Yingying Hea,
Jiang Zhu*a,
Wentao Wanga and
Haitao Ni*b
aCollege of Materials and Chemical Engineering, Chongqing University of Arts and Sciences, Chongqing 402160, People's Republic of China. E-mail: jiangzhu415@163.com
bChongqing Key Laboratory of Environmental Materials & Remediation Technology, Chongqing University of Arts and Science, Chongqing 402160, People's Republic of China. E-mail: htniok@163.com
First published on 14th November 2018
Hydrophilic cellulose nanocrystals (CNC) are typically poorly dispersed in non-polar polymer matrices and, hence, a method for the surface modification of CNC is developed for improving this dispersion. This method included an esterification reaction with acetic anhydride, butyric anhydride, and caproic anhydride. The particle size distribution (range of sizes: 80–310 nm) of CNC was determined. The SEM-EDAX indicated that (i) the structure of CNC was maintained even after incorporation of the acid anhydride and (ii) the carbon C content of modified-CNC was higher than that of pure CNC. The chemical structure of modified-CNC was identified via FT-IR and 13C NMR spectroscopy. The contact angle of CNC and modified-CNC with water and methylene iodide was measured. The surface energy of modified-CNC was lower than that of pure CNC. Thermal-property measurements indicated that the initial decomposition temperature (based on 5 wt%) of the modified-CNC was slightly higher than that of CNC. Poly(butylene succinate) (PBS) composites were obtained by mixing modified-CNC into a PBS matrix via simple melt blending of a double screw. The PBS/modified-CNC composites were investigated via DSC and XRD. Tensile testing indicated that the tensile modulus improved gradually with increasing modified-CNC content, whereas the elongation at fracture decreased.
Several properties of polymer composites depend significantly on the dispersion state of nanoparticles, i.e., the reinforcing effects result from the percolating network of CNC and excellent interfacial compatibility between the matrix and the fillers.12 Nevertheless, the hydrogen bonding between the three hydroxyl groups comprising the repeat unit of CNC occurs in aggregates. Furthermore, owing to differences in the corresponding polarities, CNC is poorly dispersed in non-polar polymer matrices. Methods for CNC modification, such as esterification, graft copolymers, acetylation, and silylation, have been developed with the aim of overcoming this problem.13–16 The dispersion of CNC may be improved, for example, by reducing the number of hydroxyl groups via esterification.
PBS is biodegradable synthetic polyester that exhibits good biocompatibility, and has the potential for widespread use in various industrial applications (such as garbage bags, packaging bags, and agricultural materials). However, PBS suffers from drawbacks such as low rigidity, high cost, and a low rate of degradation.17–19 To overcome these drawbacks, and expand the range of PBS applications, PBS composites enhanced with natural reinforcements (such as coffee shell-fiber, straw fiber, and bamboo fiber) have been proposed.20–24
According to the aforementioned studies, dispersion and interfacial adhesion of polymer composites are very important factors and, hence, improving the dispersion of CNC nanoparticles in non-polar PBS polymer matrices is essential. In the present study, CNC was obtained via the acid hydrolysis of micro-crystalline cellulose (MCC). Accordingly, CNC was modified with different length chains of acid anhydrides. Through an esterification reaction, hydrogen in the hydroxyl group of CNC was substituted by a hydrophobic ester group. The modified-CNC would promote interfacial adhesion with PBS. The chemical structures of CNC and modified-CNC were identified via Fourier transform infrared (FT-IR) and 13C nuclear magnetic resonance (NMR) spectroscopy. Thermogravimetric analysis (TGA) was performed on CNC and modified-CNC nanoparticles and the contact angles with water and methylene iodide were measured in each case. Similarly, the surface free energy of CNC and modified-CNC was calculated. The corresponding PBS/modified-CNC composites were prepared via simple melt blending of a double screw, and the morphology as well as the crystalline structure of these composites was investigated. The reinforcing effect of modified-CNC was investigated via differential scanning calorimetry and mechanical-property characterization.
The morphology of modified-CNC and PBS/modified-CNC composites was observed on via scanning electron microscopy (SEM; FEI Quanta 250) performed at an accelerating voltage of 10 kV. Prior to observation, the sample surfaces were sputtered with a layer of gold. The fraction of carbon (C) present in the modified-CNC was determined via energy dispersive X-ray analysis (EDAX; Elementar Vario EL Cube, Germany).
Infrared spectra from KBr pellets were obtained on a Fourier Transform Infrared (FT-IR) spectrometer (Nicolet T6670; Thermo Fisher Scientific, the United States) spectra were collected for wavenumbers ranging from 3700 to 500 cm−1 at a resolution of 4 cm−1.
The room-temperature structure of the modified-CNC was confirmed via carbon nuclear magnetic resonance (13C NMR) spectroscopy performed on a 400 MHz NMR spectrometer (Bruker AV II-400, Bruker, Germany).
CNC, a-CNC, b-CNC, and c-CNC powder were pressed into round pellets using a tablet press (YP-2 model; Mountains Scientific Instrument Co., Ltd, Shanghai China). The contact angles of water and methylene iodide on each material were measured using a contact angle goniometer (YIKE-360A model; Chengde Precision Test Instrument Factory, China). In addition, the surface free energies of each material were calculated from the Owen–Wendt equation and Young's equation.
Thermogravimetric analysis (TGA) of CNC and the modified-CNC was performed with a thermal analyzer (TA Instruments TGA Q500 V20.13 Build 39). During the measurements, each sample was heated from room temperature to 600 °C at a heating rate of 10 °C min−1, under nitrogen atmosphere.
The thermal properties of pure PBS, PBS/a-CNC, PBS/b-CNC, and PBS/c-CNC composites were measured via differential scanning calorimetry (DSC, Q2000 V24.11 Build 124) of 3–10 mg samples. The samples were measured at a scanning rate of 10 °C min−1 in a sealed aluminum pot under a protective nitrogen atmosphere (calibration: indium standard). The initial sample was held for 5 min at 140 °C to erase the thermal history, and then cooled from 140 to 20 °C as the first cooling run. Subsequently, the sample reheated from 20 to 140 °C during the second heating scan, and the exothermic curves were recorded for analysis.
The X-ray diffraction (XRD) patterns of the PBS/modified-CNC composites were obtained with a DX-1000 diffractometer (Dandong Fangyuan Instrument Co., Ltd, China) using CuKα radiation. The voltage and current were set to 40 kV and 25 mA, respectively. Diffraction patterns were recorded at a scanning rate of 2° min−1 for 2θ values ranging from 2° to 65°.
The mechanical properties of pure PBS, PBS/a-CNC, PBS/b-CNC, and PBS/c-CNC composites were measured on an Electronic Universal Testing Machine (UTM4204) with large rotational deformation. Tests were performed at room temperature and a tensile speed of 50 mm min−1 on dumbbell-shaped samples (gage length between grips: was 25 mm). Five samples were measured for each material, and the average of the five measurement results was reported for all samples.
The influence of acid anhydrides on modified-CNC was investigated by evaluating the C elementary percentage of CNC, a-CNC, b-CNC, and c-CNC (Table 1). During this investigation, the C elementary percentage at three points on each sample was determined via SEM-EDAX. As Table 1 shows, an average percentage value of 83.61% was obtained for CNC. This value is lower than the average values obtained for a-CNC (87.09%), b-CNC (88.31%), and c-CNC (89.60%), indicating that CNC surfaces were successfully modified by acetic anhydride, butyric anhydride and caproic anhydride, respectively. Moreover, the C elementary percentage increased with increasing chain length of each acid anhydride.
C atom | CNC | a-CNC | b-CNC | c-CNC |
---|---|---|---|---|
At% | 84.20 | 88.10 | 88.10 | 90.14 |
81.28 | 86.79 | 86.79 | 89.59 | |
85.35 | 86.38 | 90.03 | 89.07 | |
Average at% | 83.61 | 87.09 | 88.13 | 89.60 |
The esterification of CNC was characterized by the emergence of a new peak at 1738 cm−1 corresponding to the carbonyl (CO) associated with the formed ester group. Similarly, the new absorption band at 1246 cm−1 was attributed to the C–O stretch vibration of the carbonyl. Lack of changes in the peaks corresponding to CNC, suggests that the esterification reaction was successfully.
The average contact angles of water and methylene iodide on CNC and modified-CNC are summarized in Table 3. As Fig. 6 shows, the respective contact angles of water and methylene iodide on CNC, a-CNC, b-CNC, and c-CNC after 30 s.
Samples | Water (°) | Methylene iodide (°) | Dispersion force (mJ m−2) | Polar force (mJ m−2) | Total force (mJ m−2) |
---|---|---|---|---|---|
CNC | ≤5 | 40.26 | 24.30 | 47.75 | 72.05 |
a-CNC | ≤5 | 28.92 | 29.16 | 43.56 | 72.72 |
b-CNC | 76.20 | 28.24 | 41.09 | 4.28 | 45.37 |
c-CNC | 73.3 | 29.60 | 39.69 | 5.71 | 45.40 |
Fig. 6 Water contact angle of (a) CNC, (b) a-CNC, (c) b-CNC, (d) c-CNC; methylene iodide contact angle of (A) CNC, (B) a-CNC, (C) b-CNC, (D) c-CNC. |
The water contact angle of CNC and the water contact angle of a-CNC were both lower than 5° after 30 s and 90 s, respectively. However, after 120 s, the water contact angles of b-CNC and c-CNC were 76.2° and 73.3°, respectively, owing to the fact that long ester chain induces hydrophobicity in the hydrophilic CNC. Furthermore, the respective contact angles (28.92°, 28.24°, and 29.6°) of methylene iodide on a-CNC, b-CNC, and c-CNC were lower than that of methylene iodide on CNC (Table 3). Each of these angles decreased to <5° after 180 s. The surface free energies of CNC, a-CNC, b-CNC, and c-CNC were calculated from:29
Young's equation: γs = γlcosθ + γsl | (1) |
Dupréus equation: W = γs + γl − γsl | (2) |
Young–Dupré equation: W = γl(1 + cosθ) | (3) |
Owen–Wendt equation:
γsl = γs + γl − 2(γdsγdl)0.5 − 2(γdsγdl)0.5 | (4) |
γl(1 + cosθ) = 2(γdsγdl)0.5 + 2(γdsγdl)0.5 |
Calculations of the surface free energy yielded γl, γld, and γlp values of 72.2, 22.0, and 50.2 mJ m−2, respectively, for water and corresponding values of 50.8, 48.5, and 2.30 mJ m−2 for methylene iodide.29 As Table 3 shows the surface free energy values of b-CNC and c-CNC are considerably lower than those of CNC and a-CNC. This decrease in surface energy contributed significantly to the reduction of the polar component, indicating that (i) the hydroxyl groups were substituted with non-polar ester groups and (ii) b-CNC and c-CNC were well dispersed in the non-polar polymer.
Fig. 7 (a) TGA and (b) DTG curves of CNC, a-CNC, b-CNC, and c-CNC; (c and d) DSC cooling and (e and f) the second heating scans of pure PBS and PBS/modified CNC composites. |
The differential thermogravimetric (DTG) curves for CNC, a-CNC, b-CNC, and c-CNC and the decomposition temperatures based on these curves are shown in Fig. 7(b) and listed in Table 4, respectively. As the Fig. 7(b) shows, CNC exhibited greater thermal stability than a-CNC at Tmax1, but less stability than b-CNC and c-CNC. In contrast to CNC, the modified materials decompose at two ranges of temperatures, characterized by maximum temperatures of Tmax1 and Tmax2. Residues of 3.19, 1.79, and 2.24% were obtained for a-CNC, b-CNC and c-CNC, respectively. The CNC was characterized by three maximal decomposition temperatures (referred to as Tmax1, Tmax2, and Tmax3). This resulted possibly from the broad particle size distribution and molecular weight distribution of CNC i.e., the decomposition temperature would vary with the molecular weight. Table 4 reveals Tmax1 values of 227.97 °C, 221.71 °C, 246.93 °C, and 249.17 °C, as well as Tmax2 values of 317.87 °C, 489.25 °C, 484.38 °C, and 471.54 °C for CNC, a-CNC, b-CNC and c-CNC, respectively. These Tmax1 values indicate that the thermal stability of b-CNC and c-CNC is greater than that of CNC while, in the case of Tmax2, the thermal stability of all three modified materials is greater than that of CNC. At Tmax3 (491.89 °C), decomposition occurred only in the case of CNC. In fact, at decomposition temperatures lower than Tmax3, the thermal stability of CNC was lower than that of b-CNC and c-CNC. This resulted from the fact that hydrogen bonds in CNC were partially substituted by ester groups. Therefore, for graft acid anhydride, the thermal stability fluctuated with increasing length of the esterification chains, with b-CNC exhibiting better thermal stability than a-CNC and c-CNC.
Improved interfacial adhesion between the polymer matrix and nanoparticles would have a significant influence on the thermal performance and crystallization behavior of the matrix. Therefore, DSC-based investigation focused on determining the effect of a-CNC, b-CNC, and c-CNC content on the thermal performance of PBS was warranted. The DSC cooling curves of pure PBS, PBS/a-CNC, PBS/b-CNC, and PBS/c-CNC nanocomposites are shown in Fig. 7(c and d). PBS is a typical semi-crystalline polymer. The non-isothermal crystallization of pure PBS, PBS/a-CNC, and PBS/c-CNC composites is characterized by an exothermic peak that occurs at the crystallization peak temperature (Tc, in this case, 69.8 °C). The Tc of pure PBS, PBS/a-CNC, and PBS/c-CNC composites varied only modestly, indicating that the crystallization rate of PBS remained approximately constant with a-CNC and c-CNC nanoparticle filling. However, comparing with the Tc resulting from a-CNC and c-CNC filling, the Tc shifted by larger amounts after the incorporation of b-CNC nanoparticles. With the addition of 3.0 wt% b-CNC nanoparticles, the Tc of PBS decreased to 67.1 °C. This change indicated that crystallization depends possibly on the mobility of PBS molecular chains in the composites. These chains were possibly confined by the greater number of ester groups in b-CNC nanoparticles (compared with the number in a-CNC and c-CNC nanoparticles). As Fig. 7(d) shows when the b-CNC content is increased, the Tc of PBS/b-CNC composites decreased slightly to lower temperatures than the Tc of pure PBS. This decrease may be attributed to the excellent dispersion of b-CNC nanoparticles in PBS and, in turn, possible increase in the viscosity. However, the mobility of the PBS molecular chain decreased leading to a strong interaction between the well-dispersed b-CNC nanoparticles and the PBS matrix. This interaction, in turn, hindered crystallization. The DSC traces obtained in the second heating run of pure PBS and PBS/modified-CNC composites are shown in Fig. 7(e and f). The Fig. 7(e) reveals double melting peaks and small variations in the melting temperature (Tm) of pure PBS, PBS/a-CNC, and PBS/c-CNC composites. The minor peak (Tm1) and the major peak (Tm2) occur at 93.3 °C and 102.6 °C, respectively. However, the Tm of PBS/b-CNC composites decreased slightly with increasing b-CNC content to values lower than the Tm of pure PBS. The Tm1 and Tm2 of the PBS/b-CNC composites occurred at 90.8 °C and 101.5 °C, respectively. As Fig. 7(f) shows, the peak temperature of Tm1 shifted gradually to lower values with increasing content of b-CNC nanoparticles, whereas Tm2 remained constant at 101.5 °C. The double melting peaks may have resulted from the melting-recrystallization mechanism. The Tm1 peak resulted from the melting of initial crystals formed prior to the second DSC scan, and the Tm2 peak resulted from the melting of recrystallized crystals during the heating scan. In other words, the heating process consisted of competing melting and recrystallization processes. This could be explained by the fact that unstable crystals melted first and more stable crystals melted at higher temperatures than the melting temperatures of these unstable crystals. Hence, the melting rate was faster than the crystallization rate, and the crystallization rate decreased at high temperatures, resulting in an endothermic peak.
Fig. 8 XRD patterns of (a) pure PBS, PBS/a-CNC-3.0, PBS/b-CNC-3.0, and PBS/c-CNC-3.0 composites and (b) PBS/b-CNC composites with different b-CNC contents. |
Fig. 9 SEM images of (a) PBS/a-CNC-3.0, (b) PBS/b-CNC-3.0, (c) PBS/c-CNC-3.0, (d) PBS/b-CNC-0.5, (e) PBS/b-CNC-1.0, (f) PBS/b-CNC-5.0. |
As Fig. 10(a) shows, E values of 316.5 ± 4.3, 373.3 ± 3.9, 434 ± 3.1, and 350 ± 3.8 MPa are obtained for PBS, PBS/a-CNC-3.0, PBS/b-CNC-3.0, and PBS/c-CNC-3.0, respectively. These values indicate that, consistent with an enhancement of the mechanical properties, the E of the PBS mixed with modified- CNC is higher than that of pure PBS.
The highest E was obtained for the PBS/b-CNC composite. This indicated that b-CNC and PBS were better mixed (than either a-CNC or c-CNC with PBS) and the dispersion of b-CNC in PBS was the best. The three-dimensional network structure was formed by mixing modified-CNC with PBS, thereby increasing the effect of b-CNC nanoparticles (added in only low volume fractions) on the mechanical properties of PBS. Correspondingly, the E of PBS/b-CNC composites was higher than that of pure PBS [see inset of Fig. 10(b)]. E values of 382.1 ± 4.0, 399.0 ± 11, 434 ± 3.1, 424.3 ± 3.7, and 347 ± 3.9 MPa were obtained for PBS/b-CNC-0.5, PBS/b-CNC-1.0, PBS/b-CNC-3.0 PBS/b-CNC-5.0, and PBS/CNC-7.0, respectively. The optimum mass fraction of b-CNC nanoparticles was found to be 3.0 wt%. In other words, b-CNC additions exceeding 3.0 wt% resulted in agglomeration of the nanoparticles, leading to poor interface compatibility of PBS with these particles and, in turn, a decrease in the tensile modulus. Moreover, the ε decreased gradually with increasing b-CNC content. For example, ε values of 364.8 ± 4.2, 354.4 ± 4.0, 341.7 ± 3.6, 334.6 ± 3.3, 309 ± 4.8, and 286 ± 3.5% were obtained for PBS, PBS/b-CNC-0.5, PBS/b-CNC-1.0, PBS/b-CNC-3.0, PBS/b-CNC-5.0, and PBS/b-CNC-7.0, respectively. The E value of PBS/b-CNC composites was higher than that of pure PBS (consistent with enhancement of the mechanical properties) and increased with increasing b-CNC contents. The E value of pure PBS (i.e., 316 ± 15 MPa) improved by up to 26.7% with the addition of 3.0 wt% b-CNC. Therefore, the tensile modulus of PBS was significantly enhanced, owing to the incorporation of modified-CNC, whereas the elongation at fracture decreased considerably.
This journal is © The Royal Society of Chemistry 2018 |