Chemical modification of cellulose-rich fibres to clarify the influence of the chemical structure on the physical and mechanical properties of cellulose fibres and thereof made sheets

Accepted Manuscript Title: Chemical modification of cellulose-rich fibres to clarify the influence of the chemical structure on the physical and mechanical properties of cellulose fibres and thereof made sheets Authors: Ver´onica L´opez Dur´an, Per A. Larsson, Lars W˚agberg PII: DOI: Reference:

S0144-8617(17)31282-1 https://doi.org/10.1016/j.carbpol.2017.11.006 CARP 12957

To appear in: Received date: Revised date: Accepted date:

25-8-2017 16-10-2017 1-11-2017

Please cite this article as: L´opez Dur´an, Ver´onica., Larsson, Per A., & W˚agberg, Lars., Chemical modification of cellulose-rich fibres to clarify the influence of the chemical structure on the physical and mechanical properties of cellulose fibres and thereof made sheets.Carbohydrate Polymers https://doi.org/10.1016/j.carbpol.2017.11.006 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Chemical modification of cellulose-rich fibres to clarify the influence of the chemical structure on the physical and mechanical properties of cellulose fibres and thereof made sheets. Verónica López Durán1,2, Per A. Larsson1,2, Lars Wågberg1,2 1

Fibre and Polymer Technology, KTH Royal Institute of Technology, SE-10044 Stockholm, Sweden

2

BiMaC Innovation, KTH Royal Institute of Technology, SE-10044, Sweden

Corresponding authors: [email protected], [email protected], Phone: +46(0)87908296, Fax: +46(0) 879096166.

Highlights:    

Surface modification to provide different functional groups. An amorphous shell of modified cellulose surrounds the crystalline cellulose. Relation between chemical structure and performance of handsheets. The performance of the materials is determined by the functional groups.

Abstract Despite the different chemical approaches used earlier to increase the ductility of fibre-based materials, it has not been possible to link the chemical modification to their mechanical performance. In this study, cellulose fibres have been modified by periodate oxidation, alone or followed either by borohydride reduction, reductive amination or chlorite oxidation. In addition, TEMPO oxidation, and TEMPO oxidation in combination with periodate oxidation and further reduction with sodium borohydride have also been studied. The objective was to gain understanding of the influence of different functional groups on the mechanical and structural properties of handsheets made from the modified fibres. It was found that the modifications studied improved the tensile strength of the fibres to different extents, but that only periodate oxidation followed by borohydride reduction provided more ductile fibre materials. Changes in density, water-holding capacity and mechanical performance were also quantified and all are dependent on the functional group introduced. Keywords: Cellulose fibres, borohydride reduction, chemical modification, chlorite oxidation, periodate oxidation, TEMPO oxidation, structure-property relationship.

1

1. Introduction In recent years, there has been a rapid development of bio-based materials with the aim of replacing oil-based materials. Cellulose is one of the polymers that could potentially be used for this purpose, with renewability, low cost, high stiffness, high strength and low density (Bledzki & Gassan, 1999; Klemm, Heublein, Fink, & Bohn, 2005). However, their limited ductility is one of the major shortcomings that make it difficult to use cellulose fibres in more geometrically advanced applications. Several chemical procedures have been used to improve the ductility of wood fibres. Marais et al. (2014) showed that the strainability of handsheets can be increased to 6–7% when each individual fibre is coated with a polyelectrolyte multilayer composed of polyallylamine and hyaluronic acid. Matsumura et al. (2000) a produced partially derivatized cellulose ester from cellulose fibres with a combination of sulfonic p-toluene and hexanoic anhydride, and the produced cellulose hexanoate produced was thermally deformable with a modulus of 0.4 GPa and an elongation at break of 30% at a DS of 1.7. Vuoti et al. (2013) prepared hydroxypropyl cellulose ethers and found that handsheets made from such modified fibres were transparent and that freely dried handsheets had a strain-at-break of 16%. The disadvantage of the last two examples of chemical modification is that they must be performed in organic solvents. Recently, Khakalo et al. (2014) used gelatine to improve the mechanical performance of freely dried handsheets made from beaten fibres, and showed that the strain-at-break could be increased from 10% to 22%, and the tensile index also showed an increase from 59 to 70 kNm/kg after a 20 wt% addition of gelatine. Zeronian et al. (1964) showed that papers made from cotton linters could be strained to 35% after a two-step modification in aqueous media. In the first step, periodate oxidation was used to cleave the cellulose C2–C3 bond to form dialdehydes, and in the second step, a borohydride reduction was used to reduce the aldehydes and form dialcohol cellulose. Larsson et al. (2014) studied the same modification of bleached kraft fibres under heterogeneous conditions and found that handsheets made from the modified fibres with a degree of modification of 27% could be strained to 11% and had a stress at break of 90 MPa. If the fibres were mechanically beaten before the chemical modification, or modified to an even higher degree, handsheets could be prepared with a strain-at-break of up to 40%, and tensile strengths between 50 and 100 MPa. The combination of a high stress at break and a high ductility is presumed to be due to a highly crystalline core of cellulose in the fibrils surrounded by an amorphous shell of ductile, amorphous dialcohol cellulose. This shell provides a high flexibility and presumably a high molecular mobility, which facilitates a better contact between the fibres during sheet consolidation. The consolidation is in fact so great that the papers formed had a density greater than 1200 kg/m3, highly transparent and a significant oxygen barrier. Another property introduced by this two-step chemical modification is thermosplasticity, which not only gives greater transparency upon hot pressing but also greater ductility and the possibility of 3D forming complex paper structures (Larsson & Wågberg, 2016; Linvill, Larsson, & Östlund, 2017).

2

Despite the different chemical methods that have been developed to improve the ductility of paper, there is no clear understanding of the exact mechanism behind the effects observed from the introduction of different functional groups onto the cellulose. It is therefore important to gain a fundamental understanding of how chemical modification affects all these properties, and how they affect the performance of the final material. It is known that a shell of dialcohol cellulose enveloping each crystalline nanofibril core has a significant effect on the properties of the material, but little is known about other functional groups, or combinations thereof. In the present study, we have therefore carried out heterogeneous chemical modifications of cellulose fibres in aqueous media to study the influence of different functional groups on the described core–shell arrangement and how they affect the performance of handsheets prepared from the modified fibres. The fibres were partially modified by performing periodate oxidation to produce dialdehyde cellulose. Thereafter, the aldehydes were reduced, aminated or further oxidised. In addition, TEMPO oxidation was used alone and combined with periodate oxidation and a further reduction with sodium borohydride. The degree of modification was kept low, which in this case meant that only 11% of the glucose units were derivatised.

2. Experimental 2.1. Materials

2.1.1. Fibres A bleached softwood kraft pulp (K48) according to (OO)Q(OP)(ZQ)(PO) sequence, was supplied by SCA Forest Products (Östrand pulp mill, Timrå, Sweden). The fibres were first disintegrated followed by a washing procedure to remove metal ions and convert anionic charges to their sodium form (Wågberg & Björklund, 1993).

2.1.2. Chemicals Sodium metaperiodate (99%), was purchased from Alfa Aesar. Sodium carbonate (≥99.5%), hydroxylamine hydrochloride (99%), sodium borohydride (≥96%), 2-propanol (99.9%), 4-acetamido2,2,6,6-tetramethylpiperidine 1-oxyl (4-AcNH-TEMPO, ≥98.0%), sodium chlorite (80%), sodium hypochlorite solution (10-15% available chlorine), sodium acetate (99%), sodium phosphate monobasic monohydrate (98%) and hydrogen peroxide (30 wt% in water) were purchased from Sigma-Aldrich. Sodium hydroxide and hydrochloric acid standard solutions (1 M) were purchased from Merck Millipore. Tap water was used for handsheet preparation but deionised water was used in all other cases.

2.2. Methods

3

2.2.1. Chemical modification Combinations of TEMPO and either periodate oxidation alone or followed by borohydride reduction, amination or chlorite oxidation were carried out. The combinations studied are summarised in Figure 1. All the reactions were stopped by filtration and washing with water until a conductivity less than5 μS/cm was reached.

2.2.1.1. Periodate oxidation Fibre suspensions with a consistency of 20 g/l were allowed to react with 1.35 g NaIO4/g fibre in the presence of 2-propanol for 30 min in the dark at a temperature of 50 °C. This is an addition that was developed in earlier studies.

2.2.1.2. Borohydride reduction of dialdehydes The fibres were redispersed to a consistency of 8 g/L, and the dispersion was allowed to react with 0.5 g NaBH4/g fibre and NaH2PO4 0.01 M for 2 h (Larsson, Berglund, & Wågberg, 2014).

2.2.1.3. TEMPO-mediated oxidation TEMPO-mediated oxidation was carried out as previously described by Tanaka et al. (2012). In brief, cellulose fibres (non-modified or aldehydes containing) were dispersed in 0.1 M acetate buffer adjusted to pH 4.8 and 4-AcNH-TEMPO and NaClO2 were added to the dispersion followed by the addition of NaClO. The reaction was allowed to proceed for 48 h at 40 °C.

2.2.1.4. Chlorite oxidation of dialdehydes Periodate-oxidised fibres were dispersed to a consistency of 20 g/L. An amount of 2.5 mmol of NaClO2 and 2.5 mmol H2O2 were added per millimole of aldehyde. The suspension was acidified with 1 M HCl to pH 5 and allowed to react for 48 h at RT.

2.2.1.5. Reductive amination of dialdehydes with hydroxylamine Both hydroxylamine solution and suspensions of periodate-oxidised fibres were set to pH 4 before being allowed allowing them to react together at RT for 2 h. The stoichiometric amount of hydroxylamine added was five times the amount of aldehyde groups on the cellulose fibres. The final consistency of fibres was 10 g/L. An in-situ reduction of the imine group formed was then carried out by adding NaH2PO4 with a final concentration of 0.01 M and 0.5 g NaBH4/g fibre. The reaction was kept under agitation for 2 h at RT.

2.2.2. Carbonyl content determination 4

The content of aldehydes was determined by titration with hydroxylamine (Larsson, Gimåker, & Wågberg, 2008). Each determination was performed in triplicate.

2.2.3. Total charge determination The total charge was determined by conductometric titration using a Metrohm 702SM Titrino titrator according to SCAN-CM 65:02. Each determination was performed in duplicate.

2.2.4. Water retention value measurements The water retention value (WRV) of modified fibres, i.e. mass of water in wet pulp divided by the mass of dry pulp, was performed according to a simplified version of SCAN 62:00. Approximately 1– 2 g of never-dried fibres were placed in 5 ml disposable centrifuge filters (Millipore Ultrafree-CL Centrifugal Device, Billerica, USA) equipped with 5 μm polyvinylidene fluoride membranes. The fibres were centrifuged at 30 000 rpm for 30 min at RT and the dry content was determined by weighing the fibres before and after overnight drying at 105 °C. The WRV was measured in duplicate after equilibrating the samples at pH 2 and 8.5.

2.2.5. Handsheet preparation Handsheets with an approximate grammage of 100 g/m2 were prepared using a Rapid Köthen sheet former (Paper Testing Instruments, Austria). Sheets were dried between 400 mesh woven metal wires (The Mesh Company Ltd, Warrington, UK), attached to regular sheet-former carrier boards, for 15 min at 93 °C and at a reduced pressure of 95 kPa. Sheets were stored at 23 °C and 50% relative humidity until they were further tested.

2.2.6. Nitrogen analysis Elemental nitrogen analysis was performed after reductive amination to determine the amount of nitrogen present in the fibres. A sample of 10–15 mg was taken from a handsheet and the analysis was performed with an ANTEK 7000 Model 737 (Antek instruments Inc., USA). A calibration curve was prepared using hydroxylamine hydrochloride in the range of 0.005–0.04 mg (see Figure 1 in supporting information). Determinations were performed in duplicate.

2.2.7. X-ray diffraction The crystallinity of the samples conditioned at 50% RH was measured by X-ray diffraction (XRD) in an X’Pert PRO XRD (PANanalytica, The Netherlands), using copper radiation Kα (1.5418 Å), a voltage of 40 kV and a current of 45 mA. Diffraction patterns were obtained between 2θ=5° and 2θ=30° with a step size of 0.020°. The crystallinity index was calculated as previously reported (French & Santiago Cintrón, 2013; Segal, Creely, Martin, & Conrad, 1959). The Scherrer equation (Patterson, 1939) was used to estimate the crystallite width. 5

2.2.8. Scanning electron microscopy Fibre morphology was studied with a Hitachi S-4800 field emission scanning electron microscope (SEM). Prior to imaging, the samples were sputtered with a ~10 nm Pt-Pd coating in a 208 HR Cressinton sputter coater to suppress specimen charging.

2.2.9. Mechanical testing Prior to tensile testing, the structural thickness of the handsheets was evaluated according to SCAN-P 88:01. Thereafter, the handsheets were cut into strips with a width of 15 mm which were tensile tested using an Instron 5944 with a 500 N load cell. The paper strips had a free span of 100 mm between the clamps and were strained at a constant rate of 100 mm/min. Two papers from each functional group and a total of ten specimens were tested and the results were reported for those which did not fail immediately in the jaw face. For wet tensile strength testing, the sheets were soaked in water for one hour prior to testing according to SCAN-P 20:95.

3. Results and discussion

3.1. Chemical characterization of modified fibres According to Figure 1 the fibres used were modified using a combination of chemical reactions. Periodate oxidation was performed to open the anhydroglucose ring in the C2–C3 position and to partially form dialdehyde cellulose (Sample 2). Once dialdehydes had been formed, it was possible to convert them to other functional groups; borohydride reduction to produce dialcohol cellulose (Sample 3); reductive amination to give covalently attach hydroxylamine (Sample 4), and chlorite oxidation to produce dicarboxylic acid cellulose (Sample 5). The cleavage of the C2–C3 bonds, i.e. the degree of modification, was kept at a low level to avoid fibrillation of the fibres in the sequential reactions which introduce charges. Carboxylic acids were introduced on the C6 of the glucose unit through TEMPO oxidation (Sample 6), combined with the introduction of either dialdehyde (Sample 7) or dialcohol functionality (Sample 8).

The presence of aldehydes was determined by reaction with hydroxylamine hydrochloride and, as shown in Table 1, 1.36 mmol/g of aldehydes were formed after periodate oxidation (Sample 2), corresponding to a degree of oxidation of about 11%. No aldehydes were detected after reduction with borohydride. Furthermore, aldehydes known to be present after TEMPO oxidation were reduced with borohydride to avoid the unintentional effect of these aldehydes on the properties of handsheets. It was also observed that a longer periodate oxidation time was required to achieve a given carbonyl content when the fibres were first oxidised with TEMPO (Sample 7). Scott et al. (1968) suggested that the

6

presence of acids decreases the rates of reaction due to electrostatic repulsion between the acidic groups on the polysaccharides and the negatively charged periodate ion. Table 1 also shows the total charge of the modified fibres. It was found that the charge densities of dialdehyde cellulose (Sample 2) and dialcohol cellulose fibres (Sample 3), were lower than that of the original fibres. Larsson et al. (2014) and López Durán et al. (2016) showed that low molecular weight components, e.g. hemicelluloses, where charges are most likely to be found, are removed after periodate oxidation followed by borohydride reduction. Furthermore, the molecular weight distribution after borohydride reduction showed a removal of the peaks below a molecular mass of 100 kDa. When periodate-oxidised fibres were further aminated (Sample 4) a decrease in charge was observed, which could be explained as being due to a loss of low molecular weight compounds. By chlorite oxidation of dialdehyde fibres, it was possible to convert all the aldehydes to carboxylic acids (Sample 5), leading to a total charge density of 1352 μeq/g. This is a significantly greater charge density than the 783 μeq/g achieved after TEMPO oxidation (Sample 6). When the TEMPO oxidation was followed by a periodate oxidation (Sample 7), or by periodate oxidation and borohydride reduction (Sample 8), a decrease in charge density was observed. It has been shown that dialdehyde cellulose is sensitive to beta-alkoxy-elimination and that it proceeds as a competitive process during the reduction of dialdehyde cellulose by sodium borohydride (Potthast, Schiehser, Rosenau, & Kostic, 2009; Sihtola, 1959). The cleavage of the cellulose then occurs at the C5, which is in beta-position to the C3 aldehyde, while the C2 aldehyde, has no β-alkoxy substituent and no beta elimination occurs. Each of the resulting two fragments contains one aldehyde group each and the aldehyde content therefore remains constant. A clustered oxidation may also occur when two neighbouring oxidised glucose units are present. In this case, two eliminations occur starting from the C3 aldehyde and three fragments are produced. Two of them are the two halves of the cleaved cellulose chain, carrying one carbonyl each, and the third fragment is a low-molecular weight molecule with two aldehyde groups (Potthast et al., 2009).

To quantify the presence of amines in the fibres, elemental analysis was performed and it was found that the aminated fibres (Sample 4) had 1.28 ± 0.11 mmol amines/g fibre, which agrees well with the carbonyl value obtained by titration (see Table 1). Sarymsakova et al. (1998) found that hydroxylamine, in general, is a good nucleophile, due to its pKb of 5.9, and it is therefore capable of reacting to react also with the hemiacetal formed possibly formed between the aldehyde groups and hydroxyl groups. Molecular weight distributions of the partially modified fibres were also studied. After periodate oxidation, an apparent decrease in molecular weight was observed (see Figure 3 in supporting information), but when the aldehydes were reduced with sodium borohydride the molecular weight distribution shifted towards higher molecular weights and became narrower. These contradictory results indicate the shortcoming of size-exclusion chromatography for determining the 7

molecular weight of chemically modified cellulose but, to the knowledge of the authors, there is no other obvious and straightforward way of determining the molecular weight of these modified celluloses. No doubt there is a need for very accurate methods for molecular mass determination of chemically modified cellulose.

3.2. Crystallinity, water swelling and morphology of modified fibres Figure 2 shows XRD patterns of the differently modified fibres, indicating that the crystallinity of the fibres, quantified as crystallinity index, is reduced by the initial periodate oxidation, which is in agreement with earlier investigations (Kim, Kuga, Wada, Okano, & Kondo, 2000). After borohydride reduction (Sample 3), reductive amination (Sample 4) and chlorite oxidation (Sample 5), the crystallinity index increased compared with the crystallinity after periodate oxidation. No change in crystallinity was observed for the TEMPO-oxidised fibres (Sample 6), which agrees well with previous studies (Saito et al., 2009). In Figure 2, it can also be observed that, after periodate oxidation (Sample 2) and sequential TEMPO and periodate oxidation (Sample 7), the crystallinity is similar supporting the observation that TEMPO oxidation does not significantly affect the degree of crystallinity. This is further supported by the fact that the crystallinity was similar after borohydride reduction (Sample 3) and TEMPO, periodate oxidation and borohydride reduction (Sample 8).

The swelling propensity of the modified fibres in water was quantified by measuring the WRV at pH 2 and 8.5 (Figure 3). The WRV values of the low-charged fibres, i.e. dialdehyde cellulose (Sample 2), dialcohol cellulose (Sample 3) and aminated cellulose (Sample 4), did not differ significantly from the value for the reference. The slight increase in the WVR of dialcohol cellulose (Sample 3) and aminated cellulose (Sample 4) could be explained by the decrease in crystallinity, and by the presence of hydroxyl groups which are more prone to sorb water. Furthermore, Mihranyan, et al. (2004) studied the effect of crystallinity and surface area on the uptake of moisture in cellulose powders. It was found that cellulose powders with higher crystallinity showed lower moisture adsorption at relative humidity below 75%. The XRD pattern is also dependent on the amount of water present in the sample (Agarwal, Ralph, Baez, Reiner, & Verrill, 2017; Manjunath, Venkataraman, & Stephen, 1973). The apparent increase in crystallinity of dicarboxylic acid cellulose (Sample 5) could therefore possibly be explained by higher water content in the sample which translates in a broadening of the spectrum and this effect is also observed for dialcohol cellulose (Sample 3). TEMPO-oxidised fibres (Sample 6) and dicarboxylic acid cellulose (Sample 5) showed a much higher WRV, especially at high pH, i.e. when the carboxylic acids are deprotonated. It is known that the presence of carboxylic acids groups increases the osmotic pressure as a consequence of the difference in concentration between the mobile ions inside the fibre wall and the exterior solution (Grignon & Scallan, 1980) and, it has also previously been reported that WVR values increase with increasing carboxylate content (Laine, 8

Lindstrom, Nordmark, & Risinger, 2002). When a combination of TEMPO and periodate oxidation (Sample 7) and of TEMPO and periodate oxidation and borohydride reduction (Sample 8) was performed for the fibres, the WRV increased, despite the fact that the charge density decreased (Table 1). This is probably due to a combination of two different processes. First of all, the increased number of charges increases the swelling due to an increased osmotic pressure, and, as earlier discussed, a decrease in crystallinity increases the water absorption even further, resulting in the high swelling of Samples 7 and 8 although the charge density was somewhat lower.

SEM images of the modified fibres are shown in Figure 4. The fibres have a smaller width after periodate oxidation (Figure 4.2), and even more after periodate oxidation followed by amination (Figure 4.4). Dialcohol cellulose fibres are known to result in a good sheet consolidation after a significant degree of modification (Larsson et al., 2014) and this densification indeed starts to be visible in Figure 4.3. The sheets made from dicarboxylic cellulose fibres (Figure 4.5) are very dense as a consequence of large swelling of the fibres (Fält, Wågberg, & Vesterlind, 2003), and even a film formation is observed between the fibres. An even greater increase in density was observed when dialdehyde cellulose and dialcohol cellulose modifications were combined with TEMPO oxidation, as shown in Figures 4.7 and 4.8 respectively. This densification occurred even though the charge density of the fibres was lower than that of the TEMPO-oxidised fibres. This densification of the dry fibres can be explained by their swollen state of the fibres, i.e. they have still a high WRV at a lower charge density due to the change in chemical structure.

3.3. Mechanical properties of sheets made from modified fibres Handsheets made of modified fibres were tested in tension to evaluate the effect that the modifications of the fibres had on the mechanical performance of the sheets (Figure 5). After periodate oxidation (Sample 2), there was a slight decrease in the strain-at-break of the sheets, presumably as a consequence of crosslinks formed between the aldehydes and hydroxyl groups (Larsson et al., 2008). The presence of cross-links is further supported by the observed increase in wet tensile strength (Table 2). When dialcohol cellulose (Sample 3) was produced, the density increased from 585 to 720 kg/m3 (Table 2), together with an accompanying increase in strain-at-break. This effect has previously been reported by Larsson et al. (2014, 2016) and Zeronian et al. (1964) (using cotton linters), and explained by an amorphous shell of water-linking dialcohol cellulose (surrounding a core of crystalline cellulose) that provides high flexibility and molecular mobility, and improves the contact between the fibres during sheet consolidation. Presumably this molecular mobility facilitates nanofibril slippage when the fibres are strained. It should be noted that only 11% of the C2–C3 bonds were cleaved in the present study. In order to achieve a higher strain-at-break, a greater degree of modification is needed. After periodate oxidation followed by reductive amination (Sample 4), no significant increase in density or ductility was observed, which suggests that the position and the chemical environment of a 9

hydroxyl group after C2–C3 cleavage affects the interaction with neighbouring cellulose chains, as well as with water, and therefore also the plastization effect of the modified cellulose material is missing at least in the temperature intervals used in the present work. Despite the open structure of dicarboxylic acid cellulose (Sample 5) and its strong interaction with water, no increase in strain-atbreak was achieved. In fact, these fibres were very brittle and stiff (Figure 5), together with film formation between the fibres (Figure 4.5). This film formation may restrict the fibre network and thereby make the handsheets stiff and brittle. In order to fully compare these results with the earlier prepared dialcohol cellulose it would be necessary to perform this study at higher degrees of oxidation and with individual nanofibrils. The difference in behaviour between dicarboxylic acid cellulose and dialcohol cellulose is indeed very interesting and can arise both from polymer entanglement and from a difference in hydrogen bonding within the dicarboxylic acid cellulose film and the dialcohol cellulose film formed between the cellulosed fibres/fibrils in these different networks. Luo et al. (2013) recently showed that, when a composite of regenerated cellulose and polyacrylic acid is made, a semi-interpenetrating network is formed. These authors also claim that there is a strong interfacial interaction due to hydrogen bonds which determines the properties of the films. These interactions restrict segments of the polymer chain and lead to an increase in stiffness and strength.

When TEMPO-mediated oxidation was performed for the starting fibres (Sample 6), an increase in both tensile strength and strain-at-break was observed. This is generally known to be due to increased swelling of the fibres accompanied by the introduction of charges (Lindström & Carlsson, 1980). When the introduction of dialdehydes was combined with TEMPO-oxidation to introduce carboxyl groups (Sample 7), the strain-at-break decreased as a consequence of the introduction of covalent crosslinks. The existence of these cross-links is further supported by the observed increase in the wet tensile strength of sheets made from these fibres (Table 2). Interestingly, when dialcohol cellulose was combined with TEMPO-mediated oxidation (Sample 8) the strain-at-break decreased from 4 to 3.1%, while the tensile strength increased from 24 to 81 MPa. Figure 5 shows that there was in all cases an increase in the tensile strength after the different modifications, which suggests that depolymerisation was not extensive.

A decrease in crystallinity, i.e. an increase in the content of amorphous cellulose, contributes to an increase ductility (Ward, 1950). As shown in the supporting information Table 1, the crystallinity indices of the modified fibres are approximately the same, but only dialcohol cellulose exhibited a significant increase in strain-at-break. At higher degrees of modification, the strainability of dialcohol cellulose increases even further. Larsson et al. (2014) described an intact fibril core with a high degree of crystallinity surrounded by an amorphous shell of dialcohol cellulose which provides high flexibility and molecular mobility and improves the contact between the fibres during consolidation and drying. The presumed reason for this is that the dialcohol cellulose is well hydrated but still 10

bonded to the crystalline core. It is also important to stress that the dialcohol cellulose has bulk properties that are very suitable for forming highly ductile materials. Obviously the other cellulose derivatives, on the surface of the fibrils, do not have comparable properties and are therefore not able to induce the necessary properties to the cellulose based materials.

As shown in Table 2, only the dialdehyde cellulose (Sample 3) and the dialdehyde cellulose combined with carboxylic acids (Sample 7) possessed a significant wet strength. The wet strength of dialdehyde cellulose in combination with TEMPO was 16 MPa. 4. Conclusion Cellulose fibres were modified using combinations of different chemical modifications, with the aim of understanding the relation between chemical structure and mechanical properties. The surfaces of the cellulose nanofibrils constituting the fibres were modified to provide different functional groups and create a shell that surrounds each fibril. At the same degree of modification of the fibres, it was found that the functional groups in the shell largely determine the behaviour and the properties of the fibres and handsheets made from the modified fibres. Highly charged fibres showed a great increase in water-holding capacity, but, when a combination of TEMPO and dialcohol cellulose (Sample 8) was prepared, the water holding capacity increased to the same level as that of dicarboxylic acid cellulose (Sample 5), despite a lower charge density. Furthermore, despite the open molecular structure of dialcohol cellulose (Sample 3) and dicarboxylic acid cellulose (Sample 5), the fibres and handsheets prepared from them were different in terms of densification, water holding capacity and mechanical properties.

Only a combination of periodate oxidation followed by borohydride reduction, i.e. partial conversion to dialcohol cellulose (Sample 3), resulted in ductile handsheets with a strain-at-break of 4% and a Young´s modulus of 3.5 GPa, but, when carboxylic acid groups were introduced and combined with dialcohol cellulose the typical ductility decreased to 3.1% while the Youngs´s modulus increased to 6.8 GPa. On the other hand, handsheets prepared after partial conversion to dicarboxylic acid cellulose were very stiff with a Young´s modulus of 9.6 GPa, but only 1% of strain-at-break. It should, however, be noted that the fibre network has a considerable impact on the ductility of handsheets, and in order to gain a better molecular understanding, the same modifications should preferably be performed to a higher degree of modification followed by the fabrication of nanofibril films to magnify and isolate the effect of the chemical modification.

11

5. Acknowledgements The authors acknowledge VINNOVA, the Swedish Governmental Agency for Innovation Systems, through BiMaC Innovation Excellence Centre for financial support. Lars Wågberg also acknowledges Wallenberg Wood Science Centre.

12

References Agarwal, U. P., Ralph, S. A., Baez, C., Reiner, R. S., & Verrill, S. P. (2017). Effect of sample moisture content on XRD-estimated cellulose crystallinity index and crystallite size. Cellulose, 24(5), 1971–1984. Bledzki, A. K., & Gassan, J. (1999). Composites reinforced with cellulose based fibres. Progress in Polymer Science (Oxford), 24(2), 221–274. http://doi.org/10.1016/S0079-6700(98)00018-5 French, A. D., & Santiago Cintrón, M. (2013). Cellulose polymorphy, crystallite size, and the Segal Crystallinity Index. Cellulose, 20(1), 583–588. Fält, S., Wågberg, L., & Vesterlind, E. L. (2003). Swelling of model films of cellulose having different charge densities and comparison to the swelling behavior of corresponding fibers. Langmuir, 19(19), 7895–7903. Grignon, J., & Scallan, A. M. (1980). Effect of pH and neutral salts upon the swelling of cellulose gels. Journal of Applied Polymer Science, 25(12), 2829–2843. Kasai, W., Morooka, T., & Ek, M. (2014). Mechanical properties of films made from dialcohol cellulose prepared by homogeneous periodate oxidation. Cellulose, 21, 769–776. Khakalo, A., Filpponen, I., Johansson, L. S., Vishtal, A., Lokanathan, A. R., Rojas, O. J., & Laine, J. (2014). Using gelatin protein to facilitate paper thermoformability. Reactive and Functional Polymers, 85, 175–184. Kim, U. J., Kuga, S., Wada, M., Okano, T., & Kondo, T. (2000). Periodate oxidation of crystalline cellulose. Biomacromolecules, 1(3), 488–492. Klemm, D., Heublein, B., Fink, H. P., & Bohn, A. (2005). Cellulose: Fascinating biopolymer and sustainable raw material. Angewandte Chemie - International Edition, 44(22), 3358–3393. Laine, J., Lindstrom, T., Nordmark, G. G., & Risinger, G. (2002). Studies on topochemical modification of cellulosic fibres Part 2. The effect of carboxymethyl cellulose attachment on fibre swelling and paper strength. Nord Pulp Pap Res J, 17(1), 50–56. Larsson, P. A., Berglund, L. A., & Wågberg, L. (2014a). Ductile all-cellulose nanocomposite films fabricated from core − shell structured cellulose nanofibrils. Biomacromolecules, 15, 2218–2223. Larsson, P. A., Berglund, L. A., & Wågberg, L. (2014b). Highly ductile fibres and sheets by core-shell structuring of the cellulose nanofibrils. Cellulose, 21, 323–333. Larsson, P. A., Gimåker, M., & Wågberg, L. (2008). The influence of periodate oxidation on the moisture sorptivity and dimensional stability of paper. Cellulose, 15(6), 837–847. Larsson, P. A., & Wågberg, L. (2016). Towards natural-fibre-based thermoplastic films produced by conventional papermaking. Green Chem., 18, 3324–3333. http://doi.org/10.1039/C5GC03068D Lindström, T., & Carlsson, G. (1980). The effect of chemical enviroment of fiber swelling. Svensk Papperstidning, 85(3), R14–R20. Linvill, E., Larsson, P. A., & Östlund, S. (2017). Advanced three-dimensional paper structures: 13

Mechanical characterization and forming of sheets made from modified cellulose fibers. Materials & Design, 128, 231–240. López Durán, V., Larsson, P. A., & Wågberg, L. (2016). On the relationship between fibre composition and material properties following periodate oxidation and borohydride reduction of lignocellulosic fibres. Cellulose, 23(6), 3495–3510. Luo, H., Hu, J., Zhu, Y., Wu, J.-Y., Zhang, S., Fan, Y., & Ye, G. (2013). Mechanically adaptive cellulose-poly(acrylic acid) polymeric composites in wet-dry cycles. Journal of Applied Polymer Science, 127(1), 675–681. Manjunath, B. R., Venkataraman, A., & Stephen, T. (1973). The effect of moisture present in polymers on their X-ray diffraction patterns. Journal of Applied Polymer Science, 17, 1091– 1099. Marais, A., Utsel, S., Gustafsson, E., & Wågberg, L. (2014). Towards a super-strainable paper using the Layer-by-Layer technique. Carbohydrate Polymers, 100, 218–224. Matsumura, H., Sugiyama, J., & Glasser, W. G. (2000). Cellulosic nanocomposites. I. Thermally deformable cellulose hexanoates from heterogeneous reaction. Journal of Applied Polymer Science, 78(13), 2242–2253. Mihranyan, A., Llagostera, A. P., Karmhag, R., Strømme, M., & Ek, R. (2004). Moisture sorption by cellulose powders of varying crystallinity. International Journal of Pharmaceutics, 269(2), 433–442. Patterson, A. L. (1939). The scherrer formula for X-ray particle size determination. Physical Review, 56(10), 978–982. Potthast, A., Schiehser, S., Rosenau, T., & Kostic, M. (2009). Oxidative modifications of cellulose in the periodate system - Reduction and beta-elimination reactions: 2nd ICC 2007, Tokyo, Japan, October 25-29, 2007. Holzforschung, 63(1), 12–17. Saito, T., Hirota, M., Tamura, N., Kimura, S., Fukuzumi, H., Heux, L., & Isogai, A. (2009). Individualization of nano-sized plant cellulose fibrils by direct surface carboxylation using TEMPO catalyst under neutral conditions. Biomacromolecules, 10(7), 1992–1996. Sarymsakova, A. A., Nadzhimutdinov, S., & Tashpulatov, Y. T. (1998). Chemical transformation in the chains of cellulose dialdehydes and cellulose thers. Chemistry of Natural Compounds, 34(2), 170–174. Scott, J. E., & Harbinson, R. J. (1968). Periodate oxidation of acid polysaccharides inhibition by the electrostactic field of the substrate. Histochemie, 14, 215–220. Segal, L., Creely, J. J., Martin, A. E., & Conrad, C. M. (1959). An empirical method for estimating the degree of crystallinity of native cellulose using the X-Ray diffractometer. Textile Research Journal, 29(10), 786–794. Sihtola, H. (1959). Chemical Properties of Modified Celluloses. Die Makromolekulare Chemie, 35(1), 250–265. Tanaka, R., Saito, T., & Isogai, A. (2012). Cellulose nanofibrils prepared from softwood cellulose by TEMPO/NaClO/NaClO2 systems in water at pH 4.8 or 6.8. International Journal of Biological 14

Macromolecules, 51(3), 228–234. Vuoti, S., Laatikainen, E., Heikkinen, H., Johansson, L. S., Saharinen, E., & Retulainen, E. (2013). Chemical modification of cellulosic fibers for better convertibility in packaging applications. Carbohydrate Polymers, 96(2), 549–559. Ward, K. (1950). Crystallinity of Cellulose and Its Significance for the Fiber Properties. Textile Research Journal, 20(6), 363–372. Wågberg, L., & Björklund, M. (1993). Adsorption of cationic potato starch on cellulosic fibres. Nordic Pulp and Paper Research Journal, 8(4), 399–404. Zeronian, S. H., Hudson, F. L., & Peters, R. H. (1964). The mechanical properties of paper made from periodate oxycellulose pulp and from the same pulp after reduction with borohydride. Tappi, 47, 557–564.

15

Figure 1. Schematic description of the different routes chosen for chemical modification of cellulose fibres.

16

(1) Ref (2) IO-4 (3) IO-4+ BH-4 (4) IO-4+ NH2OH.HCl (5) IO-4+ClO-2

a.u.

(6) TEMPO (7) TEMPO+IO-4 (8) TEMPO+IO-4+ BH-4

10

15

20

25

30

35

2 (degrees) Figure 2. XRD diffactograms of chemically modified cellulose fibres. 4.5

1500

4.0 1250

3.0

1000

2.5 750 2.0 1.5

500

Charge density (eq/g)

Water retention value

3.5

1.0 ND pH 2 ND pH 8.5

0.5 0.0

250

0 Ref (1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Sample number Figure 3. WRV of chemically modified fibres at pH 2 and 8.5 (left axis). Charge density (open circles) is shown on the right y-axis.

17

Figure 4. SEM images of chemically modified fibres. (1) Ref (2) IO-4

80 (5)

Tensile stress (MPa)

(3) IO-4+BH-4

(8)

70

(4) IO-4+NH2OH.HCl/BH-4 (5) IO-4+ClO-2

(7)

60 50

(6) TEMPO (7) TEMPO+IO-4

(6)

40

(8) TEMPO+IO-4+BH-4

(3)

30 (4)

20

(2) (1)

10 0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Tensile strain (%)

Figure 5. Representative stress–strain curves of handsheets made of modified fibres.

18

Table 1. Carbonyl content and charge density of the modified fibres. Carbonyl content

Sample

(mmol/g)

1:Ref

0.03 ± 0.01

60 ± 3

1.36 ± 0.18

2:IO4-

Charge density (μeq/g)

(11% degree of oxidation)

48 ± 4

3:IO4-+BH4-

Nd

42 ± 8

4:IO4+NH2OH.HCl/BH4-

Nd

52 ± 1

5:IO4-+ClO2-

Nd

1352 ± 74

6:TEMPO

Nd

783 ± 27

1.38 ± 0.02

7:TEMPO+IO4-

(11% degree of oxidation)

8:TEMPO+IO4-+BH4-

Nd

646 ± 7 628 ± 2

Nd= non-detectable

Table 2. Summary of physical and mechanical properties of handsheets made of the chemically modified fibres.

Sample number

Density (kg/m3)

Young´s

Tensile

modulus

strength

(GPa)

(MPa)

Strain-at-

Wet tensile

break (%)

strength (MPa)

1:Ref

585

2.1 ± 0.1

15.3 ± 0.7

2.48 ± 0.28

-*

2:IO4-

581

2.2 ± 0.1

23.4 ± 0.8

2.33 ± 0.20

8.4 ± 0.1

3:IO4-+BH4-

720

3.5 ± 0.1

35.7 ± 0.7

4.01 ± 0.11

-*

4:IO4+NH2OH/BH4-

650

2.7 ± 0.1

23.8 ± 0.9

2.44 ± 0.17

2.8 ± 0.1

5:IO4-+ClO2-

971

9.6 ± 0.3

75.0 ± 3.5

1.02 ± 0.05

0.7 ± 0.1

6:TEMPO

767

4.8 ± 0.3

56.4 ± 2.0

3.75 ± 0.19

0.4 ± 0.1

7:TEMPO+IO4-

830

6.4 ± 0.6

63.7 ± 2.0

1.75 ± 0.11

16.0 ± 0.3

8:TEMPO+IO4-+BH4-

914

6.8 ± 0.6

80.7 ± 4.3

3.12 ± 0.28

0.9 ± 0.2

*Too weak to be measured

19