SO4 (15 mL) and Indium (2 g). The powders were dispersed on a shaker,(NH4)2SO4 solution (60 mL,1 mol/L,equal to 40 g of (NH4)2SO4) was added and the dispersion shaken violently for 180 min. The resulting gray solid was isolated by filtration and further calcined at 550 °C for 120 min. The calcined solid was immersed in concentrated sulfuric acid solution for 120 min. The solid was isolated by filtration and heated at 400 °C for 120 min. The resulting white powder was ground and sieved (100 mesh) to give SO42-/In2O3-ATP,designated as SIA hereafter. Two alternative preparations of the catalyst were undertaken in the absence of indium and omitting the calcination step are signified as SA and n-SIA,respectively.
A scanning electron microscope (SEM,SU8010; Hitachi) and transmission electron microscope (TEM,HT7700; Hitachi) were used to identify the morphology of the catalysts. X-ray photoelectron spectroscopy (XPS) was performed using an Axis Ultra spectrometer (Kratos Analytical,UK) with a monochromated Al Kα source. The X-ray diffraction (XRD) patterns of the catalysts were collected using a Bruker D8 Advance X-ray diffractometer (Cu Kα1 irradiation). The pore structures were characterized on a Micromeritics ASAP 2010 gas adsorption analyzer by N2 adsorption-desorption isotherms at liquid N2 temperature. Samples were degassed at 200 °C for 2 h under vacuum before measurements were taken. The average pore diameter was calculated using the Barrett-Joyner-Halenda method based on the desorption isotherm,and surface area was calculated by using the Brunauer-Emmett-Teller adsorption isotherm method.
The Brönsted and Lewis acid sites of the samples were determined using a TENSOR27 FT-IR spectrometer (Bruker) in a quartz cell equipped with CaF2 windows. Samples (15 mg) were pressed into self-supporting wafers and then placed in the cell. Each sample was preheated at 400 °C under a vacuum of 1×10-2 Pa for 2 h to remove water,followed by pyridine (Py) vapor adsorption for 10 min at room temperature. Pyridine was removed by degassing under vacuum for 30 min at specific desorption temperatures (150 and 400 °C) prior to recording the FT-IR spectra.
A glucose or fructose solution (30 g,2 wt%) was added into an autoclave (50 mL) at a range of temperatures for set periods of time. An acid concentration of 10 wt% SIA was used for aqueous systems. Biphasic systems consisted of GVL and an aqueous suspension containing the catalysts. At the end of the reaction,aqueous samples were collected,filtered through a 0.2-mm filter and analyzed by HPLC (Aglient 1260 system with a Bio-Rad Aminex HPX-87H column and an RI detector). The organic layer of the biphasic reactions was diluted (0.1 mL diluted in 1 mL of distilled water) and then filtered through a 0.2-mm filter prior to HPLC analysis. The conversion of the hexoses (X),yield of the products (Y) and selectivity for the products (S) were evaluated on a carbon basis using the following equations: X = (moles of glucose remaining/moles of starting amount of glucose) × 100%,Y = (moles of HMF or LA/moles of starting amount of glucose) × 100%,S = (Y/X) × 100%.
The SEM images of ATP,SA,n-SIA,and SIA are shown in Fig. 1. The catalysts were found to form fibers with an average length of about 60 nm. Regular,smooth surfaces were present on n-SIA and ATP crystals,whereas tight,rough surfaces were formed on SA and SIA during the calcination process at 550 °C. The surface of SIA was also covered with many smaller particles,which are likely present because of the immersion in H2SO4 solution and the introduction of indium [15]. The catalysts were next examined by TEM. Bundles of multiple ATP fibers were obtained after calcination at high temperature (Fig. 2(b) and (d)),which could result from aggregation of partial metal oxide particles [24].
FT-IR spectroscopy of the catalyst with chemisorbed pyridine (Py-IR) was used to distinguish the Brönsted and Lewis acid sites on the samples. Fig. 3 displays the Py-IR spectra of SA and SIA between 1300 and 1700 cm-1. The peaks at 1447 and 1449 cm-1 are assigned to pyridine adsorbed to Lewis acid sites,the peak at 1543 cm-1 characterizes pyridine adsorbed to Brönsted acid sites,while the peak at 1490 cm-1 indicates pyridine adsorbed to both Brönsted and Lewis acid sites simultaneously. The Brönsted and Lewis acid sites on SO42-/In2O3-ATP catalyze glucose isomerization and fructose dehydration,respectively. The Py-IR spectra show that the number of acid sites on SIA was much larger than that of the ATP carrier. Fig. 4 shows the XPS measurements for SA and SIA,and the magnified signals attributed to S 2p and In 3d. Both SA and SIA clearly exhibited Mg,Al and Si signals,indicating the presence of ATP. SIA also exhibited signals from In 3d,confirming the successful grafting of In(Ⅲ) onto the ATP particles. The XRD pattern of ATP shows typical peaks for pure ATP that appeared at 2θ = 20.0° and 35.2° (Fig. 5). The peaks of SIA are narrower and less intense than pure ATP,with new peaks present. However,a peak assignable to crystalline In2O3 could not be clearly seen in the SIA spectrum. This might be because In2O3 nanoparticles were highly dispersed on the surface of the ATP carrier and the diffraction peak intensity of In2O3 was very weak because of interference from ATP.
The conversion of glucose to HMF by the catalysts was trialed with various solvent systems (Table 1). Using water as the solvent afforded a relative low HMF yield (Table 1,entry 1). A catalytic activity test was performed for ATP,SA,n-SIA and SIA in a solvent system consisting of GVL and 10 wt% of water to establish the most appropriate catalyst (Table 1,entries 2-5). Unadulterated ATP had little catalytic activity,resulting in a poor yield of HMF (10.5%). The sulfated ATP,SA,showed higher activity for glucose dehydration with a 17.7% HMF yield. The yield of HMF was 18.4% and 40.2% in the presence of n-SIA and SIA,respectively. This is because the introduction of In(Ⅲ) to the ATP carrier provides more acidic sites than pure ATP. The catalytic activity of SIA is much higher than that of n-SIA,which is attributed to an elevation of the surface area (from 31.2 to 48.3 m2/g) and pore volume (from 0.103 to 0.189 cm3/g).
The production of HMF from glucose in different systems using homogeneous Lewis acid InCl3 as the catalyst was studied (Table 1,entries 6-8). An aqueous medium afforded a relatively low HMF yield. When a biphasic solvent system was used,glucose dehydration took place in the aqueous layer followed by extraction of HMF to the organic layer. The isolation of the HMF in the organic layer protects HMF from catalyst contact and minimized side reactions [27]. Saturating the aqueous layer with NaCl diminishes the solubility of both HMF and the organic solvent,improving the organic extraction efficiency [28]. Both the organic solvents used,THF and GVL,were completely miscible with water alone and formed a single-phase mixture before reaction. The use of miscible solvents is preferable because it lowers the chance of mixing problems that may be encountered when scaling up a process [23]. A higher yield of HMF was obtained in the heterogeneous SIA/GVL:H2O system than with a biphasic InCl3/NaCl-GVL:H2O system. A potential advantage of the SIA/GVL:H2O system is that the addition of salt was not necessary,allowing the homogeneous catalyst (InCl3) to be replaced by a solid acid catalyst (SIA).
The effects of reaction temperature,reaction time and acid concentration on the conversion of glucose to HMF were studied (Fig. 6). The maximum HMF yield of 40.2% was obtained at 180 °C for 1 h. Although the catalytic activity of SIA was enhanced with the increasing temperature,higher temperatures resulted in side reactions and lowered HMF production [29]. Furthermore,reacting for long periods at higher temperature promoted the decomposition of the HMF. HMF yield increased with the amount of catalyst used and reached a maximum value at 10 wt% of SIA. We concluded that SIA could be used as a potential catalyst for the preparation of HMF from glucose.
Given that SIA rendered the solution acidic and catalyzed glucose conversion,fructose dehydration was investigated to provide insight into the catalytic activity of our system. The catalysts SIA and InCl3 were compared in a range of solvents for the conversion of fructose to HMF and LA (Fig. 7). The dehydration of fructose to HMF is faster than that of glucose,with near complete conversion in 60 min for all the systems trialed. The maximum HMF yield from fructose was higher (46.2%) than that from glucose (40.2%). After HMF has been formed,it can undergo further reaction to form equimolar amounts of LA and formic acid. A common degradation product of the dehydration of hexoses is solid humins formed by the cross- polymerization of glucose and HMF. Humins were not detected in our system,which is possibly because GVL solubilizes humins,thereby minimizing deposition of carbon on the solid catalyst surface and facilitating the separation of the product [23].
We next studied the activity losses of recycled SIA. The catalyst was separated after the dehydration reaction,washed with anhydrous ethanol,and then dried at 80 °C for 12 h before use. HMF yield from glucose decreased notably with subsequent uses of SIA (Fig. 8). On the 4th cycle,the reaction yield recovered to 32.5% of HMF,but was still lower than that of first cycle. The decreased activity might be because of the loss of active groups during the regeneration process [30]. Moreover,soaking the catalyst in H2SO4 was beneficial and reclaimed some of the lost activity. An HMF yield of 35.1% was obtained after this treatment during the recycling steps. Further research will focus on the regeneration of the active groups to maintain the catalytic activity of SIA.
SO42-/In2O3-ATP was used as an excellent solid acid catalyst for the preparation of HMF from hexose. The Lewis and Brönsted acid sites on SO42-/In2O3-ATP can catalyze glucose isomerization and fructose dehydration. A 40.2% yield of HMF from glucose was obtained at 180 °C for 60 min in a miscible GVL:H2O system. The use of this system allowed replacement of a homogeneous catalyst with a solid acid catalyst for glucose isomerization.
2016,
Vol. 37 Issue (8): 1362-1368 
PDF (952 KB)
2016,
Vol. 37 Issue (8): 1362-1368 PDF (952 KB)