Polylactic acid nanocomposite films with spherical nanocelluloses as efficient nucleation agents: effects on crystallization, mechanical and thermal properties

Abstract

The low crystallization rate of commercial polylactic acid (PLA) restricts its application in food packaging materials. In this work, spherical nanocellulose formates (SCNFs) were incorporated into a PLA matrix to fabricate green nanocomposites. The addition of well-dispersed SCNFs significantly increased the crystallization rate, and thus led to obvious enhancements of the mechanical performance and thermal stability of the PLA. Compared to pure PLA, 130% and 116% improvements for the tensile strength and Young’s modulus can be obtained for the resulting nanocomposite with 10 wt% SCNFs, respectively, and the initial degradation temperature (T₀) and maximum degradation temperature (T_max) increased by 17.4 and 21.5 °C, respectively. This was ascribed to the good SCNF dispersability within the PLA matrix, the improved interfacial interaction and good crystallization ability of the PLA. Furthermore, the addition of SCNFs improved the barrier and overall migration properties of the nanocomposite as potential food packaging.

---

Introduction

Recently, there are more and more biopolymers for replacing non-renewable petroleum-based polymers as food packaging materials. Polylactic acid (PLA) is one of the most biodegradable biopolymers for food packaging, due to its origination from renewable biomass (such as corn, wheat, and sugar cane), excellent transparency, outstanding physical properties and easy processability with respect to other green polymers.^1–4 However, the slow crystallization rate of PLA as a major industrial problem leads to final PLA products with very low crystallinity during practical melt-processing conditions. Especially, it restricts their application as food packaging. Low crystallinity is also associated with poor mechanical, thermal stability, barrier and migration properties for PLA. For this reason, an efficient method to increase the crystallization rate or crystallinity, i.e. incorporation of heterogeneous nucleating agents with high nucleation density has been developed.^5,6 To broaden the utilization of PLA, fully biodegradable and biocompatible nucleating agents with excellent nucleation ability are desired.

Nanocelluloses (NCs) as important bio-based nanofillers can obviously enhance the mechanical, thermal and barrier properties of PLA, due to their good mechanical properties, high stiffness (large aspect ratio), biodegradability, and renewable character.^7–11 As expected, the mechanical strength of the resulting PLA nanocomposite was enhanced by adding rigid NCs with a rod-like shape, but compared to pure PLA, the crystallization rate and crystallinity were almost not prompted by introducing the unmodified rod-like NCs. Also, the crystallization rate and crystallinity were increased slightly by using functionalized NCs with better dispersion within a PLA matrix.^{1,4,6–8,12–15} Pei et al. used unmodified NCs and silylated NCs as bio-based nucleation agents to study their effects on the crystallization and mechanical properties of PLA.¹⁵ With addition of unmodified NCs, no obvious increase in the tensile strength and crystallinity was observed due to incompatibility and poor interfacial interaction between hydrophilic NCs with abundant hydroxyl groups and the hydrophobic PLA matrix, while the tensile strength and crystallinity of the PLA/silylated NCs were improved by 21% and 16.1%, respectively, as compared to PLA. Subsequently, they proposed that reducing the length of cellulose nanocrystals (i.e. using nanospheres instead of rods) was expected to increase the nucleation efficiency, which would greatly enhance the crystallization and other properties of PLA. Fortunati et al. also reported that surfactant-modified NCs can favor their dispersion in the PLA matrix, and thus remarkably enhance the nucleation effect for PLA crystallization.¹ This can be supported by evidence that the cold crystallization temperature of PLA was reduced from 118.6 °C to 111.5 °C for the nanocomposite with 1 wt% unmodified NCs, and 91.5 °C for the nanocomposite with 1 wt% modified NCs. Besides, the increased crystallinity can depress the migration values in food simulants of the nanocomposites as food packaging, but the maximum degradation temperature (T_max) of PLA was obviously decreased from 332 °C to 323 °C and 329 °C, because the sulfuric acid hydrolyzed the NCs and modified the surfactant on the NCs, causing degradation of PLA. From the above, the nucleation ability rate of NCs in PLA crystallization was dependent on the morphology and surface groups of NCs, the NC dispersion and loading level.^1,6,7,15 A current challenge to achieve a good crystallization ability and properties of PLA/NC nanocomposites is dependent on the well-dispersed NCs (with a spherical shape and/or hydrophobic groups) into the PLA matrix without the help of surfactants or compatibilizers.

From the above, spherical nanocellulose formates (SCNFs) with hydrophobic formates were successfully fabricated by our group,¹⁶ and the SCNFs as nucleation agents were incorporated into a PLA matrix for the production of bionanocomposites in this work. The hydrophobic formates of SCNFs can improve their dispersion in PLA, and spherical nanocrystals may show a high nucleation effect on the PLA crystallization to solve the industrial issue of PLA (slow crystallization rate). To the best of our knowledge, this is the first report about the effect of spherical nanocelluloses with functionalized groups on the mechanical, thermal, isothermal crystallization kinetics, barrier and migration properties of PLA.

Experimental

Materials

Industrial lyocell fibers without spin finishing were kindly supplied by Shanghai Lyocell Fibre Development Co., Ltd. Polylactic acid (PLA) was supplied by Bright China Industrial Co. Ltd as received without further purification. Formic acid (HCOOH), hydrochloric acid (HCl), ammonia solution (NH₃·H₂O), chloroform, isooctane and ethanol (C₂H₅OH) were purchased from Hangzhou Mike Chemical Agents Co. Ltd., China. All the materials and reagents were used as received without further purification.

Extraction of spherical nanocellulose formates

The spherical nanocellulose formates (SCNFs) were prepared through HCOOH/HCl hydrolysis of commercial lyocell fibers, and Fischer esterification of accessible hydroxyl groups in a one-pot process.¹⁶ Industrial cellulose fibers were submitted to acid hydrolysis with 90% (v/v) HCOOH/HCl at 80 °C for 8 h. After cooling to room temperature, the resultant suspension was neutralized with NH₃·H₂O solution, and washed by centrifugation until the pH of the suspension was about 7. After 10 min of exposure to ultrasonic irradiation, the suspension was freeze-dried for 48 h for obtaining the dry SCNFs.

Preparation of the PLA/SCNF nanocomposites

PLA/SCNF nanocomposite films (1–20 wt% SCNFs) were prepared by the solution-casting technique. The detailed process is described as follows: firstly, PLA pellets were dissolved in chloroform at 55–60 °C and SCNFs were dispersed in chloroform. Then the well-dispersed SCNF suspension in chloroform was slowly dropped into the PLA solution in chloroform with strong stirring. The total concentration of solid (PLA and SCNFs) in chloroform was 10 wt%. Subsequently, the mixture was casted on a glass board. After the solvent was evaporated completely at room temperature, the nanocomposite films with a thickness of approximately 70–80 μm were obtained, and further dried under vacuum at 40 °C overnight.

Field emission scanning electron microscopy (FE-SEM)

The morphologies of the lyocell fibers, SCNFs and fractured morphologies of the nanocomposite films were observed on a field emission scanning electron microscope (FE-SEM, ULTRA-55, Carl Zeiss) at 1.0 kV. The films were frozen in liquid nitrogen and fractured.

Fourier transform infrared (FT-IR) spectroscopy

FT-IR spectra were recorded on a Nicolet 5700 FT-IR spectrophotometer, operating at 64 scans and 2 cm⁻¹ resolution in the region between 4000 and 400 cm⁻¹ at room temperature. Pellets of dried SCNFs, PLA and the nanocomposites were made with KBr.

Contact angle (CA) measurement

The contact angle of water in air on the sample surface (powdered SCNF discs with a diameter of 1 cm) was measured using a CA analyzer (SL200B, USA Kino Industry Co., Ltd.) at room temperature. About 2 μL of deionized water was dropped onto the surface at a contact time of 5 s. Five independent determinations at different sites of the sample were averaged.

X-ray diffraction (XRD) measurements

X-ray diffraction (XRD) measurements were carried out on a Thermo ARL XTRA X-ray diffractometer using Cu Kα (1.5418 Å) radiation (40 kV, 40 mA) with steps of 3 min⁻¹ at room temperature. All the samples were kept for 2 weeks at room temperature to reach equilibrium crystallization before measuring.

Tensile tests

Tensile tests were measured on mechanical testing equipment (Instron 2345) at 20 °C with a relative humidity (RH) of 65% and a crosshead speed of 5 mm min⁻¹. Before measuring, all the samples were conditioned in a controlled environment for at least 24 h. Rectangle-shaped samples were used (1 × 5 mm⁻¹, 70–80 μm in thickness), and 10 replicates were tested to check repeatability.

Thermal stability

The thermal stability of the films was measured using Perkin Elmer PYRIS 1 thermogravimetric analysis (TGA). The samples (5–10 mg) were heated from room temperature to 600 °C at a rate of 10 °C min⁻¹ using a dynamic nitrogen atmosphere (30 mL min⁻¹).

Optical properties

The optical properties of pure PLA and the nanocomposite films were characterized on a UV-Vis spectrophotometer (Hitachi U-2900, Japan). Each spectrum was collected in the wavelength interval ranging from 200 to 800 nm.

Isothermal crystallization and melting behavior

For isothermal crystallization, all samples were heated from room temperature to 200 °C and kept for 5 min to completely eliminate any possible crystalline phase using TA instruments Q20 differential scanning calorimetry (DSC). Then the samples were quenched to a desired isothermal crystallization temperature (125 °C, 120 °C, 115 °C, and 110 °C) and held for sufficient time to allow crystallization completely from the quiescent melt. The exothermic curves of heat flow as a function of time were recorded.

Results and discussion

Physicochemical properties of SCNFs

Fig. 1 shows FE-SEM images of lyocell fiber and SCNFs. The lyocell fiber has a smooth surface with a diameter of 10–20 μm (Fig. 1(a)). SCNFs showed a spherical structure with an approximate size of 27 ± 2 nm (100 nanospheres were selected and measured) (Fig. 1(b)), suggesting that the acid hydrolysis could efficiently digest the amorphous cellulose domains. In Fig. 1(c), the lyocell fiber and SCNFs present crystalline II bands around 1056, 1023 and 894 cm⁻¹ attributed to the C–O–C stretching band, and motions of C-5 and C-6 atoms, respectively.^16–18 However, a new ester carbonyl stretching peak at 1720 cm⁻¹ appeared in the spectrum of SCNFs, indicating the formation of formate groups on the SCNFs through a Fischer esterification reaction between cellulose hydroxyl groups and carboxyl groups of HCOOH.¹⁶ Moreover, both the lyocell fiber and SCNFs show three main peaks at 12.2°, 20.2°, and 21.8° corresponding to the (11̄0), (110), and (020) lattice planes (Fig. 1(d)), demonstrating their crystalline cellulose II structures.¹⁶ In addition, the SCNFs with a contact angle of 33.8° ± 1.4° were larger than the 25.2° ± 0.9° for lyocell fibers,¹⁶ due to the existence of formate groups. The improved hydrophobicity of SCNFs would be beneficial for compatibility with the hydrophobic polymer matrix.

Fig. 1 FE-SEM image of lyocell fibers (a), and SCNFs (b), and FT-IR spectra (c) and X-ray diffraction (XRD) patterns (d) of lyocell fibers and SCNFs.

Mechanical properties and model approach

In general, the rod-like nanocelluloses with a large aspect ratio can contribute to significant enhancement in the mechanical strength of the polymer matrix,^19,20 while the spherical nanocelluloses with a low aspect ratio can easily disperse in the polymer matrix. So it is interesting to investigate the effect of spherical nanocelluloses with a low aspect ratio on the PLA matrix because food packaging materials need enough mechanical strength to be self-supporting and resist handling damage.^3,9,12,19 As shown in Fig. 2, surprisingly remarkable increases in the tensile strength and Young’s modulus of PLA are achieved by the addition of SCNFs (Fig. S1†). With 10 wt% loading, the nanocomposite exhibited the highest tensile strength and Young’s modulus with an increases of 130% and 116%, respectively, in comparison to the PLA matrix. The main reason was ascribed to increased crystallinity (see below), and strong hydrogen bonding interaction (or networks) at the interface resulting from good dispersion of the SCNFs within the matrix (Fig. S2†). When more SCNFs (15 and 20 wt%) were incorporated, reductions in the tensile strength and Young’s modulus were observed due to agglomeration of the SCNFs. Moreover, a gradual decrease in the elongation at break was detected for the nanocomposites (Fig. 2(a)), which was ascribed to the fact that stiff reinforcements of SCNFs will cause substantial local stress concentrations and failure at a reduced strain with respect to the PLA matrix. More importantly, compared to neat PLA, the increased magnitude in the tensile strength and Young’s modulus of the nanocomposite with 1 wt% SCNFs (spherical shape) was 45% and 27%, whereas the tensile strength of the nanocomposite with 1 wt% NCs (rod-like shape) was reduced by 9%, and the Young’s modulus was slightly enhanced by 2%.¹⁵ Besides, a more severe aggregation of rod-like NCs in the PLA matrix would weaken the mechanical strength of the nanocomposites.^4,15,21 It suggests that the spherical NCs show a stronger reinforcing efficiency in PLA than that of the rod-like NCs.

Fig. 2 Tensile strength and elongation at break (a), and Young’s modulus (b) for pure PLA and the nanocomposites.

The enhancement efficiency of the SCNFs in PLA was estimated using the models. The theoretical Young’s modulus of the nanocomposites was calculated by employing the well-known Voigt–Reuss and Halpin–Tsai models.²² The modulus estimated from the Voigt–Reuss model (E_VR) is depicted as below:

where E_f and E_m are the modulus of the nanofiller and matrix, respectively. W_f is the volume fraction of the nanofiller, where the volume contents of the nanofillers were calculated supposing PLA and SCNFs densities of 1.26 and 1.63 g cm⁻³, respectively.

The Halpin–Tsai model, which considers the aspect ratio of the reinforcing nanofillers, is depicted as below:

where E_HT is Young’s modulus of the nanocomposite, and η_L and η_T are described by:where l_f and d_f are the average length and width of the nanofiller, respectively. E_m was calculated as a value of 0.96 GPa through the tensile test data of this work. We used the E_f value of 150 GPa⁹ (we assumed that E_f of nanocelluloses with a cellulose I structure was equal to that for a cellulose II structure) and l_f/d_f as 1 for SCNFs.

Fig. 3 depicts the Young’s modulus of the PLA/SCNF nanocomposites versus the SCNF contents. The theoretical values calculated from Voigt–Reuss model were clearly higher than the modulus of the PLA/SCNF nanocomposite. This obviously implies that the SCNFs can not effectively transfer the applied stress, due to the extremely low aspect ratio of SCNFs. Therefore, the reinforcements in mechanical properties of the nanocomposites mainly originated from the nucleation effect of SCNFs and the strong hydrogen bonding interaction between the SCNFs and PLA, benefiting from the good dispersion of SCNFs in PLA. Conversely, the modulus obtained by the Halpin–Tsai model was lower than that of the PLA/SCNF nanocomposite, because the l_f/d_f value of 1 adopted led to an underestimated theoretical modulus. Consequently, a higher Young’s modulus was observed for the nanocomposite.

Fig. 3 Young’s modulus of the nanocomposites versus SCNF content; lines represent the theoretical predictions for the nanocomposites using the Voigt–Reuss model and Halpin–Tsai model.

Thermal stability

It is reported that no improvement and even a decrease in the thermal stability of PLA was achieved by adding the unmodified and modified rod-like NCs, which depended on the preparation or modification methods of the NCs.^1,6 Thus the effect of SCNFs on the thermal stability of PLA-based nanocomposites was studied by TGA under nitrogen atmosphere. The TGA and derivative DTG curves are reported in Fig. 4(a) and (b), respectively. Table 1 records the thermal parameters, such as initial degradation temperature (T₀), temperature at 5% weight loss (T_5%), maximum degradation temperature (T_max), and complete decomposition temperature (T_f). All the samples showed a single degradation peak associated to the PLA thermal degradation between 350 and 380 °C, and great increases in the T₀ and T_max values by the addition of SCNFs were observed for the nanocomposites. It suggests that the incorporation of SCNFs can induce a significant enhancement in the thermal stability of PLA. With an increase of the SCNF content, T₀ and T_max were increased to maximum values for the nanocomposite with 10 wt% SCNFs, and then slightly reduced (Fig. 4(c)). Compared to PLA, T₀ and T_max were increased by 17.4 and 21.5 °C, respectively. It hints that the nanocomposites with good thermal stability can be easily processed using high temperature melt-processing techniques.

Fig. 4 (a) TGA, (b) DTG curves, (c) thermal parameters and (d) the plot of ln[ln(W₀/W_T)] vs. θ for pure PLA, and the nanocomposites with various SCNF content.

Table 1 Thermal analysis parameters for pure PLA, SCNFs and the nanocomposites with various SCNF content

Sample	T0^a (°C)	T5%^a (°C)	Tmax^a (°C)	Tf^a (°C)	Activation energy (Ea) (kJ mol^−1)
a T0, T5%, Tmax and Tf were obtained from the TGA curves at a heating rate of 10 °C min^−1.
PLA	335.7	325.9	354.6	363.9	319.95
1 wt%	336.3	327.0	357.0	369.4	360.13
3 wt%	342.2	332.4	361.6	371.5	380.61
5 wt%	342.6	333.0	363.1	372.3	407.57
10 wt%	353.1	342.2	376.1	390.0	414.77
15 wt%	350.8	338.9	373.5	383.4	395.16
20 wt%	346.6	331.4	366.3	379.2	373.37
SCNFs	323.1	292.2	358.0	368.3	248.30

---

Apparent activation energies (E_a) for pure PLA and the nanocomposites were obtained from TGA data by applying the Horowitz and Metzger method as in eqn (4) below:²³

where W₀ is the initial weight of the polymer, W_T is the residual weight of the polymer at temperature T, T_s is the temperature determined at 36.79% weight loss, θ is T − T_s, and R is the gas constant.

Fig. 4(d) shows the plot of ln[ln(W₀/W_T)] vs. θ for the main stage of thermal degradation of pure PLA and the nanocomposites. The average apparent activation energies (E_a) obtained from the slopes are shown in Fig. 4(d) and summarized in Table 1. The higher E_a is, the faster the decomposition rate is. Table 1 illustrates that E_a of the nanocomposites was larger than that of pure PLA, implying that the onset decomposition of the nanocomposites began at a higher temperature thus leading to a quicker decomposition rate of the nanocomposites. During decomposition of the nanocomposites, the introduction of SCNFs increased the temperature of thermal decomposition so that the beginning of thermal decomposition of the nanocomposites occurred at a higher temperature than that of pure PLA. The higher temperature was the primary factor to result in the increased decomposition rate of the nanocomposites.

Fractured morphologies and optical property

The homogenous dispersion of the SCNFs is beneficial to obtain the high nucleation efficiency of SCNFs in PLA crystallization and form more hydrogen bonds between the two components. Therefore, the dispersion state of the SCNFs was estimated by observing the fracture morphologies and transparency of the PLA/SCNF nanocomposites. As shown in Fig. 5, a smooth surface was observed for the pure PLA film sample corresponding to the brittle nature of PLA. With incorporation of less than 10 wt% SCNFs, the fractured surfaces of the nanocomposites showed no visible microscale aggregation of the SCNFs in the PLA film. It suggests that the hydrophobic formate groups of the SCNFs should be helpful to improve the compatibility between the nanofillers and matrix phases, and prevent nanoparticle aggregation. When the SCNF content reached 15 and 20 wt%, rough film surfaces were observed due to the agglomerated SCNFs in the PLA. These results can be also supported by external appearance of PLA and the nanocomposite films, although the transmittance at 800 nm decreased with an increase of the SCNF content (Fig. 5(h)). These observations are in accordance with the largest improvements in the mechanical and thermal stability for 10 wt% SCNFs.

Fig. 5 FE-SEM images for the fractured morphologies of (a) PLA and the nanocomposites with a SCNF content of (b) 1, (c) 3, (d) 5, (e) 10, (f) 15 and (g) 20 wt%, and (h) UV-vis transmittance spectra.

Chemical structure

FT-IR spectra were recorded for PLA and the PLA/SCNF films in order to investigate crystallinity changes and hydrogen bonds between the SCNFs and PLA. The FT-IR spectra are shown in Fig. 6(a), and a peak at 1753 cm⁻¹ is assigned to characteristic carbonyl groups from all the PLA-based nanocomposites. The distribution of C=O in the ordered and amorphous regions of the nanocomposites is usually affected by the nanofiller loading levels and the interactions between the SCNFs and matrix.²⁴ Curve-fitting was used to divide the spectra of the pure PLA and PLA-based nanocomposites (as shown in Fig. 6(b)) from 1650 cm⁻¹ to 1850 cm⁻¹ into three peaks, i.e. peak I located at about 1775 cm⁻¹, assigned to free C=O, peak II located at about 1753 cm⁻¹, assigned to C=O in the amorphous region, and peak III located at about 1734 cm⁻¹, assigned to C=O in the crystalline domain.²⁴ Table 2 summarizes the detailed locations and fractions of peaks I, II and III for all the samples. Amorphous and free C=O components were in the majority in pure PLA. With an increase of the SCNF content (below 10 wt%), a decrease of C=O in the amorphous fraction was observed, together with an increase in the C=O fraction of the crystalline domain. In addition, the incorporation of superfluous nanofillers reduced the proportion of the crystalline domain in PLA.

Fig. 6 (a) Full ATR FT-IR spectra, (b) the curve-fitted FT-IR spectra of PLA/SCNFs with 10 wt% SCNFs in the range of 1650–1850 cm⁻¹ for reference (experimental curve; free C=O; C=O in amorphous region; C=O in ordered domain), (c) carbonyl stretching region (ν_C=O) in the infrared spectra (inset is the nanocomposite with 10 wt% SCNFs), and (d) hydrogen bond fractions (F_H–CO) of pure PLA and the nanocomposites with various SCNF content.

Table 2 Location and fraction of curve-fitting peaks for C=O absorption of the PLA/SCNF nanocomposites

Sample	Peak I^a		Peak II^a		Peak III^a
	Location (cm^−1)	Fraction (%)	Location (cm^−1)	Fraction (%)	Location (cm^−1)	Fraction (%)
a Peak I: free C=O; peak II: C=O in amorphous region; peak III: C=O in crystalline domain.
PLA	1775.5	7.8	1754.9	75.8	1734.2	16.4
1 wt%	1775.6	5.6	1755.7	75.7	1734.5	18.6
3 wt%	1775.1	5.4	1755.5	72.7	1735.0	21.9
5 wt%	1775.1	12.2	1755.0	60.7	1734.2	27.1
10 wt%	1775.2	6.2	1753.5	63.8	1734.4	30.0
15 wt%	1775.5	12.7	1753.7	63.4	1734.0	23.9
20 wt%	1775.6	5.8	1753.7	76.8	1734.5	17.3

---

The Gaussian/Lorentzian spectral function was employed to curve-fit the carbonyl band in the range from 1680 to 1800 cm⁻¹ to calculate the hydrogen bond fraction (F_H–CO). The bands of hydrogen-bonded and free C=O groups were located at around 1753 cm⁻¹ and 1765 cm⁻¹, respectively²⁵ (inset in Fig. 6(c)). Compared to pure PLA, a decrease in the hydrogen-bonded band position to 1749 cm⁻¹ was observed for all the nanocomposites with the increase of the SCNF content. This was due to the formation of intermolecular hydrogen bonding interaction between residual hydroxyl groups of SCNFs and carbonyl groups of PLA, impairing the polarity of carbonyl groups and thus leading to a decrease in the C=O band position.²⁶ In order to further investigate the variations in hydrogen bonding interaction of the nanocomposites, F_H–CO values were calculated, which can be obtained by the following eqn (5):^20,26

where A_a and A_H are the peak areas of free and hydrogen-bonded components, respectively, and r_H/a is the specific absorption ratio of the above two bands. Depending on the strength of the hydrogen bonds, r_H/a should be within a range from 1.2 to 1.75 for semi-quantitative comparison. F_H–CO of the nanocomposites as a function of the SCNF content is presented in Fig. 6(d). With an increase of the SCNF content, F_H–CO primarily increased from 0.11 to a maximum value of 0.21 for 10 wt%, and then slightly reduced to 0.16 for 20 wt%. It indicates that the nanocomposite with 10 wt% SCNFs possessed the most intermolecular hydrogen bonds due to efficient dispersion of SCNFs within the PLA, which greatly contributed to the obvious enhancements in the mechanical and thermal properties of the nanocomposites.

Crystal structure

WAXD analysis of pure PLA and PLA/SCNF nanocomposites was carried out in order to study the crystalline structure. Fig. 7 shows no obvious diffraction peaks in pure PLA, suggesting its ambiguous characteristic nature. Three typical diffraction peaks (12.2°, 20.2°, and 21.8°) of cellulose II crystalline appeared in the SCNFs. For the nanocomposites, the SCNF diffraction signals were masked by PLA crystalline diffraction at a low SCNF content (<5%), and no new diffraction peaks were found for the nanocomposites, suggesting that the crystal structure of PLA was not changed by incorporating SCNFs. A similar behavior was detected for the PLA nanocomposite with a low content (<5%) of modified NCs due to the broad shoulder of PLA at about 16°.⁸ However, when enough SCNFs were introduced (>5%), steady increases in the intensity of the peaks at 20.2° and 21.8° suggested an increase in the overall crystallinity of the nanocomposites.

Fig. 7 X-ray diffraction (XRD) patterns of pure PLA, SCNFs and the nanocomposites with various SCNF content.

Isothermal crystallization and melting behavior

In order to deeply understand the effect of the SCNFs on the bulk crystallization rate of PLA, isothermal crystallization behaviors of the PLA/SCNF nanocomposites were also studied. The isothermal crystallization exotherms of PLA and the PLA nanocomposites are shown in Fig. 8. Within the same sample, the shift of exotherms to a longer time with increasing crystallization temperature (T_c) indicated a decrease of the crystallization rate.¹⁵ Under the same T_c, with the increase of the SCNF content, the crystallization curves of the nanocomposites (Fig. 8) shifted to a shorter crystallization time, and the shortest crystallization time was achieved for the nanocomposite with 10 wt% SCNFs. It indicates that the nanocomposite with 10 wt% SCNFs possessed the highest crystallization rate due to the addition of the well-dispersed and most SCNFs as nucleating agents.

Fig. 8 Isothermal crystallization exotherms of PLA (a), and PLA nanocomposites with various SCNF content: 1 (b), 3 (c), 5 (d), 10 (e), 15 (f), and 20 (g) wt% under different crystallization temperatures (T_c), and (h) crystallization time at 110 °C and crystallinity calculated at the heating rate of 2 °C min⁻¹ of PLA and PLA nanocomposites.

The relative crystallinity (X_t) as a function of crystallization time (t) is calculated as follows:

To describe the isothermal crystallization kinetics, the classical Avrami equation was employed:

log[−ln(1 − X_t)] = log k + n log t (7)

where k is the overall kinetic constant depending on the geometry of the growing crystalline phase, and n is the Avrami exponent correlating with the nucleation mechanism and crystal growth dimension. Fig. 9(a) gives the evolution of X_t with t for pure PLA and the nanocomposites at their crystallization temperatures. In Fig. 9(a), the crystallization finished within 50.72 min for pure PLA while it was within 50.11, 47.32, 28.55, 26.63, 28.37, and 38.76 min for nanocomposites with 1, 3, 5 10, 15 and 20 wt% SCNFs, respectively. It is clear that the addition of SCNFs could improve the isothermal crystallization of PLA. Moreover, the SCNFs with a 10 wt% content were more efficient in accelerating the PLA crystallization process than the other nanocomposites. log[−ln(1 − X_t)] versus log t is plotted in Fig. 9(b), and the Avrami exponent (n) and the crystallization rate constant (k) were calculated from the slope and intercept of the linear fit. The Avrami equation is usually used to evaluate the crystal growth of PLA at the primary crystallization stage.²⁷ Thus, the initial linear portions of the plots were analyzed and the results are summarized in Table 3. The Avrami exponent, n, was related to the nature of nucleation and the dimensionality of the growing crystals. The n values of PLA were around 2.8–3.0, suggesting a three-dimensional crystallization growth and homogeneous nucleation mechanism.^24,28 The obtained n values were similar to those reported for pure PLA (2.3–3.2 at T_c = 90–130 °C by Tsuji et al.²⁹ and 2.8–3.2 at T_c = 90–130 °C by Iannace et al.³⁰). The n values of the PLA/SCNF nanocomposites were around 3.0–4.5, indicating a crystallization mode of heterogeneous nucleation.^15,31 Similar n values were reported for the PLA/nucleating agent (montmorillonite, NC) nanocomposites, where the Avrami exponent n was in the range of 3.0–4.9.^15,31

Fig. 9 (a) Plots of the relative crystallinity vs. time and (b) Avrami plots of log[−ln(1 − X_t)] vs. log t for the pure PLA samples and its nanocomposites cooled from the melt at an isothermal crystallization temperature of 110 °C.

Table 3 Avrami kinetic parameters for the isothermal crystallization of the PLA and PLA/SCNF nanocomposites

Sample	Tc (°C)	n	k (min^−n)	t1/2 (min)
PLA	110	2.98	2.29 × 10^−5	31.9
	115	2.80	2.25 × 10^−5	40.1
	120	—	—	—
	125	—	—	—
1 wt%	110	4.01	1.06 × 10^−6	28.2
	115	3.37	4.59 × 10^−6	34.1
	120	3.06	9.68 × 10^−6	38.6
	125	3.02	6.51 × 10^−6	46.2
3 wt%	110	4.03	3.28 × 10^−6	20.9
	115	3.43	8.43 × 10^−6	27.1
	120	3.19	6.66 × 10^−6	37.4
	125	3.16	4.85 × 10^−6	42.8
5 wt%	110	4.31	3.36 × 10^−6	17.1
	115	3.42	1.78 × 10^−5	22.0
	120	3.37	5.34 × 10^−6	32.9
	125	3.14	5.67 × 10^−6	41.7
10 wt%	110	4.52	2.12 × 10^−6	16.6
	115	4.12	2.95 × 10^−6	20.1
	120	3.79	4.07 × 10^−6	24.0
	125	3.48	2.99 × 10^−6	34.8
15 wt%	110	4.40	2.81 × 10^−6	16.8
	115	3.83	3.02 × 10^−6	25.1
	120	3.52	2.70 × 10^−6	34.4
	125	3.32	4.27 × 10^−6	37.1
20 wt%	110	4.37	1.26 × 10^−6	20.6
	115	3.85	2.51 × 10^−6	25.9
	120	3.39	3.71 × 10^−6	35.9
	125	3.24	3.17 × 10^−6	44.6

---

The half-time of crystallization (t_1/2) is defined as the time when X_t is equal to 0.5, which can be calculated as follows:¹⁵

The results are shown in Table 3. It is found that the strong dependence of t_1/2 on the crystallization temperature is evident. For samples crystallized from the melt, the t_1/2 value was increased with an increase of the crystallization temperature. Furthermore, at a given crystallization temperature, the nanocomposite had a smaller t_1/2 value than that of pure PLA. This indicates that the nanocomposites had higher crystallization rates (lower t_1/2) than that of PLA. For example, the t_1/2 value was 16.6 min for the nanocomposites with 10 wt% SCNFs, which was obviously lower than that of pure PLA (31.9 min) (Table 3). This result demonstrates that the crystallization rate of the nanocomposite with 10 wt% SCNFs was about two times that of pure PLA. This was ascribed to the strong nucleation effect of SCNFs in the nanocomposites. Instead, t_1/2 values of PLA nanocomposites containing 1 and 3 wt% rod-like NCs were 9.2 and 9.7 min, respectively, which are larger than the 8.6 min for PLA. It hints that the addition of rod-like NCs can not improve the PLA crystallization due to a poor nucleating effect of the aggregated NCs.¹⁵

Moreover, the crystallinity of the nanocomposite could be prompted by the incorporation of SCNFs. The crystallinity of pure PLA was only 3% at a low heating rate of 2 °C min⁻¹, while the introduction of SCNFs induced an increase in the crystallinity of the nanocomposites. Particularly, the nanocomposite with 10% SCNFs showed the highest crystallinity of 19.8% (Table S1†). Besides, the POM images show that the amounts of spherulites with a smaller diameter (in the nanocomposite with 10 wt% SCNFs) were significantly increased, compared to pure PLA (Fig. S3†), illustrating that the SCNFs as efficient nucleating agents could accelerate the folding and crystallization (rate) of PLA chains, meanwhile the entanglements of the PLA chains were enhanced (Fig. S2†),^9,25 which could also transfer efficiently applied stress upon the PLA and thus improve the mechanical strength. Moreover, the incorporation of SCNFs showed good barrier properties in water uptake and water vapour permeability, and reduced greatly the migration level of food stimulants (Fig. S4†). It illustrates that the PLA/SCNF nanocomposites can used as active food film or packaging materials.

Conclusions

In this work, we first reported a simple route to overcome the main industrial problem of PLA (slow crystallization rate) by adding SCNFs as an efficient bio-based nucleating agent. It is found that SCNFs showed good dispersion in PLA without the help of surfactants or compatibilizers, and thus significantly improved the PLA crystallization ability and interfacial interaction between two components, leading to significant enhancements in the mechanical performance and thermal stability of the resulting nanocomposites. Besides, the incorporation of SCNFs showed positive effects on the barrier and migration properties of the nanocomposite. These results showed that these high performance nanocomposites may be highly suitable as active food packaging materials.

Acknowledgements

The work had support from National Natural Science Foundation of China (51403187), the public technology research plan of Zhejiang Province, China under Grant No. 2015C33111, Zhejiang Provincial Natural Science Foundation of China under Grant No. LQ14E030007, “521” Talent Project of Zhejiang Sci-Tech University and Open fund in Top Priority Discipline of Zhejiang Province in Zhejiang Sci-Tech University (2015YXQN04). In addition, we thank the revision in the manuscript by Somia Y. H. Abdalkarim.

References

1. E. Fortunati, F. Luzi, D. Puglia, R. Petrucci, J. M. Kenny and L. Torre, Ind. Crops Prod., 2015, 67, 439–447.
2. B. Baghaei and M. Skrifvars, Composites, Part A, 2015, 81, 139–144.
3. S. K. Pankaj, C. N. Bueno-Ferrer, N. Misra, L. O’Neill, A. Jiménez, P. Bourkea and P. J. Cullena, Innovative Food Sci. Emerging Technol., 2014, 21, 107–113.
4. E. Robles, I. Urruzola, J. Labidi and L. Serrano, Ind. Crops Prod., 2015, 71, 44–53.
5. A. Mujica-Garcia, S. Hooshmand, M. Skrifvars, J. M. Kenny, K. Oksman and L. Peponi, RSC Adv., 2016, 6, 9221–9231.
6. M. P. Arrieta, E. Fortunati, F. Dominici, E. Rayón, J. López and J. M. Kenny, Carbohydr. Polym., 2014, 107, 16–24.
7. M. P. Arrieta, E. Fortunati, F. Dominici, J. López and J. M. Kenny, Carbohydr. Polym., 2015, 121, 65–275.
8. E. Fortunati, M. Peltzer, I. Armentano, L. Torre, A. Jiménez and J. M. Kenny, Carbohydr. Polym., 2012, 90, 948–956.
9. E. Fortunati, D. Puglia, F. Luzi, C. Santulli, J. M. Kenny and L. Torre, Carbohydr. Polym., 2013, 97, 825–836.
10. N. Johar, I. Ahmad and A. Dufresne, Ind. Crops Prod., 2012, 37, 93–99.
11. L. Dong, X. Zhang, S. Ren, T. Lei, X. Sun, Y. Qi and Q. Wu, RSC Adv., 2016, 6, 6436–6442.
12. E. Fortunati, I. Armentano, Q. Zhou, D. Puglia, A. Terenzi, L. A. Berglund and J. M. Kennya, Polym. Degrad. Stab., 2012, 97, 2027–2036.
13. E. Fortunatia, S. Rinaldi, M. Peltzer, N. Bloise, L. Visai, I. Armentano, A. Jiménez, L. Latterini and J. M. Kenny, Carbohydr. Polym., 2014, 101, 1122–1133.
14. A. L. Goffin, J. M. Raquez, E. Duquesne, G. Siqueira, Y. Habibi, A. Dufresne and D. Philippe, Biomacromolecules, 2011, 12, 2456–2465.
15. A. Pei, Q. Zhou and L. A. Berglund, Compos. Sci. Technol., 2010, 70, 815–821.
16. C. F. Yan, H. Y. Yu and J. M. Yao, Cellulose, 2015, 22, 3773–3788.
17. H. Y. Yu, Z. Y. Qin, B. Sun, X. G. Yang and J. M. Yao, Compos. Sci. Technol., 2014, 94, 96–104.
18. H. Y. Yu, C. F. Yan, X. X. Lei, Z. Y. Qin and J. M. Yao, Mater. Lett., 2014, 131, 12–15.
19. M. P. Arrieta, E. Fortunati, F. Dominici, E. Rayón, J. López and J. M. Kenny, Polym. Degrad. Stab., 2014, 107, 139–149.
20. H. Y. Yu, Z. Y. Qin, Y. N. Liu, L. Chen, N. Liu and Z. Zhou, Carbohydr. Polym., 2012, 89, 971–978.
21. M. R. Kamal and V. Khoshkava, Carbohydr. Polym., 2015, 123, 105–114.
22. S. Fujisawa, T. Saito, S. Kimura, T. Iwata and A. Isogai, Compos. Sci. Technol., 2014, 90, 96–101.
23. M. Bayrakci, O. Gezici, S. Z. Bas, M. Ozmen and E. Maltas, Mater. Sci. Eng., C, 2014, 42, 546–552.
24. N. Lin, J. Huang, P. R. Chang, J. Feng and J. Yu, Carbohydr. Polym., 2011, 83, 1834–1842.
25. C. C. Liao, C. C. Wang and C. Y. Chen, Polymer, 2011, 52, 4303–4318.
26. M. M. Haafiz, A. Hassan, Z. Zakaria, I. M. Inuwa, M. S. Islam and M. Jawaid, Carbohydr. Polym., 2013, 98, 139–145.
27. J. M. Zhang, H. Tsuji, I. Noda and Y. Ozaki, Macromolecules, 2004, 37, 6433–6439.
28. V. Krikorian and D. J. Pochan, Macromolecules, 2004, 37, 6480–6491.
29. H. Tsuji, H. Takai and S. K. Saha, Polymer, 2006, 47, 3826–3837.
30. S. Iannace and L. Nicolais, J. Appl. Polym. Sci., 1997, 64, 911–919.
31. G. Z. Papageorgiou, D. S. Achilias, S. Nanaki, T. Beslikas and D. Bikiaris, Thermochim. Acta, 2010, 511, 129–139.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra02768g

---