Introduction

In the simplest approach, the mechanical properties of fibre networks are determined by the fibre properties and the interactions between the fibres (Page 1969; Goutianos et al. 2018). The latter are affected by several other parameters, such as relative bonded area and shear strength of the joint. If the fibre–fibre joints are weak, the fibres will separate from each other upon loading before the fibre fails (Pöhler et al. 2020). If, on the other hand, the fibre–fibre joints are strong enough, the fibre strength will, to a larger extent, be the limiting factor creating fibre fracture rather than joint failure (Page 1969; Davidson 1972). For a certain type of fibres, the mechanical properties of the network can hence be increased by increasing the number of fibre–fibre joints (Mao et al. 2017) or by increasing the adhesive properties of the fibres (Borodulina et al. 2012; Mao et al. 2017; Kouko et al. 2019).

Several different ways of enhancing fibre adhesion have been proposed in previous studies, for example, the attachment of carboxymethyl cellulose (CMC) and/or cationic starches to fibres (Strand et al. 2017; Larsson et al. 2018; Leib et al. 2022), the adsorption of xyloglucan (XG) and cellulose nanocrystals (CNC) (Doineau et al. 2020), the mechanical beating of pulp (Motamedian et al. 2019), the addition of single component strength additives (Jocher et al. 2015; Lindström et al. 2018), the addition of cellulose nanofibrils (CNFs) (Hollertz et al. 2017), and the Layer-by-Layer (LbL) self-assembly technique (Decher 1997; Wågberg et al. 2008; Wågberg 2000; Li et al. 2018). The latter method, considered here, can be used to modify the surfaces of charged materials by consecutively adsorbing layers of oppositely charged polyelectrolytes (PEs). It was previously shown that the LbL-technique greatly impacts papers' mechanical properties and performance (Wågberg et al. 2002; Pelton 2004; Mocchiutti et al. 2011). For instance, Marais et al. (2014) show that a Layer-by-Layer treatment on the fibres improves the maximum tensile index from around 25 Nm/g for the reference to 70 Nm/g and the strain at break from 2% to an astonishing 6.5%.

In standard paper production, the different treatments of the fibres, such as beating, pressing, and drying, affect both the strength of the fibre–fibre joints as well as their volumetric concentration (Delgado-Fornué et al. 2011). By using the LbL-technique, beating of the fibres can be avoided. This results in higher press solids on the paper machine and consequently a lower steam consumption in the drier section, which in turn also allows for a higher machine speed (Marais et al. 2016). An increased network density by wet pressing might increase the molecular contact between the fibres and, hence, the joint strength (Torgnysdotter and Wågberg 2003). More significantly, it will increase the volumetric concentration of the joints. With more joints to carry the load, the network becomes stronger, inducing more fibre fractures (Deogekar and Picu 2018).

Previous studies examine the effect of densification, showing that it indeed improves the network's mechanical properties. Forsström et al. (2005) experimentally investigated pulps of different yields and found that increased yield leads to higher fibre flexibility. This increased flexibility leads to more efficient network packing and subsequently higher density, resulting in improved mechanical properties. Sehaqui et al. (2013) also observe increased properties through increasing density by adding nanofibrillated cellulose to the pulp resulting in the formation of joints between the nanofibres themselves and between nanofibres and paper fibres. Bergström et al. (2019) developed a network model and studied the deformation mechanisms in networks, concluding that the deformation is bending-dominated at low densities and then transitions to a stretching mode of the fibres when density increases, hence improving the mechanical performance of the network. High-density networks have more recently also found application in the production of high-strength moulded fibre materials, such as food trays, formed by moulding wet fibres into 3D shapes via hot-pressing and compression moulding (Nilsson et al. 2010; Oliaei et al. 2021b). High-yield pulps are particularly favoured for creating high-density papers using hot-pressing techniques (Vesterlind and Höglund 2006; Wang et al. 2018). A recent study by Joelsson et al. demonstrates the feasibility of developing high-density paper with improved dry and wet strength using hot pressing (Joelsson et al. 2020).

For paper products, the dimensional stability determined through hygroexpansion reflects a material's ability to withstand deformation amidst moisture fluctuations. Multiple investigations have been devoted to study the hygroexpansion of fibres and paper, emphasising its importance in the paper industry (Salmén et al. 1987a; Uesaka 1991; Uesaka et al. 1992; Uesaka and Qi 1994; Larsson and Wågberg 2008; Urstöger et al. 2020; Lin et al. 2022). Analysing materials under changing moisture levels helps quantifying the stresses imposed on fibres during the drying process (Salmén et al. 1987b; Lindner 2018; Jajcinovic et al. 2018).

In the present work, we have studied the combined effect of increasing the fibre–fibre joint strength by modifying the adhesive properties of the fibres and increasing the density of networks by hot pressing on the overall mechanical properties of the prepared paper sheets. By modifying the surface of fibres from an unbleached kraft pulp and a bleached softwood Chemo-ThermoMechanical Pulp (CTMP) with the LbL-technique, the joint properties were enhanced to a point where the fibre strength became the limiting factor. Different densities of laboratory prepared sheets were achieved by changing the pressure and temperature during the pressing of the wet sheets. The thickness and density of the sheets were evaluated before mechanical testing, and the mechanical properties of the paper sheets were studied by tensile testing followed by Scanning Electron Microscopy (SEM) imaging of the fracture zones. Additionally, hygroexpansion measurements were used to evaluate the effect of changing the relative humidity on the fibre network. To the best of the authors’ knowledge, this is the first investigation of controlled fibre surface modifications combined with hot pressing, in order to quantify if any effect of the LbL-treatment remains in high-density networks.

One key industrial implication of this research lies in the improved properties of the fibre–fibre joints within the paper sheet, which contribute to the development of high-performance papers and provide possibilities for broader use of bio-based fibres (Oliaei et al. 2021a; Mattsson et al. 2021; Samantray et al. 2022). This study highlights that the interfacial adhesion within these joints significantly influences the overall network performance, even at higher paper densities.

Materials and methods

Materials

Unbleached kraft and bleached CTMP softwood pulp fibres were provided by Stora Enso, Sweden. The bleaching was performed using hydrogen peroxide in one step, which is a lignin preserving bleaching process. The pulp fibres were washed with deionised water before being used to remove excess metal ions and dissolved colloidal substances. The total charge density of the pulp fibres was measured using Conductometric titration (Metrohm 856) according to the SCAN-CM 65:02 standard (Katz et al. 1984). The fibre length of the two pulps was measured using an optical fibre analysis (FiberLab™), see Fig. SI.1 in the Supplementary Information for the probability distribution. The Kappa number of the fibres was measured according to ISO 302:2015, and the results presented are the average of two measurements. Table 1 summarises the characteristics of the fibres.

Table 1 Characteristics of the fibres
Full size table

Sodium chloride (NaCl), Polyallylamine Hydrochloride (PAH) (molecular weight of 17 kDa), and Hyaluronic acid (molecular weight of 1.5–1.8·106 Da) were all purchased from Sigma Aldrich Sweden AB and used as received.

Fibre modification

Firstly, the never dried, unbleached kraft and the bleached CTMP fibres were suspended in a 5 mM NaCl solution for about 30 min, allowing the fibres to reach an equilibrium swelling. Then, the fibres were modified using the Layer-by-Layer technique by alternating the adsorption of the cationic polyelectrolyte (PAH) and the anionic polyelectrolyte (HA). The first PAH layer was adsorbed by adding 5 mg/g of fibres to the 5 mM NaCl-containing fibre suspension. The latter allows for a better distribution and adsorption of the polyelectrolytes. The adsorption time was 10 min, and an overhead stirrer was used to properly distribute the polyelectrolyte in the fibre dispersion. Next, the suspension was dewatered using Büchner funnel filtration. The excess PAH was washed away by immersing the fibres in a new 5 mM NaCl solution for 5 min under mechanical stirring, followed by another dewatering step. Similarly, the next layer of HA was adsorbed by adding 10 mg of HA/g of fibres, following the same steps. These modification steps were performed several times to finally attain the formation of 10 bilayers of PAH and HA on the surface of the fibres. The chosen addition levels were based on the previous study by Marais et al. (2014).

Nitrogen analysis

Nitrogen analysis using the ANTEK equipment (MultiTek by PAC, Houston, Texas, USA) was performed to measure the amount of polymer adsorbed onto the LbL-treated fibres. This technique is based on chemiluminescence and provides the amount of nitrogen present in the sample in counts, which can be translated into the amount of adsorbed polymer using a suitable calibration with the respective polymers. Calibration curves were prepared for the polyelectrolytes as well as the unmodified pulps and are presented in the Supplementary Information. The contribution of the nitrogen from the fibres themselves were naturally deducted from the measurements.

Sheet preparation

The paper sheets were prepared using three different methods to achieve a range of three densities (low, medium, and high). The target grammage for all the sheets was around 160 g/m2. For the low density, the paper sheets were prepared using a Rapid Köthen sheet former (Paper Testing Instruments – PTI, Pettenbach, Austria). The sheets were dried at 93 °C for 15 min under the lateral pressure of 1 bar. The medium and high-density sheets were also formed using Rapid Köthen. However, another setup was used for the pressing and drying, where a forming unit was connected to a hydraulic hot press (Fontijne). This allows us to dry the wet sheets by applying heat and pressure while simultaneously removing the water through suction. For the medium and high densities, the wet sheets were sandwiched between two metal wires to protect the sheets and allow for an even distribution of the applied pressure. Then, they were dried at 150 °C for 15 min, under 0.8 MPa and 9.5 MPa pressures, respectively. The obtained paper sheets were circular with an area of 314 cm2.

Thickness measurements

The thickness of each sheet was measured using the STFI structural thickness gauge (TJT Teknik, Sweden), which measures the thickness over a strip with a 0.1 mm sampling length and 1 μm accuracy according to SCAN-P 88:01. To account for uniformities across the plane of the sheet at least five different locations were measured, yielding an average thickness.

Tensile testing

The materials were subjected to uniaxial tensile testing to determine the mechanical properties, according to ISO 1924–2:2008. Specimens were prepared with a width of 15 mm and a testing length of 100 mm. The materials were uniaxially strained at a 100 mm/min testing rate in a ZwickRoell uniaxial testing machine with a 1 kN load cell while recording the applied force throughout the test. The strain was recorded optically using the VideoXtens software, where two points were tracked on the surface during the test. The samples were conditioned at standard climate (50 %RH, 25 °C) for at least 24 h before testing. Upon testing (not performed at standard climate), they were, for practical reasons, taken out of the climate-controlled room one by one and tested within minutes. Marin et al. (2020) investigated the effect of moisture changes on the mechanical properties of paperboard, and assuming the material considered here behaves similarly, the presently used time frame between conditioning and testing should not have a significant effect on the tested properties. To ensure comparability between samples with slight variations in density in the same family of samples, the testing data was quantified by the specific stiffness (tensile stiffness index) and the specific strength (tensile index). The region used for the tensile stiffness evaluation was determined by finding the inflexion point of the force-strain curve, see Fig. SI.3 in the Supplementary Information for an example. To further evaluate the validity of the results, the determined properties were evaluated with an analysis of variance (ANOVA), see Supplementary Information for details.

Scanning electron microscopy

Field-Emission Scanning Electron Microscopy FE-SEM (Hitachi S-4800, Japan) was used to image the paper sheets after fracture in the tensile test. Small pieces were cut from the paper sheets at the fracture zone, and the samples were coated, prior to imaging, with a thin carbon layer using a Cressington 208HR sputter coater.

Dimensional stability

To understand and distinguish the effects the densification and LbL-treatment have on the networks, the dimensional stability was investigated. The degree of restrictions imposed on the fibres during drying can be quantified by subjecting the materials to moisture changes and by measuring the resulting deformation. When the moisture content increases, the fibres start to swell. However, the total deformation of the network with the given fibre orientation will depend on fibre properties and number of joints and their properties. Increasing the number of fibre–fibre joints with densification and increasing joint strength in the wet state via LbL-treatment prevent fibre mobility during network formation and shrinkage. Consequently, the combination of densification and the addition of Layer-by-Layer (LbL) coatings leads to different loading histories on the fibres during the drying process. These distinct differences will significantly influence the final material properties, which indeed need to be quantified.

The materials were subjected to a moisture cycle while measuring the resulting expansion of the test specimens and recording the moisture content. The hygroscopic strain was evaluated as

$${\varepsilon }_{h}=\frac{l-{l}_{0}}{{l}_{0}},$$
(1)

where \({l}_{0}\) is the original sample length and \(l\) is the current length at a specified moisture content. The hygroscopic expansion coefficient can then be estimated as

$$\beta ={\frac{{\varepsilon }_{h}}{m_{c}}},$$
(2)

where \({m}_{c}\) is the moisture content in the sample, defined as the ratio of the mass of water in the sample and the total mass of the sample. The samples were subjected to a moisture cycle of 50–20-50–87-50 %RH, using the Innventia HygroexpansivityTester (Innventia HT) (Antonsson et al. 2010). Samples were prepared by cutting 15 by 100 mm strips and conditioning at 50 %RH for at least 24 h before testing. Each moisture level was maintained for at least 24 h, and the sample length was regularly recorded. To determine the accompanying moisture content, the samples were weighed at each moisture level using a scale with 0.0001 g accuracy and then subjected to oven drying at 105°C for 6 h to determine the dry weight, according to ISO 287:2009. Only the most interesting (contrasting) materials were tested: high/low density and 0/5 bilayers. To further investigate the dimensional stability of the materials, the tensile tests were repeated for the samples subjected to hygroexpansion following the same procedure as previously described in Tensile testing. However, since the samples were already cut to a length of 100 mm, the test span in these tensile tests was only 80 mm. The tensile stiffness and tensile indexes were determined as before and compared to the values before subjection to the moisture cycle.

Results

Adsorption of polyelectrolytes to the different fibres

The total amount of polymer adsorbed onto the fibres was determined using nitrogen analysis. Figure 1 shows that the adsorbed amount increases almost linearly with the addition of each layer and that both the unbleached kraft fibres and the bleached CTMP fibres can adsorb quite similar amounts of PAH and HA. The amount of polymer adsorbed after each layer is presented in Fig. SI.2 in the Supplementary Information.

Fig. 1
figure 1

Total adsorbed amount of PAH and HA on the fibres following the consecutive LbL treatment with polyelectrolytes as determined with nitrogen analysis. 10 layers in this graph is equivalent to 5 bilayers

Full size image

Sheet characterisation

The sheet thickness was measured along each sheet to characterise the prepared paper sheets and examine the effectiveness of the techniques used to obtain different densities. Figure 2a shows a thickness profile of a representative sample illustrating the uniformity of the measured thickness along five different specimens. Three different processes were used to prepare the paper sheets, as explained in the experimental part, resulting in three main ranges of densities. The obtained densities are presented in Fig. 2b, where the materials made from unbleached kraft and bleached CTMP exhibited similar results. The low densities of the papers prepared by standard Rapid Köthen sheet dryer ranged from 499 kg/m3 for the sheets from the bleached CTMP fibres to 598 kg/m3 for the sheets from unbleached kraft fibres. The medium and high densities were obtained by hot pressing the paper sheets at higher temperatures and under high pressures. This treatment resulted in densities of 817 kg/m3 and 1040 kg/m3, respectively.

Fig. 2
figure 2

(a) Representative thickness profile of one paper sheet. (b) The densities obtained for the sheets made from the untreated fibres. The error bars indicate one standard deviation (based on the spread in thickness measurements)

Full size image

The density was also measured for the papers made with LbL-treated fibres, and as shown in Fig. 3, in most cases, the surface treatment does not significantly affect the density of the prepared paper sheets. However, a slight increase in the density can be seen for the low-density materials with a higher number of adsorbed layers for the sheets from the unbleached kraft fibres.

Fig. 3
figure 3

Average densities of the paper sheets with an increasing number of adsorbed LbLs (bilayers) for (a) the unbleached kraft and (b) bleached CTMP. Multiple data points for the same material configuration indicate separate batches. The error bars indicate one standard deviation (based on the spread in thickness measurements)

Full size image

Mechanical properties

The mechanical properties resulting from uniaxial testing are summarised in Fig. 4 below. The properties are determined as described in the experimental part: the tensile stiffness index quantifies the stiffness of each material, while the tensile index is a measure of strength. The presented values are based on data for 5–12 samples of each material; due to slipping in the grips during testing, some samples had to be discarded. Typical curves from the tensile testing are presented in the Supplementary Information in Figs SI.4 and SI.5. Given the complexity of the data and the multiple treatment levels applied to the samples, an ANOVA is employed to discern the statistical significance of the observed differences in mechanical properties, presented in the Supplementary Information in Tables SI.12-SI.15.

Fig. 4
figure 4

Determined (average) specific properties for different density levels and added Layer-by-Layers, for (a) unbleached kraft pulp and (b) bleached CTMP. The error bars indicate one standard error (based on the spread in thickness measurements)

Full size image

The specific stiffness (tensile stiffness index) improves on average by about 20–60%, with the addition of LbLs, for both bleached CTMP and unbleached kraft materials, depending on number of layers and density level (see Table SI.3 in the Supplementary Information). Moreover, the specific strength (tensile index) is increased on average by at least 15% and up to 126% for unbleached kraft materials and 108% for bleached CTMP materials (see Table SI.7 in the Supplementary Information for details).

After the tensile test, the fracture zones of the specimens were imaged using SEM to investigate the failure mechanism of the modified paper sheets compared to the unmodified ones. The SEM images in Fig. 5 aid in visualising the fracture mechanisms that occur during the tensile test (for fracture images at macro scale, see Fig. SI.6 in the Supplementary Information). First, as expected, the fibres are visibly much more closely packed in the high-density sheets than in the low-density sheets. In other words, the densification process resulted in an increased volumetric concentration of fibre–fibre joints in the network, which is even more amplified by the surface treatment of the fibres. At low density and no bilayers (Fig. 5a and e), there is extensive pull-out of fibres at the fracture zone for both unbleached kraft and bleached CTMP samples. However, less pull-out is observed for the LbL-treated samples with 5 bilayers of PAH and HA already at low density (Fig. 5b and f). Additionally, fibre fracture is increased for the sheets from the treated fibres. For the high-density sheets, the SEM images show sharper, well-defined fracture lines and extensive fibre breakage (Fig. 5c and g), which is further increased by the addition of the 5 bilayers of polyelectrolytes (Fig. 5d and h). The same effect is detected for both unbleached kraft and bleached CTMP sheets, apart from some slight differences, which can be explained by the difference in fibre properties, mainly fibre length and the number of fibres/g for the different fibre types.

Fig. 5
figure 5

SEM images showing the fracture zone after tensile tests. Images (a-b) and (e–f) correspond to the low-density materials with 0 and 5 bilayers of PAH and HA. Images (c-d) and (g-h) correspond to the high-density materials with 0 and 5 bilayers of PAH and HA

Full size image

Dimensional stability

The results from the hygroexpansion tests are summarised in Fig. 6, showing the hygroscopic strain versus moisture content at different levels of relative humidity. The presented data are based on 2–3 samples per material. The hygroexpansion coefficient \(\beta\), according to Eq. (2) can be identified by observing the slope. However, when the moisture content is increased from the starting value of 50 %RH, the deformation depends highly on the degree of dried-in strains during the first cycle, which significantly affects the slope. The release of dried-in strains evidently interferes with the evaluation of the hygroexpansion coefficients. Consequently, they were evaluated during the drying phase to avoid this interference. The difference in dried-in strains can partly be seen in the (decrease in) slope between 50 and 87%, but a better measure is the residual strain after the completed moisture cycle (the final strain value after the relative humidity has decreased from 87 to 50%). Higher-density sheets showed slightly lower hygroexpansion coefficients (slope between 20 and 50%), and the dried-in strains released were significantly higher (seen by a large negative strain after completing the moisture cycle and returning to the original moisture content), particularly for the papers from bleached CTMP. Adding LbLs further lowers the slope, suggesting an increase of dried-in strains and, consequently, a better strain transfer at the joints. This is reflected in the tensile stiffness increase with added bilayers (see Fig. 4). Furthermore, the increase of mechanical properties is steeper for the bleached CTMP than the unbleached kraft, which agrees with the higher levels of dried-in strains.

Fig. 6
figure 6

Hygroscopic strain versus moisture content determined during hygroexpansion measurements for (a) unbleached kraft and (b) bleached CTMP. The reported values correspond to the RH values noted in the figure. The moisture cycle begins from the left: 20–50-87–20 (%RH)

Full size image

In order to evaluate how the dried-in strains affected the elastic modulus of the sheets, mechanical measurements post the moisture cycling were performed. The results from the tensile test for the samples after the moisture cycling are shown in Fig. 7. The presented properties were determined using the same procedure as for the values reported in Fig. 4. However, due to noise in the measured data, a linear regression was used to determine the slope up to 25% of the maximum stress. As before, the tensile stiffness index was then found by determining the maximum slope.

Fig. 7
figure 7

Determined (average) specific properties for low and high density sheets depending on 5 added bilayers and compared for samples before and after the hygroexpansion cycle (the latter indicated with (H)) for (a) unbleached kraft and (b) bleached CTMP. The error bars indicate one standard error (based on the spread in thickness measurements)

Full size image

Discussion

Previous investigations (Marais et al. 2014) have shown that the Layer-by-Layer treatment of pulp fibres is based on a consecutive addition of different polymers to saturation which leads to a linear increase in the adsorbed amount of polymers. This correlates well with our results in Fig. 1, where the adsorbed amount of Polyallylamine Hydrochloride and Hyaluronic acid increases almost linearly with each addition. Marais et al. (2014) also show that LbL-treatment with PAH and HA increases the tensile index and strain at break by a factor of three. To amplify the effect of the surface treatment of the fibres, a densification process is introduced. This process aims to optimise the modified fibre–fibre joints by increasing their volumetric concentration within the fibre network. However, as the breakage mechanism changes from fibre pull-out to fibre breakage with the increased density, it might be anticipated that the LbL-treatment will have a lower relative influence on the mechanical properties of the sheets at higher density.

The adsorption of the polyelectrolytes onto the surface of the fibres had, as such, no significant effect on the density of the prepared sheets, as shown in Fig. 3 (see also Table SI.1 in the Supplementary Information) and further supported by an ANOVA of the density levels, where no clear trends of significant differences appeared (see Tables SI.10-SI.11 in the Supplementary Information). This suggests that fibre consolidation was not affected similarly to what is achieved with a refining (Motamedian et al. 2019). The LbL-treatment primarily enhances the adhesive properties of fibre–fibre joints without altering the fibre wall flexibility (Marais et al. 2015). Consequently, the network's strength increases, shifting the primary cause of network failure from joint failure to fibre fracture. SEM images (Fig. 5) show that low-density material failure is mainly caused by fibre pull-out. However, fibre pull-out becomes nearly undetectable at higher density with added layers. The combination of surface treatment and densification significantly enhances the mechanical properties of paper sheets.

At low densities, the average of specific stiffness has an increasing trend with increased number of bilayers, see Fig. 4. However, for the unbleached kraft, the differences are not always significantly noticeable, especially for low densities (see Table SI.12 in the Supplementary Information). In contrast, the bleached CTMP demonstrates significant differences at all density levels, although differences between all levels of bilayers are not as clear (see Table SI.13 in the Supplementary Information). The difference between the results for the two pulp grades can be attributed to the differences in cross-sectional geometry. As previously observed in Brandberg and Kulachenko (2017), the effect of enhancing joint strength on network stiffness is greater for fibres with an open lumen (CTMP), compared to flat, collapsed fibres (kraft).

On the contrary, the specific strength is visibly increased for both unbleached kraft and bleached CTMP materials upon the addition of bilayers, see Fig. 4. The differences are especially noticeable for the low density level, which is also supported by the ANOVA (see Tables SI.14 and SI.15 in the Supplementary Information). At the medium and high density levels, the difference is significant when comparing any number of added bilayers to the unmodified samples (0 bilayers). However, upon comparison of the results for 1, 3, and 5 bilayers, the values are not significantly different, and the trend is occasionally inconsistent. This suggests that enhancing the joint properties has a greater effect on network strength at lower densities when there are fewer fibre–fibre joints, which is consistent with earlier numerical observations (Brandberg and Kulachenko 2017). With a greater number of joints, there is a limitation on how much the joint strength can be increased. After that limitation is reached, the effect on the mechanical performance is lower. It is however very interesting to note that the additives indeed have an effect even at the highest density showing that fibre breakage is not the only mechanism even at these densities.

The achieved improvement of specific tensile strength of 15% up to over 100% can be seen in comparison with a recent study by Negro et al. (2023), which achieved a maximum 31% improvement in tensile index by combining nanofibrillated cellulose in combination with hot-pressing. The 60% maximum increase in specific tensile stiffness in low density sheets approaches the one seen in the studies utilising nanocellulose for enhanced interactions (Sehaqui et al. 2013). Both the improvements in stiffness and strength originate from the enhanced load transfer through increased densification and improved joint strength. However, the strength particularly benefits from the associated improvements and becomes limited by fibre fracture in high density sheet papers. At the highest densities, however, the network stiffness is mainly controlled by the fibre stiffness (Brandberg and Kulachenko 2017) and the observed experimental results showing marginal improvement in stiffness in high density sheets suggests that the fibre stiffness was not significantly changed by LbL-treatment.

Hygroexpansion measurements reveal differences in networks with different densities and surface treatments. The observed slope significantly decreases when examining the effect of increasing moisture above the initial value (50–87 %RH). Drying restrictions lead to residual stresses within the network which are released when fibres absorb moisture and swell. Increased density and added bilayers further restrict network expansion. In the case of the increased density, this is attributed to increasing the number of fibre–fibre joints. As stated in Salmén et al. (1987a), the fibre expansion is restricted in high-density sheets dried under restraint. Increasing the moisture will force the fibres to expand, but the expansion in the cross direction will decrease because of the increased number of joints and imposed constraints. Later, Uesaka (1994) developed a theoretical relationship accounting for the different mechanisms controlling hygroexpansion in fibre networks, which indeed supports the hypothesis that the transverse expansion of fibres is highly affected by increased density.

On the other hand, when changing the joint properties, the network geometry is presumably unchanged. Larsson and Wågberg (2008) found that joint properties did not affect the hygroexpansion of sheets made from unbeaten pulps dried under restraint, which is consistent with the results for low density materials (see Fig. 6). However, the tensile tests show that added modifications also result in stiffer materials (see Fig. 4), suggesting increased fibre fixation in the network. This is reflected in the change in hygroexpansion for the materials with high density when comparing untreated networks to those with 5 bilayers (Fig. 6). During the forming of the sheets, the wet fibre network is dried under restraint, leading to residual stresses. LbL-treatments enhance the joint properties in the wet state (Lingström et al. 2007; Feng et al. 2009; Marais et al. 2014, 2015; Pettersson et al. 2014), further fixing fibres during sheet formation. Consequently, the dried-in strains limit the expansion at increased moisture, resulting in a lower slope. These findings align with the increased mechanical properties, as shown in Fig. 4 and the difference in fracture mechanisms, as shown in Fig. 5.

Furthermore, the moisture cycle affects the specific properties as seen from the tensile tests for the materials subjected to hygroexpansion (Fig. 7). Especially for the unbleached kraft material, the effect of the bilayers on the tensile stiffness index vanishes and the tensile index at 5 bilayers decreases significantly (Fig. 7a). The effect is visible for both the low and high densities. For the bleached CTMP, the effect of added bilayers seems to be somewhat retained after the moisture cycle, although the tensile stiffness index is visibly lower compared to the reference values. The tensile index values are similar to the references (Fig. 7b). This is surprising, considering that the bleached CTMP materials had a higher release of dried-in strains (Fig. 6b) compared to the unbleached kraft (Fig. 6a). However, Salmén et al. (1987b) conclude that there is no fundamental relation between hygroexpansion and mechanical properties, although they are both affected by the dried-in strains imposed during drying. This suggests that the drying procedures affected the kraft fibres to a greater degree, particularly at the higher density, and that a significant fraction of the increased stiffness could be due to straightening the disordered polymer chains within the fibres during constraint drying (Vonk et al. 2023).

Conclusions

Using the Layer-by-Layer (LbL) technique, fibre surfaces were modified by adding bilayers of PAH and HA, enhancing fibre–fibre joint strength. The results show that both LbL-treatment and increased density improved the mechanical properties of the paper sheets. Across all density levels, adding LbLs increased the specific tensile stiffness up to 25% for the sheets from the unbleached kraft fibres and 60% for the sheets from the bleached CTMP fibres, while the specific tensile strength was increased even by 100% for both pulp types.

For the sheets from unbleached kraft, the effect on the tensile stiffness was not always statistically significant, while the results for bleached CTMP were more convincing, which is attributed to the difference in fibres’ cross-sectional geometry and the number of fibres per gram. On the other hand, the effect on the tensile strength was statistically significant for both pulp grades, and across all density levels. However, at higher densities the differences between added number of bilayers were not statistically significant, indicating that no observable improvements could be achieved by increasing number of LbL-treatments.

SEM imaging revealed a transition from fibre pull-out to fibre fracture due to LbL-treatment and densification, indicating that stronger fibre–fibre joints influenced paper properties across all examined density levels. However, since LbL-treatments were efficient even at the higher densities it can be concluded that the fibre strength was not totally dominating even at these densities. Furthermore, hygroexpansion measurements indicate increased dried-in strains for the LbL-treated and high-density sheets. This suggests that greater residual stresses were introduced during drying because of enhanced joint strength and densification, contributing to the improved mechanical performance.

In summary, modifying the fibre surfaces using the LbL-technique, combined with network densification, allows the inherent fibre strength to be almost fully utilised by creating stronger fibre–fibre joints. The constrained drying further enhances the stiffness and strength of the dense paper sheets, with kraft fibres displaying a more pronounced effect than CTMP fibres. These findings have clear implications for industries reliant on high-density papers, such as packaging and specific-use papers. By tailoring the surface properties of the fibres, stronger interactions have been achieved, leading to enhanced utilisation of the inherent fibre properties. In future work, the presented results are suitable for network modelling. The joint properties and density can be investigated as completely separated properties, which will lead to understanding and emphasising the importance of the fibre–fibre joints in fibre-network materials even further.