Effect of Overlapping Eco-Friendly Cellulose Nanobrils and Nanoclay Layers on Mechanical and Barrier Properties of Spray-Coated Papers

This work proposes to evaluate the effect of spray-coating in papers using eco-friendly cellulose nanobrils (CNFs) and nanoclay (NC) on mechanical and barrier properties for application as reinforced bags. Sack kraft papers of 60 g m - ² (C60) were coated with CNFs + CNFs/NC in 4 layers (L5), 40 g m - ² of CNFs + CNFs/NC in 3 layers (L4), 30 g m - ² with CNFs/NC in 2 layers (L3) and 10 g m - ² of CNFs in 1 layer (L2), and compared to uncoated sack kraft papers with basis weight of 60 g m - ² (C60), 80 g m - ² (C80) and 120 g m - ² (C120). The coated papers L2; L3; L4 and L5 obtained a decrease in water vapor transmission rate (WVTR) of 3.5%; 17%; 14% and 14%, respectively, when compared to C60. Comparing L2 and L3, CNF layer induced lower contact angles on the coated paper than CNF/NC layer. When compared coated papers with C120, it was observed an increase of around 66% in tensile strength for L2, around 44% for L3, and decrease of ~ 18% for L5 coated papers. L4 achieved the same tensile strength (when divided by basis weight) than C120. L2 and L3 coated papers led to the highest values of Young’s modulus, with increase of 56% and 38%, respectively, when compared to C60. Spray-coating in the present conditions improved the mechanical and barrier properties of the coated papers, being a possible alternative to produce papers with lower basis weight and using renewable raw materials.


Introduction
Paper is one of most widely used materials and has a variety of uses. Kraft paper is produced from wood chips, which are themselves composites, and are due to their high strength, used in packaging applications. Traditionally, the function of packaging is to protect products during transport, storage and commercialization, besides providing information to the consumer about the packed product. Currently, paper is the most important material for packaging industry due to its low price, renewability, recyclability, abundance (Lavoine et al. 2014) and mechanical characteristics such as high strength and exibility (Afra et al. 2016). However, the use of standard paper in packages is limited to products that do not require some impermeability characteristics. Efforts are ongoing to reduce their water vapor permeability, and several coating techniques have been proposed in literature to modify the paper surface. Among NC is an originally powdered and layered structured material that may improve the barrier properties of polymeric lms when properly exfoliated and incorporated. Their interlayer distance are in less than 100 nm when exfoliated. CNFs may form a large network structure of hydrogen bonds due their high aspect ratio and high surface area, and consequently, may increase mechanical strength and decrease water vapor and gas permeability of the paper samples. Notably, NC and CNFs have great potential to be applied in multilayered paper-based packages. Barrier properties of nanoclays and mechanical and barrier properties of CNFs may overcome some challenges of the paper-based packages.
The NC from montmorillonite family is a hydrophilic clay formed by layers of silica tetrahedral and alumina octahedral sheets in a 2:1 ratio (Floody et al. 2009). Due to their low cost and abundance, they have been widely studied for several applications such as a drug-delivery carrier ( The aim of this paper was to evaluate the effect of spray coating for producing different layers composed of CNFs and NC, varying their order and mixture composition of CNFs + NC, on tensile and barrier properties of the coated sack papers.

Material obtainment
The paper used as the substrate for coating was the commercial sack kraft paper (extensible) with nominal weight basis of 60 g m -² (C60), 80 g m -² (C80) and 120 g m -² (C120). Sheets of commercial eucalyptus bleached kraft pulp (Klabin S/A, SP-Brazil), was the source to produce the CNFs. The nanoclay (NC) used in the coatings (mixed to CNFs) were obtained from Sigma-Aldrich Inc. NC is considered as a hydrophilic material from bentonites family (H 2 Al 2 O 6 Si) (Sigma-Aldrich Inc.) and was used along with CNF suspension.

Production of the CNFs
Eucalyptus bleached kraft pulp was hydrated and dispersed for 24 h in deionized water in 2% wt concentration and dispersed under mechanical stirring at 1000 rpm for 3 h. The suspension was submitted to 30 passages through the brillator SuperMasscolloider (Masuko Sangyo MKCA6-3, Japan), which has a xed and a rotating disc (1500 rpm) that promote the cellulose brillation by shear forces, resulting on a gelatinous aspect. The brillation process followed the guidelines suggested in previous works (Tonoli et

Preparation of the NC
A small fraction of 3% wt of hydrophilic NC from bentonites family was dispersed in deionized water (w/w) and magnetically stirred for 24 h. The mixture was soni ed for 10 min following the methodology described in Bardet et al. (2015). This step was important in exfoliating the NC plates and facilitate its incorporation in the coating mixture ( Figure 2) besides to promote better adhesion to the paper used as substrate.

CNF and CNF/NC solutions
The CNFs and CNF/NC solutions were prepared in order to spray it on the paper surface. For CNF solution, 1.4% wt CNF was added to deionized water, and then stirred at 900 rpm for 30 min. CNF/NC solution was prepared adding 1.0% of exfoliated NC and 1.4% of CNFs in deionized water, and stirred at 900 rpm for 30 min in order to intercalate the nanoclay plates in the cellulosic matrix. Some studies have shown the suitable interaction of polymers and inorganic clays (Gabr et al. 2013). A good exfoliated nanoclay can be intercalated in nanocellulosic matrix, allowing a close interaction with the nanocellulose net (Gaikwad and Seonghyuk 2015).

Experimental setup for the spray coating
CNFs and CNF/NC suspensions were sprayed onto paper, with basis weight 60 g m -², using a laboratory assembled spray coater ( Figure 3A) in multiple layers. Variables and parameters used in this process were pressure of 0.4 MPa, for 30 s of spraying time and 30 cm of distance between spray nozzle and substrate. According to Shanmugam et al. (2017), the operational range of CNF suspension concentration for spraying coating is between 1% and 2%. Below 1% wt, the suspensions are too dilute and does not form a continuous and homogenous layer over the substrate, and high concentrations of CNF suspension (above 2.0% wt) are too viscous to be applied by the spray technique. Based on that, suspensions with 1.4% wt of CNFs and 1.0% of nanoclay were prepared along with suspensions of only 1.4% of CNFs.
The paper sheets were coated over a vacuum dewatering setup to obtain a consistent removal of the water from the coated papers before oven drying. After each coating step, papers were oven-dried at 100±5 °C for 3 h. The different paper treatments are depicted in Figure 3B. The samples were conditioned at 20 °C and 50% of relative humidity (RH) prior to characterizations.

Production of CNFs and CNF/NC lms
Films were produced from the CNFs and CNFs/NC by casting method. An aliquot of 40 ml of CNF suspensions (1.4% wt) and CNFs/NC (1.4% of CNFs and 1.0% of NC, both wt) was deposited on petri dishes and oven dried at 60 °C for 48 h. These concentrations of suspensions were calculated aimed to produce lms with basis weight similar to that applied as layer on the papers which were around 20 g m -² for CNF/NC lms and 10 g m -² for CNF lms.

2.7.Thickness and basis weight of the papers and lms
The thickness of the papers was measured before and after the coating process using a digital micrometer, according to the ASTM D645-97 (2007) standard. Basis weight was determined using samples with diameter of 16 mm, previously oven dried at 100 ± 3 °C and weighed according to the ASTM D 646-96 (1996) standard. Thickness and basis weight of the lms were also calculated. The basis weight was measured before and after the spray coating process according to Eq. 1.

Bw = M / A (1)
Where: Bw is the basis weight (g m -² ); M is the mass of the sample (g); A is the area of the sample (m²).

Tensile properties of the papers and lms
Tensile properties were determined using a universal testing machine MTS Landmark servo hydraulic test system, according to ASTM D 828-97 (2002) standard. The specimens were cut parallel to the ber's direction of the paper. Data were acquired using the Station Manager software. The dimension required for the specimens was 20 mm in width and around 100 mm in length. The distance between the grip clamping zones was 50 mm and the dislocation rate was 14.1 mm/min. Tensile strength, Young's modulus and elongation at break of the lms and papers were divided by the basis weight in order to normalize the results. The average value for each property was obtained by the average of ve samples.

Water vapor transmission rate of the papers
The water vapor transmission rate (WVTR) measurements were performed according to the method described in ASTM E96 / E96M-16 (2016) standard, followed by Guimarães Jr et al. (2015). The test was conducted in a laboratorial apparatus ( Figure 4) composed of permeability cells with relative humidity of 0% placed inside a desiccator with of relative humidity of 75%. A saline NaCℓ solution was used to keep the relative humidity at 75% into the desiccator. The permeability cells were lled with silica in order to keep the initial humidity at 0%. The samples of lms and papers were placed in the cap of the cells. In addition, reference cells were allocated into the apparatus to reduce the experimental error; reference 1cell with aluminum sheet instead of the paper sample and reference 2 -cell without silica. The mass was daily monitored for eight days. The results of WVTR were obtained by the average of ve samples per treatment.

Scanning electron microscopy of the papers
Surface and cross section of the papers were assessed by scanning electron microscopy (SEM) in order to evaluate the structure of the layers deposed on the paper surface. A Zeiss LEO EVO 40 XVP microscope with secondary electron detector and accelerating voltage of 20 kV was used and the samples were coated with gold before the analysis by sputter coating technique.

Contact angle of the papers and lms
The contact angle of the papers and lms substrates was carried out by depositing drops of water on the samples surface. A Kruss Advance equipment, equipped with a CCD camera working at up to 200 images per second was used for the measurements. Results were obtained by an average of six experimental contact angles, detected in the rst microseconds where the drops become stable on the sample surface. showed a heterogeneous coating, with a porous and rough surface due to the large amount of coating material on the paper surface. Samples of the coated paper L4 are not shown here as the external surfaces are the same shown for L2 (Figures 5c and d).

Results And Discussion
SEM images of the papers cross section showed that spraying layers of CNFs and CNFs/NC on the paper surfaces increased the thickness and also creating a tortuous pathway for any substance to cross the paper. The uncoated paper C60 showed many micropores throughout the cross section, which was extended to all coated papers ( Figure 6). The coated papers designated as L2, L3, L4 and L5 had more compact layers (external or internal) due to the coatings composed by CNFs or CNFs/NC.

Thickness and basis weight of the papers and lms
Thin CNF layers are su cient to change the surface properties of the coated sack kraft paper, while thicker and uniform layers are required to change barrier properties (Brodin et al. 2014). The CNF and CNF/NC layers were deposited on the sack kraft papers by the spraying method, and the increase of thickness and basis weight were proportional to the number of layers deposited. Each layer of CNFs or CNFs/NC increased the basis weight of the paper by about 10 g m -² -20 g m -² ( Table 1).
The lms produced from pure CNFs and CNFs/NC presented similar basis weight than each coating layer deposited on the coated papers (L2, L3, L4, and L5 treatments).

Tensile strength of the multilayered papers
The coated papers had different values of basis weight, therefore the mechanical results were also divided by the basis weight value of the correspondent treatment (Table 2), in order to normalize the comparison among the properties of the papers. Table 3 presents the mechanical properties of the lms produced with CNFs and CNFs/NC.
CNFs usually form a network in nanoscale and have more brils per gram of material when compared to bers in microscale, and consequently can form more inter-bril bonds (Brodin et al. 2014). All the papers decreased in tensile strength values as compared to C60 (uncoated), when analyzing the normalized values (divided by basis weight). Lavoine et al. (2014) studied the mechanical properties of CNF coated papers by two methods of coating, bar coating and size press. The two methods studied revealed higher tensile strength for uncoated paper than CNF coated papers. The authors associated this behavior to four factors: (1) non-uniformity of coating layers; (2) insu cient coat weight of CNFs to promote improvements on mechanical properties; (3) CNFs did not penetrate into paper structure; and (4) greater amount of water penetrated in the paper structure instead CNFs.
Sack kraft paper (C120 = 120 g m -²) is usually used for production of packages for grains and seeds weighing up to about 60 kg. Then, the present results observed for coated papers were comparable to C120 uncoated paper. There was an increase of around 66% in tensile strength for L2 and around 44% for L3, when compared to C120. The L4 coated paper showed the same average value of tensile strength presented by C120. On the other hand, normalized tensile strength of L5 was around 19% lower than C120. The coating process performed for preparation of L5 treatment subjected the substrate to the highest number of wetting/drying cycles during the coating application. Successive wetting/drying cycles may induce some horni cation process to the paper bers. Swelling occurs in the cell volume when cellulose ber is placed in suspension with water, and when they are dried, their volume shrinks. However, when these bers were re-suspended in water, the original volume is not recovered, decreasing the bers capacity of water retention. Moreover, when water is removed from the bers, part of cell wall collapse, increasing inter-brillar bonding, forming irreversible and partially reversible hydrogen bonds (Ferreira et al. 2017). Besides, the pore volume is reduced modifying the pore size distribution of the paper (Hubbe et al. 2007), which leads to the increase of stiffness of the paper bers. Cao et al. (1999) compared papers produced from once dried bers and never dried bers, concluding that the once dried bers was less conformable and, consequently the paper made from them resulted in lower strength. The same behavior was observed for the L5 coated paper of the present work. Furthermore, the high standard deviation obtained for basis weight of the L5 paper suggests a heterogeneous coating that did not increase the tensile strength of the paper. Afra et al. (2016) reported an increase in tensile index for CNF coated papers proportional to the added CNF basis weight. Syverud and Stenius (2009) studied mechanical properties of CNF coated paper and reported 6% of decrease on tensile index with 2 g m -² of CNF addition compared to uncoated paper. However, the authors reported that application of thicker layers (4 and 8 g m -²) promoted 6% and 14% of tensile index increase, respectively. All coated papers showed an increase of Young's modulus in relation to the uncoated papers. Similar result was reported in Afra et al. (2016), with CNF coating layers on paper sheets, relating that stiffness is proportional to the applied CNF basis weight. On the other hand, Lavoine et al. (2014) observed a decrease in Young's modulus after CNF application on the paper. The authors observed a decrease of around 20 -45% in Young's modulus with the coating composed of 3 -14 g m -² of CNFs. The values of Young's modulus were divided by the basis weight to normalize the comparison among treatments. In this work, spraying a layer of 10 g m -² of CNFs (L2) and around 20 g m -² of CNFs/NC (L3) resulted in a great increase in Young's modulus (56% for L2 and 28% for L3). In other cases, the application of greater basis weight of CNFs/NC (57 g m -² for L5 and 42 g m -² for L4), a considerable decrease in Young's modulus (30% and 7% for L5 and L4, respectively) was observed, in comparison to C60. The elongation at break for the coated papers reduced after the coatings of around 60%, 50%, 50%, and 40% for L2, L3, L4 and L5, respectively, in relation to uncoated papers. Sack kraft paper is constituted by many micro wrinkles that confers extensibility properties to the paper. In contrast, CNFs and CNFs/NC form a layer that restrict the extensibility and decrease elongation of the paper. L4 and L5 coated papers presented curves with a second mechanical performance due to the presence of the deposited layers. During the tensile loading, the break of different layers do not occur at the same time. The layers composed by CNFs and CNFs/NC, which are stiffer, are responsible for the initial mechanical performance ( rst rupture), and then the paper substrate is nally mechanically broken (Figure 9).

Contact angle of the papers and lms
Measuring the contact angle of a water drop on the paper substrate surface gives an insight into the hydrophobicity of the surface, which can be de ned as the tendency of a water drop not to spread on the substrate. CNF coating layers induced the decrease of the contact angle (87°) for L2 coated papers, in comparison to uncoated paper C60 (111°). However, this difference was not observed for the L3 treatment that was coated with CNFs and CNFs/NC (110°) (Figure 10). Greater basis weight did not indicate lower hydrophobicity for the papers C80 and C120.
The lm produced from CNFs/NC showed greater contact angle in relation to the CNF lm, con rming that CNF/NC coating seems to be more effective for decreasing hydrophobicity. The hydrophilic nature of CNFs, especially after the cell wall brillation process is a determining factor in its application as paper coating. NC are formed by layers of tetrahedral silica sheets and alumina octahedral sheets containing Ca 2+ or Na + that have the capacity to adsorb water (Paiva et al. 2008;Silva and Ferreira, 2008), whereas when adequately dispersed into the polymeric matrices, the nanoclays presents good barrier properties (Gabr et al. 2013;Gaikwad and Seonghyuk 2015). It was observed that the L5 coated paper resulted in different contact angles, when evaluating the internal and external faces of the coated paper, even if the layer composition is the same on both sides. This may have occurred due to some heterogeneity of the layers deposition on the paper surface, with regions with excessive deposition of coating, and others with insu cient coating. The surface roughness of the different deposited layers may also have some in uence on the hydrophobicity. Although L3 and L5 have a rougher surface compared to L2, they tended to have greater contact angles and this was probably due to the presence of NC, which sometimes caused agglomerations and surface granularity. Low values of contact angles (<90°) indicate that the material have lower hydrophobicity, while high contact angles (>90 o ) indicate low hydrophobicity (Yuan and Lee 2013).

Water vapor transmission rate of the lms and papers
Films produced with CNFs/NC showed less vapor transmission compared to lms produced with only CNFs (Figure 11). The addition of NC to the composition of the CNF lms resulted in reduction of around 11% on the WVTR evaluated in the study. Generally, the incorporation of NC into CNFs is expected to retard the degradation process, due to lower moisture present due to the formation of silicate layers as a protective barrier, and/or via inducing obstacles that could delay the passage of water vapor When successfully dispersed in polymeric matrix, NC can hinder the diffusion trough the matrix due the tortoise path created by exfoliated plates that could improve barrier properties (Gaikwad and Seonghyuk 2015). The NC grains adsorb water, which expand and form barrier layers against vapor transmission (Vartiainen et al. 2010). As previously discussed, CNFs are hydrophilic materials with high aspect ratio and strong capacity to form close networks (Dufresne 2013) that may promote the decrease of WVTR.
CNF/NC coating layers led to lower WVTR than pure CNF coating layers. SEM micrographs (Figure 5a to h) demonstrated that the CNF and CNF/NC layers reduced the porosity of the paper substrate, contributing to WVTR reduction. Reduction of WVTR was observed for all multilayered papers when compared to uncoated samples (C60, C80 and C120). Analyzing the uncoated papers with different basis weight, it is noted that greater basis weights did not result in lower WVTR, indicating that the real in uence on this property is exerted by the composition of the paper/ lm and coating agent. The coating with 10 g m -² of CNFs on 60 g m -² sack kraft paper (L2 coated paper) promoted a decrease of around 4% in WVTR in relation to C60, leading to a homogeneous less porous surface that may be further coated with other hydrophobic and barrier agents in the future.
L3, L4 and L5 presented lower WVTR compared to L2 and uncoated papers (C60, C80 and C120). In this case, the greater number of layers deposited on the paper surface may have created more obstacles against the vapor passage. Results showed that L3 coated paper led to a reduction of around 17% for WVTR in relation to the 60 g m -² uncoated paper (C60). Additionally, L3 led to the greater contact angle among all the coated papers (see Fig. 15), which means a better combination of layers for this treatment composition, including for mechanical properties. The coated papers L4 and L5 presented a reduction of around 14% for WVTR in comparison to the reference uncoated paper (C60).

Conclusion
Commercial sack kraft papers were coated with overlapping layers of CNFs and CNFs/NC and were evaluated through their mechanical and barrier properties. Coated papers (except L2 treatment) which consists of only CNF layers, showed more e cient water vapor barrier when compared to the uncoated samples. SEM surface images of surface and cross-sections showed that the addition of CNFs and CNFs/NC reduced the paper surface porosity. The addition of CNFs and CNFs/NC also increased the hydrophobicity of the surface of coated papers when compared to control uncoated papers (C60, C80 and C120). The highest values of tensile strength and Young's modulus were observed for coatings with 10 g m -² of CNFs (L2) and 30 g m -² of CNFs and CNFs/NC (L3). Higher number of layers (4 to 5 layers) applied by spraying in the present conditions seems to be detrimental to mechanical strength and surface properties, and incremental to barrier properties. This is probably due to the challenges of the spray coating technology with just CNF and NC suspensions. The challenges are related to the horni cation effect of the bers in the coated samples when subjected to high number of wetting-drying cycles during the CNF/NC application, for the drying processing of the coated and multilayered papers. This work shows the high potential of CNFs and NC as coatings in packaging paper and the key is to obtain su cient exfoliation of NC, following by good dispersion of of CNFs and CNFs/CN. Further developments and optimizations on spray coating techniques or other application methods should signi cantly improve the mechanical and barrier properties of the coated papers for packaging using lower basis weight papers.

Declarations Con ict of interest statement
The authors con rm that there is no con ict of interest regarding the submission.        Representative scheme of the exfoliated NC intercalated in the CNF suspension  Typical SEM image of the surface of: A and B -uncoated 60 g m-² (C60) sack kraft paper and detail on the paper surface, respectively (white arrows indicate micropores); C and D -coated paper L2 and detail of the coated paper L2 surface, respectively; E and F -coated paper L3 and detail of the coated paper L3 surface, respectively; G and H -coated paper L5 and detail of the most external coated surface (CNFs/NC), respectively: white arrows indicate micropores on the paper surface