Video Upload Options

1. Introduction

Cellulose is the most abundant biopolymer on Earth and the principal structural polysaccharide of the plant cell wall, with an annual terrestrial production on the order of 10¹² tonnes of biomass ^[1]. Chemically it is a linear homopolysaccharide built from β-D-glucopyranose units joined by β-1,4-glycosidic bonds, so that each anhydroglucose repeat unit carries only three hydroxyl groups. This apparent simplicity conceals the difficulty of working with the material. The hydroxyl groups are arranged equatorially along rigid, ribbon-like chains and engage in a dense, cooperative network of intra- and intermolecular hydrogen bonds; the resulting fibrillar architecture is insoluble in water, in aqueous acids and bases of moderate concentration, and in virtually every common organic solvent, and it decomposes on heating well before any melt transition can be reached ^[2][3].

The media capable of overcoming this intractability are collectively termed cellulose solvents. Their importance extends well beyond the laboratory, because practically all shaping of cellulose—fibres, films, membranes, sponges, beads and monoliths—proceeds from solution. The regenerated-cellulose industry, which produces several million tonnes of fibre and film each year, is built entirely on such solvents: the viscose (xanthate) and cuprammonium processes that dominated the twentieth century, and the N-methylmorpholine N-oxide (NMMO) based Lyocell process that has expanded rapidly since the 1990s ^[4]. In research, cellulose solvents additionally provide access to homogeneous chemical modification, to the determination of molar mass and chain conformation, and to the fractionation of lignocellulosic biomass into its constituent polymers ^[5][6].

This entry summarizes the molecular and thermodynamic reasons for the insolubility of cellulose, classifies the principal solvent families, reviews what is known about the state of cellulose once dissolved, and surveys applications together with the practical constraints—recyclability, energy demand, viscosity and solvent stability—that determine which solvents are viable at industrial scale.

2. Historical Development

2.1. Derivatizing Routes and the Viscose Era

The first technically useful cellulose solutions were obtained by converting the polymer into a soluble derivative. Cellulose nitrate, prepared by nitration in 1846, became the basis of collodion and of the first semi-synthetic plastics and fibres. In 1857 Schweizer discovered that copper(II) hydroxide dissolved in aqueous ammonia—Schweizer's reagent—dissolves cellulose directly, an observation that led to the industrial cuprammonium rayon process around the turn of the twentieth century. The decisive development was the viscose process, patented by Cross, Bevan and Beadle in 1892: cellulose is treated with carbon disulfide in alkali to give soluble cellulose xanthate, which is spun into an acid bath where the xanthate decomposes and cellulose is regenerated. Commercial fibre production began in about 1905 and the process became dominant because carbon disulfide is inexpensive and the dope can be spun into strong textile filaments ^[6]. Cellulose acetate, developed industrially in the early twentieth century, follows a different logic: the polymer is derivatized to a partially hydrolysed acetate that is soluble in acetone, and the fibre is formed from the volatile organic solvent rather than from water ^[7].

2.2. Direct Dissolution in the Second Half of the Twentieth Century

The search for media that dissolve cellulose without any chemical derivatization produced three systems that remain important today. The lithium chloride/N,N-dimethylacetamide (LiCl/DMAc) system, reported in the late 1970s, became the reference solvent for cellulose analysis, including size-exclusion chromatography and light scattering, and also for homogeneous derivatization; it requires careful activation of the polymer by solvent exchange ^[8][9]. N-methylmorpholine N-oxide (NMMO) was patented as a cellulose solvent in 1969 and was developed into an industrial process—the basis of Lyocell fibres—during the 1970s and 1980s. NMMO monohydrate melts at about 76 °C, dissolves cellulose without activation or derivatization, and can be recovered almost quantitatively, which makes the process markedly cleaner than the viscose route ^[10]. A third system, cellulose in dimethyl sulfoxide containing tetrabutylammonium fluoride (DMSO/TBAF), was introduced at the end of the 1990s as an activation-free laboratory solvent particularly well suited to homogeneous reactions ^[11].

2.3. The Twenty-First Century

Two developments reshaped the field after 2000. First, Swatloski and co-workers demonstrated in 2002 that certain ionic liquids dissolve cellulose in high concentration without derivatization, initiating an intense wave of research into designable, non-volatile solvents ^[12]. Second, precooled aqueous solutions of alkali and urea—NaOH/urea and LiOH/urea—were shown to dissolve cellulose rapidly at low temperature using inexpensive, non-toxic components, ^[13] and fibres were subsequently spun from NaOH/urea solutions ^[14]. Aerogels and other porous materials were prepared from the same solvents later in the decade ^[15]. More recently, the superbase-derived ionic liquid [DBNH][OAc] has been developed into the Ioncell-F spinning process, ^[16] and deep eutectic solvents based on choline chloride and urea have been introduced as cheap, biodegradable alternatives, mainly as pretreatment media ^[17].

3. Physicochemical Basis of Cellulose Dissolution

3.1. Molecular and Supramolecular Structure of Cellulose

Each anhydroglucose unit of cellulose carries one primary (O6‑H) and two secondary (O2‑H and O3‑H) hydroxyl groups in equatorial positions, while the axial C‑H groups form two comparatively non-polar faces. Intramolecular hydrogen bonds, notably O3‑H···O5, stiffen the chain into a flat ribbon, and intermolecular O6‑H···O3 bonds link neighbouring chains into sheets that stack through van der Waals interactions.^[1] The consequence is a molecule that behaves in part as an amphiphile: its equatorial edges are hydrophilic and its axial faces hydrophobic, a duality that has direct implications for solvent selection ^[3]. Native cellulose occurs as the metastable cellulose I allomorph (with the Iα and Iβ sub-forms), whereas dissolution and regeneration generally yield the thermodynamically more stable cellulose II ^[18]. Real cellulose is only partly crystalline, with a crystallinity index of roughly 40–70 % in native sources, and the amorphous domains are far more accessible to solvents than the crystalline cores. Accessibility also depends on the drying history of the sample: drying of pulp causes irreversible aggregation of the fibrils, known as hornification, which substantially impedes subsequent dissolution ^[2].

3.2. Why Cellulose Is Difficult to Dissolve

Dissolution requires a negative free energy of mixing. The enthalpic term must compensate for breaking cellulose–cellulose hydrogen bonds and for disrupting the hydrophobic association between the ribbon faces, neither of which is offset by the solvation of the freed hydroxyl groups in a poor solvent such as water. The entropic term is unfavourable because the extremely rigid chains gain little conformational freedom on dispersion. An effective cellulose solvent therefore normally combines two features: a small, strongly hydrogen-bond-accepting species present at high effective concentration (hydroxide, fluoride, chloride, carboxylate or an N-oxide oxygen), which competes for the hydroxyl protons, and a counter-ion or co-solvent that interacts with the remaining proton-donor sites and with the hydrophobic faces of the ribbon ^[19]. Water is a particularly poor solvent for cellulose and also a strong competitor for the solvating species, which is why many effective systems are essentially water-poor or operate only within a narrow water-content window; this is a recurring theme for NMMO, ionic liquids and alkali/urea mixtures ^[10][21][22]. Because cellulose is amphiphilic, solvent quality is determined as much by the solvation of the non-polar faces as by that of the polar groups ^[3].

3.3. Kinetic Constraints and Activation

Even when dissolution is thermodynamically possible, the process is often slow and spatially uneven. Solvent penetration begins in the amorphous domains and the crystalline cores dissolve last, so swelling typically precedes dissolution; a viscous gel layer may form at the fibre surface and retard further penetration ^[2][6]. Practical dissolution therefore frequently requires activation of the substrate—the use of never-dried pulp, or solvent exchange (water to acetone to DMAc in the LiCl/DMAc system)—as well as the selection of a suitable temperature window. These windows are narrow and characteristic: precooled alkali/urea solutions must be kept well below ambient temperature, whereas NMMO and ionic liquid systems operate between about 60 °C and 120 °C ^[10][13].

4. Classification of Cellulose Solvents

Cellulose solvents are conventionally divided according to whether dissolution involves the transient formation of a covalent derivative. Derivatizing solvents convert hydroxyl groups into esters, ethers or related functions that are soluble in a suitable medium, and the parent polymer is restored during regeneration. Non-derivatizing solvents dissolve the unmodified polymer directly; they are further subdivided according to the chemical nature of the active species. The boundary between the two groups is not absolute, since some nominally non-derivatizing media, such as concentrated formic acid or urea/alkali systems, involve weak or transient acylation ^[6]. Figure 1 summarizes the principal families and their representative members.

Figure 1. Classification of cellulose solvent systems. Derivatizing media dissolve cellulose through a transient covalent derivative that is cleaved during regeneration; non-derivatizing media dissolve the unmodified polymer directly. Representative members of each family are listed.

4.1. Derivatizing Solvents

The viscose system remains the most widely used derivatizing route. Cellulose is treated with carbon disulfide in concentrated alkali to give cellulose xanthate with a degree of substitution of roughly 0.5; the xanthate dissolves in dilute aqueous sodium hydroxide to form the spinning dope, and coagulation in an acid bath reverses the derivatization. The chemistry is robust and accepts a wide range of pulp grades, but carbon disulfide is toxic and the process emits carbon disulfide and hydrogen sulfide, so only part of the reagent is recovered ^[6]. Cellulose carbamate offers a sulfur-free alternative: reaction with urea at about 130–150 °C introduces carbamate groups that render the polymer soluble in dilute aqueous alkali, and the groups are cleaved during regeneration. This route, known since the 1980s, has been revisited using urea-based deep eutectic solvents for the activation step ^[22]. Concentrated formic acid dissolves cellulose with partial esterification to the formate, which hydrolyses on coagulation; trifluoroacetic anhydride in dimethyl sulfoxide or dichloromethane gives the corresponding trifluoroacetate, dimethyl sulfoxide with paraformaldehyde gives methylol cellulose, and dinitrogen tetroxide in dimethylformamide gives cellulose nitrite ^[6][18]. The common requirement is that the derivatization must be fully reversible, and the common drawback is the additional reagent, the by-products, and the risk of an uneven distribution of substituents.

4.2. Aqueous Non-Derivatizing Systems

Aqueous sodium hydroxide alone dissolves cellulose only at high alkali concentration and low temperature, and effectively only for celluloses of low degree of polymerization or reduced crystallinity, such as steam-exploded or previously regenerated material ^[18][23]. The addition of urea or thiourea widens the practical window considerably. Precooled NaOH/urea, NaOH/thiourea and LiOH/urea solutions dissolve cellulose within minutes at about −12 °C; the components are cheap and non-toxic, but the achievable cellulose concentration is generally limited to a few weight percent, dissolution is confined to a narrow composition range, and the solutions gel on heating ^[13][24][25]. Fibres, films, membranes, microspheres, hydrogels and aerogels have all been prepared from these solutions ^{[14][15][26][27]}. Related aqueous systems based on quaternary ammonium hydroxides, or on mixtures of hydroxide bases, dissolve cellulose at higher concentration and have been studied specifically with respect to the aggregation state of the dissolved chains ^[28][29]. Molten salt hydrates such as ZnCl₂·4H₂O, LiSCN·2H₂O and Ca(SCN)₂ dissolve cellulose at high salt concentration and are inexpensive, although they are corrosive and require recovery ^[30]; aqueous zinc chloride solutions have been used to prepare all-cellulose materials in which dissolution and regeneration are combined in a single step ^[31].

4.3. Organic Electrolyte and Amine-Oxide Systems

In LiCl/DMAc the chloride ion acts as the hydrogen-bond acceptor, and the lithium ion is coordinated both by the solvent and by the cellulose hydroxyls. Dissolution requires activation of the substrate and LiCl contents of roughly 5–10 wt %, and it is usually carried out at low cellulose concentration, which is well matched to analytical applications such as size-exclusion chromatography ^[8][9]. DMSO/TBAF, in which the fluoride ion is the active hydrogen-bond acceptor, dissolves cellulose without any activation and is an excellent medium for homogeneous derivatization, but the reagent is costly and sensitive to water ^[11]. NMMO monohydrate, in which the strongly polar N‑O group acts as the hydrogen-bond acceptor, dissolves cellulose without activation or derivatization at about 90–120 °C to give dopes containing roughly 10–15 wt % cellulose; these are spun into fibres of high tensile strength, and the solvent is recovered to better than 99 % in a closed industrial loop ^[10]. Further members of this family include ammonium fluoride in DMSO and the dinitrogen tetroxide and paraformaldehyde systems mentioned above.

4.4. Ionic Liquids

Ionic liquids are salts that melt below about 100 °C, and the first report of cellulose dissolution in 1-butyl-3-methylimidazolium chloride in 2002 opened a new chapter in the field ^[12]. Their anions—chloride, formate, acetate and other carboxylates—are strong hydrogen-bond acceptors that disrupt the cellulose network, while the cations, which are frequently amphiphilic, interact with the hydrophobic ribbon faces ^[21]. Cellulose loadings of about 20–25 wt % can be reached at 60–120 °C without any activation, and the negligible vapour pressure of ionic liquids makes them attractive both for fibre formation and for homogeneous derivatization, since they can act simultaneously as solvent and as reaction medium ^{[32][33][34][35]}. The superbase-derived ionic liquid [DBNH][OAc] with water as co-solvent is used in the Ioncell-F process, in which air-gap dry-jet wet spinning of concentrated cellulose solutions yields high-strength fibres ^[16]. The limitations are equally well documented: the solvents are expensive and viscous, recovery from dilute aqueous streams is energy-intensive, they are hygroscopic so that absorbed water sharply reduces solvent quality, coloured impurities must be removed before reuse, and the toxicity and biodegradability of many ionic liquids remain open questions ^[21][34][35].

4.5. Deep Eutectic Solvents

Deep eutectic solvents are mixtures of a hydrogen-bond acceptor, typically a quaternary ammonium salt such as choline chloride, and a hydrogen-bond donor such as urea, oxalic acid or a polyol, which form a liquid with a melting point far below that of either component. They are cheap, easy to prepare, of low toxicity and readily biodegradable, and they dissolve cellulose, at least partially, although the achievable concentration and degree of polymerization are generally lower than for ionic liquids and heating is usually required. Their most convincing role to date is as pretreatment media for the production of cellulose nanofibrils, for the fractionation of biomass and as benign reaction media; they have also been used in the synthesis of cellulose carbamate ^[17][22].

4.6. Hybrid and Emerging Systems

Recent work has moved towards mixtures and process-integrated systems rather than single solvents: ionic liquids blended with polar aprotic co-solvents to reduce viscosity, aqueous alkali/urea solutions containing additives, and switchable solvents whose polarity can be altered by carbon dioxide. A parallel trend is the direct dissolution of never-dried biomass, in which one solvent both fractionates lignocellulose and provides a solution from which cellulose can be regenerated. Table 1 compares the principal systems.

Table 1. Comparison of the main cellulose solvent systems. Values are indicative ranges reported in the literature and depend on the cellulose source, the degree of polymerization and the activation history of the sample.

5. The State of Cellulose in Solution

Dissolution and molecular dispersion are not synonymous, and the difference is central to the interpretation of cellulose chemistry in solution. Even optically clear and macroscopically homogeneous cellulose solutions commonly retain a fraction of chains in an associated form, as indicated by light scattering, size-exclusion chromatography, rheology and cryogenic transmission electron microscopy; the extent of this aggregation depends on the solvent, on the cellulose concentration and degree of polymerization, and on the dissolution protocol. For this reason, a cellulose system is often better described as colloidal than as a true molecular solution ^[5][36][37]. The practical consequences are considerable: solution-derived quantities such as intrinsic viscosity or apparent molar mass must be interpreted with care, and reactions performed in such media may proceed with less than perfect homogeneity, because the interior of an aggregate is less accessible than the surrounding solvated chains. A related observation is that regenerated cellulose often shows a residual structural memory of the starting material, which is attributed to incomplete disruption of the original supramolecular order before regeneration ^[37].

Because the direct observation of dissolved chains is difficult, the occurrence of dissolution is usually inferred from the properties of the regenerated material. The disappearance of the cellulose I diffraction pattern and its replacement by that of cellulose II is widely taken as evidence that the original crystalline order was destroyed and that the polymer was reorganized from solution ^[10][14][21]. A combination of techniques is used to characterize the solution itself: carbon-13 nuclear magnetic resonance in DMSO/TBAF or in ionic liquids, size-exclusion chromatography in LiCl/DMAc or LiCl/DMSO for the molar-mass distribution, viscometry with Mark–Houwink relations for the degree of polymerization, static and dynamic light scattering, cryogenic electron microscopy, and rheology ^[8][37].

6. Properties of Cellulose Solutions

Cellulose chains are stiff and extended, with a persistence length of the order of several nanometres, so the concentration at which the chains begin to overlap is very low—fractions of a weight percent for high-molar-mass material. Consequently, almost every practically relevant dope is deep in the entangled regime, and its viscosity rises steeply with cellulose concentration and molar mass while showing pronounced shear thinning. Viscosity is therefore not a secondary property but the central engineering constraint of cellulose processing, dictating the temperatures, pressures and residence times of spinning and casting operations ^[4]. Most systems follow the usual Arrhenius temperature dependence, but precooled aqueous alkali/urea solutions behave in the opposite sense: they gel on heating, because dissolution in these media is entropy-driven and the solvent structure around the chain is destabilized as the temperature rises ^[13][26]. With increasing concentration, cellulose–NMMO and cellulose–ionic liquid dopes may pass into an anisotropic state in which the chains adopt a preferred orientation; the extent of this ordering influences the structure and mechanical properties of the regenerated fibre ^[10]. The water content of the solvent is a further critical variable: small amounts of water act as an antisolvent in NMMO, ionic liquid and alkali/urea systems, and the composition of the ternary alkali–urea–water mixture must lie inside a fairly narrow window for a stable solvent to exist at all ^[10][20][21]. Finally, cellulose is not chemically inert under dissolution conditions, and prolonged heating in high-boiling solvents can cause chain scission, so regenerated products frequently exhibit a lower degree of polymerization than the starting pulp ^[34].

7. Applications

Cellulose solvents are used wherever cellulose must be reshaped, modified or separated, and Figure 2 shows the general sequence from feedstock to product, together with the solvent recovery step that makes the sequence industrially sustainable.

Figure 2. Dissolution, shaping and regeneration of cellulose. The feedstock is dissolved, shaped in the solution state and regenerated by coagulation; the solvent is recovered and returned to the process. Industrial and laboratory implementations of the cycle are indicated.

7.1. Fibres, Films and Nonwovens

Fibre formation is by far the largest application. Viscose, cuprammonium and Lyocell fibres are all produced by extruding a cellulose solution into a coagulation bath, and the choice of solvent determines the structure of the resulting filament and the environmental profile of the process ^[4][10]. Lyocell fibres, spun from NMMO solutions, are characterized by high tensile strength and a high wet modulus, and the almost complete recovery of the solvent has made the process the benchmark for closed-loop cellulose processing ^[10]. The ionic liquid based Ioncell-F process has demonstrated fibres of comparable or higher strength using a non-volatile solvent ^[16]. Regenerated cellulose films, including cellophane produced from viscose, remain important in packaging, while nonwovens and hollow-fibre geometries are produced from the same dopes.

7.2. Homogeneous Chemical Modification

Because cellulose must normally be derivatized in the solid state, conventional industrial esterification and etherification reactions give products in which the substituent distribution is influenced by accessibility gradients within and between fibrils. Dissolution removes this constraint and allows the degree of substitution and, in favourable cases, the regioselectivity of the reaction to be controlled, while avoiding the side reactions and chain degradation associated with heterogeneous conditions. Cellulose esters, carbamates, sulfonates and ethers have all been prepared this way, and functional groups introduced in solution have been used to attach flame retardants, antimicrobial agents and biologically active molecules, or to prepare graft copolymers and block architectures ^{[7][33][34][35]}.

7.3. Porous, Shaped and Composite Materials

Regeneration from solution converts an isotropic dope into a solid with a controllable pore structure, which is the basis of a wide family of materials: hydrogels and superabsorbents, microspheres and beads for chromatography, enzyme immobilization and controlled release, aerogels with densities of a few tens of milligrams per cubic centimetre for thermal insulation, absorbents and templates, and membranes for ultrafiltration, haemodialysis and battery separators ^[15][27]. Selective dissolution of one component of a cellulose material, followed by regeneration, yields all-cellulose composites in which reinforcement and matrix are chemically identical, an approach that improves interfacial compatibility ^[38]. Solution processing also underpins the electrospinning and three-dimensional printing of cellulose structures.

7.4. Biomass Fractionation and Nanocellulose

Lignocellulosic biomass can be fractionated by dissolution-based routes: the solvent selectively dissolves cellulose or the hemicellulose fraction, and the components are recovered separately, which is the basis of several proposed cellulose-first biorefineries. Dissolution is also used to disperse cellulose nanofibrils and to prepare regenerated cellulose II nanoparticles, and deep eutectic solvents in particular have proved effective as pretreatment media for nanofibrillation ^[17].

7.5. Analytical and Fundamental Uses

Cellulose solvents are indispensable in polymer analysis. LiCl/DMAc remains the standard medium for determining molar mass and molar-mass distribution, and the combination of scattering, chromatographic and rheological data obtained in solution provides the basis for structure–property relationships that link molecular parameters to the mechanical behaviour of fibres and films ^[8][37].

8. Solvent Recovery, Green Chemistry and Practical Constraints

The industrial viability of a cellulose solvent is determined less by its dissolution power than by the cost and environmental burden of recovering it. In the Lyocell process more than 99 % of the NMMO is recovered and returned to the process, which is the principal reason for the process's favourable environmental profile relative to viscose, where carbon disulfide and hydrogen sulfide must be abated and only part of the sulfur is recovered ^[6][10]. Ionic liquids are effectively non-volatile and therefore emit no volatile organic compounds, but they must be separated from large volumes of water, either by evaporation, by membrane processes or by precipitation with an antisolvent, and this separation dominates the energy balance of the process; absorbed atmospheric moisture must also be removed, because water sharply reduces solvent quality ^[21][34]. Deep eutectic solvents are cheap enough to tolerate less exhaustive recovery, but their limited dissolution power restricts them to pretreatment, fractionation and reaction media ^[17]. Further practical considerations include the corrosivity of molten salt and acidic systems, the cost of lithium salts and the toxicity of dimethylacetamide, the high viscosity of concentrated dopes and the resulting pumping and filtration duties, and the need to avoid hornification of the feedstock by using never-dried pulp. Taken together, these requirements define the profile of an ideal cellulose solvent: a strong hydrogen-bond acceptor present at high effective concentration; a low melting point and low viscosity; thermal and chemical stability without side reactions; low toxicity and low cost; energy-efficient recovery; and the ability to dissolve never-dried, high-molar-mass cellulose without activation.

9. Current Status and Outlook

Cellulose solvents occupy a secure position in established industry through the viscose, cuprammonium and Lyocell processes, and the field is being pushed forward by two forces. The first is the demand for man-made cellulosic fibres, which continues to grow because such fibres are renewable and industrially compostable; the second is regulatory and societal pressure on the emissions and resource footprint of existing processes, particularly the sulfur-based viscose route ^[4]. Within research, several questions remain open: what exactly determines the dissolution power of a given liquid, why some salts dissolve cellulose while chemically similar ones do not, and whether the aggregates observed in nominally dissolved cellulose can be eliminated altogether. Computational screening methods are increasingly used to search for new solvent candidates, while process-oriented work concentrates on dissolving never-dried pulp directly, on integrating fractionation with dissolution, and on recovery schemes with a lower energy penalty. Because the obstacle is as much engineering and life-cycle performance as chemistry, progress over the coming years is likely to come from reconciling dissolution power with recyclability, cost and safety rather than from the discovery of an entirely new class of solvent.