Graft modification in bamboo cellulose nanofibrils: An effective option to improve compatibility and tunable surface energy


 In this work, from an endemic and non-significant value-added bamboo argentine, nanofibrils (CNFs) of 20 nm in width were obtained. These nanofibrils were chemical modified in surface with three simple steps using a noncommercial low molecular weight polylactic acid. The success of modification was confirmed by FTIR, TGA, DSC and XRD analysis. The modified nanofibrils were taken up for changing surface properties in films based on commercial PLA. The results show that dispersive (γ D/S) component of films increase of 34.7 mN/m to 36.1 mN/m after the addition of modified nanofibrils from 2 to 5% in formulation of the films, comparing with a physical blend. Interesting others result in physical, mechanical, and thermal properties of the nanocomposites, were reported.


Introduction
Renewable bio-based polymers and composites derived from natural resources are generating a great interest due to the depletion of fossil fuel resources and the negative environmental impact of fossil fuel-based plastic products (Kunaver et al. 2016 Among the most promising biodegradable polymers stands out polylactic acid (PLA), because it has been known for several decades and has attracted much attention in both the academic and industrial sectors. The advantages of this polymer are due to the fact that it is an environmentally friendly plastic, commercially available, making it the most representative biodegradable and bio-based polymer (Scaffaro et al. 2017;Mao et al. 2019). PLA can be obtained through different methods of lactic acid monomers, i.e. direct polycondensation (DP), ring opening polymerization (ROP), solid state polymerization (SSP) and direct methods like azeotropic dehydration and enzymatic polymerization (Jamshidian et al. 2010). Besides its biobased character, it exhibits an excellent melt processability, a room temperature elastic modulus comparable to that of poly(ethylene terephthalate) and a tunable crystallinity (Carlmark et al. 2012). Cellulose is another biopolymer, which is among the most widely used and most abundant natural biopolymers that can be obtained from several naturals resources such as wood, some bacteria, fungi, and some algae (Corrêa et al. 2010). It is a linear macromolecular polysaccharide consisting of a long chain of glucose units linked by β-1, 4-glycosidic bonds (Wang et al. 2017). Its excellent properties such as biocompatibility, renewability, and non-toxicity combined with a low price and good mechanical properties make cellulose an attractive resource for a broad range of applications. Cellulose can be converted into nanocellulose (NC) through different approaches such as chemical, mechanical and biological treatments. NC can be extracted as cellulose nanocrystals (or whiskers) obtained by acid hydrolysis of a celluloserich substrate and cellulose nanofibrils (or CNFs) obtained by different combinations of enzymatic, chemical and/or mechanical treatments of the starting raw material (Missoum et al. 2012). CNFs have been used as reinforcing material in a wide range of polymeric matrices, including PLA, thus preserving its bio-based nature. However, the hydrophilic nature of cellulosic plant fibers makes them difficult to disperse within the PLA matrix (Georgopoulos et al. 2005;Tingaut et al. 2010;Aitomäki et al. 2016;Ray and Sain 2016).
The hypothesis of this work is that the chemical modification of CNFs with low molecular weight PLA (PLA l) should improve the dispersion of modified CNFs in a hydrophobic polymer matrix such as PLA, promoting miscibility and interfacial adhesion between them. Therefore, the aim of this work was to prepare flexible biodegradable films from renewable natural resources and to develop composites with improved mechanical performance and thermal stability.

Materials
L-Lactic acid (85-90%) was purchased from Biopack (Argentina). All the reagents: sodium hydroxide, hydrogen peroxide, p-toluenesulfonic acid monohydrate (TSA), benzoyl chloride, pyridine, thionyl chloride and solvents were obtained from commercial suppliers (Aldrich Co). Poly (lactic acid) pellets were provided by Shenzhen Bright China Industrial Co. Ltd. (Wuhan, China) (L-LA 90%, 10% D-LA). The number average molar weight (Mn = 77,000 g mol -1 ) was determined by size exclusion chromatography (SEC). The weight average molar weight (Mw) and the polydispersity index were also determined and values of 200,000 g mol -1 and 2.6, respectively, were obtained. Prior to use, PLA was purified by dissolving it in dichloromethane and precipitating in methanol. All other materials were used without purification.

Preparation of cellulose nanofibrils (CNFs)
CNFs were produced following a procedure previously developed by Rodríguez substances from the bamboo, 100 g of the sample was dewaxed in a Soxhlet apparatus with a 2:1 (v/v) mixture of ethanol:toluene for 6 h. Subsequently, lignocellulosic substances in the samples were removed using a basic solution (NaOH 2%) at 75 °C for 1 h, and the process was repeated four times until the color of the sample became white. Thereafter, the residual material was filtered and rinsed with deionized water four times until neutral pH.
Next, residual hemicelluloses and pectin were removed by treating the sample with hydrogen peroxide solution (10%) at 60 °C for 6 h. After these chemical treatments, a bamboo cellulose dispersion (2 wt%) was prepared by adjusting the concentration of the dispersion with distilled water and mixed with an IKA high-shear mixer (T-T18 ULTRA TURRAX Basic) at 15,000 rpm speed. pretreated bleached fibers were disintegrated in a microfludizer process in a MASUKO® (MKCA6-2, Masuko Sanguo, Japan) and were passed through the device up to 60 times. The gap between the disks was adjusted to 9 µm.
The dispersion CNF were centrifuged at 10,000 rpm to reduce its water content and kept wet in the fridge.

Synthesis of poly(lactic acid) with low molecular weight (PLA1)
The synthesis of PLA1 was performed according to a previously described method (Fukushima et al. 2005;Minh et al. 2013). The synthesis was carried out by condensation of the L-lactic acid (LLA), using p-toluenesulfonic acid (TSA) as catalyst. In a three-neck round bottom flask, 100 mL LLA and 0.4 g TSA were mixed. Then, the mixture was heated to 175 °C at reduced pressure to start the polycondensation. After a period of 24 hours of polycondensation, the reaction was stopped; the products were dissolved in chloroform and precipitated in methanol. Finally, the precipitate was filtered and dried in a vacuum oven at 40 °C for 48 hours. The number average molar weight (Mn = 10,000 g mol-1) was determined by size exclusion chromatography (SEC). The weight average molar weight (Mw) and the polydispersity index were also determined and values of 14,000 g mol-1 and 1.4, respectively, were obtained. PLA1 was characterized by physical, spectroscopic, and thermal methods.

Chemical functionalization of CNFs.
A previously reported procedure (García et al. 2012) was used that consists in three reaction steps (Scheme 1). The first step is the protection of the hydroxyl groups of PLA1 through a benzoylation reaction (PLA1Bz), the second step involves the activation of carboxyl groups with thionyl chloride (PLA1BzCl), and these latter groups react immediately with the cellulose nanofibrils (PLA1Bz-g-CNF) during the third step.

Synthesis of PLA1Bz
PLA1 (1.0 g) and chloroform (20 mL) were mixed in a three-neck flask. Then, 1 mL benzoyl chloride and 1 mL pyridine were added. After 24 h of agitation at 45 °C, the mixture was subjected to few extractions with solvents and precipitated with methanol. The mixture was filtered, and the solid (PLA1Bz) was dried in a vacuum at 40 °C for 24h.

Synthesis of PLA1BzCl
The activation of carboxyl groups was achieved by dissolving 1.0 g of PLA1Bz in 60 mL of a 1:1 mixture dichloromethane:toluene at 70 °C in a three-neck flask. Then thionyl chloride (0.2 mL) and 0.5 mL of triethylamine (TEA) were added and submitted to constant stirring during 3 h. The product obtained was named PLA1BzCl and it was used without further purification.

Surface chemical modification of cellulose nanofibrils
The chemical modification of CNFs was performed in a round-bottomed reaction flask under reflux (48 h) under constant mechanical stirring in toluene medium. To suspend CNFs in toluene, a solvent exchange procedure consisting in several successive centrifugation and redispersion operations was used (water to acetone, and then acetone to toluene, both four times). Then, to perform esterification of CNFs (0.19 g) previously sonicated in toluene (10 mL) during 10 min was added to PLA1BzCl. PLA1Bz-g-CNF was purified and stored for further analysis and the preparation of nanocomposites.

Preparation of PLA/PLA1Bz-g-CNF and PLA/PLA1Bz/CNF
After purification of products PLA1, PLA1Bz and PLA1Bz-g-CNF, nanocomposites were prepared by solution casting using chloroform as solvent. Initially, high molecular PLA (Table N° 1) was dissolved in chloroform (12 mL) using a magnetic stirrer, and then PLA1Bz-g-CNF was added. The filler content (PLA1Bz-g-CNF) in nanocomposites was varied from 0 to 5% based on the total weight of PLA. Subsequently, the nanocomposites were dried completely at 50 °C under vacuum for 48 hours. For comparison purpose a blend of PLA, PLA1Bz and CNFs (PLA/PLA1Bz/CNF) was prepared using similar procedure. Table 1 shows the composition of the different materials.

Scanning electron microscopy (SEM)
Nanofibrils were sonicated for 30 seconds and deposited on microscope glasses and vacuum dried at 40°C for 24 h. Samples were coated with gold using an ion sputter coater and observed by using a scanning electron microscope Zeiss Supra 40 with field emission gun operated at 3 kV.

Molecular weight determination
PLA molecular weight was determined by Size Exclusion Chromatography (SEC) on a Waters 600 at 31 °C, through a Styragel HR 4 THF-7.8 x 300 mm -(5K-600K) column and a refractive index detector 2414 at a flow rate of 1 mL min −1 , using polystyrene standards.

Fourier-transform infrared spectroscopy (ATR-FTIR)
Attenuated total reflection Fourier Transform Infrared (ATR-FTIR) spectroscopy (Nicolet IS50) was used to follow the impact of the different reactions and to compare grafted to unmodified CNFs and nanocomposites, recording changes in functional groups induced by the modification of nanocellulose. Spectra were collected in the range 4000 cm -1 -500 cm -1 , at a resolution of 4 cm -1 .

X-ray diffraction analysis
A Siemens D 5000 X-ray diffractometer was used to observe the possible change in X-ray diffraction patterns during the reactions and to compare grafted to unmodified CNFs.
X-ray generator voltage and current were 40 kV and 30 mA, respectively. The radiation was Cu Kα with wavelength 1.54 Å. The diffraction patterns were obtained at room temperature in the range of 2θ between 14 ° and 21 ° by step of 0.02 °.
Approximately 5 mg of each sample was subjected to heating from 30 °C to 500 °C at a rate of 10 °C/min and nitrogen flow 30 mL/min. The weight loss of the materials was calculated on dry basis and the different degradation phases were noted.

Differential scanning calorimetry (DSC)
Calorimetric measurements were made on a TA Q20 differential scanning calorimeter in a dry nitrogen atmosphere. Indium standard was used for calibration.
Samples of 5−10 mg were placed in the DSC pan. Samples were first heated to 200 °C and held at that temperature for 10 min to remove the thermal history. Then, the samples were cooled to − 25 °C at a rate of 10 °C/min, held for 10 min, and heated again to 200 °C at 10 °C/min. The Tg value was taken as the midpoint of the transition. The degree of crystallinity ( ) for PLA, PLA/PLA1Bz-g-CNF and PLA/PLA1Bz/CNF nanocomposite films was calculated using Eq. (1).
Where ΔHm and ΔHc represent the melting enthalpy and cold crystallization enthalpy, respectively, ∆ is the melting enthalpy for 100% crystalline PLA which was set to 93.6 J/g (Quero et al. 2012). WPLA depicts the weight fraction of PLA in the nanocomposite sample.

Contact angle measurements
Contact angle tests were carried out with a dynamic contact angle tensiometer at room temperature, using the sessile drop measurement technique. For each test 10 µL of various liquids: diiodomethane, formamide and water, were used. The sliding angle (α) which is defined as the minimum angle from horizontal required for a water droplet to roll was determined. Measurements of α were performed trigonometrically by determining the minimum altitude for a water droplet to start rolling. For a given sample, the contact angle was measured in ten different locations on the surface (for each location, the value of the contact angle at the left and right sides of the drop was recorded). The contact angle at equilibrium for pure liquids having different polarities was used to calculate the dispersive and polar contributions to the surface energy according to the approach proposed by Owens and Wendt (Owens et al. 1969;Adamson and Gast 1990) by using Eq.
(2). The total, dispersive and polar components of these liquids are listed in Table 2.
where , , and are the total, dispersive, and polar surface energy, respectively.
Subscripts L and S refer to the liquid droplet and the solid surface, respectively, and θ denotes the contact angle between the solid surface substrate and the liquid droplet.

Morphology and structure of Cellulose Nanofibrils
In Figure 1A and 1A is shown the SEM micrographs nanofibrils morphology.
The diameter of nanofibrillated cellulose from bamboo was determined by digital image analysis from FE-SEM micrographs with its own internal analytics software of Zeiss Supra 40, as presented in the Fig. 1. The average diameter of CNFs was about 20 ± 2nm (a minimum of 50 measurements was performed).

Thermogravimetric analysis (TGA).
The TGA curves for PLA1, PLA1Bz, PLA1Bz-g-CNF and PLA1Bz/CNF are shown in  PLA1Bz-g-CNF exhibits a 50 °C increase in the degradation temperature compared to neat CNFs. This phenomenon is attributable to the fact that the grafted product generates an effective wrapping due to the hydrogen bond interactions between PLA1 chains and CNFs, which acts as a protective layer inhibiting the degradation of the material (Wu et al. 2016) and demonstrates the success of the grafting.
In the thermogram to PLA1Bz (blue line) is observed two important losses. The first one, between 304 and 310 °C, was assigned to the degradation of the carboxyl group. It is reasonable not to observe this step on the thermogram for PLA1, because in this polymer is stabilized by the formation of hydrogen bonds between the carboxyl group with the end hydroxyl group of another PLA1 chain (Kopinke et al. 1996). The second degradation process could be due to several decomposition reactions. The thermal degradation of PLA1 is mainly associated with the hydrolysis of the ester groups and it is accelerated by the end groups (-COOH) (Piemonte and Gironi 2013).
14 The TGA curve for CNF (green line) shows two degradation steps, one below 100 °C and a maximum loss rate at 285 °C. The first step corresponds to depolymerization, dehydration and decomposition of glycosidic units starting at 100 ° C and reaching its maximum value ( The original PLA1 degrades in a single stage within a narrow temperature range (307-377 °C) with a maximum degradation rate at 358 °C, in agreement with literature reports (Kopinke et al. 1996). Figure 4 shows the DSC thermograms for PLA1, protected PLA (PLA1Bz), grafted nanofibrils (PLA1Bz-g-CNF) and PLA1Bz/CNF blend. The main characteristics of the observed thermal events are reported in Table 3.  for PLA-grafted on CNFs (PLA1Bz-g-CNF) and the physical mixture (PLA1Bz/CNF) when comparing the thermal characteristics between the first and second heating scan, but a lower melting point was also noted during the second heating scan. The presence of CNFs induced an increase in both the Tg value and degree of crystallinity of PLA as already reported (Suryanegara et al. 2009;Lv et al. 2017;Shojaeiarani et al. 2019). Moreover, PLA-grafted on CNFs shows a higher Tg (62.9 °C during the first heating cycle), than that of the physical blend (59.6 °C) and that of PLA1Bz (57.9 °C). The same observation was reported for the melting temperature, which was 143.5 °C for the grafted product, 138.9 °C for PLA1Bz, and 141.4 °C for the blend. No significant variation in the crystallization temperature for PLA1Bz-g-CNF and PLA1Bz/CNF was observed, but when comparing the enthalpy of crystallization of the materials, an increase was reported when adding of CNFs.

Thermal properties and crystallinity of nanocomposites
TGA experiments showed the effect of adding PLA1Bz-g-CNF on the thermal properties of PLA films. For comparison, a PLA/PLA1Bz/CNF blend (composition 0.2% of nanofibrils) was tested and all results are illustrated in Figure 6. All samples were thermally stable up to 285°C, while above this temperature, the bionanocomposites started to undergo decomposition. A slight increase in the onset decomposition temperature was observed when adding PLA1Bz-g-CNF to PLA. When the PLA1Bz-g-CNF content was 5% the onset decomposition temperature increased by almost 7 °C compared with neat PLA. It is well-known that enhanced thermal stability of the host polymer results from favorable filler/matrix interactions. This result shows the successful modification of the nanofibrils and improved filler/matrix compatibility. On the other hand, no significant increase in decomposition temperature was reported for the composite based on PLA1Bz/CNF blend compared to PLA1Bz-g-CNF 2% (360 °C vs. 363 °C), which is attributed to the poor compatibility and low level of dispersion of the blend in the polymer matrix (Aouat et al. 2018).

DSC experiments for PLA and PLA nanocomposites
The thermal characteristics and crystallization behavior of neat PLA film and nanocomposites were investigated by DSC. The corresponding thermograms are presented in Figure 7. In addition, the values of glass transition temperature (Tg), cold crystallization temperature (Tc), melting temperature (Tm), enthalpy of crystallization (∆Hc), enthalpy of melting (∆Hm) and degree of crystallinity (Xc) of the films are given in Table 4.  PLA1Bz-g-CNF, whereas it increased 5.3% when using PLA1Bz/CNF. It is worth noting that this crystallinity corresponds to the crystalline phase formed during the cooling cycle at 10°C/min and initially present before the second heating scan according to Eq. (1). CNFs are well-known for acting as a heterogeneous nucleating agent (Zamir et al. 2018;Shojaeiarani et al. 2019), which is effectively observed for the physical blend, but it seems that PLABz-grafted CNFs did not play this role.

Mechanical properties
The mechanical properties, such as Young's modulus, tensile strength, elongation at break and work of fracture were evaluated from the stress-strain curves. Typical stressstrain curves for pure PLA and nanocomposites with different PLA1Bz-g-CNF contents and PLA1Bz /CNF blend are shown in Figure 8 and the related data are summarized in Table 5.

Fig. 8
Typical stress-strain curves for PLA and PLA nanocomposites. The results showed that the addition of PLA1Bz-g-CNF changes the mechanical properties of the PLA film. As seen in Table 5, the elongation at break as well as the work of fracture increased progressively, while both the Young´s modulus and the tensile strength decreased. The addition of PLA1Bz-g-CNF 5% raised the elongation at break by 117% and the work of fracture by 63%, while in the blend the elongation at break increased by only 26%, which is still better than for PLA1Bz-g-CNF 2% (increase by 13%). The unexpected decrease in the stiffness observed when adding CNFs to the PLA matrix could be attributed to the decrease in the degree of crystallinity of the PLA matrix (Table 4). In addition, the grafting on CNFs or physical incorporation of low molecular weight PLA could act as a plasticizer for the PLA matrix. Table 6 shows the values of contact angle measured for water, diiodomethane and formamide deposited on PLA, PLA/PLA1Bz-g-CNF, and PLA/PLA1Bz/CNF films at longer times, i.e. when the equilibrium was reached. As seen in Table 6, a clear and progressive increase in the contact angle with water and formamide was observed when adding PLA1Bz-g-CNF in PLA. This result shows that the films become more hydrophobic after the addition of modified nanofibrils. The initial contact angle value for pure PLA was approximately 6° less than for the nanocomposite film reinforced with 5% PLA1Bz-g-CNF, while the mixture showed a contact angle of 71.9°, a value even lower than for neat PLA. This increase in hydrophobicity for the films loaded with PLA1Bz-g-CNF is attributed to the fact that the nanofibrils are covered almost entirely by small chains of PLA1 grafted onto their surface. Since all the nanofibrils are wrapped, their adhesion to the matrix should be improved, and it would produce a low exposure of available OH, generating a hydrophobic PLA surface. This would not happen for the mixture, since the nanofibrils exhibit free OH available to interact with polar liquids (Bretas et al. 2015;Borkotoky et al. 2018).

Contact angle measurements
For diiodomethane, an opposite behavior to that observed for the contact angle with water and formamide was reported for PLA/PLA1Bz-g-CNF, i.e. the contact angle progressively decreased when adding PLA1Bz-g-CNF. It confirms the effectiveness of the grafting method to decrease the hydrophilicity of the material compared to the physical blending (Labet et al. 2007;Habibi et al. 2008;Paquet et al. 2010).

Surface energy measurements
To further quantify the impact of the addition of nanofibrils to PLA, we determined the total surface energy ( ), together with the polar ( ) and dispersive ( ) components of the developed films (Table 7) using the constants of the solvents used as liquid probes (  When adding cellulose nanofibrils to the PLA matrix, the total surface energy did not change significantly. However, the addition of PLA1Bz-g-CNF progressively decreases the polar contribution to the surface energy. It is attributed to changes in the different intermolecular interaction like dipole-dipole, dipole-induced dipole and hydrogen bonding interactions indicating a decrease in wettability (Dhar et al. 2016). For PLA/PLA1Bz /CNF, an increase in the polar contribution to the surface energy is observed, which is an expected finding since the surface of CNFs contains OH-rich macromolecules, the most capable to establish hydrogen bonds. These results clearly show the difference in polarity between PLA-grafted on CNFs and unmodified CNFs. It should directly affect the adhesion and dispersion of the filler in the PLA matrix, resulting in improvements in thermal and mechanical properties previously reported (Bendahou and Hajlane 2014).

Optical properties
The processing technology used in the present work allowed obtaining thin films prepared according to the formulations shown in Table 1, with thickness ranging between 26 and 27 µm.
As the filler content in the material increased, the amber placement commonly found for PLA1Bz -g-CNF and PLA1Bz/CNF nanocomposites increased (Figure 9). On the other hand, this analysis does not allow observing significant differences between PLA1Bz -g-CNF and PLA1Bz /CNF nanocomposites.

Conclusions
Chemical grafting on CNFs with short PLA chains has been efficiently achieved and used in an attempt to obtain better compatibility by modifying the polar contribution to surface energy between the hydrophobic PLA matrix and the hydrophilic CNFs. The effect of the modified nanoparticles was tested by incorporating them into PLA matrix films. The results demonstrate the effect of modifying the CNFs by incorporating it into the PLA matrix, reducing the polar contributions of the nanofibrils in the films. As a result of this successful modification, an improvement was obtained in the thermal, mechanical and surface properties of the material. In particular, improvements in fracture work have been reported. This type of modification shows that it is possible, depending on the type of polymer grafted, to configure the properties of the films. The disposition of the OH available in the nanofibrils can be adjusted according to the needs of the expected final properties. Finally, the physical-chemical, thermal, mechanical and surface properties evaluated in this study demonstrate the strong potential of biodegradable films as packaging materials in different industries, always taking into account the characteristics of the product to be packaged and the packaging requirements.