Recent advancements in the production of hydroxymethylfurfural

Abstract

Lignocellulosic materials can be utilized in the production of platform chemicals such as hydroxymethylfurfural (HMF). HMF production has been investigated in various aqueous, solvent, biphasic and ionic liquid systems. Aqueous medium usually suffers from a low HMF yield (usually 50 to 60% while using fructose as feedstock) due to the production of by-products and the decomposition of HMF. A higher HMF production was achieved by applying biphasic systems, however, these systems face some technical challenges including solvent recovery, process complexity and environmental issues, which prevent its practical implementations at industrial scales. The unique properties of ionic liquids (IL)s make them promising solvents for producing HMF from polysaccharides. In this review, the effects of various parameters such as catalysts, solvents, and process conditions on the production of HMF from various lignocellulosic feedstocks as well as systems developed for purifying HMF after production are discussed. Generally, the yield of HMF production in the IL systems was higher than 80% when fructose was used as the raw material, but was less than 50% when cellulose or other polysaccharides were used. However, the IL system is complicated and has a challenging recovery process. The proposed IL systems are also not environmentally friendly. The main emphasis of this paper is on the industrial applicability of proposed processes for producing HMF.

---

Mehdi Dashtban

Dr Mehdi Dashtban obtained his PhD from Lakehead University in 2012, and is presently a research associate at the University of Guelph in London, Ontario, Canada. His research areas include chemical, enzymatic and fungal modifications of biomass to produce biofuel and value-added chemicals such as xylitol.

Allan Gilbert

Dr Allan Gilbert is a professor of chemical Engineering department at Lakehead University in Thunder Bay, Ontario, Canada. He has more than 28 years of experience in pulp and paper as well as biorefinery processes. His main research area is to design novel processes for the chemical conversion of forest biomass to platform chemicals such as furfural and hydroxymethylfurfural.

Pedram Fatehi

Dr Pedram Fatehi is an assistant professor of chemical engineering department at Lakehead University in Thunder Bay, Ontario, Canada. His main research areas are biorefinery, colloid and surfaces as well as pulp and paper. He has developed various processes to produce value-added chemicals (e.g. furfural, hydroxymethylfurfural, dispersants and flocculants) and biofuel (e.g. ethanol) from wasted lignocelluloses of pulp and paper industry.

---

1. Introduction

The development of biorefinery technologies plays an important role in facilitating the production of value-added products. Different value-added materials are produced in a biorefinery, while cellulose fibers are the main products of traditional pulping processes.¹ Lignocellulosic materials are great resources for various biorefinery scenarios. Recent studies reported that synthetic materials (i.e. succinate, esters, levulinic acid, 2,5-diformyl furan) could be produced from lignocellulosic residues (wastes), which would in turn reduce environmental concerns associated with these wastes.^2–4 The production of nanocrystalline cellulose (NCC), phenolic compounds, alcohols, and furan-based chemicals were extensively discussed as promising value-added chemicals from biomass in different biorefinery processes.^5–7

Over the last few decades, the demand for energy has been significantly increased, while easily accessible fossil fuels are being depleted.^4,8 Biofuel production promoted new opportunities for renewable materials to be used for energy production. The first generation of biofuel uses sugar-based materials as the feedstock, which raised concerns for securing food. Lignocellulosic biomass is promising feedstock since it is abundant and sustainable for biofuel production, but it is independent of food industry.⁸

Hydroxymethylfurfural (5-hydroxymethyl-2-furfural or HMF) is a derivative of furfural with many applications in pharmaceutical and chemical industries (as a solvent) and petroleum industry (as fuel or fuel additives).^2,8 HMF (with the CAS number of 64-47-0) can be converted to a variety of other chemicals including aldehydes, alcohols and amines.⁹ It can be substituted for petroleum-based chemicals or be converted to other chemicals such as succinate, esters, levulinic acid, 2,5-diformylfuran, 5-hydroxymethylfuroic acid, 2,5-diaminomethylfuran, and 2,5-bishydroxymethylfuran. HMF can also be converted into dimethylfuran (DMFU), which has an octane rating of 119 and 40% higher energy than ethanol.^10–15 HMF can also be converted (through hydrogenation process) to HMF-based diols such as 2,5-bishydroxymethylfuran, which can subsequently be used as monomers for the production of polyesters^16,17 or to 2,5-diformylfuran (DFF) by oxidation, which can be used as a precursor of antifungal agents for pharmaceutical applications.¹⁸ The chemical conversion of hemicelluloses to value-added chemicals such as HMF, DMFU, levulinic acid (LA), levulinate (EL), succinic acid, formic acid and furfural were extensively studied in the past.^2,10,19

We previously discussed the possibilities for furfural production from pentoses of biomass using different approaches.³ In the past, the production of HMF from various feedstocks was discussed from chemistry point of view.^20–22 However, the impact of various methods on by-product production yield and kinetics, environmental aspects and industrial applicability of HMF production processes were not discussed in the past, but will be stated in this work. The aim of this review is to state the recent advancements and challenges in the production of HMF from biomass. The process conditions, kinetics and by-products of HMF production in aqueous, biphasic and ionic liquid solvents are discussed in details. Different factors including feedstock type, reaction media (i.e. aqueous, biphasic and ionic liquid), organic solvents (as extraction media), catalysts and reaction conditions (i.e. time, temperatures) affect the HMF production. More importantly, different processes proposed for purifying HMF after production, and their advantages and disadvantages are discussed in this work for the first time. This review criticizes the industrial applicability of various HMF processes proposed in literature and the challenges that should be addressed to commercialize HMF production from biomass.

2. HMF production in aqueous systems

Various processes can be developed for producing HMF from biomass. The production of HMF in aqueous systems is environmentally friendly.²³ In this regard, auto and acid catalytic, microwave assisted and solid catalytic/aqueous processes were developed for HMF production as discussed in next sections.

2.1. HMF production in autocatalytic systems

The treatment of biomass with steam or hot water at a high temperature (usually 200 °C or higher) causes biomass decomposition and some HMF formation. In this process, formic and acetic acids are formed via the hydrolysis of biomass and acts as catalysts for the HMF production.²⁴ Due to the fact that mineral acids are not used in the steam hydrolysis for HMF reactions, the aqueous wastes of the reaction are not very corrosive and toxic.²⁵

In one study, HMF production in autocatalytic (steam hydrolysis) system was practiced using a diluted glucose solution (1 wt%) at the temperature range of 150 and 250 °C for 2 h.²⁵ The maximum glucose conversion rate of 95% and HMF production yield of 10% were obtained at the temperature of 250 °C. In another study, HMF production was carried out using fructose (1 wt%) in hot compressed water under subcritical conditions of 240 °C and 3.35 MPa for 120 s,²⁶ which resulted in a fructose conversion rate, HMF selectivity and yield of 59.1%, 31.8% and 18.6%, respectively.

Overall, autocatalytic processes suffer from a low HMF production yield as well as from the formation of by-products (i.e. LA and formic acid) through the composition of HMF, which make their industrial implantation impractical. Additionally, applying a very high pressure in the HMF production may be technically challenging.

2.2. HMF production in aqueous acid catalytic systems

Different acids such as hydrochloric, sulfuric, phosphoric and formic acids were applied in the conversion of biomass into HMF in aqueous solutions.^27,28 In one study, the conversion of fructose (0.5 wt%) to HMF was performed at 210–270 °C, 4–15 MPa for 0.5–300 s in the presence of HCl as the catalyst of the reaction.²⁹ The maximum HMF yield of 64% and fructose conversion rate of 99% were achieved at 240 °C using 4 MPa after 10 s. In another study, the conversion of fructose (0.05 wt%) into HMF was assessed using different mineral (i.e. phosphoric, hydrochloric or sulfuric) and organic (i.e. oxalic, citric, maleic or p-toluenesulfonic) acids (0.1 M concentration) at 200–320 °C, 1.55–11.28 MPa, and reaction times of 75–180 s.²⁷ In the absence of any acid, a low HMF yield of 18% was obtained at 240 °C after 120 s. Among the acids used, phosphoric acid (pH 2.0) showed the highest HMF yield of 65.3%.²⁷ In another study, fructose conversion into HMF was carried out using formic or acetic acids at a high pressure (10 MPa) and temperatures ranging 180–220 °C from 10 to 100 min.²⁸ It was observed that a fructose conversion of 95% and HMF yield of 50% were obtained in the absence of any catalyst at 220 °C after 10 min of the reaction. Adding acetic acid (1 wt%) to the reaction promoted both fructose conversion (about 100%) and HMF yield (about 60%) under the same reaction conditions. The main by-product of the process was levulinic acid.²⁸

Overall, it is inferred that the HMF production yield was generally higher in acid catalytic systems (50–60% HMF yield) compared to autocatalytic systems (20–30% HMF yield). The application of strong mineral or organic acid promoted the HMF production in the homogenous systems. However, the main drawback of these systems is the environmental concerns associated with acidic wastes of the processes. Additionally, the separation, purification, and recovery of acids are costly and technically challenging.

2.3. HMF production in aqueous microwave assisted processes

The invention of microwave reactors promoted the applications of microwave in academia and industry in chemical transformations.^21,30 More recently, microwave energy has become attractive in chemical reactions due to its high heating efficiency and ease of operation. Efficient internal heating by microwave accomplishes through the direct coupling of microwave energy with molecules of solvents, reagents or catalysts, which can significantly affect the HMF production.³⁰

In one study, acidic solution (0.01 M, HCl) was used for HMF production from fructose under 2 MPa pressure with different radiation powers (90–300 W), reaction times (1–300 s) and temperatures (160–190 °C).³¹ A very low fructose conversion (5%) and HMF production (1%) were obtained in the absence of acid catalyst at 160 °C under microwave radiation (300 W for 5 min). This suggested that, similar to other aqueous systems, catalyst is a prerequisite for the selective conversion of fructose into HMF in water. The formation of acidic by-products (e.g. formic acid and LA) accelerates the conversion rate of fructose (about 65%), HMF selectivity (about 52%) and HMF yield (about 35%). In this study, the maximum fructose conversion rate of 95% was obtained at 200 °C after 60 s of the reaction. The results also showed that increasing the radiation power (from 90 to 300 W) did not affect the fructose conversion rate (remained about 95%). Overall, this process is promising because it is fast (needs less than 5 min), and a high fructose concentration (27 wt%) is achieved, while a low HCl concentration (0.01 M) is needed for the reaction.³¹ However, the HMF yield and selectivity are still low, which hinders its large scale applications.

2.4. HMF production using solid catalysis in aqueous systems

As an alternative procedure for heterogeneous acidic systems, solid catalysts was introduced and widely used in industry due to their ease of separation and recovery process.³² HMF production process can be implemented using solid catalysts in aqueous solutions. In this process, the solid catalysts can be readily filtered from the product suspensions after the reaction, which implies that the separation (and hence the recovery) of solid catalysts is easily exploitable at industrial scales.^12,32Table 1 lists the specifications of various aqueous systems studied in the literature and their maximum HMF yields.

Table 1 HMF production in aqueous systems

Raw materials	System	Process conditions	HMF/Furfural yields	By-products	Ref.
Glucose	Autocatalytic	150–250 °C, 2 h	95% (conversion)	Soluble polymers, LA and formic acid	25
Fructose	Autocatalytic	Subcritical, 240 °C, 3.35 MPa for 120 s	18.6%	Soluble polymers, LA and formic acid	26
Fructose	Aqueous acid catalytic	HCl, 483–543 K, 4–15 MPa, 0.5–300 s	99% (conversion)	N/A	29
Fructose	Aqueous acid catalytic	H3PO4,473–593 K, 1.55–11.28 MPa, 75–180 s	65.3%	Formic acid	27
Fructose	Aqueous acid catalytic	Acetic acid, 10 MPa),180–220 °C, 10–100 min	60%	LA	28
Fructose	Aqueous microwave assisted	HCl, 2 MPa, 90–300 W,1–300 s, 160–190 °C	35%	Humins, LA and formic acid	31
Fructose	Solid catalytic in aqueous systems	TiO2 and ZrO2, 200 °C, 1–10 min	38.1%	N/A	33
Sugarcane bagasse, rice husk and corncob	Solid catalytic in aqueous systems	TiO2 and ZrO2, 473–673 K, 34.5 MPa, 1–5 min	8.6% and 10.3%, respectively	Glucose, fructose, xylose and 1,6-anhydroglucose	12
Fructose	Solid catalytic in aqueous systems	Zirconium phosphate, subcritical, 240 °C, 3.35 MPa, 120 s	62.1%	Soluble polymers and furaldehyde	26

---

Different studies have been conducted to discover stable and recyclable solid catalysts for the conversion of saccharides into HMF. In one study, HMF production from fructose was studied in an aqueous system in the presence of TiO₂ and ZrO₂ under microwave-assisted reaction.³³ The reaction was carried out using a 2 wt% fructose solution at 200 °C for 1–10 min reaction time. A low fructose conversion rate of 13.2% was obtained in the absence of any solid catalyst at 200 °C after 5 min of reaction. The addition of TiO₂ led to a fructose conversion rate of 83.6% and HMF yield of 38.1% at 200 °C after 5 min of reaction. However, a lower fructose conversion rate of 65.3% and HMF yield of 30.6% were obtained using ZrO₂ as the solid catalyst,³³ which implies the importance of solid catalyst type in this reaction.

The HMF production from lignocellulosic biomass such as sugarcane bagasse, rice husk and corncob was carried out in the presence of different solid catalysts (TiO₂, ZrO₂ and mixed-oxide TiO₂–ZrO₂) under hot compressed water conditions at different reaction temperatures (200–400 °C) and 34.5 MPa for 1–5 min.¹² In this process, the hydrolysis of polysaccharides available in the lignocellulosic biomass to monomeric sugars carried out simultaneously with the conversion of monomers to two products of HMF and furfural. Among the different solid catalyst used in this study, mixed-oxide TiO₂–ZrO₂ (3 : 1 molar ratio) generated the highest HMF and furfural yields (8.6% and 10.3%) using corncob as the feedstock at 300 °C.¹² The main by-products of this process were glucose (3%), fructose (1–3%), xylose (2–3%), and 1,6-anhydroglucose (less than 1%).

HMF production from fructose was also evaluated in water under subcritical, supercritical or hot compressed conditions in the past. In one study, the conversion of fructose (1 wt%) into HMF was carried out in subcritical conditions of 240 °C and 3.35 MPa pressure in the presence of different zirconium phosphate acid catalysts.²⁶ In the absence of any catalyst, a low fructose conversion rate, HMF selectivity and yield of 59.1%, 31.8% and 18.6% were obtained after 120 s, respectively. The addition of zirconium phosphate as the solid catalyst resulted in a higher fructose conversion, HMF selectivity and yield of 80.9%, 62.1% and 50.2%, respectively. In this process, soluble polymers and 2-furaldehyde were identified as the major and minor by-products, respectively.

The main drawback of the solid catalytic systems is their relatively low HMF yield and selectivity as a result of HMF decomposition to by-products (i.e. LA, formic acid and humins). Also, exploiting a very high pressure and temperature may be technically impractical. To overcome these difficulties, different reaction systems and solvents were proposed as discussed in the following sections.

3. HMF production in biphasic systems

Biphasic systems contain aqueous and organic phases. In biphasic systems, HMF is produced in an aqueous phase, but will spontaneously be transferred to an organic phase (i.e. HMF is extracted). Therefore, the reaction medium (aqueous phase) has a limited concentration of HMF, which improves the overall HMF product yield and fructose conversion rate in the biphasic systems. The main advantage of a biphasic system is its high HMF yield, which is due to the limited side reactions occurring in the system or to the limited decomposition of HMF after formation.^3,34 The aqueous phase in this process is usually water or water–dimethyl sulfoxide (DMSO) mixture with an acid (HCl, H₂SO₄ or formic acid) as the catalyst. On the other hand, the organic phase has a great affinity for absorbing HMF and may consist of different organic solvents such as methylisobutylketone (MIBK), MIBK-2-butanol mixture or tetrahydrofuran (THF).³⁴ HMF was produced from fructose and glucose using a biphasic system containing a mixture of MIBK-2-butanol (7 : 3 w/w) at temperature of 170 °C for 5–100 min in the past.³⁵ The results showed a low HMF selectivity and glucose conversion rate (11% and 20%, respectively) in water at pH 1. The application of biphasic system (water/MIBK-2-butanol, 7 : 3 w/w) alone, however, did not improve the HMF selectivity (28%). Alternatively, DMSO was used as solvent, and HMF selectivity and glucose conversion rate were improved to 26% and 41%, respectively. Interestingly, the water–DMSO/MIBK-2-butanol system showed 53% HMF selectivity, 43% glucose conversion rate and 74% HMF yield. Consequently, it is implied that DMSO was indeed crucial to improve the HMF production yield.³⁵ The main difficulty associated with this method is that the reaction requires a large volume of extracting solvent that causes environmental concerns and makes it unattractive for a large scale industrial application. Also, the application of extractive solvent (i.e. DMSO) with a high-boiling point (189 °C) makes its distillation uneconomical.

Alternatively, 1-butanol with a lower boiling point (117.4 °C) was introduced as the organic phase in a water-solvent biphasic system to reduce the distillation cost.²³ In one study, water/butanol (1 : 3 ratio) was used as the reaction medium for the conversion of fructose (with a low concentration of 2 wt%) in the presence of NaCl (25 wt%) at 170 °C for 70 min.²³ The results showed a fructose conversion rate of 70.8% after 6 h of reaction at 170 °C. Adding the formic acid to this biphasic system (concentration of 2.5 mol L⁻¹) shortened the reaction time to 70 min with 98.3% fructose conversion and 69.2% HMF production yield.

Acetone-water reaction medium (ratio of 90 : 10) was used for HMF production under sub- and super-critical conditions.^36,37 The reaction was conducted using 1 wt% fructose concentration at different reaction conditions of 180 to 300 °C, 5 to 30 MPa and different concentrations of sulfuric acid (3 to 50 mmol L⁻¹) for 50 to 600 s. The results showed that the optimum conditions for the highest fructose conversion (99%) and HMF selectivity (77%) were acetone–water medium ratio of 90 : 10 vol%, 20 mmol L⁻¹ sulfuric acid at 180 °C for 120 s at 20 MPa.^36,37 Although the process seemed to produce a high HMF selectivity and fructose conversion (77% and 99%, respectively), the system suffered from a high pressure, low initial fructose load, and the need for high percentage of acetone and sulfuric acid.

In another study, acetone-water reaction medium (0 to 85 wt% acetone in water) and cationic ion exchange resin catalyst (containing sulfonated groups) were used under microwave radiation for different reaction times (2.5–180 min) and different temperatures (100–180 °C) at 1.73 MPa.³⁸ The results showed the resin catalyst played a significant role in HMF production and fructose conversion rate of 95.1% and HMF yield of 73.4% after 15 min reaction time at 150 °C. In this study, the maximum HMF yield of 73.4% was obtained using a 70 wt% acetone-water mixture after 15 min at 150 °C. Also, the fructose conversion and HMF yield remained the same at about 98% and 72%, respectively, when resin was recycled 5 times.³⁸ However, the resin used as the catalyst can be potentially toxic at higher temperatures (>100 °C) due to the formation of reaction products such as aromatics, hydrocarbons, organic sulfonates and sulfur oxides. Furthermore, the main drawback of this system is the risk of high pressure due to the application of acetone at 150 °C.

The low fructose concentration (usually less than 10% wt) was generally applied in the aqueous systems for HMF production, which makes the process to be infeasible for large-scale applications. Improving the selectivity of fructose conversion in efficient biphasic systems with high concentrations of feedstock (up to 50% wt) were attempted.³⁹ In this study, fructose (10–50% wt) was converted into HMF in aqueous phase containing acid catalyst (HCl or acidic ion-exchange resin) and an organic phase (a combination of MIBK and 2-butanol) biphasic system.³⁹ Increasing the fructose concentrations from 30 to 50 wt% led to a decrease in the HMF selectivity from 80% to about 62%, under the same reaction conditions. This lower HMF selectivity was attributed to the formation of solid humin polymers. To achieve a higher HMF selectivity, the amount of MIBK was doubled, which increased the HMF selectivity to 77%.³⁹ The main drawback of this system is the complexity of the reactions in different solvents, which makes the process to be unfavorable for a large scale application. In another study, 30 wt% fructose concentration was taken into account in a biphasic system of water that is saturated with inorganic salts (i.e. NaCl, KCl, LiCl or MgCl₂), different organic solvents (primary alcohols such as 1-butanol, secondary alcohols such as 2-butanol, ketones such as acetone and cyclic ethers such as THF) and acid catalysts (HCl or H₂SO₄) at 150–200 °C for 8 to 35 min.³⁴ The results showed that the highest HMF selectivity of 85% was obtained using 2-butanol at 150 °C, pH 0.6 for 35 min. The highest fructose conversion (89%) and HMF yield (84%) were obtained using KCl as the inorganic salt under the aforementioned reaction conditions.³⁴Table 2 lists the specifications of various biphasic systems studied in the literature and their maximum HMF yields.

Table 2 HMF production in biphasic systems

Raw materials	System	Process conditions	HMF/Furfural yields	By-products	Ref.
Fructose, glucose	Biphasic	Water–DMSO/MIBK-2-butanol, 443 K, 5–100 min	74%	N/A	35
Fructose	Biphasic	Water–butanol, NaCl, 170 °C, 70 min	98.3% (conversion)	Formic acid	23
Fructose	Biphasic	Acetone–water, H2SO4, 180–300 °C, 50–600 s, 5–30 MPa	99% (conversion)	N/A	36, 37
Fructose	Biphasic microwave assisted	Cationic ion exchange resin catalyst, 2.5–180 min, 100–180 °C, 1.73 MPa	73.4%	N/A	38
Fructose	Biphasic	Water–HCl, MIBK and 2-butanol, 180–220 °C,	77% (selectivity)	N/A	39
Fructose	Biphasic	Water–NaCl, 1-butanol or 2-butanol, 423–473 K for 8 to 35 min	85% (selectivity)	N/A	34

---

Overall, the biphasic systems in combination with inorganic salt showed a higher HMF selectivity compared to aqueous systems. However, the reaction required mineral acid to have an efficient yield, which raises environmental concerns. Generally, the presence of acid, salt and two phases in one process is operationally challenging. Furthermore, solvent, acid and salt should be recovered to possess an economically viable process. The complexity in the separation of organic and inorganic phases makes the process complicated and expensive at industrial scales. Consequently, the commercialization of these biphasic systems is challenging and may not be technically feasible with the available technologies.

4. HMF production in ionic liquid (IL) systems

Ionic liquids (IL)s are organic salts that consist of organic cations and either inorganic or organic anions.⁴⁰ ILs are usually liquid at a relatively low temperature (below 100 °C) and possess a high thermal and chemical stability and low volatility, but they have a high ionic conductivity and recyclability.^30,40 These unique properties make ILs suitable candidates as reaction media,⁴⁰ and are promising in addressing the difficulties associated with the solubility of cellulosic materials.⁴⁰ For instance, cellulose was reported to be soluble in 1-butyl-3-methyl- or 1-allyl-3-methylimidazaolium chloride ([C₄MIM][Cl] and [AMIM][Cl]), 1-ethyl-3-methylimidazolium acetate ([C₂MIM][Ac]), 1,3-dialkylimidazolium formates, and 1-ethyl-3-methylimidazolium phosphate via developing hydrogen bonding between anions of ILs and its hydroxyl groups.^40,41 The production of HMF in ionic liquids (IL)s has obtained great attentions these days.^42–44 Different raw materials including mono-, di-, polysaccharides and lignocellulosic residues, IL solvents, co-solvents (i.e. such as toluene) and catalysts (e.g. chromium) were used for HMF production in IL systems.^42,44

4.1. Isotherm analysis of IL systems

The effect of reaction temperature on the HMF formation was assessed in different studies.^30,45 In one study, HMF production from cellulose was investigated using metal chlorides (i.e. CrCl₃, CrCl₃/LiCl) and 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]) as the IL system under microwave radiation at different temperatures ranging from 140 to 170 °C for 10 to 50 min.³⁰ The results indicated that, by increasing the temperature of the reaction from 140 to 160 °C, the HMF yield increased from 48% to 62% after 10 min, respectively, and the highest HMF yield of 62.3% was obtained using CrCl₃/LiCl (CrCl₃, 50% mol; LiCl, 50% mol) catalysts at 160 °C for 10 min.³⁰

4.2. Impact of reaction medium on HMF production in IL systems

Several studies focused on identifying an appropriate reaction medium for the IL systems that produces HMF with a high yield and selectivity. The reaction conditions of hexose conversion to HMF in an IL medium are significantly affected by the type of solvent and catalyst. In the past, 1-butyl 3-methylimidazolium tetrafluoroborate [BMIM][BF₄] (a hydrophilic IL) and amberlyst-15 (sulfonic ion-exchanged resin as a catalyst) were used as the reaction medium for HMF production from fructose,⁴⁴ which led to 50% HMF yield at 80 °C after 3 h. Alternatively, to enhance the solubility of fructose, dimethylsulfoxide (DMSO) was added to the solvent (5 : 3 IL/DMSO ratio), and the results showed that combined [BMIM][BF₄] and DMSO (without catalyst) resulted in an HMF formation with the yield of 36% after 32 h. Furthermore, the yield was increased to 87% by adding 143 mg of the catalyst to the [BMIM][BF₄]/DMSO medium under otherwise the same reaction conditions (80 °C, 32 h). By changing the solvent to 1-butyl 3-methylimidazolium hexafluorophosphate ([BMIM][PF₆], a hydrophobic IL), an HMF yield of 80% was obtained after 24 h.⁴⁴

Some of the solvents used in the IL systems may also function as catalysts that promote the HMF production. For example, HMF production from fructose was carried out in the presence of 1-H-3-methylimidazolium chloride ([HMIM][Cl]) in the absence of any additional catalyst.⁴³ The results showed that a maximum HMF production of 92% was obtained at 90 °C after 45 min. This high HMF yield indicated that HMF decomposition in the reaction was limited.

The application of mineral acid in IL systems for HMF production has also been investigated. In one study, the conversion of fructose and glucose to HMF was studied using 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]) at a temperature range of 100–140 °C for 50–120 min.⁴⁶ In the absence of H₂SO₄ as catalyst, a maximum HMF production of 63.1% was achieved at 120 °C for 50 min. The addition of sulphuric acid (10 mol%) increased the HMF yield to 82.9% (100% fructose conversion rate) at 100 °C for 50 min. However, when CrCl₃ (10 mol%) was replaced for sulfuric acid, the HMF yield was 75.3% at 100 °C for 120 min.⁴⁶

Efforts have been made on identifying more environmentally friendly processes for producing HMF. In one study, non-metal system was replaced for chromium chloride as a catalyst.⁴⁷ The reaction was carried out in alkylmethylimidazolium chlorides and boric acid (0.44 mmol) at 120 °C for 3 h. Maximum HMF yields of 42%, 66% and 80% were obtained using glucose, sucrose or fructose under these reaction conditions, respectively, and the highest HMF selectivity of 50% was obtained using glucose. The results also showed that increasing the boric acid concentration to 1.96 mmol decreased the HMF yield to less than 10%. The formation of stronger fructose–borate chelate complexes, as by-products, was suggested to be the main reason for the lower HMF yield at higher boric acid concentrations.⁴⁷Table 3 lists the specifications of various IL systems studied in the literature and their maximum HMF yields.

Table 3 HMF production in IL systems

Raw materials	ILs	Process conditions	HMF/furfural yields	By-products	Ref.
Fructose	[NMM][CH3SO3]	DMF-LiBr, 90 °C, 2 h	74.8%	N/A	48
Fructose	[BMIM][BF4]	Amberlyst-15, DMSO, 80 °C, 32 h	87%	N/A	44
Fructose	[HMIM][Cl]	90 °C, 45 min	92%	N/A	43
Glucose	[EMIM][Cl]	CrCl3 (THF)3, 100 °C, 3 h	71%	LA, humins and formic acid	45
Glucose	[BMIM][Cl]	YbCl3, 160 °C, 30 min	24%	N/A	49
Fructose	[BMIM][Cl]	H2SO4, 100 °C, 50 min	82.9%	N/A	46
Fructose	[EMIM][Cl]	Boric acid, 120 °C, 3 h	80%	Fructose–borate chelate complexes	47
Cellulose	[BMIM][Cl]	CrCl3, H2SO4, 100 °C	12%	N/A	50
Microcrystalline cellulose	IL-1	CoSO4, MIBK, 150 °C, 300 min	24%, 17%	Dimers of furan compounds	42
Microcrystalline cellulose	IL-1	MnCl2, MIBK, 180 °C, 300 min	37%, 18%	LA	51
Cellulose (cotton linters)	[BMIM][Cl]	Ipr-CrCl2, H-Y zeolite, 120 °C, 6 h	47.5%	N/A	52
Corn stover	[EMIM][Cl]	DMA/LiCl, CrCl2, HCl, 140 °C, 2 h	48%, 37%	N/A	9

---

4.3. The effect of biomass on HMF production in IL systems

Basically, fructose is more readily converted to HMF than other hexoses possibly through a rapid cyclic dehydration process while glucose is first isomerized to fructose (usually with a low isomerization yield) and then fructose will be converted to HMF.⁴⁷ In the literature, HMF was produced from fructose and sucrose in the presence of 1-methylimidazolium- and N-methylmorpholinium-based ILs ([NMM][CH₃SO₃]) and N,N-dimethylformamide-lithium bromide (DMF-LiBr) at 90 °C for 2 h.⁴⁸ The results showed that 74.8% and 47.5% HMF yields were obtained from fructose and sucrose, respectively.⁴⁸ In another study, chromium(III) (Cr III)-based catalysts and 1-ethyl-3-methylimidazolium chloride ([EMIM][Cl]) were applied for producing HMF from glucose at 80–100 °C for 1–6 h.⁴⁵ The results showed that the maximum HMF production of 71% was achieved with applying CrCl₃-tetrahydrofuran (CrCl₃ (THF)₃) catalyst after 3 h. However, the application of chromium is the main disadvantage of the proposed process as it is not environmentally friendly and toxic for human being.

The direct conversion of glucose to HMF was also evaluated in another study using different dialkylimidazolium chlorides-based IL systems and different lanthanide catalysts as the replacement of chromium-based catalyst commonly used in IL systems which raised environmental concerns.⁴⁶ The maximum HMF yield of 24% was obtained using 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]) as the IL system and ytterbium chloride (YbCl₃) as the catalyst at 160 °C after 30 min of reaction. In this system, the maximum HMF selectivity of 32% was obtained (using YbCl₃ catalyst, [BMIM][Cl] IL system) at 160 °C after 10 min. However, the drawback of the catalyst used in this study was the low HMF yield and selectivity compared to the higher HMF yield (over 70%) and selectivity (over 50%) obtained using CrCl₃-based catalysts.⁴⁹

Biomass contains mono-, di- and polysaccharides that can be used in the HMF production. However, such a production could be more complicated than that from fructose or glucose. Several studies were conducted on converting polysaccharides to HMF in IL systems. The simultaneous hydrolysis of polysaccharides to monosaccharides and conversion of monosaccharides to HMF was investigated in the presence of acidic 1-ethyl-3-methylimidazolium hydrogen sulfate ([EMIM][HSO₄]) at 100 °C for different time intervals.⁵⁰ The results showed that this IL system ([EMIM][HSO₄]) produced HMF with 88% yield using fructose with 100% conversation rate. However, the applied IL system was not efficient when glucose (9% yield) or cellulose (no detectable HMF) was used. Cellulose was also used as the feedstock in [BMIM][Cl]/CrCl₃ IL system, however, the IL system was inefficient (i.e. less than 1% yield). Interestingly, the addition of [EMIM][HSO₄] or H₂SO₄ to [BMIM][Cl]/CrCl₃ IL system using cellulose feedstock increased the HMF yield to 12% or 8%, respectively. This may be due to the hydrolysis effect of the acid in the IL system, as acid facilitates the hydrolysis of biomass.⁵⁰ However, the main drawback of the system was the low HMF production, which makes the process inefficient.

Microcrystalline cellulose (MCC) is regarded as purified cellulose derived from mainly wood/nonwood biomass with porous structures.⁵ The productions of NCC (nanocrystalline cellulose) and MCC have obtained great attention these days, and the conversion of MCC to HMF could indicate how purified cellulose produced from wood/nonwood biomass can be converted to HMF.⁵ In other words, it could be considered as initiative for converting wood/nonwood biomass to HMF. The production of HMF from MCC was studied using CoSO₄ in 1-(4-sulfobutyl)-3-methylimidazolium hydrogen sulfate (IL-1) as the IL system and 4-methyl-2-pentanone (MIBK) as the phase modifier to extract the formed products at 150 °C.⁴² The results showed that the addition of CoSO₄ (0.2 M) to the IL-1 catalyst increased the MCC conversion rates from 70% to 84%, that of HMF from 15% to 24% and that of furfural from 7.5% to 17% at 150 °C for 300 min.

HMF production from MCC was also studied using 1-(4-sulfobutyl)-3-methylimidazolium hydrogen sulfate (IL-1) and 4-methyl-2-pentanone (MIBK) as a phase modifier (for the product extraction) at different temperatures (100–200 °C) for 300 min.⁵¹ The results showed that the HMF yield (37%) was at maximum (with 88% MCC conversion rate) under the conditions of 150 °C after 300 min, but furfural and LA were also produced under these conditions (about 18% and 6%, respectively).

The conversion of cellulose (cotton linters) to HMF was assessed using 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]) as the IL system and ipr-CrCl₂ as a catalyst at 120 °C for 1 h.⁵² Additionally, zeolite (H-Y, CVB 400) with moderated acidity was added (100 mg) to the reaction in order to promote cellulose hydrolysis and to lower the decomposition process of HMF product. The results showed that the highest HMF yield of 36% was achieved after 6 h. The analysis showed that the optimum chromium catalyst concentration for the maximum HMF yield (47.5%) was 3–5 wt% at 120 °C for 6 h. It was reported that the optimized concentration of cellulose to achieve an HMF yield of 47.5% was 6 wt%.⁵²

4.4. Production of HMF from lignocellulosic residues in IL systems

In order to find an economical and environmentally friendly process for producing HMF, an efficient and direct conversion of lignocellulosic biomass to HMF is preferred. This should decrease the number of process steps (i.e. pretreatment), possible side reactions (i.e. formation of by-products), environmental issues (i.e. ionic solvent recovery) and ultimately the cost of HMF production. In this regard, lignocellulosic biomass was used as the feedstock for the production of HMF in IL systems. The HMF formation from cellulose of biomass is through a simultaneous saccharification of cellulose to glucose monomer, glucose isomerization to fructose, and fructose conversion to HMF. Untreated lignocellulosic biomass (10% cellulose or corn stover) was converted to HMF using N,N-dimethylacetamide (DMA) (252 mg), lithium chloride (LiCl) (26 mg), CrCl₂ (4 mg) and HCl (12 μmol), in [EMIM][Cl] (159 mg) IL system at 140 °C for 2 h.⁹ The results showed a maximum HMF and furfural yields of 48% and 37%, respectively, using untreated corn stover under these reaction conditions. For comparison, untreated corn stover used for HMF production under the same reaction conditions, but in the absence of HCl catalyst, resulted in 16% HMF yield. The proposed reaction benefited from the fast reaction time (2 h) and a low temperature (140 °C) for the direct conversion of lignocellulosic biomass to HMF. However, the process suffers from the toxicity of the applied chromium catalyst, which makes a barrier for the large-scale application of the proposed process.⁹

4.5. Microwave-assisted production of HMF in IL systems

Microwave heating became an excellent choice in the reactions where ILs solvents are used due to their high anionic/cationic content. This higher conductivity enables fast and efficient heating for the HMF production.³⁰ HMF production from cellulose was investigated using solid acids (i.e. H₃PW₁₂O₄₀, Nb₂O₅, Zr₃(PO₄)₂, Cs_2.5H_0.5PMo₁₂O₄₀, SO₄²⁻/TiO₂, SO₄²⁻/ZrO₂/SBA-15, NKC-9), metal chlorides catalysts (i.e. CrCl₃, CrCl₃/LiCl) and 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]) as the IL system under microwave radiation at 100 °C for 5 min in a research study.³⁰ Then, avicel cellulosic polymer was added to the mixture and heated at 105 °C under microwave radiation and stirred for 10 min in order to dissolve cellulose in ILs. The experiment was conducted at different temperatures ranging from 140 to 170 °C under the microwave radiation for 10 to 50 min. The highest HMF yield of 62.3% was obtained using CrCl₃/LiCl (CrCl₃, 50% mol; LiCl, 50% mol) catalysts at 160 °C for 10 min. Similarly, the production of HMF was investigated using lignocellulosic residues (untreated wheat straw) and cellulose using [BMIM][Cl] IL system under the same catalytic system (CrCl₃/LiCl) at 160 °C for 10 to 17.5 min. The maximum HMF and furfural yields of 61.4% and 43.8% were obtained after 15 and 12.5 min reaction times, respectively. Interestingly, the results of straw and cellulose were comparable under the same reaction conditions, which imply that the lignin of straw did not significantly affect the HMF production.

HMF production from lignocellulosic residues was also investigated under ILs-CrCl₃ catalytic system using microwave radiation. Different lignocellulosic residues such as corn stalk, rice straw and pine wood were considered for HMF production in 1-butyl-3-methylimidazole chloride ([C₄MIM][Cl]) as the IL system and CrCl_3.⁵³ The system was subjected to microwave radiation at 400 W for different time intervals (2 to 60 min). In order to investigate the efficiency of the reaction conditions, first pure cellulose was used under the reaction conditions (2 min, 400 W), which resulted in 62% HMF yield, but the HMF yield was 45% using corn stalk under the same catalytic reaction conditions (CrCl₃, 3 min, 400 W). The yield was slightly higher (47%) for rice straw and pine wood (52%) under the same reaction conditions. The results showed that the applied method was fast and efficient, which enabled the conversion of untreated lignocellulosic residues to HMF.^53,54

Lignocellulosic biomass (corncob, pine wood or grass) was converted to HMF and furfural using different solid catalysts (i.e. H₃PW₁₂O₄₀ and Amberlyst-5) in [BMIM][Cl] IL system at different temperatures and pressures under microwave radiation (240 W) for different reaction times (1–10 min).⁵¹ The maximum HMF yield of 11.8%, 17.9% and 22.6% were obtained using corncob, pine wood and grass under the same reaction conditions (160 °C, 10 min) using NKC-9 in [BMIM][Cl] IL system, respectively.

Similarly, the effect of metal chloride catalysts in the system of [BMIM][Cl] IL (described above) was investigated under microwave radiation.⁵⁵ The results showed that when lignocellulosic biomass (corncob, pine wood and grass) were used, the HMF production of 19.6%, 24.2% and 22.5% and furfural yield of 19.1%, 33.6% and 31.4% were obtained at 160 °C for 4 min using AlCl₃ as the catalyst (replacement of the solid catalysts), respectively. The HMF and furfural yields (using corncob) were slightly increased to 37.5% and 27.4% by adding a combination of CrCl₃/AlCl₃ metal catalyst instead of either CrCl₃ or AlCl₃ catalyst.

It is inferred from the aforementioned studies that the production of HMF in ionic liquids (IL)s has a high selectivity and yield, which is due to the superior solubility of biomass in IL systems.⁴⁰ However, the separation of HMF from IL media would be difficult and require large amounts of extracting solvents. Additionally, the possibility for recycling catalysts (especially mineral acids) and IL media needs a technological breakthrough to facilitate its commercialization.

5. Kinetic analysis of HMF formation

The kinetics of HMF production is crucial when the design of reactors for HMF production is considered. The kinetic analysis of HMF was studied for different systems including aqueous, biphasic and IL. The kinetic analysis of these systems is discussed separately in the following sections.

5.1. Kinetic analysis of HMF production in aqueous systems

Kinetic studies of fructose or glucose conversion to HMF in aqueous systems was performed in the past.^28,29,56 The production of HMF from fructose was studied in the presence of organic acid catalysts (i.e. formic and acetic acids) and at 180 °C for different time intervals (10 to 100 min).²⁸ The kinetic studies showed that the activation energies of fructose conversion to HMF were 126.8 ± 3.3 kJ mol⁻¹ in the absence of any catalyst and 112.0 ± 13.7 kJ mol⁻¹ or 125.6 ± 3.8 kJ mol⁻¹ in the presence of formic or acetic acids catalysts, respectively.²⁸ In another study, the conversion of fructose (0.5 wt%) to HMF was performed in subcritical water at 270 °C, 4 MPa, and residence times of 0.5–300 s in the presence of HCl. In this case, a first order reaction model for HMF production from fructose was proposed, and the activation energy was reported to be 160.6 (kJ mol⁻¹).²⁹

5.2. Kinetic analysis of HMF production in biphasic systems

The kinetic studies of fructose to HMF in biphasic systems was conducted in the past.³⁴ In the literature, fructose (30 wt%) feedstock was used in a biphasic system using an aqueous phase that contained NaCl salt (saturated in water), acid catalysts (HCl or H₂SO₄) and different organic extractive phases (1-butanol, 2-butanol and THF).³⁴ The reaction was carried out at 200 °C for different reaction times (8 to 35 min). The kinetic analysis showed that the conversion of fructose to HMF has the activation energy of 143 kJ mol⁻¹.

5.3. Kinetic analysis of HMF production in IL systems

The kinetic of HMF formation from different feedstock (i.e. fructose or glucose as monosaccharides and cellulose as polysaccharide) in IL systems were investigated in the past.^9,43,57 In the literature, first order reaction was taken into account for investigating the reaction kinetics of fructose to HMF in IL systems. The first order reaction is expressed in eqn (1):

d[HMF]/dt = k [fructose] = d [fructose]/dt (1)

where molar concentration of each chemical is shown in [] and the rate constant for fructose conversion at a certain temperature is represented by k.

In a study on the formation of HMF from fructose (2.3 mmol dissolved in N,N-dimethylacetamide (DMA)) in LiI (concentrations of 5.1 to 72.7 mg g⁻¹) and LiBF4 (0.55 mmol g⁻¹) mixture, a rate of HMF formation was first-order with respect to halide concentration.⁹ On the other hand, the kinetic of fructose conversion to HMF was studied in 1-H-3-methyl imidazolium chloride ([HMIM][Cl]) medium at 90 °C for different time intervals ranging from 0 to 45 min.⁴³ The activation energy of HMF formation (143 kJ mol⁻¹) was determined considering the first order reaction.⁴³

The kinetic of fructose conversion to HMF was also studied in a system of functionalized mesoporous silica nanoparticles imidazole-based ionic liquid (ILs) and CrCl₂ at 90 °C for 0–6 h.⁵⁷ The results showed that, in the absence of a catalyst, HMF production reached the maximum of 25.3% at 90 °C for 3 h. Interestingly, adding functionalized catalyst (1.06 mmol g⁻¹) improved the HMF production to 73.4%. In this respect, the presence of bi-functional mesoporous silica nanoparticles increased k from 0.55 to 3.93 at 90 °C for 3 h reaction time. The activation energy of the reaction in the absence or presence of the catalyst reported to be 80.05 or 67.5 kJ mol⁻¹, respectively. This result confirmed that the addition of catalyst lowered the activation energy and accelerated the fructose conversion leading to a higher reaction rate (73.4% HMF yield).⁵⁷

The effect of reaction time on HMF formation in IL systems was investigated in several studies. For example, HMF yield was shown to increase from 49% to 62% upon increasing the reaction time from 2 to 20 min in the presence of Zr(O)Cl₂/CrCl₃ catalyst and using cellulose fiber (4 wt%) as feedstock at 120 °C.⁵⁸ In another study on using chromium(III) (Cr III)-based catalysts and 1-ethyl-3-methylimidazolium chloride ([EMIM][Cl]), the HMF formation was also improved from 57% to 71% by increasing the reaction time from 1 to 3 h, respectively.⁴⁵

6. By-product analysis of HMF formation

As discussed earlier, by-products will also be formed during HMF production. The type and amounts of by-products depend on the reaction environments, and will significantly affect the purification and ultimately the cost of HMF production. Generally, by-products are generated by some side reactions of monosugars, or by the decomposition of HMF. In the following sections, the specifications of by-products will be discussed in details.

6.1. By-product analysis of HMF production in aqueous systems

Water is the most preferred solvent for reaction media as it is environmentally friendly and widely accessible.³¹ However, the selectivity and yield of HMF production is usually low in water. The low HMF yield in aqueous systems is mainly due to the subsequent reactions of formed HMF to LA and humins (insoluble polymers). Generally, the maximum HMF yield of 50–60 mol% was achieved using water aqueous system at a fructose conversion rate of 50–95%.^28,59,60

In the literature, the conversion of fructose (0.5 wt%) to HMF was performed in subcritical water at 210–270 °C, 4–15 MPa, and the residence time of 0.5–300 s in the presence of HCl as catalyst.²⁹ In this case, a first order reaction for HMF formation and decomposition was proposed, which included the conversion of fructose to HMF (k1), decomposition of fructose to 2-furfuraldehyde (2FA) (k2), the conversion of HMF to LA and formic acid (k3), the decomposition of LA (k4) and formic acid (k5), and the formation of soluble polymers from both fructose (k6) and HMF (k7). The activation energies for k₁ and k₂ reported to be 160.6 and 132.2 (kJ mol⁻¹), respectively. However, the average activation energies for the decomposition reaction with the rate constant of k₃, k₄ and k₅ reported to be 95.6, 113.3 and 104.0 (kJ mol⁻¹), respectively. These rates were determined to be 101.9 and 114.8 for soluble polymers as k₆ and k₇ formed from fructose and HMF, respectively.²⁹

However, the decomposition of HMF (0.1 M) in water under both subcritical (175–350 °C) and supercritical (400–450 °C) conditions (time of 80 to 400 s) showed that the HMF decomposition was reduced at lower temperatures (<250 °C).⁵⁶ However, increasing the reaction temperature (300 to 450 °C) or prolonging the time (80 to 400 s) resulted in a decrease in the amount of HMF residue and increased the amounts of by-products such as LA, formic acid, CO₂, CO and H₂. This study suggested that the liquid products (LA and formic acid) are initially formed as a result of HMF decomposition, and subsequently the liquid products act as intermediate compounds to form the gaseous products. This model seemed to follow Arrhenius equation and first order reaction as follows in eqn (2):

where the carbon content in HMF is presented as A and total organic carbon (excluding HMF) is presented as B and carbon gasification efficiency is presented as C. The rate constant for HMF decomposition and the rate constant for gasification are presented as k₁ and k₂, respectively. In this case, the temperature dependence of k₁ and k₂ is explained by the Arrhenius eqn (3):

k_1,2 = k₀e^−ΔE/RT (3)

where k₀ is pre-exponential factor (s⁻¹), ΔE is activation energy (J mol⁻¹), R is the gas constant (8.314 J mol⁻¹ K⁻¹), and T represents the temperature (K).

The activation energies for the decomposed liquid (k₁) and gas products (k₂) reported to be 75.76 and 109.58 J mol⁻¹, respectively. Thus, it is implied that the HMF decomposition rate (reaction (1), A → B) was higher than the gasification rate (reaction (2), B → C) indicating HMF decomposes to the liquid product faster than does liquid intermediate products to gaseous ones.⁵⁶

6.2. By-product analysis of HMF production in biphasic and IL systems

In the literature, fructose production was carried out in a biphasic system of water–butanol (1 : 3 ratios) in the presence of formic acid (concentration of 2.5 mol L⁻¹) at 170 °C for 70 min.²³ In this process, formic acid was produced as a by-product of HMF production, which accelerated the HMF production. The results showed a lower HMF selectivity of 50% in the absence of the organic solvent (1-butanol) due to the formation of insoluble by-products (i.e. humins).

Similarly, a part of fructose is converted to other by-products such as humins, and a part of the formed HMF may decompose to LA and formic acid in the IL reaction systems.^23,39,56 Basically, the quantity of each by-product is determined by different factors such as feedstock and process conditions (i.e. media, catalyst, time and temperature). For example, when microcrystalline cellulose (MCC) was used as the feedstock for HMF production using 1-(4-sulfobutyl)-3-methylimidazolium hydrogen sulfate (IL-1) and 4-methyl-2-pentanone (MIBK) as a phase modifier (150 °C for 300 min), furfural and LA were produced as the by-products (about 18% and 6%, respectively).⁵¹ Generally, the decomposition of HMF becomes dominant at a high temperature or prolonged reaction time. This is evident by considering a high selectivity in HMF production at the beginning of the reaction.^30,45 In one study, the maximum HMF yield (62%) was achieved in a short reaction time of 10 min in a system of metal chlorides catalysts (i.e. CrCl₃, CrCl₃/LiCl) and 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]) under microwave radiation.³⁰ Generally, IL systems prevent the conversion of formed HMF to LA, which is the main advantage of IL systems in stabilizing.³⁰ This was proposed to be due to the anhydrous conditions of the reaction in IL systems.⁴⁹ For example, no LA or formic acid were detected by applying 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]) as the IL system and ytterbium chloride (YbCl₃) as the catalyst at 160 °C after 30 min of reaction. However, a substantial amount of humins was detected in this study.⁴⁹

HMF decomposition into by-products was also studied in 1-H-3-methyl imidazolium chloride ([HMIM][Cl]) medium at 90 °C for different time intervals of 0 and 45 min. The results showed a 92% conversion of fructose to HMF with the activation energy of 143 kJ mol⁻¹, while the activation energy of HMF into by-products reported to be 69 kJ mol⁻¹.⁴³ In another study, fructose conversion to HMF was conducted using functionalized mesoporous silica and imidazole-based ionic liquid (ILs) at 90 °C for 0–6 h.⁶⁰ The formation of by-products was negligible, but increasing temperature to 140 °C resulted in 15.18% HMF degradation after 6 h of reaction.

7. Processes for purification of HMF

It was stated earlier that some by-products are generated during HMF production process. Various scenarios were reported in literature to separate by-products and improve the purity of HMF. These scenarios depend on reaction media as well as by-product types and compositions. The purification processes usually involves expensive operations such as evaporation and distillation. It may also contain a solvent treatment, which makes the HMF production complicated. The purification procedures developed for aqueous and IL systems are discussed in the following sections.

7.1. Processes for purification of HMF from aqueous systems

Strategies for separating HMF from other by-products, catalysts, and solvents depend on the reaction media. Different approaches were proposed for isolating HMF from aqueous systems including filtration, solvent extraction (i.e. using ethyl acetate) and vacuum distillation.^26,34

Fig. 1 shows a process developed for HMF extraction from a suspension medium. In this case, the solid catalyst was first separated from the reaction medium by centrifugation. The HMF was separated from the reaction medium by distillation. The separated catalyst was regenerated with (2 N) phosphoric acid solution.²⁶ The recyclability of a solid catalyst (different zirconium phosphate) used in the suspension systems of water and solid catalysts, and the extraction of HMF and by-products (i.e. soluble polymers and furaldehyde) from the reaction medium were assessed in a previous work.²⁶ The results showed that the regenerated solid catalyst could be reused over 6–7 runs without losing its catalytic activity and interestingly, fructose conversion and HMF selectivity and yield remained unchanged (59.1%, 31.8% and 18.6%, respectively).²⁶

Fig. 1 A process for HMF purification from aqueous medium of water and zirconium phosphate solid catalysts.²⁶

7.2. Processes for purification HMF from biphasic systems

The purification of HMF from biphasic systems mainly requires the isolation of aqueous phase from organic phase (i.e. by decantation) followed by distillation of organic phase to separate HMF from the organic phase.³⁴ In one study, the reaction medium (aqueous phase) containing formic acid catalyst was used for converting fructose to HMF, and the formed HMF was spontaneously transferred to the organic phase (1-butanol).²³ Subsequently, HMF was distilled to produce purified HMF. To reduce the distillation cost associated with the purification of HMF, organic solvent with a lower boiling point such as 1-butanol (117.4 °C) was considered in the past. Additionally, the formic acid used as the catalyst in the study could be recycled easily and reused due to its volatility and low-boiling point (100.8 °C).²³

7.3. Process for purification of HMF from IL systems

The products in IL systems usually contain IL, HMF, by-products, other organic solvents and catalysts. Despite minimal decomposition of HMF in IL systems, the separation of HMF from IL is the main difficulty of IL systems. Fig. 2 shows an overall view of HMF production from different monosaccharides (i.e. fructose and glucose) or disaccharide (i.e. sucrose) via using different IL solvents such as [BMIM][Cl], [EMIM][Cl] and [HMIM][Cl].^45,46,49 In this process, biomass is initially converted to HMF and other by-products. Then, the solid residues (i.e. catalysts or by-products) of the reaction are separated from the system by filtration. Subsequently, the produced HMF is extracted from the IL system using organic solvents such as diethyl ether.

Fig. 2 Overview of HMF production and purification in IL systems.

To have a feasible process, the catalysts, organic solvents and ILs should be recycled and reused. Various approaches were introduced in the literature to accommodate the purification and recycling of ILs and catalysts. Fig. 3 shows a process developed for purifying HMF from other products ([BMIM][Cl] and CrCl₃). In this process, the products were first filtered, and the filtrate was then mixed with an organic solvent (toluene or MIBK).⁵⁰ Then, the mixture containing the organic solvent (carrying HMF/furfural products) and ILs were separated by decantation. HMF and furfural were distilled to produce purified HMF and furfural. Subsequently, IL phase was treated with acetonitrile followed by centrifugation in order to separate solid residues. Finally, acetonitrile/water was removed from the IL solution by vacuum distillation (Fig. 3). Then, IL was recycled to the reaction system and acetonitrile is recycled for reuse.

Fig. 3 A process for HMF/furfural purification from an IL system using toluene or MIBK.⁵⁰

However, the main drawback of this process is the application of volatile toxic solvents (i.e. MIBK) in the reaction. Alternatively, the HMF extraction process can be replaced with a flash distillation (i.e. evaporation under vacuum) if the concentration of HMF is sufficiently high in organic phase so that distillation is cost-effective and applicable.³ This will eliminate the issue associated with the application of volatile solvents.

Fig. 4 shows a process developed for HMF extraction from imidazolium-based IL systems that contained boric acid as a catalyst. In this process, the suspension (the produced HMF in alkylmethylimidazolium chlorides as the IL system) is mixed with a salt solution such as NaHCO₃, which helps the precipitation of solid residues. Then, the system is filtered to separate the solid residues.⁴⁷ HMF is then extracted from the reaction medium with ethyl acetate, and then ethyl acetate was separated from HMF via applying vacuum (i.e. flash distillation). In one study using this system, HMF was separated from the system of [NMM][CH₃SO₃] (as the catalyst) in DMF–LiBr (as the IL), which resulted in HMF with 98% purity.⁴⁸ However, the HMF purification process as well as the recovery of organic phase and catalyst seem to be time-consuming, complex and costly.

Fig. 4 A process for HMF/furfural purification from an IL system using ethyl acetate or diethyl ether.^47,48

The HMF produced from fructose in an IL system of [HMIM][Cl] was extracted based on the process described in Fig. 4 using diethyl ether.⁴³ In another study, HMF produced from cellulose was separated from the system of [BMIM][Cl], Ipr-CrCl₂ and zeolites (H-Y, CVB 400) using diethyl ether after initially separating zeolites from the system by filtration.⁵² In the same vein, diethyl ether was used for the extraction of HMF produced from MCC in the system of 1-(4-sulfobutyl)-3- methylimidazolium hydrogen sulfate, MIBK and MnCl₂.⁵¹

The recyclability of recovered solvent in Fig. 4 was evaluated in the past, as it plays an important role in the overall feasibility of the developed processes. The recyclability of CoSO₄ and IL was evaluated when diethyl ether was used as the solvent for the process. The results showed a similar catalytic activity over five repetitions.⁴² Similarly, the recyclability of MnCl₂ and IL catalysts was assessed over five consecutive runs and the results showed that the yields of HMF and furfural production were slightly reduced from 37% and 18–33% and 14%, respectively, due to the accumulation of by-products (dimer of furan compounds) in IL system after the second run.⁵¹

In the same vein, HMF produced from cellulose and untreated wheat straw was separated from IL system of [BMIM][Cl] via initially filtering the solid residues and then mixing filtrate with ethyl acetate.³⁰ Subsequently, the remaining IL solution containing CrCl₃/LiCl catalyst was subjected to vacuum distillation. Ultimately, the recyclability of [BMIM][Cl] solvent with CrCl₃/LiCl catalyst was assessed, and the results showed that the HMF yield was remained constant (61–62%) after the three runs. However, the IL/catalysts recovery required a complicated process.³⁰

8. Conclusions

Various attempts were carried out to develop an environmentally friendly process for HMF production from biomass. Aqueous system showed to be more environmentally friendly than other described solvents and thus is preferred for reaction media. However, aqueous systems have the lowest HMF production yield and selectivity, which is due to side reactions to form by-products (LA and humins). In biphasic systems, the HMF will be transferred to the organic phase after forming in an aqueous phase, which will improve the HMF production yield and selectivity and minimize the undesired reactions. However, the limitation of such processes would be the application of environmentally unfriendly catalysts, the complexity of separation processes of HMF from the reaction media as well as issues related to the recyclability of solvents. Also, not much in-depth analysis was carried out on producing HMF from lignocellulosic biomass in biphasic systems. To make the process more economically feasible, the integration of new processes into the existing processes would be essential. Furthermore, the application of lignocellulosic residues as the feedstock under the aqueous phase conditions has to be assessed. Generally, research on solvent recovery design and the application of lignocellulosic feedstock as raw materials of these systems is highly demanded. Interestingly, IL systems offered unique opportunities for the production of HMF. A high yield (>80%) of HMF was obtained using fructose in IL systems, but the production of HMF from cellulose suffers from a low production yield. More recently, lignocellulosic residues were studied in IL systems for HMF production, but with a low yield. Despite the advantages of IL systems, the amount of solvents required seems to be high. Additionally, the application of metal catalysts offered a higher HMF yield, but the toxicity of metals is one of the main concerns of their utilization. The reusability and recyclability of IL remained under questions especially when untreated biomass residues are considered as feedstock due to their high impurities. These issues are current barriers of the large-scale application of IL systems. Finding IL systems with efficient recyclability and acceptable reaction performance is the subject of on-going research. In this context, finding non-metal IL systems that are more environmentally friendly has to be investigated.

Abbreviations

HMF	Hydroxymethyl furfural
NCC	Nanocrystalline cellulose
DMFU	Dimethylfuran
DFF	2,5-Diformylfuran
LA	Levulinic acid
EL	Levulinate
MIBK	Methylisobutylketone
THF	Tetrahydrofuran
DMSO	Dimethyl sulfoxide
[AMIM] [Cl]	1-Allyl-3-methylimidazaolium chloride
[C2MIM][Ac]	1-Ethyl-3-methylimidazolium acetate
IL	Ionic liquid
[BMIM][Cl]	1-Butyl-3-methylimidazolium chloride
[BMIM][BF4]	1-Butyl 3-methylimidazolium tetrafluoroborate
[BMIM][PF6]	1-Butyl 3-methylimidazolium hexafluorophosphate
[HMIM][Cl]	1-H-3-methylimidazolium chloride
YbCl3	Ytterbium chloride
[EMIM][HSO4]	1-Ethyl-3-methylimidazolium hydrogen sulfate
MCC	Microcrystalline cellulose
DMA	N,N-Dimethylacetamide
[C4MIM][Cl]	1-Butyl-3-methylimidazole chloride
[EMIM][Cl]	1-Ethyl-3-methylimidazolium chloride.

Acknowledgements

The authors would like to acknowledge NSERC Discovery grant for supporting this research.

References

1. P. Fatehi, Biotechnol. Prog., 2013, 29, 297.
2. D. J. Hayes, S. Fitzpatrick, M. H. B. Hayes and J. R. H. Ross, Biorefineries-Industrial Processes and Product, ed. B. Kamm, P. R. Gruber and M. Kamm, Wiley-VCH, Weinheim, 2006, vol. 1, pp. 139–164.
3. M. Dashtban, A. Gilbert and P. Fatehi, J. Sci. Technol. For. Prod. Processes, 2012, 2, 44.
4. M. Dashtban, H. Schraft and W. S. Qin, Int. J. Biol. Sci., 2009, 5, 578.
5. B. L. Peng, N. Dhar, H. L. Liu and K. C. Tam, Can. J. Chem. Eng., 2011, 89, 1191.
6. P. Sannigrahi, A. J. Ragauskas and G. A. Tuskan, Biofuels, Bioprod. Biorefin., 2010, 4, 209.
7. S. Cheng, I. D'Cruz, Z. S. Yuan, M. C. Wang, M. Anderson, M. Leitch and C. B. Xu, J. Appl. Polym. Sci., 2011, 121, 2743.
8. J. P. Lange, E. van der Heide, J. van Buijtenen and R. Price, ChemSusChem., 2012, 5, 150.
9. J. B. Binder and R. T. Raines, J. Am. Chem. Soc., 2009, 131, 1979.
10. Y. Román-Leshkov, C. J. Barrett, Z. Y. Liu and J. A. Dumesic, Nature, 2007, 447, 982.
11. M. Chidambaram and A. T. Bell, Green Chem., 2010, 12, 1253.
12. A. Chareonlimkun, V. Champreda, A. Shotipruk and N. Laosiripojana, Bioresour. Technol., 2010, 101, 4179.
13. F. W. Lichtenthaler, A. Brust and E. Cuny, Green Chem., 2001, 3, 201.
14. F. W. Lichtenthaler, ACS Symp. Ser., 2003, 841, 47.
15. A. Gandini, Green Chem., 2011, 13, 1061.
16. C. Moreau, M. N. Belgacem and A. Gandini, Top. Catal., 2004, 27, 11.
17. M. Gomes, A. Gandini, A. J. D. Silvestre and B. Reis, J. Polym. Sci., Part A: Polym. Chem., 2011, 49, 3759.
18. A. S. Amarasekara, D. Green and E. McMillan, Catal. Commun., 2008, 9, 286.
19. J.-P. Lange, R. Price, P. M. Ayoub, J. Louis, L. Petrus, L. Clarke and H. Gosselink, Angew. Chem., Int. Ed., 2010, 49, 4479.
20. S. Dutta, RSC Adv., 2012, 2, 12575.
21. S. Dutta, S. De and B. Saha, Biomass Bioenergy, 2013, 55, 355.
22. L. Hu, G. Zhao, W. Hao, X. Tang, Y. Sun, L. Lin and S. Liu, RSC Adv., 2012, 2, 11184.
23. N. Jiang, R. Huang, W. Qi, R. Su and Z. He, BioEnergy Res., 2012, 5, 380.
24. K. Lamminpää and J. Tanskanen, presented in part at the 8^th World Congress of Chemical Engineering, Montreal, 2009.
25. R. Lopes de Souza, H. Yu, F. Rataboul and N. Essayem, Challenges, 2012, 3, 212.
26. F. S. Asghari and H. Yoshida, Carbohydr. Res., 2006, 341, 2379.
27. F. S. Asghari and H. Yoshida, Ind. Eng. Chem. Res., 2006, 45, 2163.
28. Y. Li, X. Lu, L. Yuan and X. Liu, Biomass Bioenergy, 2009, 33, 1182.
29. F. S. Asghari and H. Yoshida, Ind. Eng. Chem. Res., 2007, 46, 7703.
30. P. Wang, H. Yu, S. Zhan and S. Wang, Bioresour. Technol., 2011, 102, 4179.
31. T. S. Hansen, J. M. Woodley and A. Riisager, Carbohydr. Res., 2009, 344, 2568.
32. R. Rinaldi and F. Scüth, Energy Environ. Sci., 2009, 2, 610.
33. X. H. Qi, M. Watanabe, T. M. Aida and R. L. Smith Jr, Catal. Commun., 2008, 9, 2244.
34. Y. Román-Leshkov and J. A. Dumesic, Top. Catal., 2009, 52, 297.
35. J. N. Chheda, Y. Román-Leshkov and J. A. Dumesic, Green Chem., 2007, 9, 342.
36. M. Bicker, D. Kaiser, L. Ott and H. Vogel, J. Supercrit. Fluids, 2005, 36, 118.
37. M. Bicker, J. Hirth and H. Vogel, Green Chem., 2003, 5, 280.
38. X. H. Qi, M. Watanabe, T. M. Aida and R. L. Smith, Jr, Green Chem., 2008, 10, 799.
39. Y. Roman-Leshkov, J. N. Chheda and J. A. Dumesic, Science, 2006, 312, 1933.
40. Y. Chen, Y. M. Zhang, F. Y. Ke, J. H. Zhou, H. P. Wang and D. H. Liang, Polymer, 2011, 52, 481.
41. W. Lan, C. F. Liu and R. G. Sun, J. Agric. Food Chem., 2011, 59, 8691.
42. F. Tao, H. Song and L. Chou, Carbohydr. Res., 2011, 346, 58.
43. C. Moreau, A. Finiels and L. Vanoye, J. Mol. Catal. A: Chem., 2006, 253, 165.
44. C. Lansalot-Matras and C. Moreau, Catal. Commun., 2003, 4, 517.
45. S. Bali, M. A. Tofanelli, R. D. Ernst and E. M. Eyring, Biomass Bioenergy, 2012, 42, 224.
46. Q. Cao, X. Guo, S. Yao, J. Guan, X. Wang, X. Mu and D. Zhang, Carbohydr. Res., 2011, 346, 956.
47. T. Stâhlberg, S. Rodriguez-Rodriguez, P. Fristrup and A. Riisager, Chem.–Eur. J., 2011, 17, 1456.
48. X. Tong, Y. Ma and Y. Li, Carbohydr. Res., 2010, 345, 1698.
49. T. Stahlberg, M. G. Sorensen and A. Riisager, Green Chem., 2010, 12, 321.
50. S. Lima, P. Neves, M. M. Antunes, M. Pillinger, N. Ignatyev and A. A. Valente, Appl. Catal., A, 2009, 363, 93.
51. F. Tao, H. Song, J. Yang and L. Chou, Carbohydr. Polym., 2011, 85, 363.
52. M. Tan, L. Zhao and Y. Zhang, Biomass Bioenergy, 2011, 35, 1367.
53. Z. Zhang and Z. K. Zhao, Bioresour. Technol., 2010, 101, 1111.
54. L. Zhang, H. Yu and P. Wang, Bioresour. Technol., 2013, 136, 515.
55. L. Zhang, H. Yu, P. Wang, H. Dong and X. Peng, Bioresour. Technol., 2013, 130, 110.
56. A. Chuntanapum, T. L. K. Yong, S. Miyake and Y. Matsumura, Ind. Eng. Chem. Res., 2008, 47, 2956.
57. Y. Y. Lee and K. C. W. Wu, Phys. Chem. Chem. Phys., 2012, 14, 13914.
58. S. Dutta, S. De, M. I. Alam, M. M. Abu-Omar and B. Saha, J. Catal., 2012, 288, 8.
59. B. M. Kabyemela, T. Adschiri, R. M. Malaluan and K. Arai, Ind. Eng. Chem. Res., 1999, 38, 2888.
60. V. Choudhary, S. H. Mushrif, C. Ho, A. Anderko, V. Nikolakis, N. S. Marinkovic, A. I. Frenkel, S. I. Sandler and D. G. Vlachos, J. Am. Chem. Soc., 2013, 135, 3997.

---