Lignocellulose materials are abundant, cheap, and renewable, so they are considered to be ideal raw materials for producing fuels, energy products, and chemicals to replace fossil-based products; this is important because of the increasing depletion of fossil resources and growing concerns about environmental issues [1, 2]. The conversion of lignocellulosic biomass to fuels and chemicals requires effective use of the C5 and C6 sugars in hemicelluloses and cellulose, respectively [3]. Much recent research has been devoted to converting cellulose to various platform molecules such as glucose, 5- hydroxymethylfurfural (HMF), and levulinic acid [4, 5, 6]. However, because of the complex hemicellulose structure, there are relatively few studies on hemicellulose conversion to fuels and chemicals. Hemicelluloses can be used as feedstocks to produce many important chemicals such as furfural, furfuryl alcohol, furan, and tetrahydrofuran. The long-term goal should therefore be focused on exploring promising routes for hemicellulose transformation into valued-added chemicals.
Furfural, which has a wide range of application in the oil-refining, plastics, pharmaceutical, and agrochemical industries, has been identified as one of the key “green” chemicals produced from pentose-based hemicelluloses, with an annual global output of 700 000 t [7, 8]. Furfural can be further transformed to furfuryl alcohol and other five-membered oxygen-containing heterocycles such as furan and tetrahydrofuran [9]. Furfural and its derivatives are also used to make jet- and diesel-fuel-range alkanes, as a gasoline blendstock, and to develop the next generation of biofuels and bioplastics [10].
Hemicellulose conversion to furfural involves acid-catalyzed hydrolysis of polysaccharides to the constituent monosaccharides, mainly d-xylose, and dehydration of xylose to furfural [11]. Mineral acids, particularly sulfuric acid, are used as catalysts in traditional processes for furfural production, and these processes are accompanied by serious equipment corrosion, acidic waste stream production, and high energy consumption. It is necessary to consider the environmental impacts and energy requirements as well as economic factors. New eco- friendly and efficient processes for furfural production that achieve the same yields under significantly milder conditions to minimize the carbon footprint are therefore urgently needed [12].
Inorganic salts, particularly metal chlorides, have been proven to have an important influence on the conversion of lignocellulosic biomass to furans [1] and on the dehydration of glucose and fructose to HMF [13] in ionic liquids under microwave irradiation. A single-step conversion of cellulose to HMF using a pair of metal chlorides (CuCl2 and CrCl2) as catalysts in an ionic liquid was developed by Su et al. [14]. Zr(O)Cl2/CrCl3 was found to be an effective catalyst for the conversion of cellulose and sugarcane bagasse to HMF and 5- ethoxymethyl- 2-furfural in N,N-dimethylacetamide (DMA)/LiCl solvent under microwave-assisted heating [15]. Liu et al. [16] reported that NH4Cl salts could promote the conversion of d-fructose to HMF in isopropanol in 68% yield. However, there have been relatively few studies of the conversion of xylose or xylan to furfural using inorganic salts alone as catalysts. Zhang et al. [1] showed that using CrCl3 as a catalyst in ionic liquids led to a furfural yield of 63% from xylan under microwave-assisted heating at ~200 °C. AlCl3 and mineral acids were used as catalysts for the conversion of xylan in an ionic liquid and resulted in a maximum furfural yield of 84.8% at 170 °C under microwave irradiation [7]. Although ionic liquids, which have low vapor pressure and good thermal stability, are the preferred solvents for furfural production, their high viscosity and costs limit their industrial application. The use of microwave irradiation and high temperatures increases the energy consumption. Biphasic systems have been used to increase the furfural yield and inhibit the formation of by-products because the organic phase can continually extract furfural from the aqueous phase, so the selectivity is improved. vom Stein et al. [17] reported a catalytic conversion of xylose using FeCl3/NaCl as catalysts in a biphasic system consisting of an aqueous phase and a 2-methyltetrahydrofuran phase, achieving a maximum furfural yield of 71% at 140 °C. AlCl3·6H2O has also been used as a catalyst for the conversion of xylose in an H2O/tetrahydrofuran system at 140 °C, giving 75% of furfural yield [18]. However, biphasic systems require large volumes of extracting solvents, making them less attractive and possibly uneconomical on an industrial scale [19].
Water, which is the most benign solvent and therefore ideal for green and sustainable chemistry, is a poor solvent for the dehydration of xylose because many by-products are formed during furfural production, leading to low selectivity. Dimethyl sulfoxide (DMSO) with low toxicity was used as a cosolvent to suppress undesired side reactions in the aqueous phase in the conversion of fructose to HMF [19, 20]. In this work, we developed an efficient and environmentally friendly process for the dehydration of xylose to furfural using inorganic salts as catalysts in a homogeneous H2O/DMSO system at low temperature. The effects of the types of inorganic salts and the reaction conditions on furfural yields were compared.
d-(+)-xylose (98%) was purchased from the Adamas Reagent Co., Ltd. and used without further purification. AlCl3·6H2O, MgCl2·6H2O, NH4NO3, and ZnCl2 were supplied by the Guangzhou Chemical Reagent Factory (Guangzhou, China). SnCl4·5H2O, KCl, and DMSO were obtained from the Lingfeng Chemical Reagent Co., Ltd. (Shanghai, China). CrCl3 and LiCl were purchased from the Tianjin Kermel Chemical Co., Ltd. (Tianjin, China). CrCl2 and LaCl3·H2O were obtained from the Aladdin Chemistry Co., Ltd. (Shanghai, China). FeCl3·6H2O, NH4Cl, (NH4)2SO4, and CaCl2 were supplied by the Chengdu Kelong Chemical Co., Ltd. (Chengdu, China). FeCl2·4H2O, CuCl2·2H2O, Fe2(SO4)3, Fe(NO3)3·9H2O, and Al(NO3)3·9H2O were purchased from the Tianjin Damao Chemical Co., Ltd. (Tianjin, China). Al2(SO4)3 was supplied by the Tianjin Fuchen Chemical Co., Ltd. (Tianjin, China). CoCl2 was obtained from the Tianjin Yongda Chemical Co., Ltd. (Tianjin, China). MnCl2·4H2O and NiCl2·6H2O were obtained from the Zhanyun Chemical Reagent Co., Ltd. (Shanghai, China). NaCl and Cr(NO3)3·9H2O were purchased from the Guangzhou Jinhuada Chemical Reagent Co., Ltd. (Guangzhou, China).
A typical experiment was carried out as follows. Xylose (0.5 g) and a certain amount of inorganic salt were poured into a thick-walled pressure vessel. After addition of H2O (2.5 mL) and DMSO (2.5 mL), the sealed vessel containing the reaction mixture was immersed in a preheated thermostated oil bath, with a stirring rate of 500 r/min. Xylose dehydration was performed at a preset temperature of 140 °C for 4 h. The reactor was then cooled to room temperature and the mixture was filtered using a 0.22 μm filter membrane for analysis. All experiments were replicated at least three times.
The products were examined using gas chromatography-mass spectrometry (GC-MS) and identified using the NIST 08 MS library (Agilent 7890/5795 spectrometer; capillary column: Agilent DB-5, 30 m x 0.25 mm x 0.25 μm; programmed oven temperature: 60 °C, held for 3 min, ramped to 280 °C at 8 °C/min, held for 15 min; kept at 280 °C in split mode with a split ratio of 1:1; He carrier gas). The furfural content was determined using a high-performance liquid chromatograph (Agilent 1100) with a diode array detector at 277 nm, and an Agilent Eclipse XDB-C18 column. The samples were injected with a 20 μL loop, the column oven temperature was 30 °C, and the mobile phase was acetonitrile/H2O (15/85, V/V) at a flow rate of 1 mL/min.
Xylose was determined using a GPC system (Waters 2414) with a refractive index detector and a Bio-Rad organic acid column (Aminex HPX-87H, 300 mm x 7.8 mm) with H2SO4 (5 mmol/L) at a flow rate of 0.5 mL/min as the mobile phase.
The xylose conversion and furfural yield were defined as follows:
Xylose conversion = (moles of xylose reacted)/(moles of initiate xylose) x100%
Furfural yield = (moles of furfural produced)/(moles of initiate xylose) x100%
Furfural selectivity = (moles of furfural produced)/(moles of xylose reacted) x100%
The effects of the anions in salt catalysts on the dehydration of xylose to furfural were determined by performing experiments using chlorides, sulfates, and nitrates of Al3+, Fe3+, and NH4+ as catalysts. As shown in Fig. 1, metal chlorides such as AlCl3 and FeCl3 had significantly higher catalytic performance than sulfates and nitrates for furfural production under the same reaction conditions. The maximum furfural yield of 40.6% with 40.7% of selectivity was achieved over AlCl3. These results show that chlorides contributed to the dehydration of xylose to furfural. In the 1950s, Dunlop et al. [21] found that chlorides were conducive to the separation of furfural during distillation in the conventional production process. Since then, more attention has been paid to the mechanism of furfural formation. It was confirmed that chlorides increased the reaction rate, in addition to enhancement of furfural separation by the salting-out effect [7]. Marcotullio et al. [22, 23] reported that halides acted as weak bases, assisting enolization via proton transfer; they postulated that Cl- ions promoted the formation of the 1,2-enediol from the acyclic form of xylose, and accelerated the first and second dehydration steps, and the final intramolecular dehydration and ring closure, leading to furfural. However, the use of NH4Cl as the halide facilitated furfural production from xylose to a certain degree (Fig. 1). It was also found that the catalytic performance of (NH4)2SO4 was higher than that of NH4Cl because of the significantly higher acidity of (NH4)2SO4, which improves xylose conversion.
Stronger Lewis acids showed higher activity in the dehydration of xylose to furfural. Some metal chlorides act as Lewis acids, and Cl- ions have been proven to enhance furfural formation from d-xylose [22]. This result is confirmed by Fig. 1. The influences of different chlorides on the dehydration of xylose are shown in Fig. 2. Various metal chlorides, i.e., CrCl3, CrCl2, FeCl3, SnCl4, AlCl3, LaCl3, CoCl2, NiCl2, CuCl2, ZnCl2, and NH4Cl, had positive impacts on xylose conversion compared with the non-catalytic process (Fig. 2). Among these metal chlorides, SnCl4, AlCl3, and CrCl3 exhibited superior catalytic activity in the dehydration of xylose at 140 °C; SnCl4 gave the highest catalytic performance, and achieved a maximum furfural yield of 48.9%, with 52.2% of selectivity. CrCl2 and FeCl3 catalysts gave furfural yields of 22.3% and 20.4%, respectively. NiCl2, LaCl3, CoCl2, CuCl2, ZnCl2, and NH4Cl catalysts resulted in low furfural yields, and it could that many other products were formed during the reaction. This showed that these inorganic salts had low selectivity for furfural, although the xylose conversion was higher than those in the absence of a catalyst. LiCl, NaCl, KCl, FeCl2, MgCl2, MnCl2, and CaCl2 were found to be inefficient catalysts for furfural formation (data not shown in Fig. 2). Different chlorides therefore have different effects on the dehydration of xylose to furfural, based on their Lewis acid strengths.
The main products in the conversion of xylose catalyzed by different inorganic salts were also investigated by GC-MS. The results showed that furfural was the main product with the catalysts and almost no product was found without catalyst, showing that H2O could not be ionized at low temperature to form the protons that enable the production of furfural from xylose. These results indicated that the inorganic salts investigated promoted furfural formation from d-xylose. Besides the main product furfural, there was one main by-product pyrrole-2-carboxaldehyde, which was found to be present when an inorganic salt (i.e., NH4Cl, (NH4)2SO4 or NH4NO3 ) containing NH4+ ions was added to this system, indicating reaction of the nitrogen atoms in the NH4+ ions with the intermediate products during the dehydration of xylose, forming new by-products.
As discussed above, SnCl4 gave better results than the other inorganic salts, and was thus used as a catalyst in the following work. Su et al. [14] found that mixtures of two metal chlorides, e.g., CrCl2/CuCl2, significantly improved the conversion of glucose. In this work, combinations of SnCl4 with other metal chlorides were used to investigate cooperation during homogeneous catalysis. CrCl3, AlCl3, LiCl, NaCl, and KCl were used as cocatalysts; their catalytic performance is shown in Fig. 3. When SnCl4 was used as the catalyst in the dehydration of xylose, the xylose conversion was greater than 98%, so only furfural yields are given. Clearly, the furfural yield obtained with SnCl4/LiCl as the catalyst was higher than those obtained with other pairs or SnCl4 only. It is surprising that the less apparent decrease and the much obvious decrease in furfural yield using SnCl4/CrCl3 and SnCl4/AlCl3 paired catalysts, comparing to the single SnCl4, although AlCl3 and CrCl3 are strong Lewis acids. This indicates that AlCl3 and CrCl3 did not enhance the catalytic activity when mixed with SnCl4, whereas catalysts obtained by mixing SnCl4 and monovalent chlorides such as LiCl, NaCl, and KCl had higher catalytic activity. Among these monovalent chlorides, LiCl gave the best result. The cooperation between SnCl4 and LiCl afforded a furfural yield of 51.4%. These results are consistent with that reported by Binder et al. [24], who found that LiCl and LiBr had high xylose conversion efficiency, giving 56% furfural from xylose with CrCl2 (6 mol %) in DMA containing 10 wt % LiBr.
A series of experiments using SnCl4/LiCl with different SnCl4 molar ratios (χSnCl4) were conducted and the results are shown in Fig. 4. The yields of furfural consistently increased with increasing SnCl4 molar fraction. When χSnCl4 reached 0.8, the maximum furfural yield (56.9%) was achieved, 8.0% higher than that obtained using the single SnCl4 catalyst.
Experiments were carried out over SnCl4/LiCl (χSnCl4 = 0.8) at 140 °C for 4 h to study the influence of the catalyst/xylose molar ratio on the furfural yield; the results are shown in Fig 5. With increasing reaction time, the yield of furfural first increased and then decreased. Larger amounts of insoluble solids were produced at longer time periods because the long residence time contributed to polymerization and oligomerization of furfural to give furanic resins; this decreased the pentose dehydration selectivity [25]. The results also showed that the catalyst amount had a significant effect on the furfural yield. For shorter time, the catalytic performance increased with increasing catalyst amount because of enhanced accessibility of the catalyst to xylose. A maximum furfural yield of 58.5% was obtained with a molar ratio of 1:1 at 140 °C for 1 h. However, as well as increasing the cost, increasing the catalyst loading caused another problem, i.e., it favored side reactions, producing larger amounts of undesired by-products (seen in the vessel). For practical application, a reasonable ratio was 0.5:1 and a 57.7% yield of furfural was obtained within 2 h.
The influence of the xylose contents on the production of furfural was also studied; the results are shown in Fig. 6. In all cases, the furfural yields increased with increasing reaction time. The furfural yield sharply increased within 2 h, indicating that most of the furfural was produced in a short time. For a solid to liquid ratio of 1:30, the furfural yield increased from 6.3% to 60.9% as the reaction time increased from 0.25 to 6 h. Moreover, the furfural yield decreased when the time was further increased to over 6 h, and the amount of insoluble solid residue increased as a result of formation of humins [26]. It was also found that with increasing xylose content, a shorter time was needed to reach the peak furfural yield, but the peak value decreased. This was probably because a higher content of xylose would increase the number of collisions between xylose and the catalyst, accelerating xylose conversion to furfural and leading to higher furfural yields. However, over a long period of time, the collisions also led to side reactions, decreasing the furfural yields. At a ratio of 1:20, the furfural yield (59.8%) was relatively high at 140 °C for 4 h. In view of the reaction rate of furfural production and energy consumption, a solid to liquid ratio of 1:20 was used in subsequent experiments.
Experiments were carried out at different temperatures to study the temperature dependence of the furfural yield; the results are shown in Fig. 7. With increasing reaction temperature, the furfural yield sharply increased and the peak values were reached in shorter time. When the reaction time was further prolonged, the furfural yield gradually decreased, indicating that higher temperatures provided sufficient energy to promote side reactions, leading to the formation of soluble degradation products and black insoluble solids. A similar conclusion was drawn by O’Neill et al. [27]; they found that the stability of furfural at high temperatures was low, and furfural quickly formed degradation products. The optimal result for furfural formation in this work was achieved at 130 °C within 6 h, giving a maximum furfural yield of 63.0%.
DMSO has been proven to suppress side-reactions effectively and improve the selectivity [20]. In this study, DMSO was used as a cosolvent. The effect of the DMSO content in the H2O/DMSO mixture system on furfural production was investigated and the results are shown in Fig. 8. A system containing H2O/DMSO = 7/3 (V/V) gave a furfural yield of 42.6%, by 25% higher than that obtained without DMSO, but by 20% less than that obtained using a volume ratio of 5:5. It was observed that larger amounts of undesired insoluble solids were formed in the system without DMSO. These results showed that DMSO played a vital role in promoting furfural formation and significantly depressed side reactions. The results are consistent with the reported one [28].
An efficient and environmentally friendly process was developed for the dehydration of xylose to furfural using inorganic salts as catalysts in a homogeneous H2O/DMSO system at low temperature. The results showed that SnCl4 had high catalytic activity for the reaction. A paired SnCl4/LiCl catalyst (χSnCl4 = 0.8) was found to give the best catalytic performance. The highest furfural yield of 63.0% was obtained at 130 °C for 6 h, with a catalyst/xylose molar ratio of 0.5 and a solid to liquid ratio of 1:20. The mechanism of xylose conversion to furfural using SnCl4/LiCl as catalysts will be investigated and microwave irradiation will be used to further improve the furfural yields.