As described in the above two sections, ILs are promising solvents to treat plant biomass. The most important thing is the treatment capability of plant biomass under mild conditions. CDILs dissolve polysaccharides and a part of lignin at about 100 °C without any pressurization; this means that the energy-cost to treat biomass was reduced compared to other existing biomass treatment processes under heating above 150 °C and pressurization. It is very important for any industrial fields because energy-cost is directly linked to the price of the final product. The con­sumption of energy becomes of particular importance for the energy-producing industry, because the use of excess amount of energy to get comparable or less energy is meaningless. Although ILs can treat plant biomass under mild condition compared to some other methods such as kraft-pulping method, heating and long mixing time are still needed to dissolve plant biomass in ILs, too. These require­ments should further be improved to reduce energy consumption.

Some researcher use ILs for just pretreatment of plant biomass and not for dissolution of cellulose. Li and co-workers reported that enzymatic hydrolysis was significantly improved by the use of CDIL, 1-ethyl-3-methylimidazolium diethyl phosphate ([C2mim][DEP]) [66]. They investigated the effect of tempera­ture and time of the IL treatment on the hydrolysis efficiency. The pretreatment temperature was changed from 25 to 150 °C for 1 h stirring, and the hydrolysis efficiency of the pretreated wheat straw was significantly improved when the temperature was changed from 70 to 100 °C. On the other hand, the difference of the pretreatment time slightly affects the hydrolysis degree, and they reached the conclusion that only 30 min treatment was enough to accelerate the following hydrolysis. The yield of reducing sugars from wheat straw reached 54.8 % when the wheat straw was pretreated with [C2mim][DEP] at 130 °C for only 30 min. It remained only 20 % when the straw was enzymatically hydrolyzed in water for 12 h. In addition, the hydrolysis products did not show a negative effect on

S. cerevisiae fermentation. Tan and co-workers reported the IL pretreatment of palm frond after extracting the palm oil for improving conversion of cellulose into reducing sugar through subsequent enzymatic hydrolysis [67]. During the pretreatment, lignin was partly decomposed and was dissolved in [C4mim]Cl and remained in the solution after regeneration process of cellulose. In addition, hemi — cellulose was autohydrolyzed during the pretreatment. Apart from crystallinity of cellulose, cellulose digestibility should also be influenced by other factors such as DP, surface area of cellulose, as well as state of cellulose protected by lignin and hemicellulose complexes. Uju and co-workers also studied the effect of pretreatment with ILs for plant biomass [68]. They used [C4mpy]Cl as a pretreatment IL for bagasse or Eucalyptus. The pretreatment of the biomass resulted in up to eightfold increase in the enzymatic saccharification compared with the untreated biomass. At short time pretreatment, [C4mpy]Cl showed higher potential to increase the initial degree of cellulose conversion than that in [C2mim][OAc]. They suggested that the significant acceleration of enzymatic saccharification was possibly caused by the reducing of DP of cellulose by the [C4mpy]Cl pretreatment. Bahcegul and co-workers studied the correlation between the particle size of plant biomass in detail and the pretreatment efficiency with ILs for subsequent enzymatic saccharification [69]. They used cotton stalks with four different particle size pretreated in [C2mim][OAc] or [C2mim]Cl. For [C2mim]Cl, the highest glucose yield (49 %) was obtained when the biomass had the smallest particle size, while cotton stalks with larger particle size gave lower glucose yield (33 %). On the contrary, for [C2mim][OAc], the lowest glucose yield (57 %) was obtained when the cotton stalks with the smallest particle size was examined, while cotton stalks with larger particle size gave higher glucose yield (71 %). Simply consider­ing the overall surface area of the biomass particles, smaller particles gave higher glucose yield. Other unknown factor(s) should exist to affect the enzymatic saccharification. They suggested that the most suitable particle size of lignocellu — losic biomass prior to pretreatment may change depending on the IL species.

For pretreatment of lignocellulosic biomass, it is not necessary to completely dissolve cellulose but heating is still a necessary step. On the other hand, some researchers are trying to dissolve plant biomass without heating. Abe et al. found that phosphinate-type ILs dissolved plant biomass and extracted polysaccharides from plant biomass without heating [70]. Since some phosphonate type ILs as seen in Scheme 2.9 have a good ability to dissolve cellulose at ambient temperature [28], we have prepared several phosphonate type ILs and evaluated their biomass treatment ability. As a result, polysaccharide extraction degree was found to be closely related to the viscosity. This means that the IL with low viscosity had good

Scheme 2.9 Structure of alkylphosphonate type salts [70]

capacity to dissolve plant biomass within a short period of mixing time under mild condition when ILs have sufficiently high polarity. We accordingly designed a low viscosity and highly polar IL; 1-ethyl-3-methylimidazolium phosphinate (Scheme 2.10). With this IL, it became easy to extract polysaccharides rapidly from plant biomass under mild conditions (Fig. 2.8). Since this IL did not require any heating to extract polysaccharides from biomass, the energy-cost was reduced and this IL should be a promising solvent for plant biomass treatment. In addition, this IL is stable and recyclable. Thus, a closed system for biomass treatment as seen in Fig. 2.9 can be proposed.

There are a few reports about the effect of water addition on the solubility of cellulose. Padmanabhan and co-workers reported about the influence of water on the lignocellulose solubility [71]. Prior to solubility measurements, 3-5 wt% water was added to cellulose dissolving ILs, namely chloride, acetate and phosphate — based ILs. After that, powder of Miscanthus, a lignocellulosic material, was added to the ILs, and stirred the mixture at over 100 °C. However, no cellulose was extracted. This result strongly suggested that water considerably suppressed the dissolution of lignocellulose in wet ILs. They concluded that ILs should be dried well in advance to extract cellulose from biomass. Since plant biomasses also contain a certain amount of water, the biomasses have to be dried before treatment with ILs. On the other hand, as mentioned above, TBPH has a great ability to dissolve cellulose without heating even in the presence of water [47]. So, we next tried to use this novel solvent to treat wood biomass. As expected, polysaccharides such as cellulose were extracted from wood powder without heating (Table 2.10). Poplar powder was used as a wood sample, and the powder was added to reach 5 wt% against TBPH solution. In the presence of 70 wt% water, TBPH could extract polysaccharides only 4.9 % of the weight of the added poplar. On the other hand, TBPH containing 40-50 wt% water successfully extracted cellulose and other polysaccharides for 36-37 % of the weight of the added popular. These results indicate that the extraction of cellulose from powder of wood such as poplar could be carried out even in the presence of considerable amounts of water.

For the development of sustainable human societies, we have to develop new energy conversion methods based on renewable energy sources instead of fossil fuels. ILs, which dissolve renewable cellulosic biomass with low energy cost, should serve as the foundation for future development of sustainable world, espe­cially for the development of bioenergy production.

Acknowledgement Our research results mentioned here were obtained under the support of a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (No. 21225007). It was also partly supported by Japan Science and Technology Agency (JST) through the CREST program.

Research has been performed on the use of organic co-solvents in ILs. This is mainly to alter the properties of the dissolution system, such as viscosity reduction. As protic solvents like water or alcohols tend to prevent cellulose dissolution and are working as efficient non-solvents for dissolved components, the group of polar aprotic organic solvents typically will not decrease the solvation efficiency of ILs towards cellulose, in co-solvent concentrations up to 50 m% [51, 52]. The use of co-solvents may enhance the kinetics of the dissolution process, by accelerated diffusion. This allows the use of lower dissolution temperatures that in turn prevent the unwanted depolymerization reactions. Enhancement of wood dissolution kinet­ics when using co-solvents, compared to pure IL, can be notable, as demonstrated by Qu et al., who aided dissolution of milled Fir with pyridine and DMAc (as co-solvents), at the low temperature of 30 °C [53]. However, much longer dissolu­tion times were needed than for the typical high temperature dissolution. Co-solvents can also enhance wood dissolution at higher temperatures. An article by Xie et al. has demonstrated that complete dissolution of corn stover can be achieved using NMP: [emim][OAc] solutions at the higher temperature of 140 °C, in under 60 min [54]. This of course is not designed for material production, where molecular weights are maintained, but rather biofuel production where maintaining molecular weights is not critical. The majority of the available co-solvents will eventually turn into non-solvent when a limiting concentration is reached [51] and so it may even be possible to maintain binary solvent mixtures with ILs throughout the process. In fractionation processes aiming at the manufacture of derivatized products, certain co-solvents can also act as catalysts for subsequent modification reactions without need for product isolation, in between the unit processes.

Water in the IL systems may also be termed as a limiting solvent, instead of a non — or co-solvent. It has been used as a way of limiting cellulose solubility in certain ILs, while close to complete delignification with removal of hemicelluloses can still take place. The presence of acidic species have been stated to be essential for delignification in these kinds of systems and they can be added as catalysts [55] or originate from the natural acidity of the IL [56]. Depending on the anion of IL, the aqueous solutions can be relatively acidic [55, 57]. It remains uncertain whether this is related to impurities specific to pure ILs, technical preparations of ILs, as a natural property of IL-water solutions [57], or from reactions leading to acidic products. Zhang et al. have reported that from a neutral pH of the pretreatment solvent down to pH 3.4, all of the wood components are regenerating close to their natural compositions from the aqueous IL-system, without resulting in delignified pulp [55]. This is in slight contradiction with the results from Fu et al., who have used neutral aqueous solutions to basic [emim] [OAc] efficiently, without any added acid catalysts [58]. The fibrous structure of wood cells still exists in the solid cellulose enriched fractions afforded by treatments in aqueous-IL solutions [55], resembling traditional chemical pulps. This is due to the inability of ILs to dissolve crystalline cellulose, once high enough water contents are added. Thus, only the amorphous parts of the fiber are accessible to the acidic solvent.

The efficiency of the IL-water solvent system was highly dependent on the type of treated biomass as grass-type feedstocks, such as Miscanthus or Triticale, were found to be highly responsive. Nearly complete delignification and glucose digest­ibility are observed for grasses, followed by mediocre efficiencies for hardwoods and significantly lower response for softwoods [55, 56].

Compared with organic and aqueous solvents, the application of ILs for the dehydration of carbohydrates can significantly reduce the reaction temperature [20, 32]. Traditional synthetic methods for 5-HMF production from carbohydrates are normally carried out at temperatures ranging from 150 to 300 °C [4, 8, 15, 18], but the reaction temperatures required for the process in ionic liquids could be reduced to 80-150 °C [19, 41,45, 47, 85], or even room temperature in some cases [34]. Normally the conversion of fructose needs a relative low temperature ranging from 80-120 °C since it is readily converted into 5-HMF through elimination of three molecules of water. Glucose is more difficult than fructose to be transformed into 5-HMF since it tends to form a stable six-membered pyranoside structure that has a low enolization rate, and its dehydration is generally carried out at 100-140 °C [19, 41, 45, 47]. The conversion of cellulose is more difficult for efficient conversion to 5-HMF than monosaccharides and requires high tempera­tures of about 150 °C in the presence of catalysts [19, 47, 53]. In general, lower temperatures lead to low 5-HMF yield (ca. 10-20 %), whereas higher temperatures promote formation of side-products and affects the 5-HMF yield so that an opti­mum temperature exists.

The reaction temperature is an important parameter for carbohydrate conver­sions since lower temperatures allow one to reduce the energy requirements. Chan et al. [84] was able to lower the dehydration reaction temperature to below 50 °C by using a system containing [BMIM][Cl] and metal salts. Chloride salts of zirconium (IV), titanium (IV), ruthenium (III), and tungsten (IV) or tungsten (VI), the most efficient chloride salt was that of tungsten (VI) that gave a 5-HMF yield of 63 % at 50 °C.

Reactions that can be promoted at ambient conditions are considered as one of the key goals among the 12 principles of green chemistry [92, 93]. Remarkably, it was found that efficient dehydration of fructose to 5-HMF could be carried out at room temperature in ionic liquids provided that the ionic liquid was brought about its melting point and the substrate was pre-dissolved in the ionic liquid before cooling to room temperature. Qi et al. [34] developed a green catalytic system for the production of 5-HMF from fructose catalyzed by a strong cation exchange resin, by the addition of different cosolvents such as DMSO, acetone, methanol, ethanol, ethyl acetate, and supercritical carbon dioxide to the ionic liquid 1-butyl-3- methylimidazolium chloride ([BMIM][Cl]). In a typical reaction, fructose was first dissolved in [BMIM][Cl] in a water bath at 80 °C for 20 min. After the mixture was cooled down to room temperature, the solution appeared gel-like with a very high viscosity. Subsequently, the catalyst (Amberlyst-15 sulfonic ion-exchange resin) and some amount of co-solvent were added for viscosity reduction and the reaction proceeded smoothly at 25 °C. Through addition of a co-solvent, the viscosities could be greatly reduced from an estimated value of 6,800 mPa s to values of around 2,000 mPa s, with the best results being obtained for acetone (1,850 mPa s) and ethyl acetate (1,930 mPa s). Interestingly, a gaseous co-solvent, such as CO2 or supercritical CO2 (>31 °C) was tried and found to provide comparable results to the organic solvents. Thus, use of CO2 can possibly provide viscosity reduction and make it simple to regenerate the solvent system. Reductions in viscosity allowed the transformations to be carried out at close to room temper­ature. For this case [34], 5-HMF was obtained at yields of 78-82 %. The time for reaction was longer than in the previous work (6 h vs. 10 min) [20], but this method has the advantage of being performed at room temperature.

Strong hydrogen-bonding basicity ф-value in KAT values) is now recognized as the most important property of ILs with high cellulose dissolution. Viscosity of ILs is the second key factor causing cellulose dissolution at low temperature conditions. However, a rational design of ILs with a dissolution property of cellulose has not yet been established. We discuss in this chapter how to accomplish the design of ILs with high cellulose dissolution from the standpoint of nature.

Remsing et al. [23] reported based on their 13C and 35/37Cl NMR studies that there was a stoichiometric interaction between the chloride anion and the cellulose hydroxyl groups, and this might be the key driving force of cellulose dissolution in this IL (Fig. 4.3) [23].

Inspired by their result, we hypothesized that enhanced interaction of a certain anion or cation of ILs between hydroxyl groups in the cellulose molecule might be the key factor causing cellulose dissolution and we might be able to obtain a hint on how to design such anion or cation from nature. Focusing on the structure of hydrolyzing enzyme of cellulose (cellulase), we found that amino acid ILs were strongly capable of dissolving cellulose: N, N-diethyl, N-methyl, N-(2-methoxy) ethylammonium alanate ([N221(ME)][Ala]) worked as an excellent solvent for cel­lulose dissolution among ILs whose anion part was natural amino acid [24].

Hydrolysis of solid cellulose is achieved by cellulases such as endoglucanase (EGs) and cellobiohydrolases (CBHs) [25, 26]. The former can hydrolyze internal p-1,4-glycoside bonds in a cellulose polymer in the amorphous regions within the cellulose micro-fibril, and the latter can act on the free ends of cellulose polymer chains. Both types of cellulases have cellulose-binding modules that facilitate their

adsorption onto crystalline cellulose, bringing the catalytic domains physically close to their site of action (Fig. 4.4) [25, 26]. We looked at what the protein sequences of several cellulases were causing particularly in the area of substrate­binding cleft, and recognized that glucosyl-binding sites of cellulases were fre­quently formed by the exposed surface of aromatic side-chains of protein residues [25, 26]. Three of the four binding sites making up the enclosed cellulose-binding tunnel reportedly contain the tryptophan (W), asparagine (N), and isoleucine

(I) residue side-chains for Trichoderma reesei Cel6A (CBH II) [25, 26]. Therefore, it was expected that ILs made from amino acids might have an affinity toward a certain part of cellulose.

Ohno and co-workers prepared ILs that contained amino acid moieties as anion parts [21,27]. Since the hydrogen-bonding basicity of amino acid salts was reported to be high [27], amino acid ionic liquids are expected to display good cellulose dissolving ability.

Based on these results, the dissolving property of 1-butyl-3-methylimidazolium tryptophan ([C4mim][Trp]) against cellulose was tested using microcrystalline
cellulose (Avicel®) as a model compound. However, the cellulose did not dissolve at all in this IL. Further evaluation of tryptophan salts with ammonium, phospho — nium, or pyrridinium cation, revealed that choice of cation was also a key point in designing an IL with cellulose dissolution capability: N-(2-methoxyethyl),N, N — diethyl, N-methylammonium tryptophan ([N221(ME)][Trp]) dissolved cellulose (5 wt% vs. IL) at 100 °C [24]. Encouraged by the results, we prepared [N221(ME)] salts with natural amino acids and carefully evaluated their cellulose dissolution properties against the model cellulose (Avicel) (Table 4.2). Among 20 types of amino acid salts, we found that [N221(ME)][Ala] worked best to dissolve cellulose with 12 wt% versus solvent: the second solvent most effective was lysine salt ([N221 (ME)][Lys]) (11 wt%) and the third was ornitin salt ([N221(ME)][Orn]). Threonine ([N221(ME)][Thr]) and isoleucine ([N221(ME)][Ile]) salts also showed similar solubil­ity against the cellulose (7 wt%) [24]. We expected that amino acid might have affinity with a certain part of cellulose and cause its dissolution in the amino acid ionic liquid. The results reached our expectations though the details were slightly different; one of these we had anticipated, however, because there was no alanine residue near the entrance part of the cellulases [25, 26].

Many amino acid ILs dissolved cellulose, except for glutamic acid salt and all amino acids had high p-values [27]. Hence, we fixed the anionic part to alanin, and the cationic portion was re-evaluated (Fig. 4.5). It was thus confirmed that cellulose solubility was strongly dependent on the cationic part. High cellulose solubility was recorded for N, N-bis(2-methoxyethyl),N-ethyl-N-ethylammonium ([N22(ME)2]), N, N, N-tris(2-methoxyethyl),N-ethylammonium ([N2(ME)3]), N-(2-thiomethoxyethyl), N, N-diethyl, N-methylammonium ([N221(MTE)]), N-methyl-N-ethoxyethylpyr-

rolidium ([P1(ME)]) salts, in a range of 12 to 11 wt%. On the contrary, no dissolution of cellulose took place in [P444ME][Ala] or [PyME][Ala] salt. Interestingly, the presence of the methoxyethyl group on the ammonium cationic part strongly modified cellulose solubility: better dissolution was obtained for [N221ME][Ala], while poor solubility was obtained for N-butyl-N, N-diethyl, N-methylammonium alanine([N4221][Ala]). However, both the methoxyethoxymethyl substituted salt ([N221(MEM)]) and N, N,N, N-tetra(methoxyethyl)ammonium ([N(ME)4]) alanine showed poor cellulose solubility (Fig. 4.5). These results clearly indicate that cellulose solubility is determined not only by the physical characteristics of the solvent shown as KAT values, but also by the affinity of a certain group interaction of ionic liquids with cellulose might be an important factor of cellulose dissolution in the ILs.

We next investigated cellulose solubility against various ILs which have [N221 (ME)] as a cationic part. Although 10 wt% of cellulose dissolves in [N221(ME)]Cl, this salt requires a higher temperature (over 120 °C) and a longer mixing time over [N221(ME)][Ala]. Slight decomposition of IL was observed under the conditions used [24]. No dissolution of cellulose was observed when [N221(ME)]Br, hexafluor- ophosphate (PF6), N, N-bis(trifluoromethyl)sulfonylamide (NTf2), 2,2,3,3,4,4,5,5- octafluoropentyl sulfate (C5H8), or 2-aminoethylsulfonate (taurine), was used as solvent. [N221(ME)] salts with hydrogen oxide also showed poor dissolution prop­erties [24]. Poor solubility was also recorded for N, N-dimethylalaine or

N-Boc-alanine salts compared to the alanine salt (Fig. 4.6). The presence of amino group might play an important role in dissolving cellulose [24]. Therefore, we anticipate that the amino group of [Ala] may interact with a certain part of cellulose and contribute to breaking its hydrogen bond network. Liquid ammonia reportedly changes the crystalline phase of naturally occurring cellulose and dissolves it by allowing the ammonia molecules to penetrate the cellulose fibril [28, 29]. From these results, we assume that the amino group interposed the hydrogen bonding between the cellulose and caused dissolution of cellulose in amino acid ILs. However, since the cellulose solubility is also modified by the cationic part of the IL, cation might play an important cooperative role in the mechanism for cellulose dissolution, although its origin is still unclear.

Addition of an anti-solvent like water or ethanol to the cellulose/IL solution causes precipitation of the dissolved cellulose and the structure of the regenerated cellulose changes to a disordered form. Pretreatment of cellulose increases the surface area accessible to water and cellulases are believed to improve the hydro­lysis rate [30, 31]. Therefore, many researchers have attempted to hydrolyze regenerated cellulose in order to improve the hydrolysis rate by a cellulase; the dissolution of microcrystalline cellulose with [C4mim]Cl and the rapid
precipitation with water induced an increase of the amorphous region in the regenerated cellulose and enhanced the initial enzymatic hydrolysis rate [30, 31].

It was reported that cellulose regenerated from ionic liquid solution showed Type II crystalline form [31]. In fact, we confirmed that the regenerated cellulose from [N221(ME>][Ala] solution had only Type II form [24]. The X-ray diffraction patterns of the microcrystalline cellulose film (Avicel®) and that of the regenerated one were compared: the regenerated cellulose exhibited the typical diffraction patterns of Type II cellulose at 20 = 20.16° and 21.76° [32]. The results indicate that the transformation from Type I to Type II occurred after the dissolution and regeneration in [N221(ME)][Ala] [24].

Interestingly, we further found that crystal form of the regenerated cellulose was dependent on the dissolution solvent [24, 33]: 7 wt% of cellulose was dissolved in [N221(ME)][(MeO)(H)PO2] and the regenerated cellulose was a mixture of Type I and II.

Mizuno et al. [33] recently reported that the disordered chain region was increased in the order of [N221(ME)][Ala] < [C2mim][OAc] < [C2mim] [(EtO)2PO2] < [C2mim]Cl [33]: regenerated cellulose treated with [C2mim]Cl contained larger amorphous regions than the others. On the contrary, that of [N221 (ME)][Ala] had a larger amount of cellulose II crystalline structure and less suscep­tibility to enzymatic degradation than others; this suggests that the enzymatic hydrolysis rate of regenerated cellulose should increase by the same order. On the other hand, the order of degree of polymerization of the cellulose was [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 fiber.

The purification of biodiesel is an essential process towards the production of high quality fuel. The main task of which is to remove glycerol and residual catalyst. Glycerol has low solubility in FAMEs and can be separated by settling or centrifugation. The presence of residual glycerides can cause deposition of bio­diesel in internal combustion engine injectors (carbon residue) [78]. In addition, residual glycerol can initiate settling problems in the engine and, on the long term, affect human or animal health by the emission of hazardous acrolein into the environment. The presence of catalysts in biodiesel can form deposits (carbon residue) in fuel injection system, poison the emission control system, and weaken the engine [79].

Ionic liquids’ abbreviation and lull name: [С;.тіт][ТГО]: l-Ethyl-3-methylimidazolium trifluoromethanesulfonate; [C4mim][NTf;.]: l-/)-butyl-3-methylimidazolium N-bistrifluoromethanesulfonyl)imidate; [C;.mim][BF4]: l-ethyl-3-methylimidazolium tetralluoroborate; [C4mim][BF4]: l-butyl-3-methylimidazolium

tetralluoroborate; [Ci8mim][NTf4]: l-methyl-3-octadecylimidazolium te(trifluoromethylsulfonyl)imide; [Ci6mim][NTlV]: l-hexadecyl-3-methylimidazolium bis (trilluoromethylsulfonyl)imide; [C4mim][PF6]: l-butyl-3-methylimidazolium hexalluorophosphate; [C8mPy][BF4]: l-octyl-3-methyl-pyrdininium tetralluoroborate

As glycerol and methanol are highly soluble in water, water washing was widely used to remove excess contaminations (e. g. glycerol, alcohols, residual metal salts, soaps, fatty acids). However, the presence of water brings many disadvantages, including increased costs and production time, and generation of waste water [80]. Traditionally, several other methods have been used to remove glycerol from biodiesel, such as adsorption over silica, membrane reactors, and the addition of lime and phosphoric acid. Yet, technical problems remain for biodiesel produc­tion at an industrial scale [81]. To develop better process for byproduct removal, some classes of deep eutectic solvents based on mixtures of quaternary ammonium salts and compounds with hydrogen bond-donating group, have been applied in biodiesel production from rapeseed and soybean oil [82]. Deep eutectic solvents are inexpensive, non-toxic, and environmentally benign. While pure quaternary ammo­nium salts alone were inefficient, the quaternary ammonium salt-glycerol mixture solvents were successful for extraction of glycerol from biodiesel production mixtures, and a glycerol/salt molar ratio of 1:1 was found most effective. Of those salts studied, choline chloride, ClEtNMe3Cl and EtNH3Cl showed high efficiency for glycerol removal.

Deep eutectic solvents have also been used to extract glycerol from palm oil-based biodiesel production in order to meet the EN 14214, and ASTM D6751 standards. The extraction process involved different compositions of a quaternary ammonium salt to glycerol as the solvent, and a ratio of 1:1 was found most efficient. Moreover, the ratio of biodiesel to deep eutectic solvent was more important than the ratio of quaternary ammonium salt to glycerol. The used solvent can be recovered by crystallization [83].

Deep eutectic solvents based on methyl triphenyl phosphonium bromide and different hydrogen bond donors (e. g. glycerol, ethylene glycol, and triethylene glycol) were employed to remove glycerol from palm-oil-based biodiesel [84]. It was found that the solvents including ethylene glycol or triethylene glycol were successful in removing free glycerol to below the ASTM standards. These sol­vents were able to reduce the content of monoacylglycerides (MGs) and diacylglycerides (DGs), but DGs were removed more effectively than MGs. Choline chloride and methyltriphenylphosphonium bromide based deep eutectic solvents could also be used to remove residual KOH efficiently from palm oil-based biodiesel [85].

Understanding the mechanism of how ionic liquids stabilize and activate enzymes is crucial for researchers and engineers to optimize enzymatic reactions as well as synthesize the enzyme compatible ionic liquids. Many efforts with different tech­niques such as spectroscopy, molecular dynamics simulation have been paid to explore the dynamic structure and conformation of protein in ionic liquid in recent years [109—121]. In general, the ionic liquids that have strong interaction with protein such as halide containing ionic liquids tend to change the conformation of proteins and therefore inactivate enzymes while other ionic liquids such as [Tf2N]_ based ionic liquids strengthen the protein conformation resulting in enhanced enzyme stability. For example, Sasmal et al. by using fluorescence correlation spectroscopy studied the conformation dynamics of human serum albumin (HSA) protein in [Pmim][Br] and observed the denaturant effect of ionic liquids [121]. De Diego et al. used fluorescence and circular dichroism (CD) spectroscopy to analysis the a-chymotrypsin stabilization of [Emim][Tf2N] and found out that this ionic liquids shows ability to compact the native conformation of protein by great enhancement of the p-strand of protein. In our recent studies, molecular dynamics (MD) simulation was employed to investigate the structure of CALB enzyme in different ionic liquids and organic solvents and their corresponding enzyme activ­ities. The MD simulations indicate that the structure and dynamics of the cavity that holds the catalytic triad are solvent dependent: the cavity can be opened or closed in water; the cavity is open in [Bmim][TfO] and tert-butanol; the cavity is closed in [Bmim][Cl]. The closed or narrow cavity conformation observed in our simulations obstructs passage for substrates, thus lowering their probability of reaching the catalytic triad (Fig. 10.2). In addition, we observed that two isoleucines, ILE-189 and ILE-285, play a pivotal role in the open-close dynamics of cavity. Specifically, ILE-285 situated on a helix (a-10) that can significantly change conformation in different solvents. This change is acutely evident in [Bmim][Cl] where interactions of LYS-290 with chlorine anions induces a conformational switch from an a-helix into a turn (Fig. 10.3). Disruption of the a-10 helix structure results in a narrower entrance to the catalytic triad site and this change is responsible for reduced activity of CALB in [Bmim][Cl]. Moreover, the cavity profile’s size is well agreed with the enzyme activity for the synthesis of butyl acetate. The activity of the enzyme decreases with the size of the cavity in the following order: [Bmim][TfO] > tert — butanol > [Bmim][Cl].

Compared with organic solvents, most ILs have relatively high thermal stability. The decomposition temperatures reported in the open literature are generally >200 °C, and they are liquid state in a wide range of temperatures (from 70 to 300-400 °C). The decomposition temperatures of ILs that dissolving cellulose are listed in Table 1.4.

Many literature works have investigated the thermal stability of ILs on imidazolium and anion structures. The onset of thermal decomposition is similar for the different cations but appears to decrease as the anion hydrophilicity increases. Ngo et al. [73] found that the thermal stability of the imidazolium- based ILs increases with increasing linear alkyl substitution. Owing to the facile elimination of the stabilized alkyl cations, the presence of nitrogen substituted secondary alkyl groups decreases the thermal stability of ILs. They also found that the stability dependence on the anion is [PF6]~ > [Tf2N]_ ~ [BF4]~ > halides. Fox et al. [74] found that the alkyl chain length does not have a large effect on the thermal stability of the ILs.

However, the thermal stability of ILs has been revised [4]. The range of thermal stability of ILs published in the open literature is overstated. The decomposition temperature of ILs calculated from fast thermo gravimetric analysis (TGA) scans in a protective atmosphere and does not imply a long-term thermal stability below those temperatures [44]. Fox and his group have done some nice study on the thermal stability of ILs [74—76]. Compared with the data from both isothermal and constant ramp rate programs for the decomposition of 1-butyl-2,

3- dimethylimidazolium tetrafluoroborate ([BMMIM][BF4]) under N2, they found that isothermal TGA experiments may be the more appropriate method for evalu­ating the thermal stabilities of ILs [75]. Based on TGA pyrolysis data of 1, 2,

3- trialkylimidazolium room temperature ILs, They also found that although the calculated onset temperatures were above 350 °C, significant decomposition does occur 100 °C or more below these temperatures.

Singh et al. [77] analyzed the thermal stability of imidazolium based ILs [BMIM][PF6] in a confined geometry. They found that [BMIM][PF6] in confined geometry starts at an earlier temperature than that for the unconfined ILs. The loss of alkyl chain end groups of [BMIM] cation of ILs assign to the early decomposi­tion by using a phenomenological ‘hinged spring model’. The idea of ‘hinged spring’ model is that the imidazolium ring is supposed to be ‘hinged’ to the SiO2 matrix pore walls by surface oxygen interacting with the C-H group of the imidazolium ring.

Researchers have reported a new method to study the changes that occur during thermal aging of ILs. The method is potentiometric titration, which is precise, low-cost and quick analytically. To a small extent, they found that imidazolium salts start to decompose at much lower temperatures than those obtained from thermo gravimetric analysis by using this method. They also concluded that the stability of ILs is also influenced by water, except by their composition, such as anion type and alkyl substituent at the imidazolium ring. For instance, 2 wt% of water in ILs could bring about increased degradation of [BMIM]Cl at 140 °C. Furthermore, [BMIM] [BF4], [EMIM][CH3SO3], [BMIM] [CH3SO3] and [BMIM][Tf2N] are completely stable at 140 °C for 10 days [78]. Finally, from these data, long-term stability of ILs is a complicated problem with obvious and serious implications for their use as solvents media of chemical reactions.

ILs are considered as non-derivatizing cellulose solvents, i. e., the dissolution of cellulose is not due to chemical conversion of the polysaccharide [13]. Nevertheless, ILs are not necessarily chemically inert. Both cation and anion can participate in the course of chemical derivatization reactions of cellulose or react with the dissolved polysaccharide. In addition, the effect of typical IL impurities needs to be considered.

The proton at position C-2 of 1,3-dialkylimidazolium based ILs is rather acidic; the pKa-values are estimated to be about 21-24, and can be abstracted with bases to yield N-heterocyclic singlet carbenes [83, 84]. These reactive species act as nucle­ophilic intermediates in the catalytic cycles of many organic reaction, which is the reason for surprisingly high yields and/or unexpected products frequently observed for reactions performed in ILs [85, 86]. The presence of carbenes and their influence on the derivatization of cellulose in ILs needs to be considered, in particular when bases are utilized. In addition, it has been reported for low-molecular weight cellulose mimics and later on for cellulose as well that the carbene species attach to the reducing end-group in its open-chain aldehyde form (Fig. 5.6) [87, 88]. Although the conversion is accelerated in the presence of bases, it also occurs upon dissolution of cellulose in pure imidazolium acetates. To a

R: H orCH

Fig. 5.6 Reaction scheme for side reactions observed during dissolution and chemical derivati­zation of cellulose in typical imidazolium based ionic liquids (A+X_) [67, 87, 88, 91] certain extent, these particular ILs undergo self-deprotonation as result of the relatively high basicity of their anion [89]. The equilibrium is further shifted by subsequent reaction of the carbenes formed with the cellulose end-group. To avoid the effect of reactive carbene intermediates, imidazolium based ILs that were
methylated at C-2 have been proposed as cellulose solvents [14]. However, the methyl group can also be deprotonated to a certain extend [90].

In addition to the IL’s cation, the anion can undergo specific side reactions. The conversion of cellulose, dissolved in an imidazolium chloride IL, with furoyl, tosyl, and trityl chloride as well as with SO3 complexes yields the expected cellulose derivatives 8, 16, 19, and 15 (see Table 5.1) [33, 35]. In contrast, only acetylated products could be obtained when the same derivatization reactions were carried out in an EMIMAc [33, 67]. This unexpected finding has been attributed to the formation of mixed anhydrides of the anion, which is present in high concentrations and not surrounded by a solvent cage, with the reagents applied (Fig. 5.6). These intermediates subsequently react with cellulose and transfer the acetyl group. If well understood, the chemical reactivity of ILs is not necessarily a drawback of this class of cellulose solvents. As an example, the acetate anion acts as a catalyst for the ring opening of oxiranes, which could be exploited for the efficient hydroxyalkylation of cellulose in ILs [37]. Moreover, ILs can become valuable for performing derivatization reactions in which other homogeneous cellulose reaction media, such as DMA/LiCl and DMSO/TBAF, show specific side reactions and cannot be employed.

ILs often contain certain impurities, derived from the synthesis, such as unreacted educts, side products, inorganic salts, and organic acids [92]. These compounds can affect the dissolution and chemical derivatization of cellulose in ILs. N-Methylimidazole is a starting material for the synthesis of imidazolium salts, the most frequently applied type of ILs in cellulose research, and one of their major impurities. This heterocyclic base acts as catalyst, e. g., for the silyation of cellulose. Thus, highly silylated products could be obtained using common reagent grade ILs (90-95 % purity) that contain traces of N-methylimidazole (0.1-0.5 wt%), whereas no significant derivatization could be achieved when ILs of high purity (>99 %) were applied as reaction media [41].

Even hydrophobic ILs can adsorb rather high amounts of moisture from humid air atmosphere [93]. Thus, the ubiquitous presence of water should not be neglected when using ILs for processing of cellulose. However, handling under protecting gas and strictly anhydrous conditions is not necessary for most applications. Water can directly influence chemical derivatization reactions. Most reagents applied for chemical modification of cellulose are prone to hydrolysis, which leads to an apparent decrease in reaction efficiency. Water also promotes the chain degradation of cellulose especially at high temperatures or if the ILs applied contain acidic impurities [94, 95].

In addition to the influence of water on chemical reactions, water affects the solubility and state of dissolution of cellulose in ILs. In large excess, the protic non-solvent acts as ‘precipitation agent’, i. e., cellulose is regenerated from ILs upon pouring the solution into five to ten times the volume of water. It has been reported that cellulose solutions in some ILs tolerate rather high amounts of water of up to 20 wt% [96]. However, even traces of water can alter the state of dissolution of cellulose in ILs before any ‘macroscopic changes’ can be detected. It has been demonstrated that the intrinsic viscosity of cellulose/EMIMAc solutions, which is directly correlated with the size and conformation of the dissolved polysaccharide chains, first increases with increasing water content up to a maximum of 10 wt% and then decreases again until finally reaching the solubility limit [97]. Based on this finding it was concluded that a ‘micro-gel’, i. e., agglomerates of polymer coils, is formed upon the addition of water to cellulose/IL solutions. This phenomenon has significant influence on the rheological flow behavior of these solutions and might also affect the chemical derivatization of cellulose as well as the processing into cellulosic fibers by spinning processes.

Pure ILs are commonly regarded as highly thermostable with a broad liquid range; some individual representatives of this class withstand temperatures up to 400 °C [10]. Under practical lab-conditions, however, decomposition may already occur at much lower temperatures, in particular in the presence of impurities [98]. For common cellulose/IL solutions, onset temperatures (ron) for the chemical decompo­sition and liberation of gaseous compounds around 180-220 °C have been observed by means of differential scanning — and reaction calorimetry [99, 100]. The values changed only slightly upon the addition of additives such as silver or charcoal. In contrast, cellulose solutions in NMMO, employed for the production of cellulosic fibers on a technical scale, are significantly less stable (Ton « 130-160 °C) [101, 102]. Moreover, stabilizers are required in order to prevent autocatalytic thermal runaway reactions. Usually, dissolution, shaping, and chemical derivatization of cellulose in ILs is performed below 130 °C, i. e., cellulose/IL solutions are safe to handle at typical processing temperatures. However, it has been noted that the thermostability is reduced significantly when using recycled ILs. This is an indication that already below Ton, degradation products are formed. The thermal decomposition of imidazolium-based ILs, which proceeds by a dealkylation mechanisms inverse to the synthesis, yields 1-alkylimidazoles (Fig. 5.6) [103, 104]. These primary products can further decompose, e. g., into imidazole, and/or condensate with other fragments formed [91]. These compounds are highly basic and can significantly affect chemical derivatization of cellulose in ILs. Although they might initially be formed only in small amounts, these heterocyclic degradation products cannot be removed simply by evaporation. Thus, they might accumulate during multiple recycling sequences.

Biopolymers are an important area of research as a replacement for petroleum derived products. In the US, 331 million barrels, or 4.6 % of the total US petroleum consumption, were used to make polymers in 2006 (329 million as feedstock, 2 million for energy) [119]. By displacing the need for petroleum, production of biopolymers could lessen global oil demand. While cellulose, starches, and other naturally occurring biopolymers can be difficult to work with due to their chemical and physical properties, chemical modification of these naturally occurring poly­mers allows for a wide range of properties to be achieved [120]. See Table 8.1 for representative examples. Additionally, modification of biopolymers can be used to aid in analytical methods by making otherwise insoluble polymers such as cellulose soluble in a wide range of solvents [130]. Smaller molecules of interest can also be manufactured through the chemical modification of monosaccharides. In many cases, it is even possible to couple the use of ILs and enzymes to effect a biocatalytic change while maintaining the advantages of an IL system [131]. Because ILs have the ability to solubilize unmodified biomass, they are

Table 8.1 Biomass modification reactions in ionic liquids

Cellulose——— OH

well suited to the task of chemical modification of lignocellulose for a wide variety of applications. ILs are particularly well suited to chemical modification of carbo­hydrates because, generally, the sugars or biopolymers are more hydrophilic while the reactants for derivatization are more hydrophobic. The ILs are often able to solubilize both reactants and, in the case of smaller molecules, the amphiphilic product. Research into modification of biomass in ILs is also important because there is significant work focusing on using enzymes in IL based systems, which could lead to other enzymatic processing of biomass in ILs.

The modification of monosaccharides with laurates is a common method for the production of surfactants. Glucose modification with vinyl laurates in ILs has been studied by Lee et al. using 1-butyl-3-methylimidazolium trifluoromethanesulfonate (BMIMTfO) and 1-butyl-3-methylimidazolium bis(trifluorosulfonyl)imide (BMIMTf2N). Using lipase enzymes, Lee and coworkers demonstrated that a super saturated solution of glucose in a mixture of BMIMTfO and BMIMTf2N, along with ultrasound treatment produces a better conversion and yield in less time than subsaturation solutions in a pure IL without the ultrasound [121—123]. This reaction, producing sugar esters with lipase enzymes, has been demonstrated in other ILs, such as BMIMBF4 and BMIMPF6 [124].

Modification of larger saccharide chains requires ILs that are better suited to biomass solvation. Unfortunately, the ILs that solvate cellulose and lignocellulose are also destructive to enzymes [132]. Consequently, most of the research in the modification of cellulose and lignocellulose uses non-enzymatic catalysts or no catalysts at all. Success has been seen in acetylation, carbanilation, sulfation,
succinilation (Table 8.1), and benzoylation of cellulose chains in imidazolium based chlorides and bromides along with choline chloride ILs [125—129]. There has been work in combining the enzymes and ILs that will dissolve biomass. Zhao and coworkers developed ILs with Me(EtO)n— substituted imidazolium and alkylammonium (where n = 2-7) cations coupled with an acetate anion that are both effective for solubilizing cellulose and as a solvent for lipase catalyzed esterification. These ILs can dissolve cellulose up to 10 wt%, and support enzy­matic esterification with methylmethacrylate with yield up to 66 % [71].

ILs form a strong ionic matrix and the added enzyme molecules could be considered as being included rather than dissolved in the media, meaning that ILs should be regarded as liquid enzyme immobilization supports, rather than reaction media, since they enable the enzyme-IL system to be reused in consecutive operation cycles

[96] . Finally, after an enzymatic transformation process in ILs, products can usually be recovered by liquid-liquid extraction, although the organic solvents used in this step represent a clear breakdown point for the integral green design of any chemical process.

To understand the biocompatibility of ionic liquids with enzymes it is necessary to consider the structure-function relationships of enzymes in water immiscible ILs and to discern how water is partitioned between the enzyme surface and the bulk IL solution. Complementary spectroscopy measurements (e. g. fluorescence, circular dichroism, FTIR) have classically been used to investigate changes in the secondary structure of enzymes in an attempt to explain the stabilization or denaturation phenomena associated with their molecular environment. Such spectroscopic methods have been used to correlate changes in the secondary structure of monellin

[97] , CALB [98, 99] or a-chymotrypsin [100] with enzyme stability in ILs. Iborra’s group were pioneers in carrying out structural studies that revealed that the syn­thetic activity and stability exhibited by CALB in ILs was much higher than that observed in hexane, and was related with the associated conformational changes that take place in the native structure of CALB, as demonstrated by fluorescence and CD spectroscopic techniques [98]. The stabilization of CALB by hydrophobic ILs seems to be related with the observed evolution of a-helix to в-sheet secondary structures of the enzyme, resulting in a more compact enzyme conformation, that is able to exhibit high catalytic activity, suggesting that the stability of enzyme in this medium was improved by the formation of a compact, but flexible, native-like conformation of the enzyme. Turner et al. [16] described how the deactivation of the enzyme cellulase produced by water-miscible ILs (e. g., [BMIM][Cl]) is accom­panied by a fall in the fluorescence intensity maxima of the Trp parameter with respect to the native conformation in water as a result of the enhancing exposure of Trp residues to the bulk solvent and enzyme denaturation. Fluorescence spectros­copy demonstrated that monellin in a low water content (2 % v/v) in [BMpy][NTf2] resisted thermal unfolding. Fujita et al. [101] elucidated the power of hydrated [Choline][H2PO4] to maintain the activity of cytochrome c after 18 months of storage in the dissolved form at room temperature because of its ability to maintain its native secondary structure and conformation, as monitored by ATR-FTIR (attenuated total reflection Fourier transform infrared) and resonance Raman spec­troscopies. A later study found that CALB aggregates can deactivate in 1-ethyl-3- methylimidazolium-based ILs in an anion-dependent manner [42]. Studies of papain in 15 % (v/v) aqueous solutions of 1-alkyl-3-methylimidazolium-based ILs using ATR-FTIR demonstrated that the choice of anion has a significant impact

on the structure, specificity and stability of the enzyme [102]. Again, the в-sheet content in the secondary structure increased, while the a-helical content decreased.

Micaeio and Soares [103] presented a molecular dynamics simulation study of the serine protease, cutinase, in two different ILs, [BMIM][PF6] and [BMIM] [NO3]. Their work showed that the enzyme is preferentially stabilized in [BMIM] [PF6], which allows a suitable degree of hydration to be maintained at the enzyme surface and hence renders a more native-like enzyme structure, while [BMIM] [NO3] tended to be more destabilizing. These findings are in accordance with previous experimental observations [15, 86] which attributed these results to the difference in the hydrophobicity of the two ILs: [BMIM][PF6] is more hydrophobic than [BMIM][NO3] and hence is less likely to dissociate into ions to destabilize the enzyme.

A study of human serum albumin (HSA) and equine heart cytochrome c (cyt c) by CD spectroscopy and small-angle neutron scattering (SANS) demonstrated that the IL 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]) not only caused sig­nificant unfolding of the a-helical proteins when present as a cosolvent with water, but [BMIM][Cl] also changed the aggregation state of HSA, suggesting that the interaction depends on the protein sequence [104].

The secondary structure can also be analyzed with FTIR spectroscopy since proteins absorb infrared wavelengths due to peptide bond vibrations. Liu et al. [32] reported that a significant decrease in the a-helix content of lipase from Burkholderia cepacia probably affects the lipase active site: the lower the a — helix, the higher the “open” conformation of the active site, allowing easier access to the substrate.

More recently, Fan et al. [105] suggested that ILs could quench the intrinsic fluorescence of papain, probably by means of a static quenching mechanism. The calculated binding constants were very small compared with that of volatile organic solvents, indicating that only very weak interaction between ILs and papain existed. The Gibbs free energy change (AG), enthalpy change (AH), and entropy change (AS) during the interaction of papain and ILs were estimated. The negative values of these parameters obtained, indicated that the interaction between ILs and papain was a spontaneous process, also implying that hydrogen bonding and van der Waals forces played important roles in the interaction processes.

The impact of water-miscible ILs on proteins was characterized by structural changes of green fluorescent protein (GFP) in aqueous solutions containing 25 and 50 % (v/v) of [BMIM][Cl]. The SANS and spectroscopic results indicated that GFP is a great deal less compact in 50 % (v/v) [BMIM][Cl] than in neat water, suggesting unfolding from the native structure. The oligomerization state of the protein in IL-containing aqueous solution changes from a dimer to a monomer in response to the IL, but does not change as a function of temperature of the IL solution. The SANS and spectroscopic results also demonstrate that the addition of this hydrophilic ionic liquid to the solution lowers the thermal stability of GFP, allowing the protein to unfold at lower temperatures than in aqueous solution [106].

An aqueous solution of free-enzyme molecules added to the hydrophobic IL phase could be regarded as being included, but not dissolved, in the medium, the

essential water shell around the protein being preserved, and providing an adequate microenvironment for the catalytic action [14]. Usually, enzymes fold by placing the non-polar residues in a hydrophobic core, while polar residues are located on the hydrated surface. A “memory” phenomenon is observed when an enzyme is placed in a dry hydrophobic system, because the biocatalyst is trapped in the native state as a consequence of the low dielectric constant of the medium. This intensifies intramolecular electrostatic interactions and enables the catalytic activity to be maintained [103, 107, 108]. The extremely ordered supramolecular structure of ILs in solid and liquid phase has been described as an extended network of cations and anions connected by hydrogen bonds [109]. This network might be able to act as a mould, maintaining an active three-dimensional structure of the enzyme in non-aqueous environments, and avoiding classical thermal unfolding. Therefore the incorporation of molecules and macromolecules in the ionic liquid network causes changes to the physico-chemical properties of these materials and can cause, in some cases, the formation of polar and non-polar regions [109]. So, enzymes in water immiscible ILs should also be considered as being included in the hydrophilic gaps of the network, where the observed enzyme stability could be attributed to the maintenance of this strong net around the protein. ILs can clearly be considered as both solvents and liquid immobilization supports because multipoint enzyme-IL interactions (ionic, hydrogen bonds, van der Waals, etc.) may occur, resulting in a supramolecular net able to maintain an active protein conformation [98] (Fig. 11.6).

A theoretical basis for predicting the compatibility of enzymes and anhydrous ionic liquids has not yet been developed, although a number of possibly contributing factors have been discussed, such as the cation H-bond donating capability, log P, formation of hydrogen-bonded nanostructures, and solvent viscosity [8]. With regard to the compatibility of enzymes and hydrophobic ionic liquids, hydrogen bonding could be the key to understanding. It is well known that the thermal stability of enzymes is enhanced in both aqueous and anhydrous media containing polyols as a consequence of an increase in hydrogen bond interactions. Thus, both the solvophobic interactions essential for maintaining the native structure and the water shell around the protein molecule are preserved by the “inclusion” of the aqueous solution of free enzyme in the IL network, resulting in a clear enhancement of enzyme stability (Fig. 11.7).

Yang [110] maintains that an ion may affect enzyme performance by playing the role of substrate, cofactor, or even inhibitor. But more generally, the effect of specific ions could be better understood by considering an ion’s ability to alter the bulk water structure, to affect the protein-water interaction, and to directly interact with the enzyme molecules. So, the effect of ions on enzyme activity and stability has usually been linked to the Hofmeister series (or the kosmotropicity order): kosmotropic anions and chaotropic cations stabilize enzymes, while chaotropic anions and kosmotropic cations destabilize them. The influence of hydrophilic ILs on the protein activity and stability usually follows the Hofmeister series when ILs dissociate into individual ions in water [111] but, unfortunately, there are many cases in which this series is not followed, especially when there is little or no water present in the IL media, and, furthermore, some authors associate

Hofmeister effects only with anions [112]. Micaelo and Soares [103] presented a molecular simulation study of an enzyme in two ionic liquids, [BMIM][PF6] and [BMIM][NO3], observing that the enzyme structure is highly dependent on the amount of water present in the IL media and that [BMIM][PF6] significantly increases protein thermostability at high temperatures, especially at low hydration values. These ILs “strip” most of the water from the enzyme surface in a degree similar to that found in the case of polar organic solvents, while the remaining water molecules at the enzyme surface are organized in many small clusters. [BMIM] [PF6] seems to retain similar amounts of water at the enzyme surface, as acetoni­trile, and supports the evidence of the polar nature of this IL. This IL [BMIM] [NO3], in contrast, replaces almost all the water at the enzyme surface, which may be the reason for its destabilizing effect on the enzyme. A more detailed analysis of enzyme solvation by the two ILs shows that the anion species dominates the

non-bonded interactions with the enzyme, as judged by the number of hydrogen bonds observed between the enzyme and the cation and anion species of each IL. The ability of ILs to dissolve molecules depends mainly on the hydrogen bond- accepting properties of the anion, as stated by Anderson et al. [113]. Moreover, Zhao et al. [83] observed that the dissolution of lipase in most hydrophilic ILs is an indication of strong interactions between the enzyme and solvent molecules. If such interactions disturb the active sites and/or are strong enough to disrupt the protein structures, the enzyme activity is lost. However, if such interactions are not too strong but allow the enzyme’s structures to be maintained, these hydrophilic ILs do not inactivate the enzyme (such as [Et3MeN][MeSO4], or AMMOENG series ILs).

Weingartner et al. [112] observed the importance of “microheterogeneity” in ILs. The charged ionic groups and non-polar residues of cations and anions give rise to the nanoscale structural heterogeneity of ILs, which is not encountered in simple molecular solvents. The resulting hydrophilic and hydrophobic patches of the IL structure have intriguing consequences for solvation because they enable dual solvent behaviour: an IL can incorporate a non-polar solute in non-polar domains, while hydrophilic domains solvate polar solutes.

Finally, the type of reaction medium used is conditioned by the type of biotrans­formation; for example, water-immiscible ILs were found to be the most effective for the production of fermentable sugars from cellulose at low water content or in nearly anhydrous conditions. Moreover, immobilized CALB was the most efficient biocatalyst for the transesterification (alcoholysis) of vegetable oils (or animal fats). Both the IL and CALB can be recycled for at least four successive reactions without any loss of activity. Furthermore, aqueous solutions of hydrophilic ILs were necessary to produce bioethanol because the presence of water is necessary for cellulose hydrolysis. In this type reaction medium, the IL and the assayed IL-to — water concentration ratio are the key criteria.

Other strategies proposed to improve the efficiency of bioethanol and biodiesel transformations include biphasic systems based on IL and scCO2 (supercritical carbon dioxide), the addition of cosolvents and a IL coating of immobilized enzyme particles. The enormous potential of immobilized multi-enzymatic or cross-linked enzyme aggregates for bioethanol synthesis in ionic liquid media has only just been realized.

11.3 Conclusions

There is, as yet, no theoretical basis for predicting the compatibility of ionic liquids with enzymes, although key parameters for this relationship depend on the type of reaction system.

(a) Water-immiscible ILs.

The enzyme structure is highly dependent on the amount of water present in the IL medium. So, the hydrophobicity of a water-immiscible ionic liquid may be considered as a constrainer of polarity, because hydrophobicity is related to miscibility with water, and the water shell around the protein molecule is essential for maintaining the activity/stability of the enzyme. Water-immiscible ionic liquids are nevertheless hygroscopic, as noted above, and readily absorb a low percentage of water. Besides, some local hydrophobic ion-enzyme macro­molecule interactions are also important for enzyme stability. Thus, for the compatibility of enzymes and anhydrous ionic liquids, a hydrophobic effect could be the key.

(b) Water-miscible ILs.

These ILs are used as aqueous ionic liquid mixtures and the ratio of the ionic liquid-to-water used is crucial to the effect it has on the enzyme. To maintain the activity of hydrophilic ionic liquid-dissolved enzymes, a balance of mild hydro­gen bond-accepting and donating properties is required, so cation and anion size and the ability to form hydrogen bonds are important for these systems because stabilization primarily results from hydrophobic forces and hydrogen-bond. The hydrogen-bond donating ability is usually a property of the cation, while the anions act as hydrogen-bond acceptors, and it has been demonstrated that the ability of ionic liquids to dissolve complex compounds, such as sugars and proteins, mainly depends on the hydrogen-bond accepting properties of the anion. So, with regard to the compatibility of enzymes and hydrophilic ionic liquids, hydrogen bonding could be the key. Ionic liquids, in particular their anions which form strong hydrogen bonds, may dissociate the hydrogen bonds that maintain the structural integrity of the a-helices and в-sheets, causing the protein to unfold wholly or partially. As discussed above, another key property for these systems is viscosity, which is strongly influenced by cation chain length.