*3.2. Cellulose Nanofiber Isolation and Characterization*

ɳ

The OTPB pulp was bleached to eliminate the lignin content while maintaining the carbohydrates in the fiber (hemicellulose and cellulose) and thus study the effect of lignin on the production of cellulose nanofibers through different pretreatments. Unbleached and bleached pulp were used for the production of lignocellulose nanofibers (LCNF) and cellulose nanofibers (CNF), respectively, through two different pretreatments, mechanical (Mec) and TEMPO-mediated oxidation (TO). The characterization of the different cellulose nanofibers in terms of nanofibrillation yield, transmittance, cationic demand, carboxyl content and morphology is shown in Table 2.


**Table 2.** Lignocellulose nanofiber and cellulose nanofiber characterization.

ty (ɳ<sup>s</sup> −1 : nanofibrillation yield; T800: optical transmittance; CD: cationic demand; CC: carboxyl content; σ: specific surface area.

– ɳ ɳ – The nanofibrillation yield ( ty (ɳ<sup>s</sup> −1 ɳ ɳ – ) for the LCNF and CNF ranges from 13.34% to 26.44%. These low yields in comparison with those for cellulose nanofibers obtained by enzymatic hydrolysis or the TEMPO-mediated oxidation of CNF from fully bleached wood pulp show that the obtained suspension is composed of cellulose nanofibers with large widths and cellulose microfibers [48,49]. The optical transmittance (T800) of the cellulose nanofiber suspension is an indirect indicator of the nanofibrillation yield. The cellulose microfibers contained in the suspension produce a higher light scattering compared to the nanofibers, so this parameter is highly related to yield and nanometric width. As with nanofibrillation yield, only slight differences in T800 are observed between the various nanofibers, except for CNF-TO. Since the chemical compositions of LCNF and CNF are different, the transmittance of the suspensions should not be considered as a key parameter in the characterization of the suspensions since lignin affects the refractive index. It is observed that CNF-TO presents a higher transmittance due to the fact that it presents a significantly higher nanofibrillation yield than the rest of nanofibers, and in addition, it does not contain lignin in its composition.

oduced, reaching maximum values of 300 μmol/g, higher than those reached for LCNF (152.34 μmols/g), but not as high as those obtained for bleached wood pulps that can reach 1000 μmol/g – – −1 −1 Discover with a monochromatic source CuKα1 – −1 – −1 Discover with a monochromatic source CuKα1 – −1 Cationic demand (CD) refers to the ability of the anionic surface of nanofibers to capture and interact with cationic substances. This value is highly related to the specific surface of the nanofiber; the larger the surface, the greater the capacity for interaction and the carboxyl content on that surface. The values of both parameters for LCNF-Mec, LCNF-TO and CNF-Mec are similar or even higher than what has been reported in the literature for CNF obtained by mechanical pretreatment or TEMPO-mediated oxidation from fibers with high lignin content [34,50–53]. It is observed again that there are not great differences in the cationic demand and carboxyl content, except for CNF-TO. In CNF-TO, TEMPO-mediated oxidation is much more effective that when it is performed on LCNF, as revealed by the increase in carboxyl content. It is observed that CNF-TO increases the carboxyl content by more than double compared to CNF-Mec. This increase is produced by the conversion of hydroxyl groups at the C6 positions on the surface of the cellulose fibers into carboxyl groups, enabling the delamination of the fiber by the electrostatic repulsion of the charged fiber surface [54]. On the other hand, in the case of LCNF, differences between both pretreatments are negligible. The presence of lignin in the fiber can affect the effectiveness of the oxidation reaction because the reaction activator, NaClO, is also consumed as the bleaching agent, producing the oxidation and dissolution of the lignin, thus preventing the selective activation of the catalyst. In fibers previously reported in literature with a lignin content lower than 10%, a partial oxidation of the -OH groups of the cellulose is produced, reaching maximum values of 300 µmol/g, higher than those reached for LCNF-TO described in this work (152.34 µmols/g), but not as high as those obtained for bleached wood pulps that can reach 1000 µmol/g [34,49,52].

−1

The specific surface values, again, show differences in CNF-TO, which shows a considerably higher result than the other cellulose nanofibers. This is a very important parameter when using cellulose nanofibers as a reinforcing agent in materials produced from lignocellulosic materials such as paper, cardboard or fiberboards [12,55]. A larger specific surface area allows for a higher bonding

capacity with adjacent fibers, thus improving the mechanical properties of the final product. Cellulose nanofibers with similar specific surface areas produce an increase about 100% in the mechanical properties of paper and carboard with low amounts of LCNF addition (3%) [53].

Nanofiber width, despite being within the nanometric range (24–71 nm), presents some differences that are discussed. For mechanical pretreatment, the presence of lignin in the fiber (LCNF-Mec) produces greater fibrillation in the fiber, producing a smaller width than CNF-Mec. This could be due to the lignin antioxidant action that prevents the re-bonding of the covalent bonds broken during the mechanical treatments [56]. Regarding TEMPO-mediated oxidation, differences are shown with the presence of lignin, being adverse because of the effect explained above. The length of nanofibers is an important parameter when analyzing the suitability of the application of cellulose nanofibers. The lignin content can affect the effectiveness of the method used for length determination through intrinsic viscosity. However, this method allows an estimation of the effect of the different pretreatments on the length parameter. In a generalized way, decreases in the lengths were observed when the fiber was subjected to TEMPO-mediated oxidation of 68.4% and 78.8% for LCNF-TO and CNF-TO, respectively, with respect to those following mechanical pretreatment. It is caused by the degradation of the cellulose amorphous regions into gluconic acid or cellulose-derived small fragments by depolymerization and β-elimination [57]. The length of the nanofibers is strongly related to the mechanical properties of the final composites made of cellulose nanofibers. It is therefore necessary to achieve a balance between the nanometric size reached during the nanofibrillation process and the shortening of the fiber due to its degradation. The aspect ratio (L/D) is a parameter that shows the relationship between length and width. It is observed that the different cellulose nanofibers showed aspect ratios of 93.44, 20.82, 60.56 and 29.38 for LCNF-Mec, LCNF-TO, CNF-Mec and CNF-TO, respectively. It is shown that although the mechanically pretreated nanofibers present a higher width than CNF-TO, they had a higher aspect ratio due to the low degradation that they underwent in the production process. Therefore, even though CNF-TO has a larger specific surface area and is thus more suitable for application in products made from lignocellulosic material (paper, cardboard, etc.), LCNF-Mec and CNF-Mec would show better behavior when added as a reinforcing agent on polymeric matrices [58].

The chemical composition of the different cellulose nanofibers was analyzed by a FTIR technique (Figure 2). All analyzed samples, as expected, show a spectrum typical of lignocellulosic materials. The peaks at 3300 and 2900 cm−<sup>1</sup> are associated with the stretching vibration of the OH and CH groups present in the cellulose chains. The peaks in the range of 1350–1250 cm−<sup>1</sup> are attributed to the presence of chemical groups of the hemicelluloses. The peaks at 1190, 1070 and 890 cm−<sup>1</sup> are associated with the stretching and rocking vibrations of the C-O, C-H and CH<sup>2</sup> groups of cellulose [52]. However, there are some differences between the various cellulose nanofibers. It is observed that cellulose nanofibers obtained from OTPB bleached pulp (CNF-Mec and CNF-TO) do not show the peak at 1510 cm−<sup>1</sup> that is observed in lignocellulose nanofibers (LCNF). This peak is related to the C=C symmetrical stretching of the aromatic rings, characteristic of the lignin. As expected, due to the nearly total elimination of lignin content in the bleached pulp, this peak is not observed in CNF. Another difference is observed in the peak at 1610 cm−<sup>1</sup> , corresponding to the C = O stretching vibration in carboxyl groups. An important increase in the intensity of the peak is observed in the CNF-TO due to the regioselective conversion of C6 primary hydroxyl groups to carboxyl groups by the TEMPO-mediated oxidation.

The effects of the different pretreatments on the crystallinity of the cellulose nanofibers are shown in Figure 3. It is observed that the same peaks related to the 110 and 200 reflection planes of cellulose I are observed again, implying that the crystalline structure of the original fiber is maintained. The crystallinity index (CI) was calculated in the same way as for OTPB pulp. It shows that cellulose nanofibers present a lower CI (24%–49%) than the original fiber (60.26%). With regards to nanofibers obtained by mechanical pretreatment, they are produced by the disordering of the crystalline regions of the cellulose chain by the shear forces produced in the high-pressure homogenization process and during mechanical pretreatment. For TEMPO-mediated oxidized nanofibers, they are produced by the conversion of ordered cellulose structures into disordered structures by the sodium glucuronosyl

−

units during the oxidation reaction [59]. CNF is observed to have a greater crystallinity than LCNF. This is due to the lignin elimination during the bleaching process and the consequent elimination of the amorphous component of the lignocellulose matrix, increasing the total crystallinity of the fiber. In addition, it is observed that mechanical pretreatment produced a greater disordering in the cellulose chain than the TEMPO-mediated oxidation.

β

**Figure 2.** FTIR spectra of the different cellulose nanofibers.

**Figure 3.** XRD diffraction patterns and crystallinity indices of the cellulose nanofibers.

The thermal stability of the different cellulose nanofibers was studied through the analysis of the TGA and DTG curves (Figure 4). The thermal degradation behavior shows three degradation stages observed in the initial fiber: (i) moisture loss, (ii) glycosidic bond degradation and (iii) cellulose pyrolysis. It is observed that LCNF (Figure 4a) and CNF (Figure 4b) present lower values for maximum thermal degradation, i.e., a lower Tmax, than that obtained for OTPB pulp (348 ◦C). This is due to the larger specific surface of the nanometric-size fibers, which means that they are more exposed to heat, and degradation occurs more quickly than in the original fiber. It can be seen that for cellulose nanofibers obtained by mechanical pretreatment, there are no differences according to the presence or absence of lignin, both showing a Tmax = 343 ◦C. However, analyzing the total mass loss, it is observed that a residual mass at 600 ◦C of 15.14% remains for LCNF-Mec compared to 8.95% for the CNF-Mec. This fact is not indicative of a higher thermal stability, but it indicates that a greater carbonaceous residue is produced after the pyrolysis of the lignocellulosic components due to the aromatic structure of lignin. Regarding the cellulose nanofibers obtained by TEMPO-mediated oxidation, noticeable differences are observed, showing maximum degradations at 325 ◦C and 298 ◦C for LCNF-TO and CNF-TO, respectively. CNF-TO presents worse thermal stability than LCNF-TO and the products obtained by mechanical pretreatment, since in addition to the nanometric size, it has a greater number of free ends (higher cationic demand and carboxyl content), which favors thermal degradation [60]. In addition, contrarily to what has been observed in the nanofibers obtained by mechanical pretreatment, a large increase in the residual mass was produced in CNF-TO (29.76%) in comparison with the values obtained for LCNF-TO (16.54%). This fact is produced by the introduction of carboxyl groups on the surface of the fiber during TEMPO-mediated oxidation, increasing the carboxyl content, especially for CNF-TO (311.95 µmol/g) as observed through its characterization. It is therefore concluded that CNF-Mec and LCNF-Mec, in addition to presenting higher aspect ratios that can result in greater reinforcement effects in polymeric matrices, can be used in polymers with higher transition temperatures compared to CNF-TO and LCNF-TO due to their greater thermal stability.

**(a)** 

**Figure 4.** *Cont.*

Tmax

= 298 ºC

0

20

40

60

Weight loss (%)

80

100

B

100 200 300 400 500 600

Tmax

= 343 ºC

0

1

2

DTG (%/ºC)

3

4

5

Temperature (ºC)

0

20

40

60

Weight loss (%)

80

100

A

100 200 300 400 500 600

Tmax

= 343 ºC

0

1

2

DTG (%/ºC)

3

4

5

Temperature (ºC)

Tmax

= 325 ºC

**Figure 4.** TGA and DTG curves for the different cellulose nanofibers: (**a**) lignocellulose nanofibers (LCNF) and (**b**) cellulose nanofibers (CNF). Black curves for those obtained by TEMPO-mediated oxidation and grey from mechanical pretreatment.
