Possibility of Use of NCC-Reinforced Melamine-Urea-Formaldehyde Adhesive in Plywood Manufacturing

• The possibility of using nanocellulose (NCC) as a ﬁ lling substance for melamine-urea-formalde-hyde (MUF) adhesive was investigated for the process of manufacturing plywood. The adhesive mixtures were prepared with various nanocellulose concentrations. The amount of introduced ﬁ ller had a signi ﬁ cant effect on both resin and plywood characteristics. Fourier transform infrared spectroscopy (FTIR) did not show any major changes between experimental and reference variants. The viscosity of resin increased after the introduction of nanocellulose. The addition of NCC in the amount of 5 g and 10 g per 100 g of solid resin led to an improvement in bonding quality, modulus of elasticity and bending strength. Further increase of NCC concentration caused a deterioration of manufactured plywood properties. In summary, the addition of proper amount of nanocellulose resulted in manufacturing plywood with improved properties.

sized in tunicate sea animals and some bacteria (Henriksson and Berglund, 2007; Janardhnan and Sain, 2011). The attention received by cellulosic particles results from their high surface area, high stiffness and strength. Many studies on using nanocellulose (NCC) as an environmentally friendly modifi er of wood adhesives have already been carried out (Ayrilmis et al., 2016a;Gardner et al., 2008;Vineeth et al., 2019).
The effect of nanocellulose addition to amino resins was investigated by Veigel et al. (2012). They found that the modifi cation increased the viscosity of liquid adhesive and caused a signifi cant improvement in both particleboard and oriented strand board (OSB) strength properties. Studies conducted by Mahrdt et al. (2016) confi rmed that the introduction of cellulosic particles led to the increase in particleboard strength values. Furthermore, the infl uence of nanocellulose on multilayered wood-based material has also been investigated. Zhang et al. (2011) noted a major enhancement in plywood shear strength and the reduction in formaldehyde emission because of the silanized nanocellulose added to UF adhesive. However, studies conducted by Kawalerczyk et al. (2020b) did not confi rm the effect on the reduction of the harmful formaldehyde emission with the use of non-modifi ed nanocellulose. Ayrilmis et al. (2016b) concluded that cellulosic particles added to UF adhesive can reduce the VOCs (volatile organic compounds) emitted from LVL (laminated veneer lumber). The price of nanocellulose may be a limiting factor for some applications but there are many studies on fi nding the way to obtain it with more cost-effective methods (Babicka et al., 2020;Kawalerczyk et al., 2021).
Despite the fact that MUF adhesives are becoming more popular, especially for applications in kitchens, fl oors and some structural materials, and although they are increasingly replacing pure UF adhesives, no studies have examined the effect of their nanocellulose-reinforcement on the properties of manufactured plywood panels (Lei and Frazier, 2015). Thus, the aim of the present study was to investigate the effect of adding NCC to MUF adhesive on the plywood performance.

Materijali
Rotary cut birch (Betula L.) veneer sheets were purchased from the market with the dimensions of 320 mm × 320 mm × 1.3 mm, moisture content of (5 ± 1) %, without any defects, and they were used for the research purpose. The commercial MUF adhesive with the following characteristics: solid content of 64 -69 %, gel time at 100 °C of 63 s, viscosity between 1000 and 2500 mPa·s, density of 1.27 g/m 3 and pH of 9.5 -10.7 was purchased for the experiment. The 20 % aqueous solution of ammonium nitrate ( NH 4 NO 3 ) was added as a hardener taking into account the environmental aspects (Aras et al., 2015). In order to adjust

UVOD
Nanotechnology is a relatively new science that has already become the frontier of the 21 st century. It is defi ned as the fi eld focused on the control and understanding of the matter at a dimensional range between 1 and 100 nm (Hulla et al., 2015). From the beginning, nanotechnology has been playing the role of a scientifi c platform joining and linking together various disciplines (Szczesna-Antczak et al., 2012). The objects with the dimensions between 1 and 100 nm have unique mechanical, optical, magnetic and electrical properties (Wegner and Jones, 2005). Thus, the nanoscience also provides many opportunities for woodbased materials manufacturing (Candan and Akbulut, 2013). It seeks to develop advanced materials having signifi cantly improved chemical, physical properties and functions (Wegner and Jones, 2006).
Wood-based materials such as plywood, particleboards, medium-and high density fi berboards are becoming more popular in recent years. The production of plywood, which is a multilayer composite made of glued wood veneers, reached about 157 million m 3 in 2017 and it is still growing (Bekhta et al., 2020;Sydor et al., 2020). Due to its favorable mechanical properties and dimensional stability, it founds an application in many industry branches (Kawalerczyk et al., 2019a). In structural applications, plywood is used in walls, ceilings and roof constructions; in furniture as a basic material for both upholstered and case furniture (Bekhta et al., 2009b;Majewski, 2019). Moreover, it is also appreciated as a packaging material due to its ease of processing, lightness and durability, and in transport as a fl ooring material or in trailers construction. The properties of plywood and consequently its application depend on the quality of veneers and the type of used adhesive (Kawalerczyk et al., 2020c;Mirski et al., 2011).
Amino resins are synthetic adhesives widely used in wood-based material industry. Their participation in the general range of wood adhesives market is estimated at 85 % (Jóźwiak, 2018;Kamoun et al., 2003). The three main types: UF (urea-formaldehyde), MF (melamine-formaldehyde) and MUF (melamineurea-formaldehyde) may be distinguished as the most commercially available (Gonçalves et al., 2019;Mirski et al., 2020). The widespread use of MUF adhesives in recent years results from the higher bond quality, water resistance and lower formaldehyde emission in comparison with UF resin (Lei and Frazier, 2015;Tohmura et al., 2001). The adhesives mixtures applied in plywood manufacturing have to contain various kinds of fi llers. They are insoluble, non-volatile substances added in order to adjust the viscosity, reduce raw material costs and increase bonding between the wood components (Kawalerczyk et al., 2019b;Ong et al., 2018).
In recent years an interesting concept of using cellulose as a modifi er for polymers has been increasingly investigated. It is the most important constituent of the plants cell walls and, moreover, it is also synthe--shear strength after 24 h of soaking in water at (20 ± 3) °C and after preparation including boiling in water for 6 h and cooling in water at (20 ± 3) °C for 1 h according to EN 314-1 (2004), -modulus of elasticity and bending strength parallel and perpendicular to the grains of face layer according to EN 310 (1993). On the basis of the results obtained in above tests, the proper composition of adhesive mixture was selected. The experimental adhesive labeled "N5" showing the best reinforcing effect was applied in the amount of 170, 160, 150, 140 and 130 (g/m 2 ) in order to investigate the possibility of reducing the adhesive consumption in plywood production. The pressing process was conducted at 140 °C with the unit pressure of 1.3 MPa for 4 min. Manufactured plywood panels were tested in terms of shear strength both after 24 h of soaking in water at (20 ± 3) °C and after pretreatment consisting of boiling in water for 6 h and cooling in water at (20 ± 3) °C for 1 h according to EN 314-1 (2004). The obtained results were compared with a reference variant, which was labeled "170 REF" in further part of the paper.
The evaluation of mechanical properties and bonding quality involved 12 samples of each variant. The results were subjected to the multivariate statistical analysis ANOVA. The Tukey test with a significance level of α = 0.05 was applied to distinguish the homogeneous groups with the use of Statistica 13.0 software. Figure 1 illustrates the time-viscosity dependence. The results indicate that nanocellulose-reinforced MUF resins were characterized by signifi cantly increased viscosity in comparison with reference mixture.

REZULTATI
The viscosity of adhesive mixture containing the maximal amount of cellulosic nanoparticles was 46 % and 33 % higher when compared to resin only fi lled with rye fl our initially and after 6 hours of measurements, respectively. The increasing values of all resins viscosity during the test time resulted from the progressive polycondensation reactions and the constant water absorption by hydrophilic fi llers. The viscosity of the glue mixture increased with the increasing percentage of nanocellulose addition similarly to investigations reported by Damásio et al. (2017). One factor that may and regulate the viscosity of adhesive mixture, the rye fl our was introduced in accordance with the industrial formulations. Nanocellulose, added as a modifi er, was purchased from Nanografi Nanotechnology Co. Ltd. (Ankara, Turkey). As declared by the producer, the dimensions of the particles were 300 -900 nm in length and 10 -20 nm in width.

Metode
A 10 % aqueous suspension of nanocellulose was prepared with the use of magnetic stirrer (600 rpm, 10 min), because cellulosic nanoparticles had to be processed in wet state. The experimental adhesive mixtures with the addition of NCC suspension, fl our and hardener were homogenized using CAT-500 homogenizer at 1000 rpm for 2 minutes. The control adhesive was prepared in accordance with an industrial formulation. The compositions of both experimental and reference mixtures are summarized in Table 1.
No additional water was added in experimental variants because the nanocellulose was introduced in the state of water suspension. The viscosities of adhesives mixtures and their changes in 6 h were investigated using Brookfi eld DV-II + Pro viscometer. In order to assess the chemical bonding between the resin and nanocellulose, Fourier transform infrared spectroscopy was carried out. Reference and experimental adhesive mixtures were cured at 140°C and grinded. The obtained powders with a fraction of 0.125 mm were mixed with KBr at 1/200 mg ratio. Spectra were registered using a Nicolet iS5 spectrophotometer (Thermo Fisher Scientifi c) with Fourier transform at the range of 500 -4000 cm -1 at the resolution of 4 cm -1 , registering 16 scans.
The adhesive mixtures were spread on the surface of external veneers in the amount of 170 g/m 2 . The veneer sets were assembled perpendicularly to each other. Three-layer plywood panels with the dimensions of 320 mm × 320 mm were produced in a hydraulic laboratory press with the following pressing parameters: temperature of 140 °C, unit pressure of 1.3 MPa and pressing time of 4 min. Three replicate panels were produced for all the test groups. Manufactured plywood was tested in order to investigate the following properties: -formaldehyde (HCHO) emission with the fl ask method according to EN 717-3 (1996) initially and after 30 days of panels conditioning while left open at (20 ± 3) °C and (65 ± 2) % relative humidity, have contributed to this major increase in viscosity is the highly hydrophilic nature of cellulose. It tends to absorb water due to its chemical structure -the surfaces of cellulose are covered in hydroxyl groups on the equatorial positions of glucopyranose ring (Yamane et al., 2006). Another factor is the increased chemical reactivity of nanoparticles in general (Shahbazi et al., 2013). The introduction of nanocellulose probably caused a substantial interfi brillar interaction. The number hydroxyl groups located on the surfaces of individual fi brils can cause a considerable interaction between them, which can lead to the formation of temporary bonds (Iotti et al., 2011). bonding quality. The adhesive characterized by too low viscosity penetrates extensively into porous surface of the veneer during application and pressing. Consequently, the layer remaining on the veneer surface is no longer suffi cient to ensure good quality of the bond (Sellers, 1985). The viscosity may also be a factor limiting the amount of nanocellulose added to the adhesive. The lack of water in the mixture can affect the crosslinking of adhesive and moreover, the addition of highly hydrophilic fi ller can prevent water from evaporating during the pressing process (Mahrdt et al., 2016;Réh et al., 2019). Furthermore, it is also hard to evenly spread the adhesive characterized by too high viscosity and adjust it to the applying equipment. In case of this research, both reference and experimental mixtures obtained values that allowed the application without any diffi culties.
Since the reinforcing effect of nanocellulose is associated with chemical bonding, the Fourier transform infrared spectroscopy (FTIR) was carried out in order to investigate the chemical interactions. The In the case of NCC sample, the peak at 1620 cm -1 was assigned to O-H vibration of absorbed water (Wulandari et al., 2016). The peak at 1320 cm -1 corresponded to C-H and C-O vibrations contained in the polysaccharide rings and it was observed at spectra of NCC and MUF+NCC. Spectra of MUF and MUF+NCC were characterized by a peak around 1560 cm -1 , which was assigned to the C-N stretching of secondary amines. Moreover, the peak at 1360 cm -1 was assigned to the C-N stretching of CH 2 -N witch confi rmed presence of amine groups (Luo et al., 2015). The peak at 1060 cm -1 corresponded to vibration from the pyranose ring (spectra of NCC) (Wulandari et al., 2016). In the case of triazine rings, the peak at 1190 cm -1 was observed. It was assigned to C-N stretching vibrations (Yuan et al., 2016). Spectra of MUF + NCC was characterized by occurrence of a peaks at 1060 cm -1 and 1120 cm -1 , which corresponded to pyranose rings and aliphatic ring, respectively (Luo et al., 2015). These peaks were not observed at MUF spectra. Presumably, it resulted from the fact that interactions occurred between nanocellulose and MUF resin chemical. The peak at 810 cm -1 was assigned to the typical stretching of the triazine ring of melamine and it was observed at spectra of MUF resin and MUF+NCC The transmittance spectra of reference and reinforced adhesive revealed mostly the characteristic functional groups of MUF resins. According to literature, the interfacial bonding of nanocellulose with resins can be attributed to the reaction between the methylol groups of resin and hydroxyl groups of cellulose (Singha and Thakur, 2008).
Glue line quality is one of the most important properties of plywood affecting both its physical and mechanical characteristics (Bekhta et al., 2009a). Thus, the shear strength test was carried out since it is the fundamental indicator of the adhesive performance in plywood panels (Bekhta et al., 2016;Rohumaa et al., 2013). The results of bonding quality are summarized in Figure 3.
In order to fully evaluate how changes in adhesive formulation infl uenced the mechanical properties of plywood, such as bending strength (MOR) and modulus of elasticity (MOE), investigations were made both parallel and perpendicular to the grains of the face layer. The results are presented in Figure 4.
On the basis of the research conducted, it was confi rmed that the key to obtain optimum reinforcement effect was the amount of added nanoparticles (Ferreira 2017). The major improvements in shear strength values were obtained in variants where the percentages of added NCC were 5 and 10 %. The best results were observed in case of variant labeled N5 and the increase was 38 % and 41 % in comparison with reference panels after soaking and after boiling, respectively. The introduction of 10 g of cellulosic nanoparticles also led to the increase of bonding quality compared to the mixture fi lled only with fl our. As expected based on the previous research, as the amount of nanocellulose increased, the shear strength value constantly decreased (Kawalerczyk et al., 2020b, a;Veigel et al., 2012). Bonding quality of panels labeled N15 was decreased by about 8 % both after soaking and boiling in water. However, all plywood samples met the requirements of EN 314-2 (1993) and their shear strength exceeded 1 N/mm 2 .
Mechanical properties, such as modulus of elasticity and bending strength (also called modulus of rupture) of the NCC-modifi ed plywood show signifi cantly porous, more solid and compact. The deterioration observed in bonding quality, MOR and MOE resulted from too high concentration of nanoparticles, which consequently led to the formation of agglomerates. It has particularly negative effect since the nanofi llers act like carriers of stress along the glue line and the occurrence of agglomerates cause the accumulation of load at certain points of the bond (Singha and Thakur, 2008).
A major disadvantage of amino resins is a harmful formaldehyde emission (Dziurka and Mirski, 2014). According to literature, there are three main sources of emitting formaldehyde from adhesives: hydrolytic degradation of cured resin, residual formaldehyde contained in the resin and condensation reaction between hydroxymethyl groups and other aromatic carbon or two hydroxymethyl groups (Tohmura et al., 2001). Since building materials including wood-based materials are considered to be one of the most common sources of formaldehyde emission, it is important to reduce it. The results of HCHO emissions are presented in Figure 5.
Plywood panels manufactured with the use of NCC-modifi ed resin were characterized by signifi cantly lower formaldehyde emissions. As the concentration of cellulosic nanoparticles increased, the amount of emitted HCHO was substantially reduced. Thus, the best results were observed in case of variant N15, where the decrease was 23 %. Studies conducted by Zhang et al. (2011) have shown very similar effects. Authors stated that the addition of modifi ed NCC led to reduced HCHO emission due to physical adsorption and chemisorption. Moreover, Li et al. (2015) hypothesized that the decreased formaldehyde emissions can be attributed to increased viscosity of an adhesive. According to the author's theory, the glue mixture characterized by low viscosity penetrates into the wood pores, which consequently increases the emission of plywood. It was also reported that the monomeric formaldehyde and polyoxymethylene glycols contained in the resin can easily interact with OH groups of nanoparticles (Candan and Akbulut, 2013; Dudkin et al., 2006). Furthermore, cellulose itself is able to irreversibly bound formaldehyde in small quantities (Bekhta et al., 2019;Kamath et al., 1985). Another reason for this reduction can be the unique characteristics of nanoparticles such as chemical activity, tremendous surface area or physical properties. Liu and Zhu (2014) explained that the decrease in HCHO emissions results from the ability of nanoparticles to absorb free formaldehyde from adhesives.
The resin costs are more than 60 % of the total wood-based materials manufacturing costs (Cao et al., 2018). Both the modifi cation of veneer surface and the modifi cation of adhesives are investigated factors that can possibly lead to the reduction of the binding agent consumption. The possibility to reduce the amount of applied resin was determined on the basis of shear strength test and the results are presented in Figure 6.
As the amount of applied glue mixture decreased, the bonding quality also decreased. The shear strength values were reduced since the quantity of applied resin was insuffi cient to fully and evenly cover the veneer surface (Bekhta and Marutzky, 2007). Introducing nanocellulose in the amount of 5 g per 100 g of solid MUF resin had a positive effect on its properties, durability and morphology, and consequently it allowed to reduce adhesive spread rate by 30 %. The experimental plywood glued with reference resin mixture in the amount of 170 g/m 2 was characterized by equally good shear strength as panels manufactured with the use of NCC-reinforced adhesive in the amount of 140 g/m 2 . The differences between these two variants were not statistically signifi cant and the p-value was 0.999020 and 0.999987 after soaking and after boiling, respectively. However, further decrease of adhesive spread rate led to a notable deterioration in bonding quality, but the results still met the requirements of EN 314-2 (1993). The modifi cation of resin seems to be more effective than surface modifi cation, e.g. veneer compression (Bekhta and Marutzky, 2007). Corresponding results were obtained by Dukarska and Czarnecki (2016) in studies concerning the nano-SiO 2 addition to MUPF (melamine-urea-phenol-formaldehyde) adhesive. The introduction of fumed nanosilica allowed to decrease the resin spread rate in plywood production by 30 %. In future, it would be interesting to investigate how the reduction of MUF adhesive consumption infl uences the emission of free formaldehyde.

ZAKLJUČAK
The time-viscosity dependence showed a significant increase after the addition of nanocellulose. As the amount of added nanomodifi er increased, the viscosity values also increased.
The Fourier transform infrared spectroscopy (FTIR) did not show any major changes between experimental and reference samples, thus it did not explain the chemical interaction between nanocellulose and melamine-urea-formaldehyde resin.
The addition of small amounts of nanocellulose (5 g and 10 g) led to the improvement in bonding quality and mechanical properties such as modulus of elasticity and bending strength. The best results were obtained in case where the concentration of NCC was up to 5 g per 100 g of solid resin.
The MUF adhesive modifi cation with nanocellulose caused a decrease in the amount of emitting formaldehyde.
The addition of nanocellulose allowed the reduction in adhesive consumption by 30 %.