Haian Xia*ab,
Siquan Xuab,
Hong Huab,
Jiahuan Anab and
Changzhi Li*c
aJiangsu Provincial Key Lab for the Chemistry and Utilization of Agro-forest Biomass, China. E-mail: haxia@dicp.ac.cn; Fax: +86-25-85428873; Tel: +86-25-85427635
bSchool of Chemical Engineering, Nanjing Forestry University, Nanjing 210037, China
cDalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China. E-mail: licz@dicp.ac.cn
First published on 3rd September 2018
5-hydroxymethylfurfural (HMF) is a very important versatile platform compound derived from renewable biomass. The functionalized molecule with an aldehyde group, a hydroxyl group and a furan ring provides great potential for the production of a wide variety of valuable chemicals. This review highlights the latest advances in the catalytic conversion of HMF into value-added chemicals by some important reactions including (1) aerobic oxidation of HMF into furan-based aldehydes and acids such as 5-hydroxymethyl-2-furancarboxylic acid (HMFCA), 2,5-diformylfuran (DFF), and furandicarboxylic acid (FDCA), (2) reductive amination of HMF to amine, (3) the synthesis of aromatics by Diels–alder reaction followed by a dehydration reaction, (4) catalytic reduction of HMF into 2,5-bis(hydroxymethyl)furan (BHMF), and 2,5-dimethyl furan (DMF), (5) catalytic oxidation of HMF into maleic anhydride, and some other important transformations. The review mainly focuses on the recent progress in bio-catalytic, electrocatalytic, and heterogeneous catalytic transformation of HMF into high value chemicals over the past few years. Moreover, an outlook is provided to highlight opportunities and challenges related to this hot research topic.
Amongst various valuable compounds derived from biomass, HMF is identified to be a top building block chemical.6 HMF possesses a very versatile chemical activity, making it possible to be further transformed into value-added chemicals including 2,5-furandicarboxylic acid (FDCA),7,8 2,5-dimethyl furan (DMF),9,10 levulinic acid11 and other chemicals through oxidation, hydrogenation, hydrolysis, etc. (Scheme 1).12–14 Great progresses have been made in the valorization of HMF into chemicals in recent years. For example, FDCA, a very important monomer, which has been commercially produced through a homogenous catalytic system by Avatium company.15 However, there remains a huge challenge for the production of FDCA from HMF by using heterogeneous catalyst due to the stability of the catalysts and low yield upon using high concentration of substrate HMF.15,16 For DMF, its high yield (>99%) is readily obtained by various heterogeneous catalysts, but the recyclability of the catalysts and expensive HMF cost limit the large scale commercial product.17 Although there are some challenges for the commercial production of HMF-based chemicals, HMF is still deemed as a key intermediate that bridges bio-based products and carbohydrate chemistry, efficient production and rational utilization of bio-based HMF is envisaged to be promising to achieve sustainable biorefineries. It is urgent to decrease the production cost of HMF by using cheap material and to overcome the bottleneck question for low yield of HMF, especially for the use of lignocellulose as starting material in order to realize large scale production of bio-chemicals derived from HMF.
Dehydration of monosaccharides and polysaccharides, which are widely distributed in biomass, is the main synthetic pathway of HMF.18 Various substrates are used for the production of HMF: hexoses themselves, oligosaccharides as well as cellulose19–21 and the raw biomass such as wheat straw (see in Scheme 2).22 It is generally accepted that the production of HMF from biomass must include the acid-catalyzed hydrolysis of biomass to produce hexose in the first step. In the second dehydration step, starts from ketohexose (e.g. fructose) is more efficient and more selective than from aldohexose (e.g. glucose), due that the structure of aldohexose is very stable and it enolyses in a very low degree, while the enolisation is a determining factor for the HMF formation from hexoses.23,24 Over the past few years, many excellent reviews about the production of HMF from carbohydrates have been documented.3,14,25,26
Efficient utilization of HMF into chemicals mainly proceeds through homogeneous-, heterogeneous-, bio-, and electrochemical catalysis techniques. These catalysis techniques have their own advantages and drawbacks, whereas heterogeneous- and bio-catalysis are more commonly used in the transformation of HMF into chemicals compared to other strategies. Besides bio-chemicals, the generation of bio-fuels with HMF as a starting materials is an very important utilization pathway, including jet fuel hydrocarbon,27–30 ethers,17,31 and so on. Many excellent work and reviews on the synthesis of bio-fuels has been extensively reported.3,14,17,32 Therefore, in this review, we only focuses on the transformation of HMF into bio-chemicals. To the best of our knowledge, no review focused on bio-catalytic, heterogeneous, and their hybrid catalytic conversion of HMF into chemicals has been reported although there are some excellent reviews about the conversion of HMF into high value chemicals only by a single chemo-33 or bio-catalysis.34 The present review summaries the latest advances in the production of chemicals from HMF regarding heterogeneous chemo-catalysis, bio-catalysis and their hybrid approaches. In addition, a prospect was provided to highlight the challenges and opportunities for the utilization of HMF as sustainable chemicals and fuels.
The catalytic conversion of HMF to FDCA is a complex oxidation process, which forms several intermediate products such as 5-hydroxymethyl-2-furancarboxylic acid (HMFCA), 2,5-diformylfuran (DFF), 5-formyl-2-furancarboxylic acid (FFCA), as illustrated in Scheme 3. A variety of catalysts, including enzyme, homogeneous metal salts, stoichiometric oxidants and heterogeneous metal or metal oxide have been explored in the oxidation of HMF to FDCA.36–39 The corresponding results are summarized in Table 1. Among them, noble metal catalysts such as Pd, Pt, Ru and Au based catalysts exhibit excellent catalytic activity under mild conditions.8,40,41 Davis et al. achieved 79% and 68% FDCA yields with HMF fully conversion at 22 °C using Au/TiO2 and Pt/C catalysts, respectively.40 In the work of Steinfeldt and co-workers, high FDCA yield of 90% was obtained using Pd/ZrO2/LaO2 as a catalyst in aqueous media.8
Entry | Catalyst | Base | T (°C)/t (h) | Solvent | HMF con. (%) | FDCA yield (%) | Ref. |
---|---|---|---|---|---|---|---|
1 | Au/TiO2 | NaOH | 22/22 | H2O | 100 | 79 | 40 |
2 | Pt/C | NaOH | 22/22 | H2O | 100 | 68 | 40 |
3 | Pd/ZrO2/LaO2 | NaOH | 90/4 | H2O | 100 | 91 | 8 |
4 | Au–Pd/AC | NaOH | 60/2 | H2O | 100 | 95 | 48 |
5 | Mn/Fe (3:1) mixed oxides | NaOH | 90/24 | H2O | 93 | 32 | 7 |
6 | Merrifield resin-Co-Py | — | 100/24 | CH3CN | 96 | 91 | 50 |
7 | NNC | K2CO3 | 80/48 | H2O | 100 | 80 | 50 |
8 | Pt/C–O–Mg | — | 11012 | H2O | 100 | 97 | 52 |
9 | Pt/CNT | — | 95/12 | H2O | 100 | 98 | 52 |
10 | Ru/CTF | — | 140/1 | H2O | 99 | 78 | 51 |
11 | Ru/MnCo2O4 | — | 12010 | H2O | 100 | 99 | 53 |
It has been revealed that Pt and Pd nanoparticles are more active for the oxidation of an alcohol side chain, whereas Au nanoparticle is more active for the oxidation of an aldehyde side chain.15,37,42,43 It should be noted that the oxidation performances toward HMFCA of these metal nanoparticles depend on the structure of the catalysts including the support type, the size of nanoparticles, the facet-effect, etc.15,38,40 In other words, the aerobic oxidation of HMF is structure-sensitive reaction. To investigate the facet effect of Pd nanocrystals on the aerobic oxidation of HMF, Liu et al. synthesized single-crystalline Pd nanooctahedrons and nanocubes enclosed by (111) or (100) facets with controlled size using polyvinylpyrrolidone (PVP) as a capping agent.44 They found that the size-dependent effect of these Pd nanocrystals could only be attributed to the different Pd dispersions. Interestingly, they also demonstrated that Pd nanooctahedrons enclosed by (111) facet showed remarkably enhanced catalytic activity compared to Pd nanocubes enclosed by (100) facets for the aerobic oxidation of HMF, which is attributed to the reason that Pd nanooctahedrons has a lower energy in the alcohol oxidation step from HMFCA to FFCA in comparison with Pd nanocubes.44
In comparison with the monometallic catalyst, the use of bimetallic catalysts is an interesting strategy as it is able to enhance the activity and the product yield by modifying the electronic structure and the synergic effect.41,45–47 Prati et al. reported that Pd-modified Au supported on active carbon (AC) can act as an efficient and durable catalyst for the aerobic oxidation of HMF to FDCA. A series of bimetallic Au–Pd/AC with different metal ratios achieved FDCA yield higher than 95% with 100% HMF conversion.48 Apparently, the results demonstrated that the bimetallic catalyst was beneficial to improve the activity, the yield of FDCA, and the stability of the catalyst.
Despite of these progresses, most of the above mentioned catalytic systems use expensive noble metal and some of them need excessive base, these drawbacks hamper their potential for practical application. In order to address the problems, non-noble metal catalysts and metal-free catalysts in the absence of base have been developed for the conversion of HMF to FDCA. For instance, Mn/Fe (3:1) mixed oxides,7 merrifield resin supported Co(II)–meso-tetra(4-pyridyl)-porphyrin (merrifield resin-Co-Py)49 and nitrogen-doped nanoporous carbon (NNC)50 all exhibited high activity in the oxidation of HMF. It is well known that base plays an important role in the production of FDCA from HMF, and hydroxide ions intend to promote O–H and C–H bond activation of the alcohol side chain of HMF and then add directly to aldehyde intermediates to eventually form acid products.40 The employment of numerous inherently solid base catalysts avoids the addition of homogeneous base. As an example, Palkovits et al. reported that Ru clusters supported on covalent triazine frameworks (Ru/CTF) afford 78% FDCA yield at 140 °C under 20 bar air without a homogeneous base.51 Pt/C–O–Mg catalyst gave the yield of FDCA up to 97% under the optimal reaction conditions.43 Moreover, without the addition of a base, FDCA yield of 98% was obtained over functionalized carbon nanotubes (CNTs) supported Pt nanoparticles (Pt/CNTs) at 95 °C for 14 h.52 Kim and co-workers also reported that the MnCo2O4 spinels supported Ru catalyst (Ru/MnCo2O4) was exploited to afford FDCA with an excellent yield of 99% in the absence of base.53 Very recently, Zhang et al. developed a novel CoPz/g-C3N4 photocatalyst for the selective oxidation of HMF into FDCA under simulated sunlight with molecular O2 in air.54 They proposed that 1O2 species are the active oxygen species responsible for the selective oxidation of HMF, as illustrated in Scheme 4.
Scheme 4 Possible mechanism of the photocatalytic oxidation of HMF into FDCA with the CoPz/g-C3N4 catalyst. |
Although much effort has made in the aerobic oxidation of HMF into valuable chemical, there still remain a lot of challenges such as the development of more efficient non-noble metal catalyst and more green synthesized strategy without the addition of inorganic base.
A strategy is the combination of enzymes (or with chemo-catalyst) in multistep processes to carry out tandem oxidations route into FDCA. Guajardo et al. reported the oxidation of DFF with immobilized TEMPO, followed by a lipase-mediated with the assistance of peracid for further oxidation of DFF to FDCA.34 Carnell et al. used a combination of galactose oxidase with xanthine oxidoreductase to afford FDCA in promising productivities of 18 g FDCA L−1 d−1.55
Apart from those multistep methods, the development of a single enzyme that catalyzes the triple oxidation of HMF into FDCA is more attractive. Fraaije et al. identified a HMF oxidase (HMFO) which is active for the oxidation of HMF and related compounds.56 It was found that HMF oxidase (HMFO) is capable of oxidizing 5-(hydroxymethyl)furan-2-yl-methanol into FDCA, and can also produce FDCA from HMF with high yield at ambient temperature and pressure. The underlying mechanism presents that HMF oxidase (HMFO) acts on alcohol groups only and depends on the hydration of aldehydes for the oxidation reaction to produce FDCA.
The results of HMF hydrogenolysis to produce DMF in recent years are summarized in Table 2. A variety of hydrogenolysis catalysts have been reported for the production of DMF from HMF, such as Pd, Pt and Ni based catalysts. In the work reported by Wang et al., a 98% of DMF yield was obtained in the 1-BuOH with PtCo@HCS catalyst at 180 °C.60 Huang et al. reported that Ni–W2C/AC was used as a catalyst to obtain 96% DMF yield in THF.61 Rauchfuss and co-workers used formic acid as hydrogen donor and Pd/C as the catalyst for the production of DMF from HMF, affording an excellent yield (>95%).62 However, to obtain high DMF yields, formic acid and sulfuric acid must be used simultaneously for these catalysts, which is not environmentally friendly. Subsequently, in the absence of formic acid, Chidambaram and Bell obtained a 32% DMF yield with 47% of HMF conversion over Pd/C in ionic liquids, which avoided the use of unfriendly additives.63 But the lower solubility of hydrogen in ionic liquids significantly restrained the conversion efficiency, which is a drawback of this method. Recently, the application of supercritical carbon dioxide–water combined with Pd/C improved mass transfer and turned into a green DMF production pathway, affording a DMF yield of 100%.64
Entry | Catalyst | Solvent | H2 source | T (°C)/t (h) | HMF con. (%) | DMF yield (%) | Ref. |
---|---|---|---|---|---|---|---|
1 | PtCo@HCS | 1-BuOH | H2 (10 bar) | 180/2 | 100 | 98 | 60 |
2 | Ni–W2C/AC | THF | H2 (10 bar) | 180/3 | 100 | 96 | 61 |
3 | Pd/C | THF | Formic acid | 150/2 | 100 | >95 | 62 |
4 | Pd/C | [EMIM]Cl-MeCN | H2 (62 bar) | 120/1 | 47 | 32 | 64 |
5 | Pd/C | ScCO2–H2O | H2 (10 bar) | 80/2 | 100 | 100 | 64 |
6 | CuRu/C | 1-BuOH | H2 (6.8 bar) | 220/– | – | 70 | 13 |
7 | Ru/Co3O4 | THF | H2 (7 bar) | 130/24 | 100 | 94 | 67 |
8 | Ru-HT | 2-Propanol | H2 (10 bar) | 220/4 | 100 | 58 | 69 |
9 | Cu–PMO | Sc methanol | MeOH | 260/3 | 100 | 48 | 69 |
Besides Pd and Pt-based catalysts, Cu and Ru-based catalysts also have been shown to preferentially catalyze HMF C–O bond hydrogenolysis that minimize aromatic hydrogenation.65 A pioneering work reported by Román-Leshkov et al. developed a catalytic process for the production of DMF from fructose, bimetallic CuRu/C catalyst provided 70% DMF.13 However, this CuRu/C catalyst showed poor tolerance to the presence of chloride ions. A similar catalytic system was reported using non-noble bimetallic Cu–Co catalysts supported on CeO2, ZrO2, and Al2O3 for the selective hydrogenolysis of HMF to DMF. High selectivity of 78% was obtained.66
The emergence of various stable and reusable catalysts which are not susceptible to chloride ions effectively addressed this challenge. Wang et al. prepared Ru/Co3O4 catalyst that maintained high activity over five cycles, resulting in up to 94% DMF yield under mild reaction conditions.67 Nagpure et al. used a reusable Ru doped hydrotalcite (HT) catalyst also obtained a modest yield of 58% for the formation of DMF.68 Hansen and co-workers employed a stable Cu-containing mixed metal oxide for the catalytic hydrogenation of HMF to achieve 48% DMF yield in supercritical methanol.69
Recently, the catalytic transfer hydrogenation of HMF by Meerwein–Ponndorf–Verley reduction has been developed and provided a better alternative for the hydro-upgrading of HMF.73–75 Formic acid, 2-propanol, and methanol are commonly used as in situ hydrogen sources for the hydrogenation of biomass-derived platform molecular. Lin et al. developed an efficient process for the catalytic hydrogenation of HMF to BHMF using ethanol as both hydrogen source and solvent on cheap ZrO(OH)2.73 A high BMHF selectivity up to 88.9% with a HMF conversion of 94.1% were achieved at 423 K in 2.5 h. The external H2-free process is cost-efficient due to the usage of non-noble metal in the absence of external H2.
Scheme 5 Detailed reaction network of DMF and ethylene at 528 K.87 |
Williams et al., reported that a 75% selectivity toward p-xylene has been obtained using H–Y zeolite as the catalyst for the cycloaddition of ethylene and DMF and subsequent dehydration with an aliphatic solvent at 300 °C.93 Kim et al. used mesoporous beta zeolite with a nanosponge-like morphology to catalyze the reaction of DMF with ethylene and produced p-xylene with a yield of 80%.94 Its appreciable yield is due to the fact that the external surfaces and internal micropores of the catalyst possess a large number of Brønsted acid sites. In addition to the traditional zeolite catalyst, other novel heterogeneous catalysts have also been applied to cyclization of DMF to p-xylene. Xin and Zhang reported an original route for the direct synthesis of p-xylene from 2,5-dimethylfuran catalyzed by scandium(III) triflate (Sc(OTf)3) in 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ionic liquid under mild conditions.95 In this process, the selectivity to p-xylene reached 63% with 90% DMF conversion. Feng et al. designed mesoporous aerosol with sulfonic acid groups (SiO2–SO3H) as the catalyst for highly selective production of p-xylene from DMF and investigated the effect of active site location on catalytic performance.96 Toste et al. have developed an innovative pathway to convert DMF and acrolein to p-xylene for bio-renewable PET production through these steps consisting of a Diels–Alder reaction, oxidation, dehydrative aromatization, and decarboxylation. However, the process uncovered in this work would certainly not be immediately feasible because of the low temperature conditions required in the Diels–Alder reaction step and the moderate yield of the aromatization step.90 Recently, Tsang et al. reported a novel catalytic transformation of biomass-derived furans and ethylene produced in situ by dehydration of bio-derived ethanol to aromatics over zeolite catalysts.97 They demonstrated that ethanol can act as a dienophile source for furan cycloaddition instead of gaseous ethylene, leads to remarkably higher reaction rate and higher selectivity into aromatics because of lower activation barriers. Obviously, the innovative process is completely renewably and is safer to operate compared to the use of ethanol. With similar strategy, Li and co-workers reported a highly atom-economic route for the continuous production of p-xylene from biomass-derived building blocks.85 This new process uses W2C/AC as a robust non-noble metal catalyst to mediate a cascade dehydroaromatization and hydrodeoxygenation reactions in the absence of external redox species, providing excellent p-xylene yield of 90%. Notably, the new process is readily applicable to the synthesis of various (multi)methylated benzenes from bio-based building blocks.84,85
In another scenario, a number of studies demonstrated successful conversion of HMF derivatives into other aromatics besides p-xylene.79,83,98–102 Davis et al. have developed a new alternative pathway to terephthalic acid (PTA) by the reaction of oxidized products of HMF and ethylene over solid Lewis acid catalyst.103 It was found that the partially oxidized HMF, 5-hydroxymethylfuroic acid (HMFCA) is reacted with high-pressure ethylene over Sn-beta to yield 4-hydroxymethylbenzoic acid with 31% selectivity at 61% HMFCA conversion after 6 h at 190 °C, as shown in Scheme 6.103 Moreover, no any reaction occurred between 2,5-furandicarboxylic acid (FDCA) and high-pressure ethylene even under high temperature (220 °C) and longer reaction times (14 h), showing that the strong deactivating effect of two carboxyl groups hampered the Diels–Alder reaction with ethylene.103
V-based catalysts are found to be the most common and effective for the formation of maleic anhydride until now. It has been reported that the process for the transformation of HMF into MA afforded a MA yield of 52% and 64% over VO(acac)2 and vanadium-containing heteropoly acid catalysts, respectively.111,113 However, from the practical point of view, readily separated and recovered heterogeneous catalytic systems can decrease the production cost and therefore are more suitable than homogenous catalytic systems. Yin et al. explored a catalytic aerobic oxidation of HMF into maleic anhydride and maleic acid with vanadium-substituted heteropolyacid as the catalyst.113 Under the optimal conditions, a combined yield of 64% for maleic anhydride and maleic acid was obtained. Moreover, the mechanistic studies excluded FDCA, DFF, HMFCA, and FFCA as the reaction intermediates in the pathway of HMF oxidation to maleic anhydride. They proposed a new mechanism that the oxidation of HMF is initiated by the C–C bond cleavage between the hydroxymethyl group and furan sketch of HMF by H5PV2Mo10O40 catalyst.113 Recently, Li et al. found that a series of V-containing oxides such as VOHPO4, (VO)2P2O7 and V2O5 showed relatively good catalytic activity (more than 75% yield of MA) for the selective oxidation of HMF.114 They suggested that the free radical reaction could be involved in the aerobic oxidation of HMF to MA according to the case that a free radical inhibitor 4-tert-butylphenol could significantly reduce HMF conversion and MA yield.114 The recycle experiments demonstrated that V2O5 was supported on fumed silica as a heterogeneous catalyst compared to bulk V2O5, and thereby exhibiting better recyclability. In comparison, bulk V2O5 leached fast during the reaction process.114
In view of poor catalyst recyclability of supported transition-metal oxide catalysts, Li et al. developed a facile metal catalyst-free system for the conversion of furfural into maleic acid with H2O2 as an oxidant and formic acid as the solvent, which obtained an unprecedented 95% MA yield.115 In addition, this strategy is also effective for the production of MA from other platform molecules such as HMF and HMF derivatives, as shown in Scheme 7. It can be seen that almost quantitative yield (99%) of MA was obtained from furan, and good to excellent yields of MA (77–91%) were also achieved if the furan ring was attached with aldehyde groups such as using HMF and DFF as the substrates. When the furan ring was substituted with carboxylic acid and methyl groups, lower MA yields were obtained, showing that carboxylic acid and methyl groups have a negative effect on MA production.
Scheme 7 Oxidation of various furan derivatives to MA with H2O2 as the oxidant in formic acid solvent. |
It is of great importance to develop continuous-flow gas phase reactions in chemical industry due to the advantages of product separation and catalyst reusability.115 However, the gas phase fixed-bed catalytic oxidation of furfural to MA received very little attention. V2O5/γ-Al2O3 catalyst was used in the gas oxidation of furfural into MA, but the catalyst was unavoidable deactivated by deposited of maleates and/or resin like product especially at low reaction temperature. At a given high temperature, an initial MA yield of 68% was achieved and the yield was still greater than 50% after 15 h on stream with the use of 1.0 vol% furfural and O2/furfural mole ratio equal to 10 at 593 K.116 Recently, Zhang et al. reported a plate vanadium phosphorus oxide as an efficient catalyst for gas phase oxidation of furfural to MA. A 90% yield of MA was achieved at furfural concentration of 10 vol% with air as an oxidant.117 Moreover, the catalyst exhibited good stability over the duration of 25 h of time-on-steam.
The mechanism of HMF aerobic oxidation to MA is still under debate, and some novel, robust and highly efficient catalyst need to be further explored in order to meet the large scale application. The alternative strategy with the use of renewable HMF provides a sustainable pathway to produce MA, which would alleviate the dependence on petroleum-based feedstocks.
The production of caprolactam (the monomer for nylon-6) or 1,6-hexanediol (monomer for high performance polyesters, polyurethane resins, and adhesives) in industry is currently a benzene-based process containing tedious seven steps. Recent elegant work by de Vries and co-workers proposed a relatively simple multistep route to caprolactam from HMF, proceeding with 2,5-bishydroxymethyl-tetrahydrofuran, 1,6-hexanediol and caprolactone as the key intermediates.118 It is valuable because this technology provided an overall selectivity of 86% caprolactam from HMF, meaning that the production of 1 kg of caprolactam would require 1.44 kg of HMF. This progress paves a new avenue for the transformation of renewable biomass resources to produce caprolactam, albeit the productivity of each step still needs to be further improved from the practical application viewpoint.
DMTHF is the deep reduction product of DMF, and it contains a higher energy content than DMF and may provide additional stability on storage over extended periods of time because of its fully hydrogenated furan ring.13 DMTF can also serve as an ideal substitute for tetrahydrofuran (THF) in chemical industry. DMTHF can be generated with HMF as an intermediate from carbohydrates, subjected to hydrolysis, dehydration and subsequently selective hydrogenation processes. Sen and co-workers provided a facile route of the one-pot conversion of carbohydrates to DMTHF.119 In their study, a homogenous catalytic system consisting of soluble Rh salt as a catalyst, HI and chlorobenzene as an additive, was employed in water under mild conditions. The highest DMTHF yields of 86% and 70% were achieved, with fructose and glucose as substrates in a homogenous system, respectively.119 Li et al. reported a novel process for one pot production of DMF and DMTHF from fructose by optimizing the synergic effect of Ru/C catalyst assisted by an ionic liquid. DMF and DMTHF can be produced by the dehydration of fructose followed by in situ hydrodeoxygenation of the resulting HMF on Ru/C catalyst with the use of ionic liquid and a biphasic [BMIM]Cl/THF solvent, and an optimal total target product yield of 67% was afforded.120 The above catalytic system of one-pot conversion of carbohydrates to furan based fuels features a viable process and it gives a high efficiency.
Dihydroxymethyltetrahydrofuran (DHMTHF) is commonly utilized as a precursor for the manufacture polyols and polymers or as a solvent for the conversion of carbohydrates.121,122 To produce DHMTHF, the aldehyde functional group and furan rings in the HMF should be completely hydrogenated without further hydrogenolysis. Thus, a desired catalyst should have suitable hydrogenation activity toward the aldehyde group and CC double bond of furan ring and avoids the breakage of C–O and C–C to obtain high yield of DHMTHF. In 2013, Tucker et al. studied that HMF was completely hydrogenated over the Pd/C catalysts, achieving 90% DHMTHF yield in the saturated 1-butanol with water.123 Compared to monometallic catalysts, bimetallic catalysts tend to exhibit higher activity for the completely hydrogenation of HMF due to the synergistic effect between metals. For instance, when a silica-supported bimetallic Pd and Ni catalyst (Pd–Ni/SiO2) was applied, DHMTHF yield was up to 96.0% in water at 40 °C for 2 h under 80 bar H2, which was higher than that over silica-supported palladium (Pd/SiO2) or silica supported nickle (Ni/SiO2). The excellent catalytic activity of Pd–Ni/SiO2 was ascribed to the synergy of Pd and Ni in promoting the hydrogenation of the aldehyde group and furan ring of HMF and suppressing the subsequent hydrogenolysis of DHMTHF. To simplify the preparation pathway of DHMTHF, Yang et al. proposed a novel catalytic system that performed a one-pot tandem process to obtain DHMTHF directly from fructose in a biphasic system.124 The combination of Amberlyst-15 and silica-supported ruthenium modified by trimethylchlorosilane (Ru/SiO2-TM) afforded a 34% yield of DHMTHF directly from fructose in the cyclohexane/H2O biphasic system.
Levulinic acid is not only widely used in the production of resins and biological active materials, but also can be employed as dyes and pesticides intermediates.125 Traditionally, levulinic acid can be produced by the rehydration reaction of HMF catalyzed by mineral acid. To date, some novel and efficient solid acid catalysts has been applied in the production of levulinic acid, such as zeolite, solid superacid, pillared clay and others.126,127 In the work reported by Ramli et al., Fe/HY catalyst exhibited the highest catalytic performance with 62% levulinic acid yield at 180 °C in 180 min.126 Dumesic and co-workers unfolded considerable results of LA production (>70%) with furfural alcohol as a starting material in monophasic THF–H2O medium over H-ZSM-5 catalyst.128 Following this research, Requies and co-workers presented the improved LA yield (77%) from furfural alcohol using ZSM-5 under an optimized reaction condition.129 Lourvanij and Rorrer studied a pillared clay as a solid acid catalyst for the dehydration of glucose with typical LA yields of 20%.130 Recently, Chen and co-workers reported the application of a solid superacid (S2O82−/ZrO2–SiO2–Sm2O3), which yielded 23% of LA from rice straw at high temperature.131
Although great advances have been achieved in the transformation of HMF into valuable chemicals, it is still a big challenge to realize the commercial large-scale production of high-value chemicals from HMF in biorefineries. Future research can be focused on the following aspects: (1) although the high yield of HMF can be achieved from edible materials including fructose and glucose, it still faces tremendous problems for the transformation of non-edible lignocellulosic materials such as straw and wood to HMF, especially the use of high concentration of lignocellulose as a starting material. Thus, robust catalytic reaction system is to be developed so as to achieve a large-scale and economic production of HMF. (2) Most of the existing transformation processes generally conduct with very low substrate concentration (<10%), which lead to large-volume solvents handling and therefore increasing separation costs. Much effort should be devoted to the conversion of the substrate with high concentration. (3) The chemo-catalytic conversion of HMF into valuable chemicals may afford an efficient route to produce chemicals and bio-fuels. However, the design and development of robust heterogeneous catalysts with high activity and stability, especially under harsh conditions, need to be further explored. The life and stability of the catalyst still need to be further improved in order to meet a large-scale production. In addition, avoiding the usage of environmentally unfriendly additives in the reaction system is a trend for a green, sustainable process. (4) The bio-catalytic conversion of HMF could offer an environmentally friendly, fossil-independent alternative production pathway to produce high value chemicals. However, the low productivity, low substrate concentration, and even the diluted and large volume fermentation broth could cause extensive energy consumption for product separation and concentration in the downstream processes. (4) It is promising necessary to develop some more efficient and novel processes by integrating the bio-catalysis, chemo-catalysis even photo- and electro- catalysis in order to address some technical challenges in the valorization of biomass into value-added chemicals.
This journal is © The Royal Society of Chemistry 2018 |