NC State
Hubbe, M. A., Maitland, C., Nanjiba, M., Horst, T., Ahn, K., and Potthast, A. (2023). “Archival performance of paper as affected by chemical components: A Review,” BioResources 18(3), 6430-6498.


For about two millennia, paper has served as a main medium for preservation of people’s ideas, stories, contracts, and art.  This article reviews what is known about the various components that make up paper from the perspective of their long-term stability under typical storage conditions.  Literature evidence is considered relative to the susceptibility of different paper components to embrittlement, acid hydrolysis, microbiological attack, and discoloration, among others.  The cellulose that makes up a majority of most paper items is demonstrably stable enough to persist for many hundreds of years on the shelves of archival collections, though it is susceptible to acid-catalyzed hydrolysis, which can be accelerated by byproducts of decomposition.  Though less attention has been paid to the archival performance of various minor components of modern paper products, evidence suggests that at least some of them are subject to likely breakdown, embrittlement, or decay in the course of prolonged storage. Based on these considerations, one can envision different categories of paper that can be expected to meet different levels of storage stability: ancient recipes for handmade papermaking, e.g. washi and hanji, archival-grade paper products, ordinary modern alkaline paper products, and paper manufactured without concerns for its longevity.

Download PDF

Full Article

Archival Performance of Paper as Affected by Chemical Components: A Review

Martin A. Hubbe,a,* Crystal Maitland,b Moumita Nanjiba,a Tali Horst,a Kyujin Ahn,c and Antje Potthast d

For about two millennia, paper has served as a main medium for preservation of people’s ideas, stories, contracts, and art. This article reviews what is known about the various components that make up paper from the perspective of their long-term stability under typical storage conditions. Literature evidence is considered relative to the susceptibility of different paper components to embrittlement, acid hydrolysis, microbiological attack, and discoloration, among others. The cellulose that makes up a majority of most paper items is demonstrably stable enough to persist for many hundreds of years on the shelves of archival collections, though it is susceptible to acid-catalyzed hydrolysis, which can be accelerated by byproducts of decomposition. Though less attention has been paid to the archival performance of various minor components of modern paper products, evidence suggests that at least some of them are subject to likely breakdown, embrittlement, or decay in the course of prolonged storage. Based on these considerations, one can envision different categories of paper that can be expected to meet different levels of storage stability: ancient recipes for handmade papermaking, e.g. washi and hanji, archival-grade paper products, ordinary modern alkaline paper products, and paper manufactured without concerns for its longevity.

DOI: 10.15376/biores.18.3.Hubbe

Keywords: Permanence of paper; Aging; Decomposition; Hydrolysis; Storage stability; Conservation; Paper properties

Contact information: a: North Carolina State University, Department of Forest Biomaterials, Campus Box 8005, Raleigh, NC 27695-8005 USA; b: Canadian Conservation Institute, 1030 Innes RD, Ottawa, Ontario K1B 4S7; c: National Archives of Korea (NAK), Conservation and Restoration Division, 30 Daewangpangyo-ro 851beon-gil, Sujeong-gu, Seopngnam-si, Gyeonggi-do, Korea (13449); d: University of Natural Resources and Life Sciences, Vienna, (BOKU), Department of Chemistry, Institute for Chemistry of Renewables, Muthgasse 18, A-1190 Wien or Konrad-Lorenz-Straße 24, A-3430 Tulln;

* Corresponding author:


The premise of this article is that there must be a relationship between the components that are included in a paper specimen and the overall archival performance of that paper material. If one accepts this premise as valid, then it makes sense to consider various possible ingredients of paper in turn, asking what is known about their archival performance. Thus, the present review article has attempted to find out what has been reported regarding a wide range of substances that have been used in the production of paper items. Concerns about the durability of paper items are not new (Johnson 1891; MacAlister 1898). Kantrowitz et al. (1940) list approximately 300 annotated articles on the subject of permanence and durability of paper from 1885 to 1939. The two terms permanence and durability have been used variously; often they specify the retention of properties such as strength and color during aging (permanence) as opposed to retention of the physical/mechanical properties during use (durability) (McCrady 1989). The concerns of these writers were not small. For instance, Johnson (1891) regarded inferior paper as a threat to the permanency of literature. Already in 1824 it was suggested that the roots of such problems may lie in the composition of the paper (Murray 1824).


In the context of the present article, the term “paper” will generally refer to typical old or modern grades of printing and writing paper, including xerographic copy paper. In the case of modern printing paper, the components will include cellulosic fibers, mineral particles (most likely calcium carbonate), starch products, and a variety of minor substituents. The focus will skew towards considerations of modern paper (i.e. a focus on wood pulp, kraft pulp preparation, and modern fillers, etc.), though historic materials and practices will also be considered. This modern focus is because there is obviously less natural aging data available for more recent materials, and, modern papers have increasing quantities of materials other than cellulosic fibers as fractions of their total composition. In principle, a high enough level of degradation in any one of these components of paper – especially the major components – would weaken the structure of the paper as a whole.

Table 1 lists review articles that deal with the storage stability of paper from various perspectives, along with some text indicating their main focus. As shown, many of the articles have focused on (a) mechanisms that can account for the degradation of paper’s properties over time as it is being stored, and (b) strategies for slowing down or even reversing the decline in paper properties. Notably, none of the listed articles has as its main focus the archival quality of different components of typical paper.

Most manufactured paper is not intended for long-term storage. Libraries and museums store only a minor fraction of what has been produced. In exceptional cases, where a paper document or artwork is intended to be permanent, there are standards that can be followed with respect to its composition (ANSI 1984; Anon 1993). Further discussion of what constitutes permanent paper can be found in the ISO 20494 and ISO 9706 standards, and in the textbook by Allscher and Haberditzl (2019). Paper products designated as “permanent” will contain a certain amount (e.g. 2%) of an alkaline reserve component such as calcium carbonate. The Canadian standard for permanent paper can be regarded as unique due to the fact that it states mechanical and optical aspects of permanence separately (Canadian General Standards Board 2016). The content of lignin in paper is capped by the standard only in cases where the optical properties are of concern. The preparation and properties of permanent or archival papers have been considered in other articles (Clapp 1977; Shahani and McComb 1987; Bird 1999; Basta et al. 2006). The present article comes at the topic from a different direction, considering a wide variety of components that are often present in paper products that can be used for printing and writing.

Table 1. Review Articles and Chapters Dealing with Aspects of the Archival Performance of Paper

Components of Cellulosic Plant Materials

Cellulosic fibers often comprise 70 to 90% of the dry mass of printing paper, a fact that can justify a major focus on what is known about their archival characteristics. This subsection will consider various chemical substances – both major and minor – that make up typical papermaking fibers. A following subsection will consider issues from the perspective of the fiber types derived from different pulping and bleaching methods.


A variety of fibers have been used in papermaking throughout time depending on factors of geography and availability. Currently, however, the majority of machine-made paper is made using wood-derived fibers (Hunter 1947; Collings and Milner 1990; Baker 2010). Plant-derived cellulose can be isolated from wood by conventional pulping and bleaching methods (Dence and Reeve 1996; Sixta 2006; Fardim and Tikka 2011). Bleached kraft fibers, which are widely used in high-quality books, have been processed under conditions that maintain a substantial proportion of hemicellulose, in addition to the cellulose component of the wood. However, the conditions of pulping and bleaching can be adjusted to yield relatively pure cellulose (Chen et al. 2016). Such cellulose can be briefly described as a linear polymer of anhydroglucose units (French 2017) in which the units are joined by β-glycosidic bonds. The detailed structure of cellulose (see Fig. 1), having three –OH groups per anhydroglucose unit, is conducive to the development of fibrillar structures, both at the nano-scale and in the macroscopic material. Cellulose from wood or cotton has crystallinity in the range of about 50 to 90, depending on a variety of factors, including the method of assessment (Ahvenainen et al. 2016; French 2022).

Fig. 1. Features of parts of a cellulose molecular chain having different susceptibility to steps in decomposition (French 2017)

Dry, pure cellulose is relatively stable under ambient conditions. For example, thermogravimetric analysis studies have shown that pure cellulose has a rather narrow range of mass loss at about 295 to 380 °C (Ramiah 1970). The fact that cellulose-based paper items sometimes can last for more than 1000 years (Hunter 1947; Jeong et al. 2014b) can be taken as further evidence of the stability of the cellulose itself. Jeong et al. (2014b) attributed the exceptionally high storage stability of traditional Hanji paper to an initial high molecular mass of the cellulose and a low rate of chain scissions compared to other paper specimens that were tested. Even more remarkable storage stability has been documented in the case of papyrus samples that had been prepared during Egyptian dynasties. Such specimens, which have cellulose as a main component, have lasted for upwards of 3000 years (Lojewska et al. 2017; Bausch et al. 2022a,b). Lojewska et al. (2017), however, found evidence of substantial degradation of the cellulose in the ancient papyrus. Gel permeation chromatograms show a mean degree of polymerization (DP) of about 2100 in the cellulose obtained from fresh papyrus plants, whereas the ancient material had cellulose with a DP of about 800. Potthast et al. (unpublished) measured a DPw of 2000 +/- 400 for historic samples aged 1100 to 2300 years.

A factor that can help account for the relatively high storage stability of cellulose is its notably high density of hydrogen bonding, which occurs within (intra-) and between (inter-) the molecular chains (Wohlert et al. 2021). Such bonds, though they are readily reversed in the presence of water, can be expected to constrain the motions of molecular segments, thus making the material less susceptible to chemical degradation than it otherwise might be. The relatively high density and highly regular arrangement in the crystalline domains renders those parts of the cellulose structure especially resistant to various enzymatic or chemical attacks (Meng and Ragauskas 2014; Jerome et al. 2016). In addition, as shown by Jeong et al. (2014b), the molecular mass may sometimes be as high as 1.4 million g/mol.


Both the content and the chemical composition of hemicellulose in wood and other plant fibers depend on the species of origin. Though hemicellulose has many similarities to cellulose, it differs in certain respects that can be expected to render it less storage stable. For one thing, the molecular mass is lower, e.g. having a weight average of 10,000 to 40,000 g/mole (Bai et al. 2012; Krawczyk et al. 2013). Figure 2 shows typical hemicellulose structures often present in hardwood xylem (4-O-methyl-D-glucurono-D-xylan) and softwood (galactoglucomannan) (Ebringerova et al. 2005; Mäki-Arvela et al. 2011). In order to maintain a relatively high yield of chemical pulping processes, paper manufacturers have a strong incentive to adjust pulping and bleaching conditions to maintain a relatively high content of the original hemicellulose in most paper grades, including printing papers. In addition, hemicellulose contributes to inter-fiber bonding. The lower degree of polymerization and the lack of crystallization of hemicellulose may be essential so that it can fulfill its task in helping to bind adjacent cellulosic and lignin-related moieties together (Molin and Teder 2002). Also, it has been found that the presence of hemicellulose tends to make paper items more resistant to hornification, which can be defined as a loss in ability to swell again in water after having been dried (Wan et al. 2010).

Fig. 2. Structures of some common hemicellulose types (Figures adapted and redrawn from Ebringerova et al. (2005) and Mäki-Arvela et al. (2011))

The non-crystalline nature of hemicellulose is known to make it vulnerable to swelling in water (Pere et al. 2019). The swollen conditions can make cellulosic materials susceptible to such decomposition mechanisms as acid hydrolysis and enzymatic attack (Zhang et al. 2006; Kumar et al. 2012). The carboxylic acid groups present on certain hemicellulose molecules (see the first structure in Fig. 2, for instance), can contribute acidity to the material, and therefore to its deterioration (Feller et al. 2002). Carboxyls are also responsible for catalyzing the formation of chromophoric groups (Ahn et al. 2019). In addition, hydrolysis of acetate ester groups (as in the second example shown in Fig. 2) can release acetic acid in the course of storage (Potthast et al. 2022). This mechanism is further shown in Fig 3. In addition, acetic acid is also formed from carbohydrate degradation products in the course of natural aging. Hemicellulose also may be responsible for thermal and photochemical discoloration of paper (Lee and Feller 1986). Under more severe conditions of heating, aqueous suspensions of cellulose have been found to decrease in pH, which can be attributed to the same reaction; such acidity contributes to autohydrolysis during certain treatments of biomass for the production of monomeric sugars and various biofuels (Garrote et al. 2001; Nabarlatz et al. 2004; Liu et al. 2015a; Surek and Buyukkileci 2017).

Questions remain about the extent to which these mechanisms are the same or different under typical dry storage conditions of paper. Though no liquid water is present, typical paper will contain about 4 to 8% water under a wide range of storage conditions. Kramer et al. (2015) recommended using an HVAC system to maintain relatively low values of relative humidity in the range of about 35% to 55%, with seasonable adjustments to avoid very large contrasts between the storage conditions and ambient conditions.


Fig. 3. Acetyl groups in a common hardwood hemicellulose, which can be the source of acetic acid as a result of acid-catalyzed hydrolysis. The acetylglucuronoxylan structure is based on a drawing by Biely et al. (2014).


When attempting to make high quality paper for printing and writing applications, papermakers through the ages have generally preferred to use fibers that have little or no lignin content.

Fig. 4. Some common structures within lignin macromolecules (Adapted from Balakshin et al. 2020)

A major reason for that preference is that upon light exposure lignin can be subject to color changes during storage (Wong et al. 1995; Bird 1999; Paulsson and Ragauskas 2000; Smith 2012). Lignin also is much more susceptible to oxidation reactions in comparison to cellulose (Malachowska et al. 2020). If paper with high brightness stability is needed, the presence of lignin should be avoided. Color change can be deemed acceptable for archival permanence of certain types of (textual) records where it is the information that is important, as long as some contrast between printing/writing and paper surface can be maintained. Optical permanence / brightness stability obviously plays a significant role for other types of cultural records, however (images, art, etc.). The situation can become muddy if a record that was intended to be merely useful as an information record turns out to be so significant that someone would like to put it on exhibition. Figure 4 shows common structures and linkages that are present in lignin (Balakshin et al. 2020).

An adverse effect of lignin on the storage-stability of paper has been suspected by many authors (Clements 1987; Arnold 1998; Łojewski et al. 2010). Bogolitsyna (2011) developed methodology to detect lignin-derived byproducts of the type that may result from the aging of lignocellulosic materials. Dupont et al. (2007) observed such aromatic degradation products from lignin in aged papers. However, Zou et al. (1993) found that the presence of as much as 28% of lignin in paper fibers did not have a significant accelerating effect on the loss of strength of paper subjected to accelerated aging. Schmidt et al. (1995) proposed a sacrificial antioxidant role for lignin after comparing fully-bleached and lignin-containing pulps. Lignin-containing pulps exhibited poor optical permanence, but better retention of paper strength and cellulose DP than their bleached counterparts. One has to keep in mind that lignin cannot be treated as a material with a uniform chemical structure. Thus, the effects on aging highly depend on its origin and the processes involved during pulping and papermaking. One example for a different behavior based on lignin origin during light-induced aging is given by Potthast et al. (2004).

Bégin et al. (1998) observed that as long as there was sufficient calcium carbonate present, lignin content did not appear to affect the aging behavior of either lab-made handsheets or commercial paper products. They attributed such effects to the buffering of pH to about 7.5 or more. All of the papers tested exhibited substantial loss in tear strength and folding endurance when the extract pH was below 5, but strength was maintained in the case of all sheets having extract pH of neutral or higher. Malachowska et al. (2020) likewise found no evidence of faster accelerated aging of paper sheets that contained lignin in the case of specimens prepared at neutral pH.

Clements (1987) mentioned the degradation of lignin as a likely contributing cause of acidification, leading to acid-catalyzed hydrolysis of cellulose. It is worth noting that the presence of lignin is often associated with mechanical pulp grades, and these can be expected to have greater proportions of both hemicellulose and various extractives in comparison to kraft pulps. The acid content of the extractives can lower the extract pH, unless suitably buffered (Verenich et al. 2004; Leiviska and Ramo 2008). Hydrolysis of acetyl groups of the hemicellulose is another source of acidity. If “lignin” is being blamed for the poor storage stability of papers that contain higher amounts of extractives and/or hemicellulose, this can be regarded as a kind of guilt-by-association rather than a true effect of the lignin.

Further evidence that lignin does not necessarily contribute to rapid breakdown of paper-like items is apparent from the fact that papyrus specimens, which contain lignin, have lasted for more than 3000 years (Lojewska et al. 2017). Natural papyrus is known to contain about 12 to 15% lignin (Bausch et al. 2022b). However, these authors suggest that a further contribution to the stability of papyrus may come from its content of natural flavonoids. These compounds, which can act as lignin precursors and not as just extractives, have the potential to serve as internal antioxidants (Rosado et al. 2021; Rencoret et al. 2022).

Wood extractives

Figure 5 shows the structures of some common wood extractives, which together often make up from 1% to 3% of the dry mass of the wood species most used as the source for papermaking fibers. Among the chemicals related to wood extractives, by far the most research related to the storage stability of paper has been focused on rosin, which is an extractive component present in the wood of pine and other conifers. Many studies have reported relatively poor storage stability of paper that had been manufactured with rosin as a sizing agent (Wilson and Parks 1980; Clements 1987; Shahani et al. 1989; El-Saied et al. 1998; Basta 2003, 2004; Basta and Fadl 2003; Basta et al. 2006; Jacob et al. 2017). However, essentially all of the studied rosin-containing papers mentioned in the cited articles had been prepared with papermaker’s alum (aluminum sulfate), which is highly acidic. Barrett et al. (2016) showed a strong adverse correlation between the alum content of European handmade papers in the period between the 14th and 19th centuries and their permanence characteristics. It is notable that alum was also historically added for purposes other than alum-rosin sizing (Brückle 1993).

Fig. 5. Chemical structures of some common wood extractives

Bialczak et al. (2011) carried out accelerated aging tests on newsprint specimens and focused on the changes in extractives and wetting properties. It was found that the content of extractable material declined very quickly during heating at 60 °C. After two days of such accelerated aging, only 20% of the initial extractives were still capable of being extracted. Results were attributed, at least in part, to polymerization of some of the extractive substances. Evidence also pointed to the oxidation of extractive substances as a likely mechanism leading to lower extractive amounts. It makes logical sense to expect that some portion of the extractives may have been directly volatilized from the paper during the accelerated aging. Another key observation was an increased hydrophobic character of the paper following accelerated aging (Bialczak et al. 2011). This is in agreement with earlier research by Swanson and Cordingly (1959). These authors were concerned with mechanisms leading to the self-sizing of paper sheets that contained wood extractives. In other words, the migration of wood extractives to the surface of paper can be a mechanism leading to increased hydrophobicity over time.

Some unique evidence about the chemical stability of certain wood extractives was obtained under more extreme conditions in a study of palm oil fatty acid ester and natural ester as a component for the dielectric fluid within electrical transformers (Tokunaga et al. 2016). The synthetic ester was found to have favorable stability in comparison to mineral oil after heating at 170 °C for 20 to 120 days. Likewise, Suzuki et al. (2014) found that kraft paper degraded more slowly in a palm fatty acid ester in comparison to when mineral oil was the suspending medium.

Effects of Pulping and Bleaching on Archival Performance

The preceding subsection dealing with the archival performance of cellulose, hemicellulose, lignin, and extractives tells only part of the story. What is missed by such a focus on the individual main chemical components of papermaking fibers is the potentially large effects of such processes as pulping and bleaching. Those aspects will be considered here.

Alkaline pulping and fiber quality

The use of alkaline conditions as a means to obtain long cellulosic fibers from plant materials with a minimum of mechanical damage has been known since ancient times (Hunter 1947; Barrett 2005; Hubbe and Bowden 2009). The lye obtained from wood ash was found to be beneficial in separating the fibers. Such alkaline conditions for the pulping of fibers such as hemp and the bast fibers from the bark of paper mulberry (Broussonetia papyrifera) appear to have resulted from the persistent efforts of local artisans to improve their products. The modern alkaline pulping methods, including kraft pulping and soda pulping, thus appear to be rooted in the early cultural history of the technology. Baker (2010) documents the shifts in the western paper-making technology as papermaking was being industrialized in the 19th century. Mizumura et al. (2017) show correlations between the amounts of alkali used for preparing Japanese washi and the resulting paper qualities. While caustic soda (sodium hydroxide) permits shorter cooking and fiber cleaning times, it also results in shorter fibers and lower quality papers than soda ash (sodium carbonate), and slaked lime (calcium hydroxide). Wood ash (lye/potash) is considered to produce the softest fiber, but it requires the most time and labor during the papermaking period.

Kraft pulp archival performance

Because the kraft process has become the dominant modern pulping method for preparing pulp fibers for paper (exceeding sulfite processes in the 1950s (NAA 2022)), it is worth considering its likely effects on the archival performance of the resulting paper. Kraft pulping differs from the traditional alkaline pretreatment of pulp fibers with respect to such factors as concentration of the alkali, temperature, pressure, and the presence of sodium sulfide as a co-reagent (Fardim and Tikka 2011). For example, the following conditions may be regarded as typical: active alkali 17% to 28%, sulfidity 15% to 28%, and temperature 150 to 170 °C (Mendonça et al. 2002; Daniel et al. 2003). One of the known consequences of kraft pulping is a decrease in degree of polymerization (DP) of the cellulose. For instance, it was reported that the intrinsic viscosity (a measure of DP) of cellulose solutions decreased from an initial value of about 1300 cm3/g after kraft pulping to about 950 cm3/g, depending on the extent of delignification (Tao et al. 2011). More severe losses in intrinsic viscosity, at comparable levels of delignification, were observed in cases where kraft pulping had been followed by oxygen delignification. Alkaline conditions are able to catalyze a so-called peeling reaction, starting from the ends of cellulose chains, in addition to random scissions elsewhere in the chains (Fengel and Wegener 1989). The latter, which requires higher temperature, is more of a concern, since they can cause a more drastic drop in DP.

The extent of the peeling reaction is limited due to an alternative pathway, as illustrated in Fig. 6. Note that in this so-called “stopping reaction,” the end product is a stable carboxylic acid group at the end of the chain. Once this end group has been established, the peeling reaction no longer proceeds.

Fig. 6. Mechanism of the peeling reaction. BAR= Benzylic Acid Rearrangement (Fengel and Wegener 1989)

An important consequence of chemical pulping of wood is that the micro-domains once occupied by lignin are converted to mesopores, i.e. roughly corresponding to the IUPAC defined range of 2 to 50 nm in diameter (Stone and Scallan 1965, 1968). While the pulp is kept in the wet state, the presence of these mesopores contributes to the conformability of the fibers, especially after the pulp has been mechanically refined (Steadman and Luner 1985). This renders the fibers highly conformable and capable of developing relative bonded area between adjacent fibers upon drying of the paper. But upon drying, many of the pores close up as a result of capillary forces, and a high density of hydrogen bonding can develop between what used to be the adjacent walls of mesopores (Hubbe et al. 2007). As a result, the material loses some of its ability to swell again if the paper is placed back in water, an effect that is called hornification (Weise and Paulapuro 1999; Welf et al. 2005; Law et al. 2006). This is relevant to archival concerns, since hornification contributes to a more brittle nature (Kato and Cameron 1999), as does the aging of paper. Implications of hornification will be further discussed in a later section, “Degraded Conditions of Cellulose in Paper”

Bleaching of kraft pulps

In a broad sense, bleaching treatments of papermaking pulp can be regarded as a continuation of pulping. However, whereas chemical pulping (such as the kraft process) has a primary aim of removing lignin, bleaching processes have an additional goal of either removing or decolorizing chromophoric groups (Engwall et al. 1997; Suess 2009). To make matters somewhat more challenging, the residual lignin remaining in the fibers after pulping may be inherently more difficult to break down and remove, in comparison to the lignin already removed (del Rio et al. 2001). An ideal bleaching treatment ought to be selective – attacking and solubilizing lignin-related and other chromophoric species with a minimum of damage to cellulose or hemicellulose. However, it is typical for the DP of cellulose after bleaching to be substantially lower than it was at the end of the pulping stage (Pouyet et al. 2013).

The cleavage of bonds within the cellulose macromolecule are of greatest concern when they occur close to the middle of the chain, thereby having a large effect on its degree of polymerization. Oxidative processes that are aimed at destruction of residual lignin moieties and chromophores to some extent also affect the cellulose and hemicellulose (Gierer 1997). Ney (1982) proposed a hydrogen abstraction from an anhydroglucose unit as an initial step in the oxidative degradation of cellulose during bleaching of pulp with hydrogen peroxide under alkaline conditions. This was also refined by Gratzl (1992), Zeronian and Inglesby (1995), and Gierer (1997) (see Fig. 7). Hence, the action of reactive oxygen species (hydroxyl radicals, dioxygen, or superoxide radicals among others) in chlorine-free bleaching sequences leads to keto groups along the cellulose chain (Fig. 8). Especially, but not exclusively, under alkaline conditions the glycosidic bond in beta-position to this keto group can easily be cleaved, which results in an immediate drop in DP, depending on the overall position of this keto group (Fig. 9). Based on the model reactions, beta elimination becomes significant above pH 9 (Potthast et al. 2006; Hosoya et al. 2018).

Fig. 7. Mechanisms by which oxidative bleaching may cleave cellulosic chains, thus reducing the degree of polymerization of cellulose in bleached pulps. Structures redrawn

Fig. 8. Formation of carbonyl structures as a consequence of hydrogen abstraction by hydroxyl radicals

Fig. 9. Cleavage of the glycosidic bond by β-elimination at oxidized structures along the cellulose chain, according to Lewin (1997)

Also ozone bleaching results in the formation of additional keto groups along the cellulose chain (Potthast et al. 2003), which results in less stable pulps also with regard to brightness. A study by Lemeune et al. (2004) found that oxidation by ozone led to much greater decreases in the degree of polymerization of cellulose, compared to ClO2. Hypochlorous acid is an example of a bleaching agent that is widely known to be less selective than ClO2 and therefore generally more damaging to the cellulose. In particular, hypochlorous acid has been shown to bring about oxidative damage to the pulp (Zhou et al. 2008; Chenna et al. 2013). For details on the effect of different bleaching agents on pulp the reader is referred to the Handbook of Pulp by Sixta (2006).

In modern pulp and paper mills there has been a gradual evolution in bleaching technologies, which generally involve a series of bleaching stages in succession (Dence and Reeve 1996; Hart 2012). For example, a common bleaching sequence in 1970 would include chlorine gas in the first stage, followed by an alkaline (NaOH) extraction stage, and then a couple of stages with sodium hypochlorite as the bleaching agent. However, such practices were shown to generate highly toxic and persistent chlorinated aromatic species in the effluent from such mills (Engwall et al. 1997; Axegård 2019). By the 1990s, the chlorine treatment had been replaced mainly by bleaching sequences employing chlorine dioxide (Gellerstedt et al. 1995). In a well-run bleaching plant, when relying upon chlorine dioxide, it is possible to reduce the production of chlorinated dioxins and furans to below detectable levels (Axegård 2019). Another shift has been the implementation of an initial oxygen gas bleaching stage in many mills (Tao et al. 2011; Hart 2012). Oxygen bleaching is less selective than, for instance, chlorine dioxide as a delignification agent. A consequence of this lower selectivity is a greater tendency to decrease the DP of the cellulose (Pouyet et al. 2013; Brännvall and Walter 2020). A few mills, mainly in the Nordic countries, have been running without any chlorine-containing bleaching agents. Such totally chlorine free (TCF) bleaching systems rely heavily on oxygen delignification as an initial stage of bleaching.

Carboxylic acid groups

For carboxyl group determination of pulp and paper specimens, titration approaches are available to measure the overall content (Barbosa et al. 2013, TAPPI Method T237). Also, adsorption of methylene blue is frequently applied (Weber 1955; Philipp et al. 1965; Fardim and Holmbom 2003). In some cases a differentiation between C6-carboxyl groups, e.g. mainly present in 4-O-methylglucuronic uronic acids of hemicellulose (cf. Fig. 3) and C1 carboxyls, i.e. an oxidized reducing end group, is desirable. For the first case, the FDAM method can be used. Besides the total content of uronic acids, it also provides their distribution along the molar mass distribution of cellulose and can be used to trace hemicelluloses in a pulp or paper specimen (Bohrn et al. 2006). To quantify the C1-Oxidation at cellulose is more difficult. Currently, one approach based on selective fluorescence labeling is being developed (Budischowsky et al. 2022).

Relevant data connecting different bleaching methods to the titratable charge demand of cellulosic pulps was reported by Laine (1997). Various oxidizing agents were found to affect the bleaching of kraft pulps, including sequences containing oxygen, ozone, chlorine dioxide, and hydrogen peroxide. A final hydrogen peroxide stage has a greater tendency to convert carbonyls to carboxyl groups, which adds to brightness stability of pulps.

Letnar (2002) considered the question of bleaching or not bleaching relative to the permanence and archival durability of library materials on paper. Pulp was prepared by the kraft and the sulfite processes, such that the kraft fibers were brown in color. Alkaline pH conditions were maintained, due to the presence of calcium carbonate filler. In terms of mechanical properties, good archival performance was observed regardless of whether the pulps had been bleached or not. On the other hand, the unbleached kraft fibers did not have suitable optical properties for graphic papers.

Carbonyl groups

Carbonyl groups in cellulose other than at the reducing end represent a deviation from the perfect cellulose structure and may generate labile spots that affect stability. Although the carbonyl content typically is in the lower µmol/g range, i.e. 12 µmol/g, which corresponds to a ratio of approximately one extra carbonyl per 1000 anhydroglucose units, their presence matters with regard to brightness reversion and overall stability, as mentioned before. Hence quantifying carbonyls is a way to judge the effectiveness of a stabilizing treatment as well as the status of a historic paper. The degree of yellowing of a paper may not always give conclusive answers.

If sufficient amounts of material are available, the hydroxylamine method can be performed (Cyrot 1957; Szabolcs 1961; Rehder et al. 1965). For smaller sample quantities and for localizing carbonyls in relation to the DP of cellulose, the CCOA method can be applied (Röhrling et al. 2002, 2003; Potthast et al. 2003). If only aldehyde groups are of interest, mainly in form of reducing end groups, the copper number (TAPPI T-430 om94) or the TTC method (Szabolcs 1961; Horn and Eijsink 2004) can be applied. It has to be kept in mind that both methods operate at strong alkaline conditions; hence, the presence of carbonyls along the cellulose chain may induce chain degradation and new reducing ends.

Mechanical pulp archival performance

So-called high-yield pulps, which typically contain over 90% of the original dry content of the wood, have long come under suspicion as sources of instability of paper (Wong et al. 1995; Paulsson and Ragauskas 2000; Seki et al. 2005; Bialczak et al. 2011; Jablonsky et al. 2012). Lignin-containing fibers are known to be susceptible to yellowing, especially when exposed to light (Leary 1994; Yuan et al. 2013; Yun and He 2017). As noted by Zervos (2010), the two primary mechanisms of chemical degradation of paper made from high-yield pulps are the same as those for lignin-free paper: acid hydrolysis and oxidation by air. The decomposition of paper made from high-yield pulp during accelerated aging has been tracked by analysis of the volatilized organic compounds (Hrivnak et al. 2009b). It has been found that the usage of high-yield pulp, in the presence of calcium carbonate filler particles is able to resist aging (Eggle et al. 1984). Likewise, deacidification by spraying aqueous suspensions of magnesium carbonate, an alkaline buffering agent, was found to slow down the aging of such papers. However, paper made from high-yield fibers has been found to be susceptible to yellowing, even after deacidification (Bukovský 1997).

Regenerated cellulose

Though regenerated cellulose, e.g. rayon or lyocell fibers, is not ordinarily used for manufacture of paper, its permanence characteristics are likely to interest some readers of this article. In addition, regenerated cellulose is sometimes used indirectly. For instance, rayon fabric has been used as the structural part of a facing to protect fragile archival paper specimens during their repair (Bedenikovic et al. 2018a,b). Rayon also has been used for decorative Japanese papers (Japanese Paper Place 2022). Due to the fact that regenerated cellulose has a different crystal structure (cellulose II) compared to native cellulose (cellulose I), it is reasonable to expect there to be differences in such issues as archival performance. One has to keep in mind that different regenerated celluloses are on the market today. Their properties differ in crystallinity, DP, and fiber characteristics. The degree of crystallinity of classical viscose fibers has been reported as lower in comparison to the starting cellulose I materials (Zhang et al. 2001; Röder et al. 2006a,b; Gao et al. 2011; Azubuike et al. 2012; Liu et al. 2015b). The overall DP of man-made cellulosic fibers is much lower compared to paper pulp, but the high orientation of the fibers gained in the spinning process generates a different type of stability, also due to extra hydrogen bonding, which cannot be directly compared to ordinary pulp. Man-made cellulosic fibers contain much less hemicelluloses compared to paper pulp. Little is known about the aging behavior of man-made celluloses fibers in direct comparison to pulp or paper. This suggests a need for more study before any version of regenerated cellulose would be recommended for archival paper preparation.

Mineral Fillers

Mineral fillers and coatings have been used throughout the history of papermaking to modify the surface properties and opacity, as well as to prepare paper for printing, drawing, and writing media. For example, “prepared” paper surfaces with lead white are described in Cennini’s 14th century treatise; in the era of old master metalpoint drawings, a variety of minerals were used to fill/coat/create a smooth drawing surface. The reason for placing mineral fillers second in order of discussion after cellulosic fibers is that a majority of modern printing papers contain about 10% to 25% of mineral content, and sometimes even more (Fairchild 1992; Hubbe 2014a). Mineral particles, especially at higher proportional contents, have a negative impact on the strength properties of paper, but on the other hand they often cost a lot less than the fibers they replace. The strength issue is important, since the aging of paper will result in a further loss of strength. At some point the strength will be insufficient to allow handling of an item without high risk of damage. Though papermakers employ starch and other chemical additives to gain back some of the sacrificed strength (Hubbe 2014a), due to the usage of relatively large amounts of mineral, such practices may just shift the concern to the archival performance of those other additives, and these issues will be considered in subsequent subsections of the article.

An expected debonding and weakening effect of mineral fillers on paper is illustrated in Fig. 10. Item “A” in the figure depicts a typical shape of kaolin clay particles that have been used for many years by papermakers. Item “B” portrays a typical shape of a scalenohedral precipitated calcium carbonate (PCC) particle, which is often called a rosette. Frames “C” and “D” of the figure illustrate an observed “bulking” effect of the rosette particles, in comparison to the clay particles. In either case, the imposition of mineral particles between cellulose fibers impedes the direct development of hydrogen bonding between the cellulosic surfaces.

Fig. 10. Schematic depicting likely role of mineral filler particles in preventing close approach of some adjacent fiber surfaces during the formation and drying of paper, thus contributing to lower paper strength

Calcium carbonate

By far the dominant type of mineral used in modern paper is calcium carbonate, and the majority of it that is used in printing paper is PCC (Fairchild 1992; Hubbe and Gill 2016). Commercial usage of calcium carbonate as a filler for paper has been documented back to 1925 (NAA 2022). When a sufficient amount of CaCO3 is placed into aqueous media, it can effectively prevent the pH from remaining below about 7.5. In the presence of acidity, one of the following reactions may take place:

CaCO3 + H3O+ 🡪 Ca2+ + HCO3 + H2O (1)

CaCO3 + 2H3O+ 🡪 Ca2+ + CO2↑ + 3H2O (2)

In addition to consuming the acid, the second reaction can yield carbon dioxide, which will either remain in solution or come out of solution as bubbles. Because the usage of calcium carbonate maintains a non-acidic pH, the use of calcium carbonate in papermaking has been called alkaline papermaking. The usage of calcium carbonate as a filler increased in the 1970s and 1980s, paired with the increasing popularity of alkaline sizing agents (NAA 2022).

Barrow (1974) tracked the relative frequency of paper specimens produced with alum (acidic) or calcium carbonate (alkaline) over the period from 1500 to 1974 and showed a steep rise in the use of alum from 1500 to 1700, with a continuing more gradual rise up to 1974. Barrow attributes a corresponding decrease in pH to this increase in alum. The most recent sets of paper specimens tested by Barrow were over 95% made under acidic conditions. Barrow also showed a strong correlation between acidity and time of production, such that the acidic papermaking conditions seemed to be implicated in the loss of folding endurance. Barrett et al. (2016), however, counter this narrative using XRF data to survey nearly 1600 paper samples from the 14th to 20th centuries, showing relatively consistent alum concentrations, but tracking significant decreases in gelatin and calcium that would correlate in the same manner to the decreasing pH. Later, starting in the mid-1980s and continuing through the 1990s, there was a major shift in commercial papermaking practices, such that calcium carbonate displaced clay as the main filler used in printing papers (Hubbe 2005). Because the rosin-alum sizing system does not work effectively under the weakly alkaline pH conditions, papermakers had to turn to other internal sizing agents that do not rely upon the use of alum (see later discussion of ASA and AKD).

Some of the chemistry associated with the production of PCC versions of CaCO3 have the potential to affect the resulting pH, which is a variable that can affect paper permanence in various ways. To minimize the costs of shipping, it is a common practice to convert mined limestone to burnt lime, CaO, at the mine site by strong heating of the CaCO3. Then, after transporting the lime to a satellite facility, usually right beside a paper mill, the following reactions are carried out:

CaO + H2O 🡪 Ca(OH)2 (slaking) (3)

Ca(OH)2 + CO2 🡪 CaCO3 + H2O (carbonation) (4)

The final reaction (the carbonation) is not necessarily carried out to 100% completion. If some of the Ca(OH)2, which is called “milk of lime,” is still present within the resulting PCC, then the paper machine system’s pH may rise to 9 or higher. Though such high pH values are not common in paper mills, various authors have expressed concern that they may later accelerate oxidation of the cellulosic material in the dry paper (Kolar and Novak 1996; Malesič et al. 2002).

Titanium dioxide

Titanium oxide can be considered for usage as a filler (or “pigment) in paper products that need to have relatively high opacity, e.g. thin paper sheets for religious texts. The high refractive index values of the rutile (2.7) and anatase (2.55) crystal forms of TiO2 give rise to a high efficiency of light scattering, especially when the particle diameter is optimized at about 0.2 to 0.3 μm (Thiele and French 1998).

The catalytic properties of TiO2 may raise concern relative to paper permanence. According to Perez et al. (1998), TiO2 undergoes the following reaction in the presence of oxygen and UV light,

TiO2 + e + UV 🡪 TiO2* (5)

TiO2* + O2 🡪 TiO2 + O2⋅ (6)

where TiO2* denotes an activated species of titanium dioxide. Photocatalysts have been prepared by adding TiO2 nanoparticles to clay particle surfaces (Liu and Zhang 2014). Qin et al. (2021) recently prepared TiO2 nanoparticles having junctions between rutile and anatase phases. The heterojunctions between the crystal phases were credited with a high efficiency of creating oxygen radical species upon exposure to UV-containing solar light. In the cited work, this was judged to be a positive outcome, since it conferred an antibacterial effect. However, in addition to killing bacteria, the generated free radical species might be expected to initiate damage to chemical structures within the paper (Perez et al. 1998; Campanella et al. 2005). It has been shown, for instance, that in the presence of UV light, TiO2 particles can catalyze the breakdown of lignin (Machado et al. 2000). The breakdown appears to involve both degradation of phenolate groups and hydroxylation of aromatic rings. Perez et al. (1998) observed relatively minor decreases in the viscosity of cellulose solutions obtained from cellulose pulp that had been exposed to UV light in the presence of TiO2 and hydrogen peroxide. However, viscosity losses of about the same magnitude were observed in control tests not including the TiO2. Thus, the main concern is likely to arise in archival paper specimens that contain either lignin or various dyes that have aromatic structures.

Fortunately, there is a way to be able to benefit from the very high light scattering ability of TiO2 without having to suffer the risk of free-radical-induced degradation pathways. As is commonly done in preparing commercial TiO2 products for paints, the TiO2 particles can each be coated with a thin layer of another mineral, such as SiO2 (van Driel et al. 2016).

Kaolin clay

Kaolin clay has been in general usage as a paper filler since the 1780s, with increased usage in the 1870s (NAA 2022). The usage of kaolin clay within the main plies of printing paper products has declined greatly since the early 1980s (Hubbe 2005). However, it is still a major component in many coatings for paper. Kaolin can be generally described as a platy aluminosilicate mineral. Commercial kaolin products mainly are mined from metamorphic deposits that had been formed from the deposition of silt as glacial run-off emptied into ancient oceans (Dill 2016). As would be expected based on such a geologic origin, kaolin particles have a stable and generally unreactive chemistry relative to the concerns of paper’s permanence.

In common with other mined products, such as ground limestone calcium carbonate, kaolin products used in the paper industry will contain dispersants, which are usually polyacrylates or their copolymers (Loginov et al. 2008). The dispersants render a strong negative charge to the wetted surfaces. The resulting inter-particle repulsion is helpful during the mining, shipping, and storage. Papermakers employ cationic additives such as papermaker’s alum and cationic copolymers of acrylamide to retain not only the negatively charged kaolin particles, but also such items as cellulosic fines, PCC particles, and various additives to the papermaking process. The effects of dispersants and other such additives on archival performance do not appear to have been reported.

Other minerals

Some other mineral products are used by papermakers, but generally at lower quantities. Talc, which is a highly platy magnesium silicate mineral, is primarily used as a control agent for pitch and stickies in paper machine systems (Sutman and Nelson 2022). A variety of precipitated silicate and silica products, generally having much higher specific surface areas than ordinary kaolin, can be used in coatings for high-end printing papers, for instance to achieve high resolution ink-jet images (Cawthorne et al. 2003). Specialty grades of paper (glossy, coated, filled, etc.) historically have contained a variety of inorganic components in addition to those described above, such as zinc sulfide, zinc oxide, lead carbonate, etc. (NAA 2022). No information was found in the present literature searching work relative to the archival performance of such products.

Starch Products

Next in order of decreasing amounts in typical modern printing papers are various starch products. Starch has a long history of use in the paper industry as a sizing agent, strengthening agent, or as an adhesive for mineral-based coatings. Hunter (1947) reports use of starch to size paper as early as 700 A.D. Early Persian papermakers employed rice and wheat starches, as well as a variety of plant mucilages to prepare paper for writing (Barkeshli 2003). As papermaking traditions passed from the Middle East to Europe, these materials continued in use. For instance, Anguera (1996) reports that the earliest paper made in Valencia (early 14th century) was sized with “vegetable glue” of rice or wheat origin, or in other words starch. A primary function of these starch products is to confer strength to the paper, either by supplementing the inter-fiber bonding or by applying the starch as a film to the paper surface. In modern papermaking, internal or “wet-end” starch is typically a cationic derivative of starch, and the levels of addition are usually less than about 1.5% on a mass basis (Howard and Jowsey 1989; Jancovicova et al. 2012). Cationic starches were introduced to papermaking in the 1950s (NAA 2022). Greater amounts of starch, sometimes comprising 5% of the dry mass of paper, can be applied to the paper surface at a size press (Brouwer 1997; Shirazi et al. 2004). Essentially all of the starch that is used in printing papers has first been solubilized by cooking in water before its application (Horie 2010). In paper conservation, it is well known that starch films made from flour sources (i.e. gluten protein- and lipid-containing) age much more poorly than purer starch sources. These films stiffen and yellow to a greater degree. The type of starch (i.e. amylose to amylopectin ratio / different granule size, etc.) also affects how they age – rice vs. corn vs. potato vs. wheat starch films age differently.

For reasons of minimizing costs, a majority of the starch that is applied to the paper surface during commercial papermaking consists of unmodified starch, which is often called pearl starch. Such starch, when it comes from natural maize, potato, wheat, or tapioca sources, is invariably a mixture of two contrasting molecular forms. The amylopectin form, which has a branch point at each third amylose group on the chain, has some crystallinity in the native state, but it resists recrystallization (retrogradation) when dried from a solution (Gudmundsson 1994; Chang et al. 2021). The amylose form, which is linear and has a relatively low molecular mass often near one million g/mole (Ong et al. 1994), can adopt various crystalline forms. Figure 11 depicts two forms of structure within starch that can be present during the preparation of paper. Native starch can be in the form of helical structures. Notably, as shown in an end-wide view at the left of the fibers, V-type starch helices have been shown to have a more hydrophobic character toward the inside, making it possible for the starch to serve as a carrier of relatively hydrophobic compounds (Immel and Lichtenthaler 2000; Bildik Dal and Hubbe 2021). Most notably, the V-complex is able to enclose fatty acids or iodine (Tolstoguzov 2003; Putseys et al. 2010).

Cooking of the starch, followed by the passage of time in a heated solution, leads to a process called retrogradation (Andersson et al. 2008). This effect becomes prominent if an amylose-containing solution is held a long time within a temperature range of about 70 °C or higher, such that the amylose portions of it undergoes irreversible crystallization (Keetels et al. 1996). At a molecular scale, this entails formation of the macromolecules into a different form of helix (A-type) and precipitation of clusters of those helices (Conde-Petit et al. 2006; Liu et al. 2007). Though such changes indicate potential instability in starch structures, it is unclear the extent to which the helical forms and clustered forms of starch might continue to change during the subsequent storage of dried paper.

Fig. 11. Representation of a progressive crystallization process within starch, as has been observed in the aging of starch films over relatively short time periods

When starch is used as a binder in aqueous pigment formulations for the coating of paper, the starch type of choice is hydroxyethylated starch (Jauregui et al. 1995). Starch that has been hydroxyethylated is noted for its tough, non-brittle films and the absence of retrogradation phenomena. Since an ether bond is formed in the hydroxyethylation process, such products are expected to have good storage stability (Wu et al. 2018).

Evaluations of the storage stability of starch invariably have involved time periods very much shorter than those of interest to paper archivists. For instance, Mali et al. (2006) reported significant increases in crystallinity of various plasticized starch films during 90 days at room temperature. Fama et al. (2007) observed significant increases in crystallinity of starch films that had been plasticized with either glycerol or sorbate within a 4-week test period. Significant decreases in moisture content were apparent after eight weeks of storage. Thirathumthavorn and Charoenrein (2007) showed that the crystallinity of starch films increased during a month of storage. Panaitescu et al. (2015) found that nanofibrillated cellulose could be used to form a network within films comprised of starch and polyvinyl alcohol, thus hindering starch crystallization. Because all of these listed changes took place during a relatively short period, relative to the storage of books, one might anticipate more substantial changes to starch in paper over longer periods. In paper conservation, it is empirically known that starch films made from flour sources (i.e. gluten protein- and lipid-containing) age much more poorly (i.e. stiffen and yellow to a greater degree) than purer starch sources (Wills 1984). The type of starch (i.e. amylose to amylopectin ratio / different granule size, etc.) also affects how they age, such that rice vs. corn vs. potato vs. wheat starch films age differently (Indictor et al. 1978; Van Steene and Masschelein-Kleiner 1980).

Other Natural Polymers


The polysaccharide chitosan, which is depicted in Fig. 12, is remarkably similar to cellulose except that what would have been an –OH group on the C2 carbon of the anhydroglucose unit is an amine group instead. Chitosan can be obtained by treating the shells of shrimp or other crustaceans with strong solutions of NaOH. Various authors have proposed the use of chitosan as an additive for papermaking (Rohi Gal et al. 2023). In particular, Basta (2003) evaluated the possible use of chitosan to enhance the aging resistance of paper that had been treated with rosin and alum. It is well known that chitosan is soluble under acidic aqueous conditions, as in the presence of acetic acid. In the cited work, paper specimens were dipped in solutions of 0.4% chitosan and 1% acetic acid, which was followed by treatment with alkaline solutions of sodium hydroxide or sodium silicate as precipitators. The type of precipitating agent was found to be important, with the sodium silicate providing better results. The reported results are generally consistent with what would be expected based on a deacidification treatment (Baty et al. 2010).

Fig. 12. The structure of chitosan (depicting a fully deacetylated form)


Proteins, especially gelatin, have been widely used as sizing agents during the period of hand-papermaking in Europe (Lang et al. 1998). As shown in Fig. 13, proteins are products of the formation of peptide bonds between amino acids.


Fig. 13. Peptide bonding structures as the linkages within proteins

It is important to note that alum was not first introduced to the papermaking process for alum-rosin internal sizing, but as an additive during external sizing of paper with the protein gelatin (tub sizing). Potassium aluminum sulfate was introduced by the mid-17th century to assist in hardening the gelatin size, and, in increasing amounts throughout the week to prevent putrefaction of the gelatin solution (Clapp 1972; Baker 2010), such that gelatin use and alum use cannot always be detangled. Chen et al. (2003) showed faster degradation of paper that had been sized with a combination of alum and animal glue when subjected to accelerated aging. Gelatin, which is mainly composed of the protein collagen, is also considered as a strengthening agent for the conservation of weakened paper specimens (Basta and Fadl 2003; Hummert 2019). The word “resizing” is used to mean treatment of a paper artifact with solutions of either gelatin or cellulose ethers to increase strength or replace sizing removed during the course of aqueous treatment (Henry 1986; Hummert 2019).

Basta and Fadl (2003) reported that paper specimens strengthened by treatment with solutions of gelatin were subject to rapid loss of strength during accelerated aging. The effects were especially evident when testing paper having a relatively low basis weight. The results were surprising in the light of work reported by Barrett and Mosier (1994), who found a positive correlation between the gelatin content of historic paper specimens and their lightness (L*) value from color measurements. The observed beneficial effect of the gelatin had been attributed to an alkaline buffering capability of gelatin (Baty and Barrett 2007). Lang et al. (1998) reported correlations between the gelatin content of historic paper specimens and several other parameters, including surface pH, fluorescence, sulfur content, and the L* value. The cited work by Barrett and Mosier (1994) and Lang (1998) did not, however, include any evaluation of strength of the historical papers that were evaluated. It is possible that even if the gelatin treatment had contributed to raising the pH of the historical papers when they were new, that gelatin might have decomposed, in part, during the ensuing years, leading to the strength losses corresponding to those reported by Basta and Fadl (2003). Another possibility is that the gelatin material or the procedures employed by Bast and Fadl (2003) were significantly different from those that had been employed in originally preparing the historical specimens studied by Barrett and Mosier (1994). Further testing is needed to resolve such issues.

Gelatin has also been shown to have stabilizing effects on the oxidative and hydrolytic damage that can be caused by iron gall inks, either as initial sizing material, or as re-sizing applied during a conservation treatment (Kolbe 2004; Poggi et al. 2016; Gimat et al. 2021). Gimat et al. (2021) demonstrated that in gelatin sized or re-sized samples there is less iron to be found inside the cellulosic fiber.

Other proteins, such as animal glue and casein, and later soy proteins have been used as coating adhesives (NAA 2022).

Synthetic Water-soluble Polymers

In modern papermaking a variety of petroleum-derived chemical additives are used. Water-soluble polyelectrolytes widely used by papermakers include acrylamide copolymers, glyoxylated polyacrylamide, polyvinyl alcohol, and polyamidoamine-epichlorohydrin, among others. An example is shown in Fig. 14. No accounts of accelerating aging tests were found corresponding to the acrylamide products mentioned.

Perhaps because it is widely used in paper conservation, several studies have considered the permanence characteristics of polyvinyl alcohol (PVOH). Bicchieri et al. (1993) found that treatment of paper with low-mass PVOH for restoration purposes was effective for restoration of strength. Selected accelerated aging tests were carried out on a paper prepared by the authors, treated by immersion in PVOH, and then subjected to the hot conditions for specified time periods. The beneficial effect of the PVOH on paper properties did not deteriorate as a function of accelerated aging time. The PVOH treatment also appeared to resist biological attack according to sources cited by those authors.

Fig. 14. Chemical structure of an acrylamide copolymer using in papermaking

Basta (2004) prepared a mixture of PVOH with or without borax. The best results, with respect to paper treatment, were with a PVOH treatment level of 0.5%. Borax treatment, as a supplement to the PVOH treatment, led to slight further improvements at low levels of PVOH. Notably, by adding borax, the beneficial effects of PVOH did not fall off even when the PVOH level was increased to 0.75% or higher. Because borax has an alkaline pH, it might be expected that the beneficial effect on the paper was related to maintaining an alkaline pH in the paper.

Hydrophobic Sizing Agents

Due to contrasting sets of issues, the following discussion regarding hydrophobic internal sizing agents will be divided into four main topic areas: rosin-alum sizing, alkaline sizing, surface-application of hydrophobic copolymers, and the use of wax-like materials. Gelatin, starch, and plant mucilages applied as external sizing agents have been discussed in other sections of this paper.

Rosin sizing agents

There are many reports documenting the adverse effects of rosin-alum sizing on the permanence of paper (Clements 1987; El-Saied et al. 1998; Baty and Sinnott 2004; Area and Cheradame 2011).

Fig. 15. Preparation of fortified rosin from levopimaric acid (a rosin component) and maleic anhydride

Kim et al. (1988) reported degradation of rosin by UV illumination, leading to discoloration of paper. El-Saied et al. (1998) found that paper that had been sized with rosin and alum also had a lower thermal decomposition temperature, which can be taken as evidence of some molecular breakdown having occurred before the testing. Figure 15 illustrates the preparation of fortified rosin, which is a main component of most rosin products currently used by papermakers.

There are two main approaches that papermakers employ to render paper hydrophobic with usage of rosin products. The oldest technology is based on the saponification of rosin with alkali, which converts the carboxylic acids to their soap form. After adding the soap to a papermaking fiber suspension, the rosin is set (precipitated onto the fiber surfaces) by addition of aluminum sulfate (Ehrhardt and Leckey 2020). This reaction is largely accomplished in the aqueous phase, even before the resulting paper sheet is dried. The second approach is to heat up the rosin mixture (including fortified rosin) to its melting point and prepare an emulsion in the presence of a stabilizing polymer, such as cationic starch or polyamidoamine-epichlorohydrin. When such an emulsion is added to a paper machine furnish, the emulsified droplets coated by the cationic polymer are retained onto fiber surfaces, but the rosin does not become molecularly spread and anchored onto the fiber surfaces until the paper is being dried (Ehrhardt and Leckey 2020).

A recent review article by Jablonsky et al. (2020) suggests that the aluminum has a more dominant effect on paper degradation, in comparison to the rosin. In addition to the well-known effects of low pH, the cited authors provided evidence of involvement of aluminum species in redox radical oxidation of cellulose carboxyl or carbonyl groups. Because alum has a high buffering capacity, at a given pH in the range of about 4 to 5.5, the titratable acidity in the presence of alum will be significantly higher than in a solution of a strong acid such as sulfuric acid. Notably, it has been reported that deacidification treatments do not necessarily remove the adverse effects of aluminum compounds on paper permanence (Jablonsky et al. 2020). The cited authors suggested that such effects may be due to non-uniformity of the chemical compounds, such that acidic regions may still persist.

Because rosin compounds such as abietic acid and levopimaric acid are carboxylic acids, one can expect that they too might contribute to the acidity of paper. The fallacy in that argument is that the pH of the added rosin product added during papermaking can be very different, depending on the type of rosin product that is employed. The traditional rosin-alum system, going back to the invention by Illig (1807), employed the soap form of rosin, which has an alkaline pH. Rosin free-acid emulsion sizing agents, which have become increasingly popular in recent years, have an inherently acidic pH, but they tend to be employed in papermaking processes at a somewhat higher value of pH (e.g. 4.5 to 5.5) than the traditional rosin soap sizing agents (e.g. 4 to 4.5) (Ehrhardt and Leckey 2020).

Another factor worth considering is the effect of the rosin sizing on the strength of the original paper, even before being subject to the effects of time. Recent work by Korpela et al. (2021) showed that sizing with rosin and alum decreased the initial strength of paper handsheets. The effects were consistent with the rosin having become spread over the fiber surfaces before formation of inter-fiber bonding. No strength loss was observed in cases where the sizing was carried out using alkylketene dimer (AKD) as the sizing agent. Those results are consistent with a mechanism in which an emulsified sizing agent does not spread until the paper has become substantially dry and the bonded areas have become established (Hubbe 2014b).

Alkaline sizing agents

When paper is manufactured in the presence of calcium carbonate filler, it is common to employ the hydrophobic sizing agent alkylketene dimer (AKD) or alkenylsuccinic anhydride (ASA). Introduced in the 1950s and 1970s, respectively (NAA 2022), both of these products need to be emulsified using some form of polymer, such as cationic starch or a cationic synthetic polymer. Because the latter are likely to benefit the strength of the paper, once again there can be uncertainty regarding the direct effects of the AKD or ASA on archival properties. Perhaps more importantly, the calcium carbonate (as discussed earlier) will have controlled the pH to be higher than 7.5 and more likely to be between 8 and 8.5. Thus, it is not surprising that Basta et al. (2006) found that AKD-sized paper had much higher tolerance of accelerated aging in comparison to paper than had been sized with rosin and alum. This is consistent with the earlier finding of El-Saied et al. (1998) and Basta and Fadl (2003) that handsheets prepared at neutral pH were stronger than those prepared at acid pH with rosin and alum. A search of the literature did not reveal any work having been done relative to effects of ASA on the archival properties of paper. Figure 16 represents the reactions associated with paper sizing with ASA and AKD.

Fig. 16. Hydrophobic sizing systems based on treatments with alkenylsuccinic anhydride (ASA) and alkylketene dimer (AKD)

Fig. 17. Chemical structure of styrenemaleic anhydride (SMA) and a representation of its tendency to orient itself so that the styrene groups face outwards toward the atmosphere at a paper surface during the drying process

Surface-applied hydrophobic copolymers

Hydrophobic copolymer agents applied at the size press of a paper machine were considered in a review article (Bildik Dal and Hubbe 2021). As shown in Fig. 17, a representative of this type of agent is styrenemaleic anhydride (SMA) copolymer. Due to the content of maleic acid functional groups, it is reasonable to expect some formation of ester bonds, depending on the temperature and time of curing as the paper is being dried. Effects relative to the archival performance of paper do not appear to have been considered in published literature.


Surface treatment or impregnation with wax can be used for making wax paper or preparing a barrier coating for food packaging (Spence et al. 2011). Figure 18 shows representative structures for waxes derived from natural sources or from petroleum. The direct effect on paper properties may be hard to predict, since one might expect that a hydrophobic material such as wax would interfere with the formation of hydrogen bonding between the adjacent fibers. However, Renee et al. (2004) observed that treatment of kraft bagasse pulp fibers with a petroleum wax emulsion resulted in higher strength of the resulting handsheet paper. The authors concluded that the wax was able to cover the surfaces of individual fibers, creating a hydrophobic surface. The fact that a hydrophobic, waxy surface sometimes can contribute to bonding is consistent with the results of Bildik et al. (2016), who treated blotter paper with solutions of waxy AKD sizing agent in heptane. It was proposed in that work that the AKD was functioning as the matrix phase of a composite structure.

Fig. 18. Chemical structures of typical wax components from natural sources (adapted from Christie 2022)

With respect to archival performance related to wax, interesting results were presented by Jeong et al. (2014b), who evaluated Hanji paper from the Annals of King Sejong from about year 1300 that had been coated with beeswax. These specimens had suffered a much more severe loss in degree of polymerization in comparison to a conventional Hanji paper specimen. In this case, the reason for degradation was not the wax or components thereof and also not degradation products that could have increased the acidity of the wax-coated papers, but the microorganisms growing on the wax secreting cellulolytic enzymes that caused severe hydrolysis. The presence of citric acid, not a degradation product of cellulose/paper, but a classical metabolite from microorganisms, provided the evidence. Natural waxes are known to be comprised of fatty acids (Tulloch and Hoffman 1972), such that their hydrolytic breakdown would give rise to carboxylic acids. This concept is consistent with the findings of Regert et al. (2001), who found that the heating of beeswax yielded palmitic acid, among other compounds. However, their pKa is less detrimental compared to the action of a hydrolytic enzyme in the splitting of a glycosidic bond. Barrow (1953) reported instances in which paper sheets appeared to be damaged due to their location adjacent to natural material. In that case, groundwood material was identified as the source of acid substances. The damaging effects were attributed to migration of acid-containing material into an adjacent structure composed of delignified fibers.

Inorganic Coagulants and Iron Gall Ink

The topic of coagulants deserves to be considered more deeply, in light of the importance of alum, which is not only a coagulant but also an important source of acidity that affects the permanence of paper items. Another important inorganic coagulant is polyaluminum chloride (PAC). Some examples of coagulants are shown in Fig. 19. Iron salts also would fall in the same category, but due to their rust-like color, they seldom are intentionally used by papermakers.

A potential contributing problem associated with the usage of inorganic coagulants, in addition to their acidity, would be their tendency to reduce the swelling of cellulosic materials (Nedeltschewa 1977; Kato et al. 2000). As shown by Kato et al. (2000) the presence of aluminum compounds in pulp reduced the water retention value of carboxymethylcellulose (CMC) powders that had been dried at either 20 or 105 °C. This implies partial and irreversible loss of swelling ability occurring in the course of drying of the material. In other words, the aluminum treatment may have promoted hornification of the fibers during the drying process. This is a concern in the light of the previously mentioned relationship between hornification and the accelerated aging of paper (Kato and Cameron 1999). It is worth noting, however, that CMC it quite different from the cellulose fibers present in paper documents.

The topic of inorganic coagulants appears to bear a relationship to the widely studied problems associated with the use of iron gall inks on certain archival documents (Kolar et al. 2006; Potthast et al. 2008; Melo et al. 2022). Iron gall and carbon black inks were the most commonly used black inks in early European manuscripts because of their durability. The main ingredients of iron gall ink are gallic acid, iron source (FeSO4), gum arabic, and water (Stijnmann 2004; Çakar and Akyol 2022). Kolar et al. (2006) found strong correlations between the width of an ink line, decreasing pH, and decreasing basis weight of the paper on the likelihood of apparent damage to historical documents upon which iron gall inks had been applied. When copper salts are present as either an ingredient or a trace contaminant in iron salts, increased oxidation has been observed, but the same amount of hydrolysis (Kanngießer et al. 2004; Potthast et al. 2008). Neevel (1995) found that documents could be protected by treating them with natural complexing agents, such as calcium phytate or calcium hydrogen carbonate in aqueous solution. Presumably these agents were able to complex with iron species. Follow-up work by Henniges et al. (2008) showed that treatment with a mixture of calcium phytate and calcium hydrogen carbonate inhibited further hydrolysis and oxidation. Importantly, this stability is demonstrated both in areas of paper carrying ink, but also in regions of un-inked paper. Tse et al. (2010) used micro-fading to show that the combined calcium phytate and calcium hydrogen carbonate treatment also serves to make iron gall inks less sensitive to change with light exposure. Research by Rouchon et al. (2011) also supported oxidation as a main mechanistic pathway for damage of paper due to the presence of iron gall inks. The cited authors suggested that the mechanism may be promoted by the ability of gallic acid, a tannin component, to reduce iron (III) to iron (II).

Fig. 19. Some inorganic coagulants (papermaker’s alum and polyaluminum chloride (PAC)), as well as chemical structures of the high-charge cationic organic polymers dimethylamine-epichlorohydrin, poly(ethyleneimine), and poly-(diallyldimethylammonium chloride)

Organic Coagulants

The organic coagulants employed in modern papermaking can be described as intermediate molecular mass polyamines having a high density of cationic charge. Examples include poly(dimethylamine epichlorohydrin), poly(diallyldimethylammonium chloride), poly(ethyleneimine), and in recent years poly(vinylamine). Some examples are shown in the lower part of Fig. 19. There does not appear to have been research attention focused on the effects of such additives on permanence characteristics of paper. However, a relevant effect, analogous to what has been reported for treatment with aluminum species (Kato et al. 2000), has been reported. Zhang et al. (2002) reported that treatment of never-dried unbleached kraft pulp with poly(dimethylamine epichlorohydrin) before drying at 105 °C led to relatively large losses in the strength of paper made from the same fibers after they had been reslurried in water. The observations were consistent with the hornification mechanism already discussed. It follows that treatment of fiber suspensions with sufficient amounts of such coagulants might be expected to embrittle paper, though further testing would be needed to confirm such a mechanism.

Dyes and Fluorescent Whitening Agents

Dyes may represent a point of vulnerability of archival papers. Dyes necessarily absorb the energy of light in order to perform their function, and such absorbed energy has the potential to break down various chemical structures, depending on their susceptibility (Colombini et al. 2007; Gervais et al. 2014). There is a wide range of color-fastness relative to light exposure among different dyestuffs used in papermaking (Lips 1981).

According to one report, certain dyes can affect the storage stability of paper items (Anon 1957). It was reported that direct dyes can have a protective effect on paper, whereas acid dyes can be harmful. Anthraquinoid dyes were said to have a “photo-rendering” effect on cellulose (El-Saied and Basta 1998). When such paper samples were treated with non-aqueous antioxidant and deacidification solution along with a natural dye, there were decreases in the rate of loss of the degree of polymerization and decreasing pH of the papers (Çakar and Akyol 2022).

Fluorescent whitening agents, which have chemical and structural aspects similar to those of some direct dyes, have been noted for their tendency to lose their fluorescent whitening ability during long exposure to light (He et al. 2015). The mechanism of energy absorption of ultraviolet light, partial dissipation as heat, followed by emission of visible light, is illustrated in Fig. 20. Such a mechanism suggests that the molecule may be susceptible to decomposition when it is in its activated state, while an electron is in an antibonding orbit. Several authors have explored the history and deterioration of these “optical brighteners” in cultural heritage materials, confirming their light-sensitivity (Leclerc and Flieder 1992; Mustalish 2000, Connors-Rowe et al. 2007). Mustalish (2000) notes additional pH and heat sensitivities and identifies that their deterioration may alter appearances not only by a quenching of their emitted fluorescence, but possibly also by a darkening of yellowing of the degraded dye compound, along with any adverse interactions between degrading dye and other paper materials. Conservators are becoming increasingly aware that optical brighteners may be moved or removed by aqueous treatments (Wetzel 2005; Engelke 2023).