In the previous section, the dissolution of plant biomass was mentioned. Those studies dissolve cellulose, hemicellulose, lignin, and some other materials altogether. On the other hand, some researchers are interested in isolating target or useful materials selectively from plant biomass. In the case of dissolution and collection process of cellulose, some separation processes of lignin and other polysaccharides are required. When one can design ILs suitable to extract only lignin selectively, the ILs should become a promising solvent in the pulping method of plant biomass.

Pu and co-workers reported about lignin dissolution using several ILs in 2007 [57], and Lee and co-workers reported on a similar process in 2009 [58]. Pu and co-workers investigated several imidazolium type ILs have a good ability to dissolve softwood kraft pulp lignin. In their study, 344 g/L lignin was dissolved in methylsulfate type ILs, and it was suggested that the selection of anions for ILs is important to dissolve lignin. In the case of [C4mim] salts, the order of lignin solubility was the function of anion species as follows: MeOSO3~ > Br~ > Cl_ > BF4~. Lee and co-workers also investigated the lignin dissolving ability of imidazolium salts having several anions. According to their study, CDILs such as [C2mim][OAc] have good ability to dissolve lignin, and less polar ILs which have non-coordinating anions such as BF4~ dissolve a small amount of Kraft lignin.

Table 2.8 Cellulosic biomass treatment ability of several CDILs

Ionic liquid

-N©N’

—os ,p

Vv

Selective extraction of not only lignin but also other components from biomass is important to construct an effective biomass conversion process. Anugwom and co-workers constructed the selective extraction process for hemicellulose from spruce, a typical plant biomass, using switchable ILs [62]. A switchable IL (Scheme 2.8) was investigated as dissolution/fractionation solvents for plant bio­mass. After the treatment for 5 days without stirring, the amount of hemicellulose in the undissolved fraction was reduced by 38 wt% as compared with that before treatment. They stated that the recovered hemicelluloses were very important in many industrial fields, because the spruce hemicellulose was mainly galactoglu — comannans, which could be used as bioactive polymers, hydrocolloids, or paper­making chemicals [63].

ILs are useful to extract not only the main components of biomass such as polysaccharides and lignin but also some valuable materials. For example, a pharmaceutical ingredient, shikimic acid, was extracted from Ginkgo biloba with [C4mim]Cl [64]. Shikimic acid is the starting material for the synthesis of oseltamivir phosphate (Tamiflu®), which is used as an antiviral agent for the H5N1 strain of influenza [65]. As seen in Fig. 2.7, using [C4mim]Cl at 150 °C, the extraction yield of shikimic acid reached to 2.3 wt%, which was 2.5 times higher than that extracted with methanol at 80 ° C. Usuki and co-workers also established the isolation process using an anion-exchange resin. They clarified that CDILs are useful and important to collect valuable materials from plant biomass.

Alterations to the polymeric structures, arising from reactions where ILs act as catalysts or even a reactive species to form adducts, may have an important role to play for the complete solubility of the composite structures of plant polymers. Recently a number of publications have demonstrated various types of reactivities for ILs, which were at first commonly thought to be relatively inert (with the dialkylimidazolium acetates being a prime example), under the conditions that are typically used in biomass treatments. These results may explain the observed differences between many ILs classes when it comes to mechanisms of cell wall dissolution.

Acid catalyzed reactions seem to be detrimental for both carbohydrates and lignin. Both of these wood components have been shown to partially depolymerize during acidic IL treatments [41]. These include reactions where the presence of the acids in the systems is intentional and where acid has originated from impurities or side reactions, at elevated temperatures. One destructive dissolution IL class are the so-called protic ILs, such as 1-H-3-alkylimidazolium ILs. Cox et al. have found this IL class to readily hydrolyze the p-O-4 ether bonds, in model compounds. Yields were found to be dependent on the anion in the protic IL [34]. The same was also observed with isolated oak wood lignin under conditions above 110 °C [35].

Similar reactions to those observed with protic ILs, degradation has been found to take place even in aprotic ILs, that shouldn’t in theory contain significant concentrations of acidic protons. At a temperature of 120 °C, Kubo et al. have found [36] that P-O-4 type model compounds undergo elimination at the a-p-position [36]. In two chloride based ILs, 1-butyl-3-methylimidazolium chloride ([bmim]Cl) and [amim]Cl, enol ether formation, without cleavage of the P-O-4 bond, was the predominant reaction. Part of the original structure was left intact after 72 h incubation. Conversely, in acetate IL 1-ethyl-3-methylimidazolium acetate ([emim][OAc]) all of the original P-O-4 structures were gone after 72 h, forming only a low amount of enol ethers and the majority of materials as other unidentified structures. In agreement with the aforementioned data, George et al. [37] have found that acetate and sulfate ILs dramatically reduce the molecular weight of organosolv lignin, while only minor changes were observed with several chlorides tested under the same conditions. In any case, based on the mechanism suggested by Cox et al., enol or vinyl ethers may hydrolyze further, even in the chloride ILs. In the acidic environment presence of trace moisture levels may lead to depolymerisation [34]. Formation of condensed structures in lignin have also been detected by HSQC-NMR after [emim][OAc] treatment at 155 °C [38].

There is a notable difference in depolymerization between different lignin preparations, such as alkali and organosolv lignins, where the p-O-4 type of linkages seem to be among the most reactive linkages [37]. Based on structural features it may be reasonable to expect lower reactivity of softwood lignin, compared to hardwood lignin, as softwood lignins have a higher abundance of condensed linkages and hardwoods are rich in the more labile p-O-4 ethers [39].

Impurities remaining from the IL synthesis are likely catalysts for certain reactions. According to Li et al. [32], the delignification of bagasse in [emim] [OAc], at temperatures reported to be above the lignin glass-transition temperature, does not happen anywhere near the same efficiency when recycled IL is used. Loss of the ILs delignification capability, after recycling [13], may allude to the presence of reactive species in the IL as an impurity, which are consumed during the first treatment step.

Cellulose is known to be labile towards acid hydrolysis in the dissolved state. The IL environment has been demonstrated to be effective for such reactions [40, 41]. The fact that rapid depolymerisation can also take place in technical ILs, even without added acid catalyst, should be surprising. Gazit and Katz have demon­strated that cellulose hydrolysis can happen under relatively mild conditions, in commercial-grade dialkylimidazolium chlorides and acetates [42]. The higher purity grades of ILs were also degrading cellulose even faster than lower grades. The catalyst for the reaction was found to be a trace amount acid, formed during the treatment. This could be scavenged by 1-methylimidazole, which is a very typical impurity in low grade commercial ILs. Such observations about acid formation strongly suggest the use of mild temperatures in IL treatments.

Fine control of temperature and dissolution atmosphere may be necessary to reduce the depolymerization of wood during the treatments. According to Miyafuji et al., the depolymerization of carbohydrates in 1-ethyl-3-methylimidazolium chlo­ride ([emim]Cl) can be mostly prevented using temperatures below 90 °C [43]. On the other hand, the use of mild conditions effects negatively to the degree of wood dissolution. Lignin showed much higher resistance towards degradation and any low molecular weight fragments were observed only at temperatures above 110 °C. Reactions in chloride-based ILs, induced by atmospheric impurities, have been investigated by Nakamura et al. [44]. Oxygen was found to facilitate the dissolution of lignin, in addition to solubilization of carbohydrates.

Depolymerization is not the only reaction type taking place when lignocellulose is treated with ILs. Addition of dialkylimidazolium cations, at the C2 position, to the reducing ends of polysaccharides has been reported from acetate ILs [45—47]. The same effect has been observed with isolated lignin [31]. The reaction follows from self-deprotonation of the imidazolium cation, forming a carbene. The carbene formation and following reactions with electrophiles will lead to conver­sion of anions to their conjugate acid form [48, 49]. Acid species formed in these types of reactions are suggested to be responsible for the depolymerization of cellulose [42], but basic impurities may also capture the released proton. Alterna­tively, acetylation of cellulose by the IL anion has been observed to happen to low degree in pure [emim][OAc] at high temperature (150 °C)[47]. The mechanism of formation of such structures is still controversial. Cetinkol et al. [50] have reported deacetylation of hemicelluloses and acetylation of lignin taking place when wood is treated in [emim][OAc]. This suggests a transacetylation mechanism [48, 49], but it is still unproven and alternative mechanisms may prevail.

Reactivity of ILs with wood polymers is certainly an important issue related to green processing of wood in such media. Not only for its effects to recycling and atom economy of the processes, but also to the yield, purity, and overall quality of the resulting materials. The use of mild conditions and possible additives, or co-solvents, in IL systems may help to gain control over unwanted side reactions.

Although Ranoux et al. reported that the synthesis of 5-HMF from carbohydrates is autocatalytic that does not require any additional catalyst [67], the proper use of catalyst not only promotes the reaction rate, but also can improve product selectiv­ity. Catalysts play a crucial role in the conversion of carbohydrates into 5-HMF. The production of 5-HMF from carbohydrates in ionic liquids has been broadly studied in the presence of a variety of catalysts such as mineral and organic acids [36, 44, 68—70], acidic ionic liquids [29, 30, 71—73], transition metal ions [41, 43—45, 50, 51, 53, 74—79] or solid acid catalysts such as ion-exchange resins [20, 28, 34, 80], molecular sieves [31], zeolite [81], or carbonaceous materials [65, 82, 83], as summarized in Table 9.1.

The dehydration of fructose into 5-HMF is an acid-promoted reaction so that homogeneous acid catalysts might be thought of as being efficient for the process. However, homogenous acids have serious practical problems related to product separation, solvent recycle and equipment corrosion. Therefore, solid acid catalysts are better candidates for 5-HMF preparation from fructose than homogeneous catalysts. Qi et al. proposed an efficient process for the 5-HMF production from fructose in [BMIM][Cl] ionic liquid by using a sulfonic ion-exchange resin as catalyst. These types of catalysts are limited to temperatures below 130 °C due to the thermal stability of the resin [20]. To overcome the temperature limitations of the polymeric resins, Qi et al. [82] synthesized a novel carbonaceous solid catalysts with — SO3H, -COOH, and phenolic — OH groups by incomplete carbonization of cellulose followed by either sulfonation with H2SO4 to give carbonaceous sulfonated solid (CSS) material or by both chemical activation with KOH and sulfonation to give activated carbonaceous sulfonated solid (a-CSS) material (Fig. 9.5). The catalysts were shown to be effective for the catalytic conversion of fructose into 5-HMF, and a 5-HMF yield of 83 % could be obtained in [BMIM] [Cl] with CSS catalyst at 80 °C for 10 min. Catalyst a-CSS exhibited a somewhat lower activity than that of CSS, even though a-CSS had a much larger surface area than that of CSS (0.5 VS. 514 m2/g). The authors ascribed the lower activity of the a-CSS catalyst compared with that of the CSS catalyst to the lower concentration of —SO3H groups of the a-CSS catalyst (0.172 vs. 0.953 mmol/g).

Many Lewis acids such as metal chlorides have been used for the transformation of 5-HMF from carbohydrates in ionic liquids. The most frequently used Lewis acid catalyst, chromium (II or III) chloride, has been extensively employed for the production of 5-HMF from different carbohydrates such as fructose, glucose, sucrose, cellobiose, and cellulose [41, 45, 47, 53, 87, 91]. However, there are other metal salts that have been used for efficient conversion of sugars into 5-HMF in ionic liquids. Zhang et al. [77] established a new catalytic system based on germanium(IV) chloride (GeCl4) for the conversion of carbohydrate into 5-HMF in [BMIM][Cl], and this system exhibited excellent catalytic activity for fructose and moderate activity for other carbohydrates such as glucose, cellobiose, sucrose, and cellulose, in terms of 5-HMF yield. Chan et al. [84] developed an efficient tungsten salt catalytic system for the conversion of fructose into 5-HMF with a yield of 63 % at low temperatures (RT to 50 °C) in [BMIM][Cl] ionic liquid. When the THF-[BMIM][Cl] biphasic system was applied as a continuous reaction process, a 5-HMF yield of above 80 % was obtained, indicating that the system could be suitable for the large-scale synthesis of 5-HMF from fructose.

Liu et al. [38] reported some exciting results for fructose dehydrated to 5-HMF that was achieved with a ChCl/CO2 system. Yields of up to 72 % could be obtained with the environmental strategy and 5-HMF was found to be highly stable in the presence of ChCl, presumably through the formation of a hydrogen-bonding struc­ture with ChCl to form a DES. The catalytic process allowed 5-HMF yields higher than 60 % with a high fructose initial content (up to 100 %), whereas such high 5-HMF yields could be obtained only from a fructose initial content lower than 20 % with usual methods. In usual processes, after extraction, 5-HMF is unavoid­ably contaminated with Lewis and Br0nsted acids, however, in this process, a decrease of pH upon addition of CO2 allows circumventing this problem because carbonic acid is readily converted to CO2 and water when the CO2 pressure is released.

Dissolution of cellulose in an ionic liquid (IL) was first achieved using imidazolium chloride as mentioned before [13]. Since then, many researchers have used chloride salts as cellulose dissolving ILs [1, 3]. However, most chloride salts have serious disadvantages: high Tm, high viscosity, especially when cellulose is dissolved in the IL, and high corrosive nature. Since these drawbacks are particularly critical when chloride salts are applied in industrial use, ILs must be developed that enable

efficient processing of cellulose biomass while minimizing the cost of the extra energy.

Cellulose dissolution properties of ILs depend on the hydrogen-bonding char­acteristics for which the role of the anion is important for loosening hydrogen bonds in the crystalline region of cellulose [1—3]: the hydrogen bond-accepting ability of the anions of ILs seems to be closely linked to the solubility of the cellulose, and the solubility of cellulose in ILs increases almost linearly with the increasing hydrogen bond-accepting ability of the anions [14]. To estimate hydrogen-bond-accepting ability, it is important to know the hydrogen-bonding characteristics. The Kamlet — Abboud-Taft parameters (KAT values) specify three distinct solvent polarities: hydrogen-bonding acidity (a), hydrogen-bonding basicity (p), and dipolarity/polar — izability (n*)[15, 16]. Ohno and co-workers pointed out that the p-values obtained in a solvatochromatic study might be a better indicator of the ability to dissolve cellulose than the pKa values [17, 18]. Hydrogen-bonding basicity is considered to be necessary to dissolve cellulose, because high basicity weakens the inter — and intra-molecular hydrogen bonds in cellulose crystal [14].

Table 4.1 is a list of several ILs that have the capability to dissolve cellulose and their reported KAT values. As shown, p-values depend on the anion which is a good indicator of the cellulose dissolving property.

ILs containing carboxylate anions were reported to show strong hydrogen­bonding basicity: Bonhote et al. [19] reported that [C2mim][acetate] displays strong hydrogen-bonding basicity [19] and that this salt dissolves cellulose well [8]. Imidazolium salts with carboxylic anions, such as lactate [20] and amino acid

[21] are also reported to show high hydrogen-bonding basicity. All carboxylic acid salts dissolve cellulose very well and have better dissolution ability than the corresponding chloride salts. The IL [C4mim][RCO2] shows stronger hydrogen­bonding basicity than chloride salts and can dissolve cellulose [14]. Ohno and Fukaya [14] established that the solubilization temperature of cellulose in imidazolium carboxylates depends on the length of alkyl side chains of imidazolium cations and increases for larger alkyl side chains [14]. Since viscosity of the ILs generally increases with imidazolium salts that have long alkyl chain lengths, this clearly indicates that decreased viscosity of ILs greatly affects the high dissolution of cellulose in IL; the solubilization temperature of cellulose, in fact, decreases with decreasing viscosity (Table 4.1)[14]. [C4mim][t-C4H9CO2] has the highest в-value (see Table 4.1), and so it might be expected to dissolve cellulose readily. However, the solubilization temperature of this liquid was higher than [C4mim][HCO2] and this is probably due to increased viscosity of the [t-C4H9CO2] salt.

Ohno and co-workers conducted a very detailed study of this issue and prepared various [C2mim] salts combined with various sulfonium and phosphonium anions [17]. They found that all ILs have identical n* values, a values and в values. A comparison of sulfonium salts and phosphonium salts showed the latter to have higher в values and higher cellulose solubility. In particular, it was found that [C2mim][(MeO)HPO2] dissolves cellulose even at room temperature [17]. Although в-value of [C2mim][(MeO)HPO2] is smaller than that of [C2mim][(MeO)MePO2], the former liquid dissolved cellulose better at lower temperature than the latter. This can be explained by the decreased viscosity of the former liquid [17].

The recovery of glycerol and the removal of inorganic salts remain cost in-efficient in traditional biodiesel production process catalyzed by acid or base catalysts [58]. Biocatalytic processes by lipases, especially immobilized biocatalysts, offer a promising route to improve the greenness of biodiesel production, as these processes can be done with high activity and selectivity under mild reaction conditions [59]. However, the traditional use of immobilized lipases also has several disadvantages, because lipids and methanol constitute a two-phase system which severely impedes mass transfer and enzyme catalysis. Some ILs with good enzyme compatibility were also good solvents for biocatalytic conversion [60, 61]. The design and application of lipase-compatible ILs have been investigated [62—71]. The main task of the introduction of ILs into these areas was to balance enzymatic compatibility and miscibility of ILs with lipids and methanol, thus to facilitate catalyst recycling and reuse, and product separation and purification.

For example, lipases from Candida Antarctica and Pseudomonas cepacia have been successfully immobilized in different ILs. The methanolysis of lipids was performed at room temperature, and product separation was realized by simple decantation, resulting in a facile reuse of the ILs/enzyme catalytic system. However, it was also found that many hydrophobic ILs have poor capability in dissolving lipids, while hydrophilic ILs tended to cause enzyme inactivation [62, 72]. A new type of ether-functionalized ILs carrying anions of acetate or formate was prepared. These ILs dissolved oils at 50 °C, at which temperature lipases maintained high catalytic activities even in the presence of high concentrations of methanol (up to 50 % v/v). High conversions of miglyol oil were observed in mixtures of ILs and methanol (70/30, v/v) when the reaction was catalyzed by a variety of lipases and different enzyme preparations (free and immobilized). The preliminary study on the transester­ification of soybean oil in ILs/methanol mixtures further confirmed the potential of using oil-dissolving and lipase-stabilizing ILs in biodiesel production [69].

Homogeneous systems involving ILs have also been known for biocatalytic production of biodiesel [65]. Three hydrophobic ILs capable of dissolving triolein-methanol mixtures were devised. These hydrophobic ILs were based on cations attached with long chain alkyl groups. The [C18mim][NTf2] was able to dissolve vegetable oil to form a monophasic system and provided an excellent microenvironment for catalysis. It has also been shown that polystyrene divinylbenzene porous matrix covalently attached with 1-decyl-2-methyimi — dazolium cation could be used as carrier to immobilize C. antarctica lipase B. The immobilized lipase was applied for methanolysis of triolein in both tert — butanol and supercritical (sc) CO2 as reaction media. It was found that the use of modified supports with low ionic-liquid loading led to the highest yields (up to 95 %) and operational stability (85 % biodiesel yield after 45 cycles) in scCO2 at 45 °C, 18 MPa. The presence of tert-butanol as an inert co-solvent in the scCO2 phase at the same concentration as triolein was key to avoid poisoning the biocat­alyst through the blockage of its active sites by the polar byproduct (glycerol) [65]. Additionally, the anion also played an important role on the efficiency, which was improved by increasing its hydrophobicity (i. e., [NTf2] > [PF6] > [BF4] > [Cl]). More hydrophobic microenvironment provided by NTf2 on the surface could facilitate a more efficient enzyme conformation as well as a higher accessibility of the substrate to the enzyme active site, leading to an increase in the observed activity. A large alkyl chain in the cation resulted in a clear improvement of the efficiency for the biotransformation in monophasic liquid systems [73]. Many other progresses are summarized in Table 7.5.

Although enzymes in ionic liquids have shown enhance in activity and stability, the relatively low solubility and therefore low activity and stability of lyophilized or free enzymes in ionic liquids are still the major drawback of enzymatic reactions in ionic liquids. As with organic solvents, proteins are not soluble in most of the pure ionic liquids. Although some ionic liquids can dissolve enzymes through the weak hydrogen bonding interactions, they often induce enzyme conformational change resulting in enzyme inactivation [41, 61]. That might limit their potential applica­tions in biotechnology. Various attempts have been made to improve the activity and stability of enzyme in ionic liquids and in general these attempts can be divided into two categories: (1) modification of enzymes and (2) modification of solvent environment (Table 10.3).

Among these methods for modification of enzymes, immobilization is the most common technique. Immobilization of enzymes on the solid carriers or supports is a routine method for improving the enzyme stability in organic solvents as well as in ionic liquids. The supports can be resins [14] (e. g., Novozym 435, lipase B from

Table 10.3 Methods for activation and stabilization of enzymes

Modification of solvent environment Water-in-ionic liquid microemulsions Using surfactants to form aqueous micelles in ionic liquids

Coating enzymes with ionic liquids

Manipulation and design of ionic liquids structures

Having large molecule structure in order to minimize the H-bond basicity and nucleophilicity of anion Containing multiple ether and/or hydroxyl groups to optimize the solvent properties for enzymes

Candida antartica immobilized on acrylic resin, the most commercially available well-known immobilized enzyme), carbon nanotube [62—64], agarose hydrogel film [65—67], and magnetic silica particles [68]. Furthermore, enzymes can be immobilized by sol-gel encapsulation which is a technique for entrapping bio­molecules in polymer matrix via non-covalent interactions between the polymer network and biomolecules [69, 70]. To overcome the limitations of sol-gel process such as gel shrinkage, pore collapse and inhibition effect of released alcohol during the preparation of sol-gel materials, different additives such as sugars, amino acids, carbon nanotubes, and ionic liquids have been added to reduce the gel shrinkage, adjust the protein hydration and to increase the activity and stability of enzyme [71, 72]. Interestingly, the polymerized ionic liquid microparticles could be also used to encapsulate enzymes, e. g. horseradish peroxidase, with activity more than two times higher than that encapsulated in polyacrylamide microparticles [73]. More­over, enzymes can be self-immobilized by using cross linking agent such as glutaraldehyde. Several different techniques have been tried for self­immobilization of enzyme using glutaraldehyde such as cross-linked enzymes (CLEs) [74, 75], cross-linked enzyme crystals (CLECs or CLCs) [76—78] and cross-linked enzyme aggregates (CLEAs) [79—81]. Among these techniques, CLEAs are shown to have more advantages such as ease of preparation and recycling, enhanced activity and stability. CLEAs can be used for variety of enzymes, and higher enzymes stability of CLEAs in ionic liquid than free enzyme have been reported [82—84]. In addition, the modification of enzymes with PEG (having both hydrophilic and hydrophobic characteristic) either by physical adsorp­tion or covalent attachment in order to make the modified enzymes soluble in organic solvents and ionic liquids [85—92] as well as pretreatment of enzyme and supports/carriers with organic solvents such as n-propanol before immobilization also result in enhanced enzyme activity and stability [84, 93—96].

The activity and stability of enzymes can be improved by the modification of solvent environments. In this approach, the micro-environment surrounding the enzymes or ionic liquids media are modified to become more compatible with enzymes. Several approaches have been done in recent years such as using surfac­tant to form aqueous micelles in ionic liquids [97—100], coating enzyme with ionic liquids [101—104] and manipulating or designing enzyme compatible with ionic liquids [105—107]. Regarding the activation and stabilization of enzymes in ionic liquids, several excellent reviews are available in literature [16, 55, 108].

Many density correlations of ILs have been reported because it is an important fundamental property [49]. IL database, such as The UFT/Merck Ionic Liquids Biological Effects Database, IUPAC Ionic Liquids Database and Tohoku Molten

Salt Database, provide up-to-date information on densities of ILs [26]. Generally, imidazolium-based ILs are widely used as solvents for ILs applications owing to their excellent physical properties, such as low viscosities and high thermal and aqueous stability. Their densities are more available in the open literature than other properties ILs.

The usual densities of ILs vary between 1.12 g • cm-3 ([(n-C8H17)(C4H9)3N] [(CF3SO2)2N]) and 2.4 g • cm-3 (a 34-66 mol% [(CH3)3S]Br/AlBr3 ionic liquid) [64, 65]. However, large cations have densities lower than water, such as aliquat (in Table 1.6), since their long alkyl chains have higher flexibility. From IPE IL database [26], the densities of most ionic liquids tend to have low sensitivity to variations in temperature. Furthermore, the impact of impurities on densities is much less dramatic than for viscosities.

One can make some conclusions on the cation effect and also to study the effect of alkyl chain length on the density and derived properties. Gardas et al. [66] found that as the alkyl chain length in the pyrrolidinium cation increases, the density of the corresponding IL decreases, similar to that observed for imidazolium-based ILs [67]. Generally, the order of increasing density for ILs with a common cation is N(CN)2- < BF4- < CF3CO2- < CF3SO3- < PF6- < Tf2N- (shown in Fig. 1.4) [26, 53, 69]. The higher densities of Tf2N—containing ILs arise from the much higher mass of the anion [68].

Since it is difficult to measure densities of all ILs at different conditions, it is necessary to find a method to estimate densities of ILs. So far, a large amount of models about prediction of densities of ILs have been proposed in the open literature. The common models are group contribution methods (GCMs), quantita­tive structure property relationships (QSPRs) and artificial neural networks (ANNs) [70]. For instance, Lazzus [71] used the method of ANNs to estimation the density of imidazolium-based ionic liquids at different temperature and pressure. The method has a better estimation results, but it is not convenient to be used in prediction.

The volumetric properties for the ILs can be estimated by the values of density. For example, [CnMIM][BF4] (n = 2, 3, 4, 5, 6) are common ILs, many researchers have previously reported many properties in wide ranges of temperature. We can plot of values of in p against T, a straight line was obtained for given IL, and its empirical linear equation is [72]

Lnp = b — aT (1.7)

Where b is an empirical constant, the negative value of slope, a = —(3lnp/3T)p, is thermal expansion coefficient of the IL [CnMIM][BF4].

We can also obtain the molecular volume (Vm) of ILs. The value of Vm was calculated using the following equation [72]

Vm = M/(Np) (1.8)

Where M is molar mass of ILs, N is Avogadro’s constant.

In comparison to common molecular solvents, the viscosity of ILs, in particular those that can dissolve cellulose, is very high. As an example, at 25 °C the viscosities of DMSO (2 mPa s) and EMIMAc (200 mPa s) differ by two orders of magnitudes. The viscosity further increases when cellulose is dissolved in the ILs. For cellulose solutions in EMIMAc and BMIMCl, transition from a diluted to a semi-diluted regime, in which polymer coils start to entangle, occurred at a polymer content of 1-2 % [75, 76]. Above these concentrations, viscosity is proportional to cn with n being about 4.0 at 0-40 °C and about 2.5 at 60-100 °C (Fig. 5.4 left). Comparable values for the exponent have been reported for cellulose solutions in DMA/LiCl (n = 3-4) and NMMO (n = 4.6) [77, 78].

With decreasing temperature, the viscosity of cellulose/IL solutions strongly increases (Fig. 5.4 right), which is expected for a polymer solution. However, the increase does not follow an Arrhenius-like behavior (Eq. 5.1) that is usually observed for polymer solutions, including cellulose solutions in NMMO or NaOH/water [79, 80]. As indicated by the strong temperature dependence of the exponent n, the deviation is especially pronounced below 40 °C. The temperature behavior of cellulose/IL solutions can be fitted best by a Vogel Fulcher Tamman

Fig. 5.4 Viscosity of cellulose solutions in 1-ethyl-3-methylimidazolium acetate (EMIMAc, left and right) and 1-butyl-3-methylimidazolium chloride (BMIMCl, right) at different cellulose concentrations (left) and temperatures (right). Left: straight lines represent fit according to a linear (c < 1 %) or exponential equation (c > 2 %). Right: dotted lines represent extrapolation from T = 60-100 °C according to an Arrhenius equation, straight lines represent fitting according to a Vogel Fulcher Tamman equation (Adapted with permission from [75], Copyright 2009, American Chemical Society, and from [8], Copyright 2012, MDPI AG)

equation (Eq. 5.2) [75, 76]. Since, many pure ILs show the same curved slopes, it can be concluded that the tremendous viscosity increase is caused by specific interaction within the ILs itself and not with the dissolved polysaccharide [81, 82].

ea

П = A • eRT (5.1)

П = B • VT • e7^ (5.2)

The high viscosity of cellulose/IL solutions significantly affects the mass trans­fer in all steps involved in the processing of the polymer. For efficient homogeneous derivatization of cellulose, reactions are usually performed in solutions having a polysaccharide concentration in the range of 2-10 %. Above 80 °C, viscosity of these systems is sufficiently low to guarantee efficient stirring and even distribution of the reactants in the reaction mixture. Thus, homogeneous derivatization reactions performed in ILs at elevated temperatures (e. g., acylation of cellulose) usually yield uniform products with well reproducible properties. In contrast, at temperatures below 40 °C interference of reaction kinetics and mass transport phenomena might occur as result of the significant viscosity increase, in particular if the reagents applied are solids and/or highly reactive (Fig. 5.5a, b).

It has been demonstrated that sulfation and tosylation of cellulose in imidazolium chlorid based ILs, performed at or below 25 °C to prevent undesired side reactions, only yields product mixtures composed of highly functional poly­saccharide derivative and unmodified cellulose [33, 35]. In both cases, a rather long period of time is required for a complete mixing and even distribution of the

solid reagent

Ч-

Fig. 5.5 Cellulose solutions in ionic liquids with different agents, depicting the high viscosity (a, b) and hydrophilicity (d, e) of the systems as well as the effect of co-solvents (c, f)

reagents within the highly viscous system. Although the cellulose/IL solution and the derivatization reagent might eventually form a homogeneous mixture if given sufficient time, they remain as two separated phases due to kinetic hindrance, i. e., the systems is not in a state of thermodynamic equilibrium. Thus, these derivatiza — tion reactions can be considered as heterogeneous; cellulose molecules in direct vicinity to the derivatization reagent are converted to derivatives with a high DS whereas other fractions of the polysaccharide remain unmodified. However, effi­cient homogeneous derivatization of cellulose in ILs could be achieved at room temperature by using co-solvents to diminish the high viscosity (see Fig. 5.5c and Sect. 5.3.4.2).

The chemical modification of cellulose, dissolved in an IL, is a completely homogeneous process only if all reaction partners, including reagents, bases, co-solvents, and intermediate — and final product, remain dissolved in one phase over the whole reaction time. However, cellulose dissolving ILs are known to be rather hydrophilic compounds, which are consequently not miscible with various hydrophobic compounds frequently involved in the chemical derivatization of cellulose (Fig. 5.5d, e). Depending on whether phase separation occurs at the very beginning or later during the reaction, the corresponding reactions proceed either completely or partly heterogeneous. It needs to be considered that this change
in the reaction course might affect the reactivity of the derivatization reagent, the regioselectivity, or the distribution of substituents along the polymer chains in comparison to a completely homogeneous conversion of cellulose.

The esterification of cellulose with lauryl chloride in BMIMCl starts homoge­neous but results in precipitation of the hydrophobic derivative formed upon increasing substitution [14]. As result of the decreased reactivity in the heteroge­neous system formed, no products with DS > 1.5 could be obtained. Due to the hydrophilic nature of common cellulose dissolving ILs, homogeneous preparation of trimethylsilyl cellulose (TMSC) is not feasible [41, 42]. The hydrophobic reagent applied for derivatization is not miscible with the polysaccharide solutions and the silylated products formed precipitate from the reaction mixture at DS > 2. A heterogeneous procedure for silyation of cellulose has been described using toluene, which is immiscible with cellulose/IL solution, to solubilize the hydro­phobic reagent [42]. The reaction involves transition of (a) the reagent into the hydrophilic IL phase and (b) the TMSC with a certain DS into the hydrophobic toluene phase, where the derivative is further silylated in a homogeneous reaction. Thus, this particular procedure is most suitable for obtaining highly functionalized TMSC. In contrast, complete homogeneous preparation of TMSC in the DS range of 0.4-2.9 has been achieved in ILs by using chloroform as a hydrophobic but IL-miscible co-solvent [41].

There has been some research into deoxygenation and pyrolysis of carbohydrates in ILs. Generally, the thermal stability of ILs precludes their use under the conditions needed for deoxygenation catalysts and pyrolysis. In some cases, however, ILs have been shown to support these reactions. Sheldrake et al. demonstrated the use of dicationic ILs as a medium for pyrolysis of cellulose to dehydration products of glucose. The ILs used were composed of two imidazolium rings connected with a

4- , 6-, or 9-carbon alkyl chain. Generally, pyrolysis occurs at temperatures above 300 °C, but the use of these dicationic ILs allowed the production of levoglucosenone, 1-(2-furanyl)-2-hydroxyethanone, and HMF along with trace amounts of levoglucosan at 180 °C. Only 5.5 % combined yield was obtained for these products. Monocationic imidazolium ILs and dicationic pyridinium ILs were not thermally stable at this temperature and no products, other than IL decompo­sition products were detected [60].

Chidambaram and coworkers demonstrated the production of HMF and subse­quent deoxygenation into 2,5-dimethylfuran in EMIMCl. While a number of acids were tested as catalysts, it was found that heteropoly acids gave a better combina­tion of conversion and selectivity than any of the simple acids tested (such as sulfuric, fluoroacidic, nitric, HCl, and phosphoric). The best of the heteropoly acids, H3PMo12O40 (12-MPA), produced a 71 % conversion of glucose and a 89 % HMF selectivity. To investigate the next step of processing HMF in ILs, metal catalysts were used to deoxygenate the HMF into 2,4-dimethylfurfural in the IL. Palladium, platinum, ruthenium, and rhodium catalysts supported by carbon were added to IL/HMF solutions at 120 °C under 62 bar of H2. The most effective of these catalysts was Pd/C which gave 19 % conversion and 51 % selectivity. When a small amount of acetonitrile was added to prevent the formation of humins, a conversion of 47 % with 36 % selectivity was obtained [118]. Other work has been done on the hydrodeoxygenation of lignin specifically and will be discussed later in this chapter.

Several groups have disagreed on whether or not a given enzyme is active in a particular ionic liquid. For example, Schofer et al. [19] reported that CALB had no activity in [BMIM][BF4] or [BMIM][PF6], but other groups reported good activity during transesterification or ammoniolysis in the same ionic liquids. Such incon­sistencies may be the result of impurities.

By measuring the effect of some additives on lipase catalyzed acetylation, Park and Kazlauskas [84] concluded that the most likely causes of slow or no reactions in some ionic liquids are traces of silver ion or acidic impurities. More recently, Lee et al. [90] reported that the activity of Novozym 435 in [OMIM][NTf2] decreased linearly with the chloride content, and that 1 % (wt.) increase in [OMIM] [Cl] (^1,540 ppm [Cl-]) caused a 5 % decrease in enzyme activity.

Washing with water followed by vacuum drying can be used to purify water — immiscible ionic liquids. However, the purification of water-miscible ionic liquids, such as [BMIM][BF4] involves filtering through silica gel followed by washing with aqueous sodium carbonate solution.

Although these purified ionic liquids worked reliably, purification still needs further research. Besides, it has to be borne in mind that the effects of impurities may vary from enzyme to enzyme in ionic liquids.

Most ionic liquids based on common cations and anions should be colourless, with minimal absorbance at wavelengths greater than 300 nm. In practice, the salts often take on a yellow hue, particularly during the quaternization step. Impurities or unwanted side reactions involving oligomerization or polymerization of small amounts of free amine are a major limitation for studying enzyme structure by UV/visible spectroscopic techniques.

Clearly, the impurity most likely to be present in high concentrations in ionic liquids is water. Other reaction solvents are generally easily removed by heating the ionic liquid under vacuum. This observation is very important when ILs (as pure solvents) in nearly anhydrous conditions are used as reaction media.