Reflections on the Solubility of Cellulose - Industrial & Engineering

Ind. Eng. Chem. Res. 2010, 49, 11121–11130

11121

Reflections on the Solubility of Cellulose Wood cellulose can be used for producing biofuels and biopolymers, thus offering a solution to global concerns on the excessive use of fossil fuels. This requires a cellulose solvent that also allows the ecofriendly processing of selective wood components. Some ionic liquids (ILs) have shown promising results as cellulose solvents with many advantages over traditional approaches. It is agreed that their ionic nature is responsible for cleaving hydrogen bonds between cellulose chains, resulting in dissolution of the biopolymer. However, it is still necessary to establish a structural relationship between IL cations and anions, which explains why only certain ion combinations show the ability to dissolve cellulose. This work aims to analyze the structural similarities displayed by common cellulose solvents focusing on requirements for ionic liquids to qualify as such. A mutual relationship between IL anions and cations is postulated that offers an explanation for the ability or disability of certain ion combinations to dissolve the biopolymer. 1. Introduction: The Quest for Alkahest A solvent is commonly defined as a substance (usually a liquid) that possesses the power of dissolving other substances. Water is undoubtedly the most universal solvent, because it is involved in all of the natural biological processes that enable life to continue on our planet. However, the discovery of novel solvents with advantageous properties has always been of interest to mankind. The alchemists of the Middle Ages were already aspiring for an universal solvent for everything, the socalled “Alkahest”. Although the Alkahest was never found, a solvent for gold (aqua regia) had been discovered a millennia earlier than a suitable solvent for cellulose.1,2 This fact certainly reflects the changing aspirations of scientists. The necessity to face present problems such as a shortage of fossil fuels, environmental changes, and a rapidly growing population, has attracted extensive research to find a suitable solvent for dissolving cellulose from natural biomaterials, such as wood. Cellulose is the most abundant biopolymer on Earth, it regenerates quicker than fossil fuels, and it does not represent a human food source.3 Once extracted, the cellulosic biomass is a desirable feedstock for a variety of applications. One example is the production of biofuel by fermenting the monomer sugar units after hydrolyzing cellulose.4 Another option is the manufacturing of biocompatible high-tech composites in medical engineering. Some cellulose composites are desirable materials for tissue replacements in brain and heart surgery or as scaffolds for joint replacement and bone regeneration.5–8 Especially with regard to biocomposites production, it is important to dissolve and process cellulose without degradation.9,10 As a consequence, harsh dissolution conditions that could disrupt the highly ordered structures in native cellulose that are responsible for the biopolymer’s recalcitrance toward dissolution in most solvents must be avoided. Several traditional solvent systems possess the potential to dissolve cellulose, but concerns regarding their toxicity and recyclability have motivated the search for alternatives.10 In 2002, Swatloski et al. reported the solubility of cellulose in an ionic liquid (IL).11 ILs are molten salts with a melting temperature below the boiling temperature of water. Many are liquids at room temperature, which allows them to act as solvents. Their ionic nature is responsible for several desirable properties that distinguish them from most molecular solvents. Common characteristics of ILs include negligible vapor pressure, nonflammability, and, in some cases, reasonable thermal stability.12 Such solvent properties are very * To whom correspondence should be addressed. Tel.: +1 480 802 8911. Fax: +64 3364 2063. E-mail: [email protected].

desirable in industrial processes, because they promise easy handling and recyclability. However, not all ILs possess the ability to dissolve cellulose and their dissolution efficiency can vary considerably.10 It is very likely that the most effective IL cellulose solvents have not yet been discovered among the postulated 1012 possible ionic liquids.13 The lack of understanding of the dissolution process of native cellulose is still a major barrier in developing more-effective cellulose solvents. The motivation to improve such solvents does not only originate in the desire to optimize possible industrial applications, but also in the ability to prepare cellulose composites with enhanced mechanical properties.14 The objective of this commentary is to analyze literature data to provide new perspectives for the understanding of the dissolution process of cellulose, especially with a focus on ILs as solvents. The authors hope that this work will stimulate further discussions and experiments that will eventually result in the discovery of novel ILs with an enhanced dissolution power for cellulose. 2. Facts about the Solubility of Cellulose Native cellulose is recalcitrant toward dissolution because of its high crystallinity.10 Its ordered structure is the result of a three-dimensional hydrogen-bond (H-bond) network between the hydroxyl groups attached to each polymer chain. Successful cellulose solvents must possess the ability to compete for the existing intermolecular H-bond interactions, in order to separate the polymer chains from each other, resulting in the dissolution of the biopolymer.15 On this note, we must mention that a desirable cellulose solvent also must comply with other requirements beyond its good dissolving power, such as low viscosity and toxicity, ease of recycling, low melting temperature, and high thermal stability. However, those additional attributes are not expected to have an influence on the dissolution mechanism of cellulose and, therefore, are omitted in the subsequent discussions. The success of the dissolution process depends on variables such as mixing time, mixing temperature, and the crystallinity and degree of polymerization (DP) of the biopolymer. In addition, an efficient cellulose solvent should be able to reduce the crystallinity of cellulose during the process of its dissolution and regeneration. For this reason, the authors decided to focus on solvents with the ability to dissolve cellulose, regardless of its crystallinity or DP, without pretreatment, and without degrading or derivatizing the biopolymer. Furthermore, those solvents should be efficient at ambient pressures and

10.1021/ie1006596  2010 American Chemical Society Published on Web 08/13/2010

11122

Ind. Eng. Chem. Res., Vol. 49, No. 22, 2010

Figure 1. Simplified visualization of the polar characteristics of traditional cellulose solvents: N,N-dimethylacetamide + lithium chloride (DMA + LiCl), dimethylsulfoxide + tetrabutylammonium fluoride (DMSO + TBAF), and N-methylmorpholine-N-oxide (NMMO).

Figure 2. IL ions with the potential to dissolve cellulose: 1,3-dialkylimidazolium, 1,3-dialkylpyridinium, acetate, formate, dialkylphosphate, thiocyanate, chloride, and bromide.

moderate temperatures (99 %. Depending on the water content, it can dissolve up to 0.3 mass fraction of cellulose.23 However, the use of NMMO in commercial fiber processing (Lyocell process) leads to side reactions that affect the recyclability and toxicity of the solvent. In addition, some of the side reactions are exothermic and can result in runaway reactions during the process.23 Those and other drawbacks explain the ongoing search for alternative cellulose solvents, such as ILs. Although a wide range of different ionic liquids has been synthesized and characterized for specific purposes, only limited data are available on their ability or disability to dissolve cellulose. To shed light on the dissolution mechanism of cellulose in ILs, it is of equal importance to know which ILs cannot dissolve the biopolymer. It proves difficult to compare the reported cellulose-dissolving ability of ILs, because many factors (vide supra) influence the efficiency of the dissolution process. Especially important factors are the viscosity and the water content of the IL. Both increasing water content and increasing viscosity inhibit cellulose dissolution.11,24–27 This can

be explained by the regenerating effect of water as an antisolvent for cellulose and with the increased ion mobility in low-viscosity ILs. A comprehensive overview of the literature data reporting on the ability or disability of ILs to dissolve cellulose, with respect to the mixing conditions and types of cellulose, was published recently.10 From this data overview, it is obvious that only certain combinations of IL ions are effective cellulose solvents. To give an example, the 1-butyl-3-methylimidazolium cation is able to dissolve cellulose if paired with a Cl- anion, but not if paired with bis(trifluoromethanesulfonyl)amide [Tf2N]- anion. On the other hand, the Cl- anion cannot dissolve cellulose if paired with a pyrrolidinium or piperidinium cation.28 Based on these observations, it is the authors’ opinion that both IL ions (cation and anion) participate in the dissolution process, although other research has attributed most of the dissolution ability to the anion.25,29–35 However, the idea that both IL ions participate in the dissolution process has been mentioned in the literature.15,36,37 Some data have even suggested that hydrophobic interactions between the IL cation and the cellulose are responsible for the dissolution of cellulose.38,39 Figure 2 summarizes ions that can be found in ILs that have the potential to dissolve cellulose.10 In general, ILs are nonderivatizing solvents for cellulose, although it has been reported that IL acetate anions can react with cellulose if activated with p-toluenesulfonyl chloride.40 The most common cations involved in cellulose dissolution are imidazolium and pyridinium.10 The nature of the alkyl side chains on the imidazolium cation does not play an active role in the dissolution of cellulose, but it can assist in decreasing the viscosity of the IL, resulting in enhanced dissolving ability for the biopolymer.26,36,41 In the case of imidazolium chlorides, it was shown that even-numbered alkyl side chains on the imidazolium cation are better-suited for dissolving cellulose than odd-numbered chains, but this could not be confirmed for the corresponding bromides.42 It has been argued that the delocalized 3-centers-4-electrons configuration of the imidazolium cation might be responsible for its cellulose dissolution capacity, but trials with nonaromatic cyclic cations, such as pyrrolidinium or piperidinium, failed.28,43 Common IL anions with a strong potential to dissolve cellulose are chloride, acetate, formate, and dialkylphosphate.10,33,42,44 Bromide and thiocyanate are less efficient but are known to be able to dissolve cellulose with low degrees of polymerization.11,26,28 Only limited data are available for IL fluorides.42 On the other hand, some IL ions are known for their inability to dissolve cellulose (see Figure 3). Those include pyrrolidinium and piperidinium cations, as well as several anions, such as tetrafluoroborate, hexafluorophosphate, bis(trifluoromethanesulfonyl)amide (Tf2N), and tosylate (Ts).10,28 There have been several reports that ILs containing noncyclic ammonium- and phosphonium-based cations have dissolving capacity for


Figure 3. IL ions without good cellulose dissolving abilities: pyrrolidinium, piperidinium, tetrafluoroborate, hexafluorophosphate, bis(trifluoromethanesulfonyl)amide (Tf2N), and tosylate (Ts).

Figure 4. Comparison of cellulose-dissolving IL ions: planar ring cations, as well as trigonal planar, tetrahedral, linear, and spherical anions.

cellulose.45,46 However, their dissolution power is very poor, especially for higher degrees of polymerization of the biopolymer. Turbid dispersions of up to a cellulose mass fraction of 0.3 in 2-hydroxyethylammonium formate has been reported.47 However, further studies on alkanolammonium ILs with organic acid anions have shown that these types of ILs are only capable of forming fine dispersions of undissolved cellulose without any dissolution.48 As mentioned before, it is the authors’ belief that both IL ions must be suitable for the efficient dissolution of cellulose. The poor cellulose solubility of the aforementioned ammonium- and phosphonium-based ILs is, according to the view of the authors, to be attributed to the potent anion and not to the cation. The reasons for this will be discussed later in this work. 3. Comparing Cellulose-Dissolving Ionic Liquids Based on the temporary classification of IL ions, one can now look for structural similarities that distinguish ions with good cellulose-dissolution ability from those with poor ability. Dissolving cations seem to consist of planar, nitrogen-containing rings with the ability to delocalize their positive charge within their aromatic π-system (see Figure 4). This is in contrast to poorly dissolving cations (see Figure 5). The nonaromatic ring structure of these compounds results in a tetrahedral-like nitrogenium heteroatom without the ability to delocalize the positive charge. Poorly dissolving IL cations, such as noncyclic ammonium and phosphonium cations (vide supra), also display their localized positive charge at an sp3 hybridized nitrogen center. A wide structural variety of cellulose-dissolving anions can be found, covering trigonal planar, tetrahedral, linear, and spherical anions (see Figure 4). Nondissolving anions are primarily hexagonal or tetrahedral (see Figure 5).

11123

Figure 5. Comparison of noncellulose-dissolving IL ions: tetrahedral cations, as well as hexagonal and tetrahedral anions.

The obvious differences between dissolving and nondissolving cations originates from their structure, which governs their ability to delocalize their positive charge. Such differences are not as obvious with regard to the anions. Tetrahedral structures and charge delocalization can be found within both groups of anions. Differences can also be found in terms of their H-bonding ability. All of the nondissolving anions display poor to negligible H-bond donor or acceptor characteristics. This is especially true for the tetrafluoroborate and the hexafluorophosphate anions. The bis(trifluoromethanesulfonyl)amide (Tf2N) and tosylate (Ts) anions are theoretically better suited for H-bonding interactions but their bulky substituents probably impede such interactions, especially with respect to the tightly cross-linked cellulose H-bond network. This is in contrast to cellulosedissolving anions. All of these anions are moderate to excellent H-bond acceptors, and this ability is partially reflected in their dissolution power:31–33,49,50 thiocyanate < bromide < dialkylphosphate < formate ≈ acetate ≈ chloride. Comparing cellulose-dissolving anions with the ability to delocalize their charge between two O atoms, dialkylphosphate ILs do not perform as well as formates or acetates. This is possibly due to the larger anion size, resulting from additional substituents on the P atom that hinders efficient interaction with the H-bond network of cellulose. Similar charge delocalization can be found within the SO3 - of the tosylate anion, with the bulky phenyl substituent facing the other direction. However, both the hydrophobic character of the phenyl group and its rigidity are not favorable for interacting with the interwoven H-bond network of hydrophilic cellulose. A difference in H-bonding ability can also be found between dissolving and nondissolving cations. Because of the aromatic character of the cation, all of the ring protons are capable of hydrogen bonding to some extent. This is not true for the tetrahedral, nondissolving cations. Another aspect to consider is the polarity of the cations. Although difficult to quantify in some cases, it is without doubt that the heteroatoms in the imidazolium and pyridinium rings induce a dipolar character into these cations.51–54 Water molecules within the ILs can decrease the effective dipole moment of the solvent, resulting in the precipitation of dissolved cellulose.55 This is no surprise, but it is a question whether dipoles inherited in the cations are favorable for the dissolution of cellulose. However, alkanolammonium-based ILs can be very polar and they are able to participate in hydrogen bonding.48 Nevertheless, they cannot dissolve cellulose. Hence, several requirements must be met for

11124


Figure 6. Numbering of the imidazolium cation.

Figure 8. Hypothetical ability of traditional nonderivatizing cellulose solvents to arrange in cyclic formations: tetrabutylammonium fluoride + dimethylsulfoxide (TBAF + DMSO), N-methylmorpholine-N-oxide (NMMO), and N,N-dimethylacetamide + lithium chloride (DMA + LiCl). Figure 7. Piccolinic acid N-oxide.

cations to act as good cellulose solvents. A summary of the conclusions drawn so far is given below. • Cation characteristics favorable for cellulose dissolution: (i) aromatic heterocycle with an sp2 nitrogen heteroatom; (ii) ability to delocalize the positive charge; (iii) ability to participate in hydrogen bonding; (iv) dipolar ring character. • Anion characteristics favorable for cellulose dissolution: (i) ability to act as a H-bond acceptor; (ii) substituents should not be bulky or hydrophobic. 4. Attempts To Find Similarities among Cellulose-Dissolving Solvents The objective of this section is to identify structural similarities between traditional cellulose solvents and cellulose-dissolving ionic liquids. As mentioned above, ILs are nonderivatizing solvents for cellulose. For this reason, only nonderivatizing traditional cellulose solvents are considered for comparison (vide supra). The proposed similarities are derived from three observations, which are outlined below. First, nuclear magnetic resonance (NMR) studies of imidazolium-derived ILs with varying water content indicate that the chemical shift of the H-2 proton on the imidazolium cation can vary according to the water content of the IL (see Figure 6).56 A low water content can result in a greater downfield shift of the H-2 signal, compared to a higher water content. Consequently, an increasing water content can result in a decreased acidity of the imidazolium H-2 proton. This can be explained by the competing effect of water molecules as H-bond partners for the anions. Considering the fact that a water mass fraction of 0.01 in cellulose-dissolving ILs can be sufficient to prevent dissolution of the biopolymer,11 it is likely that the acidic H-2 proton of imidazolium-derived ILs is involved in the cellulose dissolution process.32,37 Second, NMR and density functional theory (DFT) studies by Balevicius et al. showed that the acidic proton of piccolinic acid N-oxide (see Figure 7) can experience chemical downfield shifts of up to 20 ppm, depending on the surrounding bulk media. In addition, they observed that the H-bond state in IL media is largely governed by the dielectric properties of the bulk media.57 This means that polar water molecules as immediate H-bond competitors can hinder the formation of intermolecular H-bonds between the imidazolium cations. Furthermore, Balevicius et al. postulated the formation of sixmembered cyclic states that involve the acid proton of piccolinic acid N-oxide as a shared bridge proton.57 The question arises as to whether cyclic states with highly polar character can result in the increased H-2 downfield shift in dry, imidazolium-derived ILs.

Third, it has been suggested that the permittivity of ILs can correlate with their ability to dissolve cellulose.10 Although measuring the permittivity of ILs has been proven to be difficult, because of their ionic nature, the latest simulation studies support this idea by demonstrating that the molecular polarizations of IL ions can be of crucial importance in understanding solvation processes.58 Solvent polarity seems also to be a key factor for the cellulose dissolution, as discussed above. The fact that all cellulose solvents display highly polar characteristics is hardly a coincidence (see Figures 1 and 2). Inspired by these three observations (vide supra), it was determined that all nonderivatizing cellulose solvents can hypothetically be arranged in cyclic structures of varying ring size. It is important to stress at this point that the discussed ring formations (vide infra) are not intended to represent the true nature of these solvents; instead, they are to be regarded as an attempt to assign a structural similarity between them. The objective of Figure 8 is to depict the hypothetical ring formations of the previously considered traditional nonderivatizing cellulose solvents. The ring sizes vary between five and six atoms, and small ions or polar characteristics assist in the theoretical formations.59 Larger cations, such as NMMO or tetrabutylammonium, seem to be less-suited for this purpose; however, one must consider the three dimensions that allow a more favorable arrangement to reduce steric hindrance, when compared to two-dimensional drawings. Although this fact might hold true for NMMO with its inherited N-O dipole, it is more difficult to fit the tetrabutylammonium cation into that scheme. However, fluoride is a very small anion and the involvement of an additional dipolar molecule (DMSO) could compensate for the unfavorable steric geometry of the cation. Consequently, one expects to find that larger anions, such as Cl- or Br-, result in the inability of the solvent to dissolve cellulose.60 The inherited ionic nature of cellulose-dissolving ILs simplifies their arrangement in hypothetical ring structures (see Figure 9). Involving the acidic H-2 proton of the imidazolium cation for the ring formation results in six- or eight-membered cyclic structures, depending on the number of H-2 protons involved. Three ion pairs are necessary for the [BuMePy]Cl to form a six-membered ring, but only two pairs are needed in the case of imidazolium chlorides. If paired with anions, such as formate, acetate, or phosphate derivatives, the proposed cyclic imidazolium arrangements tend to be even more stable. This is because both anions possess a second O atom that can interact with the imidazolium cation. But again, less steric hindrance is favorable and this is reflected in the enhanced dissolution ability of formate or acetate, compared to phosphate derivatives. A valid point for criticism at this point is the sterically hindered structure of the pyridinium cation. This issue will be resolved later in this discussion, but it is important to note that the dissolving


11125

Figure 11. Simplified axial view of interchain H-bonds in crystalline cellulose microfibrils.

Figure 9. Hypothetical ability of IL cellulose solvents to arrange in cyclic formations: 1,3-dialkylimidazolium formates (R3 ) H) and acetates (R3 ) Me), 1,3-dialkylimidazolium dialkylphosphates (R4 ) OMe, OEt), 1-butyl3-methylpyridinium chloride ([BuMePy]Cl), and 1,3-dialkylimidazolium chlorides.

Figure 10. Intramolecular and intermolecular hydrogen bonds (H-bonds) in cellulose.

efficiency of [BuMePy]Cl must be questioned, because this solvent significantly degrades cellulose during dissolution.27 Imidazolium cations are better-suited for participation in hypothetical cyclic formations and the acidic character of the imidazolium H-2 proton initially seems to play an important role within it.37 The higher its acidity, the more polar the resulting imidazolium components and the easier the formation of the cyclic arrangements. Complete dissociation of the H-2 proton would result in Arduengo carbene-like structures with high polarity. Arduengo carbenes are known to be relatively stable. However, it is known that imidazolium cations with methyl substituents at the C-2 position are also able to dissolve cellulose.46,61–63 This shows that the presence of the imidazolium H-2 proton is actually not a requirement for such ILs to dissolve cellulose, but the experimental observations discussed above did assist in developing a theory dealing with the mechanism of cellulose dissolution. 5. Structural Investigations on the Cellulose Crystal Lattice After identifying a common characteristic between traditional and IL cellulose solvents, a closer look at the structure of crystalline cellulose is required to explore whether the proposed similarities can be responsible for the dissolution of the biopolymer. Figure 10 shows the intramolecular and intermolecular H-bonds between two parallel cellulose chains. Because of the fact that cellulose consists of cellobiose monomers, every second glucose unit is rotated by 180°.10 The direct result is that only the hydroxy groups at the C-3 and C-6 position are responsible for attractive hydrogen bonding that hold the cellulose chains together. The hydroxy group at the C-2 position

Figure 12. Simplified projection of the (020) plane of native cellulose, only showing the H-bonds at the relevant C-3 and C-6 positions. This results in 9-membered cyclic structures within each polymer chain and 14-membered rings between the cellulose chains.

only reinforces the intrachain interaction to the neighbored glucose molecule. Consequently, cellulose solvents need only attack the H-bond associated with the C-3 and C-6 hydroxy groups to dissolve cellulose without degradation. In the natural state, 40 to 70 cellulose chains are associated within the microfibrils, which consist of crystalline and noncrystalline regions.10 In wood cellulose, these crystalline regions can reach a length of up to 5 µm.64 Figure 11 shows a simplified axial view of the interchain H-bonds in crystalline cellulose microfibrils. Imagining a rectangular shape of the crystalline regions, the chains at the corner positions of the rectangle are most prone to solvent attack. This is because they are most exposed to the surrounding environment, and, in addition, they display the least amount of interchain interactions with adjacent hydroxy groups. The crystal structure of native cellulose has been determined by X-ray diffraction (XRD) experiments.65 Gardner and Blackwell have stated that the H-bond network in native cellulose is completely contained within the (020) plane and that it is compatible with both parallel and antiparallel chains.65 Although the bond angles and atom distances are known, it is difficult to envisage the precise three-dimensional structure of aligned cellulose chains in the crystal lattice. For better visualization purposes, Figure 12 displays a simplified projection of the (020) plane of native cellulose, only focusing on the H-bond network associated with the relevant C-3 and C-6 positions. This results in 9-membered cyclic structures

11126


Figure 13. Proposed dissolution mechanism of cellulose in an IL. Figure 15. Traditional nonderivatizing cellulose solvents ordered according to their number (n) of bridging atoms between their dipoles: N-methylmorpholine-N-oxide (NMMO), N,N-dimethylacetamide + lithium chloride (DMA + LiCl), and tetrabutylammonium fluoride + dimethylsulfoxide (TBAF + DMSO).

Figure 14. Scheme for the possible interaction of a cellulose hydroxy group with a solvent dipole: the amount of bridging molecules n determines the size of the resulting cyclic state.

within each polymer chain and 14-membered rings between the cellulose chains. It has been suggested that, during the dissolution of cellulose in ILs, the cellulosic interchain hydrogen bonds are broken apart, because of the formation of new H-bonds with the solvent (see Figure 13).15,32,37 However, the cellulose hydroxy groups themselves already do form very strong hydrogen bonds between each other. Consequently, the only feasible reason why cellulose solvents can successfully compete with the cellulose hydrogen bonds is that the newly formed H-bonds between the solvent and the cellulose chains are more stable than the initial interchain interactions of the biopolymer. This could be due to two reasons: (i) the interactions between the cellulose polymer chains and the solvent result in different cyclic states with a favored geometry compared to the initial ones in native cellulose; and/ or (ii) the cellulose solvent can form hydrogen bonds that are at least as stable as the hydrogen bonds between the cellulose hydroxy groups. The objective of the next section is to explore these possibilities. 6. Speculations on Cellulose-Dissolving Ionic Liquids The previous discussions revealed structural similarities among cellulose solvents and offered explanations for the general ability of cellulose solvents to compete successfully with the strong H-bond network in native cellulose. The objective in this section is now to link these two components together to provide a theory on the dissolution mechanism of cellulose. As mentioned earlier (vide supra), the solvent is only required to attack the C-3 and the C-6 hydroxy groups of the cellulose to separate its polymer chains (see Figure 12). However, considering that the C-6 hydroxy group also interacts with intrachain C-2 hydroxyls, additional H-bond attack at C-2 can assist in breaking the interchain H-bonds that involve the C-6 hydroxy group. However, prior to discussing selected cellulose solvents, three assumptions are made: (i) cyclic arrangements consisting of six members are favorable; (ii) at least two cellulose hydroxy groups are involved in solvent interactions; and (iii) each cellulose hydroxy group does not interact with more than two H-bond partners at the same time. Figure 14 shows schemes for the possible interaction of a cellulose hydroxy group with a

Figure 16. Cellulose-dissolving ILs ordered according to their number (n) of bridging atoms between their dipoles (top). However, their ability to delocalize their charges enables them to adopt a flexible dipolar geometry (bottom).

solvent dipole. Necessary is a dipolar character with favorably n ) 2 bridging atoms between the dipoles to form a sixmembered cyclic state. Fewer bridging atoms result in fivemembered (n ) 1) or four-membered (n ) 0) arrangements. The question arises as to whether any nonderivatizing cellulose solvents fit into this scheme. First of all, traditional solvent systems are considered with respect to their number (n) of bridging atoms between their dipoles (Figure 15). One can find everything from zero to two bridging atoms, resulting in the hypothetical formation of four- to six-membered cyclic arrangements. The bulky tetrabutylammonium cation is expected to be dependent on both the presence of a very small anion and the presence of an additional solvent molecule that is polarizable and unbulky, to provide a geometrically suited dipole for the interaction with cellulose. Admittedly, it does seem odd that NMMO, as an effective cellulose solvent, is only able to form a four-membered cyclic arrangement, because of its lack of bridging atoms between its dipoles. However, to resolve this puzzle, it is first necessary to take a closer look at cellulosedissolving ILs. With respect to cellulose-dissolving ILs, only n ) 0 or 2 bridging atoms are found at a first glance (Figure 16). However, the possibility of delocalizing the positive charge in the cation heterocycle allows the considered ILs to adjust their dipolar geometry appropriately, according to its needs. The acidity (or the presence) of the imidazolium H-2 proton is not important in this respect. This is postulated to be a crucial factor for the successful dissolution of cellulose. As shown before, small ions that are capable of hydrogen bonding are crucial for any cellulose solvent. The ability to hydrogen-bond is important to break down the intermolecular H-bonds in cellulose. Ion size


11127

Figure 17. Possible H-bond interactions between N-methylmorpholine-N-oxide (NMMO) and cellulose: three cellulosic hydrogen bonds can theoretically be attacked simultaneously.

governs ion mobility and, therefore, influences the ion’s ability to penetrate and attack this network. The smaller the ion, the more favorable. This explains why the addition of salts composed of small ions, such as lithium chloride, for example, is favorable for cellulose dissolution. It is not only that many traditional cellulose solvents rely on those salts, but ILs also increase their cellulose dissolution power if such ions are present.32,66 The disadvantage of ILs, in that respect, is due to their nature. To be molten salts at low temperatures, they must meet certain structural criteria for decreasing the attractive Coulombic forces between their cations and anions. This requires that at least one ion cannot be small anymore. Requiring that one ion, usually the cation, must be of larger size can result in a hindered dissolution power for cellulose, because of its decreased ability to penetrate the cellulosic H-bond network and to adjust its dipole in a favorable position. The important advantage of the ILs shown in Figure 16, compared to traditional cellulose solvents, is their ability to delocalize the charge of their more bulky ion and, hence, improve their ability to adapt the geometrical arrangement of their dipoles for a favorable H-bond interaction with the cellulose. This is supported by Sashina and Novoselov, who stated that the degree of charge delocalization in IL ions influences their cellulose dissolution behavior.55 Traditional nonderivatizing cellulose solvents have the disadvantage that their active components (small ions) are usually dissolved in, or associated with, an antisolvent for cellulose, such as DMSO or water, for example.16,60 Formate, acetate, and phosphate derivatives can also delocalize their negative charge, to a certain extent; this maybe helps to compensate for their bigger size. The presence of two good H-bond acceptor atoms (oxygen) in intermediate proximity on one side of the molecule and the nonfavorable substituents facing away on the opposite side can explain why those ions are still found in cellulose-dissolving ILs. However, it will always be a tradeoff between ion size and its ability to present its H-bond site at a favorable location to form thermodynamically stable H-bonds with cellulose. The possible existence of the proposed concerted arrangements between IL ions and cellulose polymer chains must be validated. Spectroscopic techniques, such as NMR or Raman experiments, can possibly give hints about the changes in the immediate environment of the solvent. Furthermore, the energetic aspects of the proposed interactions between solvent and polymer must be considered. Entropic effects are expected to influence the dissolution process as well as the thermodynamic stability of the newly formed H-bonds between the solvent and dissolved cellulose. The conclusions drawn from IL cellulose solvents shed a different light on the NMMO dissolution process. According to the proposed dissolution mechanism, it is the flexible arrangement of dipoles that is crucial for efficient attack of the

native H-bond network between cellulose polymer chains. The special arrangement of heteroatoms in the morpholine ring, together with its nonaromaticity, can explain the ability of NMMO to interact favorably with cellulose. First, the lack of a delocalized π-system within the ring keeps the heterocycle sufficiently flexible to adopt various conformations. Second, the oxygen in the position para to the N atom in the heterocycle is postulated to have the right geometry to interact with two hydroxy groups at positions C-2 and C-3 of one particular glucose unit at the same time, thus competing for both intramolecular and intermolecular interactions (see Figure 17). As mentioned previously, attack of the C-2/C-6 intrachain H-bond can assist in cleaving the corresponding C-6/C-3 interchain H-bond. Hence, the simultaneous interaction of NMMO with three hydroxy groups of the same polymer chain (C-2, C-3, and the neighbored C-6) can facilitate the separation of adjacent cellulose chains. It is possible that the C-2 and C-3 hydroxy groups are most prone to solvent attack, because they only participate in one interaction with adjacent hydroxy groups. Once deprived of its intrachain bonding partner, the hydrogen of the C-6 hydroxy group is now available to interact with the oxygen substituent of NMMO. The final step can be the cleavage of the C-6 interchain hydrogen bond and subsequent interaction with the new bonding partner NMMO. Assuming that one hydroxy group does not take part in more than two hydrogen bond interactions at the same time, the resulting cyclic state is likely to be more stable than the initial one, because only three hydrogen bonds are broken and four new ones are formed between NMMO and one cellobiose unit. Although computer studies do not support the idea of additional H-bond interactions that involve the ring oxygen in NMMO, it is interesting that the cyclic state postulated between the heterocyclic core of NMMO and the cellulose hydroxy groups consists of 14 atoms (see Figure 17).67 This is identical to the interchain ring size in native cellulose (see Figure 12) and is another indication that the resulting cyclic arrangement is most likely to be at least of equal stability, compared to the geometry inherited in native cellulose. As a consequence, NMMO can also be regarded as a dipole with n ) 2 bridging atoms and one can conclude that a second heteroatomic species at a favorable position in the IL cation possibly enhances cellulose dissolution in ILs. Based on the analysis of NMMO, one can further conclude that it is more likely for an IL to attack more than one cellulose hydroxy group at once. After refining the initial approach, dealing with bridging atoms between the solvent dipoles to a model that supports the concept of a flexible dipole arrangement due to charge delocalization, it is logical that an IL can attack more than one cellulose hydroxy group at once. Other studies support this assumption.37 Furthermore, dipolar characteristics

11128


Figure 20. Mesomeric structure of dimethylcarbamate.

Figure 18. Dipolar character in aromatics can be introduced through heteroatoms. Electronegativities increase from phosphor to oxygen. Figure 21. Structure of the phosphate anion.

Figure 19. Heterocycles with varying dipolar character: nitrogen and oxygen are both more electronegative, and phosphorus is less electronegative, than carbon.

already inherited in the nature of the cation (i.e., N-O in NMMO) can assist in breaking cellulose hydrogen bonds. To summarize the conclusions derived from the discussions above, one can now refine the initial IL ion characteristics that are favorable for enhanced cellulose dissolution. • Cation characteristics favorable for cellulose dissolution: (i) aromatic heterocycle; (ii) ability to delocalize the positive charge; (iii) second heteroatom in an aromatic ring can be favorable; (iv) inherited dipolar characteristics can be favorable. • Anion characteristics favorable for cellulose dissolution: (i) ability to act as an H-bond acceptor; (ii) substituents should not be bulky or hydrophobic; (iii) either small in size or ability to offer several H-bond acceptor sites. 7. Possible Alternatives for Cellulose-Dissolving Ionic Liquids Reflecting upon the concluded ion characteristics for cellulose-dissolving ILs is the most interesting part of this work. The objective of this section is to take advantage of the newly gained insights and apply them for proposing alternative cellulose solvents. Alternative cations and anions for cellulosedissolving ILs are proposed and discussed. Suitable cations are favorably heterocycles that can at least partially delocalize their positive charge. The presence of an inherited dipole is expected to be favorable. Dipolar characteristics can be introduced with heteroatoms positioned at opposite sides of the ring. The larger the difference in electronegativity of the heteroatom, the larger the resulting dipole (see Figure 18). It remains to be seen what is most beneficial for dissolving cellulose: (a) two heteroatoms that are more electronegative than the connecting C atoms (i.e., nitrogen and oxygen) or (b) two heteroatoms with the largest difference in electronegativity possible (i.e., phosphor and oxygen). The imidazolium cation is an example for case (a), whereas NMMO follows case (b) by exploiting its N-O dipole. An unsaturated equivalent to morpholine is 1,4-oxazine, but oxazine as a six-membered ring is not an aromatic compound, because of the presence of an sp3 ring carbon. Similar difficulties are encountered if one wants to pair phosphorus and oxygen in a six-membered heterocycle to introduce a large difference in electronegativities. However, suitable heterocycles are found in the five-membered oxazole and 1,3-oxaphosphole (see Figure 19). Suitable anions are good H-bond acceptors, preferably small in size, and without any hydrophobic substituents. The presence

of several H-bond acceptor sites can be favorable. Possibilities include fluoride (F-), chloride (Cl-), formate (HCO2-), acetate (MeCO2-), phosphate derivatives (RxRyPOR2), peroxide (O22-), superoxide (O2-), hydroxide (OH-), nitrite (NO2-), or nitrate (NO3-). Only limited data are available for IL fluorides, and their cellulose-dissolution ability is not convincing.42 A possible explanation would be that the fluoride-containing IL has a highly ordered structure in the liquid state, because of the small and very electronegative F- anions. This could inhibit the penetration of the cellulosic H-bond network, and other studies have shown that a high degree of dissociation in ILs is favorable to dissolve cellulose.37 Most IL chlorides are good cellulose solvents, but their relatively high viscosity and their corrosive behavior is not desirable.10 Alkali superoxides are known to be unstable and react with traces of water to oxygen, hydrogen peroxides, and hydroxide. In addition, peroxides must be handled with care, because of their combustive nature, and IL peroxides also are most likely to be unstable. Because of the nature of the hydroxide anion, corresponding ILs are expected to be hygroscopic and viscous. IL formates and acetates pose an alternative to the chlorides, although their long-term stability at elevated temperatures still must be established. However, the CO2synthon can assist in finding anions with similar dissolution ability but increased thermal stability. Dimethylcarbamate (see Figure 20) could qualify as such, and the mesomeric effect of the N atom also offers the possibility of two H-bond acceptor sites. It remains to be seen whether the bulkiness of the anion is an issue. The nitrogen equivalent to the CO2- synthon is NO2-; thus, nitrite and nitrate are additional ions that must be considered as alternatives. Furthermore, anions based on phosphate derivatives are known to show moderate dissolution ability for cellulose and exhibit reasonable thermal stability.68 The phosphate anion can offer multiple sites that qualify as H-bond acceptors (see Figure 21). The triple negative charge enables the anion to be paired with three cations. Even mixtures of cellulose-dissolving heterocyclic cations and smaller cations (i.e., lithium), paired with the phosphate counterion, could prove beneficial for cellulose dissolution. The possibilities are manyfold and must be explored. Figure 22 summarizes the suggested alternatives for IL ions that are expected to show enhanced cellulose solubility. 8. Conclusion Until now, the search for ionic liquids (ILs) with cellulosedissolution ability has mainly been based on trial and error. The objective of this work was to identify general structural similarities among cellulose solvents to propose alternative ILs with enhanced dissolution power for the biopolymer. It is reasoned that both IL cations and anions are important for efficient cellulose dissolution. IL ions that are found in cellulosedissolving ILs were identified and compared to nondissolving ILs, as well as to traditional cellulose solvents. It was found


Figure 22. Proposed IL ions for enhanced cellulose dissolution: oxazolium, 1,3-oxaphospholium, phosphate, dimethylcarbamate, nitrate, and nitrite.

that all nonderivatizing cellulose solvents share one similarity: they can form hypothetical cyclic arrangements with themselves. Structural analysis of the native cellulose crystal lattice reveals that only one specific hydrogen bond (H-bond) between the glucose monomers in cellulose must be broken to achieve polymer chain separation. It is suggested that efficient cellulose solvents are able to form cyclic arrangements with the cellulose during dissolution. The resulting cyclic arrangements are expected to be of similar stability to the interchain H-bonds in native cellulose. Cation characteristics that are expected to enhance cellulose dissolution are (i) an aromatic heterocycle with the ability to sufficiently delocalize the positive charge, and (ii) a second heteroatom in the aromatic ring that results in a dipolar character of the cation. Anion characteristics that are expected to enhance cellulose dissolution are (i) small-sized H-bond acceptors, (ii) substituents that should neither be bulky nor hydrophobic, and (iii) the ability to offer several H-bond acceptor sites. These characteristics enable the resulting IL to offer a flexible arrangement of its dipoles to arrange in such a way that an energetically favored cyclic arrangement with cellulose can be formed. Proposed alternatives for cellulosedissolving IL ions are oxazolium, 1,3-oxaphospholium, dimethylcarbamate, and phosphate. Furthermore, it is suggested that mixtures of ILs and salts with small-sized ions, such as lithium chloride (LiCl) for example, could result in highly efficient cellulose solvents that are more environmentally friendly than traditional cellulose solvents. The objective of this work is to provide guidance for the development of highly effective IL cellulose solvents. Future studies are needed to confirm, reject, or refine the postulates of this work. The discovery of an “Alkahest” for cellulose would not only boost the ecofriendly extraction of wood cellulose in industry, but it would also allow the production of cellulose composites with superior mechanical properties. Both are of equal importance if we intend to become more sustainable through biofuel and biocomposite production. Acknowledgment The authors would like to express their gratitude to John M. Prausnitz, Robin D. Rogers, and Joa˜o Coutinho for providing critical and helpful comments during the preparation of the manuscript. The New Zealand Foundation for Research, Science and Technology (FRST) is acknowledged for financial support. Literature Cited (1) Hassan, A. Y. Transfer of Islamic Technology to the West, Part III, Technology Transfer in the Chemical Industries, http://www.history-sciencetechnology.com/Articles/articles 72.htm (accessed November 2009). (2) de Vains, A. R. British Patent GB189561, 1921. (3) Argyropoulos, D. S.; Xie, H. World Patent WO/2008/098037, 2008.

11129

(4) Huber, G. W.; Iborra, S.; Corma, A. Synthesis of Transportation Fuels from Biomass: Chemistry, Catalysts, and Engineering. Chem. ReV. 2006, 106, 4044. (5) Barbosa, M. A.; Granja, P. L.; Barrias, C. C.; Amaral, I. F. Polysaccharides as scaffolds for bone regeneration. ITBM-RBM 2005, 26, 212. (6) Czaja, W. K.; Young, D. J.; Kawecki, M.; Brown, R. M., Jr. The future prospects of microbial cellulose in biomedical applications. Biomacromolecules 2007, 8, 1. (7) Lee, K. Y.; Mooney, D. J. Hydrogels for Tissue Engineering. Chem. ReV. 2001, 101, 1869. (8) Ramakrishna, S.; Mayer, J.; Wintermantel, E.; Leong, K. W. Biomedical applications of polymer-composite materials: a review. Compos. Sci. Technol. 2001, 61, 1189. (9) Li, Q.; He, Y.-C.; Xian, M.; Jun, G.; Xu, X.; Yang, J.-M.; Li, L.-Z. Improving enzymatic hydrolysis of wheat straw using ionic liquid 1-ethyl3-methyl imidazolium diethyl phosphate pretreatment. Biores. Technol. 2009, 100 (14), 3570. (10) Pinkert, A.; Marsh, K. N.; Pang, S.; Staiger, M. P. Ionic liquids and their interaction with cellulose. Chem. ReV. 2009, 109 (12), 6712. (11) Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. Dissolution of cellulose with ionic liquids. J. Am. Chem. Soc. 2002, 124, 4974. (12) Ngo, H. L.; LeCompte, K.; Hargens, L.; McEwen, A. B. Thermal properties of imidazolium ionic liquids. Thermochim. Acta 2000, 97, 357. (13) Forsyth, S. A.; Pringle, J. M.; MacFarlane, D. R. Ionic LiquidssAn Overview. Aust. J. Chem. 2004, 57, 113. (14) Duchemin, B. J. C. Structure, property and processing relationships of all-cellulose composites, Ph.D. thesis, University of Canterbury, Christchurch, New Zealand, 2008. (15) Feng, L. Chen, Z.-l. Research progress on dissolution and functional modification of cellulose in ionic liquids. J. Mol. Liq. 2008, 142, 1. (16) Heinze, T.; Koschella, A. Solvents applied in the field of cellulose chemistry: A mini review. Polı´meros 2005, 15, 84. (17) Saalwachter, K.; Burchard, W.; Klufers, P.; Kettenbach, G.; Mayer, P.; Klemm, D.; Dugarmaa, S. Cellulose Solutions in Water Containing Metal Complexes. Macromolecules 2000, 33, 4094. (18) Heinze, T. New ionic polymers by cellulose functionalization. Macromol. Chem. Phys. 1998, 199, 2341. (19) Leipner, H.; Fischer, S.; Brendler, E.; Voigt, W. Structural changes of cellulose dissolved in molten salt hydrates. Macromol. Chem. Phys. 2000, 201, 2041. (20) Potthast, A.; Rosenau, T.; Sixta, H.; Kosma, P. Degradation of cellulosic materials by heating in DMAc/LiCl. Tetrahedron Lett. 2002, 43, 7757. (21) Le Moigne, N.; Spinu, M.; Heinze, T.; Navard, P. Restricted dissolution and derivatization capacities of cellulose fibres under uniaxial elongational stress. Polymer 2010, 51, 447. (22) Ramos, L.; Frollini, E.; Koschella, A.; Heinze, T. Benzylation of cellulose in the solvent dimethylsulfoxide/tetrabutylammonium fluoride trihydrate. Cellulose 2005, 12, 607. (23) Rosenau, T.; Potthast, A.; Sixta, H.; Kosma, P. The chemistry of side reactions and byproduct formation in the system NMMO/cellulose (Lyocell process). Prog. Polym. Sci. 2001, 26, 1763. (24) Sun, N.; Rahman, M.; Qin, Y.; Maxim, M. L.; Rodriguez, H.; Rogers, R. D. Complete dissolution and partial delignification of wood in the ionic liquid 1-ethyl-3-methylimidazolium acetate. Green Chem. 2009, 11, 646. (25) Zhu, S.; Wu, Y.; Chen, Q.; Yu, Z.; Wang, C.; Jin, S.; Ding, Y.; Wu, G. Dissolution of cellulose with ionic liquids and its application: A mini-review. Green Chem. 2006, 8, 325. (26) Cuissinat, C.; Navard, P.; Heinze, T. Swelling and dissolution of cellulose. Part IV: Free floating cotton and wood fibres in ionic liquids. Carbohydr. Polym. 2008, 72, 590. (27) Heinze, T.; Schwikal, K.; Barthel, S. Ionic liquids as reaction medium in cellulose functionalization. Macromol. Biosci. 2005, 5, 520. (28) Kosan, B.; Schwikal, K.; Meister, F. Solution states of cellulose in selected direct dissolution agents. Cellulose 2010, 17, 495. (29) Remsing, R. C.; Swatloski, R. P.; Rogers, R. D.; Moyna, G. Mechanism of cellulose dissolution in the ionic liquid 1-n-butyl-3methylimidazolium chloride: a 13C and 35/37Cl NMR relaxation study on model systems. Chem. Commun. 2006, 1271. (30) Remsing, R. C.; Hernandez, G.; Swatloski, R. P.; Massefski, W. W.; Rogers, R. D.; Moyna, G. Solvation of Carbohydrates in N,N-Dialkylimidazolium Ionic Liquids: A Multinuclear NMR Spectroscopy Study. J. Phys. Chem. B 2008, 112, 11071. (31) Brandt, A.; Hallett, J. P.; Leak, D. J.; Murphy, R. J.; Welton, T. The effect of the ionic liquid anion in the pretreatment of pine wood chips. Green Chem. 2010, 12, 672.

11130


(32) Xu, A.; Wang, J.; Wang, H. Effects of anionic structure and lithium salts addition on the dissolution of cellulose in 1-butyl-3-methylimidazoliumbased ionic liquid solvent systems. Green Chem. 2010, 12, 268. (33) Fukaya, Y.; Hayashi, K.; Wada, M.; Ohno, H. Cellulose dissolution with polar ionic liquids under mild conditions: Required factors for anions. Green Chem. 2008, 10, 44. (34) Youngs, T. G. A.; Holbrey, J. D.; Deetlefs, M.; Nieuwenhuyzen, M.; Gomes, M. F. C.; Hardacre, C. A Molecular Dynamics Study of Glucose Solvation in the Ionic Liquid 1,3-Dimethylimidazolium Chloride. ChemPhysChem 2006, 7, 2279. (35) Youngs, T. G. A.; Hardacre, C.; Holbrey, J. D. Glucose Solvation by the Ionic Liquid 1,3-Dimethylimidazolium Chloride: A Simulation Study. J. Phys. Chem. B 2007, 111, 13765. (36) Zhang, H.; Wu, J.; Zhang, J.; He, J. 1-Allyl-3-methylimidazolium chloride room temperature ionic liquid: A new and powerful nonderivatizing solvent for cellulose. Macromolecules 2005, 38, 8272. (37) Zhang, J.; Zhang, H.; Wu, J.; Zhang, J.; He, J.; Xiang, J. NMR spectroscopic studies of cellobiose solvation in EmimAc aimed to understand the dissolution mechanism of cellulose in ionic liquids. Phys. Chem. Chem. Phys. 2010, 12, 1941. (38) Liu, H.; Sale, K. L.; Holmes, B. M.; Simmons, B. A.; Singh, S. Understanding the Interactions of Cellulose with Ionic Liquids: A Molecular Dynamics Study. J. Phys. Chem. B 2010, 114, 4293. (39) Lindman, B.; Karlstro¨m, G.; Stigsson, L. On the Mechanism of Dissolution of Cellulose. J. Mol. Liq. 2010, DOI: 10.1016/j.molliq. 2010.04.016. (40) Scho¨bitz, M.; Meister, F.; Heinze, T. Unconventional Reactivity of Cellulose Dissolved in Ionic Liquids. Macromol. Symp. 2009, 280, 102. (41) Luo, H.-M.; Li, Y.-Q.; Zhou, C.-R. Study on the dissolubility of cellulose in the functionalized ionic liquid. Gaofenzi Cailiao Kexue Yu Gongcheng 2005, 21, 233. (42) Vitz, J.; Erdmenger, T.; Haensch, C.; Schubert, U. S. Extended dissolution studies of cellulose in imidazolium based ionic liquids. Green Chem. 2009, 11, 417. (43) Weingaertner, H. Understanding ionic liquids at the molecular level: Facts, problems, and controversies. Angew. Chem., Int. Ed. 2008, 47, 654. (44) Yang, F.; Li, L.; Li, Q.; Tan, W.; Liu, W.; Xian, M. Enhancement of enzymatic in situ saccharification of cellulose in aqueous-ionic liquid media by ultrasonic intensification. Carbohydr. Polym. 2010, 81, 311. (45) Zavrel, M.; Bross, D.; Funke, M.; Bu¨chs, J.; Spiess, A. C. Highthroughput screening for ionic liquids dissolving (ligno-)cellulose. Bioresour. Technol. 2009, 100 (9), 2580. (46) Zhao, H.; Baker, G. A.; Song, Z.; Olubajo, O.; Crittle, T.; Peters, D. Designing enzyme-compatible ionic liquids that can dissolve carbohydrates. Green Chem. 2008, 10, 696. (47) Bicak, N. A new ionic liquid: 2-hydroxy ethylammonium formate. J. Mol. Liq. 2004, 116, 15. (48) Pinkert, A.; Marsh, K. N.; Pang, S. Alkanolamine ionic liquids and their inability to dissolve crystalline cellulose. Ind. Eng. Chem. Res. 2010, DOI: 10.1021/ie101250v. (49) Fukaya, Y.; Sugimoto, A.; Ohno, H. Superior solubility of polysaccharides in low viscosity, polar, and halogen-free 1,3-dialkylimidazolium formates. Biomacromolecules 2006, 7, 3295. (50) Baker, S. N.; Baker, G. A.; Bright, F. V. Temperature-dependent microscopic solvent properties of “dry” and “wet” 1-butyl-3-methylimidazolium hexafluorophosphate: Correlation with ET(30) and Kamlet-Taft polarity scales. Green Chem. 2002, 4, 165. (51) Weingaertner, H.; Sasisanker, P.; Daguenet, C.; Dyson, P. J.; Krossing, I.; Slattery, J. M.; Schubert, T. The Dielectric Response of RoomTemperature Ionic Liquids: Effect of Cation Variation. J. Phys. Chem. B 2007, 111, 4775.

(52) Weingaertner, H. The static dielectric constant of ionic liquids. Z. Phys. Chem. 2006, 220, 1395. (53) Chiappe, C.; Pieraccini, D. Ionic liquids: solvent properties and organic reactivity. J. Phys. Org. Chem. 2005, 18, 275. (54) Crowhurst, L.; Mawdsley, P. R.; Perez-Arlandis, J. M.; Salter, P. A.; Welton, T. Solvent-solute interactions in ionic liquids. Phys. Chem. Chem. Phys. 2003, 5, 2790. (55) Sashina, E.; Novoselov, N. Effect of structure of ionic liquids on their dissolving power toward natural polymers. Russ. J. Gen. Chem. 2009, 79, 1057. (56) Marsh, K.; Pinkert, A.; Pang, S.; Staiger, M. Solubility of cellulose and wood in ionic liquids. Presented at the 17th Symposium on Thermophysical Properties, Boulder, CO, 2009. (57) Balevicius, V.; Gdaniec, Z.; Aidas, K. NMR and DFT study on media effects on proton transfer in hydrogen bonding: Concept of molecular probe with an application to ionic and super-polar liquids. Phys. Chem. Chem. Phys. 2009, 11, 8592. (58) Krekeler, C.; Dommert, F.; Schmidt, J.; Zhao, Y. Y.; Holm, C.; Berger, R.; Site, L. D. Electrostatic properties of liquid 1,3-dimethylimidazolium chloride: Role of local polarization and effect of the bulk. Phys. Chem. Chem. Phys. 2010, 12, 1817. (59) Spange, S.; Reuter, A.; Vilsmeier, E.; Heinze, T.; Keutel, D.; Linert, W. Determination of empirical polarity parameters of the cellulose solvent N,N-dimethylacetamide/LiCl by means of the solvatochromic technique. J. Polym. Sci., Part A: Polym. Chem. 1998, 36, 1945. ¨ stlund, A.; Lundberg, D.; Nordstierna, L.; Holmberg, K.; Nyde´n, (60) O M. Dissolution and Gelation of Cellulose in TBAF/DMSO Solutions: The Roles of Fluoride Ions and Water. Biomacromolecules 2009, 10, 2401. (61) Barthel, S.; Heinze, T. Acylation and carbanilation of cellulose in ionic liquids. Green Chem. 2006, 8, 301. (62) El Seoud, O. A.; Koschella, A.; Fidale, L. C.; Dorn, S.; Heinze, T. Applications of Ionic Liquids in Carbohydrate Chemistry: A Window of Opportunities. Biomacromolecules 2007, 8, 2629. (63) Kosan, B.; Michels, C.; Meister, F. Dissolution and forming of cellulose with ionic liquids. Cellulose 2008, 15, 59. (64) Walker, J. Primary Wood Processing: Principles and Practice, 1st ed.; Chapman and Hall: London; 1993. (65) Gardner, K. H.; Blackwell, J. Structure of native cellulose. Biopolymers 1974, 13, 1975. (66) Roder, T.; Potthast, A.; Rosenau, T.; Kosmsa, P.; Baldinger, T.; Morgenstern, B.; Glatter, O. The effect of water on cellulose solutions in DMAc/LiCl. Macromol. Symp. 2002, 190, 151. (67) Marhoefer, R. Computersimulationen zum Loeseverhalten von Cellulose in aliphatischen Amin-N-oxiden, Ph.D. Thesis, University of Technology, Darmstadt; Germany, 2004. (68) Kosmulski, M.; Gustafsson, J.; Rosenholm, J. B. Thermal stability of low temperature ionic liquids revisited. Thermochim. Acta 2004, 412, 47.

Andre´ Pinkert, Kenneth N. Marsh,* and Shusheng Pang Department of Chemical and Process Engineering, UniVersity of Canterbury, Christchurch, New Zealand ReceiVed for reView June 8, 2010 IE1006596

Reflections on the Solubility of Cellulose - Industrial & Engineering

Recommend Documents