Hydrophobic molecular treatments of cellulose-based or other polysaccharide barrier layers for sustainable food packaging: A Review

Abstract

Paper, nanocellulose, and other polysaccharide-based materials can be excellent candidates for food packaging barrier layers, except that they tend to be vulnerable to moisture. This article reviews published research describing various chemical treatments having the potential to render hydrophobic character to such layers. Emphasis is placed on systems in which hydrophobic monomers are used to treat either particles or sheets comprised largely of polysaccharides. A goal of this review is to identify combinations of materials and procedures having promise for scale-up to industrial production, while providing effective resistance to moisture. The idea is to protect the underlying polysaccharide-based barrier layers such that they can continue to impede the transfer of such permeants as oxygen, greases, flavor compounds, and water vapor. A further goal is to minimize any adverse environmental impacts associated with the treatments. Based on the research articles considered in this review, promising hydrophobic treatments can be achieved involving silanes, ester formation, other covalent interactions, plasma treatments, and to some extent by various treatments that do not require formation of covalent bonds. The article is designed such that readers can skip ahead to items of particular interest to them.

Full Article

Hydrophobic Molecular Treatments of Cellulose-based or Other Polysaccharide Barrier Layers for Sustainable Food Packaging: A Review

Dorota B. Szlek, Autumn M. Reynolds, and Martin A. Hubbe *

Paper, nanocellulose, and other polysaccharide-based materials can be excellent candidates for food packaging barrier layers, except that they tend to be vulnerable to moisture. This article reviews published research describing various chemical treatments having the potential to render hydrophobic character to such layers. Emphasis is placed on systems in which hydrophobic monomers are used to treat either particles or sheets comprised largely of polysaccharides. A goal of this review is to identify combinations of materials and procedures having promise for scale-up to industrial production, while providing effective resistance to moisture. The idea is to protect the underlying polysaccharide-based barrier layers such that they can continue to impede the transfer of such permeants as oxygen, greases, flavor compounds, and water vapor. A further goal is to minimize any adverse environmental impacts associated with the treatments. Based on the research articles considered in this review, promising hydrophobic treatments can be achieved involving silanes, ester formation, other covalent interactions, plasma treatments, and to some extent by various treatments that do not require formation of covalent bonds. The article is designed such that readers can skip ahead to items of particular interest to them.

DOI: 10.15376/biores.17.2.Szlek

Keywords: Water resistance; Contact angles; Polysaccharides; Silanes; Paper; Films

Contact information: Department of Forest Biomaterials, North Carolina State University, Campus Box 8005, Raleigh, NC 27695-8005; Email: hubbe@ncsu.edu

INTRODUCTION

Motivating Goals

Protect the barrier layer

The motivating goal of this review article is to identify propitious approaches to achieve effective and eco-friendly water-barrier capability for use in single-use food packaging systems. As has been described elsewhere, it is possible to form various polysaccharides, including nanocellulose and related composites, into dense layers having promising ability to block the transport of oxygen (Aulin et al. 2010; Hubbe et al. 2017). Because of the high density of hydrogen bonding that can be achieved within highly refined paper-like structures, especially when coated with materials such as nanocellulose or starch, it is possible to achieve high resistance to grease, even without usage of fluorochemicals (Hubbe and Pruszynski 2020). Such structures are highly regarded from an environmental standpoint because they are mainly composed of photosynthetically renewable materials that are inherently biodegradable. On the other hand, such structures composed of polysaccharides, including cellulose, are susceptible to effects of aqueous solutions, which can cause the films and structures to swell, weaken, and become non-resistant to oxygen, water vapor, and other permeants (Fotie et al. 2020).

A second goal of this review is to consider a parallel strategy in which the surfaces of cellulose nanomaterials are molecularly treated to render them hydrophobic. In such cases, a hydrophobic molecular treatment has the potential to render polysaccharide-based material compatible with various hydrophobic plastics (Dufresne 2011). In such a form, the nanocellulose (or other polysaccharide-based particles) could perform better in various water-resistant composite films in combination with suitable oleophilic matrix polymers.

As represented by the illustration in Fig. 1, the present review focuses on processing strategies that involve molecular treatments of surfaces – often at the monomolecular or sub-monomolecular coverage level.

Fig. 1. Conceptual illustration of an ideal hydrophobic molecular treatment of a polysaccharide substrate. The orange circle represents a part of the hydrophobic compound that is anchored to the substrate. The green “stick” represents the hydrophobic part. Dotted lines suggest hydrogen bonding, which would be vulnerable to wetting by water.

In principle, it is possible to convert the hydrophilic surfaces of cellulose and various other biopolymers to a highly hydrophobic state by surface derivatization or adsorption (Cunha and Gandini 2010a,b; Samyn 2013; Hubbe et al. 2015b; Farhat et al. 2017). Such approaches are appealing from a general environmental perspective, since the effects often can be achieved with a minimum amount of material. For instance, a monomolecular layer or sublayer of a suitably anchored hydrophobic compound may be sufficient to achieve a high level of water repellency (Hubbe 2007; Oh et al. 2011; Kumar et al. 2016; Krishnamurthy et al. 2020). Questions to be considered in this review article include (a) whether such treatments are sufficient to meet the needs for various packaging applications and (b) whether such treatments can meet expectations for eco-friendliness.

Minimal adverse environmental impact

In present production, a leading strategy to protect paper and other polysaccharide-based structures from wetting involves application of a polyethylene laminate film (Borch 1991; Vinayagamoorthy 2017). For instance, most milk cartons are assembled as a sandwich, with a paperboard layer between two polyethylene laminate layers (Kirwan 2013). To prevent failure of such cartons at cut edges (e.g. at the bottom interior), and at pinholes in the laminate film, the paperboard will have been hydrophobically treated – usually with alkylketene dimer sizing agent (Dumas 1981; Ehrhardt and Leckey 2020). The format is illustrated in Fig. 2.

Fig. 2. Basic layered cross-sectional format of a typical milk carton. The green coloration indicates a hydrophobic nature, whereas the pink coloration indicates a weak negative ionic charge when in contact with an aqueous solution.

It is technically possible to recycle milk cartons, separating and recovering both the papermaking fibers and the plastic (Srivatsa and Markham 1993). However, most cartons currently are not recycled. This may be due to such factors as the wet condition of many used food packages, concerns about contamination by food, and the complex nature of such waste materials. Rather, a high proportion of food-related waste becomes landfilled (Kakadellis and Harris 2020). Some other modern single-use food containers can be even more difficult to recycle, due to strongly-adhering layers of different materials, such as aluminum foil (Keles and Dundar 2007) or non-biodegradable plastics (Mulakkal et al. 2021), which are often firmly attached to a layer of paper. When such packaging material becomes tossed out as litter or follows storm drains into water bodies, it contributes to the load of non-degrading matter present in the environment (Eriksen et al. 2014; Jambeck et al. 2015; Avio et al. 2017). As such, it can interfere with marine life in profound ways (Gregory 2009). In addition, the fluorochemical treatments that have been relied upon to achieve greaseproof characteristics in products for fast foods have raised environmental concerns (Curtzwiler et al. 2021). The perfluorochemical treatments are bioaccumulative and resistant to breakdown in the environment (Houde et al. 2011; Kabadi et al. 2018; Trier et al. 2018). By contrast, certain polysaccharide-based films that have been molecularly surface-treated to render them hydrophobic have shown rapid biodegradation (Chen et al. 2021). There is a need for more research of such issues.

Bioplastic films, some of which can be melt-extruded, have received much attention as a promising option for replacing petroleum-based polyolefin films. Though such films can serve as effective barriers to aqueous media (Singha and Hedenqvist 2020; Attallah et al. 2021), concerns have been raised regarding their biodegradability (Emadian et al. 2017). For instance, it appears that the rate-determining step for degradation of poly(lactic acid) (PLA) in the environment is abiotic, and that a temperature over about 55 °C is needed to bring about meaningful degradation (Agarwal et al. 1998; Karamanlioglu et al. 2014; Hubbe et al. 2021). Even though PLA can be prepared from plant-based materials, it persists for a very long time without degrading in natural environments, such as soils and seawater.

In many studies, relatively thick layers of hydrophobic polymers have been prepared by casting from nonaqueous solution, followed by evaporation (Rhim and Ng 2007). Though such technologies can be effective for preparation of water-resistant plastic film layers, there can be extra expenses involved in recovery of the solvents (Kim et al. 2014) and there are concerns about environmental effects of the solvents (Chemat et al. 2019; Fadel and Tarabieh 2019). So-called green solvents can be employed as a means to decrease such concerns (Clarke et al. 2018; Sheldon 2019). However, the macroscopic nature (many times thicker than a monolayer) suggests a much greater time required for biodegradation.

Yet another approach involves mixing a polysaccharide-based aqueous solution with suitable water-soluble but relatively hydrophobic copolymers such that the resulting film is hydrophobic. This approach can involve addition of such copolymers as styrene maleic anhydride (SMA) or styrene acrylate (SA) to a solution of starch that is applied to paper’s surface at a size press of a paper machine (Iselau et al. 2015, 2018; Bildik Dal and Hubbe 2021). Upon drying, the amphiphilic copolymer becomes oriented at the surface in such a way as to resist wetting by water. Though such approaches have merit, bulk coatings such as these will be regarded as outside of the scope of the present article.

To address the need for truly biodegradable food packaging systems, while still providing effective protection against penetration of aqueous solutions, it will be assumed in the present review article that promising solutions are likely to involve (a) systems that are mainly based on photosynthetically renewable materials, (b) hydrophobic treatments of surfaces such that water resistance is achieved with approximately a monolayer of coverage, and (c) application systems that mainly avoid the utilization of organic solvents.

The idea that protection against water can be achieved with approximately a monolayer of well-chosen and anchored monomers is not new. The paper industry has relied for many years on such internal sizing treatments as alkylketene dimer (AKD) and alkenylsuccinic anhydride (ASA) to achieve a range of water resistance (Dumas 1981; Hubbe 2007). Commercial specimens of both ASA and AKD are understood to contain a mixture of alkyl chain lengths. These additives are conventionally added to the cellulose fiber slurry before the formation of the paper sheet, and they spread and become covalently bound to the surfaces of the cellulosic fibers during the drying process. For example, AKD is the most common internal sizing treatment for achieving a hydrophobic paper structure in milk cartons (Dumas 1981; Ehrhardt and Leckey 2020). Such cartons, as already mentioned, ordinarily are protected on both sides by laminated films of polyethylene. However, the AKD treatment is needed due to cut edges of the sandwich structure (inside of each carton) as well as pinholes in the plastic layers (Tufvesson and Lindström 2007). In addition to conventional internal sizing agents used by papermakers, a wide range of treatment options have been considered in the scientific literature. These treatments, which are tabulated in the appendix of this article, are a main focus of this review.

Not interfere with a barrier layer

A key requirement for any treatment intended to impart hydrophobic character to an eco-friendly packaging system may be that it should not interfere with or defeat other required functions of the same layer of material. Two such requirements can be critical. First, the treatment ought not to defeat the oxygen-blocking ability of a film layer that is intended to have that capability. Second, the treatment ought not to harm the strength of the layer, especially in cases where the layer is intended to provide strength to the package. Such interferences are represented schematically in Fig. 3, where it is suggested that the presence of various compounds might interfere with the hydrogen-bonded structures within polysaccharide-based films.

Fig. 3. Sketch illustrating the concept of interference of hydrogen bonding within a polysaccharide-based structure due to the presence of certain hydrophobic compounds

Studies have shown that incorporation of various hydrophobic components or plasticizers into nanocellulose films, during their preparation, have the potential to degrade the oxygen barrier performance of that film (Lagarón et al. 2004; Hansen et al. 2012). The hydrophobic groups can interrupt the patterns of hydrogen bonding, thus providing less resistant paths for diffusion of the nonpolar molecules. In addition, water itself tends to swell and plasticize such structures, greatly increasing their oxygen permeability (Aulin et al. 2010).

Limited studies have considered effects on strength when hydrophobic materials have been reacted on the surface of paper or other polysaccharide-based materials. It has been observed in industry that internal sizing of paper with such agents as ASA and AKD generally does not interfere with paper strength. This is despite the fact that the hydrophobic groups would be expected to get in the way of potential hydrogen bonds that otherwise could form across the zone between adjacent cellulosic fibers in molecular contact. It had been proposed that this is because the emulsified hydrophobic agents mainly remain in the form of droplets or waxy particles and do not spread to a significant extent over the fiber surfaces until near the end of the paper drying process (Hubbe 2014). By that point in the process, the bonded areas between the fibers are already well established. The increased temperature of the paper, after most of the water has been evaporated, can be expected to increase the vapor pressure of the ASA or AKD molecules, allowing them to migrate. Therefore, the mobilized AKD or ASA molecules, transported by surface diffusion (Shen et al. 2002; Shen and Parker 2003) or diffusing in the vapor phase (Akpabio and Roberts 1987; Yu and Garnier 2002; Zhang et al. 2007), are restricted to the remaining air-solid interfaces. This concept was recently confirmed by Korpela et al. (2021), who compared AKD and rosin soap sizing systems. The AKD sizing system had no adverse effect on paper strength. By contrast, rosin soap sizing, which is known to involve spreading of the sizing agent already in the wet state before formation of the paper sheet, thus decreased paper strength. In other relevant work, Bildik et al. (2016) showed that when AKD was dissolved in heptane, applied to an existing sheet of paper, followed by drying, the strength of the paper actually increased. In other words, rather than interfering with paper strength, the AKD material appeared to function as a kind of matrix phase within a paper-based fiber composite. These findings are consistent with the hypothesis that already-established bonded areas between cellulosic fibers tend not to be affected by the migrating sizing compounds.

The concerns just mentioned often can be overcome by use of a multi-layer structure. In fact, the use of multiple layers, each contributing different attributes, is a common strategy in creating packaging solutions (Ferrer et al. 2017; Helanto et al. 2019; Reichert et al. 2020). Thus, it will be tentatively assumed, in the presentation of this review, that a high-performing, low-cost, eco-friendly barrier to aqueous fluids will have high value even in cases where one or more additional layers may be required to provide strength or to block the transfer of oxygen, oils, fragrances, or water vapor, etc.

Factors Affecting Speed, Scalability

To be interesting to industrialists, each candidate unit operation needs to be suitable for scale-up to industrially relevant speeds and dimensions. Each unit operation, as well as the related equipment, must be considered relative to any limitations in maximum speeds. In addition, any required chemical reactions need to be compatible with continuous processing at high speed. Some types of unit operations of interest include vapor-phase application, plasma treatments, and the use of aqueous emulsions. These three approaches are illustrated schematically in Fig. 4. As will be discussed later, these three general approaches have been shown to offer favorable combinations of relatively quick hydrophobization, avoidance of organic solvents, and suitability for scale-up.

Fig. 4. Three solvent-free approaches (set off by the vertical bold lines) to hydrophobic molecular treatment of polysaccharide substrates. Note that the black circles represent odd-electron (free radical) groups. The light-blue wavy lines represent cationic polyelectrolytes.

Vapor-phase application

It is desirable if a hydrophobic compound to be conveyed in gaseous form to and reacted at the surface of a target material. Some of the silane-based treatment procedures to be considered in later sections of this article involve vapor-phase exposure and reaction (Fadeev and McCarthy 2000; Cunha et al. 2010b; An et al. 2011; Oh et al. 2011; Glavan et al. 2014; Lazzari et al. 2017; Yu et al. 2019; Jankauskaite et al. 2020; Zhao et al. 2020; Wulz et al. 2021). Acid chlorides, likewise, are sufficiently reactive that they can be transported as vapor to a surface that bears –OH groups and allowed to react, forming esters (Berlioz et al. 2009; Fumagalli et al. 2013; Wulz et al. 2021). Zhang et al. (2007) likewise achieved successful vapor-phase sizing with ASA, which is an anhydride of a dicarboxylic acid; parallel tests attempted with AKD failed, apparently due to the decomposition of AKD when it was heated. Yuan et al. (2005) likewise achieved successful vapor-phase reduction of the surface energy of a cellulose surface by treatment with an anhydride.

Plasma application

Plasma treatments of polysaccharide-based surfaces can be regarded as a distinct class of vapor-phase treatment (Siow 2018; Zhang et al. 2018). A key distinction is that the treatment is carried out in the presence of highly energized gas molecules, which induce reactivity to the hydrophobic species and to the surfaces. Belgacem and Gandini (2005), Kalia et al. (2009), and Belgacem et al. (2011) reviewed work involving such systems. In the case of silane-based treatments, plasma application can be used to activate hexamethyldisiloxane so that it reacts effectively and becomes bound to the substrate (Creatore et al. 2001, 2002; Mai and Militz 2004; Schneider et al. 2007, 2009; Deilmann et al. 2008a,b; Avramidis et al. 2009; Siliprandi et al. 2011; Chen et al. 2017; Mitschker et al. 2018; Yang et al. 2018; Kakiuchi et al. 2019; Cerny et al. 2021; Ma et al. 2021). Plasma also has been used to graft perfluorinated alkyl chains onto cellulose (Kong et al. 1992; Sahin et al. 2002, 2007; Navarro et al. 2003; Zhang et al. 2003; Balu et al. 2008; Toriz et al. 2008; Mirvakili et al. 2013; Siro et al. 2013). Starostin et al. (2015, 2016) used tetraethyl orthosilicate as an additive material for plasma treatment of polymers. Samanta et al. (2012) used a plasma system to react butadiene at cellulosic surfaces. In summary, plasma is able to achieve effective grafting of a wide range of compounds, not relying on the presence of reactive groups such as silanes, acid chlorides, or anhydrides.

Aqueous treatments involving emulsions

From an environmental standpoint, water-based formulations are generally regarded with favor. As a means to impart hydrophobicity, there are two main challenges. First, one has to figure out a way to convey a water-reactive hydrophobic compound in a water phase. The second challenge is that energy needs to be expended in a subsequent drying step. The first goal can be met by preparing the mixture as an emulsion, in which the droplets of hydrophobic compound are suspended in the water. In the case of AKD, such emulsions are formed above the melting point of the AKD and then immediately cooled, forming a dispersion of solid particles. Liang et al. (2013) used such an approach to treat cotton fabrics with a fluorinated trimethoxysilane. Related methods are widely used within the paper industry to treat the fiber furnish before the formation and drying of the sheet (Hubbe 2007; Ashish et al. 2019). In paper machine applications, the wet web needs to be dried in any case. Thus, the drying process can be used to induce ester formation between hydrophobic compound and a paper surface. In the case of ASA size, the reaction goes to completion, even before the end of the drying process. AKD is less reactive (Lindström and Larsson 2008). Thus, AKD may continue to react with the paper surface as the paper is in the form of a large reel or individual rolls, the interiors of which remain hot inside due to their size. Cationic polyelectrolytes such as cationic starch can be a good choice as an emulsifying agent. The starch can serve as a steric stabilizer, keeping the hydrophobic particles from colliding and coalescing. The cationic charge helps to retain the emulsion particle on fiber surfaces during paper’s formation before it is dried. Though the general procedure is not as inherently fast as vapor-phased treatments, it can be appropriate when a material needs to be dried anyway during its manufacturing process.

Options for continuous treatment

The papermaking process, as just discussed, is an example of a continuous industrial process. There are other related continuous unit operations that can be used for rendering such material hydrophobic. For industrial applications, it is important that such operations can be carried out at a large scale and relatively high speed. For example, plasma treatment (Starostin et al. 2015) and chemical vapor deposition technology (Alf et al. 2010) have been considered for roll-to-roll application (Alf et al. 2010). Continuous application also has been studied for silane treatment (Yu et al. 2019). However, there is a continuing need for more research to be carried out at a pilot scale, thus providing a bridge between theoretical studies and commercial production.

Review Articles

This review builds upon earlier progress, much of which already has been reviewed in previous articles. Selected reviews are listed in Table 1, along with their areas of focus.

Table 1. Selected Review Articles Dealing with Hydrophobization of Cellulosic or other Polysaccharide-based Surfaces

THEORETICAL BACKGROUND

To prepare for later discussion of specific types of hydrophobic modifications of polysaccharide surfaces, this section will review some principles that pertain in many situations. This will include a discussion of the most widely considered polysaccharides (cellulose, hemicellulose, starch, and chitosan), concepts related to hydrophobicity and wetting, procedures for relatively simple removal of various compounds from the surfaces (i.e. removal of contaminants), categories of treatment processes, some chemical principles, durability, biodegradability, and issues related to the use of multilayer structures in food packaging.

Polysaccharide Surface Chemistry and Barrier Properties

Cellulose

Sustainable single-use packaging can be based on polysaccharides as a main category of ingredients. The word polysaccharide refers to sugar polymers, i.e. photosynthetically renewable and biodegradable materials. The component monomeric sugars, e.g. glucose in the case of cellulose, are highly soluble in water. Cellulose can be described as a linear polymer of glucose in which the anhydroglucose units are connected by β-1,4 glycosidic linkages. The weight-average degree of polymerization of cellulose within cotton was reported as 3335 (Ling et al. 2019). For cellulose in wood, the corresponding values are about 9500 or 9600 and possibly as high as 15,000 (Goring and Timell 1962). The fact that cellulose is quite insoluble has been attributed to its structural regularity, linear form, and its tendency to form crystalline zones having high physical density and a highly regular pattern of hydrogen bonding both within (intra-) and between (inter-) the adjacent cellulose chains (Lindman et al. 2010). As explained in the cited article, the dense pattern of hydrogen bonding is supplemented by van der Waals forces, which play a dominant role in certain crystal planes of natural cellulose.

An intermediate degree of hydrophilic character of purified cellulose surfaces can be attributed mainly to the presence of some –OH functional groups, which are polar and which are able to form hydrogen bonds with water molecules (Hatakeyama and Hatakeyama 1998). The effects of these –OH groups depend on their orientation. This phenomenon was demonstrated by Yamane et al. (2006), who regenerated solid cellulose from solutions by addition to contrasting fluid media. When the cellulose was regenerated in the presence of water, it developed a hydrophilic surface, i.e. a low contact angle with water. By contrast, if the cellulose was regenerated within a non-polar medium, it developed a relatively hydrophobic surface. Figure 5, which uses the atom locations shown by Yamane et al. (2006), shows how the hydrophilic –OH groups are located above and below the cellulose chain when it is presented in the “front” view in the cited article. The hydrogens that are bonded to oxygen are shown in a brighter blue to represent their partial positive charge that is induced by the electronegative oxygens. By contrast, hydrogens not associated with oxygen mainly are presented above and below the chain shown in the “edge” view. These findings are consistent with the relatively high levels of crystallinity in typical samples of cellulose, in combination with the fact that the –OH groups on a cellulose molecule are oriented in a planar manner. Thus, different crystal planes of cellulose have different wettability.

Fig. 5. Representations of cellulose chains in “front” and “edge” view (redrawn based on Yamane et al. 2006), showing that hydrophilic –OH groups are mainly on certain faces, and at least one face of the cellulose-I crystal generally lacks –OH groups

Plant-derived cellulose has relatively few carboxylic acid groups (-COOH) before industrial processing. Strong alkaline pulping, as in the case of the kraft process, can develop –COOH groups (metasaccharinic acid) at the reducing ends of cellulose macromolecular chains (Van Loon and Glaus 1997). The –COOH groups of polysaccharides typically have pK_a values in the range of about 3.3 to 5, meaning that at neutral pH most of them will be in dissociated form, giving them a negative ionic charge (Laine et al. 1996). The high polarity of charged groups, depending on their frequency at a surface, can provide a relatively large contribution to water-wettability (Hansen 2007; Notley and Norgren 2010).

The barrier properties of cellulose have been well demonstrated in the case of nanofibrillated cellulose (NFC), which is also referred to as cellulose nanofibril. A review article by Aulin et al. (2010) documents the development of high resistance to both oxygen and oils when aqueous suspensions of NFC are formed into a film. The impenetrability to these non-polar substances, when dry, has been attributed to a high film density and a high cohesive energy density, both of which can be attributed to a high density of hydrogen bonding (Lagarón et al. 2004). However, these barrier properties have been found to deteriorate when such films are exposed to liquid water or high levels of relative humidity.

Hemicelluloses

The hemicelluloses present within woody materials typically provide the strongest contribution to hydrophilic character. There are two main roles that can be envisioned for hemicellulose in packaging. First, hemicellulose serves as a bonding agent between the fibers in paper products. Second, hemicellulose can be considered as an option for preparing thin films that might be applied onto paper surfaces (Hansen et al. 2012; Borjesson et al. 2019; Shao et al. 2020). Trees commonly used for papermaking have about 25 to 35% of hemicelluloses (Pettersen 1984). Hemicelluloses can be briefly described as copolymers of two or more types of sugar unit. Degrees of polymerization of hemicellulose in wood are typically in the range of 100 to 200 (Pettersen 1984). In contrast to cellulose, hemicelluloses have irregular structures, including side groups or acetylation along the chain. The plural form of the word hemicelluloses is used within scientific literature due to the presence of two or more copolymer structures in a typical type of wood. The irregular structure of hemicelluloses implies a higher accessibility of water to –OH groups, leading to a higher tendency to wet and to swell in water, compared to cellulose. This concept is supported by the finding that the swelling of wood is decreased progressively during the thermal destruction of hemicellulose (Repellin and Guyonnet 2005). A contribution to negative charge, in the range of 2 to 50 μeq/g of pulp, can be attributed to the susceptibility of hemicelluloses to hydrolytic cleavage of acetyl groups during alkaline treatment, and the value can rise to 100 μeq/g with peroxide bleaching (Pranovich et al. 2003). Various researchers, as discussed in earlier review articles, have considered the incorporation of hemicellulose into films that can be potentially useful in packaging (Cunha and Gandini 2010a,b; Farhat et al. 2017). As described in the cited articles, hydrophobic treatments or hydrophobic extruded layers are required when using hemicellulose-based films in such applications. Figure 6 provides examples of hemicelluloses present in softwood (Eronen et al. 2011) and hardwood (Ebringerova and Heinze 2000) pulps.

Fig. 6. Major components of softwood and hardwood hemicelluloses

Lignin

Relative to cellulose and hemicellulose, the lignin component of woody materials is more hydrophobic, especially in its natural state (Kang et al. 2019). Lignin can be generally described as the product of various reactions among monolignol compounds, such a coniferyl alcohol, sinapyl alcohol, and p-coumaryl alcohol (Boerjan et al. 2003). These become joined during the biosynthesis of lignin by oxidative coupling. The lignin in softwoods (conifers) is mainly derived from coniferyl alcohol, yielding guaiacyl lignin groups. The lignin in hardwood species has a majority of lignin structures derived from sinapyl alcohol, yielding syringyl lignin groups. Figure 7 shows examples of two of the most common linkage structures present in native softwood lignin (Chang and Jiang 2019).

The generally hydrophobic nature of lignin can be attributed to its high content of aromatic groups, thus giving a relatively high ratio of carbon to oxygen atoms (Maksimuk et al. 2020). The cited article says that carbohydrate-free lignin can have a composition of 70% carbon, 6.5% hydrogen, and 23.5% oxygen on a mass basis. By contrast, the chemical formula for cellulose (and various other polysaccharides) implies 40.00% carbon, 6.67% hydrogen, and 53.33% oxygen. Note that in the examples shown in Fig. 7, though some hydrophilic phenolic –OH groups will be present, a majority of such positions will be occupied by the more hydrophobic methoxyl (-OCH₃) groups. Also, though some aliphatic –OH groups will be present in lignin, the proportion of such groups is much lower than in polysaccharides. Within a tree, the lignin provides rigidity, protection against compression, and relatively strong resistance to natural biodegradation.

Fig. 7. Examples of the two most common linkage structures within native softwood lignin

Extractives

In addition to the three main ingredients of wood (cellulose, hemicellulose, and lignin), the trees commonly used in papermaking commonly contain about 1 to 2% of extractives. Higher levels, e.g. 2 to 29%, can be present in tropical woods (Choong and Achmadi 1991). As illustrated in Fig. 8, the extractives can include fatty acids, triglyceride fats comprising fatty acid groups, various unsaponifiable compounds, and, in the case of conifers, resin acids, including rosin and related compounds in the terpene family (Back and Allen 2000).

Fig. 8. Common forms of extractives present in softwood material

The hydrophobic nature of the extractives is evident from their name, since the solvents used for the extraction of wood are oleophilic. Although the relative content is generally low, the hydrophobic nature of extractives provides a driving force for them to migrate to the air-solid interfaces of wood-based materials during the passage of time (Swanson and Cordingly 1956). In addition, the drying of wood has a tendency to draw extractives to the point where evaporation is taking place, often near or at the surface (Mottonen and Karki 2010). Extractives also have a strong tendency to self-associate (Hubbe et al. 2020). The likely presence of extractives in cellulosic materials can pose a challenge when one attempts to determine the reasons for hydrophobicity of cellulosic specimens after various surface treatments; it is always possible that residual extractives are contributing to the observed effects, regardless of what other compounds have been added.

Chitosan

Due to its widespread availability as a byproduct of crustacean harvesting, chitin is widely regarded as a promising polysaccharide for eco-friendly product formulations (Elsabee et al. 2009; Aranaz et al. 2010; Deng et al. 2017; Bhardwaj et al. 2020). By treatment of crustacean shells with strong base, chitin can be converted to chitosan, which has a chemical structure almost identical to that of cellulose. The main difference is that amine groups are present in the C2 positions in place of the –OH groups that would be there in the case of cellulose (Dash et al. 2011). Due to the presence of the amine groups, chitosan can be solubilized in weakly acidic water, especially after addition of acetic acid, such that the pH is below the pKa of the functional amines. As illustrated in Fig. 9, the acidity results in protonation of the amine groups. The solubilized chitosan can subsequently be formed into films, which can be considered for medical (Rinaudo 2006; Elsabee et al. 2009; Dash et al. 2011) or food-contact purposes (van den Broek et al. 2015; Deng et al. 2017; Vikele et al. 2017). It has be observed that chitosan films, even when relatively pure, can resist wetting by water (Cunha et al. 2008; Vikele et al. 2017; Bhardwaj et al. 2020). Such hydrophobicity sometimes has been attributed to the presence of hydrophobic extractives (Cunha et al. 2008). Recently it was suggested, however, that such hydrophobicity comes about due to the self-orientation of the films during their drying; the idea is that the non-polar sides of the chitosan chains become exposed outwards toward the air phase, thus decreasing the free energy of the system (Hubbe 2019).

Fig. 9. Chemical structure of chitosan, also showing its transformation to a cationic species when exposed to acidic conditions

Starches

When plants have a need to store energy in a recoverable form, starch biosynthesis is a prime evolutionary option. When the energy is needed by the plant, the starch can easily be converted to glucose by the action of amylase enzymes (van der Maarel et al. 2002). Starch availability has subsequently become greatly amplified by the cultivation practices of humans, and it has become widely used in various formulations, including films for food packaging (Jimenez et al. 2012; Prabhu and Prashantha 2018). Due to its water-solubility, some differences in the way it crystallizes, and its abundance hydroxyl groups, starch is widely regarded as a hydrophilic material (Herman et al. 1989). Cunha and Gandini (2010a,b) reviewed modifications to convert starch to a hydrophobic material. In addition, the wettability of starch surfaces has been shown to be affected by the orientations of the macromolecular segments (Sundari and Balasubramanian 1997; Immel and Lichtenthaler 2000; Shrimali et al. 2018; Bildik Dal and Hubbe 2021). The challenge remains, however, that starch-based films are generally prone to swelling in water, and their barrier properties suffer due to the effects of high humidity, especially if a plasticizer is present (Mali et al. 2005).

Hydrophobicity and Wetting

To provide background for later discussions in this article, some aspects regarding surface hydrophobicity will be summarized here, including literature references to fuller descriptions. The principles to be discussed in this section are well established, and they can be helpful in understanding the role of the chemical treatments to be considered later in this review. Key areas of focus will be contact angles of water, morphological aspects affecting contact angle, and issues related to the compatibility of different phases that come into contact. These topics were reviewed in a more general manner by Sengupta and Han (2014), who focused on the best-established theories.

Contact angles

The ability of a material to resist penetration by liquid water or an aqueous solution can be estimated based on a knowledge of its contact angles, in combination with estimates of the size of its pores (Hubbe et al. 2015a). When tests are conducted on relatively flat surfaces, the contact angle, as well as the differentiation between wettable and nonwettable surfaces, can be defined as shown in Fig. 10.

Fig. 10. Schematic illustration defining the contact angle of a liquid on a flat substrate, where surfaces having contact angle values below 90 are defined as wettable (A), and those with contact angles above 90 are non-wettable (B)

Theoretical descriptions to account for contact angles generally start by assuming a perfectly smooth, flat, nonporous substrate having no chemical variabilities. On such an ideal surface, one can expect that the angle of contact would be a function of the interfacial tensions acting in three directions, as illustrated in Fig. 12. Young (1805) proposed such a concept and introduced Eq. 1,

γ_SV – γ_SL = γ_LV cos θ (1)

where γ_SV is the interfacial tension at the solid-vapor boundary, γ_SL is the interfacial tension at the solid-liquid boundary, and γ_LV is the well-known surface tension of the liquid, in the presence of its own vapor in the gas phase.

Fig. 11. Representation of the interfacial force balance proposed by Young

Whereas the interfacial tension in at the air-liquid interface (γ_LV) can be measured directly (Hartland 2004), there is no direct way to measure either of the other two interfacial tension terms in Eq. 1, i.e. the solid-liquid interfacial tension and the solid-air (or solid to vapor) interfacial tension. Because the solid will have a very limited ability to stretch reversibly, it can be questioned whether the two unknown terms can be regarded as thermodynamic quantities (Hubbe et al. 2015a). As discussed in the cited work, there have been protracted efforts to resolve the theoretical issues and to obtain practical results, but no comprehensive theoretical approach has received full acceptance. It follows that there is a need for fresh, creative thinking and a focus on general principles.

In general, it is known that wettability of a liquid on a certain smooth solid will be favored by a high degree of molecular interactions between the liquid and the solid. On the other hand, relatively poor wettability will be expected in cases where the molecules within the fluid interact more strongly with each other than with the molecules at the solid surface. This situation can be illustrated by the example of a droplet of pure water on a flat slab of pure paraffin wax. Within the water phase, the molecules interact strongly by means of hydrogen bonding, polar interactions (which may be treated as part of the hydrogen bonding interaction), and van der Waals – London dispersion forces (Liang et al. 2007). Among those forces, the interaction between the two phases involves only the dispersion forces.

Paraffin wax can be regarded as a prototypical non-wettable surface. It is composed just of alkyl chains. These are completely nonpolar and have only a moderate Hamaker constant, which governs the magnitude of the dispersion force interactions (Visser 1972, 1995). Surfaces that are coated by a monolayer for perfluorinated alkyl groups, e.g. polytetrafluoroethylene, have even lower values of the dispersion content due to the low polarizability of the electrons in the outermost molecular orbitals (Visser 1972). By contrast, if a surface contains exposed polar hydroxyl groups, as in the case of typical cellulosic materials, the surface free energy is expected to be much higher. The cellulosic surface would be expected to interact with a water droplet through hydrogen bonding and polar interactions in addition to the van der Waals – London dispersion forces. Based on these considerations, if the goal is to impede the spreading and permeation of water, then it makes sense to treat the cellulosic or other polysaccharide surface either by derivatizing the –OH groups or by covering them up. Such approaches will be discussed in this article.

Rates of permeation

If the pores are modeled as simple cylinders, perpendicular to the plane of a porous film, then the rate of penetration and the distance penetrated at time equal to t are given by (Lucas 1918; Washburn 1921),

dl/dt = γ_LV R cos θ / (4η l ) (2)

l = [(R γ_LV cos θ] t /(2η)]^1/2 (3)

where r is the equivalent radius of the pores (based on the cylindrical pore model), η is the dynamic viscosity, v is the average velocity of fluid flow into the capillary, L is the distance of permeation at time t, and t is the elapsed time after the initial wetting. These equations, though highly idealized, have been found to be useful as a starting point in understanding the rates of permeation of liquids through paper and related substrates (Aspler et al. 1984; Aspler and Lyne 1984). The model is illustrated in Fig. 12.

Fig. 12. Sketch illustrating the Lucas-Washburn model of the wetting of a porous solid

Morphological aspects

The measuring of contact angles of water onto cellulosic surfaces is prone to relatively high levels of scatter of the data. Much of that scatter usually can be attributed to various scales of roughness, or of porosity, on typical surfaces. Roughness can be expected to affect wetting contact angles in two general ways – equilibrium effects and hysteresis effect. The equilibrium effects were first described and estimated by Wenzel (1936). Wenzel essentially made a correction to Young’s equation. The purpose was to account for the greater amount of surface area, per unit of planar area, when the material is rough. Thus, Wenzel’s equation can be expressed as:

r_w [γ_SV – γ_SL ] = γ_LV cos θ (4)

In Eq. 4, r_w is the roughness coefficient (ratio of real surface area to planar surface area), and the other parameters have the same definitions as before. The practical effect of the roughness coefficient in Eq. 4 is that it predicts contact angles to be further away from 90 degrees, compared to a smooth surface. Thus, systems that would be classed as “wetting” (with contact angles below 90) would be predicted to be even more wetting when a realistic level of roughness is assumed, rather than using the model of a perfectly smooth surface. Likewise, systems that would be classed as “nonwetting” (angles > 90 on an ideal smooth surface) would be even more non-wetting when roughness is taken into consideration.

When the surfaces have large roughness or pores, relative to the scale of random wave action at the liquid surfaces (de Gennes 1985; Piao et al. 2010), then it is reasonable to anticipate hysteresis effects. Strong evidence of the importance of hysteresis effects is provided by observed differences in advancing and receding contact angles on real surfaces, including cellulosic surfaces (Gardner et al. 1991). Two idealized situations can impede the movement of a three-phase contact line on a wetted surface – relatively large, oriented features and deep pores. As described in more detail elsewhere (Piao et al. 2010; Hubbe et al. 2015a), when there is a ridge of roughness that exceeds the scale of wavelike thermal motion at the liquid surface, the local value of contact angle will not be the same as the contact angle relative to the overall plane of the surface. It makes intuitive sense that certain patterns of roughness can contribute to hysteresis. However, one usually does not know enough detail of the surface features to make accurate estimates of advancing or receding contact angles on rough surfaces.

To estimate the effects of pores or deep valleys on a surface, relative to contact angles and wettability, most researchers have based their calculations on the work of Cassie and Baxter (1944). The governing equation can be expressed as,

cos θ = f₁ cos θ ₁ + f₂ cos θ ₂ (5)

where θ is the observed contact angle, θ₁ is what the value of contact angle would have been if all of the surface had been of type “1”, θ₂ is what the value of contact angle would have been if all of the surface had been of type “2”, and f₁ and f₂ are proportional amounts of interface corresponding to the two compositions, each in comparison to a hypothetical planar surface area. When using the approach of Cassie and Baxter, it is important to be aware that there are some incorrect mathematical expressions that are in common usage (Milne and Amirfazli 2012). By setting one of the theta terms in Eq. 5 to 180, it is possible to account for the complete non-wettability of an empty pore. The equation then can be used to account for super-hydrophobic systems, which arise when small pillars of hydrophobic solid are surrounded by suitably deep valleys. The same equations also can account for instances where superhydrophobic systems fail, meaning that the valleys within such structures become filled with the wetting liquid (Hubbe et al. 2015a). Once that occurs, such systems are readily wetted by water.

Compatibility

In addition to affecting wettability and permeation by fluids, the hydrophobic treatment of a cellulosic surface also has potential to affect the compatibility of the material when it is being considered as a reinforcement within a composite. Some examples are listed in a recent review article that considered the use of cellulose-based reinforcements of different size within a series of different plastic phases (Hubbe and Grigsby 2020). Whereas the data reported in the literature failed to show any strong relationship between composite properties and the size of the reinforcing particles, the hydrophobic treatment of those particles has been repeatedly shown in the literature to provide a strength advantage. Another situation in which bonding at interfaces can be critically important in packaging applications is when a cellulosic film needs to adhere to an adjacent film layer, which might be oleophilic (Mittal 2010; Lee et al. 2019).

Two conditions need to be fulfilled in order to achieve strong adhesion at a phase boundary within a polymer matrix or between a cellulosic film and an oleophilic film. The first of these is that during the formation of that boundary, often at a temperature above the melting point of the matrix, one of the materials needs to fully wet the other surfaces (Good 1992). Due to the relatively low surface tensions of common polymers such as melted polyethylene, polypropylene, polyesters, etc., in combination with the relatively high surface free energy of dry cellulosic surfaces, it is reasonable to expect full wetting to occur. Studies have shown that such polymer melts have a low contact angle with clean, dry cellulosic surfaces, which is consistent with the concepts pioneered by Young (1805) and those who further developed theories related to wettability. The second condition, which is harder to achieve, is that the two phases either need to be connected by covalent bonds or there has to be a least a moderate amount of three-dimensional molecular overlap at the boundary, i.e. mixing among molecular segments from each adjacent phase (Aradian et al. 2000). This concept is illustrated in Fig. 13. Such overlap is likely to be absent at an interface formed between a hydrophobic polymer phase (for instance when it is cooling form a melted condition) and a polysaccharide phase. The reason is that the strongly hydrogen bonded material resists association with the molecular chains of the nonpolar phase (Blokzijl and Englberts 1993). Thus, the mutual solubility of the two phases will be low. A higher number of hydrogen bonds are able to form if the polysaccharide material bonds mainly with itself rather than having its segments mix to a significant extent with those of a polyolefin in an interfacial region. Such effects contribute to an inherent reluctance for different types of polymer chains to intermix with each other (Flory 1942).

Fig. 13. Concept of inter-mixing of polymeric segments at the boundary between two well-adhering phases

As will be seen later in this article, when discussing strategies for surface modification of cellulosic surfaces, various covalent reactions at surfaces of polysaccharide films and particle surfaces have the potential to attach a variety of alkyl or other functionalities. As will be reported, not only can such treatments render the surfaces hydrophobic, but they also can provide the polysaccharide surfaces with a means to locally intertwine with the adjacent matrix material, thus forming bonds strong enough to resist delamination at those interfaces (Hubbe and Grigsby 2020). Such findings are consistent with the intermixing of polymers at the interface, especially in the melt condition (Zhang et al. 2011).

Removal of Loosely Bound Substances

Before reviewing some fundamentals related to surface derivatization and grafting of polysaccharide materials, it is first important to consider issues related to the likely presence of loosely bound hydrophobic substances already present at those surfaces. Natural materials can contain a variety of waxes, oils, and fatty substances. As mentioned earlier, these may have already migrated to air-solid interfaces (Swanson and Cordingly 1956; Bildik Dal and Hubbe 2021). Though such unbound, generally monomeric substances can be expected to be compatible with various oleophilic polymer melt materials, they provide a point of weakness at the interface. Adhesion scientists often refer to weak boundary layers when they are explaining instances of low adhesion strength (Gardner et al. 2008).

As noted in a review article by Bajpai (2017), adhesion at cellulosic surfaces often can be improved by removal of loosely bound materials. This can be done with alkaline solutions, which have the ability to saponify esters, as well as to convert fatty acids to their soap forms (Mercantili et al. 2014) so that they are readily washed from the surfaces. It is likely that some of the benefit of mercerization, involving treatment of cellulosic materials with relatively strong alkali (Albinante et al. 2013), is due to removal of saponifiable monomeric compounds. Such effects also can be achieved by extraction with organic solvents (Belgacem and Gandini 2005).

Treatment Classes

In the next main section of this article, which deals with different chemical modifications, a variety of process options will be mentioned. Since each such option could be considered for a range of different chemistries, it makes sense to compare and describe the process options in advance. To begin, Fig. 14 provides a pictorial summary of the options, which include vapor-phase transfer, plasma, aqueous solvent, aqueous emulsion, non-aqueous solvents, and enzymatic treatments.

Fig. 14. Depiction of process options for hydrophobization of polysaccharide-based surfaces

Vapor-phase

Vapor-phase treatment can be an attractive option for industrial processing because it offers a potential way to avoid the handling of solvents, including the possible evaporation or collection of those solvents after the treatment. A key requirement is that the substance to be distributed must have high chemical stability above its boiling point. In addition, it has to be sufficiently reactive in order to achieve a suitable level of covalent bonding with the treated solids, presumably during a continuous processing operation.

Hydrophobization of polysaccharide-type surfaces by vapor-phase treatment has been widely reported (Fadeev and McCarthy 2000; Yuan et al. 2005; Oh et al. 2011; Fumagalli et al. 2013; Glavan et al. 2014; Jin et al. 2015; Lazzari et al. 2017; David et al. 2019; Yu et al. 2019; Jankauskaite et al. 2020; Leal et al. 2020; Zhao et al. 2020; Shang et al. 2021; Wulz et al. 2021). Fumagalli et al. (2013) noted that surface esterification, brought about by treatment with palmitoyl chloride, was restricted to surface reactions. The resulting cellulose nanocrystal aerogels were readily dispersible into organic media. Lindström and Larsson (2008) reviewed relevant literature and reported effective vapor-phase treatment of paper with alkenylsuccinic anhydride (ASA), leading to a high degree of hydrophobicity. A similar approach was found not to work for the alkylketene dimer (AKD) sizing agent, an effect that was attributed to decomposition of the AKD when it was heated (Zhang et al. 2007; Lindström and Larsson 2008).

Chemical vapor deposition is a gas-phase treatment in which the selected monomers are able to react with each other to form a polymeric layer on the treated surface (Alf et al. 2010). Such treatment, often carried out under vacuum, can be tailored to provide programmed patterns, i.e. a type of printing on the surface. Chemical vapor deposition has been used as a means to develop hydrophobicity on polysaccharide-based surfaces (Balu et al. 2008; Cunha and Gandini 2010a,b).

Plasma

Though plasma treatments can be regarded as vapor-phase treatments, they take place in the presence of energetic gas species. Free radical compounds present within a plasma can activate compounds present both within the plasma and on the treated surfaces (Andreozzi et al. 2005; Saleem et al. 2021). Hence, various components of the mixture can be caused to react covalently with a treated substrate (Belgacem and Gandini 2005). Plasma treatments are a favored way to treat surfaces with certain types of silane compounds, such as hexamethyldisiloxane (Avramidis et al. 2009; Starostin et al. 2016; Kakiuchi et al. 2019). Likewise, Balu et al. (2008) reported the effective preparation of superhydrophobic cellulose fibers by plasma treatment in the presence of pentafluoroethane, which is an otherwise non-reactive compound.

Aqueous emulsions

As a means of minimizing environmental impacts, there is an incentive to employ liquid water as a medium for conveying hydrophobic compounds to the surfaces of polysaccharide materials. A challenge is presented by the fact that the materials to be placed onto the surfaces generally are insoluble in water. One way to overcome this dilemma is to prepare an emulsion, using a suitable stabilizer (Mai and Militz 2004; Peydecastaing et al. 2006; Yuan et al. 2006; Dankovich and Hsieh 2007; Liang et al. 2013; Ganicz et al. 2020). Thus, the material to be distributed is in the form of tiny droplets (e.g. 1 µm). This is a very common approach used in modern papermaking when adding hydrophobic agents to the fiber suspension (Dumas 1981; Hubbe 2007; Lindström and Larsson 2008; Bildik Dal et al. 2020). In those applications, it can be advantageous to employ cationic starch or a cationic acrylamide copolymer as a stabilizer, since the positive charge will help to retain the sizing agent at fiber surfaces during the formation of the sheet. As another alternative, emulsions can be stabilized with nanoparticles; such formulations are known as Pickering emulsions (Bayer et al. 2009; Li et al. 2021b).

Non-aqueous solvent

Hydrophobic agents suitable for treatment of polysaccharide-based surfaces are often soluble in non-polar solvents. Some ideal attributes when selecting such a solvent might include high solubility of the compound of interest, relatively low boiling point (e.g. in the range 35 to 100 °C), adequate chemical stability, and sufficiently low viscosity. Such properties make it likely that the solvent can be used for casting of the solution on a specimen of interest, followed by evaporation (Kalia et al. 2014). The following studies employed casting from a nonpolar solvent as a way to place hydrophobic compounds at cellulosic surfaces (Siqueira et al. 2010; Bayer et al. 2011; Wang et al. 2015; Rukmanikrishnan et al. 2020a,b). An advantage of many organic solvents is that they do not react with such functional groups as chlorosilanes, acid chlorides, and anhydrides. By contrast, the possibility of premature reaction with water is a drawback of the emulsion systems just discussed. The use of solvents in the application system may incur additional costs for the solvent itself, the energy needed for its evaporation, and the capital and operating costs for solvent recovery (Seyler et al. 2006).

Enzymatic

An enzyme can be defined as a protein structure having the ability to catalyze a chemical reaction (Engel 2020). Because enzymes are biological products, they function in aqueous systems and generally are effective at convenient temperatures. Kudanga et al. (2010) used laccase enzyme to catalyze the reaction of fluorophenols to lignocellulosic surfaces. Dong et al. (2014) employed laccase enzyme as a means to attach dodecyl gallate to the surfaces of jute fibers. Likewise, Cusola et al. (2015) used an enzymatic approach to induce hydrophobicity of paper-based materials. All three sets of authors proposed a phenolic ether structure as a means of holding long-chain alkyl groups at the cellulosic surfaces. More recently, lipase was used to connect laurate alkyl chains to cellulose nanocrystals with the formation of ester linkages under aqueous conditions (Yin et al. 2020). Stepan et al. (2013) earlier had demonstrated the attachment of stearate groups to hemicellulose in the presence of lipase. An aqueous enzymatic derivatization of ethylcellulose with poly-hydroxybutyrate chains had been reported by Iqbal et al. (2014). Thus, it appears likely that enzymatic approaches will become more widely utilized for such surface treatments.

Chemical Principles

As an organizing concept, it will be assumed in this research that the most favorable hydrophobic surface treatments of polysaccharide surfaces, depending on the details of what is needed, will have a lot to do with energies. First to be considered will be the activation energies needed to form covalent bonds. Next will be the energies of adsorption, which may be important in cases where no covalent bonding is anticipated. Third, there may be a role of energy in hydrophobic associations that contribute to stabilization of a hydrophobic surface on a polysaccharide surface. The adsorption of polyelectrolytes has energy implications that differ from those of smaller molecules.

Activation energies

The importance of activation energies was already apparent in the earlier discussion of enzymes. As was noted, enzymes can promote the formation of specific covalent bonds. The rates of each of the covalent reactions to be considered in this review article will depend at least partly on its activation energy. Such issues will determine the degree to which some important reactions of interest take place and what can be done to promote the desired reactions. The rate-limiting step of many reactions, including those involved in derivatizations at polysaccharide surfaces, is often constrained by a need to reach a transition state. That energy of activation (E_a) generally corresponds to an energy saddle that lies between the reagent system and the bonded state (Gandour and Schowen 1978; Engel 2020). The situation is shown graphically in Fig. 15. It has been argued that the role of an enzyme is to help stabilize a transition state, which may effectively lower the energy of that state (Schramm 1998).

Fig. 15. Typical diagram of how the potential energy of interaction is likely to change as a function of the progress of a reaction, i.e. the reaction of a hydrophobic compound with the surface of a polysaccharide substrate

The concept of activation energy also can provide a framework for thinking about the types of covalent bonds that are likely to meet the requirements for effective and efficient modification of polysaccharide surfaces. The required input of energy to achieve reaction needs to be in the right range to be achieved at convenient temperatures and short time intervals. It has been argued that the tension within the five-membered ring of the anhydride group in the ASA sizing agent raises the energy of that species; thus, less energy is needed in order to reach the transition state for ester formation with the –OH groups of a polysaccharide surface (Hill et al. 1998; Shi et al. 2016). Thus, the rate of reaction of ASA with polysaccharide surfaces, such as cellulose, is relatively high. However, high reactivity of a certain anhydride compound with –OH groups on polysaccharides is often correlated with relatively high reactivity with water as well, leading to wasteful hydrolysis. Thus, acid chlorides, which have very high reactivity with the surface –OH groups, are usually restricted to water-free treatment systems, such as solvent application or vapor-phase application (see Table C in the Appendix).

Adsorption and desorption energies

Energy likewise may play a key role in systems aiming to modify surfaces by adsorption phenomena. Unlike the situations just discussed, there is usually little or no activation energy to overcome when the added compounds are becoming adsorbed from solution onto a target substrate. In typical cases, there is a favorable free energy of adsorption when the adsorbing species, initially present in the water phase, contains a hydrophobic group (Al-Ghouti and Al-Abisi 2020; Bildik Dal and Hubbe 2021), as is being considered here. In some cases the adsorption may be further favored by the presence of cationic groups, which are attracted to the negative ionic charges on typical polysaccharide-based substrates (Biswas and Chattoraj 1997; Al-Ghouti and Al-Abisi 2020). Only in cases where both the adsorbate and the solid substrate have the same (usually negative) sign of net charge is there likely to be an energy barrier opposing spontaneous adsorption of aqueous dissolved species that contain substantial hydrophobic groups (Wernert and Denoyel 2016). In the absence of such an energy barrier, rates of adsorption onto cellulose-based surfaces are typically governed by diffusion mechanisms (Douven et al. 2015; Hubbe et al. 2019).

Once a hydrophobic compound has adsorbed from aqueous solution onto a polysaccharide-based substrate, then the energy of adsorption can be instrumental in keeping it there. In general, the likelihood that a monomeric species escapes from the surface will be determined by a summation of adhesion energy contributions, including van der Waals, electrostatic, and entropic terms related to changes in the randomness of the system (Werth and Reinhard 1997; Ghosh et al. 2001; Enell et al. 2005). Because the attachment of a polymer involves the summation of contributions from many connected segments in the chain, spontaneous desorption is much more difficult. A sufficient energy of adsorption also is expected to be critical to the structural integrity of hydrophobic layers applied by adsorption to polysaccharide type surfaces intended for packaging applications.

Self-assembly

In addition to the substrate-adsorbate energies of interaction, the stability of a hydrophobic monomolecular layer at a surface exposed to water can also be strongly affected by tendencies for self-assembly among the adsorbing molecules. Self-assembly can be defined here as the tendency of certain hydrophobic groups to organize themselves into continuous condensed monolayers at various interfaces (Hubbe et al. 2020). As described in an earlier review article, such contributions to formation of hydrophobic layers are especially prominent in the case of compounds having long-chain alkyl groups (Hubbe et al. 2020). The contribution of hydrophobic groups to self-assembly of hydrophobic layers has been proposed in several studies (Renneckar et al. 2006; Aranaz et al. 2010; Li et al. 2011; Aarne et al. 2013; Hu et al. 2017; Wan et al. 2017; Malakhova et al. 2018; Cai et al. 2021). Within such layers it is common for the hydrophobic molecules to be pressed tightly together, i.e. condensed, as a compact monolayer. In principle, the presence of a condensed monolayer of nonpolar groups at a surface tends to resist not only water, but also it tends to be unfavorable for adhesion of other materials to that surface (Garoff and Zauscher 2002; Faucheux et al. 2004).

Macromolecular redundant interactions

A polymer chain can have multiple points of attachment. In addition, a polymer in solution, due to its connected nature, has only a relatively small change in randomness when comparing the polymer in solution to its adsorbed state. As a consequence, the adsorption of polymers onto surfaces tends to be more favored thermodynamically than it would be based merely on the enthalpy of interaction (Fleer et al. 1993). Strong binding to cellulosic surfaces can be expected especially when contributions from entropy are bolstered by the presence of hydrophobic groups. For example, styrenemaleic anhydride (SMA) is well known as a water-dispersible copolymer that is used in combination with size-press starch to make paper hydrophobic (Bildik Dal and Hubbe 2021). As discussed in the cited review article, SMA and similar size-press additives appear to migrate to the polysaccharide surfaces during the process of drying from a water solution.

Durability

The energy-related issues addressed in the previous subsection can be viewed as aspects related to the desired durability of hydrophobic effects to be incorporated into packing systems. A typical single-use container may need to resist continual or intermittent moisture both inside and out. The external moisture might consist, for instance, of condensate during its storage within a refrigerator. There is a need for eco-friendly, low-weight systems that are able to match the durability of such materials as glass, polycarbonate, or aluminum foil, etc. Not only must the hydrophobic effects be achieved initially, but they also need to persist in spite of abrasive action (Milionis et al. 2016).

Lability

At the same time that one needs a food packaging system to be able to endure long-term exposure to water, along with the stresses and abrasion of scuffing and handling, it also needs to break down. Two goals need to be met simultaneously – resisting aqueous solutions during the period of usage, and being susceptible to relatively fast biodegradation if and when the package is discarded to the environment. One important tactic for designing a material that breaks down is to incorporate molecular weak links into it, i.e. labile bonds. The silicon-carbon bond is a prime candidate for such effects. The chemical approaches to forming such bonds are well known (Bei et al. 2012), but they do not occur in nature. Apparently, there may not be a sufficient evolutionary advantage of creating an inherently weak bond as part of a biological process. On the other hand, it could be proposed that natural breakdown of such a bond could constitute a planned decomposition if the item were to end up in a natural environment.

Three-dimensionality

The development of a three-dimensional structure can be used as part of a strategy to achieve sufficiently durable structure, relative to the planned usage of a packaging system, while at the same time being able to incorporate weak or labile bonds, favoring the package’s natural breakdown. This scenario appears to describe certain layered structures formed by reaction of alkoxysilanes (Jankauskaite et al. 2020). This topic will be considered more deeply when discussing that type of surface treatment.

Chemical crosslinking in the course of curing of surface layers can be achieved during chemical vapor deposition, which contributes to durability (Alf et al. 2010). Crosslinking also occurs during the curing of unsaturated oils (Dankovich and Hsieh 2007).

Biodegradability

Most currently used food packaging systems meet or exceed their required barrier properties and durability requirements, but they lack biodegradability. Further discussion of reported findings related to biodegradability appear later in this review article, but some general points can be made at the outset. First, one needs to keep in mind that there is only a weak correlation between “bio-based” and “biodegradable” (Robertson 2014). The molecular modification of any material has the potential to block its susceptibility to biological degradation. Attributes such as hydrophobicity generally tend to make materials less biodegradable, regardless of their origin (Luckachan and Pillai 2006; Yamano et al. 2014). A likely explanation is that enzymes are relatively large molecules, such that they cannot penetrate below the outer surface. Hydrophobic character makes it less likely that a packaging material will swell in water. In the absence of such swelling, the enzymes lack access to the bulk phase. However, some initial breakdown, leading to fracturing or swelling of the surface layers, can provide the needed access to enable them to cleave various bonds. Another key point is that abiotic changes are sometimes required before significant enzymatic biodegradation can begin, as was mentioned in the case of PLA in the Introduction of this article.

Layer Options

Before considering the effects of various specific hydrophobic treatments of surfaces, there are some general points related to layered structures that have general applicability. First, multilayered films already can be regarded as a well-established strategy in food packaging (Stasiek 2005; Lim et al. 2008; Lamnawar and Maazouz 2009; Garofalo et al. 2018; Mizielinska et al. 2020). Second, sometimes the addition of material to a barrier film is able to plug defects, such as holes (Martinpolo et al. 1992). The overall barrier properties often can exceed those of the individual components acting alone (Garofalo et al. 2018). However, the melt flow compatibility and interfacial integrity can be challenging (Lamnawar and Maazouz 2009). The references cited in this paragraph generally relate to the use of non-biodegradable, synthetic polymers. The challenge to be considered in this work is to achieve comparable effects by application of essentially molecular layer treatments in combination with photosynthetically renewable polysaccharides in the form of films or paper-like structures.

REPORTED EXPERIMENTAL FINDINGS AND INTERPRETATION

General Considerations

Having outlined, in the previous section, some key factors related to the creation of barrier layers having the ability to resist water penetration, the next step will be to consider literature dealing with different chemical treatments. As noted earlier, attention in this article is focused on strategies that rely on as little as one molecular layer of a hydrophobic substance to achieve such resistance. The goal of this section is to describe progress that has been achieved in development of surface treatments that can serve as the main barrier to water transport into and through a package or wrapper that is intended to protect food.

Though it is possible to contemplate a single, uniform layer of material meeting all of the barrier needs of a package (for instance in the case of a glass bottle), the focus of this article is on the achievement of satisfactory barrier performance using systems that are all or mostly composed of renewable materials. Existing review reports already have addressed important related topics such as the blockage of oxygen permeation (Paunonen 2013; Hubbe et al. 2017; Wang et al. 2018). While a heavy, monolithic structure such as glass can be effective for denying transport of a wide range of permeants (e.g. water, water vapor, oils, aroma compounds, and flavors), the focus here is what can be done while relying mainly on renewable, organic compounds. In principle, the goal of meeting the water-barrier goals with near-monolayer molecular coverage represents an opportunity to make food packages that are much lighter than glass.

As indicated in earlier review articles, cellulose-based packaging plies that are effective for resisting permeation by oxygen and grease are often harmed by the presence of high relative humidity or liquid water (Ferrer et al. 2017; Hubbe et al. 2017; Hubbe 2021). For this reason, it will be assumed here that an excellent barrier layer against liquid water and that water vapor will be a very important factor in the development of paper-based multi-layer packaging for foods. It is proposed here that the packaging layer that is charged with the responsibility of blocking the transport of water molecules has the potential to preserve some or all of the ability of under-lying layers in a packaging structure to block transport of other important permeants, including flavors, oils, oxygen, and contaminants.

Tabulation of Reported Findings

Factors considered

Due to their length, tables summarizing information obtained from various reviewed articles are placed in the Appendix of this article. Separate tables are provided for different classes of hydrophobic treatments. In each case, the categories considered include the chemical agent used in the treatment (silane, ester, amine, ion pair, adsorbed monomer, adsorbed polymer plasma click chemistry, and other), the medium (e.g. aqueous, ethanol, gas phase), the type of substrate (e.g. cellulose, starch, natural fibers, nanofibrillated cellulose), the water contact angle (WCA), various details (e.g. temperature, time, concentration), and the author-year information.

Order of topics

The order of the subsections that follow is meant to reflect levels of recent research interest, as well as the degree of progress that has been achieved in preparing effective hydrophobicity under industry-friendly processing conditions. Thus, silicone-based treatments are considered first, followed by other covalent reactions with the –OH and amine groups of various polysaccharide-based barrier layers. Such reactions include esterification, amine formation, and click chemistry. Next to be considered are non-reactive strategies, such as the usage of ionic associations as a means to anchor hydrophobic groups. Last to be considered will be the usage of various hydrophobic substances, especially naturally derived materials such as waxes, which can contribute hydrophobicity to the surfaces despite their lack of specific anchoring mechanisms.

Silane-type Treatments

General issues

Silane, the precursor for many of the other compounds to be discussed here, consists of a central silicon atom with four bonds of hydrogen in a tetrahedral configuration similar to methane. Silane compounds can complex with transition metals and may spontaneously combust in the air. Siloxanes, on the other hand, have a central oxygen bonded with two silicon atoms to form a Si-O-Si linkage and may be either straight chained or branched molecules that form the backbone of silicone-based compounds and polymers. The major route to form siloxanes is through the condensation reaction of two silanols. This will be described in greater detail below.

Based on the number of relatively recent scientific articles, it seems that researchers regard silane-type surface treatments as a highly promising field (Witucki 1993; Owen and Williams 1991; Xie et al. 2010). The present literature search identified over 70 articles dealing with hydrophobization of polysaccharide-based surfaces by means of silane-related compounds. A likely reason for the popularity of this research topic may be the variety of chemical species that can be used, as will be discussed in this section. Though various silane treatments can be applied from the gas phase or from organic solvents, there is increasing interest in aqueous-based treatment strategies, which have the potential to minimize environmental impacts. The layers formed by these treatments are rich in Si-O bonds, and there can be a wide range of pendant groups, including hydrophobic alkyl groups.

Siloxanes are useful for surface coatings that induce water repellency through silanization, which occurs when organofunctional alkoxysilane molecules cover the surface of a substrate (Tilley and Fry 2015). It has been widely reported that reactions involving formation of Si-O-cellulose bonds can impart water repellency to the surface of paper, typically by reducing the surface energy of paper and increasing the water surface contact angles between controlled and treated samples. Cappelletto et al. (2012) reported findings from their research on the effect of increasing methyl substitution of silane/siloxane paper coatings and its effect on hydrophobicity using a simple and direct sol-gel application process of single and double layered coatings of different methyl-functionalized alkoxysilanes. Their data suggests that increasing methyl substitution increases the hydrophobicity of paper primarily due to its ability to be evenly coated across the surface of cellulose (Cappelletto et al. 2012).

Some literature sources have referred to siloxanes as glassy, glass-like, or glass-hybrid materials. This is likely due to the idea that strong Si-O-Si linkages correspond to the structure of silica glass and glass-like films formed through the polycondensation of alkoxysilanes that is followed by curing (Iwamiya et al. 2020). While some Si-O-Si linkages have been observed in siloxane modified cellulose paper substrates, usually by FTIR, this will typically only occur when there is three-dimensional condensation of the compounds on the cellulose surface, which is beyond the scope of this review. In the case of siloxane monolayers on cellulose, the conventional sol-gel process mentioned above is too harsh and would degrade untreated cellulose paper before the film can form (Iwamiya et al. 2020). Therefore, alternative methods used to apply a siloxane coating onto paper are often employed, which yield similar, but not identical, results. Often, in the case of cellulose substrates, the siloxane coating being referred to as glassy is based on quantitative thermal analysis and mechanical strength testing; this suggests the formed siloxane matrix significantly increases the paper’s thermo-oxidative stability and brittleness (Cappelletto 2012; Iwamiya et al. 2012).

While siloxanes have been used in commercial products for over a half a century, members of this family of compounds have been characterized in recent literature as emerging contaminants (Fijalkowski et al. 2017; Coralli et al. 2021). This is likely due to their documented persistence in the natural environment, which may lead to accumulation. Siloxanes that bear organic groups beyond methyl could reasonably be expected to display characteristic organic behavior, but there is not such data to support this. Despite the abundance of each element on Earth and carbon’s essential role in supporting living organisms, the bond of Si-C is not known to occur naturally. There is no known biological process or enzyme that naturally catalyzes the bonding or cleavage of Si with C, nor any methyl or organic group (Rucker et al. 2015). The biodegradation of Si-C bonded organosiloxanes occurs too slowly to be observed under standard test conditions, while alkoxysilanes, silyl esters, and D3 and D4 may be hydrolyzed during test time parameters (Rucker et al. 2015). The cited authors report that the water hydrolysis that drives Si-O cleavage, by comparison, may be observed in di-, oligo-, and polysiloxanes, alkoxysilanes, and silyl esters according to chemical reaction Schemes [1] through [3]. It is important to keep in mind that the rate of such a mechanism depends on chemical structure and related physical physicochemical properties and environmental conditions in which the reaction is set.

R₃SiOSiR’₃ + H₂O 🡪 R₃SiOH + HOSiR’₃ [1]

R₃SiOR’ + H₂O 🡪 R₃SiOH + HOR’ [2]

RC(=O)OSiR’₃ + H₂O 🡪 RC(=O)OH + HOSiR’₃ [3]

Schemes 1, 2, and 3. Initial hydrolysis steps of some silane-related compounds

Polydimethylsiloxanes, such as those often used in industry, are not known to hurt the composting process, but they are not considered compostable either, since they do not fully biodegrade within a reasonable period. It has been suggested that the biodegradation of such molecules is highly dependent on environmental moisture content. While polydimethylsiloxanes do not naturally biodegrade quickly in moist compost conditions, they have been shown to degrade by over half their original mass in 4 months after being placed into soil that is allowed to dry periodically (Lehmann et al. 2001; Rucker et al. 2015). With increasing drying of the soil, these compounds were characterized by increased rates of biodegradation, with some samples reaching an end mass 20 to 30% that of the original, though some degradation products were still present in the soil (Lehmann et al. 2001). More research should be conducted on improving the biodegradability of these compounds used in commercial products.

Tri-alkoxysilanes

Siloxane copolymers may be classified depending on their structure and microstructure of various units of siloxane. The alkoxysilane class of compounds has attracted the most research attention for such purposes as hydrophobically modifying polysaccharide-type surfaces. The conditions that have been studied to achieve such effects are summarized in Table A in the Appendix.

The most commonly used forms of alkoxysilanes have three ethoxy or three methoxy groups, as illustrated in Scheme 4. These groups are susceptible to hydrolysis upon exposure to water. As a rule, two steps are incorporated into processes aimed at using alkoxysilanes to impart hydrophobic effects on plant-based materials. In the first step, as illustrated in Scheme 4, the alkoxide groups are hydrolyzed to an intermediate silanol form. In the second step, sometimes involving an adjustment of pH, the polysaccharide-based material is introduced, allowing the interfacial reaction to proceed.

Scheme 4. Trialkoxysilanes and their initial hydrolysis

Scheme 5. Self-condensation of hydrolyzed trialkoxysilanes

Figure 16 illustrates a two-step process by which hydroxysilane species may become bound to a polysaccharide substrate (Xie et al. 2010). As shown, the first step is likely to involve hydrogen bonding. Next, depending on such factors as temperature, time, and loss of water through evaporation, covalent bonding and further condensation of the silanol functions can be expected. Note in the figure that such reactions, which might be called “curing”, involve the loss of water.

Fig. 16. Reaction of hydrolyzed silane compounds with hydroxylated surfaces, with the loss of water

For the first step, by adjusting the pH to an acidic range, the hydrolysis of alkoxide groups can take place while the subsequent condensation among these entities can be suppressed (Bel-Hassen et al. 2008). Table 2 lists some reported pH values that have been selected in reported studies when the goal has been to hydrolyze triethoxysilanes but to discourage their mutual condensation before they have had a chance to react with the polysaccharide surface. Thus, pH values in the range of about 3 to 5 are often chosen to allow formation of relatively stable intermediate silanol compounds, which then have a chance to individually react with the –OH groups on the surfaces of the polysaccharide substrate.

Table 2. Selected pH Conditions for Prehydrolysis of Alkoxysilanes before their Exposure to Polysaccharide Surfaces