Effect of The Enzyme Charge On The Production And Morphological Characteristics of Cellulose Nanobrils

The effect of cellulase enzymes on the degree of polymerization of cellulose and the mechanical brillation process has been widely reported. However, the available information does not allow to establish specic relationships between the applied enzymatic-mechanical treatment, the degree of polymerization, and the characteristics of the cellulose nanobrils (CNFs) produced. This work aims to establish specic relationships between the intensity of enzymatic treatment, the degree of polymerization of the cellulose, the morphology of CNFs, and the tensile strength of the lms. It was determined that the decrease in the degree of polymerization plays an important role in the brillation processes of the cell wall to produce CNFs and that there is a linear relationship between the degree of polymerization and the length of CNFs, which is independent of the type of enzyme, enzyme charge, and intensity of the applied mechanical treatment. In addition, it was determined that the percentage of decrease in the degree of polymerization of CNFs due to mechanical treatment is irrespective of the applied enzyme charge. Else ways, it was shown that the aspect ratio is a good indicator of the eciency of the brillation process, and the degree of polymerization in not. Finally, it was shown that the resistance of CNF lms is positively related to the degree of polymerization up to a maximum value which corresponds to the maximum of the aspect ratio.


Introduction
Due to abundance and sustainability, plant cellulose and cellulosic nanomaterials have attracted increasing interest as an alternative to synthetic materials, especially as llers and reinforcement for composite materials (Qing et al., 2013). Cellulose nanomaterials (CNM) comprise a broad spectrum of materials produced by the deconstruction of the cell wall. Due to the surprising and promising characteristics of cellulose nanomaterials: biocompatible and transparent material, excellent mechanical behavior at low weight, and very reactive due to the hydroxyl groups present on its surface (Lavoine et al., 2012;Grüneberger et al., 2014;Nechyporchuk et al., 2016) several researchers have focused their interest on the study of this nanomaterial and its potential uses at an industrial level: coating, polymeric reinforcement, 3D printing, rheological modi er, among others (Heggset et al. 2017; Albornoz-Palma et al. 2020b). The term '' cellulose nano brils (CNF) '' refers to brils with nanoscale widths (less than 100 nm) (ISO 2017), whose average length values are estimated in the order of several micrometers (Tanaka et al., 2015).
In previous decades, the challenge associated with the isolation of CNFs was to reduce the high energy demand required by the mechanical brillation process, reporting values between 5 y 70 kWh/kg (Spence et al., 2011). However, with the incorporation of pretreatment methods, such as chemical pretreatments (carboxymethylation, carboxylation, quaternization, sulfonation, and oxidation) or enzymatic pretreatments, energy consumption and production costs decreased, making CNFs a more attractive material for commercial applications (Saito et al., 2006;Siró and Plackett 2010;Nechyporchuk et al., 2016). However, research continues to focus on optimizing existing techniques and developing alternative methods that can bene t the production process or provide CNFs with new properties (Nechyporchuk et al., 2016).
The use of cellulase enzymes as pretreatment for the production of CNF has been one of the most studied, since it favors the process of deconstruction of the cell wall, reduces energy consumption, and facilitates the production of nano brils with more homogeneous and controlled dimensions (Pääkkö et al., 2007;Delgado-Aguilar, 2015;Tarrés et al., 2016;Hu et al., 2018). In addition, with the use of enzymatic hydrolysis, high yields are favored, the environmental impact is reduced, and it is possible to produce materials suitable for biomedical applications (Ribeiro et al., 2019).
Cellulases are groups of enzymes that catalyze the breakdown of cellulose polymer into smaller polymer chains or even cellobiose and glucose. Traditionally, these enzymes are divided into three groups: endoglucanase, exo--1,4-glucanase or cellobiohydrolase (CBH), and -glucosidase (cellobiase) (Bhat and Bhat 1997). In terms of CNFs production with enzymatic pretreatment, endoglucanases are the most interesting, considering their action is randomly focused on the amorphous regions of cellulose (speci cally on the β-1,4 bonds corresponding to C1 of the rst glucose unit and C4 of the after unit), breaking the cellulose chain into polymers of shorter length (Delgado-Aguilar, 2015; Ribeiro et al., 2019). Nechyporchuk et al. (2015) compared CNF production using monocomponent endoglucanases and a mixture of endoglucanase, exo--1,4-glucanase, and cellobiase and showed that monocomponent endoglucanase has a better effect on the separation of nano brils while inducing less depolymerization of cellulosic chains.
The degree of polymerization (DP) is de ned as the number of times that the monomeric unit that forms the polymer chain repeats (Delgado-Aguilar 2015). This is an important parameter that evaluates the length of cellulose chains and is frequently used to evaluate produced CNFs (Qing et al., 2013). Endoglucanases randomly cleave amorphous regions of cellulose chains and are characterized by drastically decreasing the degree of polymerization, but they slowly release soluble sugars from crystalline regions (Ek et al., 2009). and constant stirring at 800 rpm with a Stirrer Type BS. After the reaction time, the enzyme was denatured at 80°C for 20 minutes. The pulp was re ned at 46,000 revolutions in a PFI mill at 10% consistency. After the mechanical treatment, the pulp was separated and subjected to 5 enzymatic pretreatments with the same conditions of the rst enzyme pretreatment, varying only the enzyme percentage: 0%, 0.025%, 0.05%, 0.075%, and 0.1% with respect to the dry weight of the pulp.

High pressure homogenization
Suspensions in water of the treated bers were prepared, according to the sequence of mechanical and enzymatic pretreatments described in point 2.2.1 at a consistency of 0.5%. These were subjected to a homogenization process at a pressure 700 bar in a Gea Niro Soavi Panda Plus 2000 homogenizing equipment provided with S-type impact head. The treatment consists of passing the ber suspension through the equipment 15 times, in order to produce 5 types of CNFs that differed only in the enzyme charge of the second enzyme pretreatment

Transmittance of CNF dispersions
The transmittance was determined on CNF dispersions at 0.1%, using a Genesys UV10 spectrophotometer, at a wavelength of 800 nm, using distilled water as a reference, according to the methodology presented by Delgado-Aguilar (2015).

Intrinsic viscosity
Intrinsic viscosity was determined from the viscosities of CNF dispersions, according to Albornoz-Palma et al. (2020a). For measurements, a Brook eld LVDV-I + viscometer was used with a ULA spindle (Ultra Low Adapter) spindle, which con guration corresponds to a double-cylinder geometry, from sample at different concentrations (0.02% (w/v) to 0.08% (w/v)). Each of the samples was heated in a Julabo SW22 thermal bath at 23ºC for 2 hours before the measurement. Samples did not show thixotropic behavior and reached a steady state in less than 20 s. The measurement conditions were to 23ºC and shear rate of 73.38 s -1 .

Morphological characteristics of CNFs
The average length of CNF dispersions ( L ̅ ) was measured using S3500 Laser Diffraction Particle Size where p is the density of NFCs (1.6 g/ml) and is the aspect ratio of CNFs (L ̅ /d ̅ ).

Mechanical properties of CNFs
The maximum load stress (N) supported by rectangular samples (2 cm wide, 8 cm long) made from CNFs lms (30 g/m 2 ) was determined in the tension testing equipment (TestResources Inc, USA). The separation of the clamps in the specimens during the measurement was 2 cm.

Mechanical-enzymatic pretreatment
Cellulolytic enzymes are glucoside hydrolases that break the β-(1à4)-glucosidic bonds of carbohydrates by inversion or retention of the anomeric carbon con guration (Ek et al. 2009). Table 1 shows the percentage distribution of the solid residue and the fraction of solubilized carbohydrates from the eucalyptus pulp after the enzymatic pretreatments. As expected, the enzyme complex solubilized part of the hemicelluloses and cellulose of the raw material, whose initial chemical composition was: 77.7 ± 0.5% cellulose, 21 , 2021). The rst enzymatic treatment (2º-0.000%), which was identical for the 5 CNFs, solubilized a lower amount of cellulose and xylan than the pulps with double enzymatic treatment, so the carbohydrate yield in the solid fraction was higher. This is due to the fact that in the rst enzyme treatment the accessibility of the enzyme to the substrate was lower than the second. Both the mechanical re ning process, as well as the rst enzymatic treatment caused greater internal and external brillation of the bers, increasing the exposed surface area of the ber and the absorption of water, which facilitated the accessibility of the enzyme for second hydrolysis enzymatic.
The percentage of carbohydrates in the solid fraction decreases as the enzyme charge of the second enzyme treatment increases from 0% to 0.1%, which is mainly due to the hydrolysis of amorphous cellulose (Pääkko et al., 2007). Furthermore, in the second enzymatic treatment, the degradation of xylans was statistically the same for all samples, so the difference in the decrease in solid yields is clearly due to the hydrolysis of cellulose.

Cellulose nano brils production
Determining the morphological characteristics of CNFs is relevant to understand the nal properties of this nanomaterial. Table 2 shows the morphological characteristics of the CNFs produced from pulps with different enzyme charges and the same mechanical treatment. Regarding the width, length, and degree of polymerization of the CNFs, a decrease in these characteristics is observed as the enzyme charge increases, which can be attributed to a greater hydrolysis of the cellulose chains due to an increase enzyme concentration. Endoglucanases decrease drastically the degree of polymerization, which would increase the frequency of rupture or breakpoints within the bers, facilitating the access of water to the interior of the cell wall and consequently the process of mechanical brillation. The transmittance of light at a speci c wavelength through a CNF suspension indicates the presence of smaller and/or more homogeneous nano-objects and is often used as an indirect method to estimate the degree of brillation of CNF dispersions (Delgado-Aguilar, 2015; Tarrés et al., 2016). According to the above, the results of Table 2 show that the increase in the enzyme charge effectively allows greater brillation of the BHKP, which is re ected in the decrease in the lengths and widths of the CNFs. From a technological point of view, it is often interesting to produce brils with morphological characteristics that favor certain applications. For example, it is expected that a higher aspect ratio will favor the reinforcing properties of composite materials.
When observing the aspect ratio of the nano brils (Table 2) it can be clearly seen that this parameter does not have the same tendency as the transmittance of light and that its variation is affected by the enzyme charge during treatment, increasing with the enzyme charge a maximum value and then decreases. With the aim of getting a better understanding of how enzyme charge affects the morphological characteristics of CNFs, the relationships are shown in Figure 1.
For the length of the CNFs (Figure 1a) a negative linear relationship with the enzyme charge is observed, whose differences are statistically signi cant between the different samples (LSD method, 95% con dence). These results prove that a higher enzyme charge can generate signi cant changes in the lengths of the CNFs for the same mechanical treatment. In this case, a charge of 0.1% of enzyme (in the second enzymatic treatment) generates a 33% decrease in CNFs lengths.
As mentioned by Taheri and Samyn (2016), the minimum width of the brils produced during brillation depends on the operating conditions and the equipment used for mechanical brillation. Analyzing the widths of the CNFs (Figures 1b), it is observed that there is a drastic and signi cant decrease (p value <0.05) at low enzyme charges (<0.0005 ml/g). For enzyme charges greater than 0.0005 ml/g, the decrease in width is much less pronounced and there are no statistically signi cant differences (LSD method, 95% con dence). These results suggest that, for the same mechanical treatment, an increase in the enzyme charge facilitates brillation up to the minimum width given by the equipment. In this sense, there is evidence of the need to de ne a control parameter that encompasses the morphological properties of CNFs and optimize the enzymatic pretreatments for each mechanical process.
The morphological parameter that best re ects the concept of brillation corresponds to the aspect ratio of the CNFs, since it has been shown that an increase of the enzyme charge favors brillation as a result of the transversal break of the brils, reducing the length of these, and the longitudinal break that decrease the width until a minimum value. As a consequence of the above, the aspect ratio of the CNFs presents a maximum (p = 311) (Figure 1c) for an enzyme charge, in the second enzyme treatment, of 0.0005 ml/g. An e cient brillation process seeks to produce a homogeneous material, with small widths (<50 nm), and attempting to decrease transversal break of the brils (length). Therefore, the aspect ratio is the parameter that best re ects these characteristics and corresponds to a good indicator of the effectiveness of the brillation process.
On one hand, Andrade et al. (2021) determined that for CNFs produced from the same raw material and with the same enzymatic-mechanical treatment, the aspect ratio was 330. On the other hand, Albornoz-Palma et al. (2020a) in their study produced CNFs with an aspect ratio of 303, from the same raw material and mechanical treatment, but with a different type and charge of enzyme.

Effect of mechanical-enzymatic treatment on the degree of polymerization of cellulose
The degree of polymerization is an important parameter that evaluates the length of cellulose chains and is frequently used to evaluate CNFs (Qing et al., 2013). Table 3 shows the variation in the degree of polymerization as a function of the applied treatment. As mentioned above, endoglucanase enzymes are characterized by a rapid decrease in the degree of polymerization (Ek et al. 2009).
The "Variation 1" in Table 3 indicates the decrease in DP with respect to the initial raw material (P), which decreases as the intensity of the mechanical and/or enzymatic pretreatment applied increases. The DP of cellulose decreases between 72 and 81% for the CNF produced, with DP values between 278 and 447, which coincides with that reported by various authors who produced CNF with similar enzymatic and mechanical pretreatments. Albornoz-Palma et al. bar) and produced CNFs with a DP of 280 when dosing 3 FPU of enzyme/g ber. From the above, it seems that the lower limit of the degree of polymerization is 220 for CNFs with enzymatic pretreatment with a high degree of brillation. The "Variation 2" in Table 3 represents the decrease in the DP of the CNF after the homogenization process, with respect to the pulps with mechanical-enzymatic pretreatment. The results show that the decrease in DP due to the homogenization process is independent of the enzyme charge, since for the same mechanical treatment (15 passes through the homogenizer) the decrease in DP is close to 40% for all cases. As the relationship between length and degree of polymerization is positive linear (Figure 2a) and the relationship between length and enzyme charge is negative linear (Figure 1a), width and aspect ratio show the same trends as a function of the degree of polymerization than with the enzymatic charge, but in a specular way. Because of the above, the width of the CNFs (Figure 2b) at DP less than 362 did not show statistically signi cant differences (LSD method, 95% con dence). On the other hand, at DP greater than 362, the differences are signi cant, and the DP of the samples varies by up to 38%. The relationship between the aspect ratio and the DP has a maximum at DP = 362 (Figure 2c). This maximum is due to the fact that DP less than 362 there is a change in the lengths of CNFs but not in the widths, so the aspect ratio decreases.
In the literature, it has been observed that a higher degree of polymerization is strongly related to the improvement in the mechanical properties of nano brils (Zimmermann et al., 2010;Tarrés et al., 2016).
Regarding the maximum load supported by the CNFs lms (Figure 2d), it is observed that for DP values greater than 362, there are no statistically signi cant differences (Bonferroni method, 95% con dence). For DP less than 362, the maximum load decreases, and the values are statistically different (Bonferroni method, 95% con dence). When relating these results to the enzymatic charge and the aspect ratio, it can be seen that for charges less than or equal to 0.0005 ml/g (DP 362), where the aspect ratio is increasing to its maximum, the resistance of the CNF lms remains constant. For enzyme charges greater than or equal to 0.0005 ml/g (DP 362), where the aspect ratio is decreasing, the resistance of the CNF lms decreases. The above shows that the resistance of CNFs lms is positively related to the degree of polymerization up to a maximum value, which corresponds to the maximum value of the aspect ratio. This is why the aspect ratio is the parameter that best predicts the nal mechanical properties of the CNFs.

Conclusions
The enzymatic pretreatment facilitates the deconstruction of the cell wall of the bers. However, at high enzyme charges, the aspect ratio of the CNFs shows a drastic decrease, due to ine cient brillation due to the transversal break of CNFs.
Moreover, it was shown that the DP is linearly related to the length of the CNFs, independent of the type of cellulase enzyme, enzymatic charge, and intensity of the mechanical treatment applied. Furthermore, the percentage decrease in DP, for the same intensity in the mechanical treatment, is independent of the enzymatic charge.
Finally, the maximum load supported by CNF lms is linearly related to the DP up to a maximum value, corresponding to the maximum value of the aspect ratio. The aspect ratio proved to be the parameter that best re ects the nal mechanical properties of the CNFs and the e ciency of the brillation process.