Structural properties of coated papers with cellulosic nanofibres using different metering systems and drying technologies

Progress in Organic Coatings 179 (2023) 107543 Available online 20 March 2023 0300-9440/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Structural properties of coated papers with cellulosic nanofibres using different metering systems and drying technologies Úrsula Fillat * , Priscilla Vergara, Juan Carlos Villar, Nuria G´ omez Forest Research Centre (INIA, CSIC), Crta. de la Coru˜ na, Km 7.5, 28040 Madrid, (Spain) ARTICLE INFO Keywords: Cellulose nanofibrils Nanocellulose coating Air permeance Paper packaging Coating metering Drying technology ABSTRACT Cellulose nanofibres (CNFs) can improve the quality of cardboard packaging. This work evaluates the ability of CNFs to impart barrier properties to commercial paper used in packaging. Three CNFs were tested: mechanical (m-CNFs), by TEMPO mediated oxidation (T-CNFs) and carboxymethylated (c-CNFs). Two metering systems (wound rod and blade micrometer) were used to apply one, five and 10 layers of CNFs suspensions, and two drying methods (hot air jet and contact with a hot polished surface in a speed dryer) were evaluated. The quality of the CNFs coated papers was measured by structural (thickness, roughness and air permeance) and optical properties (gloss) and visual appearance. c-CNFs coatings obtained the lowest air permeance values, from 0.1 to 0.01 μ m/Pa⋅s depending on the number of coating layers, and were below those of commercial starch. T-CNFs also reduced air permeance at low weights, reaching 1.1 μ m/Pa⋅s with one layer and 0.1 when five layers were applied. Five layers of m-CNFs were needed for good results. SEM images showed good coverage at low coat weight in c-CNFs and T-CNFs, whereas m-CNFs was unable to cover the base paper. With increased layers of mCNFs, an adequate film was formed. c-CNFs had the highest fibrillation degree and acid group content, and was the CNFs coating with most potential as a coating for paper. 1. Introduction In 2018, the recycling rate of plastic packaging waste for the EU27 was 41.4 %, compared to a much higher rate of 84.2 % for paper and cardboard packaging [1]. The diversity of plastic waste further complicates the recycling process, leading to lower quality and a higher cost of recycled material compared to unrecycled material [2]. This means that most plastic waste is incinerated or ends up in landfill, with the consequent associated environmental problems [3]. Paper and cardboard can successfully replace plastic packaging, because they are natural, renewable, biodegradable materials and are more easily recycled than plastic. Paper and paperboard are permeable to water, aqueous solutions, greases and gases. They can also acquire barrier properties and extended functional performance through lamination with plastics or aluminium foil, among other materials [4], but this slows biodegradation and makes recycling difficult. As an alternative, coated papers/cardboards that can be fully recycled or composted have been proposed [5,6], although their properties do not reach those obtained with extrusion or lamination [6]. The use of biopolymers, such as cellulose nanofibres (CNFs) obtained from biomass, improves the biodegradability and recyclability of paper products [7]. CNFs are produced when cellulose fibres are subjected to high shear forces [8]. Because disintegration methods are energyintensive processes [9], various alternatives including mechanical, enzymatic and chemical pre-treatments (TEMPO mediated oxidation and carboxymethylation) have been applied to reduce energy consumption [10–12]. CNFs properties depend on the cellulose source [13], pre-treatment, and the intensity of the final disintegration method [14]. In papermaking, CNFs have been added to pulp slurry as flocculants and dry strength agents [15–20], on the paper surface in the size press [21–23], and also to improve the properties of coating and sizing emulsions in the manufacture of printing papers [24,25]. More recently, CNFs have been studied as coatings to improve paper barrier properties of packaging paper [7,26–32]. To assess the viability of CNFs coated paper, it is necessary to evaluate the runnability of CNFs in the different scenarios involved in surface coating and paper converting. Factors affecting final coating quality are associated with the properties of the base paper, the coating emulsion, the metering system and coating and drying operating variables [5,33]. Coating technology is divided in two stages: the application of coating to cover paper surface * Corresponding author. E-mail address: [email protected] (Ú. Fillat). Contents lists available at ScienceDirect Progress in Organic Coatings journal homepage: www.elsevier.com/locate/porgcoat https://doi.org/10.1016/j.porgcoat.2023.107543 Received 1 December 2022; Received in revised form 1 March 2023; Accepted 8 March 2023 Progress in Organic Coatings 179 (2023) 107543 2 and metering of coating to a desired coat weight. The roll applicator is the typical system for coating, while blade and rod technologies are more frequently used to meter a specific amount of coating onto the paper [33]. Both coating technologies are successful for applying coatings with solids content higher than 10 % or even as much as 70 %. In the literature on CNFs applications, various methods have been tested to obtain the desired CNFs coat weight, including casting on the base paper or producing a top layer using a dynamic hand sheet former [32], spraying [34], size press [22], rod coating [22,35], cylindrical laboratory coater [23], and roll to roll [30]. These studies revealed limitations to obtain a homogenous coating surface mainly associated with the high viscosity of CNFs suspensions, even at low solids content, usually 1–3 % o.d. (oven dry weight). After the coating stage, drying wet CNFs films is also critical because dewatering rates of the coating slurry affect the quality of the final coated layer. Several drying methods have been applied, including contact drying [22,23], hot air [18,25,36], infra-red [18,29], oven [28,37] and room temperature drying [35]. In some cases, coated samples have been dried under tension [35]. The results of these works reveal that a too fast evaporation rate of CNFs coating causes disruption of the coating layer as the viscosity of coating increases faster. Therefore, it would be worthwhile to determine how the coating and drying methods in CNFs coated papers affect the integrity and uniformity of the paper surface and its final properties. Final paper properties are established by finishing operations in papermaking. With coated papers, the final roughness will determine the size and number of pores remaining on the coated surface. Therefore, coated papers are usually subjected to smoothing and calendering to achieve the target roughness. Some authors calendered CNFs coated papers [26,27,30,32,36], but because they did not define roughness targets, it is not possible to determine the extent of the changes in air permeance due to CNFs coatings or the roughness differences between samples. To the best of our knowledge, few studies have compared the performance of different types of CNFs applied as paper surface coatings under the same conditions and little information is available about CNFs application using different coating methods. Suitability of CNFs paper coatings with different metering systems and drying methods were addressed; three types of CNFs — mechanical, TEMPO and carboxymethylated were applied at different pickups to improve paper properties. Two metering systems for applying the coating film (wound rod and blade micrometer) and two drying methods (contact with a hot polished surface and hot air jet) were evaluated. As a control, CNFs coated papers were compared with a paper coated with cationic starch, usually applied in the size-press of the paper machine. The quality of the CNFs coated papers was evaluated by barrier, structural and optical properties: air permeance, SEM images, gloss, and visual appearance. 2. Materials and methods 2.1. Raw materials and chemicals 2.1.1. Commercial CNFs and starch Commercial CNFs samples comprised a slurry of cellulose nanofibrils mechanically prepared by super mass collider (m-CNFs), freeze-dried carboxymethylated nanofibrils (c-CNFs), and freeze-dried TEMPO nanofibrils (T-CNFs) sourced from Cellulose Lab (Canada). According to data provided by the manufacturer, the average widths were 30–80 nm, 10–13 nm and 50 nm for m-CNFs, c-CNFs and T-CNFs, respectively, and the average lengths were up to several hundred microns for m-CNFs, 1–3 μ m for c-CNFs, and 0.5–80 μ m for T-CNFs. After adding water to achieve 2 % o.d., the three CNFs were stirred in Ultraturrax (IKA, Germany) at 10,000 rpm at room temperature for 10 min. The 2 % o.d. of solid concentration was chosen as the highest level to ensure the fluidity of coating suspensions onto the paper surface. Commercial CNFs were characterised by fibrillation yield, acidic group content, and water retention value (WRV). Cationic Amylofax® HS A starch (AS) was kindly provided by Avebe (Netherlands) and cooked at a concentration of 10 % o.d. at 90 ◦C for 30 min under continuous agitation. 2.1.2. Recycled paper A commercial recycled paper supplied by Smurfit Kappa (Spain) was chosen as the base paper for this study. The two recycled papers used, RF1 and RF2, were characterised by basis weight, thickness, gloss, Gurley air resistance, Bendtsen air permeance and Bendtsen roughness (Table 1). All tests were conducted following ISO standards at standard laboratory conditions (23 ±1 ◦C and 50 ±2 % RH). The characterisation showed high similarity between the two papers. 2.2. CNFs characterisation 2.2.1. Fibrillation yield (FY) A dispersion of CNFs in water at 0.1 % (w/v) was prepared, then centrifuged (2200 g, 20 min) to separate the nanofibrillated material (supernatant) from non-fibrillated and partially fibrillated fibres (sediment) [38]. The sediment was dried (104 ◦C) and the fibrillation yield (FY) was calculated as the ratio between the mass of cellulose nanofibers in the supernatant and the mass of centrifuged cellulose, using Eq. 1: FY (%) = 100⋅(1− (WS/W0) ) (1) where W 0 is the dry weight (g) of CNFs before centrifugation and W S is the weight of the dry sediment after centrifugation (g). 2.2.2. Water retention value Water retention value (WRV) was measured following Cheng et al. [39], but using an acetate filter of 0.45 μ m pore size. CNFs suspensions were centrifuged at 3000g for 20 min at room temperature. WRV was calculated using Eq. 2: WRV (%) = (WRF−WR0)*100/WR0(2) where WR F is the weight (g) of the wet sample remaining in the filter after centrifugation and WR 0 is the dry weight (g) of the test sample. 2.2.3. Acidic group content Acidic group content (AG) was measured by the conductometric titration method [40], with modifications. Firstly, sodium carboxylate groups were converted to free carboxyl groups by ion-exchange treatment. A sample of 0.15 g o.d. was dispersed in 185 mL of 1 mM NaCl solution, then 0.1 M HCl solution was added to pH 3. The suspension was stirred under nitrogen atmosphere for 30 min at room temperature. After agitation, titration 0.05 M NaOH was added every 30 s and conductivity was recorded using a COND 8+conductometer (XS Instruments, Italy). Carboxylate content was calculated using Eq. 3: AG (mmol/g) = c⋅(V2−V1)/w(3) where V 1 and V 2 are the NaOH volume (mL) consumed at the first and second intersection point, respectively, c is the NaOH concentration (M), and w is the sample dry weight (g). 2.3. Coating procedure: coating metering systems, drying methods and calendering The first study (#1) addressed the effect of CNFs film wet thickness on the coated paper properties, and the second (#2) addressed the effect of multilayer application. Both studies evaluated the influence of the coating metering system and drying method on the resulting coated papers. Before CNFs coating, the base paper samples were preconditioned to a moisture content of about 18 %, between the range of 15 % and 20 % used in the size-press in papermaking. Before coating, the starch solution and CNFs suspensions were pre-heated to 40 ◦C to Ú. Fillat et al. Progress in Organic Coatings 179 (2023) 107543 3 improve runnability. Paper samples were coated using K Control Coater (K202), from RK Printcoat Instruments (UK). This equipment includes several wound rods (R) and an adjustable blade micrometer (B) to apply the coating suspension. The grooves of wound rod precisely control wet film thickness. The micrometer applicator incorporates an adjustable spreading blade using a micrometer to accurately set the paper substrate/blade gap in 10 μ m increments. Coating was always applied with pressure between the metering unit and the coating bed, and the loading pressure was adjusted manually following the manufacturer's instructions. The coating beds were chosen following the manufacturer's suggestions: three-part foam/rubber/melinex for rod applications and float glass for blade applications. The rod metering speed was set to 5 m/ min. All paper samples were coated on the smoothest side. To evaluate the effect of the drying method on the structure and barrier properties of the coated papers, the coated layers were dried using a Speed Dryer (SD24E) from Labtech (Italy), in which the coated samples were dried against a highly polished surface (PS) at 85 ◦C for 5 min, and an air jet dryer (AJ), where a hot air jet at 70 ±5 ◦C was directed onto the wet coating layer. Table 2 shows the operating conditions established in study #1 and study #2 and the theoretical coat weights (g/m 2 ) calculated from wet film thickness and dry solids content of CNFs dispersions. Three coating samples were obtained from each operating condition with each CNFs suspension or starch solution. Lastly, coated samples were smoothed using a metal iron at 160 ± 5 ◦C to a Bendtsen roughness of 1600 ±300 mL/min. The CNFs coated papers were therefore compared under similar roughness values, as this parameter is directly related to the porous structure of paper. To increase coat pick-up in CNFs-coated samples, the coating and drying steps were repeated to apply five (5F) or 10 layers (10F) of each CNFs using the operating conditions described in Study #2 in Table 2. Coated samples were finally smoothed only when the five (5F) or 10 layers (10F) of CNFs coating had been applied. The influence of pre-conditioning, wetting, drying and smoothing operations in base paper was assessed by subjecting test pieces of the original papers (RF1 and RF2) to each experimental condition described in Table 2, replacing CNFs and starch suspension application by spraying the paper with water until it became completely wet. These samples were used as control papers (RF1-c and RF2-c). 2.4. Coated paper characterisation Control papers and papers coated with CNFs suspensions or starch solution were characterised as original papers. All tests were performed at standard laboratory conditions (23 ±1 ◦C and 50 ±2 % RH). To study surface coating uniformity, coated samples were scanned (Epson Expression 1000 XL, Japan) with image resolution set at 1400 pixels/in. Morphological analyses were performed by Scanning Electron Microscopy (SEM) equipped with an energy dispersive spectrometer (EDS) using a JSM 6400 (JEOL, Tokyo, Japan) at 25 kV (maximum resolution 3.5 nm and 8 mm working distance). For SEM observation, samples were stored at 50 ◦C for 24 h, followed by evaporation with graphite and metallisation with gold to prevent accumulation of surface charge. 3. Results and discussion 3.1. CNFs characterisation c-CNFs nanocellulose had the highest fibrillation yield, water retention value, and acidic group content, followed by T-CNFs, as shown in Table 3. m-CNFs, produced by mechanical treatment, showed the lowest fibrillation yield and negligible acidic group content. This characterisation was consistent with the average width and length reported by the supplier. c-CNFs exhibited the lowest fibril dimensions, followed by T-CNFs and m-CNFs. These differences between the chemically modified CNFs (c-CNFs and T-CNFs) and m-CNFs were expected, as mCNFs was subjected only to mechanical treatment. Fibril size distribution and acidic group content influenced the visual appearance of the three CNFs. c-CNFs and T-CNFs suspensions had a gellike appearance (transparent-translucent), whereas m-CNFs is opaque and tended to flocculate and separate in phases. WRV is a measure of the intrinsic ability of cellulose products (in this case CNFs suspensions) to hold water. The amount of water includes the water not only in pores, but also in fibril external surfaces and even inter-fibre spaces. WRV has been associated with fibrillation degree [27] and carboxyl content [41]. This agrees with the characterisation results of the three commercial CNFs, as c-CNFs showed the highest fibrillation yield and acidic group content, also resulting in the highest WRV. Differences in fibrillation yield between m-CNFs and T-CNFs were not high, but the largest concentration of charged groups in T-CNFs resulted in five times more WRV than in m-CNFs. The gel-like appearance and high WRV in c-CNFs and T-CNFs samples was expected to affect the runnability of the coating process and the visual appearance of coated samples. Table 1 Base paper properties. Property (units) Basis weight (g/m 2 ) Thickness ( μ m) Gloss (%) Gurley air resistance (s) Bendtsen air permeance (mL/min) Bendtsen roughness (mL/min) Standard ISO 536 ISO 534 ISO 8254-1 ISO 5636-5 ISO 5636-3 ISO 8791-2 Side A Side B RF1 122 ±2 179 ±4 4.1 ±0.2 60 ±6 223 ±25 1766 ±97 2139 ±142 RF2 116 ±1 174 ±5 4.2 ±0.6 33 ±2 355 ±18 1382 ±160 1712 ±62 Table 2 Operating conditions used in K Control Coater and theoretical coat weights. Operating conditions Theoretical coat weights (g/m 2 ) with: Study, layer and base paper Rod (R)/ blade (B) Thickness wet film ( μ m) Drying method 2 % o. d. CNFs 10 % o. d. starch Study #1. One layer (1F). RF1 RF2 R-2 12 PS 0.24 1 R-5 50 PS 1 5 R-8 100 PS 2 10 B-5 50 PS 1 5 B-10 100 PS 2 10 Study #2. One (1F), five (5F) and 10 (10F) layers. RF2 R-5 50 PS/AJ 1 5 10 5 B-5 50 PS/AJ 1 5 10 5 Table 3 CNFs characterisation. FY (%) WRV (%) AG ( μ mol/g) m-CNFs 4 ±1 452 ±14 75 ±28 T-CNFs 17 ±2 2118 ±27 1010 ±35 c-CNFs 89 ±1 4127 ±358 2831 ±150 Ú. Fillat et al. Progress in Organic Coatings 179 (2023) 107543 4 3.2. Effect of wet film thickness on coated paper properties In Study #1, CNFs wet films were applied to the commercial papers (RF1 and RF2) using three rods and two gaps with the blade. Wet coated film was dried by the polished surface method. The theoretical coat weight of CNFs was 0.2–2.0 g/m 2 and the results were compared with those of starch coated paper (1.2–10.0 g/m 2 ), as shown in Table 2. No significant differences were detected in thickness. All samples showed mean values of 175–181 μ m for RF1 and 175–185 μ m for RF2 coated samples. These differences are within the uncertainty of the method. Air permeance and gloss of controls RF1-c and RF2-c were similar to those of the base papers, differences were negligible compared to the increases produced by CNFs and starch coatings. As shown in Fig. 1a, a slight increase in air resistance was observed when starch was applied to RF1, similar for all rods (expressed as film thickness at top). The increase was more noticeable when using blade coating on RF2 paper (Fig. 1b). cCNFs increased air resistance in all samples. The increments were higher than those observed with starch, even taking into account that c-CNFs suspension was applied at a lower concentration. The improvement was progressively higher when the amount of c-CNFs suspension applied with rods in RF1 was higher (Fig. 1a). Air resistance of RF2 is much higher in c-CNFs samples coated with blade (Fig. 1b) than in samples corresponding to RF1 and rods, even though the coat weight of the wet film was similar. No significant differences were observed after coating RF1 with T-CNFs and rods, and only the RF2 coating with blade and gap B-10 caused meaningful air resistance, similar to that obtained with starch coating. m-CNFs, obtained using only mechanical energy, was not very effective. Neither of the coating methods (rod or blade) imparted air resistance and only the rod method had a minimal effect on RF1 paper. No significant differences in gloss were observed when mechanical m-CNFs was applied with rods or blade, while some increase was observed when using starch and the two chemical CNFs (Fig. 2). In general, when a higher thickness of wet film was transferred to the base paper, both on RF1 and RF2, an increase in gloss was observed using rods and blade. Depending on the coating technology, c-CNFs was more advisable for rod coating, while application with blade gave the best coating gloss using T-CNFs. Due to the low CNFs content in the dispersions, it was not possible to obtain a high coat weight with only one layer when using different rods or gaps with the blade. The low coat weight applied could explain the poor results for air resistance obtained with m-CNFs and T-CNFs despite the better results obtained with c-CNFs coatings. In addition, due to the higher water content, application of the highest wet film thickness could lead to paper swelling and open the fibre network, resulting in lower air resistance of the coated papers, as reported by Lavoine et al. [22]. Moreover, drying could be very difficult, producing cockles and wrinkles in the final coated papers. It was therefore worthwhile determining whether higher coat weights with the three CNFs could improve the barrier properties. To avoid excessive soaking of the base paper and ensure an adequate coat weight, we decided to apply several coating layers (Study #2) instead of one coating with a higher wet film thickness. In this case the no. 5 R-5 rod and B-5 gap with blade were chosen for the next study on applying different layers. Both of these options apply a film with a thickness of 50 μ m. 3.3. Influence of coating and drying technology. Multilayer study 3.3.1. Thickness and roughness In Study #2 coat weight was increased by applying one, five and 10 films to obtain theoretical CNFs coat weights of 1, 5 and 10 g/m 2 , calculated from the concentration of the suspensions and the thickness of coating wet films applied. Fig. 3a and b shows the average values of the thickness and roughness of the papers coated under the different conditions. The properties correspond both to rod and blade coated samples, as no appreciable differences were found between the two application methods. The thickness value of the samples coated with one to five layers (1F to 5F) is the same as the value of control RF2-c: 180 ±10 μ m, as shown in Fig. 3a. This was expected, due to the low coat weights and superficial calendering. However, in samples coated with 10 layers (10F), the thickness increase was significant, especially with T-CNFs and m-CNFs nanocelluloses dried with hot air jet (10F-AJ). Although the roughness target was set at 1600 ±300 mL/min, it was not possible to obtain this value with 10 layers of T-CNFs or m-CNFs dried with hot air jet (10F-AJ) (Fig. 3b). Roughness and thickness results with ten layers can be explained by the presence of flocs of fibres in mCNFs, and the successive wetting and drying cycles that caused swelling on the coating and the base paper and created cockles and wrinkles on the T-CNFs coating surface, as detailed below. 3.3.2. Air permeance In Study #2, air permeance results were obtained from measurements of Gurley air resistance, because this method has a lower detection limit (0.1 μ m/Pa⋅s) than the Bendtsen method (0.35 μ m/Pa⋅s). Fig. 4a, b and c, respectively, show the air permeance when one, five and 10 films of CNFs were applied. The bars of each figure represent each type of CNFs (and one starch film in Fig. 4b for comparison) applied with rod (light bars) or blade (dark bars) and dried with polished surface (plain bars) or hot air jet (striped bars). In all the samples coated with the same number of layers, the lowest air permeance always corresponded to the samples coated with c-CNFs. Air permeance with a single layer of c-CNFs (0.1–0.3 μ m/Pa⋅s) was even lower than that obtained by applying starch at a higher coat weight (0.1 to 0.7 μ m/Pa⋅s). Compared to previous studies, one layer of c-CNFs produced lower air permeance than those obtained at similar carboxymethylated CNFs coating grammage on unbleached kraft paper without further calendering [42] or on recycled paperboard by slot die with calendering [30]. When applying Fig. 1. Increments in Gurley air resistance in Study #1. a) Rod coated RF1 papers and b) blade coated RF2 papers. Ú. Fillat et al. Progress in Organic Coatings 179 (2023) 107543 5 five c-CNFs layers (5 g/m 2 ) onto RF2, an air permeance of 0.01 μ m/Pa⋅s was obtained, a value considered very low in agreement with other authors who used coat weights of around 6 g/m 2 [30,43]. Increasing the number of layers to 10 did not lead to improvements, but rather a slight increase (in some samples) that could be attributed to irregularities caused by the successive application of layers, as each layer comprises a wetting and drying cycle that destructures the material. No clear effect of the method of application or drying of the carboxymethylated CNFs was observed. When one layer of T-CNFs was applied, this CNF was able to decrease air permeance to 1.1–3.6 μ m/Pa⋅s. As in the case of c-CNFs, the similar results of the coating and drying methods did not permit inference of an effect on air permeance. Air permeance obtained with 5 T-CNFs layers falled to 0.1–0.4 μ m/Pa⋅s, similar to the values obtained with starch at the same coat weight and with a single layer of c-CNFs. Air permeance values were higher than those obtained by Ottesen et al. [30] when TCNFs was applied by slot die and calendered. These better results can be attributed to differences in calendering or in T-CNFs fibrillation degree. However, the high variability in the data obtained should be noted. The variability increased as the number of layers increased, suggesting that increased manipulation associated with more layers, with successive wetting and drying cycles, is responsible for the greater variability caused by the appearance of surface defect irregularities. Lastly, mechanical m-CNFs was unable to decrease air permeance when a single layer was applied, but caused a substantial decrease with five or 10 layers. Air permeance with five layers was even lower in mCNFs samples than the values obtained with T-CNFs and the reference starch. Poor results with mechanical CNFs coatings at low coat weights reported in the literature [23,28] improved when coat weights were increased from 2.3 to 9 g/m 2 [22,29,34,35]. Similar air permeance (below 0.03 μ m/Pa⋅s) was obtained by Kumar et al. [18,26] and Kumar, Bousfield and Toivakka [27] by applying 5 g/m 2 of mechanical CNFs on liner board and paper board using a rotary coater with slot die and calendering. Increasing the layers from five to 10 did not improve air permeance in any of the CNFs samples, indeed air permeance increased. This may be related to a loss of integrity of the sample due to wetting and drying cycles, as reported by other authors [29]. For the same coating metering system, no conclusive differences were observed by drying method using the drying parameters applied in this work, except for starch coatings. Better efficiency of the hot air jet was observed when drying one layer of starch (Fig. 4b), both for rod and blade application. When one layer of TCNFs was applied using the rod method, hot air jet drying yielded lower air permeance than the polished surface method (Fig. 4a), although no further differences were found when using five or ten layers. No definitive conclusion can be drawn regarding the effect of the drying method when applying m-CNFs nor c-CNFs. The lowest air permeance observed with c-CNFs agrees with previous studies using other coating and drying methods and base substrates [30,37,43]. Film formation is typically influenced by factors including particle size distribution, rheological and chemical properties of the suspension, water evaporation rate and drying temperature [5]. Air permeance has also been associated with the mean diameter of the CNFs used [36,43]: a high aspect ratio increases the amount of potential hydrogen bonds between nanofibrils [30], forming a denser layer with compact packing that increases tortuosity [43], leading to lower gas permeability [42,44]. Moreover, the degree of carboxylation increases water retention and viscosity in CNFs suspensions, and high WRV in coating prevents water absorption inside the paper [5]. Therefore, the lower air permeance obtained with c-CNFs is consistent with its higher fibrillation, WRV and carboxylation degree compared with T-CNFs and m-CNFs, and also explains its better runnability in coating and film Fig. 2. Gloss in Study #1. a) Rod coated RF1 papers and b) blade coated RF2 papers. Fig. 3. RF2 papers coated with starch and CNFs. Mean values for a) thickness and b) roughness. Average values of rod and blade coated samples. Application of one, five or 10 layers (1F, 5F and 10F) and drying by hot air (10F-AJ) or polished surface method (10F-PS). Ú. Fillat et al. Progress in Organic Coatings 179 (2023) 107543 6 formation during drying. One layer of c-CNFs allowed to significantly decrease the air permeance, then this CNFs can be considered an adequate material for air barrier coatings in packaging. T-CNFs and m-CNFs showed good results when an adequate coat weight was applied. 3.3.3. Gloss Gloss measurements showed similar behaviour in all coated papers, regardless of the coating metering or drying system (data not shown). Fig. 5 shows gloss values in samples coated by blade with one, five and 10 layers and dried using the polished surface method. m-CNFs did not improve paper gloss with one layer and improved it only slightly with five or 10 layers. c-CNFs slightly increased gloss when one layer was applied (similar to the gloss obtained with cationic starch) and with five or 10 layers the surface gloss was 11.9 % and 20.6 %, respectively. The best results were obtained with T-CNFs: a single layer of T-CNFs obtained a similar gloss to five layers of c-CNFs and more than 10 layers of m-CNFs. By increasing the number of layers to five or 10, the gloss values were 23.8 % and 30 %, respectively, improving the visual perception of the coated papers. A decrease in gloss when coating with mechanical CNFs was observed by other authors [23,30,32,37], but an increase was observed when coating carboxymethylated and TEMPO CNFs [30]. Gloss of coated papers is a parameter that depends on surface properties, and is related to the surface roughness and surface texture of the paper [45]. Non-homogeneous surfaces can reduce gloss, which is also affected by the shape and particle size distribution of coating particles and the refractive index of the material [46,47]. c-CNFs and T-CNFs suspensions exhibited lower fibril dimensions than m-CNFs and a translucent appearance, also contributing to increased gloss in the coated paper. 3.3.4. SEM images of coated samples When examining the surface of RF2 corrugated paper, the SEM image (Fig. 6a) shows an open, porous network of randomly crossed fibres with a presence of mineral fillers. Comparison of starch coating methods (metering and drying) showed differences in the sample surfaces (Fig. 6b and c), indicating that the method affects the coated surface. The main differences were found in samples coated with rod and dried with hot air jet (AJ), where a greater coverage of the paper surface was observed compared to application with the blade and drying with the polished surface method (PS). This is in accordance with the smaller values of air permeance measured in starch-coated samples dried with hot air jet (Fig. 4b). Fig. 7 shows SEM images of CNFs coated samples using the wound rod (R) and polished surface (PS) methods. When a layer of mechanical m-CNFs was coated (Fig. 7c — 1F), the surface of the paper was not completely covered. The SEM image revealed the presence of micro/ nano fibrils adhered to the fibres of the base paper that are most probably m-CNFs (indicated with arrows), showing large pores. By increasing the number of layers (Fig. 7c — 5F), a CNFs film completely covered the base paper. Poor coverage can explain the absence of differences between control RF2-c and coated paper with one m-CNFs layer; moreover, air permeance decreased significantly when five layers were applied, and a non-porous m-CNFs film was formed. A lack of the coverage of the mechanical and enzymatic CNFs coatings with low coat weights has been reported in the literature [28,43,48], as well as fully Fig. 4. Air permeance measured by the Gurley method in RF2 coated paper. a) One layer of CNFs, b) one layer of starch and five layers of CNFs and c) 10 layers of CNFs. Fig. 5. Gloss on RF2 blade coated papers dried on a polished surface. One (1F), five (5F) and 10 layers (10F) of CNFs. Ú. Fillat et al. Progress in Organic Coatings 179 (2023) 107543 7 covered surfaces with higher layer weights [22,27,30]. When m-CNFs coatings were applied, better coverage was observed when one layer was coated by blade rather than by rod (Fig. 8). A wound rod follows the unevenness of the paper, and the m-CNFs suspension was inserted between the fibres of the surface base paper. However, blade coating applied a greater amount of mechanical CNFs in the valley areas of the base paper and less to the hill areas. Then coating thickness was less uniform. On the other hand, when one layer (1F) of the chemically modified CNFs (T-CNFs and c-CNFs) was coated (Fig. 7a and b), the surface of the paper was well covered and the network structure of the fibres in the base paper was less evident (Fig. 6a), also in agreement with the decrease in air permeance with a single c-CNFs or T-CNFs layer. When the number of layers (5F) was increased, with a corresponding increase in coat weight, the original fibres of the base paper were barely visible (Fig. 7a and b). No differences were observed when the chemical CNFs (c-CNFs and T-CNFs) were applied by rod or blade and dried with the polished surface method (images of c-CNFs coated by blade not shown). Comparison of the three CNFs showed that m-CNFs coated paper with five layers had a rougher appearance than the corresponding surfaces obtained by applying T-CNFs and c-CNFs. This can be explained by the greater average width and length of the micro and nanofibrils of mCNFs compared to those of the chemical CNFs. In addition, Fig. 9 shows the presence of holes in T-CNFs coated samples. T-CNFs exhibited a lower fibrillation degree than c-CNFs and thus the crack defects in coatings may have been formed during drying due to the presence of large fibrils [43] and can also explain the poorer results in air permeance in T-CNFs than in c-CNFs samples (Fig. 4). 3.3.5. Visual appearance of coated samples Images of uncoated paper and blade coated samples with each CNFs suspension with one, five or 10 layers are shown in Fig. 10. Each image shows a 280 mm ×280 mm surface corresponding to a consumer's visual perception when observing the sample. As seen in Fig. 10, considerable differences were observed among the CNFs blade coated surfaces. Similar behaviours were observed in rod coated samples (data not shown). The chemical CNFs (T-CNFs and cCNFs) barely modified the visual appearance of the base paper surface after coating. Only a slight gloss gain was noticeable when the quantity Fig. 6. SEM images of starch-coated paper at ×500. a) Recycled paper RF2, b) starch-coated paper by micrometer blade dried by polished surface (AS-B-5-PS), and c) starch-coated paper by wound rod dried by hot air jet (AS-R-5-AJ). Fig. 7. SEM images of CNFs-coated RF2 paper by no. 5 wound rod dried using the polished surface method (R-5-PS) with one layer (1F) and five layers (5F) at ×1000. a) TEMPO CNFs (T-CNFs), b) carboxymethylated CNFs (c-CNFs) and c) mechanical CNFs (m-CNFs). Fig. 8. SEM images x500. One layer (1F) of m-CNFs coating dried using the polished surface method. a) Coated by no. 5 wound rod (m-CNFs-R-5-PS) and b) by micrometer blade (m-CNFs-B-5-PS). Ú. Fillat et al. Progress in Organic Coatings 179 (2023) 107543 8 of nanofibres transferred to the paper surface was increased, in accordance with the gloss results (Fig. 3). In contrast, m-CNFs coated surfaces had an uneven white hue, nanofibre flocs can be seen spread out unevenly across the surface, as other authors reported [27]. This rough appearance was more pronounced at higher coat weights, especially with 10 layers. This could be explained by the low fibrillation yield and higher average fibril dimensions of mechanical m-CNFs compared to chemical CNFs, causing m-CNFs to form white and opaque aqueous suspensions, whereas and T-CNFs suspensions were translucent. Comparison of the images of papers coated with each CNFs suspension and different applications (1F, 5F and 10F) by drying method showed that coated papers dried by hot air jet produced a wavy surface. This defect was especially evident in T-CNFs coated samples when the number of CNFs layers was increased. This unwanted phenomenon, known as cockling, appears on the paper surface as wrinkles that are usually much longer than wider, and is related to wet end operating conditions and drying conditions in papermaking [49]. It is usually induced by a moisture content change and is irreversible in many cases; i.e., the flat surface of the sheet is not recovered even when the initial moisture content is restored. Lipponen et al. [50] demonstrated that moisture gradient is a crucial factor in irreversible cockling during cyclic moisture content change. In our work, paper samples were subjected to multiple drying processes after applying the CNFs wet films, but this defect was noticeable only in multilayers of coated samples dried by air jet. It can be attributed to the hot air used to dry wet coating films, which caused a higher moisture gradient than hot contact drying in the speed dryer. In addition, during hot contact drying, the coated papers were under slight pressure applied by the cover of the equipment, which might have reduced the formation of cockling. With regard to visual appearance, c-CNFs was the most suitable of three CNFs tested for coating paper. Carboxymethylated CNFs formed translucent, uniform coats with a visual appearance similar to the base paper and gloss values that resembles papers laminated with plastics. No defects due to runnability (e.g., cockling) were detected. 4. Conclusions Different metering systems and drying methods were used to coat a recycled paper with mechanical, TEMPO and carboxymethylated CNFs. In the first study, wet CNFs films were applied using three wound rods and two gaps with the blade and were dried by contact with a hot polished surface. Best results in air permeance were obtained with c-CNFs, besides an increase in the thickness of the film affected air permeance in all c-CNF samples and in a lesser extent in T-CNFs samples coated by blade. The highest gloss values were obtained by T-CNFs when coated by blade with the highest gap. In the second study, two metering systems (wound rod and blade micrometer) and two drying methods (hot polished surface and hot air jet drying) were used to apply a multilayer CNFs coating. c-CNFs, which showed the highest fibrillation degree and acid group content of the three CNFs, exhibited the best performance in runnability with all the methods studied and was also the coating that yielded the lowest air permeance values, reaching 0.01 μ m/Pa⋅s, the best paper coverage, and the best visual appearance. Samples coated with TCNFs showed a slight decrease in air permeance, mainly when five layers were applied, values of 0.1 μ m/Pa⋅s were obtained. A lack of coverage was observed in SEM images after application of a single layer of mCNFs, which was unable to decrease air permeance. However, when an adequate coat weight was applied, a thin film completely covered the Fig. 9. SEM images ×200. One layer of T-CNFs coating dried using the polished surface method a) coated by no. 5 wound rod (T-CNFs-R-5-PS) with one layer (1F) and b) by micrometer blade (T-CNFs-B-5-PS) with one layer (1F). Fig. 10. Scanned images of uncoated paper RF2-c and CNFs blade coated (B-5) papers, with one, five and 10 layers (1F, 5F and 10F), respectively. Dried by hot air jet (AJ) and polished surface method (PS). a) T-CNFs, b) c-CNFs and c) mCNFs. Sample size: 280 mm ×280 mm. Ú. Fillat et al. Progress in Organic Coatings 179 (2023) 107543 9 base paper and a remarkable decrease in air permeance were obtained. The metering and drying methods affected air permeance and surface morphology in samples coated with one layer of T-CNFs and starch, but no differences were observed when more than five layers were applied. The three CNFs studied will be included in future studies to assess barrier properties that must be considered in paper packaging intended to come into contact with food. Funding This work was supported by the Spanish Ministry of Science and Innovation (CTQ2017-84963-C2-2-R and PID2020-114365RB-C22) and Madrid Regional Government (RETO PROSOST S2018/EMT-4459). Ethics approval and consent to participate Not applicable. No research have been conducted involving human participants and/or animals. Consent for publication Not applicable. Availability of data and materials All of the material is owned by the authors and/or no permissions are required. Authors adhere to discipline-specific rules for acquiring, selecting and processing data. CRediT authorship contribution statement All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Úrsula Fillat, Priscilla Vergara, Juan Carlos Villar and Nuria G´ omez. The first draft of the manuscript was written by Úrsula Fillat and all authors commented on previous versions of the manuscript. The manuscript is an original work of the authors and has not been published in another journal or previously submitted to “Progress in Organic Coatings”. All authors read and approved the manuscript for publication to “Progress in Organic Coatings”. Declaration of competing interest The authors have no relevant financial, non-financial or competing interests to disclose. Data availability Data will be made available on request. Acknowledgements This work was supported by the Spanish Ministry of Science and Innovation (CTQ2017-84963-C2-2-R and PID2020-114365RB-C22) and Madrid Regional Government (RETO PROSOST S2018/EMT-4459). Morphological analysis of coated surface samples was carried out in the National Centre of Electronic Microscopy at the Complutense University of Madrid. The authors thank P. Pereira, P. Meza and P. Castellano for their contribution to the experimental work. References [1] Eurostat, Eurostat statistics explained, in: Packaging Waste Statistics, 2021. https ://ec.europa.eu/eurostat/statistics-explained/index.php?title=Packaging_waste_s tatistics. (Accessed 9 October 2021). [2] European Parliament, Plastic waste and recycling in the EU: facts and figures, in: European Parliament News, 2021. https://www.europarl.europa.eu/news/en/he adlines/society/20181212STO21610/plastic-waste-and-recycling-in-the-eu-facts -and-figures. (Accessed 9 March 2021). [3] Eurostat, Eurostat data browser, in: Recycling rate of Packaging Waste by Type of Packaging, 2021. https://ec.europa.eu/eurostat/databrowser/view/CEI_WM020__ custom_354860/bookmark/table?lang=en&bookmarkId=bc39f400-65cd-40a 8-bf14-c995c729e2a5. (Accessed 9 December 2021). [4] U. Thoden van Velzen, L. Lisanne De Weert, K. Molenveld, Flexible laminates within the circular economy, in: Technical Report WFBR 2037. Project: Sustainable Packages, Wageningen University & Research, 2020, https://doi.org/10.18174/ 519019. [5] T. Kimpim¨ aki, A.V. Savolainen, Barrier dispersion coating of paper and board, in: Surface Application of Paper Chemicals, Springer, Dordrecht, 1997, https://doi. org/10.1007/978-94-009-1457-5_12. [6] R. Raditya, Parameters that influence the performance of dispersion barrier coatings 2990, University of Maine, 2019. Electronic theses and dissertations. [7] M.A. Hubbe, A. Ferrer, P. Tyagi, Y. Yin, C. Salas, L. Pal, O. Rojas, Nanocellulose in thin films, coatings, and plies for packaging applications: a review, Bioresources 12 (1) (2017) 2143–2233, https://doi.org/10.15376/Biores.12.1.2143-2233. [8] A.F. Turbak, F.W. Snyder, K.R. Sandberg, Micro-fibrillated cellulose and process for producing it, in: Patent n◦CH 648071 (A5), 1985. [9] T. Saito, S. Kimura, Y. Nishiyama, A. Isogai, Cellulose nanofibers prepared by TEMPO-mediated oxidation of native cellulose, Biomacromolecules 8 (2007) 2485–2491, https://doi.org/10.1021/bm0703970. [10] A. Isogai, T. Saito, H. Fukuzumi, TEMPO-oxidized cellulose nanofibers, Nanoscale 3 (1) (2011) 71, https://doi.org/10.1039/C0NR00583E. [11] T. Isogai, T. Saito, A. Isogai, Wood cellulose nanofibrils prepared by TEMPO electro-mediated oxidation, Cellulose 18 (2) (2011) 421–431, https://doi.org/ 10.1007/s10570-010-9484-9. [12] A. Naderi, T. Lindstr¨ om, J. Sundstr¨ om, Repeated homogenization, a route for decreasing the energy consumption in the manufacturing process of carboxymethylated nanofibrillated cellulose? Cellulose 22 (2) (2015) 1147–1157, https://doi.org/10.1007/s10570-015-0576-4. [13] U. Fillat, B. Wicklein, R. Martín-Sampedro, D. Ibarra, E. Ruiz-Hitzky, C. Valencia, A. Sarri´ on, E. Castro, M.E. Eugenio, Assessing cellulose nanofiber production from olive tree pruning residue, Carbohydr. Polym. 179 (2018) 252–261, https://doi. org/10.1016/j.carbpol.2017.09.072. [14] N. Lavoine, I. Desloges, A. Dufresne, J. Bras, Microfibrillated cellulose – its barrier properties and applications in cellulosic materials: a review, Carbohydr. Polym. 90 (2012) 735–764, https://doi.org/10.1016/j.carbpol.2012.05.026. [15] A. Balea, E. Fuente, M.C. Monte, N. Merayo, C. Campano, C. Negro, A. Blanco, Industrial application of nanocelluloses in papermaking: a review of challenges, technical solutions, and market perspectives, Molecules 25 (526) (2020) 1–30, https://doi.org/10.3390/molecules25030526. [16] P.R. Charani, M.H. Moradian, Utilization of cellulose nanofibers and cationic polymers to improve breaking length of paper, Cell. Chem. Technol. 53 (7–8) (2019) 767–774, https://doi.org/10.35812/CelluloseChemTechnol.2019.53.75. [17] Ø. Eriksen, K. Syverud, Ø. Gregersen, The use of microfibrillated cellulose produced from Kraft pulp as strength enhancer in TMP paper, Nord. Pulp Pap. Res. J. 23 (3) (2008) 299–304, https://doi.org/10.3183/npprj-2008-23-03-p299-304. [18] V. Kumar, A. Elfving, H. Koivula, D. Bousfield, M. Toivakka, Roll-to-roll processed cellulose nanofiber coatings, Ind. Eng. Chem. Res. 55 (12) (2016) 3603–3613, https://doi.org/10.1021/acs.iecr.6b00417, doi:10.15376/Biores.12.4.7656-7679. [19] A.F. Lourenço, J.A.F. Gamelas, T. Nunes, J. Amaral, P. Mutj´ e, P.J. Ferreira, Influence of TEMPO-oxidised cellulose nanofibrils on the properties of fillercontaining papers, Cellulose 24 (2017) 349–362, https://doi.org/10.1007/s10570016-1121-9. [20] T. Taipale, M. ¨ Osterberg, A. Nyk¨ anen, J. Ruokolainen, J. Laine, Effect of microfibrillated cellulose and fines on the drainage of Kraft pulp suspension and paper strength, Cellulose 17 (5) (2010) 1005–1020, https://doi.org/10.1007/ s10570-010-9431-9. [21] Y. Boissard, MFC for Paper Surface Treatment, Luleå University of Technology, 2017. Master’s thesis. [22] N. Lavoine, I. Desloges, B. Khelifi, J. Bras, Impact of different coating processes of microfibrillated cellulose on the mechanical and barrier properties of paper, J. Mater. Sci. 49 (2014) 2879–2893, https://doi.org/10.1007/s10853-013-7995-0. [23] F. Richmond, Cellulose Nanofibers Use in Coated Papers, University of Maine, 2014. PhD thesis. [24] A.F. Lourenço, J.A.F. Gamelas, P. Sarmento, P.J. Ferreira, Cellulose micro and nanofibrils as coating agent for improved printability in office papers, Cellulose 27 (2020) 6001–6010, https://doi.org/10.1007/s10570-020-03184-9. [25] H. Song, M. Ankerfors, M. Hoc, T. Lindstr¨ om, Reduction of the linting and dusting propensity of newspaper using starch and microfibrillated cellulose, Nord. Pulp Pap. Res. J. 25 (4) (2010) 495–504, https://doi.org/10.3183/NPPRJ-2010-25-04p519-528. [26] V. Kumar, V.R. Koppolu, D. Bousfield, M. Toivakka, Substrate role in coating of microfibrillated cellulose suspensions, Cellulose 24 (2017) 1247–1260, https://doi. org/10.1007/s10570-017-1201-5. [27] V. Kumar, D.W. Bousfield, M. Toivakka, Slot die coating of nanocellulose on paperboard, TAPPI J. 17 (1) (2018) 11–19, https://doi.org/10.32964/TJ17.01.11. [28] S.M.M. Mousavi, E. Afra, M. Tajvidi, D.W. Bousfield, Dehghani, M. Firouzabadi, Cellulose nanofiber/carboxymethyl cellulose blends as an efficient coating to improve the structure and barrier properties of paperboard, Cellulose (2017) 3001–3014. Ú. Fillat et al.