ILs are being investigated as a solvent for not only pure cellulose but also other cellulosic biomass. Cellulosic biomass, such as wood, is composed of several hardly soluble polymers and many other materials. Other polysaccharides are also an attractive target to be extracted from biomass. In 2007, Fort and co-workers reported wood biomass treatment by ILs, and they clarified that a mixture of [C4mim]Cl and DMSO partially dissolved wood biomass at 100 °C (Fig. 2.5) [48]. The dissolving degree was achieved to about 70 % (wt/wt at added biomass). They analyzed the extracted materials and clarified that they were a mixture of polysaccharides and lignin. Shortly after this, Kilpelainen and co-workers reported on wood dissolution by [C4mim]Cl and [Amim]Cl [49]. They treated soft — and hard-wood such as Norway spruce sawdust and southern pine thermomechanical pulp at temperatures between 80 and 130 °C for 8 and 13 h, respectively, and

observed that the biomass samples were partially dissolved. When the dissolution of the same lignocellulosic samples was soaked in 1-benzyl-3-methylimidazolium chloride ([Bnmim]Cl), transparent amber solutions were obtained. Wang and co-workers used a room temperature IL, [Amim]Cl to extract cellulose-rich mate­rial from several wood chips such as pine, poplar, Chinese parasol, and catalpa [50]. They showed that pine was one of the most suitable wood species for cellulose extraction with ILs, and its cellulose extraction degree reached to 62 %.

Miyafuji and co-worker observed the state of woodchips from softwood, Cryptomeria japonica, during the ILs treatment using light microscope [51]. Figure 2.6 shows micrographs of latewood, earlywood, and the latewood/ earlywood boundary after treatment with [C2mim]Cl at 120 °C. The cell walls in latewood became disordered after 0.5 h treatment. In addition, some destruction or flaking was observed in the cell walls after 4 h treatment. By contrast, no significant change was observed in earlywood even after 4 h treatment. They suggested that latewood swells easier than earlywood because of the difference in the density.

Although ILs could dissolve only a part of wood biomass in an early stage, the complete dissolution of wood was achieved by Sun and co-workers in 2009 with carboxylate salts under heating [52]. After that the following separation methods were also investigated. Kilpelainen and co-workers also reported the complete dissolution of lignocellulose materials [49]. That process helps to break some of interchain chemical bonds such as lignin-carbohydrate bond, and the lignocellulose material was used after mechanical pulping. Sun and co-workers clarified that [C2mim][OAc] completely dissolved softwood (southern yellow pine) and hardwood (red oak) after 46 and 25 h heating at 110 °C for pine and oak, respectively. In addition, they suggested that carbohydrate-free lignin and cellulose-rich materials

were obtained by adequate precipitating process by the addition of acetone and water. On this basis, they developed the biomass treatment process as shown in Scheme 2.7.

Regarding lignin regeneration, Casas and co-workers also studied and reported some interesting results [53]. They collected regenerated lignin from Pinus radiata and Eucalyptus globulus woods dissolved in imidazolium-type ILs. Lignin was successfully regenerated by precipitation with methanol from wood solutions in [Amim]Cl, [C4mim]Cl, or [C2mim]Cl. Against this, lignin was not regenerated from acetate-type ILs. In addition, contents of different functional groups in the regenerated lignin were found to depend on the species of IL employed as well as wood species dissolved.

In the next section, direct lignin extraction from wood is mentioned. Sun and co-workers investigated the effect of particle size of the added biomass [52]. For [C4mim]Cl, the particle size was observed to have a significant influence on the extraction of lignin. The IL dissolved 52.6 % of the finely milled biomass (<0.125 mm), but only 26.0 % of coarser biomass (0.25-0.50 mm). It is easy to comprehend that smaller particles have larger gross surface area and lignin is easier to be solubilized. On the other hand, for [C2mim][OAc], the particle size of biomass did

not affect the results significantly. The [C2mim][OAc] dissolved more than 90 % of the added wood even from the particles as large as 0.5-1.0 mm. Sun et al. also evaluated the effects of some pretreatments, i. e., microwave or ultrasound irradiation (Table 2.7) [52]. These pretreatments accelerated the lignocellulose dissolution. With 60 x 3 s microwave pulses, the time for complete dissolution (tcd) was reduced to shorter than half of that without pretreatment. As seen in Table 2.7, ultrasound pretreatment also accelerated the dissolution. In spite that these pretreatments are effective, it should not be ignored that these steps also consume energy.

In 2011, Sun et al. reported that complete dissolution of lignocellulose was carried out with shorter mixing time at temperature above the glass transition temperature of lignin [54]. Complete dissolution of 0.5 g bagasse in 10 g of [C2mim][OAc] requires more than 15 h heating at 110 °C, by contrast, it dissolves completely in the IL within 5-15 min heating at 175-195 °C. In addition, processing bagasse in the IL at 185 °C for 10 min gave higher yields of both recovered lignin and carbohydrate than the previous methods using lower temperatures and longer times (e. g., 110 °C, 16 h). There was an associated problem with the thermal stability of [C2mim][OAc], because about 15 % of the IL degraded after processing at the higher temperature.

Miyafuji and co-workers reported that cellulose dissolving ILs work as not only a solvent for plant biomass but also a reaction medium. They found that [C2mim]Cl

dissolved wood and that the solubilized polymers such as cellulose were depolymerized to low molecular weight compounds just by mixing [55]. Japanese beech wood flours (0.09 g) were added to 3 g of [C2mim]Cl, and the mixture was heated to 90-120 °C under gentle stirring. After that, the molecular weight distri­bution of the solubilized compounds in [C2mim]Cl was studied by gel permeation chromatography. As a result, the molecular weight of the solubilized compounds was found to decrease as the treatment time was extended, and such depolymeri­zation was more enhanced at higher temperature. They suggested that [C2mim]Cl penetrated into wood and liquefied polysaccharides such as cellulose at the initial stage of the reaction, and the crystal structure was gradually broken down.

The cellulosic biomass dissolving ability of several CDILs is summarized in Table 2.8.

While the action of new designer ILs on wood is still under investigation, many details about the mechanisms of wood solubility in commonly utilized dialkylimi- dazolium ILs have been revealed. In the early publications covering the area of wood dissolution, a complete dissolution is mentioned to take place [7, 8, 12], at least for certain wood species. Since it has been demonstrated that all of the wood components are soluble in ILs in their individual purified forms [5, 21, 22], the naive simplification was made that wood is soluble in such ILs. However, later it was determined that wood dissolution under mild dissolution conditions was partial, rather than complete. In light of our results and other recent publications, it seems clear that mild dissolutions (<100 °C) in chloride-based ILs, such as [amim]Cl, are actually not able to provide the driving force to completely dissolve the wood cell wall, in its native state [14, 23—25].

Even if each of the polymeric components, in their purified form, has good solubility in the selected ILs, this does not ensure efficient solubilization of intact fibers. This is because of the complexity of the native wood. The structure of the plant cell wall is complex and highly orientated with many physically and chem­ically distinct regions, such as primary and secondary (S1, S2, and S3) cell walls and the presence of middle lamella. The variety of polymeric backbone structures, the presence of multiple functionalities and in particular the covalent/physical interactions between the three main wood components make things considerably more complicated. Nature’s design of the cell wall is resistant towards physical stress and controls the diffusion of fluids inside the fiber. Inefficient mass transfer of IL or solvated polymers can greatly hinder the dissolution process, even in ILs having high capability for dissolution [26]. This is true especially in the cases where single components are isolated by an extraction type of mechanism. For ILs unable to dissolve cellulose, delignification has been observed to take place mainly on the outer surface of the fibers, due to low accessibility inside the bulky secondary wall [27]. This is analogous to kraft pulping where initial delignification occurs on the fiber surface in the middle lamella and lumen. The different cell types in corn stover show drastic differences in relative lignin solubility, during IL treatment, which further demonstrates the significance of cell wall ultrastructure or composition [26]. Even the early — and late woods in the same sample piece of Sugi wood showed very different response to swelling in IL [28]. They showed that the empty lumen allows for flexibility that protects the fiber from swelling induced physical defects. There also seem to be fundamental differences in the overall solubility, or at least in the kinetics of the dissolution, between soft — and hardwood species [6, 8, 24]. Apparently the natural design of tracheids and fibers between soft — and hardwood offers different resistance towards dissolution. All these reports demon­strate that there is a need to focus on revealing the degrees of recalcitrance, related to physical (macroscopic cell wall structure and crosslinking) and chemical factors (polymer interactions and covalent bonding) prior to making generalizations about the efficacy of fractionation of a single IL system on a single species.

Evidently the distribution and structure of lignin in cell walls has a crucial effect on wood solubility. This is not surprising as lignin is considered to be a branched polymer formed by random radical driven crosslinking, thus resembling a complex networked structure [29]. A fundamental property of crosslinked polymer structures is their inability to form true solutions. There are several publications about lignin isolation from lignocellulose via extraction-type mechanisms, using various ILs [9, 10, 20, 30, 31]. It is also not surprising that depolymerization of the isolated lignins have been observed [9, 20, 31]. Depolymerization can be a result of several mechanisms, including covalent bond scission via pulverization operations. This can be controlled by the conditions and choice of IL, as discussed in the following section.

The incomplete solubility of cell wall in an unreactive IL does not rule out component fractionation. It seems that dialkylimidazolium chlorides have a much lower reactivity with lignin than other commonly used IL types, such as dialkyli — midazolium acetates. A group of publications has shown that this property can be utilized in isolation of cellulose rich fractions [14, 23—25]. In particular hemi — celluloses and lignin are not separated completely in such systems. Physical or chemical methods to alter cell wall ultrastructure and polymeric networks, prior to or simultaneously with dissolution, may be crucial for complete fractionation. Support of this hypothesis can be found from studies where lignin structure was altered by (1) excessively heating the mixtures until thermal decomposition reac­tions start to take place [32], or (2) use of an oxidative catalyst to partially degrade the lignin polymer backbone [33]. Enhanced separation of lignin from polysaccha­ride was achieved, compared to a previous method [8]. What is not known and not taken into account are the above enumerated considerations and the molecular weight distributions of the resulting pulps.

Inulin, which is also called fructan, is a carbohydrate consisting of fructose units that vary in the degree of polymerization (DP) from 2 to 60, or higher. The fructosyl units in inulin are linked by P(2 ! 1) linkages with the polymer chains terminating with a glucose unit [62]. Inulin is indigestible. The production of 5-HMF from inulin in ionic liquids has not been studied as extensively as cellulose, since inulin is not so abundant in nature. However, the carbohydrate provides a different feedstock that might give insight into mechanism of its transformation into 5-HMF. The hydrolysis of inulin to fructose followed by the dehydration of fructose to produce 5-HMF is a possible two-step reaction pathway. Because both reaction steps are catalyzed by acid catalysts, it is interesting to consider the production of 5-HMF from inulin as a one pot reaction, since this would avoid the separation of the fructose in the intermediate step.

Hu et al. [63] developed a process for the direct conversion of inulin to 5-HMF in choline chloride (ChoCl)/oxalic acid and ChoCl/citric acid deep eutectic solvents (DES), for which a 5-HMF yield of 56 % was obtained at relatively low temper­atures (80 °C). When a biphasic system with IL and ethyl acetate was used for the in-situ extraction of 5-HMF, the 5-HMF yield was enhanced to 64 %, since the product 5-HMF is soluble in ethyl acetate, while the reactant inulin and fructose are insoluble in the DES, and ethyl acetate is only slightly soluble in the DES, which is favorable for the reaction in the biphasic system [63]. Although SnCl^EMIM] [BF4] system is efficient for glucose dehydration, there is no advantage in using the system for inulin conversion to 5-HMF since only moderate 5-HMF yields of 40 % are obtained [50].

Qi et al. [64] used the characteristics of two kinds of ionic liquids to develop an efficient process for the direct conversion of inulin to 5-HMF in one pot with two reaction steps under mild conditions. In the first step, the ionic liquid 1-butyl-3- methyl imidazolium hydrogen sulfate ([BMIM][HSO4]) was employed as both solvent and catalyst for the rapid hydrolysis of inulin into fructose with 84 % yield in 5 min reaction time at 80 °C. In the second step, 1-butyl-3-methyl imidazolium chloride ([BMIM][Cl]) and a strong acidic cation exchange resin were added to the mixture to selectively convert the generated fructose into 5-HMF, achieving a 5-HMF yield of 82 % in 65 min, which is the highest 5-HMF yield reported to date for an inulin feedstock. These authors proposed a conceptual process that the ionic liquid, resin, catalyst, ethyl acetate, and the supplied chemicals are internally recycled and represent an efficient method for producing 5-HMF from inulin (Fig. 9.4).

Xie et al. [65] studied the catalytic conversion of inulin into 5-HMF in ionic liquids by lignosulfonic acid catalyst, and found that that consecutive hydrolysis of inulin was very fast, with a maximum yield of 47.0 % achieved in only 5 min, and the yield could be maintained even with prolonging the reaction time to 90 min. The effect of water addition was examined and it was demonstrated that the addition of

H2O does not affect the reaction, with comparative yields obtained since the dehydration of one molecule of fructose produces three molecules of H2O, which is sufficient for the previous hydrolysis of inulin into fructose, therefore, additional water does not seem to be necessary for this reaction system. Bennoit et al. [66] tried to partially substitute ionic liquids with glycerol or glycerol carbonate as cheap, safe and renewable sourced co-solvents for the acid-dehydration of inulin to 5-HMF. They found that a system that used [BMIM][Cl]/glycerol carbonate in a 10:90 ratio as solvent and a wet Amberlyst 70 acidic resin as catalyst resulted in a 5-HMF yield of 60 % at 110 °C. A major part of the ionic liquid was substituted with glycerol carbonate which allows the cost and environmental impact of the process to be potentially be lower than methods that petroleum derived co-solvents.

Although ionic liquids have been investigated as solvent for the synthesis of 5-HMF, they still have some shortcomings such as price and toxicity that hamper their industrial applications. Liu et al. [38] developed an interesting choline chlo­ride (ChCl)/CO2 system, where the addition of CO2 could decrease the pH of the system through formation of carbonic acid and played as an efficient catalyst for the hydrolysis of inulin and the dehydration of fructose into 5-HMF. The inulin conversion with or without water addition in ChCl/CO2 system was studied. The yield of 5-HMF was rather disappointing (12 %) in the absence of water (120 °C, 4 mPa CO2, 90 min). The authors attributed the low 5-HMF yield to the low content of water at the initial stage of the reaction that does not favor the hydrolysis of inulin to fructose. Hence, water was initially added (16 wt%), and the yield of 5-HMF improved to 38 % when the reaction time was prolonged to 6 h. The 5-HMF was recovered with 41 % yield after 15 h of reaction, demonstrating the stability of 5-HMF in ChCl as compared to other solvents. This process has many advantages: (1) it is not necessary to remove the trace amounts of catalyst contained in 5-HMF after extraction; (2) the use of cheap and safe CO2 as a source of acid catalyst;

(3) the low ecological risk of the ChCl/CO2 system that is renewable; and (4) the system is still efficient at high loadings of fructose (up to 100 wt%).

Toshiyuki Itoh

Abstract Cellulose consists of linear glucose polymer chains that form a very tight hydrogen-bonded supramolecular structure making it highly resistant to enzymatic degradation. The ionic liquid, 1-butyl-3-methylimidazolium chloride ([C4mim]Cl), has been found to dissolve cellulose and the regenerated cellulose from the IL solution is less crystalline. To design ionic liquids that dissolve cellulose, Kamlet — Abboud-Taft p-values can be used as a solvent indicator. Amino acid anions have strong interactions between hydroxyl groups in the cellulose molecule: N, N-diethyl, N-methyl, N-(2-methoxy)ethylammonium alanate ([N221(ME)][Ala]) thus they are studied in this chapter for cellulose dissolution. Addition of an anti-solvent like water or ethanol to the cellulose/IL solution caused precipitation of cellulose dissolved and the structure of the regenerated cellulose to change to a disordered form. Crystal form of the regenerated cellulose depends on the dissolution solvent; the disordered chain region seems to increase in the order of [N221(ME)][Ala] < [C2mim][OAc] < [C2mim][(EtO)2PO2] < [C2mim]Cl. On the other hand, the order of degree of polymerization of the cellulose is [N221(ME)][Ala] > [C2mim] [OAc] > [C2mim][(EtO)2PO2] > [C2mim]Cl. Treatment with [N221(ME)][Ala] is therefore much more suitable to use in preparing regenerated cellulose fibers than other commonly used ionic liquids.

Keywords Cellulose dissolution • Amino acid ionic liquids • Regenerated cellu­lose • A mixed solvent

T. Itoh (*)

Department of Chemistry and Biotechnology, Graduate School of Engineering, Tottori University, 4-101 Koyama Minami, Tottori, Japan e-mail: titoh@chem. tottori-u. ac. jp

Z. Fang et al. (eds.), Production of Biofuels and Chemicals with Ionic Liquids, Biofuels and Biorefineries 1, DOI 10.1007/978-94-007-7711-8_4,

© Springer Science+Business Media Dordrecht 2014

4.1 Introduction

Cellulose is an important renewable resource for production of biocomposites and biofuel alcohols. However, since it consists of linear glucose polymer chains that form a very tight hydrogen-bonded supramolecular structure, cellulose resists enzymatic degradation. There has been growing interest in the development of a means of modifying cellulose structure to an easily digestible form by biodegrada­tion [1]. Multiple hydrogen bonding among cellulose molecules results in the formation of highly ordered crystalline regions [2]. Therefore, cellulose does not dissolve in water and common organic solvents at ambient conditions. The chal­lenge for dissolving cellulose has a long history [3]. The first attempt was reported early in the 1920s and some mixed solvent systems for cellulose dissolution were developed [1,3]: sodium hydroxide/carbon disulfide (CS2) [4] and sodium hydrox — ide/urea [5] are well known as commercial cellulose derivatizing solvents. Rosenau et al. [6] reported using N-methylmorphorine-N-oxide monohydrate (NMMO) as a solvent for direct dissolution of cellulose in an industrial fiber-making process [6]. Combination of a polar molecular solvent with a salt was also reported to dissolve cellulose: N, N-dimethyl acetoamide (DMA) in combination with LiCl [7], a mixture of DMSO and tetrabutylammonium fluoride (TBAF) [8] were thus developed as cellulose dissolution solvents. Recently, Ohno and co-workers reported an interesting cellulose dissolution system of a mixed solvent of tetrabutyl- phosphonium hydroxide (TBPH) containing 40 wt% water [9]. In all cases, an appropriate combination of organic salts and polar solvents was essential to realiz­ing high dissolution of cellulose. Fischer et al. [10] reported that molten salt hydrates (LiXmH2O; X = Г, NO3“ CH3CO2 , ClO4~) dissolved cellulose [10]. It is now well recognized that very high polarity of the solvent system might be the key to breaking down the cellulose network and dissolving cellulose even if these solvents have no ionic character [3] (Fig. 4.1). However, there is a serious environmental drawback to such traditional solvent systems: they generally require large quantities of hazardous chemicals and high temperatures. From the standpoint of green chemistry, development of a safe and efficient cellulose disso­lution process can be anticipated [3].

Ionic liquids (ILs) usually melt below 100 °C and are becoming attractive alternatives to volatile and unstable organic solvents due to their high thermal stability and nearly non-volatility. The most fascinating nature of ILs is their structural diversity. We are able to design their physicochemical properties, includ­ing viscosity, polarity, and hydrophobicity. Numerous papers and several reviews on the ILs have been published [11], and it is now widely recognized that ILs are applicable to the media for many types of chemical reactions [11] and even for enzymatic reactions [12]. ILs have consequently show a unique solubility in many inorganic and organic materials, and it is anticipated that ILs might dissolve insoluble compounds, including cellulose, which is impossible with conventional molecular liquids [11].

Swatloski et al. [13] reported a breakthrough on this issue using ionic liquid technology: they found that cellulose dissolved in an ionic liquid, 1-butyl-3- methylimidazolium chloride ([C4mim]Cl), and that the regenerated cellulose from the IL solution was less crystalline [13]. An increased reaction rate of cellulase — mediated hydrolysis was realized when regenerated cellulose from the ionic liquid solution was subjected to the enzymatic reaction resulting from reduced crystallin­ity [13]. Since then, extensive investigations have been carried out to develop an IL that possesses the capability to dissolve cellulose and reduce its crystallinity [1,3,

14] . Most reported ILs are imidazolium based or alkyloxyalkyl-substituted ammo­nium salts with chloride, formate, acetate, propionate, or phosphate as counter anion (Fig. 4.2) [1, 3, 14].

Currently, most commercial biodiesel is produced from plant lipids using homoge­neous basic catalysts such as NaOH or KOH [20]. Nevertheless, these catalytic systems have a number of drawbacks: (I) catalysts cannot be reused and have to be neutralized which produces wastewater; (II) formation of stable emulsions that makes FAMEs separation difficult; (III) glycerol is obtained as an aqueous solution with low purity; (IV) the process is sensitive to residual water and free fatty acids [33]. ILs are recognized as green solvents due to their special properties comparing with traditional organic solvents, such as tunability, non-detectable vapor point and performance benefits over molecular solvents. Their properties can be designed to suit a particular need by changing the structures of the cation, anion or both acidic or basic for special synthesis [34, 35]. The principle was widely used for biodiesel production from lipids [11—13].

In consideration of good solubility of ILs to inorganic and organic compounds, and tunable miscibility with organic solvents, the simplest way is to use ILs for biodiesel production as solvents to immobilize traditional acidic or basic catalysts, such as K2CO3, NaOH, hydroxide salts of ammonium cations, sodium methoxide, lithium diisopropylamide, and H2SO4 [36]. Usually, a two-phase system (a glycerol-methanol-ILs-catalyst phase and biodiesel phase) forms due to the immiscibility of biodiesel with ILs after the reaction is done. The catalytic system can be reused after decanting the biodiesel directly. For example, under basic conditions, the combination of 1-n-butyl-3-methylimidazolium bis(trifluoromethyl — sulfonyl)imide (BMI • NTf2), alcohols, and K2CO3 (40 mol%) results in production of biodiesel from soybean oil in high yield and purity. H2SO4 immobilized in BMI • NTf2 efficiently promotes the transesterification reaction of soybean oil and various primary and secondary alcohols. In this multiphase process the acid is almost completely retained in the IL phase, while the biodiesel forms a separate phase. The recovered IL containing the acid catalyst could be reused for six times without significant yield or selectivity loss [36].

It is known that the acidity and basicity of ILs can be tuned by changing the composition of cationic and anionic species. Some acidic or basic ILs have been used as both catalysts and solvents for the synthesis of biodiesel (Scheme 7.2) These ILs can be synthesized by introduction of acidic functional groups into either the cation or anion, or adding a Lewis acid catalysts in ILs to form a catalytic active Lewis acid ILs [12, 37]. No matter the use of ILs as solvents or catalysts, the processes are usually efficient and facile for biodiesel production (Table 7.4). Inexpensive materials such as non-edible oils and waste cooking oils contain high free fatty acid contents, which are not suitable for base-catalyzed biodiesel produc­tion process. Therefore, free fatty acids should be converted into FAMEs, for which acidic ILs have been better than traditional mineral acids. For example, the dicationic IL N, N,N, N-tetramethyl-N, N-dipropanesulfonic acid ethylenediammonium hydro­gen sulfate could be used as efficient and recyclable catalyst for the synthesis of biodiesel from long-chain free fatty acids or their mixtures [44]. The reaction was accomplished in a monophase at 70 °C for 6 h, while the products were separated by liquid/liquid biphase separation at room temperature with yields of 93-96 %. The work-up process was simple, and the catalysts could be reused for six times with little activity loss. This novel and clean procedure offered advantages including short reaction time, high yield, operational simplicity, and environmental friendliness. To achieve a better catalyst separation, acidic ILs-based catalysts have been covalently immobilized onto SBA-15. The immobilized catalysts displayed relatively high activity in esterification of oleic acid with short-chain alcohols because of the synergistic effects of both Lewis and Br0nsted acidic sites. Under the optimal reaction conditions (molar ratio of methanol to oleic acid 6:1,5 wt% catalyst loading, and 363 K for 3 h), the conversion of oleic acid reached 87.7 %. It was found that some metal chloride-based ILs could efficiently convert un-pretreated Jatropha oil with high-acid value (13.8 mg KOH/g) to biodiesel. For example, when FeCl3 was added to [1-butyl-3-methyl-imidazolium][CH3SO3], a biodiesel yield of 99.7 % was achieved at 120 °C [53].

The basic ILs can be designed by introducing a strong basic anion or an organic basic moiety. The principle was used widely to synthesize ILs for the transester­ification of lipids with methanol and ethanol. Most recently, three novel dicationic basic ILs were prepared for synthesis of biodiesel from soybean oil. Among them, 1,2-bis(3-methylimidazolium-1-yl)ethylene imidazolide showed the highest bio­diesel yield of 99.6 %. When the acidity of soybean oil was 0.49 mg KOH/g, the yield of biodiesel was 99.6 %. However, when it was 1 mg KOH/g, the yield of biodiesel dropped to 82.5 %. Thus, basic ILs had limited capacity to use high acidity feedstock for biodiesel production [54].

It is well recognized that organic bases (e. g. 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU), guanidine) are important catalysts for the transesterification of lipid with alcohols [27, 55—57]. As both organic bases and alcohols are important components in the reaction, a novel phase-switching homogeneous catalysis was devised for clean production of biodiesel and glycerol (Scheme 7.3). It was found that the FAMEs can be decanted from the system and the yields were up to 95.2 % [27].

The produced glycerol was extracted from FAMEs completely by the “switchable solvents”, and recovered with high purity after recycling DBU by another extraction process. This system has been tested for integrated production of biodiesel from cell mass of the oleaginous yeast Rhodosporidium toruloides Y4. While intracellular lipid was successfully extracted, only about 21.9 % of the lipid was converted into FAMEs. Nonetheless, such systems offered significant advantages.

It is well known that the performance of enzymes in organic solvents as well as in ionic liquids is affected by common factors such as water activity, pH as well as the existence of excipients and impurities. Moreover, the physicochemical properties of ionic liquids such as viscosity, polarity, hydrophobicity, kosmotropicity, etc. also have strong effect on the activity and stability of enzymes in ionic liquids. The effect of these physicochemical properties of ionic liquids on enzyme activity and stability are summarized in Table 10.2.

In general, enzymes are optimized in nature to perform best in aqueous envi­ronment, at neutral pH, temperature below 40 °C and at low osmotic pressure. When enzymes are used in either pure organic solvents or ionic liquids, the

minimum amount of water, which is best described by the water activity, is of crucial importance to maintain the enzyme activity. For ionic liquids, the same methods can be used to maintain a constant water activity as those established for organic solvents [33]. However, in some ionic liquids (e. g. fluoride contained ionic liquids), the water present in the reaction system may cause hydrolysis of ionic liquids and result in reduced enzyme activity and stability. This is attributed by the change in pH of system and the inhibition effect of the hydrolyzed products. The pH strongly effects the activity and stability of enzyme in ionic liquids media as shown in work of Tavares et al. [48]. In this study, the activity of laccase was well maintained at pH 9.0 for 7 days (with activity loss about 10 %) in aqueous solution of 1-ethyl-3-methylimidazolium 2-(2-methoxyethoxy) ethylsulfate, [Emim] [MDEGSO4], while significantly reduced at pH 5.0. Moreover, the impurities present in ionic liquids are also important factor that need to be taken into account when carrying out enzymatic reactions in ionic liquids. The impurities may influ­ence the physicochemical properties of ionic liquids and hence, enzymatic reaction. For example, the activity of immobilized CALB (Novozym 435) in [Omim][Tf2N] decrease linearly with chloride content while the activity of immobilized lipase from Rhizomer miehei ( Lipozyme IM) drastically decreases in the presence of [Omim][Cl] [49].

Since ionic liquids are composed only of ions, the effect of ions on the enzyme activity and stability is also an important factor. Ions can affect the stability of proteins through the interactions between the ions and charged amino acid groups in the proteins [50]. Some enzymes require metal ions such as cobalt, manganese, zinc, etc. for their activity. If these ions are removed or interfered by interaction

m—————— Stabilizing Destabilizing———————————— ►

Anion F"> PO43′> SO42′> CH3COO’> Cl’> Br’> I’> SCN’

Cation (CH3)4N+>(CH3)NH2+> NH4+> K+> Na+> Cs+> Li+> Mg2+> Ca2+> Ba2+

Fig. 10.1 The Hofmeister series as an order of the ion effect on protein stability (Reprinted with permission from Ref. [43]. Copyright © 2013, Elsevier) with ionic liquids, enzyme might be inactivated. Several researchers have proposed Hofmeister series to explain the behavior of enzymes in aqueous solution of ionic liquids. Hofmeister series (Fig. 10.1) which indicates the kosmotropicity of individual cations and anions of ionic liquids may be a good guide for choosing ionic liquids for enzyme activity and stability in aqueous solutions [43, 44, 46]. The anion such as PO43~, SO42~, CO32~ (kosmotropic anions), and cations such as NH4+, K+, Cs+ (chaotropic cations) stabilize enzymes, while chaotropic anions and kosmotropic cations destabilize them. However, the influence of Hofmeister series is complicated when enzymes are present in nearly anhydrous ionic liquids.

Properties of ionic liquids such as viscosity, polarity, and hydrophobicity also affect the enzymes in ionic liquids. Ionic liquids are well recognized to have higher viscosity than conventional organic solvents. As a general trend, enzymes are more active in less viscous ionic liquids that can be attributed to the mass transfer limitation in high viscous media. For example, a-chymotrypsin maintained higher activity in less viscous [Emim][Tf2N] (34 mPa • s) than that in high viscous [MTOA][Tf2N] (574 mPa • s) for the synthesis of N-acetyl-L-tyrosine propyl ester [26]. Ionic liquids are considered as highly polar account on their ionic nature. The polarity of ionic liquids has been empirically determined by means of a variety of solvatochromic probe dyes [51]. For instance, betaine dye no. 30 or Reichardt’s dye has been used to characterize ionic liquid polarity by the solvent polarity parameter ET(30) and the corresponding normalized polarity scale £^. In addition, the Kamlet-Taft parameters (a, p and n*, which quantify hydrogen-bond donating ability (acidity), hydrogen-bond accepting ability (basicity) and polarity/polariz — ability, respectively), determined by set of dyes, have also been used to quantify the polarity of ionic liquids [52, 53]. Based on solvatochromic probes studies, ionic liquids have polarity close to low chain alcohol or formamide [54]. Several studies have demonstrated that activities and stabilities of enzyme in ionic liquids were increased with increasing ionic liquids polarity [24, 26] although there were some reports showing no clear effect of polarity on enzyme [21,45]. Hydrophobicity may be considered as narrow concept of polarity. However, it is practically important to separate hydrophobicity to polarity since the former is often related to the misci­bility with water [55]. Depending on the structure of cation and anions, ionic liquids can be hydrophilic or hydrophobic. As a general rule, the structure of anions has strong influence on the hydrophobicity of ionic liquids than the contribution of cations. Hydrophilic anions such as halides, carboxyl groups, having high hydrogen bonding capability, strongly interact with enzyme resulting in the conformational change and deactivation of enzymes. On the other hand, the hydrophobicity of ionic
liquids increases as the alkyl chain length in cations increases. In some reports, ionic liquids with long alkyl chain in cations may behave as surfactants in aqueous solution and have strong stabilizing impact on enzyme [56, 57]. However, in some cases, the long alkyl chain might have negative effect on the activity and stability of enzyme due to the resulted higher viscosity [58]. In practice, the log P (logarithm of partition coefficient between octanol and water) scale can be used to quantify the hydrophobicity of ionic liquids, and in some reports this scale can be used to optimize enzyme activity and stability in ionic liquids [21, 46, 59, 60].

Viscosity is one of the most important physical properties when considering ionic liquid applications. The viscosities of many ILs are much higher than most organic solvents at room temperature. Generally, the viscosity of ILs is 10-1,000 mPa s. A low viscosity is generally desired to use IL as a solvent, to minimize pumping costs and increase mass transfer rates while higher viscosities may be favorable for other applications such as lubrication or use in membranes [49]. Viscosity can be fitted with the Vogel-Tammann-Fulcher equation although it usually follows a non-Arrhenius behavior. Viscosities of ILs remain constant when the shear rate increases so that they have Newtonian and non-Newtonian behaviors [44].

The viscosity of ILs is usually affected by the kind of the anion, cation and substituents on the cation and anion of the imidazolium-based ILs. Generally, for ILs with the same anion, the alkyl substituents on the imidazolium cation is larger,

the viscosity of ILs is higher. For example, for the 1-alkyl-3-methylimidazolium hexafluorophosphate and bis((trifluoromethyl)sulfonyl)imide series ([Rmim][PF6] and [Rmim][Tf2N]), viscosity increases with increasing the number of carbon atoms in the linear alkyl group [50]. Furthermore, branching of the alkyl chain in 1-alkyl-3-methylimidazolium salts usually result in lower viscosity. Finally, a reduction in van der Waals interactions can also attribute to the low viscosity of ILs bearing polyfluorinated anions. Hydrogen bonding between counter anions and symmetry can also affect viscosity. In short, the viscosity of ILs based on the most common anions decreases in the order Cl_ > [PF6]_ > [BF4]_ > [TfO]_ > [Tf2N]“ > [dca]“ (shown in Fig. 1.3) [26, 51—53].

The impurities in the ILs greatly affect their viscosities [4]. In one study [54], a series of ILs were prepared and purified by many kinds of techniques. Then their impurities were analyzed and physical properties were evaluated. The results showed that chloride concentrations of up to 6 wt% were found for some of the preparative methods whereas chloride concentrations of between 1.5 and 6 wt% increased the observed viscosity by between 30 and 600 %. Studies also found that the non-halo aluminate alkylimidazolium ILs absorbed water rapidly from the air. As little as 2 wt% (20 mol%) water could reduce the viscosity of [BMIM][BF4] by more than 50 %. Therefore, purities and handling should be carefully considered when viscosities of ILs are measured.

Owing to the widely application of ILs, the experimental measurement and theoretical modeling of viscosities of ILs and mixtures are essential in the devel­opment and design of processes [55]. There are several models used for the prediction of ILs viscosities.

Abbott [56] proposed a theoretical model for prediction of viscosities by mod­ifying the “whole theory”. In that model, 11 ILs mainly based on imidazolium at three temperatures (298, 303 and 364 K) were investigated. The model had low reliability despite its theoretical interpretation and therefore it has limited applica­tion for practical processes.

Han et al. [57] proposed a QSPR method for prediction of the viscosity of imidazolium based ILs. In that work, a database of 1,731 experimental data values at various temperatures and pressures were used for 255 ILs, that included 79 cat­ions and 71 anions. As for the viscosity of imidazolium-based ILs, the cation-anion electrostatic interactions have important effects.

The most useful viscosity estimation models for complex molecules are those based on group contributions. The methods usually use some variation of temper­ature dependence proposed by de Guzman [58].

The Orrick-Erbarmethod [59] proposed employs a group contribution technique to estimate the A and B parameters in the following equation [49]:

n B

ln— = A +- (1.6)

pM T v ‘

Where the n and p are the viscosity in mPa • s units and density is in g • cm~3 units, respectively.

Viscosities calculated by the following method are in good agreement with experimental literature data. The model could predict the viscosity of new ILs in wide ranges of temperature and could be extended to a larger range of ILs as data for these become available. It is also shown that an Orrick-Erbar-type approach was successfully applied to estimate of the viscosity of ILs by a group contribution method.

In 2002, it was reported that ILs can dissolve biomass materials [23]. Viscosity plays a role in cellulose solvation, because it considered that ILs with low viscosity are more efficient and easier to handle in dissolving cellulose [60]. When an IL has a low viscosity, cellulose can be dissolved at room temperature. For example, microcrystalline cellulose was dissolved at a lower temperature in 1-ethyl-3-methyl imidazolium methylphosphonate [EMIM][CH3PO4] in compared with 1-ethyl-3- methylimidazolium dimethylphosphate [EMIM][(CH3)2PO4] [61]. However, vis­cosities of ionic liquids are not the only important parameter in biomaterial disso­lution. In 1-benzyl-3-methylimidazolium chloride [PhCH2MIM]Cl, researchers have found that it was a rather powerful solvent no matter its dicyanamide anion, the cation-anion pair resulted in reasonably low viscosity [62]. Nevertheless, it was found that ILs containing alkyloxy or alkyloxyalkyl groups have low viscosities and that they are beneficial for dissolving cellulose. Especially, a powerful solvent for cellulose has been found to be 1-(3, 6, 9-trioxadecyl)-3-ethylimidazolium acetate [Me(OEt)3-Et-Im][OAc] (in Table 1.6) [63].

ILs have proved to be efficient solvents for the dissolution of high amounts of cellulose in rather short time. As has been summarized in Sect. 5.2, numerous chemical derivatization reactions have been realized successfully in these novel homogeneous reaction media. However, with increasing gain of knowledge it became apparent that ILs also possess some specific disadvantages that can hamper their use as cellulose solvents in technical scales. Some of these issues are techno­logical aspects that are currently studied and can be expected to be solved in the near future; e. g., IL cost, purity, and recycling availability. Others are intrinsic restrictions related to the physical and chemical properties of ILs, i. e., limitation that cannot be avoided but only be attenuated.

5- Hydroxymethylfurfural (HMF) can be produced from poly — and monohexoses and is a valuable platform chemical that can be used to make polymers, fuels, and commodity chemicals (Fig. 8.7). HMF is the product of the dehydration of

6- carbon sugars such as fructose, glucose, and mannose. Because ketoses are furanoses when in their cyclic form, they are much easier to convert into HMF than aldoses. Polymers of six carbon sugars can also be used to produce HMF. The following procedure is generally used in the conventional production of HMF:

(1) hydrolyze polyhexoses into monomers, (2) isomerize aldoses into ketoses, and

(3) use an acid catalyst to dehydrate ketoses into HMF [100]. Once HMF is produced, it can be processed into fuels, resins, solvents, alkanes, fuel additives, or polymers as a replacement for petroleum resources. The utility of HMF as a renewable biomass based platform chemical has led to significant research in the production of HMF in ILs, although the technology has still not matured into an industrial process [101, 102].

Acid catalyzed dehydration of a ketose, such as fructose, is the easiest method for production of HMF. This method has been implemented successfully using ILs as solvents. Lansalot-Matras and Moreau demonstrated up to 80 % yield of HMF from fructose in the ionic liquids 1-butyl-3-methylimidazolium tetrafluoroborate (BMIMBF4) and 1-butyl-3-methylimidazolium hexafluorophosphate (BMIMPF6) with added DMSO to solubilize fructose. The advantage of the IL was demonstrated in a reaction without an added catalyst. An HMF yield of 36 % was obtained after 32 h in a DMSO/IL solution while only trace amounts of HMF were detected in pure DMSO after 44 h [103]. Later, the same research group demonstrated the use of the acidic IL 1-H-3-methylimidazolium chloride as a combination solvent and catalyst by producing 92 % yield of HMF from fructose and nearly quantitative amounts of HMF and glucose from sucrose. It was also noted in this article that there was no observed degradation of the HMF after the conclusion of the reaction [104]. Qi and coworkers tested sulfuric acid, HCl, phosphoric acid, acetic acid,

CuCl2, PdCl2, Dowex resin, and Amberlyst 15 as catalysts for production of HMF from fructose in BMIMCl. The Amberlyst catalyst was the best of the catalysts, producing an 83.3 % yield of HMF after only 10 min [105]. Others have taken the idea of acid catalyzed dehydration of fructose in ILs and worked to make it more environmentally friendly by using ILs made from renewable materials. Hu et al. tested a number of ILs and found that choline chloride coupled with citric acid was the most effective system for the fructose to HMF conversion, achieving over 90 % yield [106]. Acid catalyzed dehydration was tested on a number of different 6-carbon sugars by Sievers et al. through the use of added sulfuric acid in BMIMCl. Since glucose and mannose can both be isomerized into fructose, both sugars should be viable feed stocks for HMF production. As has been demonstrated previously, fructose gave high yields of HMF (>90%). Glucose only produced up to 12 % HMF yield while only very small amounts of HMF were detected when mannose was used as a substrate. Additionally, xylose, a 5-carbon sugar, was shown to undergo an analogous reaction to form furfural with up to a 13 % yield [107].

Recently, it was discovered that some metal chlorides can catalyze the conversion of aldoses such as glucose and mannose into HMF in ILs. Zhao et al. demonstrated the conversion of glucose to HMF using CrCl2 in 1-ethyl-3-methylimidazolium chloride with a yield of almost 70 %. Both CrCl2 and CrCl3 were found to be effective in this system, although CrCl2 demonstrated the highest catalytic activity [108]. This discovery has led to many studies investigating metal chloride promoted production of HMF in ILs. Pidko et al. performed work using a combination of experimental techniques and computational modeling to show that CrCl2 and CrCl3 effect the dehydration of glucose into HMF through ring opening and hydrogen shift catalyzed by CrCir ions in solution [109, 110]. Binder et al. worked to elucidate the mechanism further through the use of glucose, mannose, galactose, lactose, tagatose, psicose and sorbose along with isotopic labeling. In this study, it was demonstrated that the chromium catalyst causes a 1,2-hydride shift which leads to a furanose that can be dehydrated [111]. In these studies, the efficacy of the CrCl2/IL system could be enhanced through the use of microwave irradiation and by supporting the CrCl3. The addition of microwave irradiation instead of simple heating in an oil bath allowed HMF to be recovered with up to a 40.2 % yield after only 2.5 min while convention­ally heated reactions obtained 32.5 % yield after 32 min [112]. In addition to CrCl2 and CrCl3, it has been shown that SnCl4 can catalyze the dehydration of sugars, inulin, and starch into HMF in EMIMBF4. In this system, and HMF yield of between 40 and 65 % could be obtained, depending on the substrate [113].

While production of HMF from monosaccharides is a useful technology, by producing HMF directly from lignocellulose, the saccharification step of biomass processing would be removed. Su et al. demonstrated the use of coupled metal chlorides to produce HMF from cellulose in a single step. By using CuCl2 and CrCl2 in EMIMCl, cellulose could be converted to HMF with a yield of 55.4 % that stayed constant over several recycles of the IL/catalyst system. Additionally, the metal chlorides were shown to work in a synergistic manner, with almost no HMF production with either metal chloride on its own [114]. Binder and Raines demon­strated a similar system in which CrCl2 or CrCl3 in a solution of N, N-dimethylformamide (DMA) with LiCl and the IL EMIMCl. With this system, fructose, glucose, cellulose and lignocellulose from corn stover and pine wood could all be converted into HMF. The yields were dependent on conditions and substrate, although even the lignocellulose produced up to a 48 % yield [115]. In a study by Zhang et al. that was previously mentioned in the section on cellulose hydrolysis, the EMIMCl/water mixtures that could be used to hydrolyze cellulose were also demonstrated to be effective at HMF production when a CrCl2 catalyst was added [94]. By incorporating the use of microwave irradiation with the CrCl3/ IL system, glucose and cellulose could be converted to HMF with 90 and 60 % yields, respectively [116]. Expanding the use of microwave treatments, lignocellu — lose from corn stover, rice straw and pine wood could be converted to HMF and furfural with yields of 45-52 and 23-31 %, respectively, with CrCl3 in BMIMCl in 3 min or less [117].

Another property of ILs is hydrophobicity, which should not be confused with solvent polarity. The terms hydrophilic/hydrophobic ions are often used synony­mously with water miscibility based on the miscibility of ILs with water, so that ionic liquids can be divided into two categories: hydrophobic (water-immiscible) and hydrophilic (water-miscible). The hydrophobicity of ILs is usually quantified by the log P scale, a concept derived from the partition coefficient of ILs between 1-octanol and water. Russell’s group measured the log P values (< —2.0) of several ILs, and suggested that they are very hydrophilic in nature based on Laane’s scale [31]. They also observed that free lipase (Candida rugosa) was only active in hydrophobic [BMIM][PF6] (log P = —2.39), but inactive in other hydrophilic ILs, including [BMIM][Ac] (log P = —2.77), [BMIM][NO3] (log P = —2.90) and [BMIM][TfA] [15]. Through a systematic investigation of lipase-catalyzed transesterification in over 20 ILs, it was observed that lipase activity increased with the log P value to reach a maximum, and then decreased as log P further increased (a bell-shaped dependence). These examples implied that the high hydro — phobicity (high log P) of ILs could be beneficial for enzyme stabilization [34]. How­ever, many exceptions to this rule have been reported, and ionic liquids have not been treated according to this log P concept to relate them to enzyme activity.

The water miscibility of ILs generally depends on the anions they contain, and the solubility of water an ionic liquid can be varied by changing the anion from [Cl] to [PF6]. However, this behaviour varies widely and sometimes unpredictably; for example, [BMIM][BF4], does not dissolve simple sugars to an appreciable degree and [BMIM][Cl], in contrast, dissolves massive amounts of cellulose. And yet, these ionic liquids are of similar polarity on the Reichardt’s scale. It was demonstrated that the ability of ionic liquids to dissolve complex compounds, such as sugars and proteins, mainly depends on the H-bond accepting properties of the anion. A recent measurement of the H-bond accepting properties of such ILs revealed that [BF4] or [MeSO4] were better H-bond acceptors (в = 0.61 and 0.75, respectively) than [PF6] (в = 0.50), which can be considered a reasonable expla­nation for the difference in water miscibility [85]. The high hydrogen-bonding basicity and overall hydrophilic nature of water-miscible ILs enable to dissolve enzymes (to a greater or lesser degree) while enzymes are barely soluble in hydrophobic ILs [86]. The Sheldon group [87] maintains that hydrogen bonding could be the key to understanding the interactions of proteins and ionic liquids. Water is a powerful hydrogen bonding medium and an ionic liquid must mimic water in this respect to dissolve proteins, in particular as regards the hydrogen bond — accepting properties of the anion. The interaction should not be too strong, how­ever, because, otherwise the hydrogen bonds that maintain the structural integrity of the a-helices and в-sheets will dissociate, causing the protein to unfold. So, to maintain the activity of ionic liquid-dissolved enzymes, a balance of mild hydrogen bond-accepting and donating properties is required. In contrast, in ionic liquids that do not dissolve enzymes, the enzyme preserves its native structure in this ionic liquid just as it preserves its catalytic activity. Besides, enzyme-compatible anions exhibit lower hydrogen bond basicity, which minimizes interference with the internal hydrogen bonds of an enzyme [84]. They also exhibit lower nucleophilicity and thus a lower tendency to change the enzyme’s conformation by interacting with the positively charged sites in the enzyme structure [87].

The approximate ionic association strength in aprotic solvents is listed below in increasing order [88]:

[NTf2] < [PF6] < [СЮ4] < [SCN] < [BF4] < [TfO] < [Br] < [NO3] < [TFA] < [Cl]

This order represents the strength of an anion in its interaction with solvated cations through ionic attraction, or may even represent the strength of interactions between the anion and the changed surface of macromolecules (such as proteins). Dupont [10] suggested the strength of hydrogen-bond basicity in the similarly increasing order of:

[BPh4] < [PF6] < [BF4] < [TFA]

These sequences confirm that enzyme activity is probably related with the hydrogen-bond acceptor strength of the anion: anions with low hydrogen-bond basicity are enzyme stabilizing.

In regards to IL cations, these are usually accepted to show a lower dominant effect than anions of the same charge density, because anions are more polarizable and hydrate more strongly [43]. So, cations seem to interact indirectly via interac­tion with anions, depending on the degree of coordination and the length of the alkyl chain of the cations.

Other authors suggest that the stability of the enzyme depends on the nucleophi- licity of the anion. Kaar et al. [15] observed that free Candida rugosa lipase was only active in hydrophobic [BMIM][PF6], but inactive in all hydrophilic ILs based on [NO3], [Ac] and [TFA] during the transesterification of methylmethacrylate with 2-ethyl-1-hexanol. They indicated that the last three anions are more nucleophilic than [PF6], and thus could interact with the enzyme causing protein conformational changes. However, in another study, a contradictory result was reported. Irimescu and Kato [89] carried out the CALB-catalyzed enantioselective acylation of 1-pheny- lethylamine with 4-pentenoic acid, and found that the reaction rate relied on the type of IL anions (reaction rate in a decreasing order of [TfO] > [BF4] > [PF6], keeping the cation unchanged). This suggests higher anion nucleophilicity, correlating with higher enzymatic activity. On the other hand, Lee et al. [90] measured the initial transesterification rates of three lipases (Novozym 435, Rhizomucor miehei lipase, and Candida rugosa lipase) in different ILs with the same water activity, and observed that the anion effect on the initial rates followed a decreasing order: [NTf2] > [PF6] > [TFO] > [SbF6] — [BF4]. They suggested that [TFO] and [BF4] are more nucleophilic than [PF6], although these results could be better explained by the anion hydrophobicity of IL.

An interesting phenomenon observed by Zhao et al. [34] during the transester­ification of ethyl butyrate with 1-butanol was that when the solvents (dichloromethane or ionic liquids) and substrates were dried but the lipase was not totally dried (-3 % wt water) higher reaction rates were observed with microwaves than in a water-bath. However, when the enzyme was also dried, the differences in reaction rates became insignificant. This interesting behaviour has actually been reported in a number of papers, the authors of which maintain that in a fairly dry hydrophobic solvent and substrate environment, the enzyme particle is surrounded by (at least) one layer of water molecules. The solvent is hydrophobic so it does not strip off the water layer. In this microenvironment, the water layer near the enzyme surface has a much higher relative dielectric constant (er = 80.1 at 20 °C) than the surrounding IL. Therefore, under microwave irradiation, the enzyme surface is likely to have a higher temperature than the bulk solvent due to the superheating of the water layer.

For bioethanol production, hydrophilic ILs are effective in dissolving cellulose, but the activity of cellulases decrease significantly in their presence. Therefore, IL residues should be entirely removed after cellulose regeneration [83].

When dealing with ILs in nearly anhydrous conditions, at low water content (<2 % v/v), all the assayed water-immiscible ILs (i. e. [BMIM][PF6], [BTMA] [NTf2], etc.) were shown to be suitable reaction media for biocatalytic reactions, because water-immiscible ionic liquids are nevertheless hygroscopic, as noted above, and readily absorb a small percentage of water [91], but sufficient to maintain the active conformation of the enzyme. For example, in the case of the lipase B (Novozym 435 from Novo) catalyzed transesterification of vegetable oil for biodiesel production, the best enzyme activity and a biodiesel yield of 97-98 % was obtained with [EMIM][PF6] and [BMIM][PF6], while hydrophilic ILs were poor solvents for this methanolysis for two reasons: oil insolubility (heterogeneous system) and enzyme deactivation [22, 32]. Besides, ILs with long alkyl chains (e. g., [C16MIM][NTf2] and [C18MIM][NTf2]) have been used in a homogeneous one-phase system for lipase-catalyzed biodiesel production which this avoids direct interactions between the enzyme and pure methanol and allows for the stable reuse of lipase in the ILs. These long chain and lipophilic ILs create a non-aqueous system for oil transesterification, and at the end of the reaction, a triphasic system was formed as a result of lowering the temperature, which facilitates biodiesel extraction [27, 28].

The remarkable results obtained for enzymatic reactions in water-immiscible ILs in nearly anhydrous conditions underline the suitability of these solvents as a clear alternative for biodiesel production. In this strategy, enzymes (immobilized or not) are simply suspended in the ILs, and the resulting a mixture can be used for biocatalytic reactions. Enzymes coated with ionic liquids can also be used. For example, for citronellyl ester synthesis, immobilized lipase from C. antarctica coated with ionic liquids at a very high temperature (95 °C) in hexane and solvent-free conditions was used [11].