Abstract
Factors affecting the swelling of cellulosic fibers are considered in this review. Emphasis is placed on aqueous systems and papermaking fibers, but the review also considers cellulose solvent systems, nanocellulose research, and the behavior of cellulosic hydrogels. The topic of swelling of cellulosic fibers ranges from effects of humid air, continuing through water immersion, and extends to hydrogels and the dissolution of cellulose, as well as some of its derivatives. The degree of swelling of cellulose fibers can be understood as involving a balance between forces of expansion (especially osmotic pressure) vs. various restraining forces, some of which involve the detailed structure of layers within the fibril structure of the fibers. The review also considers hornification and its effects related to swelling. The expansive forces are highly dependent on ionizable groups, pH, and the ionic strength of solution. The restraining forces depend on the nature of lignin, cellulose, and their detailed structural arrangements.
Download PDF
Full Article
Swelling of Cellulosic Fibers in Aqueous Systems: A Review of Chemical and Mechanistic Factors
Martin A. Hubbe,a Björn Sjöstrand,b Magnus Lestelius,b Helena Håkansson,b Agne Swerin,b and Gunnar Henriksson b,c
Factors affecting the swelling of cellulosic fibers are considered in this review. Emphasis is placed on aqueous systems and papermaking fibers, but the review also considers cellulose solvent systems, nanocellulose research, and the behavior of cellulosic hydrogels. The topic of swelling of cellulosic fibers ranges from effects of humid air, continuing through water immersion, and extends to hydrogels and the dissolution of cellulose, as well as some of its derivatives. The degree of swelling of cellulose fibers can be understood as involving a balance between forces of expansion (especially osmotic pressure) vs. various restraining forces, some of which involve the detailed structure of layers within the fibril structure of the fibers. The review also considers hornification and its effects related to swelling. The expansive forces are highly dependent on ionizable groups, pH, and the ionic strength of solution. The restraining forces depend on the nature of lignin, cellulose, and their detailed structural arrangements.
DOI: 10.15376/biores.19.3.Hubbe
Keywords: Osmotic pressure; Donnan equilibrium; Dissociation; Hydrophilicity; Crystallinity; Water retention value
Contact information: a: Department of Forest Biomaterials, North Carolina State University, Campus Box 8005, Raleigh, NC 27696-8005, USA; b: Pro2BE, The research environment for Processes and Products for a circular Biobased Economy, Department of Engineering and Chemical Sciences, Karlstad University, SE-65188, Karlstad, Sweden; c: KTH Royal Institute of Technology, School of Engineering Sciences in Chemistry, Biotechnology and Health, Department of Fiber and Polymer Technology, Division of Wood Chemistry and Pulp Technology., SE-10044 Stockholm, Sweden; *Corresponding author: hubbe@ncsu.edu
INTRODUCTION
The swelling of cellulosic fibers can have profound effects on such applications as absorbent paper products (Botkova et al. 2013; Hubbe et al. 2013; Ma et al. 2015), nanocellulose products (Torstensen et al. 2022), and products made from dissolving grades of cellulosic fibers (Okugawa et al. 2020, 2021). Less dramatic but still very important effects are observed in other paper products (Larsson and Wågberg 2008; Figueiredo et al. 2016). During the manufacture of paper, there is often a positive correlation between the swelling of cellulosic fibers (especially for kraft fibers) and the development of paper strength (Jayme and Büttel 1968). This phenomenon is mainly connected to the increased fiber flexibility that is associated with the swelled structure and consequently the increased number of fiber-fiber bonds in the fiber network. On the other hand, with increasing swelling of cellulosic fibers, the amount of water carried past the press section of a paper machine is increased (Hubbe and Heitmann 2007; Hubbe et al. 2020). The need to evaporate such water can slow down the rate of production and increase the net expenditure of energy.
Background: Structural Character of Cellulose and Matrix Polymers
The swelling of cellulosic fibers and their networks is normally quantified as liquid (water) mass uptake per mass of fiber, rather than quantified by a length or volume increase. A main reason is that several important steps in fiber swelling do not involve geometric or dimensional changes. To be able to discuss swelling behavior of cellulose and pulp fibers, one needs to take account of the structure of not only cellulose, but also cell wall matrix polysaccharides, i.e., hemicelluloses and pectins. Cellulose has a simple primary structure consisting of anhydrous β-glucopyranosides connected with 1,4 glycosidic bonds, and a degree of polymerization that can reach several thousands (Heinze 2015). The structure of this monomer can be described as being like a discus, in which there are hydroxyl groups peripherally oriented, whereas there are apolar groups facing the tops and bottoms (Fig. 1) (Seeberger 2015).
Fig. 1. Representation of the repeat unit in a cellulose chain: (a) β-glucopyranoside, (b) 1,4 glycosidic bonds, depicting also intra-chain hydrogen bonds
The glycosidic bond in the cellulose chain allows the possibility to form two stabilizing hydrogen bonds on each side, and these stabilize the structure into a two-fold helix, i.e., a flat and almost straight structure, with hydroxyls exposed to the sides and a hydrophobic character facing upwards and downwards (Kovalenko 2010). In natural cellulose, i.e. cellulose I, chains with the same direction, i.e., having the reducing endo towards the same direction, are lined up beside each other. They are connected with multiple hydrogen bonds, forming sheets in a secondary structure. These sheets are located on the top of each other and held together by van der Waals bonds in a tertiary structure. For steric reasons, one residue is not located exactly over another but displaced both in chain direction and side direction. Thus, cellulose is forming a crystalline structure (Glasser et al. 2012), which is important for the development of superior mechanical properties of the polysaccharide (Ioelovich 2016).
The exact shape of the crystal, the quaternary structure, is a matter of debate; cellulose is synthesized by an enzyme complex located in the cell membrane, and the chains are present in aggregates, meaning that several chains are synthesized in parallel, thereafter forming extended crystals, which together constitute microfibrils (Saxena et al. 2005). The size, i.e., the number of cellulose chains, and maybe also the shape of the microfibril, therefore depend on the size and shape of the aggregate of cellulose synthases. In some non-plant sources of cellulose, i.e., bacteria, tunicates, and certain algae, the microfibrils can be very large and sometimes have a rather flat structure (Moon et al. 2011). In higher plants, the microfibrils are believed to be rather small; recent results suggest 18 chains (Jarvis 2018), but earlier both 28 and 36 chains have been suggested (Cosgrove 2014).
What then is the cross-sectional shape of a fibril in higher plants? The answer is not known with certainty, but a common suggestion is the honeycomb shapes presented in Fig. 2. Although other shapes also have been suggested, it is inevitable that the microfibril will have both hydrophobic and hydrophilic surfaces exposed (Mazeau 2011). These two surfaces have distinctly different properties, and it seems as if the hydrophobic side is attractive for adsorption of hemicelluloses, as will be further discussed below (Heinonen et al. 2022; Kong et al. 2022). Proteins that specifically bind to cellulose, such as the cellulose binding module of fungal aerobic cellulases, seem to bind to the hydrophobic surface by forming sandwich structures between the hydrophobic top of glucopyranside residues and aromatic amino acids, i.e. tyrosine and tryptophan (Arola and Linder 2016). Issues related to the differing affinities of different crystal planes of cellulose will be considered in more depth later in the article.
Fig. 2. Likely cross-sectional shapes that have been proposed for the microfibrils of cellulose within higher plants. a) Honeycomb (diamond) model. This model is favored by many scholars; note that both hydrophilic and hydrophobic surfaces are exposed. b) An alternative “double fibril” model. Here the hydrophobic surfaces are smaller, but instead “pockets” are created.
The contrasting hypothetical cross-sectional shapes of the initial cellulose microfibrils that form during biosynthesis can be expected to have implications related to how the units self-assemble into larger structures at later stages of biosynthesis. A recent review of the fundamentals of self-assembly of lignocellulosic materials showed how idealized shapes, not unlike those shown in Fig. 2, might be expected to govern patterns of self-assembly (Hubbe et al. 2023). For instance, one might speculate that the “pocket” depicted in Part B of the figure might provide some kind of evolutionary advantage in the ability of plants to form strong cell walls. However, many fundamental questions remain unresolved concerning the mechanisms of cell wall formation in plants.
The cellulose microfibril should not be understood as being a stiff, static, and perfect crystal; rather, cellulose consists of a mixture of highly crystalline and more unordered cellulose domains. The unordered cellulose is partly located on the surface of the fibril. In the presence of water or similar liquids, there is probably some movability of the superficial cellulose chains on the fibril – but there are indications that there also are segments of the fibrils in which the unordered structure goes deeper into the fibril than the single layer indicated in Fig. 3 (Brestkin and Frenkel 1971; Nishiyama et al. 2003; Fernandez et al. 2011). The unordered cellulose is generally more reactive towards hydrolysis with acid or cellulolytic enzymes (cellulases) and probably also for chemical derivatization (Wang et al. 2014a; Ioelovich 2021). Most likely, an unordered segment in a fibril might allow the fibril to bend more easily, and the movability of the superficial cellulose chains might play an important role for interaction with water or another component. A honeycomb organization of the microfibril with moving superficial cellulose chains will give the microfibril a cylindrical appearance – something that seems to be supported by electron microscopic pictures of cell walls.
Fig. 3. The central part of the microfibrils might in general be rather crystalline, whereas surficial cellulose might be more flexible and unordered. There are most likely also short segments where the unordered cellulose goes deeper (Brestkin and Frenkel 1971; Fernandes et al. 2011).
With some exceptions, such as the seed hairs of cotton, cellulose does not appear as a single polysaccharide component in higher plants; the plant cell wall generally contains other polymeric carbohydrates, which are often called matrix polysaccharides. These are divided into two groups, the hemicelluloses and the pectins (Scheller and Ulvskov 2010). Hemicelluloses have a main chain in which the monosaccharide residues are connected with β-1,4 glycosidic bonds. The hydroxyl group on the C4 is peripherally located (as in glucose, Fig. 1). This allows the polysaccharide to adapt to the same two-fold structure as in cellulose, although other conformations also are possible. Examples are glucomannans, arabinoxylans, and xyloglucans. Pectins are cell wall polysaccharides that cannot adapt to such a two-fold helix. Examples include β-1,4 galactans, -1,5 galacturonans, and rhamnogalacturonans (Alkorta et al. 1998; Ebringerová et al. 2005; Scheller and Ulvskov 2010).
Recent data suggest that both hemicelluloses and pectin can interact, such that they bind to cellulose surfaces, but that they do so in different ways; hemicelluloses can align themselves both parallel and antiparallel with cellulose chains in the microfibrils. In a sense, they may appear to “crystalize” onto the cellulose domains. By contrast, pectins bind in more nonspecific way. Not surprisingly, the hemicelluloses appear to be more firmly attached to the cellulose microfibrils. Generally, matrix polysaccharides appear to prefer to bind to the hydrophobic surfaces of fibrils (Fig. 2) (Heinonen et al. 2022; Jarvis 2023).
The composition of matrix polysaccharides differs greatly between different phylogenic groups, but also between different cell wall layers. These differences start with the secondary wall, which is dominant in terms of both mass and volume (Harris 2006). The most studied of these are the woody eudicolyledons (hardwoods, i.e. “eucots”), monocotyledons, and conifers (softwoods, i.e. the dominating phylum of gymnosperms) (Daly et al. 2001). In woody eucots, xylans are the most common secondary cell wall polysaccharides, but glucomannans also are present in significant amounts (Salmén 2022). In monocots, xylans dominate almost totally, whereas in conifers, galactoglucomannans are the most common, followed by rather large amounts of arabinoxylans (Moreira and Filho 2008; Peng and She 2014). The primary cell walls have a more complex composition, and in addition to xylans, also xyloglucans and pectins (rhamnogalacturonan and polygalacturonic acid) are common (Hayashi 1989; Brett et al. 2005; Caffall and Mohnen 2009). The latter polysaccharides can also be present in the middle lamella. Berglund et al. (2020) presented evidence that different hemicelluloses can influence the elastic modulus and stretch to breakage of cellulose-based films in contrasting ways.
In summary, the cellulose itself probably mainly exposes surfaces rich in hydroxyls, thereby opening up for the ability to interact with water and other molecules, often leading to hydrogen bonding. In addition to this, there are possibilities for hydrophobic interactions and van der Waals interactions on the hydrophobic sides of a fibril (Fig. 2). A more unordered cellulose, when it is in a swollen state, will is likely to have proportionally more hydrophobic surfaces exposed. Charges, on the other hand, do not occur on pure natural cellulose (which is a homoglucan), so salt bridges do not occur directly between natural cellulose molecules (Kontturi et al. 2006). Anionic structures, i.e., carboxylic acids, hypothetically may be induced on the cellulose polymer, by technical processes, such as bleaching. The matrix polysaccharides (the hemicelluloses) will most likely be forming layers around microfibrils, connecting them, or embedded in them, adding more of unordered structures with richer possibilities for hydrophobic interactions and van der Waals forces. Furthermore, certain matrix polysaccharides, i.e., xylans, polygalacturonic acid, and rhamnogalacturonan, contain anionic carboxylic acids, thereby making plant fibers negatively charged. Xylans have been suggested to act as a kind of dispersant for cellulose microfibrils by adding charge to them by associating with the surfaces (Berglund et al. 2020). A similar role has been observed for glycerol (Moser et al. 2018b). Especially the cationic groups in pectins often form salt bridges to calcium ions, thereby crosslinking chains by means of “eggbox structures” (Fig. 4) (Cao et al. 2020). These structures function as crosslinkers to polysaccharide chains and have importance for the rigidity of pectin-rich tissues, such as middle lamella of the phloem in certain herbs, e.g. flax and hemp. Removal of calcium from pectin leads to a swelling and weakening of pectin rich tissues, and it increases their accessibility for enzymatic degradation; this approach has been considered for chemical retting of bast fibers (Henriksson et al. 1997). Addition of calcium salts may on the other hand decrease swelling (Thakur et al. 1967). In line with this, Eucalyptus wood with high content of calcium is more recalcitrant to kraft pulping (Vegunta et al. 2022).
Fig. 4. Complexes formed between carboxylic anions and calcium(II) ions are known as “egg boxes”. They are believed to play a central role in the rigidity of pectin-rich tissues (Heinze 2016). a) Chemical structure of an eggbox, i.e., complex between two uronic acids and a calcium ion, thereby crosslinking two polysaccharide chains; b) Schematic presentation of the role of egg boxes in pectin structures; the calcium ions crosslink the “smooth” regions rich in galacturonic acids (Cao et al. 2020)
Steps in Swelling
The process of swelling of cellulosic fibers can be envisioned as a continuum of overlapping steps. To begin, equilibration of dry cellulosic fibers, paper products, or other such materials with increasing levels of relative humidity results in increasing moisture uptake and slight changes in macroscopic dimensions (Larsson and Wågberg 2008; Gamstedt et al. 2015; Joffre et al. 2016). Greater proportional effects of relative humidity on dimensions have been found when examining the thickness dimension of thin films of nanocellulose (Rehfeldt and Tanaka 2003; Shrestha et al. 2017; Torstensen et al. 2018). Wetting by pure water results in much larger mass uptake and swelling of dimensions (Scallan 1983; Solhi et al. 2023). A key focus of this review article will be on ways in which the pH and ion contents of such solutions affect the degree of swelling and how such effects can be understood based on chemistry, physics, and colloid science. Further aspects of the swelling mechanisms are revealed by considering evidence from more extreme situations, including the behavior of cellulose-based hydrogels (Ganji et al. 2010; Chang and Zhang 2011; Ma et al. 2015) and solvent systems for cellulose (Budtova and Navard 2016). For instance, some of the greatest observed degrees of swelling of cellulosic material are associated with the last steps before dissolution in the presence of specialized solvent systems (Cuissinat and Navard 2006; Zhang et al. 2013; Budtova and Navard 2016).
The Riddle of Cellulose’s Insolubility in Water
In addition to the practical importance of swelling, as already described, the results of investigations into the swelling of cellulose fibers also can provide evidence into fundamentals of how the cellulosic material is being held together. As will be described further in this article, hydrogen bonding plays a central role not only in the bonding within cellulosic fibers, but also with respect to their interactions with water (Medronho et al. 2012; Norgren et al. 2023; Sjöstrand et al. 2023). If that were the only type of bonding, however, then it would be hard to explain the insolubility of cellulose in most aqueous conditions (Bergenstråhle et al. 2010; Budtova and Navard 2016). As been discussed above, cellulose chains and microfibrils display hydrophobic surfaces at specific planes, and these might interact with each other by van der Waals forces. When water is present, the cellulose is kept together, in part, by hydrophobic interactions, which involves an entropic effect of the solvent (Ghosh et al. 2002; Glasser et al. 2012). Thus, this article will consider effects such as the orientation of cellulose macromolecular chains, which have differing affinity for water on different sides. This amphiphilic character is demonstrated by the differing wettability characteristics of the different faces of cellulose crystals (Yamane et al. 2006). In addition, the contrasting microfibril angles in different sublayers of a cellulosic fiber, as will be shown later, sometimes tend to restrain the swelling of other layers within the same fibers.
Sections of this Article and Recognition of Earlier Review Articles
This review article is organized into the major sections of quantification of swelling, factors affecting the swelling of cellulosic matter, theoretical aspects of swelling of cellulosic matter, applications of swelling of cellulosic material, and some closing comments. From a developmental and materials science perspective, the articles to be cited point to two main strategies that have been used to modify the swelling degree of cellulose-based materials intended for different applications. Chemical approaches, including the forming or attachment of different ionizable functional groups on cellulosic surface, can be used to increase the osmotic pressure that tends to increase swelling, subject to various aqueous conditions. Chemical delignification, along with physical treatments such as mechanical fibrillation (i.e. refining), can be used to alter or decrease physical restraints to the swelling of the same materials. However, as will be discussed, the restraining tendencies of cellulosic structures are also highly dependent on any history of drying, which can decrease the swelling ability, i.e. effects of hornification (Fernandes Diniz et al. 2004; Salmén and Stevanic 2018; Sjöstrand et al. 2023; Sellman et al. 2023).
Aspects of the swelling of cellulosic fibers have been considered in some earlier review articles, and these are listed in Table 1. Based on the items in this list, there appears to have been an acceleration of interest within the last two years.
Table 1. Earlier Review Articles Dealing with Aspects of Swelling of Cellulosic Fibers and Related Materials
QUANTIFICATION OF SWELLING
The swelling of cellulosic fibers and networks is often defined as the amount of liquid (often water) mass uptake per mass of fiber. Another way to define swelling is based on dimensional changes. However, such dimensional changes may be insignificant or hard to quantify in many situations of interest. Thus, in the literature, various different definitions and evaluations methods have been employed.
Commonly used methods to quantify the swelling of cellulosic fibers and related materials can be placed into the categories of water retention value (and related tests), methods involving measurements of the concentrations of soluble polymers in solution, methods based on observed changes in dimensions of objects, methods based on permeability, and a variety of lesser-used approaches (Gallay 1950). These are described in the subsections below. As a general summary, each of the existing methods to be considered for evaluation of swelling involves areas of uncertainty and approximation. Methods that rely upon the dimensions of swollen cellulosic material typically suffer from uncertainty in defining the boundary between the material and the surrounding aqueous (or other) medium. Methods that depend on the sizes of selected probe molecules tend to be blind to the presence of some pores that may be larger than the probe. And some, such as the WRV tests to be described next, can be regarded as practical, but not firmly grounded in theory.
Water retention value
The water retention value (WRV) test (Welo et al. 1952; Marsh et al. 1953; Jayme and Büttel 1964, 1968; Cheng et al. 2010), which involves centrifugation of a damp plug of fibers, is specified in standard methods (SCAN-C 62:00, 2000; TAPPI UM 256 1981; ISO 23714:2014). The basic assumption is that when a specified centrifugal acceleration is applied to the damp plug of fibers, the acceleration will remove water that was either between the fibers or within the fiber lumens. The idea is that the effective forces acting on the fibers will squeeze each of the fibers, such that any water contained within the lumen space will come out via the pit openings. An analogy can be made to the squeezing of a tube of toothpaste. By contrast, it is assumed that any water contained within the mesopores or micropores within the cell walls of fibers will be prevented from escaping by sufficiently large capillary forces (Welo et al. 1952). Different fiber starting materials will of course showcase varying fiber properties, and the remaining water will be located differently based on fiber morphology.
As a general rule, WRV tests have a reputation of providing useful comparisons between chemically delignified pulp samples that have been refined to different levels. However, high levels of external fibrillation can contribute to the WRV values in a way that does not fit with the usual definition of swelling. For example, a well-fibrillated thermomechanical pulp (TMP) fiber may have a WRV value that suggests about twice the level of swelling in comparison to the results of soluble exclusions tests (see later discussion). In addition, TMP tends to maintain its WRV over cycles of papermaking and repulping (Alanko et al. 1995). Because TMP has a composition similar to wood, its cell walls lack the porosity and swelling ability of kraft fibers, especially after the latter have been refined.
A related test method of water removal is drainage, which can be measured in Degrees Schopper Riegler (°SR) or Canadian Standard Freeness (CSF). Drainage is more related to liquid water passing between fibers but has been shown to correlate linearly with WRV. However, it is important to note that slopes of linear correlations are different for different pulp compositions (Sjöstrand et al. 2019).
To carry out a WRV test, a well-defined amount of pulp fibers is formed into a plug, which is then centrifuged under specified conditions of centrifugal force, time, and temperature. One weighs the material immediately after centrifugation. The procedure is done in such a way as to allow water released from the fibers to pass through a screen and not return. A second weight is determined after the same plug of fibers has been completely dried, preferably according to ISO 638 (2008) or a similar method.
Different results can be obtained from WRV tests, depending on the level of centrifugal acceleration (Cheng et al. 2010). The WRV values tend to decrease with increasing centrifugal acceleration and time of spinning. As described in the cited article, smaller particles of cellulose gave rise to larger WRV values, which is consistent with the presence of generally smaller capillary opening within the material. Based on this principle, WRV has been employed as a way to characterize the degree of fibrillation of highly fibrillated cellulose materials, i.e. micro- or nanocellulose (Gu et al. 2018). Effects of carboxymethylation (Laine et al. 2003) and TEMPO-oxidation (Maloney 2015) of cellulosic fibers also have been evaluated with WRV.
Although the WRV test has been used for many years as a practical way to estimate the amount of water held within the cells walls of ordinary cellulosic fibers, especially when they are being used to manufacture typical grades of paper, there is evidence that some other categories of water can be involved, especially when polymeric chemical additives are being employed. For example, if one were to assume that the WRV test senses only water present in mesopores, then one would expect that relatively small cationic polymers, having easy access into such pores, would have a big effect on the WRV results. In principle, the small cationic polymers would be expected to diffuse into tiny pores in the cell walls and neutralize surface charges, leading to pore contraction due to the decreased osmotic pressure. However, the opposite was found by Swerin et al. (1990), and Ström and Kunnas (1991). They observed that a high-molecular-mass cationic polymer product had a much bigger effect on WRV than a low-mass version. It has been argued that such results are consistent with a charged patch type of interactions among external fibrils and polyelectrolytes at the fiber surfaces (Hubbe and Panczyk 2007b; Hubbe et al. 2008). Alternatively, the same effects might be attributed to complexation among oppositely charged polyelectrolytes, where the cellulosic fibrils are regarded as if they were anionic polyelectrolyte strands (Hubbe 2019). In either case, the WRV results appear to be sensitive to details concerning such charged polymer additives.
Some modified versions of the WRV test have been demonstrated, seeking to broaden the range of information that can be obtained. For example, Hubbe and Panczyk (2007a) developed a modified WRV method (MWRV) in which a mass was placed on top of the damp plug of fibers before the start of centrifugation. Such a modification makes the test results somewhat more relevant to such situations as water removal in a wet-pressing operation. Notably, however, variations in pH and salt concentration were not found to have major effects on the results of the MWRV tests. Rather, increased refining increased the MWRV results, whereas drying history of the fibers decreased the MWRV values.
Some researchers have obtained WRV values when evaluating the properties of nanocellulose suspensions, but the justification and meaning of such tests can be questioned. To carry out such tests, a membrane is used to support the cellulosic material during centrifugation (Cheng et al. 2010; Gu et al. 2018). This approach makes it possible to evaluate nanocellulose materials having small enough particle size to pass through the usual supports used in WRV tests. Since nanocellulose suspensions do not have any lumen spaces or inter-fiber spaces, one cannot claim that the applied centrifugation is meant to remove water from such spaces. Rather, it seems likely that the WRV results from testing of nanocellulose suspensions will be related to surface area and fineness of the material, which are expected to contribute to both capillary forces and resistance to flow.
Solute Concentration Tests
Another basic way to gain information about the amounts of water that are contained within the mesopore structure of cellulose fiber cell walls involves the evaluation of concentrations of probe macromolecules in solution (Stone and Scallan 1966, 1967, 1968; Scallan and Carles 1972; Alince 1991). The idea is based on an assumption that a water-soluble polymer of a certain molecular mass will be prevented from entering a pore if it is smaller than a critical size, perhaps related to the radius of gyration of that macromolecule. Three ways of carrying out such tests are possible, as described below.
Fiber saturation point
To determine the fiber saturation points of cellulosic materials in aqueous suspension, one selects a water-soluble polymer having a lack of attraction to cellulosic surfaces and a suitably high and narrow distribution of molecular mass such that it can be assumed that little of the macromolecule will be present in the mesopores of a typical chemical pulp (kraft, sulfite, etc.) specimen (Stone and Scallan 1966, 1967; Alince 1991). Dextran molecules having molecular mass of about one million Daltons or greater have been used for such tests. Rather than providing a physical barrier to entrance of the probe polymers, one relies upon the fact that the probability that such a polymer will diffuse into a given space depends strongly on whether or not the freedom of motion of polymer segments will become constrained, thus affecting the entropy content and the free energy (Alince 2002). The concentration of the probe polymer in the bulk of solution will tend to be increased, compared to what can be calculated based on the known added quantity. The calculation involves the total amount of solution present, including solution that happens to be within mesopores and micropores of the cell walls of fibers. One then calculates the volume that is associated with the pores too small to permit access. As noted by Scallan and Carles (1972), the obtained fiber saturation value results are often in reasonable agreement with WRV test results, at least when testing ordinary papermaking fibers within typical ranges of mechanical refining.
Inverse size-exclusion chromatography
Size exclusion chromatograph (SEC) has emerged as an effective and practical way to evaluate the molecular mass distributions of polymers in solution (Burgess 2018). The method is based on the principle that the time required for a given macromolecule to elute through a packed column will depend on whether or not it is small enough to spend time within the pores of a selected packing material. Very small probe molecules will require the longest elution times, since they will be spending more time, on average, within parts of the liquid that are not available for larger macromolecules. Inverse size exclusion chromatography (ISEC) measurements are similar, except that one employs a known probe macromolecule, having a narrow molecular mass distribution, and the quantity to be determined is the pore size or pore size distribution of a solid material, e.g., a swollen cellulosic substance (Yao and Lenhoff 2004). The ISEC method was employed, using a series of different molecular mass fractions, by Berthold and Salmén (1997a) to help understand how the process of kraft pulping and bleaching affect the pore size distribution. In general, it was found that there was a substantial increase in the volume of mesopores (i.e., pores from 2 to 50 nm in diameter) following chemical pulping, with even greater porosity after both pulping and bleaching. Follow-up work showed how such distributions were shifted due to partial closure of intermediate-sized pores (mainly mesopores) in the course of drying and rewetting (Berthold and Salmén 1997b).
Polyelectrolyte adsorption
Because cellulosic materials typically have a negative charge – due to the presence of axabinoxylan or pectin or by chemical modification of cellulose (natural cellulose is, as discussed above, uncharged) – cationic polyelectrolytes will be expected to interact strongly with them. It has been proposed to use cationic polyelectrolytes of known molecular mass to probe the pore structures of cellulosic fibers (Alince and van de Ven 1997; Alince 2002). By employing a highly branched polymer, the cited work reached the conclusion that mesopores within the swollen cell walls of bleached kraft fibers were relatively uniform and about 80 nm or larger in diameter. However, much smaller sizes have been determined by other methods, such as NMR (Li et al. 1993). Studies have shown that factors affecting permeation of cationic polyelectrolytes from aqueous into porous materials are complex (Wu et al. 2009). As described in the cited review article, cationic polymers at first may be drawn into small pores in a snake-like fashion; however, the progress of cationic polymers into such pores is likely to be strongly affected by kinetic factors, including essentially trapped non-equilibrium conditions. Further information about cellulose porosity was obtained indirectly by sensing electrokinetic effects that originate from those areas of bleached kraft fibers (Hubbe et al. 2007a).
Table 2. Methods to Evaluate Swelling Based on Dimensional Changes
Methods Based on Dimensional Measurements
Reported swelling methods involving dimensional measurements of cellulose-based materials have tended to be diverse, rather than following a standard. It appears that different kinds of specimens, or possibly different orientations of those specimens, will require different procedures. Table 2 summarizes such work, with attention to the nature of the specimens and the methods. Note that some of these methods were designed in such a way as to capture rapid changes in dimensions upon introduction of changes in humidity or immersion (Qing et al. 2013; Jablonsky et al. 2014; Shrestha et al. 2017). In addition, it has been possible to capture dramatic and to a large extent reversible changes in dimension by wetting and redrying by suitable methods (Kontturi et al. 2011).
Permeability
Though the degree of correlation between permeability and porosity (or fractional volume content of pores in a material) needs to be established in each case, permeability is sometimes used as a proxy for porosity measurements. In paper specimens, there is a general expectation that air permeability will decrease with increasing apparent density. However, the effect is not linear. This is shown in Fig. 5, which for the first time plots data reported by Vänskä et al. (2016). The cited authors varied the lignin content as well as the extent of mechanical refining. As shown in the figure, the logarithm of the time required for permeation of 100 cm3 of air was found to be roughly proportional to apparent density. However, all of the results shown in the figure for the highest Gurley seconds values were beyond the upper limit of measurement. The densest sheets, representing high levels of refining, tended to act as near-perfect seals, suggesting molecular contact between adjacent fibers in the structure. In other cases, the correlation between permeability and density may be poor due to detailed differences in pore structure (Hantel et al. 2017). Nonlinear and complex relationships between paper permeability and density are also suggested by studies of the relative ease of removal of water during the paper forming process (Hubbe et al. 2020). The cited review article showed, for instance, that densified surface layers may have a dominant effect on permeation behavior.
Fig. 5. Plot prepared here for the first time from the data of Hantel et al. (2017), showing the logarithm of air permeation times for paper prepared with different delignification degree and mechanical refining levels, leading to the apparent density values shown for the resulting paper
Surface Area and Porosity Using Specialized Drying Approaches
One of the most trusted approaches to determining the surface area and pore sizes of materials is by adsorption of very cold nitrogen or argon gases at low pressures (Bardestani et al. 2019). When the goal is to study the pore structures of cellulosic materials, one of the key challenges is the fact that such tests require the usage of completely dry samples. It is well known, however, that substantial and partly irreversible closure of pores (i.e. hornification) occurs when cellulosic materials are dried using conventional evaporation, especially in cases where the lignin has been removed (Fernandez Diniz et al. 2004; Hubbe et al. 2007b). Geffertova et al. (2013) found that repeated cycles of paper forming, drying, and reslurrying of the pulp in preparation for the next cycle without additional refining tended to increase the air permeability of sheets. This is consistent with the increasing stiffness and lack of compliance of hornified fibers. To minimize such effects during determination of pore sizes in wet specimens, it is recommended to replace the water in the specimens with another solvent and then employ critical point drying (Kang et al. 2018). In other words, evaporation takes place at the location within a phase diagram where solid, liquid, and gas phases meet each other. Such approaches can avoid development of high surface tension forces and extensive hydrogen bonding in the course of drying. Once the specimens have been dried, then one can use the Brunauer-Emmett-Teller (BET, 1938) method to obtain the surface area, based on gas adsorption isotherms (Sing 2001; Bardestani et al. 2019). Related methods can provide estimates of pore-size distributions (Barrett et al. 1951; Bardestani et al. 2019), though such determinations require simplifying assumptions (Groen et al. 2003).
These BET-related methods require that the sample is dried. Even when using specialized drying methods, there still will be some risk of pore collapse, which will influence the result. Therefore, there is interest in methods for measuring the cellulose surface area without drying. One such method is based on adsorption of the colorants methylene blue and Congo red (Inglesby and Zeronian 1996). This method requires a high content of sulphate ions to make the adsorption possible. This is problematic, since it can lead to aggregation of fibers and collapse of pores. Moser et al. (2018a) developed a method based on adsorption of the polysaccharide xyloglucan on the cellulose, which does not require increased ionic strength. The content of non-bound xyloglucan can easily be detected, since it forms a strongly colored complex with iodine.
FACTORS AFFECTING THE SWELLING OF CELLULOSIC MATTER
The emphasis of this section is to consider evidence about what parameters affect swelling of cellulosic fibers. Though some theoretical aspects may be unavoidable, in light of such emphasis, the goal will be to consider articles that employed an empirical approach, rather than to focus on theories, which will be covered in the subsequent main section.
Aspects related to experimental design and selection of conditions need to be considered carefully when attempting to understand published work related to the swelling of cellulosic fibers. With many exceptions, typical experiments related to the swelling of cellulosic materials have involved one-at-a-time variation of certain parameters e.g., pH. The conditions that are held constant, in such experiments, can be numerous and quite specific to the interests of the researchers. Those conditions may not necessarily match those employed by other researchers. As will be illustrated by some of the cases considered in this article, some findings will appear to have general validity, whereas other findings may be limited to special circumstances.
In general, as will be described in this section, there has been great progress in understanding different aspects that contribute to swelling behavior (Benselfelt et al. 2023; Sellman et al. 2023; Sjöstrand et al. 2023; Solhi et al. 2023). Factors that influence the extent of swelling include not only the chemical composition and aqueous conditions, but also the details of the layered structures that make up a cellulosic fiber.
Fiber Structural Details Resulting from its Biosynthesis
Many key attributes affecting the swelling behavior of cellulosic fibers have their origin in the nanostructures and microstructures that are established during biosynthesis. For a full understanding of those details, readers are referred to other sources (Brown and Saxena 2000; McNamara et al. 2015; Manan et al. 2022).
Some important aspects, relative to swelling behavior, can be emphasized. First, it has been shown that cellulose chains during their biosynthesis are essentially extruded towards the outside of cell membranes in parallel, essentially continuous manner (Brown and Saxena 2000; Tobias et al. 2020). As mentioned earlier, they almost immediately join together in nascent crystalline form in groups of about six, thus becoming microfibrils. Although the term “microfibrils” does not contain the term “nano,” it has become well established in the literature (Doblin et al. 2002) as a descriptor for these features that have nano-sized thicknesses. The process is depicted schematically in Fig. 6, part A. The nanofibrils may then join in parallel with other such groups, thus becoming larger fibrillar structures. In close succession, hemicellulose is also formed, and then lignin, thus surrounding and connecting the cellulose nanofibrils or microfibrils. In that manner, nature forms a nanocomposite structure in which the relatively stiff cellulose fibrillar elements are embedded in a double matrix of hemicellulose and lignin. Figure 6, part B illustrates the diffusion of monolignol compounds into the nascent cellulosic substance during the biosynthesis process.
Fig. 6. Schematic diagram of key biosynthetic events in the formation of lignocellulosic fibers. A: cellulose biosynthesis; B: lignin biosynthesis occurring within spaces of the just-formed cellulose nanofibrils or microfibrils.
Another key aspect that will affect the swelling behavior of cellulosic fibers, related to their biosynthesis, is the layered nature of the cell walls, and in particular the microfibril angles within the cell walls. Figure 7 depicts a typical cellulosic fiber, as it might be present in a wood specimen. As shown, each fiber has a thin primary (P) layer, and within that are three secondary (S) sublayers, of which the S2 sublayer is by far the thickest, such that it tends to dominate the structural aspects of fibers. The microfibril angle within the S2 sublayer often is within the range of 2 to 25 degrees, i.e., close to being aligned with the fiber and with the axis of the plant (Barnett and Bonham 2004; Donaldson 2008). Substantially higher values of the S2 sublayer fibril angle are found in such wood species as juniper (Hanninen et al. 2012); such wood species tend to be much more flexible and resistant to breakage, but less able to grow high.
Fig. 7. Sketch of lignocellulosic fiber, including layered structure, typical microfibril angles, and lignin distribution
The thinner S1 sublayer wraps itself around the S2 sublayer, such that it tends to restrict outward swelling of the S2 sublayer (Barnett and Bonham 2004). On the inside of the structure, adjacent to the lumen space, the S3 sublayer is likely oriented such that it tends to prevent inward collapse of the S2 sublayer. Between adjacent fibers, the “middle lamella” region is rich in lignin, thus providing a stiff, decay-resistant matrix to the composite structure. Studies suggest that at least part of the hemicellulose within a typical cellulosic fiber acts as a covalently bonded connector between lignin and cellulose (Lawoko et al. 2003).
Disassembly of Woody Material
In view of the theme of restraining factors and forces that tend to oppose the swelling of cellulosic fibers, it is important to keep in mind some of the widely employed technologies that are being used to prepare such fibers for papermaking, nanocellulose, and other applications. Many such processes can be regarded as disassembly processes, whereby either some components are removed or the fibers are mechanically separated from each other. Such treatments may weaken the ability of structures within the fiber to restrain swelling tendencies.
Delignification
Terms including “pulping” and “chemical pulping” are used to denote the widely employed methods of removing lignin (often with partial removal of hemicellulose) as part of the preparation of cellulosic fibers for papermaking and other applications (Mboowa 2021). It has been shown that removal of lignin from a cellulosic fiber, still in its undried state, leaves behind a structure having substantial mesoporous character (Stone and Scallan 1967; Berthold and Salmén 1997a,b). In other words, there are pores having diameters in the range of about 2 to 50 nm. Berthold and Salmén (1997a), who employed inverse size exclusion chromatography, found that delignification resulted in development of about 1.1 to 1.2 mL/g of new pore volume for water-swollen never-dried bleached and unbleached kraft pulps. Confirmatory evidence of increased pore volume after chemical pulping has been obtained by WRV testing (Jayme and Büttel 1968; Laivins and Scallan 1993). For example, Carlsson et al. (1983) showed that the WRV, a measure of swelling, increased from an initial value near to 115% (mass water per mass solids) up to the range 160 to about 250 %, depending on the pH. Scallan and Tigerström (1992) showed that such results could be interpreted as being due to a decreased elastic modulus of the cellulosic fiber material with the removal of the lignin, thus decreasing the ability of the fiber structures to resist swelling.
Presumably this development of mesopore spaces in wood fiber cell wall as a result of chemical pulping could be attributed to the presence of lignin-rich nano-domains having such dimensions in the original wood. Such domains are illustrated in Fig. 8, which is a redrawn version inspired by Fengel and Wegener (1989). Note that lignin is envisioned in the sketch as surrounding a microfibril (i.e., an assemblage of nanofibrils), the center portion being mostly cellulose, and with hemicellulose acting as an intermediary between the cellulose and the lignin. As mentioned in the Introduction, it should be understood that pectins also are expected to play a role (Alkorta et al. 1998; Ebringerová et al. 2005; Scheller and Ulvskov 2010). The idea of lignin occupying nano-sized domains is supported by the relatively large size of the lignin macromolecules that have been isolated by alkaline pulping methods (Gupta and Goring 1960).
Fig. 8. Depicture of woody material at the nano-scale, whereby cellulose and lignin tend to occupy different nano-domains, and the hemicellulose tends to be located between the cellulose and lignin
Lignin-carbohydrate complexes in intact wood
Another important feature that is shown in Fig. 8 is a type of linkage that can be expected to play a contributing role in inhibiting swelling of intact wood specimens. Note the label “LP-Linkage” that appears at the lower right of the figure. This label refers to covalent attachments between hemicellulose and lignin, i.e., lignin-carbohydrate complexes (LCCs) (Lawoko et al. 2003). The presence of such bonds helps to explain how it is possible for the hemicelluloses, which are the most hydrophilic main component of wood, to remain securely adherent, at a molecular level, to lignin, which is much more hydrophobic. Thus, wood retains its inherent strength and integrity, even when it is soaked during rainy seasons or immersion in water. In intact wood, lignin forms connections with different matrix polysaccharide molecules in a way that crosslinks the structure. This prevents wood fibers from extensive swelling in water (Henriksson 2017). Such limitations to swelling can be regarded as an attribute of “woody” matter, i.e. secondary xylem. During chemical pulping, the lignin-polysaccharide networks are in large part dissolved (Lawoko et al. 2005).
The moisture content of intact wood specimens has been shown to rise in approximately linear fashion with increasing relative humidity (RH) from about zero to about 0.3 mass fraction at about 95% RH (Fredriksson 2019). Further exposure of the material to liquid water gave rise to a dramatic rise to about 2.5 mass units of water per mass unit of solid. It is reasonable to expect that such values are constrained by such factors as the pore volume within cell lumens, vessels, some swelling of hemicellulose, and the presence of LCCs that prevent what otherwise might be a form of delamination at the nanoscale of wet wood. As noted by Fredriksson (2019), under dry conditions, much of the water is associated with wood material via hydrogen bonding, whereas under saturated conditions capillary forces draw water into the structure.
Mechanical refining of pulp
To prepare cellulosic fibers for papermaking, their aqueous suspensions are typically passed through mechanical devices called refiners (Gharehkhani et al. 2015). Often this operation is carried out at consistency (filterable solids) levels of 4 to 10%. In a modern pulp refiner, flocs of fibers experience repeated compression and shearing events as the rectangular “bars” of a rotor repeatedly cross a corresponding set of bars on a stator.
There are two main classes of refining practices – refining of mechanical pulps and refining (sometimes called “beating”) of chemical pulps such as kraft pulps. A key difference lies in the fact that mechanical pulps are generally fed into the refining system in the form of wood chips, whereas the chemical pulps already have been liberated as individual fibers due to the breakdown and dissolution of the lignin. The equipment, processes, and various options for pulp refining have been discussed elsewhere (Li et al. 2011; Gharehkhani et al. 2015; Kerekes 2015). As noted by Lahtinen et al. (2014), both main classes of pulp refining generally result in a fibrillated appearance of the material. In other words, some unraveling takes place of the outer fibrillar layers due to the repeated compression and shearing event. Kraft pulps were found to fibrillate relatively easily and quickly, yielding large values of WRV (e.g. 380%). This effect is consistent with the breakdown and removal of cellulosic strands encircling the dominant S2 sublayer of pulp fibers. Increases in WRV due to refining can be largely attributed to delamination within the cell walls, i.e., internal fibrillation (Przybysz et al. 2017). By contrast, mechanical pulps did not reach as high levels of WRV and there was substantial fragmentation of fibers, in addition to visible fibrillation (Lahtinen et al. 2014). The fines fraction of refined pulp has been shown to hold onto a disproportionate amount of the water detected by WRV and related tests (Laivins and Scallan 1996; Olejnik et al. 2017).
Figure 9 presents a schematic view of how the cross-section of a kraft pulp fiber is expected to change over the course of extensive mechanical refining. The manner in which fibers are packed together in the woody tissue of a tree gives rise to the rectangular shape of unrefined wood-derived fibers, as depicted on the left side of the figure.
Fig. 9. Schematic view of (left) the cross-section of an unrefined kraft pulp fiber, showing the layered structure of the cell wall and (right) an intentionally exaggerated view of the same fiber’s cross-section after mechanical refining, featuring swelling of the partly delaminated S2 sublayer, mechanical breakdown and either removal of fibrillation of the outer layers, and partial or complete collapse and closure of the lumen space. Figure previously used by Debnath et al. (2022)
Repeated shearing and compression events as the fiber passes through one or more mechanical refiner stages is expected to impart swelling. In this illustration, the bright green S2 sublayer is depicted as becoming thicker, presumably in the course of internal partial delamination, with the intake of water. The breakdown of the outer layers of the kraft fiber, as well as some of the S2 sublayer, may partly result in attached cellulosic fibrils extending outward from the main part of the fiber. When such fibril material becomes detached during the course of refining, they become part of the fines component of the fibrous slurry. These narrow cellulosic materials will be in contact with large amounts of water, leading to the question as to whether this can be regarded as another contribution to swelling.
Preparation of nanofibrillated cellulose
Considerably greater increases in apparent swelling can be achieved if mechanical compression and shearing of cellulosic material is continued way beyond the levels associated with ordinary papermaking, i.e., with the production of microfibrillated (MFC) or nanofibrillated cellulose (NFC) (Eichhorn et al. 2010; Lavoine et al. 2012; Hubbe et al. 2017). The resulting materials, often consisting of highly branched and diverse strands of cellulose having diameters less than about 100 nm, typically have no restraints on their swelling. When the amount of water is kept in a suitable range, hydrogels are formed (Hubbe et al. 2013). However, in violation of the usual expectation for a proper hydrogel, such mixtures can continue to take up more water until they become viscous non-gelled suspensions (Hubbe et al. 2017). By contrast, a proper hydrogel will be composed of hydrophilic (usually negatively charged) polymer segments with sufficient covalent cross-linking to prevent their dissolution (Ganji et al. 2010; Hubbe et al. 2013).
A variety of mechanical devices, as well as different chemical and enzymatic treatments, can be employed to prepare nanofibrillated cellulose and related products from a variety of cellulosic materials. From a practical standpoint, there can be advantages of simply using ordinary pulp refining equipment, which is a mature technology employed at paper mills throughout the world (Gharehkhani et al. 2015). By carrying out very large numbers of repeated passages through such refiners, gel-like MFC suspensions can be obtained (Chen et al. 2016; Shafiei-Sabet et al. 2016). At a laboratory scale, various homogenizer devices can be used effectively to achieve different degrees of fibrillation by varying the number of passes. For instance, Pääkkö et al. (2007) used a high-pressure homogenizer to convert oxidized cellulose into nanocellulose suspensions having exceptionally high viscosity at a given solids content. Others have used devices in which a pair of jets collide directly, causing suspended cellulose particles to become fibrillated (Dimic-Misic et al. 2016). Finally, friction-grinding equipment can be used as a way to prepare highly fibrillated cellulose suspensions at the lab scale (Hassan et al. 2011). Products of such shearing devices, as long as they are kept in their wet state, can swell with water without any upper limit. However, such mixtures may pass from being gel-like to being liquid-like at some point, since typically there are no permanent attachments between the adjacent particles while they remain in a wet condition.
Aqueous Conditions
Colloidal science considers the forces of interaction between materials at very short distances in aqueous media (Hubbe and Rojas 2008; Benselfelt et al. 2023). Thus, many of the factors related to water that affect swelling of cellulosic materials fall within that field. Here the attention will be on the evidence, considering in what ways the swelling has been found to depend on pH, the nature of charged groups on cellulosic materials, the concentrations of salts, and effects due to various multivalent ions.
Table 3. Reported Effects of pH on the Swelling of Cellulosic Materials